Adobe InDesign CS4 Guide

ADOBE® INDESIGN® CS4
ADOBE INDESIGN CS4 PRODUCTS
PROGRAMMING GUIDE
© 2008 Adobe Systems Incorporated. All rights reserved.
Adobe InDesign CS4 Products Programming Guide
If this guide is distributed with software that includes an end user agreement, this guide, as well as the software described in it, is furnished under license and may be
used or copied only in accordance with the terms of such license. Except as permitted by any such license, no part of this guide may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, electronic, mechanical, recording, or otherwise, without the prior written permission of Adobe Systems
Incorporated. Please note that the content in this guide is protected under copyright law even if it is not distributed with software that includes an end user license
agreement.
The content of this guide is furnished for informational use only, is subject to change without notice, and should not be construed as a commitment by Adobe Systems
Incorporated. Adobe Systems Incorporated assumes no responsibility or liability for any errors or inaccuracies that may appear in the informational content
contained in this guide.
Please remember that existing artwork or images that you may want to include in your project may be protected under copyright law. The unauthorized incorporation
of such material into your new work could be a violation of the rights of the copyright owner. Please be sure to obtain any permission required from the copyright
owner.
Any references to company names in sample templates are for demonstration purposes only and are not intended to refer to any actual organization.
Adobe, the Adobe logo, Acrobat, Bridge, Creative Suite, Illustrator, InCopy, InDesign, Photoshop, and Reader are either registered trademarks or trademarks of
Adobe Systems Incorporated in the United States and/or other countries. Microsoft and Windows are either registered trademarks or trademarks of Microsoft
Corporation in the United States and/or other countries. Apple and Mac OS are trademarks of Apple Computer, Incorporated, registered in the United States and
other countries. Java is a trademark of Sun Microsystems, Incorporated in the United States and other countries.All other trademarks are the property of their
respective owners.
Adobe Systems Incorporated, 345 Park Avenue, San Jose, California 95110, USA. Notice to U.S. Government End Users. The Software and Documentation are
“Commercial Items,” as that term is defined at 48 C.F.R. §2.101, consisting of “Commercial Computer Software” and “Commercial Computer Software
Documentation,” as such terms are used in 48 C.F.R. §12.212 or 48 C.F.R. §227.7202, as applicable. Consistent with 48 C.F.R. §12.212 or 48 C.F.R. §§227.7202-1
through 227.7202-4, as applicable, the Commercial Computer Software and Commercial Computer Software Documentation are being licensed to U.S. Government
end users (a) only as Commercial Items and (b) with only those rights as are granted to all other end users pursuant to the terms and conditions herein.
Unpublished-rights reserved under the copyright laws of the United States. Adobe Systems Incorporated, 345 Park Avenue, San Jose, CA 95110-2704, USA. For U.S.
Government End Users, Adobe agrees to comply with all applicable equal opportunity laws including, if appropriate, the provisions of Executive Order 11246, as
amended, Section 402 of the Vietnam Era Veterans Readjustment Assistance Act of 1974 (38 USC 4212), and Section 503 of the Rehabilitation Act of 1973, as
amended, and the regulations at 41 CFR Parts 60-1 through 60-60, 60-250, and 60-741. The affirmative action clause and regulations contained in the preceding
sentence shall be incorporated by reference.
Contents
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
For experienced InDesign developers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
For new InDesign developers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Persistent Data and Data Conversion . . . . . . . . . . . . . . . . . . . . . . . . 29
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Persistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Databases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Persistent objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Using persistent objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Implementing persistent objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
IPMStream methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Implementing a new stream . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Missing plug-ins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Warning levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Missing plug-in alert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Guidelines for handling a missing plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Data handling for missing plug-ins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Conversion of persistent data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
When to convert persistent data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Converting data with the conversion manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Converting data without the conversion manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
PluginVersion resource, format numbers, and their macros . . . . . . . . . . . . . . . . . . . . . . 53
Setting up resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Schemas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
SchemaList . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
DirectiveList . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Advanced schema topics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Arrays of values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
FieldArray . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Conditional-field inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3
Contents
Commands
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Command pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Databases and undoability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Command parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Command undoability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Command processing and the CmdUtils class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Command sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Command managers, databases, and undo support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
The command processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Scheduled commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Snapshots and interface implementation types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Command history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Merging changes with an existing step in the command history . . . . . . . . . . . . . . . . . . . 81
Undo and redo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Extension patterns involved in undo and redo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Notification within commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Error handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Protective shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Key client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Command facades and utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Command-processing APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Error string service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Persistent interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Persistent boss. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Snapshot interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Inval handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Snapshot view interface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4
Adobe InDesign CS4 Products Programming Guide
Contents
Notification
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101
Observer pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101
Responder pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .102
Observers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .102
Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103
Observers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
Message protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
Subject and observer types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105
Regular and lazy notification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106
Observers and undo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111
Relating observers to subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112
Document notification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113
Observers and the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113
Responders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115
Signal manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .116
Responder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117
Responders and the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117
Responders and global error state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .118
Ordering of responders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .118
Responders and undo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
Key client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .120
User-interface widget observer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .120
Model observer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .120
Selection observer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122
Document observer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122
Active context observer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122
Subject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124
Responder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124
Selection
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125
Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125
Selection format and target . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125
Design patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126
Selection architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .127
Abstract selection bosses and suites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130
Adobe InDesign CS4 Products Programming Guide
5
Contents
Concrete selection bosses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .131
Layout selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132
Table selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133
Text selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .134
Galley text selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .135
Story-editor text selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136
Note text selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137
XML selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138
Document defaults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .140
Application defaults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141
Integrator suites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142
CSB suites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142
Encapsulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143
Suites and the user interface: an example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143
Responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144
Basic client-code responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144
Selection-observer responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144
Custom-suite responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144
Custom suites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145
Selection extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146
Caches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146
Selection change notification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146
Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147
Communication with Integrator suite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147
Selection observers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .147
Selection-utility interface (ISelectionUtils) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148
Layout Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .149
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .150
Documents and the layout hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152
Parent and child objects and IHierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155
Spreads and pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157
Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .161
Layers in a basic document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .162
Layer options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .165
Navigating spread content using ISpread. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166
6
Adobe InDesign CS4 Products Programming Guide
Contents
Master spreads and master pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .168
Master spreads and master pages in a basic document . . . . . . . . . . . . . . . . . . . . . . . . .169
Master-page item overrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172
Basing one master page on another . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172
Page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173
Frames and paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173
Graphic page items. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177
Text page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177
Interactive page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .178
Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .178
Abstract page items and kPageItemBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179
Guides and grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180
Ruler guides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180
Margin and column guides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180
Document grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Baseline grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Snap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
Layout-related preferences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .182
Coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .183
Transformation matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .183
Pasteboard coordinate space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .184
Inner coordinate space and parent coordinate space . . . . . . . . . . . . . . . . . . . . . . . . . .185
Spread coordinate space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .188
Page coordinate space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .189
Page-item coordinate space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .190
Bounding box and IGeometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .190
Transformation and ITransform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .191
Measurement units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .192
Geometrical data types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .194
The layout presentation and view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .194
Layout presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .194
Layout view. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .196
Current spread and active layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .197
Layout-presentation and layout-view coordinate spaces . . . . . . . . . . . . . . . . . . . . . . . .198
Key client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .199
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .202
New-page-item responder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .202
Custom page item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203
Custom unit-of-measure service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203
Adobe InDesign CS4 Products Programming Guide
7
Contents
Commands that manipulate page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204
Page-item creation commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204
Page-item update commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .205
Page-item deletion commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .209
Graphics Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211
Path concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211
Paths data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213
Path operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .215
Graphic page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217
Graphic page-item types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217
Graphic page-item settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219
Graphic page-item data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .223
Graphic page-item examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .226
Graphics import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .238
Export to graphics file format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .243
Colors and swatches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246
Swatches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .250
Solid colors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .250
Gradients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251
Swatch lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251
Inks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251
Color management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .252
ICC profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Color-management workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Data model for color management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Graphic attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256
Graphic-attribute data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256
Representation of graphic attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .258
Graphic styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .258
Graphic state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259
Mapping graphic attributes between domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .260
Rendering attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .261
Color-rendering attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .261
Gradient attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262
8
Adobe InDesign CS4 Products Programming Guide
Contents
Stroke effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262
Path stroker. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262
Path corners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263
Path-end strokers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263
Transparency effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .263
Flattening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .272
Transparency data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .272
Data model for drawing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276
Presentation views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276
Graphics context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .277
Viewport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .277
Dynamics of drawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .278
Drawing the layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .278
Drawing page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .281
Drawing in user-interface widget windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .283
Offscreen drawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284
Client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .285
Path-related client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .285
Graphic page-item client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .286
Key color-related client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .287
Graphic-attribute client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .288
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .289
Custom graphic attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .289
Custom path-stroker effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .289
Custom corner effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .290
Custom path-end effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .290
Custom page-item adornments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .291
Custom drawing-event handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .294
Swatch-list state. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .296
Initial state of swatch list and ink list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .296
State of swatch list and ink list after adding a custom stop color . . . . . . . . . . . . . . . . . . .297
Swatch list and ink list after adding a gradient swatch. . . . . . . . . . . . . . . . . . . . . . . . . .299
Swatch list and ink list after applying an unnamed color to an object . . . . . . . . . . . . . . . .300
Color spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .300
Catalog of graphic attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .301
Mappings between attribute domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .306
Spread-drawing sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .307
Adobe InDesign CS4 Products Programming Guide
9
Contents
Controlling the settings in a graphics port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .309
Drawing sequence for a page item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .310
Text Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .313
Text content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .315
Stories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .315
Text formatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .324
Class associations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330
Text presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331
Text layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331
Text frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331
Frame list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .333
Threading and text frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .336
Parcels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .339
Span . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342
Text frames and the wax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .343
Text-frame options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .345
Text-frame geometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .345
Text Inset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .346
Text wrap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .347
Text on a path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .350
The wax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .354
Wax strand, wax line, and wax run . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .354
Examples of the wax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355
Text adornments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .358
Text composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .364
Phases of text composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .366
Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .367
Recomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .369
Wax strand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .370
Paragraph composers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .371
Shuffling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .371
Vertical justification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .371
Background composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .371
Recomposition transactional model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .372
Recomposition notification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .372
Implementation notes for paragraph composers . . . . . . . . . . . . . . . . . . . . . . . . . . . . .372
10
Adobe InDesign CS4 Products Programming Guide
Contents
Fonts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .389
Font-subsystem architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .390
Fonts within the document. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .392
Composite fonts and international-font issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .395
Tables
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .397
Table structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .397
Design and architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .402
Table model versus text model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .402
Table data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .404
Cell data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406
Table attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .407
Table and cell styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .409
Formatting tables, cells, and table text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .413
Essential APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .414
Table commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .414
ITableSuite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .414
ITableStyleSuite and ITableStylesFacade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .415
ICellStyleSuite and ICellStylesFacade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .415
Printing
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .417
Printing is simply drawing to the printer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .417
Control can be shared . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .417
Inks and colors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .417
Overprinting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .418
Trapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .418
Color management and proofing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .418
Preflight and packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .418
Exporting to EPS and PDF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .419
Printing data model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .419
Print settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .419
Print preset styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .420
Trap styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .420
Utility APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .420
The print action sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .421
Common print interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .422
Adobe InDesign CS4 Products Programming Guide
11
Contents
Print user interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .423
Print dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .423
Print Presets dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .435
Extending the Print dialog box or the Print Presets selectable dialog box . . . . . . . . . . . . .435
Printing extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .435
Print-setup provider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .435
Print-insert-PostScript proc provider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .436
Print-data helper-strategy provider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .437
Draw-event handlers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .438
Printing solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .438
Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .438
Working with print-preset styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .439
Working with trap styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .441
Participating in the print process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .443
Bosses that aggregate IPrintData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .448
Print-action and supporting commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .448
Japanese page-mark files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .449
Exporting to EPS and PDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .450
Exporting to EPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .450
Exporting to PDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .453
PDF Import and Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
PDF import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .455
PDF export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .459
InDesign/InCopy document export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .459
InDesign book export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .468
Selected page-items export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .469
PDF-style import and export. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .470
Adding, deleting, and editing styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .471
Frequently asked questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .472
How does the PDF export provider determine whether it should start the viewer
after the export? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .472
How do I set the PDF clipboard setting as seen in the File Handling preferences? . . . . . . . .472
How do I control which layer of a document should be exported? . . . . . . . . . . . . . . . . . .472
How do I make the two-page spreads in my document export as two separate PDF pages? .473
Why does kPDFExportCmdBoss give me an assert after the command is processed
(ASSERT 'db != nil' in PDFExportController.cpp)? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .473
How do I set up line ranges for output in InCopy Galley or Story mode? . . . . . . . . . . . . . .473
Is it possible to export only selected text from an InDesign document? . . . . . . . . . . . . . .473
12
Adobe InDesign CS4 Products Programming Guide
Contents
Implementing Preflight Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .475
About preflight in InDesign CS4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .475
About rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .475
Rule IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .476
Rule service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .476
IPreflightRuleService example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .477
Rule bosses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .478
IPreflightRuleVisitor interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .479
IPreflightRuleVisitor method examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .480
IPreflightRuleVisitor::GetClassesToVisit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .480
IPreflightRuleVisitor::Visit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .481
IPreflightRuleVisitor::AggregateResults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .483
IPreflightRuleVisitor::UpdateRuleData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .490
IPreflightRuleVisitor::ValidateRuleData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .490
More on specific objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .491
Native, UID-based objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .491
Artwork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .492
Text runs and ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .494
Tables, rows, columns, and cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .495
XML Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .497
XML-based workflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .497
Using XML with InDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .498
Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .498
XML features at a glance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499
XML extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499
Tagging in tables and inline graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499
Throw away unmatched existing (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499
Throw away unmatched incoming (left). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Importing repeating elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Importing CALS table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Support for DOM core level 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Support table- and cell-styles import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Support XML-rules processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Snippets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .500
Adobe InDesign CS4 Products Programming Guide
13
Contents
The user interface for XML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .501
Structure view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .501
Tags in layout view and story view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .501
Tags panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .502
Mapping between tags and styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .503
Validation window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .504
Other. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .504
XML model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .505
Native document model and logical structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .505
Elements and attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .505
Content items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .506
References to elements and content items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .507
Document element and root element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .508
Backing store. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .509
Persistence and the backing store . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .510
Importing XML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .511
Import architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .512
Importing a minimal XML file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .515
Unplaced content versus placed content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .518
XML template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .519
Matching against an XML template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .519
Importing repeating elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .519
Throwing away unmatched existing elements on import (delete unmatched right). . . . . . .520
Throwing away unmatched incoming elements on XML import . . . . . . . . . . . . . . . . . . .522
Attribute-style mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .522
Creating links on XML import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .523
Sparse import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .523
Importing a CALS table as an InDesign table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .524
Support table and cell styles when importing an InDesign table . . . . . . . . . . . . . . . . . . .524
Exporting XML. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .524
Export architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .525
Document order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .526
Tagged graphic placeholder, exported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .527
Tagged text range, exported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .528
Tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .529
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .530
Tag-to-style mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .531
Style-to-tag mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .532
14
Adobe InDesign CS4 Products Programming Guide
Contents
Elements and content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .534
Tagged graphic placeholder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .534
Tagged images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .536
Tagged stories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .537
Tagged text ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .538
Tagged inline graphics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .542
Tagged tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .543
DTD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .545
Processing instructions and comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .547
Structural (container) elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .549
XML-related preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .549
Workspace-level XML preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .549
Service-level XML preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .551
Key client API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .552
XML suites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .552
Command facades and utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .553
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .553
XML acquirer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .553
XML transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .554
XMl-import matchmaker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .555
Post-import responder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .555
Custom suite for the structure view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .556
SAX-content handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .556
SAX DOM serializer handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .557
Custom-tag service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .557
SAX-entity resolver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .557
XML-export handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .558
Commands and notification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .558
Backing store and notification of changes in logical structure. . . . . . . . . . . . . . . . . . . . .558
Entities supported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .559
Assets from XSLT example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .560
Limitations of the InDesign XML architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .561
Adobe InDesign CS4 Products Programming Guide
15
Contents
Scriptable Plug-in Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . 563
Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .563
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .565
Scripting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .565
Benefits of making a plug-in scriptable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .565
Scripting and IDML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .566
Making a plug-in scriptable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .566
Tools for making a plug-in scriptable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .567
Scripting architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .567
Scripting DOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .568
Versioning the scripting DOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .569
Scripting plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .570
Script interaction with the scripting DOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .571
Scripting process overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .572
Script managers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .573
Scriptable plug-ins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .573
Scripting resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .574
Script providers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .574
Scriptable boss classes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .574
How to make your plug-in scriptable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .575
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .575
Defining IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .576
Adding a new property to an existing script object . . . . . . . . . . . . . . . . . . . . . . . . . . .576
Adding a new event to an existing script object . . . . . . . . . . . . . . . . . . . . . . . . . . . . .576
Adding a new script object to make preferences scriptable . . . . . . . . . . . . . . . . . . . . . .577
Adding a new singleton script object . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .578
Adding a new script object to make a boss with a UID scriptable. . . . . . . . . . . . . . . . . . .578
Adding a new script object to make a boss with no UID scriptable. . . . . . . . . . . . . . . . . .579
Adding a new script object to make a C++ object with no boss scriptable . . . . . . . . . . . . .580
Adding a new script object to make a panel scriptable . . . . . . . . . . . . . . . . . . . . . . . . .581
Adding an error-string service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .582
Handling multiple concurrent requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .582
Reviewing scripting resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .583
Running versioned scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .586
Supporting IDML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .588
Maintaining IDML forward and backward compatibility . . . . . . . . . . . . . . . . . . . . . . . .589
Verifying your plug-in’s data is round-tripped through IDML . . . . . . . . . . . . . . . . . . . . .590
Tips for debugging the scripting architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .591
16
Adobe InDesign CS4 Products Programming Guide
Contents
Scripting resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .591
VersionedScriptElementInfo resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .592
Object element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .593
Event element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .595
Property element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .597
Struct element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .601
TypeDef element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .602
Enum element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .603
Enumerator element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .604
Metadata element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .604
Provider element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .606
Suite element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .609
ScriptElementIDs, ScriptIDs, names, descriptions, and GUIDs . . . . . . . . . . . . . . . . . . . . .610
ScriptID/name registration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .612
Scripting data types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .615
Script-object inheritance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .619
Overloading an existing event or property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .621
Versioning of scripting resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .621
Client-specific scripting resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .632
Elements that are not applicable to a particular object . . . . . . . . . . . . . . . . . . . . . . . . .638
Key scripting APIs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .639
Scripting DOM reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .641
Scripting DOM versioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .641
Dumping the scripting DOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .642
Snippet Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
Conceptual overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .643
Snippets as self-contained assets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .643
Snippet types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .644
Features that depend on snippets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .644
User interface for snippets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .644
Drag and drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .644
Asset library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .645
Export snippet from selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .645
Snippet model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .645
INX, IDML, and snippets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .645
Boss DOM overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .646
Scripting DOM overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .647
From boss DOM to scripting DOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .647
Adobe InDesign CS4 Products Programming Guide
17
Contents
Snippet types and policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .648
Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .648
Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .650
Snippet examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .652
Filled-rectangle snippet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .652
Image-item snippet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .657
XML element snippet, tagged placeholder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .660
Client API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .663
Suites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .663
Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .663
Extension patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .664
Adding persistent data to snippet files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .664
Frequently asked questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .664
What is a snippet, and how do I create one? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .664
What happens if I export a snippet of a placed image? . . . . . . . . . . . . . . . . . . . . . . . . .664
What features are based on snippets? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .665
How accurately is data round-tripped through snippets?. . . . . . . . . . . . . . . . . . . . . . . .665
When do I have to care about snippets?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .665
Can I export spreads or pages as snippets?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .665
Should we generate snippet files from scratch? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .665
Can I import a snippet directly into a library? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .666
Can I import a snippet into the scrap database? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .666
Can we add our own new snippet types? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .666
Shared Application Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . 667
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .667
Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .667
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .667
How it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .668
ISnippetExport. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .668
IAppPrefsExportDelegate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .669
ISnippetImport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .669
IAppPrefsImportOptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .669
IAppPrefsImportDelegate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .670
18
Adobe InDesign CS4 Products Programming Guide
Contents
Working with snippet APIs: frequently asked questions . . . . . . . . . . . . . . . . . . . . . . . . . . .670
How do I create streams for reading and writing snippets?. . . . . . . . . . . . . . . . . . . . . . .670
How do I limit my export to those items in the preference panel? . . . . . . . . . . . . . . . . . .670
How do I export all text styles, object styles, XML tags, or swatches in the
application workspace? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .671
How do I import a snippet into the application? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .671
How do I control whether existing objects like paragraph styles are replaced or
deleted on import. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .671
How do I determine the correct ScriptID to use for a preference I’m trying to include
or exclude? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .671
How do I know which list element types will be exported by default?. . . . . . . . . . . . . . . .671
User-Interface Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .673
Key concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .674
Design objectives for user-interface API. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .674
Idioms and naming conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .674
Abstractions and re-use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .675
Widgets versus platform controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .676
Commands, model plug-ins, and user-interface plug-ins . . . . . . . . . . . . . . . . . . . . . . .676
Suites and the user interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .677
Finding widgets in the API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .678
Notifications about control events or changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .678
Sample user interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .679
Type binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .680
Factorization of the user-interface model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .683
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .683
Control views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .684
Control data models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .684
Event handlers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .685
Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .685
Relevant design patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .686
Observer pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .686
How event handlers implement controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .687
The role of MVC in the user-interface model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .688
Chain of responsibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .689
Facade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .689
Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .690
Widget-observer pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .690
Persistence and widgets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .691
Adobe InDesign CS4 Products Programming Guide
19
Contents
Resource roadmap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .692
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .692
OpenDoc framework (ODF) resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .693
Top-level framework resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .694
Localizing framework resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .694
Resource compilers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .695
Customizing a widget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .696
Advanced event handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .696
Writing a proxy event handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .696
Watching events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .697
Key abstractions in the API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .697
Suppressed User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 699
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .699
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .699
XML-based implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .700
XML file format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .701
SuppressedWidget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .702
SuppressedAction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .702
SuppressedMenu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .703
SuppressedDragDrop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .703
SuppressPlatformDialogControl. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .704
SuppressedUI tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .704
SuppressedUI tool limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .706
Working with the ISuppressedUI API. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .706
Other user-interface “suppression” mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .707
Using Adobe File Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .709
Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .709
Adobe File Library architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .711
Adobe File Library classes and utilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .711
Common file API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .713
File API specific to InDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .714
Debugging IDFile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .716
Porting guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .716
Creative Suite 2 porting concerns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .716
Creative Suite 3 porting concerns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .716
20
Adobe InDesign CS4 Products Programming Guide
Contents
Frequently asked questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .717
Why should I use Adobe file library? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .717
Does Adobe file library support cross-platform path conversion? . . . . . . . . . . . . . . . . . .717
Should I still use the ICoreFileName interface? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .717
Why was the GetSysFile method renamed GetIDFile in SDKFileHelper? . . . . . . . . . . . . . . .717
How do I navigate between IDFile and IDPath? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .717
What are the differences between a file and a directory? . . . . . . . . . . . . . . . . . . . . . . . .718
What are the relationships between IDPath and IDFile? . . . . . . . . . . . . . . . . . . . . . . . . .718
Why should IDFile not be treated as PMString? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .718
How do an invalid path and a nonexistent path differ? . . . . . . . . . . . . . . . . . . . . . . . . .718
Can I construct an AString from PMString? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .718
How can I convert a relative path to an absolute path? . . . . . . . . . . . . . . . . . . . . . . . . .719
Performance Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 721
Use profiling tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .721
Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .721
Mac OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .721
Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .722
Commands should operate on lists of inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .722
Notify on the document subject instead of the page item object . . . . . . . . . . . . . . . . . .722
Mark commands that do not require undo support . . . . . . . . . . . . . . . . . . . . . . . . . . .723
Observers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .723
Attach to documents, not page items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .723
Do work only when the command is done . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .723
Do not update the user interface from an observer . . . . . . . . . . . . . . . . . . . . . . . . . . .724
Watch for lazy notification whenever possible. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .724
File input/output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .724
Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .724
Use memory caches for items accessed often . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .724
Avoid allocating too much memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .725
Idle tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .725
Honor the RunTask flags. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .725
Do a small amount of work during each RunTask call . . . . . . . . . . . . . . . . . . . . . . . . . .725
Adobe InDesign CS4 Products Programming Guide
21
Contents
Tools
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727
Key concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .727
The toolbox and the layout view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .727
Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .728
Cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .728
Tool tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .729
Trackers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .729
Tracker factory, tracking, and event handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .729
Beyond the toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .730
Drawing and sprites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .730
Documents, page items, and commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .731
Line-tool use scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .731
Trackers with multiple behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .734
Tool manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .735
Toolbox utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .735
Tool type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .735
Custom tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .736
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .736
Class diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .736
Partial implementation classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .737
Default implementations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .737
ToolDef ODFRez type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .738
Icons and cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .738
InDesign trackers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .739
Working with tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Catching a mouse click or mouse drag on a document . . . . . . . . . . . . . . . . . . . . . . . . .740
Implementing a custom tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Displaying a Tool Options dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Finding the spread nearest the mouse position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Changing spreads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Performing a page-item hit test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .740
Setting or getting the active tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .741
Observing when the active tool changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .741
Changing the toolbox appearance from normal to skinny . . . . . . . . . . . . . . . . . . . . . . .741
Using default implementations for trackers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .741
Suppressing the application's default tracker for a custom toolbox . . . . . . . . . . . . . . . . .741
Tool-category information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .742
Default implementations of tool-related interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .745
Tracker listings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .746
22
Adobe InDesign CS4 Products Programming Guide
Contents
Diagnostics
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 751
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .751
Using the diagnostics plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .751
Diagnostics menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .752
Diagnostics > Command menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .752
Diagnostics > Document Structure menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .753
Diagnostics > INX DTD menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .754
Diagnostics > Object Model menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .755
Diagnostics > Scripting DOM menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .756
Running the Diagnostics plug-in in indesign/incopy with a script . . . . . . . . . . . . . . . . . .756
Running the Diagnostics plug-in in InDesign Server on Windows with a script . . . . . . . . . .757
Running the Diagnostics plug-in in InDesign Server on Mac OS with a script . . . . . . . . . . .757
Frequently asked questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .758
Where is the Diagnostics panel?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .758
What is trace? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .758
Why don’t I get trace messages? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .758
InCopy: Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 759
About InCopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .759
Developing for InCopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .759
Using the combined InDesign/InCopy SDK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .759
The InCopy API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .760
Compiler settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .760
Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .760
Synchronization of design and architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .760
File relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .761
Stories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .762
Page geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .762
Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .762
The document model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .763
User-interface differences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .763
Checking the feature set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .763
Workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .764
Design and architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .764
InCopyBridge plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .766
InCopyBridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .766
InCopyBridgeUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .767
Adobe InDesign CS4 Products Programming Guide
23
Contents
InCopy: Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 769
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .769
End-user requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .769
Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .771
Data model for notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .773
Essential APIs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .779
Useful commands and associated notification protocols . . . . . . . . . . . . . . . . . . . . . . . .783
Working with notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .786
Adding a note at the current insertion-point position . . . . . . . . . . . . . . . . . . . . . . . . . .786
Inserting text into a note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .787
Converting text to a new note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .787
Converting note content to text. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .787
Navigating among notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .788
Splitting a note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .788
Expanding and collapsing notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .788
Selecting a note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .789
Getting kNoteDataBoss, given a text index whose position is anchored to note . . . . . . . . .789
Deleting notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .790
Changing notes-palette content to reflect particular note data . . . . . . . . . . . . . . . . . . . .790
Checking note spelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .790
Observing a note that is being modified . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .790
Using notes in InDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .790
InCopy: Track Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .791
User interface for Track Changes feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .791
Data model for Track Changes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .794
Redline strand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .794
Tracking text insertion and deletion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .795
Example: Track Changes in action. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .799
Track Changes preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .809
Key client APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .809
Track Changes utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .809
Suite interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .809
RedlineIterator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .811
kDeletedTextBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .812
24
Adobe InDesign CS4 Products Programming Guide
Contents
Useful commands and associated notification protocols . . . . . . . . . . . . . . . . . . . . . . . . . . .814
kSetRedlineTrackingCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .814
kSetTrackChangesPrefsCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .814
kActivateRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .816
kDeactivateRedlineCmdBoss. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .816
kRejectAllRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .816
kRejectRangeRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .817
kRejectRedlineCmdBoss. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .817
kAcceptAllRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .818
kAcceptRangeRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .819
kAcceptRedlineCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .819
kMoveRedlineChangeCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .820
kRedlinePreserveDeletionCmdBoss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .821
Working with Track Changes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .821
Navigating tracked changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .821
Accepting and rejecting tracked changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .821
Understanding multiple change records in one location . . . . . . . . . . . . . . . . . . . . . . . .822
Avoid insignificant tracked changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .822
Undoing accepted deleted text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .822
Determining whether a location in a story is in deleted text . . . . . . . . . . . . . . . . . . . . . .822
Determining whether a primary story-thread location is at a deleted-text anchor . . . . . . . .822
Getting kDeletedTextBoss, given a text index having deleted text . . . . . . . . . . . . . . . . . .823
Removing deleted text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .823
Moving a change record from one story to another . . . . . . . . . . . . . . . . . . . . . . . . . . .823
Maintaining kRedlineStrandBoss text-run information . . . . . . . . . . . . . . . . . . . . . . . . .823
InCopy: Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 825
Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .825
Assignment workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .826
Assignment-export options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .826
Assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .827
Assignment data model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .829
Assignment hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .829
Object structure of an assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .830
Moving assignment content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .832
Assignment files and links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .834
Assignment files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .834
Comparing assignment files to InDesign and InCopy files . . . . . . . . . . . . . . . . . . . . . . .835
Assignment-file format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .835
Adobe InDesign CS4 Products Programming Guide
25
Contents
The assignment API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .838
IAssignmentMgr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .838
IAssignedDocument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .838
IAssignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .838
IAssignedStory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .838
IAssignmentSelectionSuite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .839
IAssignmentUtils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .839
IAssignmentUIUtils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .839
IAssignmentPreferences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .839
Common commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .839
Glossary
26
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 841
Adobe InDesign CS4 Products Programming Guide
Introduction
For experienced InDesign developers
Introduction
This guide provides detailed information on the Adobe® InDesign® CS4 plug-in architecture.
This C++-based SDK can be used for creating plug-ins compatible with the CS4 versions of
InDesign, InDesign Server, and Adobe InCopy®.
This guide contains the most detailed information on plug-in development for InDesign products. It is not designed to be a starting point. It picks up where Learning Adobe InDesign CS4
Plug-in Development leaves off, and it is more commonly used to understand particular subjects deeply.
For experienced InDesign developers
If you are an experienced InDesign plug-in developer, we recommend starting with Adobe
InDesign CS4 Porting Guide.
For new InDesign developers
If you are new to InDesign development, we recommend approaching the documentation as
follows:
1. Getting Started With the Adobe InDesign CS4 Products SDK provides an overview of the
SDK, as well as a tutorial that takes you through the tools and steps to build your first plugin.
2. Learning Adobe InDesign CS4 Plug-in Development introduces the most common programming constructs for InDesign development. This includes an introduction to the InDesign
object model and basic information on user-interface options, scripting, localization, and
best practices for structuring your plug-in.
3. The SDK itself includes several sample projects. All samples are described in the “Samples”
section of the API reference. This is a great opportunity to find sample code that does something similar to what you want to do, and study it.
4. Adobe InDesign CS4 Solutions provides step-by-step instructions (or “recipes”) for accomplishing various tasks. If your particular task is covered by the Solutions guide, reading it
can save you a lot of time.
5. This manual provides the most complete, in-depth information on plug-in development for
InDesign CS4 products.
Introduction
27
Introduction
For new InDesign developers
28
Persistent Data and Data Conversion
Concepts
Persistent Data and Data Conversion
This chapter describes how an application stores and refers to persistent data as objects and
streams. The chapter also provides background information and implementation guidelines
related to data conversion.
This chapter has the following objectives:
z
Describe how Adobe InDesign® stores data.
z
Explain how an object is made persistent.
z
Show how to refer to a persistent object.
z
Define a stream and explain how to create a new stream.
z
Identify when data conversion is used.
z
Describe the types of conversion providers.
z
Show how conversion providers are defined.
z
Describe schemas and their use.
For common procedures and troubleshooting related to converting persistent data, see the
“Versioning Persistent Data” chapter of Adobe InDesign CS4 Solutions.
Concepts
Persistence
A persistent object can be removed from main memory and returned again, unchanged. A persistent object can be part of a document (like a spread or page item) or part of the application
(like a dialog box, menu item, or default setting). Persistent objects can be stored in a database,
to last beyond the end of a session. Non-persistent objects last only until memory is freed,
when the destructor for the object is called. Only persistent objects are stored in databases.
Databases
The application uses lightweight databases to store persistent objects. The host creates a database for each document created. The host also creates a database for the clipboard storage
space, the object model information (the InDesign SavedData or Adobe InCopy® SavedData
file), and the workspace information (the InDesign Defaults or InCopy Defaults file).
Each document is contained in its own database. Each persistent object in the database has a
UID, a ClassID, and a stream of persistent data.
The stream with the persistent object data contains a series of records that correspond to the
persistent object. The object’s UID is stored with the object, as a key for the record. The vari-
Persistent Data and Data Conversion
29
Persistent Data and Data Conversion
Persistent objects
able-length records have one segment for each persistent interface. Every segment has the same
structure:
ImplementationID tag
int32 length
<data>
Figure 1 is a conceptual diagram of this structure. The figure is a conceptual diagram of the
database contents, and does not represent the actual contents of any database. The format and
content of the data are determined by the implementation of the interface. See “Reading and
writing persistent data” on page 34.
FIGURE 1
Conceptual diagram of an object’s data
k<Foo>Boss object persistent data record
UID
ImplID
Length
Data
ImplID
Length
Data
ImplID
Length
For each object, the ClassID value is stored in the table, and the ImplementationID values are
stored in the stream, but the InterfaceID value is not stored with either. Adding an existing
implementation (i.e., an implementation supplied by the SDK) to an existing class can cause
problems if another software developer adds the same implementation to the class: one of the
two plug-ins will fail on start-up. To avoid this collision, create a new implementation for any
persistent interface to be added to a class, using ImplementationAlias. For an example, see
<SDK>/source/sdksamples/dynamicpanel/DynPn.fr.
Persistent objects
This section discusses how persistent objects are created, deleted, and manipulated.
The application stores persistent objects in a database, and each object has a unique identifier
within the database. The methods for creating, instantiating, and using a persistent object are
different from the methods for non-persistent objects.
Using persistent objects
When a persistent object is created, its record exists in memory. Certain events, like a user’s
request to save a document, trigger the writing of the record to storage. (See Figure 2) A count
is maintained of the number of references to each persistent object. Any object with a reference
count greater than zero remains active in memory. If the reference count for a persistent object
reaches zero, the object may be asked to write itself to its database and be moved to the instance
cache, which makes the object a candidate for deletion from memory. Events that require
access to the object, like drawing or manipulating it, trigger the host to read the object back
30
Persistent Data and Data Conversion
Persistent objects
into memory. Because individual objects can be saved and loaded, the host does not have to
load the entire document into memory at once.
FIGURE 2
Creating and storing persistent objects
Host
Main memory
Storage
NewUID
Instantiate
(no storage)
Dirty
ReadWrite
Reference count
reaches zero
Instantiate
(with storage)
Creating a persistent object
Creating a new instance of a persistent object differs from creating a non-persistent object,
because it requires a unique identifier to associate the object with its record in the database. For
the CreateObject and CreateObject2 methods used to create persistent objects, see
<SDK>/source/public/includes/CreateObject.h.
For examples of the creation of persistent objects, see the CreateWidgetForNode method in
<SDK>/source/sdksamples/paneltreeview/PnlTrvTVWidgetMgr.cpp or the StartupSomePalette method in <SDK>/source/sdksamples/dynamicpanel/DynPnPanelManager.cpp.
You also can call IDataBase::NewUID method to create a new UID in the database. It adds an
entry to the database relating the UID to the class of the object being created; however, the
object is not instantiated yet, and no other information about it exists in the database.
Instantiating a persistent object
Before you instantiate a new object, use the InterfacePtr template to retrieve one of the object’s
interfaces. Pass the returned InterfacePtr to the IDatabase::Instantiate method, which calls the
database to instantiate the object.
There is only one object in memory for any UID and database. This method checks whether
the object already is in memory and, if so, returns a reference to that object. If the object is
Persistent Data and Data Conversion
31
Persistent Data and Data Conversion
Persistent objects
marked as changed, it is written to the database. For more information, see “Reading and writing persistent data” on page 34. If the object is not in memory, the method checks for previously stored data in the database. If data is found, the method instantiates the object from this
data. Otherwise, the method instantiates the object from the object’s constructor. Each implementation has a constructor, so the boss and all its implementations are instantiated.
An object is stored to the database only if it is changed, making the default object data invalid.
A single object could exist and be used in several user sessions without ever being written to
the database. A persistent object whose data is not stored in the database is constructed from
the object’s constructor.
Whether it instantiates a new object or returns a reference to an object previously instantiated,
IDatabase::Instantiate increments the reference count on the object and returns an IPMUnknown* to the requested interface, if the interface is available.
Using commands and wrappers
There are commands for creating new objects of many of the existing classes (e.g., kNewDocCmdBoss, kNewPageItemCmdBoss, kNewStoryCmdBoss, and kNewUIDCmdBoss). When you
need to create an object, first look for a suite interface, utility interface, or facade interface to
make creating your object safe and easy. If no such interface exists, use a command if one is
available, rather than creating the object with general functions.
Using a command to create an object protects the database. Commands are transaction based;
if you use a command when the application already is in an error state, the command performs
a protective shut-down, which quits the application rather than permitting a potentially corrupting change to be made to the document. Commands also provide notification for changes
to the model, allowing observers to be updated when the change is made, including changes
made with undo and redo operations.
When you implement a new type of persistent object, also implement a command to create
objects of that type, using methods outlined in “Implementing persistent objects” on page 33.
Types of references to objects
There are four types of reference to a persistent object: UID, UIDRef, InterfacePtr, and
UIDList. Each type of reference serves a different purpose. Understanding these reference
types makes working with persistent objects easier.
32
z
UID is the type used for a unique identifier within the scope of a database. The UID value
by itself is not sufficient to identify the object outside the scope of the database. Like a
record number, a UID value has meaning only within a given database. UID values are useful for storing, passing, and otherwise referring to boss objects, because UID values have no
run-time dependencies, and there is an instance cache ensuring fast access to the instantiated objects. kInvalidUID is a value used to identify a UID that does not point at a valid
object. Any time a UID is retrieved and needs to be tested to see if it points at a valid object,
the UID should be compared to kInvalidUID.
z
A UIDRef object contains two pieces of information: a pointer to a database and the UID of
an object within this database. A UIDRef is useful for referring to objects, because it identifies both the database and the object. Using a UIDRef object is a common means of referring to a persistent object, especially when the persistent object is to be passed around or
Persistent Data and Data Conversion
Persistent objects
stored, since a UIDRef does not require the referenced object to be instantiated. A UIDRef
object cannot itself be persistent data, because it has a run-time dependency, the database
pointer. An empty or invalid UIDRef object has kInvalidUID as its UID and a nil pointer as
its database pointer.
z
A InterfacePtr object contains a pointer to an interface (on any type of boss object) and
identify an instantiated object in main memory. While an InterfacePtr object on the boss is
necessary for working with the boss, it should not be used to track a reference to a persistent
object, because this forces the object to stay in memory. In many cases, a nil pointer
returned from InterfacePtr does not indicate an error state but simply means the requested
interface does not exist on the specified boss.
z
A UIDList object contains a list of UIDs and a single pointer to a database. This means all
objects identified by a UIDList must be within the same database. A UIDList is a class
object and should be used any time a list of objects is needed for a selection or a command.
Destroying an object
There are commands for deleting persistent objects. Use the command rather than calling the
DeleteUID method directly, to be sure of cleaning up all references to the object or item references contained in the object.
When you implement a command to delete a persistent object, after you remove the references
to the object, use DeleteUID to delete the object from the database, as follows:
IDataBase::DBResultCode dbResult = database->DeleteUID(uid);
Implementing persistent objects
To make a boss object persistent, add the IPMPersist interface. Any boss with this interface is
persistent and, when an object of the boss is created, it is assigned a UID. Even though the
object has a UID, and the UID has an entry in the (ClassID, UID) pairings in a database, the
object does not automatically store data. It is up to the interface to store the data it needs. To
implement this, add the ReadWrite method to the interface (see “Reading and writing persistent data” on page 34), and make sure the PreDirty method is called before information is
changed (see “Marking changed data” on page 34).
NOTE:
If you add an interface to an existing persistent boss, the interface also may be made
persistent. If so, you must obey the following implementation rules for it.
Adding the IPMPersist interface to a boss
All instances of IPMPersist must use the kPMPersistImpl implementation. For an example, see
the kPstLstDataBoss boss class definition in <SDK>/source/sdksamples/persistentlist/PstLst.fr
Creating an interface factory for a persistent interface
For a persistent implementation, use the CREATE_PERSIST_PMINTERFACE macro.
Persistent Data and Data Conversion
33
Persistent Data and Data Conversion
Streams
Reading and writing persistent data
To store data, your interface must support the ReadWrite method. This method does the actual
reading and writing of persistent data in the database. The method takes a stream argument
containing the data to be transferred. Read and write stream methods are generalized, so one
ReadWrite method handles transfers in both directions. For example, XferBool reads a boolean
value for a read stream and writes a boolean value for a write stream. For an example, see the
BPIDataPersist::ReadWrite method in <SDK>/source/sdksamples/basicpersistinterface/BPIDataPersist.cpp.
Marking changed data
When data changes for a persistent object that resides in memory, there is a difference between
the current version of the object and the object as it exists in the database’s storage. When this
happens, the object in memory is said to be dirty, meaning it does not match the version in
storage. Before a persistent object is modified, you must call the PreDirty method to mark the
object as being changed, so it is written to the database. For an example, see the BPIDataPersist::Set method in <SDK>/source/sdksamples/basicpersistinterface/BPIDataPersist.cpp.
The PreDirty method called from within BPIDataPersist::Set is implementation-independent,
so you can rely on the version provided by HELPER_METHODS macros defined in HelperInterface.h (DECLARE_HELPER_METHODS, DEFINE_HELPER_METHODS, and
HELPER_METHODS_INIT).
Streams
This section discusses how streams are used to move information into and out of a document.
Streams are used by persistent objects to store their information to a database. Streams also are
used by the host application, to move data like placed images, information copied to the clipboard, and objects stored in the database. IPMStream is the public interface to streams. Implementations of IPMStream typically use the IXferBytes interface to move data.
Stream utility methods (in StreamUtil.h) are helpers for creating all the common types of
streams used to move information within or between InDesign databases. The stream utility
methods and general read, write, and copy methods are needed any time you work with a
stream.
IPMStream methods
IPMStream is a generalized class for both reading and writing streams. Any particular stream
implementation is either a reading stream or a writing stream, and the type of stream can be
determined with the IPMStream::IsReading and IPMStream::IsWriting methods.
Any persistent implementation has a ReadWrite method, which uses a set of data-transferring
methods on the stream to read and write its data. (See “Reading and writing persistent data” on
page 34.) The IPMStream methods starting with the Xfer prefix are used for transferring the
34
Persistent Data and Data Conversion
Streams
data type identified in the method name. For example, XferByte transfers a byte, XferInt16
transfers a 16-bit integer, XferBool transfers a boolean value, and so on. All transferring methods are overloaded, so they can take a single item or an array of items. (The XferByte(uchar,
int32) version typically is used for buffers.) Streams also handle byte swapping, if required. If
swapping is not set (SetSwapping(bool16)), the default is to not do byte-order swapping.
Additional IPMStream methods, XferObject and XferReference, transfer boss objects and references to objects. XferObject transfers an owned object, and XferReference transfers a reference to an object not owned by the object using the stream. To decide which method to use,
think about what should happen to the object if the owning object were deleted. If the object
should still be available (as, for example, the color a page item refers to), use XferReference. If
the item is owned by the object and should be deleted with the owner (as, for example, the page
a document refers to), use XferObject.
Implementing a new stream
If you must read from or write to a location the host application does not recognize, you must
create a new type of stream. For example, you might need to create a new stream type to import
and export files stored on an FTP site or in a database.
Stream boss
The first step in implementing a new stream is to define the boss. Typically, a stream boss contains IPMStream and any interface required to identify the type of information in the stream,
the target or source of the stream, or both. Example 1creates a pointer-based read-stream boss:
EXAMPLE 1 kExtLinkPointerStreamWriteBoss, a pointer based stream boss
Class
{
kExtLinkPointerStreamWriteBoss,
kInvalidClass,
{
IID_IPMSTREAM, kExtLinkPointerStreamWriteImpl,
IID_IPOINTERSTREAMDATA, kPointerStreamDataImpl,
}
};
IPMStream is the only interface all stream bosses have in common. In Example 1, the IPointerStreamData controls a stream that writes out to memory; it contains a buffer and a length.
Persistent Data and Data Conversion
35
Persistent Data and Data Conversion
Missing plug-ins
Example 2 is another example.
EXAMPLE 2 kFileStreamReadBoss, a stream commonly used in importing
Class
{
kFileStreamReadBoss,
kInvalidClass,
{
IID_IPMSTREAM, kFileStreamReadLazyImpl,
IID_IFILESTREAMDATA, kFileStreamDataImpl,
}
};
IPMStream interface and the IXferBytes class
When implementing your own stream, take advantage of the default implementations of IPMStream, CStreamRead, and CStreamWrite. These default implementations use an abstract base
class, IXferBytes, to do the actual reading and writing. To implement a stream for a new data
source, you must create an IXferBytes subclass that can read and write to that data source.
Missing plug-ins
This section discusses how to open a document that contains data saved by a plug-in that is no
longer available.
Plug-ins you create can add data to the document. When your plug-in is present and loaded, it
can open and interpret the data; however, if the user removes the plug-in and then opens the
document, or gives the document to someone who does not have the plug-in, the plug-in is not
available to interpret the data.
You have two ways to handle such situations:
z
Control what warning is shown when the document is opened without the plug-in.
z
Implement code to update the data the next time the document is opened with the plug-in.
The rest of this section describes these options.
Warning levels
The application can give a warning when it opens a document that contains data created by a
plug-in that is not available. There are three warning levels: critical, default, and ignore. By setting the warning level, the plug-in can specify the relative importance of its data. Data created
by the plug-in has the “default” warning level unless you override the setting and identify the
data as more important (critical) or less important (ignored). This importance settings can be
modified by adding resources to the plug-in’s boss definition file:
z
36
CriticalTags — A “critical” warning tells the user the document contains data from missing
plug-ins and strongly advises the user not to open the document. If the user continues the
Persistent Data and Data Conversion
Missing plug-ins
open operation, the application opens an untitled document that is a copy of the original, to
preserve the original document. Use this level when the data is visible in the document or
contributes objects owned by another object in the database, like text attributes, owned by
the text model.
z
DefaultTags — A “default” warning tells the user the document contains data from missing
plug-ins and asks whether to continue the open operation. If the user continues the open
operation, the application opens the original document. Use this level when the data is selfcontained and invisible to the user, but the user might encounter missing function that
would have been provided by the plug-in.
z
IgnoreTags — An “ignore” warning provides no warning message at all; the application proceeds with the open operation as if there were no missing plug-ins. Use this level when the
data is invisible to the user and completely self-contained. In this case, the user does not
need to know the plug-in was involved in the construction of this document. If the plug-in
stored data in the document, but that data is used only by this plug-in and does not reference objects supplied by other plug-ins, the user sees no difference in the document when
the plug-in is missing. For example, the plug-in might store preferences information in
every document for its own use.
You can set these warnings to use ClassID (when the plug-in creates new bosses) or ImplementationID (when the plug-in adds interfaces to existing bosses) values as triggers. Use kImplementationIDSpace to specify a list of ImplementationID values, and kClassIDSpace for
ClassID values. You can put any number of IDs in the list, but all the IDs must be of the same
type. Use a second resource to mark IDs of another type. Example 3 and Example 4 set the
warning level to ignore data stored by the PersistentList plug-in in the SDK by adding two
resources to PstLst.fr:
EXAMPLE 3 Marking implementation IDs as ignored
resource IgnoreTags(1)
{
kImplementationIDSpace,
{
kPstLstDataPersistImpl,
kPstLstUIDListImpl,
}
};
EXAMPLE 4 Marking boss classes as ignored
resource IgnoreTags(2)
{
kClassIDSpace,
{
kPstLstDataBoss,
}
};
You do not need to mark any IDs that do not appear in the document (for example, data that
was written out to saved data) or implementations that are not persistent.
You do not need to mark IDs if you want the default behavior.
Persistent Data and Data Conversion
37
Persistent Data and Data Conversion
Missing plug-ins
Missing plug-in alert
This alert is activated when a document is opened and contains data from one or more missing
plug-ins that cannot be ignored. The document contains a list of the plug-ins that added data to
it. Each piece of data added has an importance attached to it; this may be critical, default, or
ignorable. Data marked as ignorable does not cause the alert to be activated. Data marked as
critical or default causes the alert to be activated. In the case of critical data, the alert works
more strongly; this is the only difference between critical and default data.
The alert tells the user data is missing, presents a list of missing plug-ins, and allows the user to
continue or cancel the open operation. Each missing plug-in has the chance to add a string to
the alert that specifies additional useful information (e.g., a URL for purchasing or downloading the plug-in). The alert is modeled on the missing-font alert.
The “Don’t Warn Again For These Plug-ins” option is de-selected by default. If this option is
selected, the alert is not activated the next time a document is opened and any subset of the
listed plug-ins is missing (and no other plug-ins are missing). This allows users accustomed to
seeing (and ignoring) alerts concerning specific plug-ins to automatically bypass the alert,
while still getting warned about data from any plug-ins newly found to be missing. The alert is
activated again if a document is opened that uses other missing plug-ins. The alert is activated
again if the “Don’t Warn Again For These Plug-ins” option is de-selected.
Guidelines for handling a missing plug-in
If a plug-in creates persistent data in a document, these guidelines ensure the document
behaves gracefully if a user tries to open it when the plug-in is missing or if the document is
edited by a user who did not have the plug-in:
If your plug-in does not store data in documents, you do not need to take any special precautions.
If the data stored by your plug-in does not reference other data in the document and does not
appear visually in the document, mark the data as ignorable.
If editing the document without your plug-in could corrupt the document, mark the data as
critical. You can specify a string that is displayed when the plug-in is missing and the user
opens a document that contains data added by the plug-in. See the ExtraPluginInfo resource,
which may provide information like the URL of a site from which the missing plug-in can be
obtained. See an example of the use of this resource in <SDK>/source/sdksamples/transparencyeffect/TranFx.fr.
If there are checks your plug-in can do to restore the data’s integrity on opening a document
edited without the plug-in, supply a FixUpData method. For an example, see
<SDK>/source/sdksamples/persistentlist/PstLstPlugIn.cpp.
If you want your plug-in to handle the storage of its own data, use the application-supplied
mechanism that treats the plug-in’s data like a black box. (See “Black-box data storage” on
page 39.)
38
Persistent Data and Data Conversion
Missing plug-ins
Data handling for missing plug-ins
If a document contains data placed there by a plug-in that is not available, the user can choose
to open the document anyway. If the data is completely self-contained, there may be no problem; however, if the plug-in’s data depends on anything else in the document, undesirable
things can happen.
Missing data not copied
InDesign maintains most data in a centrally managed model in which the core-content manager keeps track of what information is added to the document, handles conversion of the data,
and provides a convenient mechanism for instantiating objects based on the data. This
approach does not allow the data to be copied when the plug-in is missing, however, because
the content manager would not be able to provide these services for the missing plug-in’s copied data, and that potentially can leave the document in an invalid state. With the exception of
those objects that hold onto only ClassID values, InDesign blocks copying data associated with
missing plug-ins.
This means no attribute is copied if the plug-in that supplies the attribute is missing. This
applies to all attributes: text, graphics, table, cjk, etc. Furthermore, if a plug-in attached a data
interface to an existing attribute, and the plug-in is missing, the attribute is copied but the addin data interface is not. This is consistent with how InDesign handles UID-based objects.
Likewise, if a data interface is added to an XML element, the data interface is not copied if the
plug-in that supplied it is missing.
There are several features based on an object from a required plug-in holding a ClassID from
an optional plug-in, including adornments, text composer, pair-kern algorithm, section numbers, and unit defaults. In these cases, the consequences of losing track of the plug-in that supplied the data is much less severe. Conversion of these ClassIDs is quite unusual and could be
handled if necessary by issuing new ClassIDs. Error handling in the user interface when the
plug-in is missing is much more graceful.
Black-box data storage
A second, simpler data-model storage mechanism was added for software developers requiring
that data (like text attributes) is copied with the text, and for developers who want to attach
data to other ID objects, such that it gets copied even when the source plug-in is missing. This
mechanism is black-box data storage.
In the simplest case, the black box is just a new persistent interface that sits on the object. A
plug-in can store data in the box or fetch data out of the box. The data is keyed by a ClassID,
which is supplied by the plug-in. Multiple plug-ins can store their data in the same black box,
and each plug-in gets its own, unique, streamed access to its data. The black box just keeps
track of the key, the length of the data, and the data stream. The software developer is responsible for handling everything else—conversion, swapping, etc. Users do not get a missing plug-in
alert for data placed in a black box.
Any UID-based object could have a black box. In addition, attributes and small bosses (used
for XML elements) also can have black boxes.
Persistent Data and Data Conversion
39
Persistent Data and Data Conversion
Conversion of persistent data
The following objects support black boxes:
z
kDocBoss — The root object of the document does not get copied.
z
kPageItemBoss — This includes all page item objects, including spreads, master pages,
splines, frames, images, and text frames.
z
Attributes — This includes text, graphic, table, cjk, etc. (i.e., everything that appears in an
AttributeBossList).
For more information on the black-box mechanism, see IBlackBoxCommands, IBlackBoxCmdData, and IBlackBoxData in the API reference documentation.
FixUpData
Suppose a hyperlink attribute is linked to another frame, and the user can double-click the link
to go to the frame. A plug-in supplies an observer, so if the frame is deleted, the link is severed.
Now suppose you give the document containing the hyperlinks to someone who does not have
the hyperlink plug-in. This person edits the document, deletes the frame, saves the document,
then returns the document to you. The document is now corrupted, because your plug-in was
unable to delete the associated link, which now points to an undefined frame.
To restore the integrity of a document in this case, the plug-in can override the IPlugIn::FixUpData method. This method is called when the document is opened, if the plug-in was used to
store data in the document and the document was edited and saved without the plug-in. In this
case, the hyperlinks plug-in could override FixUpData to scan all its objects, checking whether
the linked frame UIDs were deleted; when the document is opened with the plug-in, the
method correctly severs the links.
Conversion of persistent data
This section describes types of conversion providers and the advantages of each type.
Converting persistent data from an old document to a new one is complex, because each plugin can store data independently in the document. When the host opens a document created
with an older version of the application or an older version of any plug-in used in the document, you must convert and update the older data to match the new format.
Versioning persistent data is the process of converting persistent data in a database from one
format to another. The data in different formats usually resides in different databases; for
example, data in a database (document) from a previous version of InDesign versus that in a
database (document) from a newer version of InDesign. Just as each plug-in having a persistent
data implementation is responsible for the order of reading and writing its own data (thus
implicitly defining a data format), each plug-in also is responsible for converting its own data
from one format to another. Whether data in a database requires conversion usually is determined when the database is opened.
40
Persistent Data and Data Conversion
Conversion of persistent data
There are two approaches to converting persistent data:
z
You can use the host’s conversion manager to manage the conversion process. This
approach is the most common. See “Converting data with the conversion manager” on
page 42.
z
The plug-in that owns the data can manage when and how data is converted, using version
information or other data embedded with the object data. See “Converting data without the
conversion manager” on page 52.
When to convert persistent data
As a plug-in developer, you want to ensure your users can open documents with persistent data
from an older version of your plug-in. To do this, your plug-in must provide data conversion
function. The best time to consider your data-conversion strategy is when you realize your
plug-in will store some data to a database. You need to implement data-conversion utilities in
the following cases:
z
You change the order of IPMStream::Xfer* calls in the ReadWrite method.
z
You change a persistent object’s definition.
z
You re-number (change the value of) an ImplementationID or ClassID identifier.
z
You remove a plug-in and data from the removed plug-in might be in a document a user
wants to open.
In any of these cases, the conversion manager needs to be notified how to convert the persistent
data for use by the loaded plug-in.
Specifying how the persistent data format changed is somewhat different from adding persistent data to or removing it from a document. Besides telling the conversion manager to add or
delete any obsolete data, the conversion provider has to be able to tell the conversion manager
about every implementation in the plug-in that was ever removed, to keep the content manager
up to date about the various persistent data formats.
NOTE:
To provide a document for use with an earlier version of InDesign, use the InDesign
Interchange file format.
At the very least, the resources required to support data conversion can help you keep a log of
how your persistent data format has changed. Such a log can be useful.
Sample conversion scenario
Consider a plug-in with two released versions, 1 and 2, which use the same data format; in this
case, both versions of the plug-in have the same format version number, 1. The new release of
the plug-in, version 3, stores additional data, such as a time stamp. You must update the format
version number to match the current plug-in version number; so, for plug-in version 3, the format version also would be 3. Because you changed the format version number, you must create
a converter that converts from version 1 to version 3. See Table 1.
Persistent Data and Data Conversion
41
Persistent Data and Data Conversion
Conversion of persistent data
TABLE 1 Version changes example
Plug-in version
Format change
Format version
1
N/A (new plug-in)
1.0
2
No
1.1
3
Yes
3
4
Yes
4
For a fourth version of the plug-in, you again change the format, allowing a date stamp to be
signed. Change the plug-in and format version numbers to 4, and add an additional converter
to convert from version 3 to version 4. Conversions from version 1 to version 4 are done by the
conversion manager, which chains the converters together; the first converts from format version 1 to format version 3, and the second converts from format version 3 to format version 4).
Converting data with the conversion manager
Each document contains header information about the content, which includes a list of all
plug-ins that wrote data to the document and the version number of the plug-in last used to
write that data. When a document is opened, the application checks whether any plug-ins are
missing or out of date. If a plug-in is missing, it might provide an alert embedded in the document. (See “Missing plug-ins” on page 36.)
If a plug-in is out of date, data written by the old plug-in must be updated to match the format
required by the loaded plug-in. A plug-in can register a conversion service to do the update.
The InDesign conversion manager (IConversionMgr) determines which conversions are
required for opening the document and calls the appropriate plug-in to do the conversion.
When the persistent data for any plug-in changes, this is a document format change. Any of the
following can change the document format:
42
z
Changes to the ReadWrite method — If the ReadWrite method is used to stream data to the
document, changing the ReadWrite method might change the document format; however,
an implementation might have a ReadWrite method that works with some other database
or other data source, not with the document itself. For example, a widget has a ReadWrite
method used for streaming to and from resources and to and from the SavedData file.
Changes to a method that does not work with the document database do not require any
special conversion.
z
Changes to an object’s definition — If you add an implementation to (or remove an implementation from) the definition of a persistent boss in the framework resource (.fr) file, you
change how the object is streamed. If you add a new implementation, an old version of the
object will stream, but it will not contain the data normally appearing for the implementation you added. This is fine if the data can be initialized adequately from the implementation’s constructor; otherwise, you may need to add a converter. If you change the
implementation of an interface from one ImplementationID to another, you must convert
the data. If you remove an ImplementationID from a class, you should add a converter to
Persistent Data and Data Conversion
Conversion of persistent data
strip the old data from the object; otherwise, the obsolete data is carried around with the
object indefinitely.
z
Re-numbering an ImplementationID or ClassID — If an ImplementationID or ClassID
changes, you must register a converter so occurrences of the ImplementationID or ClassID
in old documents can be updated. In practice, re-numbering a ClassID or ImplementationID is a source of many bugs and typically leads to corrupt documents, so we strongly
recommend you not re-number an ImplementationID or ClassID.
When you make a format change, you must do two things to maintain backward compatibility:
z
Update the version number of the plug-in whose data format you changed.
z
Provide a converter that can convert between the previous format and the new format.
Updating version numbers
Each plug-in has a plug-in version resource of type PluginVersion. The PluginVersion resource
appears in the boss definition file. The first entry of this resource describes the application's
build number; on the release build, this entry is the final-release version number. The second
entry of the resource is the plug-in's ID, followed by three sets of version numbers, followed by
an array of feature-set IDs. If any of the IDs in this list matches the current feature-set ID, the
plug-in is loaded. To see an example, open any example .fr file in the SDK.
Each version number has a major number and a minor number. The first version number is the
version of the plug-in, which gets updated for every release of the plug-in. The second version
number is the version of the application with which the plug-in expects to run. This ensures
the user does not drop a plug-in compiled for one version of the application into an installation
of another version of the application. The last number is the format version number, which
indicates the version of the plug-in that last changed the file format; this is the version number
written into the document, and it is the number the conversion manager checks to see whether
conversion is required.
The format version number does not always match the plug-in’s version number. The format
version number does not generally change as often as the plug-in version number. The plug-in
version number changes for every release of a plug-in, but the format version number changes
only if the format for the data the plug-in stores in the document has changed and the conversion manager is required to convert the data.
Adding a converter
Converters can be implemented as conversion services. InDesign supports two types of service-provider-based data conversion:
z
Schema-based provider — Schema-based converters are configured through resources, are
easier to use than code-based converters, and cover most format-change needs. Use this
type of converter unless it cannot handle the special needs of your plug-in.
z
Code-based provider — If your implementation uses a special data-compression algorithm
or other storage optimizations, involves data of variable length, or uses virtual object store
(VOS) objects, schema-based converters cannot handle the necessary data format conversions. In this case, you must implement your own custom conversion provider.
Persistent Data and Data Conversion
43
Persistent Data and Data Conversion
Conversion of persistent data
A conversion service is responsible for all conversions done by the plug-in. A converter might
at first handle only a single conversion, from the first format to the second. Later, if you change
the format again, you can add another conversion to the converter, to convert from the second
format to the third.
Suppose you market a plug-in that supplies a date stamp. You had released two versions (version 1.0 and version 2.0) of the plug-in without changing the persistent data created by the
plug-in; so both releases have the same format version number (1.0). For the third released version of the plug-in, you add a time stamp. You must update the format version number to
match the current plug-in version number; i.e., for plug-in version 3.0, the format version also
must be 3.0.
Suppose for the fourth released version of the plug-in (version 4.0), you again change the format, allowing a date stamp to be signed. You then change the format version number to 4.0 and
add an additional converter, capable of converting from format version 3.0 to format version
4.0. Conversions from format version 1.0 to format version 4.0 can be done by the conversion
manager, which chains the two converters together, using the first converter to convert from
format version 1.0 to format version 3.0, and using the second converter to convert format version 3.0 to format version 4.0.)
Adding a converter (either schema-based or code-based) to your plug-in means adding a new
boss with two interfaces, IK2ServiceProvider and IConversionProvider. The
IK2ServiceProvider implementation, kConversionServiceImpl, is provided by the application.
You need to supply the IConversionProvider implementation. Here is a sample boss:
/**
This boss provides a conversion service to the conversion manager
to use the schema-based implementation.
*/
Class
{
kMyConversionProviderBoss,
kInvalidClass,
{
IID_ICONVERSIONPROVIDER, kMySchemaBasedConversionImpl,
IID_IK2SERVICEPROVIDER, kConversionServiceImpl,
}
},
Most of the work is done in the conversion provider supplied with the SDK.
The default implementation of IConversionMgr calls IConversionProvider::CountConversions
and IConversionProvider::GetNthConversion to determine which conversions are supported
by the converter. When you implement a new converter, CountConversions returns 1, and GetNthConversion returns fromVersion set to the version number before your change and toVersion set to the version number having your change in it. VersionID is a data type in the Public
library; it consists of the PluginID value, the major format version number, and the minor format version number.
For a new converter, fromVersion should be VersionID(yourPluginID, kOldPersistMajorVersionNumber, kOldPersistMinorVersionNumber). The toVersion should be the new format version number.
44
Persistent Data and Data Conversion
Conversion of persistent data
Your new conversion is added as conversion index 0. For a new converter, it looks like this:
int32 TextConversionProvider::CountConversions() const
{
return 1;
}
void TextConversionProvider::GetNthConversion(
int32 i, VersionID* fromVersion, VersionID* toVersion) const
{
*fromVersion = VersionID(kTextPluginID, kOldPersistMajorVersionNumber,
kOldPersistMinorVersionNumber);
*toVersion = VersionID(kTextPluginID, kNewPersistMajorVersionNumber,
kNewPersistMinorVersionNumber);
}
When adding another change after a changed version already exists, the change numbers
should chain together so the conversion manager can do changes across multiple formats. So, if
your plug-in used three formats—starting with version 1, then changed in version 3, and
changed again in version 4—your plug-in should register one converter that handles the conversion from format version 1 to format version 3 and another converter that handles the conversion from format version 3 to format version 4. If necessary, the conversion manager can
chain them together to convert a document from version 1 to version 4. The methods would
look like this:
const int32 kFirstFormatVersion = 1;
const int32 kSecondFormatVersion = 3;
const int32 kThirdFormatVersion = 4;
const int32 kFirstChange = 0;
const int32 kNewChange = 1;
int32 TextConversionProvider::CountConversions() const
{
return 2;
}
void TextConversionProvider::GetNthConversion(
int32 i, VersionID* fromVersion, VersionID* toVersion) const
{
if (i == kFirstChange)
{
*fromVersion = VersionID(kTextPluginID, kMajorVersionNumber,
kFirstPersistMinorVersionNumber);
*toVersion = VersionID(kTextPluginID, kMajorVersionNumber,
kSecondPersistMinorVersionNumber);
}
else if (i == kNewChange)
{
*fromVersion = VersionID(kTextPluginID, kMajorVersionNumber,
kSecondPersistMinorVersionNumber);
*toVersion = VersionID(kTextPluginID, kMajorVersionNumber,
kThirdPersistMinorVersionNumber);
}
}
Persistent Data and Data Conversion
45
Persistent Data and Data Conversion
Conversion of persistent data
Next, tell the conversion manager which information you changed. The default implementation of IConversionMgr calls IConversionProvider::ShouldConvertImplementation with each
implementation in the document supplied by the plug-in. Depending on the data status of the
implementation, the plug-in must return kMustConvert when the content must be modified,
kMustRemove when the data is obsolete and must be removed, kMustStrip when the content
must be stripped, or kNothingToConvert for all other cases.
Using kTextFrameImpl as an example, define a method for IConversionProvider::ShouldConvertImplementation, returning kMustConvert when the tag is kTextFrameImpl and kNothingToConvert in all other cases. The plug-in should do this when the conversion manager is
converting that plug-in’s data and not performing another conversion, so check the conversion
index and make sure it is the one that you added. It might look like the following:
IConversionProvider::ConversionStatus
TextConversionProvider::ShouldConvertImplementation(
ImplementationID tag, ClassID context, int32 conversionIndex) const
{
IConversionProvider::ConversionStatus status =
IConversionProvider::kNothingToConvert;
switch (conversionIndex)
{
case 0:
if (tag == kTextFrameImpl)
status = IConversionProvider::kMustConvert;
break;
default:
break;
}
return status;
}
Next, implement ConvertTag to do the actual conversion. Suppose the TextFrame ReadWrite
method writes an integer value and a real value, and you are adding a boolean. The ConvertTag
implementation might look like the following, which illustrates what most converters look like:
ImplementationID TextConversionProvider::ConvertTag(
ImplementationID tag, ClassID forClass, int32 conversionIndex, int32 inLength,
IPMStream* inStream, IPMStream* outStream, IterationStatus whichIteration)
{
ImplementationID outTag = tag;
switch (conversionIndex)
{
case 0:
if (tag == kTextFrameImpl)
{
if (inLength > 0)
{
int32 passThruInt; inStream->XferInt32(passThruInt);
outStream->XferInt32(passThruInt);
int32 passThruReal; inStream->XferInt32(passThruReal);
outStream->XferInt32(passThruReal); // Adding new field bool16 smartQuotes =
kTrue;
outStream->XferInt32(smartQuotes);
46
Persistent Data and Data Conversion
Conversion of persistent data
}
} break;
default:
break;
} return outTag;
}
If the converter needs to convert a class, it implements ShouldConvertClass and ConvertClass.
This is necessary only if the class is being deleted or is a container for some other data. (See
“Containers and embedded data” on page 48.)
Removing classes or implementations
If you remove a ClassID or an ImplementationID from a version of your plug-in, the identifier
also must be removed from documents as they are converted. The conversion manager
removes an identifier for you if you implement your converter to return invalid status. For
example, to remove a tag, implement ConvertTag as follows:
ImplementationID TextConversionProvider::ConvertTag(
ImplementationID tag, ClassID forClass, int32 conversionIndex, int32 inLength,
IPMStream* inStream, IPMStream* outStream, IterationStatus whichIteration)
{
ImplementationID outTag = tag;
switch (conversionIndex)
{
case 0:
if (tag == kTagToRemove)
outTag = kInvalidImpl;
break;
default:
break;
}
return outTag;
}
Deleting a class is similar: implement ConvertClass to return kInvalidClass.
Changing an implementation in one class
Suppose a boss is defined with an IBoolData interface, implemented as kPersistBoolTrue,
meaning it defaults to true. Suppose you change this to kPersistBoolFalse. When you read an
old document with the new boss, you want it to use the new implementation. You would then
write a converter to take the data stored as kPersistBoolTrue and switch it to be stored as kPersistBoolFalse. When you first access the boss, it has the old value. If you do not make a converter, the boss will not read the old data, because there is no interface in the new boss using
the kPersistBoolTrue ImplementationID. Instead, the boss looks for kPersistBoolFalse but does
not find it, because the old boss did not have it; therefore, the value is false, because it is the
default value instead of the old value.
To change the ImplementationID of the data in the document, make a conversion method to
catch the old ImplementationID and change it to the new one. Do this only for your boss; do
not change other bosses using kPersistBoolTrue. Your method might look like this:
Persistent Data and Data Conversion
47
Persistent Data and Data Conversion
Conversion of persistent data
ImplementationID TextConversionProvider::ConvertTag(
ImplementationID tag, ClassID forClass, int32 conversionIndex, int32 inLength,
IPMStream* inStream, IPMStream* outStream, IterationStatus whichIteration)
{
ImplementationID outTag = tag;
switch (conversionIndex)
{
case 0:
if (tag == kPersistBoolTrue && forClass == kMyBoss)
outTag = kPersistBoolFalse;
bool16 theBool;
inStream->XferBool(theBool);
outStream->XferBool(theBool);
break;
default:
break;
}
return outTag;
}
forClass is the boss in which the data was found. The conversion should be done only if
forClass is the one you want changed.
Also implement ShouldConvertImplementation. forClass is either the class in which the data
was found or kInvalidClass if any class should match. The implementation of ShouldConvertImplementation looks like this:
IConversionProvider::ConversionStatus
TextConversionProvider::ShouldConvertImplementation (
ImplementationID tag, ClassID forClass, int32 conversionIndex) const
{
IConversionProvider::ConversionStatus status =
IConversionProvider::kNothingToConvert;
switch (conversionIndex){
case 0:
if (tag == kTextFrameImpl &&
(forClass == kMyBoss || forClass == kInvalidClass))
status = IConversionProvider::kMustConvert;
break;
default:
break;
}
return status;
}
Containers and embedded data
InDesign does not have many instances of containers. One example is in the text attribute code.
A text style is a UID-based boss object containing an implementation, TextAttributeList, that
contains a list of attribute bosses (Bold, Point Size, Leading, and so on). These attribute bosses
are not UID-based bosses; instead, TextAttributeList streams the ClassID for a boss, followed
by the boss’s data, then streams the next boss in the list. TextAttributeList, therefore, must call
the stream’s content tracker to notify the content manager that a new class was added to the
document.
48
Persistent Data and Data Conversion
Conversion of persistent data
The following example illustrates why this is important.
Suppose a new Red-Line plug-in adds an attribute to the style and the text does not compose
correctly if the Red-Line plug-in is removed. Red-Line can list the attribute as critical, so the
host provides a strongly worded warning when the user tries to open the document without the
Red-Line plug-in. However, if TextAttributeList does not register the attribute boss with the
content manager, the application never knows the attribute is in the document and cannot
warn the user. Suppose the Red-Line plug-in is updated, and the attribute boss in particular is
updated to store its data differently. The Red-Line plug-in registers a converter handling the
format change. If the host does not know the attribute appears in the document, the Red-Line
converter is never called to perform the conversion. The document appears to open correctly,
but it crashes on the attribute’s ReadWrite method the first time the style is read.
For the Red-Line attribute to be converted, TextAttributeList must register a converter. If the
content manager was notified when the attribute was streamed, the conversion manager knows
the attribute needs to be converted. It even knows the attribute was streamed by TextAttributeList. But the conversion manager has no way of knowing where the attribute is inside the
data streamed by TextAttributeList; therefore, TextAttributeList should register a converter that
calls back to the conversion manager to convert each piece of embedded data. Otherwise, the
embedded data is not converted.
Content manager
The host has a content manager that tracks what data is stored in the document, which plug-in
stored it, and the format version number of the plug-in last used to write the data.
Think of this as a table attached to the root of the document: every ImplementationID part of
the document appears in the table. This table also includes all ClassID values in the document.
Table 2 and Table 3 are an example.
TABLE 2 Version information
ImplementationID
PluginID
Format version number (only
minor number)
kSpreadImpl
kSpreadPluginID
0
kSpreadLayerImpl
kLayerPluginID
0
kTextAttrAlignJustImpl
kTextAttrPluginID
1
kCMSProfileListImpl
kColorMgmtPluginID
3
...
Persistent Data and Data Conversion
49
Persistent Data and Data Conversion
Conversion of persistent data
TABLE 3 Class IDs in the document
ClassID
Plug-in ID
Format version number (only
minor number)
kSpreadBoss
kSpreadPluginID
1
kSpreadLayerBoss
kLayerPluginID
1
kTextAttrAlignBodyBoss
kTextAttrPluginID
3
kDocBoss
kDocFrameworkPluginID
1
When an object is streamed to a document, the document receives the ClassID of the object
and the data streamed by each of the object’s ReadWrite methods. The data written by a ReadWrite method is marked with the ImplementationID of the ReadWrite implementation; this
way, when the data is read later, the system knows which C++ class to instantiate to read the
data. IContentMgr maintains an overall list of the ClassID and ImplementationID values used
in the document. When a new ClassID or ImplementationID is added to the document, IContentMgr checks which plug-in supplied the ClassID or ImplementationID, notes the plug-in ID
and the format version number, and adds the new entry to the table.
The plug-in supplying a ClassID is the one supplying the class definition. Only the plug-in supplying the original class definitions is considered; add-in classes do not count. The plug-in supplying an ImplementationID is the one that registered a factory for the ImplementationID.
This data is used to detect missing plug-ins and manage document conversion. When the user
opens a document containing data supplied by a missing plug-in, the host alerts the user that
the document contains data from a missing plug-in. The host detects missing plug-ins by looking in the content manager to see which plug-ins supplied data to the document and checking
this list against the list of loaded plug-ins to see whether any are missing.
This data also is used for document conversion. When the user opens a document, the default
implementation of IConversionMgr checks the format version number of the plug-ins supplying data to the document and compares it with the format version number of loaded plug-ins.
Any mismatch means a conversion is required before the document can be opened. If there is a
format version change without a supplied data converter, the document will not open.
Conversion manager
When a document is opened, the conversion manager is called to check whether any data in
the document requires conversion. If the data requires conversion and a converter is available,
the document is converted and opens as an untitled document. If a converter cannot be found,
the host displays an alert message that warns the user the document cannot be opened because
it cannot be converted.
This initial check, implemented in IConversionMgr::GetStatus, happens very early in the process of trying to open the document, before any objects in the document are accessed (i.e.,
before any of their ReadWrite methods are called). This is critical, because a ReadWrite
method will not succeed if the object needs to be converted. The conversion manager accesses
the content manager (converting it if necessary), and uses the content manager to find out what
50
Persistent Data and Data Conversion
Conversion of persistent data
plug-ins supplied data to the document. If any plug-ins used in the document have different
formats than the formats of the loaded plug-ins, conversion is necessary. Next, the conversion
manager sees whether there is a converter to convert from the format in the document to the
format associated with the loaded plug-in. If such a converter is available, conversion is possible; the document may be closed and opened again as an untitled document.
If the GetStatus method returns with a result indicating conversion is not necessary, the open
operation proceeds without calling the conversion manager again. If the GetStatus method
returns with a result indicating conversion is necessary but impossible, the open operation is
aborted. If the GetStatus method returns with a result indicating conversion is required and a
converter is available, the conversion manager’s ConvertDocument method is called to perform the conversion.
The first thing ConvertDocument does is compile a list of the classes and implementations in
the document that must be converted. A plug-in may have many different implementations
that wrote data to the document, but only one of these implementations might require conversion. The conversion manager uses the content manager to iterate over classes and implementations supplied by the plug-in, and the conversion manager calls the converter to find out
which classes and implementations require conversion. Each class or implementation in the
document that the converter says must be converted is added to the list of classes or list of
implementations to be converted. Other data written by the plug-in is ignored by the converter;
it is not converted.
The next step is to iterate over the UIDs in the document. For each UID, the conversion manager gets the class of the UID. If the class is on the list of classes to be converted, the conversion
manager calls a converter for the class. If, as is more common, the class contains an implementation needing to be converted, the conversion manager opens an input stream on the UID and
iterates through the implementations in the UID. If any implementation in the UID requires
conversion, an output stream is opened on the UID. Any implementation not requiring conversion is copied to the output stream. If an implementation requires conversion, its converter
is called. The converter gets passed the input stream and the output stream. The converter
reads from the input stream in the old format and writes to the output stream in the new format.
After the UID iteration is completed, the content manager is called to update the format numbers for all plug-ins that were outdated, to show their data was converted and is up to date.
After this, it is safe for ReadWrite methods to be called on objects in the document.
If any converter requires access to another object to do its conversion, it is called in the last
phase of conversion. For example, suppose there was one document-wide preference setting for
whether text in frames has smart quotes turned on, but you want to make this a per-frame setting. The text frame must store a new flag that indicates whether smart quotes are on or off.
The converter adds a new flag in the first phase of conversion, but the converter cannot set the
correct value for the flag until the preferences are converted. So, the converter needs to be
called once again, after the first phase is complete and it is safe to instantiate the preferences, to
get the value of the document-wide smart-quotes setting so the converter can set the frame’s
new smart-quotes setting accordingly.
Persistent Data and Data Conversion
51
Persistent Data and Data Conversion
Conversion of persistent data
Converting data without the conversion manager
To perform format conversion without the conversion manager, the persistent data itself must
identify the version of the plug-in that originally wrote the information. To use this method,
add a version number to the implementation classes, and implement the ReadWrite method to
write the version number first whenever the method writes data, as shown in Figure 3. This
method has the advantage of forward compatibility.
FIGURE 3
Persistent data with version number
k<Foo>Boss persistent data record
ID
Length
Data
Vers. | Other
ID
Length
Data
Vers. | Other
For example, suppose the first version of your plug-in stores two-byte values, fOne and fTwo,
and uses a byte value to store the data’s version number. The ReadWrite method for this implementation would be something like Example 5:
EXAMPLE 5 ReadWrite method supporting multiple versions (version 1 of the plug-in)
FooStuff::ReadWrite(IPMStream* stream, ImplementationID implementation)
{
stream->XferByte(0x01);
stream->XferByte(fOne);
stream->XferByte(fTwo);
}
The second version of this plug-in modifies the data format by adding a 16-bit integer value,
fThree. In this version, your ReadWrite method must be able to read data in either format, but
it always must write data back in the new format. See Example 6:
EXAMPLE 6 ReadWrite method supporting multiple versions (version 2 of the plug-in)
FooStuff::ReadWrite(IPMStream* stream, ImplementationID implementation)
{
uchar version = 0x02; //Always *write* version 2 data
stream->XferByte(version);
if (version == 0x01)
{
stream->XferByte(fOne);
stream->XferByte(fTwo);
}
else
{
stream->XferByte(fOne);
stream->XferByte(fTwo);
stream->XferInt16(fThree);
}
}
52
Persistent Data and Data Conversion
Resources
This code both reads and writes all data for this implementation. It could check whether the
stream is a reading or writing stream and perform only the needed operation, but the dual-purpose code actually is simpler and accomplishes the same thing. If the stream is a ReadStream,
the version is initialized as 0x02 but immediately replaced by the contents of the first byte in
the stream, and the rest of the stream is processed according to the version number found. If
this plug-in encounters data claiming a version number greater than 2, only the data it understands (processed by the else clause) is read. This method allows the version 2 plug-in to work
with data from both earlier and later versions. Each new version of the plug-in using this
method must preserve the portion of the stream previous versions created and add new information only to the end.
When using this approach, the plug-in’s data version number must not change between versions of the plug-in, but only when a data converter is being supplied to convert the data from
one version to another.
Resources
This section explains the fundamental elements and ODFRez resources you need when incorporating a converter in your plug-in.
PluginVersion resource, format numbers, and their macros
PluginVersion is a resource included in every InDesign and InCopy plug-in. Part of the
resource is the declaration of a persistent data format number, like the following:
kSDKDefPersistMajorVersionNumber, kSDKDefPersistMinorVersionNumber,
For an example, see the resource definition in <SDK>/source/sdksamples/basicdialog/BscDlg.fr.
Each plug-in specifies a different format number. These format numbers are stored in documents, so when a document is opened, the content manager can determine whether data conversion is needed. This determination is made by comparing the format numbers stored in the
document for each plug-in with the format numbers declared by loaded plug-ins. If the format
numbers are different, the conversion manager is called upon to do a data conversion.
Format number values must meet the following criteria:
z
Be greater than or equal to zero.
z
Increase with each format change.
z
Increment if the data format of any persistent class in a plug-in changes.
NOTE:
If multiple persistent classes in a plug-in change their data formats at the same time, the
data-format version number needs to increment only once. How much you increment
data format version numbers is up to you.
Format-number values are just numbers; however, you might find it easier to use #define macros for format numbers instead of using their values directly.
Persistent Data and Data Conversion
53
Persistent Data and Data Conversion
Resources
Table 4 lists format-number macros and their values from previous InDesign SDKs. See
<SDK>/source/sdksamples/common/SDKDef.h.
TABLE 4 Format number macros from prior InDesign SDKs
Version
Macro: Major
Macro: Minor
Tuple
1.0
kSDKDef_10_PersistMajorVersionNumber
kSDKDef_10_PersistMinorVersionNumber
{0, 307}
1.5
kSDKDef_15_PersistMajorVersionNumber
kSDKDef_15_PersistMinorVersionNumber
{1, 0}
1.0J
kSDKDef_1J_PersistMajorVersionNumber
kSDKDef_1J_PersistMinorVersionNumber
{2, 1}
2.0 / 2.0J
kSDKDef_20_PersistMajorVersionNumber
kSDKDef_20_PersistMinorVersionNumber
{4, 1}
CS / CS
Japanese
(3.0 /
3.0J)
kSDKDef_30_PersistMajorVersionNumber
kSDKDef_30_PersistMinorVersionNumber
{5, 1}
Setting up resources
With the first format change for a plug-in, you must add a converter (either schema-based or
code-based) to your plug-in (only one converter per plug-in). See “Adding a converter” on
page 43.
Schemas
The schema resource defines which formats of which implementations the converter should
handle. The schema resource is defined in your .fr file and compiled using the ODFRC (Open
Document Framework Resource Compiler). See Example 7 and <SDK>/public/includes/Schema.fh.
EXAMPLE 7 Schema resource
resource Schema(uniqueResourceID)
{
ImplementationID,
{ schemaFormatMajorNumber, schemaFormatMinorNumber },
{ // [0..n] SchemaFields go here (see SchemaFields.fh) }
};
Examples
When you define a schema resource, you explicitly state which format number of which implementation it defines. It knows which plug-in contains the implementation, because it is defined
implicitly to be the plug-in that contains the schema resource. In other words, all schemas
defined in plug-in A are for implementations provided by plug-in A. The schema-based converter uses the information in this resource (or the SchemaList resource) to determine how to
map persistent data from one format to another.
54
Persistent Data and Data Conversion
Resources
In Example 8, (1) identifies the schema resource ID. The value does not matter, as long as it is
unique among all schema resources compiled into the plug-in. The first item in the schema
specifies the implementation ID, and the second item specifies the format number as a tuple {1,
0}. Next comes the schema field list. Curly brackets ( { } ) delimit the list, and commas separate
the individual fields. Each field has a type, name, and default value. It is extremely important
each field name is unique and these values are not re-used when a field is deleted. Field names
must be unique among the schemas that describe different formats of the same implementation, because this is what allows data mapping between different format numbers.
NOTE:
Although it is not done in the example below, we recommend you use #define macros
for each field name, to enhance readability.
EXAMPLE 8 Schema resource from <SDK>/sdksamples/snapshot/Snap.fr
resource Schema(1)
{
kSnapPrefsDataPersistImpl, // ImplementationID
{RezLong(1), RezLong(0)}, // format number
{
{PMString {0x0001, ""}}, // dialog default file name
{ClassID {0x0002, 0}}, // format class ID, fFormatClassID
{Real {0x0003, 1.0}}, // fScale
{Real {0x0004, 72.0}}, // fResolution
{Real {0x0005, 72.0}}, // fMinimumResolution
{Real {0x0006, 0.0}}, // fBleed
{Bool16 {0x0007, 0}}, // fDrawArea
{Bool16 {0x0008, 0}}, // fFullResolutionGraphics
{Bool16 {0x0009, 0}}, // fDrawGray
{Bool16 {0x000a, 0}}, // fAddAlpha
{Int32 {0x000b, 512}}, // fDrawingFlags, set default to IShape::kPrinting == 512
{Bool16 {0x000c, 0}}, // fIndexedColour
}
};
In Example 9 (based on the preceding Snap.fr example), there is a new schema resource ID (2).
Inside the schema, the following changes occurred:
z
The implementation ID has not changed, but the format number is specified as a tuple {1,
1}.
z
schemaFormatMinorNumber changed from 0 to 1.
z
In the schema field list, the fifth field (0x0005, fMinimumResolution) was deleted. During
conversion, the fifth field (Real) is stripped from the existing data.
z
The seventh field changed its type. It uses the same name but is now of type Bool8. This
requires an implicit conversion. (For a table of valid and invalid type conversions, see the
“Versioning Persistent Data” chapter of Adobe InDesign CS4 Solutions.)
z
A new field (0x000d, fMaximumResolution) was added. On conversion, an element of type
Int32 (with a value of 3) is appended to the end of the existing data.
Persistent Data and Data Conversion
55
Persistent Data and Data Conversion
Resources
EXAMPLE 9 Hypothetical schema resource
resource Schema(2)
{
kSnapPrefsDataPersistImpl, // implementation ID
{RezLong(1), RezLong(1)}, // format number
{
{PMString {0x0001, ""}}, // dialog default file name
{ClassID {0x0002, 0}}, // format class ID, fFormatClassID
{Real {0x0003, 1.0}}, // fScale
{Real {0x0004, 72.0}}, // fResolution
{Real {0x0006, 0.0}}, // fBleed
{Bool8 {0x0007, 0}}, // fDrawArea
{Bool16 {0x0008, 0}}, // fFullResolutionGraphics
{Bool16 {0x0009, 0}}, // fDrawGray
{Bool16 {0x000a, 0}}, // fAddAlpha
{Int32 {0x000b, 512}}, // fDrawingFlags, set default IShape::kPrinting == 512
{Bool16 {0x000c, 0}}, // fIndexedColour
{Int32 {0x000d, 3}}, // minimum resolution enum
}
};
Default values
Default values in a schema are used only when the associated field is added to the data stream
by conversion. For example, if an implementation originally contained only one int32 value but
now also needs a bool16 value, the first schema lists only the int32 and the second schema lists
both the int32 and the bool16. When the conversion runs, the schema converter writes a
bool16 into the output stream, using the value specified as the default. Because the int32
already existed, the default value specified by the schema is not used.
Suppose you have an implementation with two int32 values. In your constructor, you give them
values of 11 and 52. The schema should reflect this. If you later decide 53 is a better default
value for the second one, change the schema to match. In this case, however, you do not need a
new schema.
SchemaList
The SchemaList resource allows you to specify several schemas in one resource, for convenience. Example 10 shows the two previous schema resources combined in one SchemaList.
Even if initially you have only one Schema inside your SchemaList, it is a good idea to create
this resource because, over the course of development, this SchemaList resource acts as a persistent data format log.
56
Persistent Data and Data Conversion
Resources
EXAMPLE 10 Hypothetical SchemaList resource
resource SchemaList(uniqueResourceID)
{{
Schema // schemaTypeIdentifier
{
kSnapPrefsDataPersistImpl,
{RezLong(1), RezLong(0)},
{
{PMString {0x0001, ""}},
{ClassID {0x0002, 0}},
{Real {0x0003, 1.0}},
{Real {0x0004, 72.0}},
{Real {0x0005, 72.0}},
{Real {0x0006, 0.0}},
{Bool16 {0x0007, 0}},
{Bool16 {0x0008, 0}},
{Bool16 {0x0009, 0}},
{Bool16 {0x000a, 0}},
{Int32 {0x000b, 512}},
{Bool16 {0x000c, 0}},
}
};
Schema // schemaTypeIdentifier
{
kSnapPrefsDataPersistImpl,
{RezLong(1), RezLong(1)},
{
{PMString {0x0001, ""}},
{ClassID {0x0002, 0}},
{Real {0x0003, 1.0}},
{Real {0x0004, 72.0}},
{Real {0x0006, 0.0}},
{Bool8 {0x0007, 0}},
{Bool16 {0x0008, 0}},
{Bool16 {0x0009, 0}},
{Bool16 {0x000a, 0}},
{Int32 {0x000b, 512}},
{Bool16 {0x000c, 0}},
{Int32 {0x000d, 3}},
}
};
}}
The possible types for schemaTypeIdentifier (see Example 10) are listed below. See Schema.fh
for their definitions.
z
ClassSchema — Schema for classes in the current plug-in.
z
OtherClassSchema — Like ClassSchema, but you can specify it for another plug-in by an
additional PluginID field.
z
ImplementationSchema — Schema for implementations in the current plug-in.
Persistent Data and Data Conversion
57
Persistent Data and Data Conversion
Resources
z
Schema — Another name for ImplementationSchema, because this is the most common
type.
z
OtherImplementationSchema — Like ImplementationSchema, but you can specify it for
another plug-in by an additional PluginID field.
DirectiveList
To instruct the schema-based converter to add or remove implementations or an entire boss,
specify a DirectiveList resource to your resource file, as shown in Example 11. (The DirectiveList resource formerly was known as BossDirective.) The DirectiveList resource is defined
in your .fr file and compiled using ODFRC.
EXAMPLE 11 DirectiveList syntax
resource DirectiveList(uniqueResourceID)
{
{ // [0..n] Directives go here }
};
A DirectiveList resource is required each time you do any of the following:
z
Add a new boss or remove an existing one.
z
Add an implementation to a boss or remove one from a boss.
z
Change the ID of a boss or implementation.
The DirectiveList resource serves several purposes:
z
Helps the conversion manager delete unnecessary data from a document when a boss or
implementation is removed.
z
Keeps the content manager up to date regarding a document’s contents.
z
Allows the conversion manager to do its job correctly, even when a boss or implementation
moves to a different plug-in.
The possible list of directives is in Table 5.
TABLE 5 Possible list of directives
58
Directive
Description
IgnorePlugin
Mark a plug-in as ignorable.
MoveClass
Moves a boss class from one plug-in to another.
MoveClassToPlugin
A boss was moved from one plug-in to another.
MoveImplementation
Moves an implementation from one plug-in to another.
MoveImplementationToPlugin
An implementation was moved from one plug-in to another.
RemoveAllImplementation
Removes an implementation from all boss classes.
RemoveAllOtherImplementation
Removes an implementation from all boss classes.
Persistent Data and Data Conversion
Advanced schema topics
Directive
Description
RemoveClass
Removes a boss class from a plug-in.
RemoveImplementation
Removes an implementation from a boss class.
RemoveOtherClass
Removes a boss class from another plug-in.
RemoveOtherImplementation
Removes an implementation from a boss class.
RemovePlugin
Removes an entire plug-in.
RenumberPlugin
Re-numbers an entire plug-in.
ReplaceAllImplementation
Replaces one implementation with another in all plug-ins.
ReplaceClass
Replaces one boss class with another.
ReplaceImplementation
Replaces one implementation with another in a specific plug-in.
NOTE:
See Schema.fh. Fields vary by type of directive.
Advanced schema topics
Arrays of values
Suppose an implementation contains three boolean flags followed by four uint32 values. The
schema could contain seven separate fields, or it could define two array fields, as in
Example 12:
EXAMPLE 12 Schema resource with array fields
#define kBarOptions 1
#define kBarValues 2
resource Schema(2)
{
kBarImpl,
{1, 0},
{
{Bool16Array{kBarOptions, {kTrue, kFalse, kTrue}}},
{Uint32Array{kBarValues, {0, 0, 0, 0}}}
}
};
The number of default values statically determines the number of elements in each array. Note
that the set of default values is enclosed in braces.
Persistent Data and Data Conversion
59
Persistent Data and Data Conversion
Advanced schema topics
FieldArray
If the quantity of array elements is dynamic or an array consists of structures rather than single
elements, use a FieldArray, which is a type of field, like Bool16, Uint32Array, or any of the
other types in Schema.fh.
In the following example, the Bar implementation differs slightly from the previous example.
Instead of having three flags followed by four values, it has a value associated with each flag,
and the number of (flag, value) pairs is dynamic:
EXAMPLE 13
#define kBarPairs 1
#define kBarOption 2
#define kBarValue 3
resource Schema(3)
{
kBarImpl,
{1, 0},
{
{FieldArray{kBarPairs, {Uint16{0}}, {{Bool16{kBarOption, kFalse}},
{Uint32{kBarValue, 0}}}}}
}
};
The syntax looks slightly complicated because of ODFRC limitations, but it is fairly straightforward. The schema contains only one field; its type is FieldArray, and its name is kBarPairs.
Following the field's name is its iteration count. Because this is an attribute of the FieldArray, it
has a type but not a name. It also has a default value, which usually is zero. The counter might
be a signed or unsigned integer that is 8, 16, or 32 bits wide.
NOTE:
The iteration count immediately precedes the iterated values. (If your implementation's
persistent data requires the count be elsewhere, you must write your own conversion
provider.)
Following the iteration count is the list of iterated fields. Each has a type, name, and default
value, like any other field. The default value is not used unless the iteration count has a nonzero
default. In the preceding example, the output stream would simply be “0.” If the default iteration count were 2, the output would be “2, kFalse, 0, kFalse, 0.”
FieldArrays can be nested up to three levels deep.
Conditional-field inclusion
An important variant of the FieldArray type is the FieldIf type. Use this construct to include a
block of fields zero times if a condition is not met or once if the condition is met. The conditional value can be of type Bool8, Bool16, ClassID, or ImplementationID. If the conditional
value is of type ClassID or ImplementationID, the fields are included if the ID is valid.
60
Persistent Data and Data Conversion
Advanced schema topics
Example 14 is a slight variant of Example 13:
EXAMPLE 14
resource Schema(4)
{
kBarImpl,
{1, 0},
{
{FieldIf{kBarPairs, {Bool16{kFalse}}, {{Bool16{kBarOption, kFalse}},
{Uint32{kBarValue, 0}}}}}
}
};
In this case, kBarOption and kBarValue are included zero or one times, depending on the
Bool16 value.
FieldIf constructs can be nested up to three levels deep.
Persistent Data and Data Conversion
61
Persistent Data and Data Conversion
Advanced schema topics
62
Commands
Concepts
Commands
This chapter describes InDesign’s command architecture. It also outlines how to find and use
the commands provided by the InDesign API, how to implement new commands of your own,
and other extension patterns associated with commands.
Concepts
This section introduces the generic command pattern, databases that provide persistence to the
application’s objects, and models that represent the application’s objects in memory.
Command pattern
The intent of the command pattern is as follows: “Encapsulate a request as an object, thereby
letting you parameterize clients with different requests, queue or log requests and support
undoable operations.” (This is described in Gamma, Helm, Johnson, and Vlissides, Design Patterns, Addison-Wesley, 1995.) The structure of this pattern is shown in Figure 4.
FIGURE 4
Client
Command Pattern Structure
Invoker
Command
+ Execute() : void
Receiver
+ Action() : void
ConcreteCommand
+ Execute() : void
receiver->Action();
This pattern decouples the object that invokes an operation from the object that implements it.
The client wants an operation to be performed, the receiver knows how to perform the operation, and the two are decoupled by this pattern. The command declares an interface for executing an operation. The client does not send messages directly to the receiver; rather, the client
creates a concrete command object and sets its receiver. The concrete command implements
execute by calling the corresponding operation on the receiver. The invoker asks the command
to carry out the request, which causes the receiver to perform the operation.
Within the InDesign API, an implementation of the command pattern is used in situations
where preventing data corruption is paramount and users must be able to undo or redo
changes. See “Commands” on page 68.
Commands
63
Commands
Concepts
Databases and undoability
A persistent object has its state maintained outside the application runtime. It can be removed
from main memory and returned again, unchanged. The application maintains persistent
objects in a database. A database on disk is an opaque binary file in the operating system’s file
system. This database file stores the serialized state of a collection of objects. The set of objects
stored in a particular database is collectively known as a model. For example, an InDesign document file like Untitled-1.indd is a database file that represents the state of an instance of a document model. See “Models” on page 64.
Objects are stored in a database in the following format:
z
Objects that persist are stored as a tree of objects, each of which is associated with and persisted by a database.
z
Each persistent object has a identifier, the UID, that identifies it uniquely within its database.
z
Each persistent object (except the root object) is owned and referred to by another object,
the parent. A persistent object also can be referenced by other objects, though this does not
indicate any form of ownership.
The class that represents a database is IDataBase.
For more detail, see the “Persistent Data and Data Conversion” chapter.
Databases must be consistent and stable. Some databases also need to support undoability—the
ability for a user to undo or redo changes. For example, the ability to undo changes made to a
document is required; however, there is no requirement to undo changes made to the database
that persists the user interface state.
Table 6 shows the databases that exist and their support for undoability. If a database supports
undoability, that support cannot be turned off. All changes to objects that persist in the database must be undo-able. Turning off undo support is not allowed, because it would disable
error handling and increase the risk of document corruption.
NOTE:
To change objects that persist in a database that supports undo, we recommend using
commands. If, however, you need to change data without showing something in the Edit >
Undo menu, you can bypass commands and wrap your changes in calls to
CmdUtils::BeginAutoUndoSequence and CmdUtils::EndAutoUndoSequence.
Models
A model is a collection of objects backed by a database for persistent storage. A model is a treestructured graph of objects. The ownership relationships between objects in a model defines
this tree structure. Ownership relationships are just parent-child relationships within the tree.
64
Commands
Concepts
Document model
Documents are represented by the document model. Figure 5 shows an example.
FIGURE 5
Objects in a document model (instance diagram)
Document content
such as layout and
text.
UID 60 :kSpreadBoss
UID 1 :kDocBoss
ISpreadList
IStoryList
ISpread
UID 41 :
kTextStoryBoss
Document
preference settings
such as styles and
swatches.
IDocument
IDocument::GetDocWorkSpace
ITextModel
UID 75 :
IWorkspace
kDocWorkspaceBoss
IStyleGroupManager
UID 125 :kStyleBoss
IStyleInfo
ISwatchList
UID 91 :
kPMColorBoss
IColorData
ITextAttributes
The document (kDocBoss) is the root object in the document model. It owns a collection of
spreads (kSpreadBoss), stories (kTextStoryBoss), a workspace (kDocWorkspaceBoss), and so
on. These objects may own further objects specific to their domain. The document’s workspace
owns objects like styles and swatches, which can be used throughout the document.
The database file that provides persistence for a document model is an end-user document file;
for example, a file Untitled-1.indd saved from InDesign.
Commands
65
Commands
Concepts
Defaults model
Global preference settings are represented by the defaults model. Figure 6 shows an example.
FIGURE 6
Objects in the defaults model (instance diagram)
Global default preference
settings such as styles and
swatches.
UID 1 :
kWorkspaceBoss
IStyleGroupManager
UID 27 :kStyleBoss
IStyleInfo
IWorkspace
ISwatchList
UID 42 :
kPMColorBoss
IColorData
ITextAttributes
The workspace (kWorkspaceBoss) is the root object in a defaults model. It owns the objects
that represent default preference settings like styles, swatches and so on. The defaults model is
global and is accessed from the session (kSessionBoss) via the ISession interface. When a new
document is created, preference settings can be copied from the defaults model into the document’s workspace (kDocWorkspaceBoss).
The database file that provides persistence for this model is an application defaults file. For
example, the file named InDesign Defaults is the database file used by InDesign to persist the
defaults model.
66
Commands
Concepts
Session model
A session of an application is represented by the session model. Figure 7 shows an example.
FIGURE 7
Objects in the session model (instance diagram)
UID 1 :kSessionBoss
ISession
UID 2 :kAppBoss
IApplication
UID 3 :
kActionManagerBoss
UID 4 :
kPanelManagerBoss
UID 5 :
kToolManagerBoss
UID 6 :
kDocumentListBoss
The session (kSessionBoss) is the root object in the session model. This model owns application-related objects that must persist from session to session. For instance, user interface
objects are found here, together with other objects that store the application’s type system (the
boss classes provided by all registered plug-ins).
The database file that persists objects in this model is the application’s saved data file. For
example, the file named InDesign SavedData is the database file used by InDesign to persist the
session model.
NOTE:
Objects in the session model can be modified by calling mutator methods on their
interfaces directly. There is no need to modify them using commands. The database that
persists this model does not support undo.
Books, asset libraries, and other models
Several other features in an application have their own dedicated model, together with a database that provides that model with persistence. Prominent examples are books, asset libraries,
and the scrap. See Table 6 for the level of undo support provided by their databases.
Commands
67
Commands
Commands
Commands
This section describes how commands are used in the application. It includes sections on the
CmdUtils class, which is fundamental to the processing of commands, and how to use command sequences to group multiple commands into one logical unit.
Within the application, the most prevalent use of commands by client code is to modify persistent objects in the document model or defaults model. For example, plug-ins use commands to
create frames in a document or modify the default text style.
Commands encapsulate the changes made to persistent objects into a database transaction.
Encapsulating changes in this way helps prevent corruption. Furthermore, any change to the
state of persistent objects that needs to support undo must be made using a command.
Commands change persistent objects by doing the following:
z
Calling mutator methods on interfaces that exist on persistent objects.
z
Processing other commands (that call mutator methods on interfaces that exist on persistent objects).
z
Calling utilities that process commands (that call mutator methods on interfaces that exist
on persistent objects).
z
A mixture of the above.
A command is processed by the application; this means an instance of the command is passed
to the application to be executed. A command can change persistent data within only one database each time it is processed. The processing of a command moves the database from one consistent state to another.
When a command is used to modify objects that persist in a database that supports undo, that
command can manifest on the undo and redo menu items, and the database automatically
reverts the state of affected objects on undo and restore the changed state on redo.
NOTE:
Before InDesign CS4, commands were directly responsible for reverting or restoring
the changes they made to persistent objects on undo and redo. The application has
taken over this responsibility.
The InDesign API provides commands for plug-ins to use. See “Command processing and the
CmdUtils class” on page 70 and “Key client APIs” on page 84.
Plug-ins also can introduce new commands using the command extension pattern. See “Command” on page 86. This is required when a plug-in adds custom persistent data to a document,
defaults, or other objects that persist in a database that supports undo. The extension patterns
named persistent interface and persistent boss are ways in which custom persistent data can be
added to a database. See “Persistent interface” on page 89 and “Persistent boss” on page 91.
68
Commands
Commands
Command parameters
Commands implement the protocol used to modify objects that persist in a database that supports undo. A client initiates this protocol by instantiating a command and passing it off to the
application for processing. The client also is responsible for initializing the state of the command, including the setting of parameters. In general, the command pattern supports two
mechanisms for parameter passing:
1. The UIDList in the ICommand interface (the “item list”), which identifies a set of persistent
objects; see ICommand::SetItemList. On input, this might identify the set of page items on
which the command operates. For example, a page item move command (kMoveAbsoluteCmdBoss) would use this to identify the set of items to be moved. On output, the item
list might identify the set of object manipulated by a command. For example, a page item
create command (kNewPageItemCmdBoss) would provide the UIDs of the objects created
as a result of the command being processed. The use of the command’s item list is specific to
each command; a command is not required to use the item list.
2. Data-carrying interfaces on the command’s boss class. For example, the command used to
apply a particular text style aggregates an IBoolData interface to define whether character
styles should be overridden, and an IRangeData interface to indicate the range of text that
should be updated by the command.
These two approaches for passing parameters are shown in Figure 8. There is a command boss
class showing two aggregated interfaces, ICommand and a data-carrying interface (ISomeData). Parameters can be passed through the item list in ICommand, the data-carrying interface, or both.
FIGURE 8
Passing parameters to commands
«boss»
kSomeCommandBoss
«interface»
ICommand
+ GetItemList() : const UIDList*
+ SetItemList(const UIDList&) : void
Commands
«interface»
ISomeData
+ GetData()
+ SetData()
69
Commands
Commands
Command undoability
Each command has a property called undoability (see ICommand::Undoability), which determines whether the command name can appear in undo and redo menu items (i.e., whether the
changes made by the command can be undone or redone in a distinct step).
z
An undoability of kRegularUndo is used by commands whose changes need to appear as a
separate named step in undo and redo menu items. This is the default behavior.
z
An undoability of kAutoUndo is used by commands whose changes do not appear as a separate named step; for example, commands whose changes should be undone or redone with
an existing step.
Changes made to objects that persist in a database that supports undo must be undoable. To
achieve this, the various extension patterns involved, such as commands and persistent interfaces, must be implemented following the rules described in this chapter. The database then
automatically reverts the state of affected objects on undo and restores the changed state on
redo. The undoability of a command (kRegularUndo / kAutoUndo) has no effect on this behavior.
It affects only whether the command appears as a distinct step in undo and redo menu items.
Command processing and the CmdUtils class
Client code uses commands when changing objects that persist state in any database that supports undo. For example, commands are used to change objects in the document model or the
defaults model. The level of support for undo of each database is given in Table 6.
To change the objects, client code creates instances of one or more commands and submits
these instances to the application for execution. For example, to create frames in a document, a
plug-in either creates commands to be processed and passes them to the application or calls a
helper class that encapsulates both the creation and request for processing. The helper classes
provided by the InDesign API are introduced in “Command facades and utilities” on page 84.
Client code that must create commands uses the CmdUtils class. CmdUtils::CreateCommand
creates an instance of a specific command. The client code parameterizes the command, using
the command’s item list and data interfaces. The client code submits the command to the application for execution, by calling CmdUtils::ProcessCommand. Client code can group the commands it processes into command sequences, using methods and classes provided by
CmdUtils. Guidance on processing commands and command sequences is in “Command-processing APIs” on page 86 and the “Commands” chapter of Adobe InDesign CS4 Solutions.
The sequence of calls involved in processing a command is shown in Figure 9. The client code
creates the command, populates the item list and other data interfaces, then passes the command off for processing.
70
Commands
Commands
FIGURE 9
Client code
Client processing a command
CmdUtils
A Command
A Persistent
Interface
CreateCommand
Some client code creates an
instance of a command using
CmdUtils::CreateCommand
construct
SetItemList
The client sets the command
parameters, both through the
item list (shown here) or
through data carrying
interfaces.
ProcessCommand
Do
Set
PreDirty
The processing of the
command is initiated through
the client calling
CmdUtils::ProcessCommand.
Client code
creates
command, sets
parameters and
passes it for
processing. The
client should
also test error
state on return,
before
processing
further
commands.
CmdUtils should
be used for all
client
interactions with
the commands
sub-system. The
rest of the
internals are not
shown in this
diagram.
The update to some
interface on a (UID
based) boss object.
The call to PreDirty
allows the
application to track
the changes made
so that the
application can
undo or redo the
changes
automatically.
Command sequences
A command sequence (see ICommandSequence) groups a set of commands into a single
undoable step that changes the model from one consistent state into another. The command
sequence manifests on the undo and redo menu items as a single item.
Commands
71
Commands
Commands
Command sequences can be nested. For example, say you begin a sequence in one method,
then call a second method that begins a second sequence. The second sequence is said to be
nested within the first. When sequences are nested, only one sequence—the outermost one—
appears on undo and redo menu items.
Command sequences assimilate all commands are processed within their scope, whether the
command is processed directly from within the sequence or indirectly through subcommands
(commands processed by a command). See the “Commands” chapter of Adobe InDesign CS4
Solutions for how to write code that uses command sequences.
A typical scenario for a command sequence is shown in Figure 10. Between the BeginCommandSequence and EndCommandSequence calls, all commands processed act as a single set of
changes; only one element will appear on undo and redo menu items.
FIGURE 10
Calls made for a command sequence
Client code
CmdUtils
BeginCommandSequence
ICommandSequence*=
SetName
ICommandSequence
The client creates a new
command sequence.
The client gives the sequence
a name, otherwise the
sequence uses the name of
the first command processed.
ProcessCommand
ProcessCommand
Multiple commands are
processed within the scope of
a sequence.
EndCommandSequence
The client indicates the
sequence is complete.
A command sequence can change persistent objects in more that one database. (A command,
on the other hand, can change objects only in one database each time it is processed.) Operations that change objects in two or more documents can be implemented using a command
sequence. Such a set of modifications manifests in undo and redo menu items for all affected
documents.
72
Commands
Command managers, databases, and undo support
A command sequence has limited support for error handling: the sequence either succeeds
entirely or fails. An abortable command sequence (see IAbortableCmdSeq) allows more
sophisticated error handling, to allow for fail/retry semantics. Abortable command sequences
incur a significant performance overhead and should be used only where absolutely necessary.
Guidance on using these types of sequence is in the “Commands” chapter of Adobe InDesign
CS4 Solutions.
Command managers, databases, and undo support
This section describes how command managers relate to the application’s databases.
The application object model realizes a tree-structured graph of boss objects. Within this tree
are several distinct models, notably the session model, the defaults model, and one or more
document models (see “Models” on page 64). Each model has a distinct database that provides
persistence for its objects (see “Databases and undoability” on page 64). Each database has an
associated boss, called a command manager. When commands are used to change objects in a
model, the command manager is responsible for executing the commands.
Figure 11 shows that the application object model has several distinct databases that provide
persistence for its objects. The diagram shows some of the objects in an InDesign session (with
two open documents) and the databases with which these objects are associated. Each UIDbased object refers to its associated database through its IPMPersist interface. The defaults
model, represented by an instance of kWorkspaceBoss, persists in the InDesign Defaults database file. Each document model, represented by an instance of kDocBoss, persists in an InDesign document database file. The session model, represented by the instance of kSessionBoss
and the child objects that do not belong to any other model, persists in the InDesign SavedData
database file.
Each database has an associated command manager (see objects in Figure 11 that have an
ICommandMgr interface). The command manager is responsible for managing database
transactions and executing the commands that change the objects that persist in the database.
If an object has an ICommandMgr interface, it is the root object of its associated database. This is
shown in Figure 11. Key objects that are command managers are listed in Table 6; for the complete list, see ICommandMgr in the API Reference.
NOTE:
Commands
The class that represents a database is IDataBase. This class is an abstract C++ class, but
it is not an IPMUnknown interface.
73
Commands
Command managers, databases, and undo support
FIGURE 11
Models and databases in an indesign session (object diagram)
UID 1 :kSessionBoss
ICommandMgr
ISession::QueryApplication
UID 2 :kAppBoss
ISession::QueryWorkspace
IApplication::QueryDocumentList
UID 3 :kDocumentListBoss
UID 1 :kWorkspaceBoss
ICommandMgr
IDocumentList::GetNthDoc(0)
IDocumentList::GetNthDoc(1)
IPMPersist
UID 1 :kDocBoss
IPMPersist
ICommandMgr
IPMPersist
InDesign SavedData :
IDataBase
74
IPMPersist
Untitled-1.indd :
IDataBase
UID 1 :kDocBoss
ICommandMgr
IPMPersist
Untitled-2.indd :
IDataBase
IPMPersist
InDesign Defaults :
IDataBase
Commands
Command managers, databases, and undo support
TABLE 6 Frequently Used Databases and their Support for Undoability
Command manager /
root boss class
Database
Undo
support
Example filename and use
kSessionBoss
Application SavedData
None
InDesign SavedData User interface objects like
tools, menus, panels, dialogs, controls, and
string translations. Objects that define the
application object model, like plug-ins and boss
classes.
kWorkspaceBoss
Application Defaults
Full
InDesign Defaults Default resources like
swatches, fonts, and styles inherited by new
documents.
kDocBoss
InDesign Document
Full
Your.indd Document content like spreads,
pages, page items, and text.
kBookBoss
InDesign Book
Partial
Your.indb A collection of InDesign documents
and associated resources.
kCatalogBoss
InDesign Asset Library
Partial
Your.indl A collection of page items.
Clipboard/scrap
Application ClipboardScrap
Partial
Each database has a level of support for undo which is fixed when the database is created and
can be one of the following:
z
Full — The database fully supports undoable operations. The changes made by a transaction can be reversed automatically on undo and restored on redo. Undo and redo menu
items are fully supported.
z
Partial — The database can undo only the most recent operation. If an error occurs, the
database is rolled back to its state before the transaction began; however, once a transaction
is committed, it is irreversible. There is no support for undo and redo menu items.
z
None — The database does not support undoable operations. There is no support for error
handling, abortable command sequences, or undo and redo menu items.
NOTE:
Undo support varies by product. In InDesign and InCopy, full undo support is
provided for documents and defaults. In InDesign Server, only partial undo support is
provided for these databases; the server has no concept of user undo and redo.
Objects that persist in a database with full or partial support for undo must be modified using
commands. For example, a plug-in must process commands to change document content such
as spreads, pages, or page items.
Objects that persist in a database without support for undo can be modified by calling mutator
methods on persistent interfaces directly. Commands are not required. For example, a plug-in
can call interfaces on user interface objects, such as widgets, that set state directly.
Commands
75
Commands
The command processor
The command processor
This section describes how a command is passed through the application until it is processed.
The session (see the ISession interface) has a singleton instance of the command processor (see
kCommandProcessorBoss and the ICommandProcessor interface). The command processor is
the core application component to which all commands are submitted via CmdUtils.
The command processor’s structure is shown in Figure 12. A command targets one or more
objects for change, and these objects persist in a particular database. Normally, client code creates a command instance and sets this target by passing parameters to the command (see
“Command parameters” on page 69). The client code submits the command instance for processing using CmdUtils. Figure 13 shows the internal collaboration between the objects
involved.
FIGURE 12
Command processor (class diagram)
«boss»
kSessionBoss
1
ISession::QueryCommandProcessor
1
«boss»
kCommandProcessorBoss ICommandProcessor
1
1..*
«boss»
kSomeCommandBoss
ICommand
«boss»
kSomeModelRootBoss
IDataBase::GetRootUID
1
IPMPersist::GetDataBase
1
ICommandMgr
changes objects that persist in
IDataBase
The command processor examines the command and determines which database contains the
objects the command changes. Each database has an associated command manager boss (see
ICommandMgr interface), shown in Figure 12 as the boss class named kSomeModelRootBoss.
For example, the root of the document model is kDocBoss, and kDocBoss is a command manager. The command processor calls the associated command manager to execute the command.
76
Commands
The command processor
Figure 13 shows the internal processing of a command. The call to process a command is
passed to the session-wide command processor. This determines which database is targeted by
the command, and the command is passed to the command manager for the specified database. The command manager processes the command, and the command updates an interface
on a persistent (UID based) boss object. On completion, control returns to the client that initiated the process.
FIGURE 13
CmdUtils
Internals of command processing (sequence diagram)
CommandProcessor
A Command
Manager
A Command
A Persistent
Interface
ProcessCommand
ProcessCommand
Do
DoImmediate
Set
PreDirty
The command is
passed to the
session wide
command
processor which
determines the
target database
and calls
ICmdManager::Do
for that database.
NOTE:
Commands
The command
manager
invokes the
command.
The command
updates the
persistent
interface.
Generally the
command will
derive from the
Command class,
providing an
implementation
of Do().
The persistent
interface calls
PreDirty before
making any
update to the
data.
Third party plug-ins should not interact directly with the command processor or
command manager interfaces. CmdUtils provides all methods needed by third parties
for processing commands.
77
Commands
Scheduled commands
Scheduled commands
Commands can be scheduled for later processing, using CmdUtils::ScheduleCommand. The
command is placed in a queue and processed when the application is idle as part of the main
event loop, before idle tasks are processed. The sequence of calls involved in processing a
scheduled command is shown in Figure 14. Compare this with the more normal case of immediately processing a command, shown in Figure 13. In Figure 14, the scheduled command is
passed to the command processor (through the CmdUtils::ScheduleCommand call). At some
later time, after all other processing is complete and control is returned to the main application
event loop, all scheduled commands are processed.
Guidance on using scheduled commands is in the “Commands” chapter of Adobe InDesign CS4
Solutions.
FIGURE 14
CmdUtils
Processing a scheduled command
CommandProcessor
A CmdManager
A Command
A Persistent
Interface
ScheduleCommand
ProcessScheduledCmds
ProcessCommand
Do
DoImmediate
Set
PreDirty
Some client code schedules the
command. The command is
processed as part of the main
event loop after all other
processing has completed, and
before idle tasks are processed.
78
Processing of a scheduled command is identical to that of a non-scheduled
command. The only difference is with a non-scheduled command, client code is
responsible for the initiation, upon completion, control returns to the client. With
a scheduled command the event loop initiates processing, upon completion
control returns to the main event loop. If a scheduled command returns an error,
it is reported to the user, the error condition is cleared and processing resumes.
Commands
Snapshots and interface implementation types
Snapshots and interface implementation types
A snapshot is a copy of the state of an interface at a particular time. Undo and redo are achieved
automatically by the application, which keeps snapshots of the interfaces changed within an
undoable transaction. The snapshots are kept in the command history (see “Command history” on page 80) of the database with which the interface is associated.
The way in which an IPMUnknown interface is implemented determines whether its data is
persistent and/or needs a snapshot to support undo or redo:
Commands
z
Regular interfaces store transient data that is not maintained on undo or redo. They are
declared to the object model using CREATE_PMINTERFACE, and their state is lost if they
are removed from memory. Data in a regular interface is not persistent, and no snapshot is
taken.
z
Persistent
interfaces
are
declared
to
the
object
model
using
the
CREATE_PERSIST_PMINTERFACE macro. They serialize their state to their associated
database via the ReadWrite method and can be removed from memory and returned again
unchanged. Persistent interfaces that are part of a database that supports undo also are
called to serialize their state (again via their ReadWrite method) to a snapshot. This mechanism is required to support undo and redo. See “Persistent interface” on page 89.
z
Snapshot interfaces store transient data that must be maintained on undo and redo. They
are declared to the object model using the CREATE_SNAPSHOT_PMINTERFACE macro,
and they serialize their state in a snapshot via the SnapshotReadWrite method. Snapshot
interfaces are aggregated on a boss that persists in a database that supports undo; however,
their data is not persistent. Data maintained by a snapshot interface is lost whenever the
associated database is closed or when the command history for the database is cleared. See
“Snapshot interface” on page 92.
z
Snapshot view interfaces are similar to snapshot interfaces, which also store transient data
that must be maintained on undo and redo. The distinction is that a snapshot view interface
depends on model objects that persist in another database. A snapshot view interface is part
of a user interface object and must explicitly identify the database containing the model of
interest. Also, the database on which a snapshot view interface depends can change. For
example, a widget that tracks some state in the front document must change the database on
which its snapshot view interface depends when the front document changes. Snapshot
view
interfaces
are
declared
to
the
object
model
using
the
CREATE_VIEW_PMINTERFACE macro. See “Snapshot view interface” on page 99.
z
It also is possible to create a hybrid interface to allow the data that persists in the database to
differ from the data of which a snapshot is taken for undo and redo. This can be used by
complex data structures, like collections that need to optimize performance. Such an interface
is
declared
to
the
object
model
using
the
CREATE_PERSIST_SNAPSHOT_PMINTERFACE macro. Data is serialized to the database via the ReadWrite method and to the snapshot via the SnapshotReadWrite method.
For example, consider this approach if you have a collection of some sort and want to serialize only those objects that changed relative to the snapshot (rather than the entire collection).
79
Commands
Command history
Command history
Consider a database to be the state of the persistent object model. For example, a document
database is the state of the document object model (kDocBoss and all its dependents) at a particular time. Persistent interfaces (on dependent objects) are called to serialize their state to the
database when necessary, via their ReadWrite method.
The command history (ICmdHistory) provides a history of the undoable operations that were
performed on each database. It records the state changes made by a particular command or
command sequence, for the purposes of undo and redo. The name of the first command or
command sequence processed to perform an operation on the model appears as a named step
in the command history. These steps manifest as undo and redo menu items. The command
history is used to automatically revert the state of affected objects on undo and to restore the
changed state on redo.
The database and its command history are related but distinct states.
To maintain the command history, the database monitors which of its objects change when
commands are processed. It tracks the UIDs that are deleted, created, or un-deleted. It tracks
persistent interfaces, snapshot interfaces and snapshot view interfaces that are modified, and it
takes snapshots of them. This information is kept in the command history as a set of CmdStackItem objects. (CmdStackItem is opaque on the SDK; third parties cannot access its data.)
When undo is invoked, the application automatically reverts the database state to its previous
revision, using this information. When redo is invoked, the application automatically restores
the database state to its next revision.
See Figure 15. The command processor (kCommandProcessorBoss) aggregates the command
history (interface ICmdHistory), which is responsible for maintaining the state changes made
by a particular command or command sequence (which manifests on the undo and redo
menu). This state, represented as CmdStackItems, is used to move the model to previous and
next states on undo and redo. The CmdStackItem encapsulates all data required for an undo
and redo that occurred while processing a command or command sequence.
80
Commands
Undo and redo
FIGURE 15
Command history
«boss»
kCommandProcessorBoss
The command
processor
aggregates a
command history
interface.
1
1
«interface»
ICmdHistory
1
The command history maintains a set of
CmdStackItems (opaque in the SDK). These items
represent the set of UIDs added, deleted and modified.
For the modified persistent interfaces, the changes are
captured through a call to ReadWrite.
Any UID based object that has an associated snapshot
interface that has been modified, will have its state
captured through a call to its SnapshotReadWrite.
0..*
CmdStackItem
On undo/redo, the state that is maintained in the
CmdStackItem (opaque in the SDK), which is used to
reset the UID based objects (through calls to
ReadWrite and SnapshotReadWrite).
Merging changes with an existing step in the command history
Sometimes it is desirable to extend the scope of an existing step shown in the undo/redo menu
to include functionality that occurs after the step finished. This can be done using either of the
following:
z
A command with undoability of kAutoUndo.
z
A command sequence with undoability of kAutoUndo.
Undo and redo
To enable a user to undo or redo a change, the objects changed must:
z
Persist in a database that supports undo (see Table 6).
z
Be modified by processing commands.
The application maintains the state required for undo/redo of changes made to persistent
objects, as commands are processed. Each database that supports undo has a command history
in which the necessary information is recorded, as described in “Command history” on
page 80. Each persistent interface on a UID-based boss object has its ReadWrite method called
to add the state to the command history (for undo and redo) and to the database (to persist its
Commands
81
Commands
Undo and redo
state). On undo, the ReadWrite method is called to reset the state back to that from before the
action, and on Redo, the ReadWrite method is called again to set the state back to that from the
initial action. The sequence of calls to a persistent interface is shown in Figure 17. Snapshot
interfaces are called on their SnapshotReadWrite method using a similar approach; see
Figure 18.
The command processor records in the command history the set of undoable operations it has
performed (see the ICmdHistory interface on kCommandProcessorBoss). The steps in the
command history appear in undo and redo menu items and are modelled internally using
CmdStackItems. (These are internal only. Discussion is provided here to give some notion of
how the command processor works.) This is shown in Figure 16.
FIGURE 16
Command history internals (instance diagram)
kCommandProcessorBoss
CmdHistory Internals
1
1
«interface»
ICmdHistory
DB1:
CmdStackItem1
DB2:
CmdStackItem1
DBN:
CmdStackItem1
DB1:
CmdStackItem2
DB2:
CmdStackItem2
DBN:
CmdStackItem2
DB1:
CmdStackItemN
DB2:
CmdStackItemN
DBN:
CmdStackItemN
The command history maintains a stack of CmdStackItem objects for each database that supports undo; see Figure 16. A CmdStackItem (opaque in the SDK) represents the set of changes
made by the sequence or command that manifests on the undo/redo menu. It contains data
gathered during a database transaction for the purposes of undo and redo. On undo/redo, the
“current” CmdStackItem is used to revert the state of the model. The CmdStackItem encapsulates all data required for an undo/redo that occurred while processing a command or command sequence. This includes all changes to persistent interface and snapshot interface data
that were pre-dirtied. The CmdStackItem also maintains any inval handler cookies (see “Inval
handler” on page 94) created as part of the sequence.
82
Commands
Notification within commands
Each step has a name of either a command or a command sequence that performed the operation. On undo, the application automatically reverts the database to its state before the operation was performed. On redo, the application automatically restores the database to its state
after the operation was performed.
Extension patterns involved in undo and redo
The following extension patterns are called at undo or redo:
z
Persistent interfaces; see “Persistent interface” on page 89.
z
Snapshot interfaces; see “Snapshot interface” on page 92 (introduced in InDesign CS4).
z
Snapshot view interfaces; see “Snapshot view interface” on page 99 (introduced in InDesign
CS4).
z
Inval cookies; see “Inval handler” on page 94 (introduced in InDesign CS4).
z
Observers, via their LazyUpdate method (introduced in InDesign CS4); see the “Notification” chapter.
Notification within commands
A command can initiate notification, so observers interested in changes to the objects it modifies are notified. See the “Notification” chapter.
Commands also can be processed within observers and responders. Take care with commands
that modify the model in this way. See the “Notification” chapter for further discussion of
model changes within the context of notification.
If you are implementing your own command, see “Command” on page 86 for additional information on notification in the command extension pattern.
Error handling
In response to a user gesture or another event, the application calls a plug-in to carry out an
action. For example, the user creates a text frame. In this case, the plug-in creates and processes
the commands that perform the action. On detecting an error, a plug-in normally sets the global error code (see ErrorUtils) to something other than kSuccess and returns. When control
returns to the application, the global error code is checked. If it is set, the database is reverted to
the state it had before the modifications began. Extension patterns that participated in the
transaction being rolled back are called as follows:
Commands
83
Commands
Key client APIs
1. Persistent interfaces, snapshot interfaces, and snapshot view interfaces modified by the
transaction are called to revert their state. See “Persistent interface” on page 89, “Snapshot
interface” on page 92, and “Snapshot view interface” on page 99.
2. Inval cookies created during the transaction are called to undo. See “Inval handler” on
page 94).
When the application returns to its main event loop, it informs the user through an alert, using
the string associated with the error code that is set (see “Error string service” on page 88). The
global error code is then cleared, and the application continues.
Plug-in code that processes commands or command sequences must check for errors. CmdUtils::ProcessCommand returns an error code that should be checked for kSuccess before continuing. Alternatively, you can check the global error code using ErrorUtils. Details are in the
“Commands” chapter of Adobe InDesign CS4 Solutions. If a plug-in tries to process a command
while the global error code is set, protective shutdown occurs to protect the integrity of the
document or defaults databases (see “Protective shutdown” on page 84).
Command implementation code also must check for errors. If a command encounters an error
condition within the scope of its Command::Do method, it should set the global error code
using ErrorUtils and return. Details on error handling is in the command extension pattern;
see “Command” on page 86.
Plug-in code that requires more sophisticated error handling, to allow for fail/retry semantics,
must use an abortable command sequence. For information on using abortable command
sequences, see the “Commands” chapter of Adobe InDesign CS4 Solutions.
Protective shutdown
Protective shutdown is a mechanism that helps prevent document corruption. Any attempt to
process a command when the global error code is set (to something other than kSuccess)
causes a protective shutdown. The application creates a log file describing the problem it
encountered and then exits.
Key client APIs
This section summarizes the APIs a plug-in can use to interact with commands.
Command facades and utilities
There are many command-related boss classes named k<whatever>CmdBoss, but in many
cases you do not need to instantiate these commands, as there are command facades and utilities in the API that encapsulate parameterizing and processing these commands.
Interfaces like those in Table 7 are examples of command facades and utilities. These encapsulate processing of many commands required by plug-in code. These interfaces are aggregated
84
Commands
Key client APIs
on kUtilsBoss; the smart pointer class Utils makes it straightforward to acquire and call methods on these interfaces. You can call their methods by writing code that follows this pattern:
Utils<IDocumentCommands>()->MethodName(...)
TABLE 7 Useful command facades and utilities
API
Used for manipulating...
IDocumentCommands
Documents.
IPathUtils
Frames and splines. See the “Layout Fundamentals” chapter
for other APIs that help with layout.
IGraphicAttributeUtils
Graphic attributes. See the “Graphics Fundamentals”
chapter for other APIs that help with graphics.
ITextModelCmds
Text. See the “Text Fundamentals” chapter for other APIs
that help with text.
ITableCommands
Tables. See the “Tables” chapter for other APIs that help
with tables.
IXMLUtils
XML. See the “XML Fundamentals” chapter for other APIs
that help with XML.
kUtilsBoss
See kUtilsBoss in the API Reference for a complete list of
command facades and utilities.
These utility classes abstract over low-level commands. Before using commands, always look
for such a utility to see if there is a method that serves your purpose on one of these interfaces.
By doing so, you avoid the increased chance of confusion and error that comes with processing
commands. Your use case may require you to process some commands, if the utilities do not
provide all of the functionality you require.
As used in the application, command “facade” and “utility” are just different names for the
same thing, a class that reduces and simplifies the amount of client code you need to write to
change objects in the model. They decouple client code from command creation, initialization,
and processing.
When implementing custom commands, it is advisable to provide your own command facade
or utility. If you want to share your class with other plug-ins, implement it as an add-in interface on kUtilsBoss; for example, see XDocBkFacade. If the class is used only by one plug-in, a
C++ class is sufficient; for example, see BPIHelper.
Commands
85
Commands
Extension patterns
Command-processing APIs
The classes used when writing client code to process commands are listed in Table 8.
TABLE 8 Useful classes for processing commands
API
Description
CmdUtils
Provides methods that create, process, or schedule commands and
manage command sequences.
PersistUtils
Provides methods like GetDataBase, GetUID, and GetUIDRef, which
are used to identify the objects to be operated on by a command.
ErrorUtils
Provides access to the global error code.
NOTE:
You must use CmdUtils to process commands. Do not use the ICommandProcessor or
ICommandMgr interfaces for this. Misuse of these interfaces can easily cause
document corruption.
For information on finding and processing the commands provided by the API, see the “Commands” chapter of Adobe InDesign CS4 Solutions.
Extension patterns
This section summarizes the mechanisms a plug-in can use to extend the command subsystem.
Command
Description
Suppose:
z
You added a new persistent boss to a document or a persistent interface to an object in a
document, and you need to set custom data in these objects.
z
The API does not provide a command that sets the model data you need to change.
z
You want do some processing, and you concluded that a command provides the best pattern
in which to encapsulate it.
Architecture
To implement a command, follow these steps:
86
z
Define a new boss class in your plug-in that aggregates IID_ICOMMAND. If you require
input parameters over and above the command’s item list, aggregate further data interfaces
to the boss through which the parameters can be passed.
z
Provide an implementation of ICommand using Command as a base class.
Commands
Extension patterns
z
Add client code that processes your new command. See the “Commands” chapter of Adobe
InDesign CS4 Solutions.
z
For more guidance, see the Command page in the API Reference and “Best Practice” below.
Best practices
Commands
z
The Do method contains the code that modifies the model. The model is changed by calling
mutator methods on model interfaces, processing commands, processing commands in a
command sequence, or calling utilities that process commands for you.
z
If you encounter an error condition within your Do method, set the global error code
(ErrorUtils::PMSetGlobalErrorCode) and return. The application is responsible for reverting the model back to its state before the Do method was called and informing the user of
the error.
z
If you need more sophisticated flow control that allows for fail/retry semantics, use an
abortable command sequence (see IAbortableCmdSeq).
z
The Do method can change objects in only one database each time it is called. To change
objects in more than one database in one undoable step, use a command sequence (ICommandSequence).
z
Normally, the database containing the objects to be changed is passed using the command’s
item list. Alternatively, a command can target a predetermined database, by calling Command::SetTarget in its constructor. It also can override Command::SetUpTarget and determine the database containing the objects to be changed when the command is processed.
z
Normally, the DoNotify method contains the code that initiates notification, should you
require it. Initiate notification by calling ISubject::ModelChange (not ISubject::Change).
Calling ISubject::ModelChange when model objects are changed broadcasts regular and
lazy notification (see the “Notification” chapter). The application calls DoNotify after the
Do method of the command is called; however, notification need not be restricted to this
method.
z
Notification can be performed for each object that was modified using the ISubject interface of affected objects. Notification also can be performed centrally. For example, many
commands that modify the document model notify change using the ISubject interface of
the document (kDocBoss). Sometimes commands use both these approaches. The benefit
of using a centralized approach is that an observer needs to attach to only one subject to
receive the notification.
z
If you need to pre-notify observers before the model is changed, call your DoNotify method
from your Do method before making any changes to the model.
z
Notification can result in the processing of further commands, which can result in global
error code being set. If further commands are processed, application shutdown occurs. If
multiple subjects are notified by a command, the global error code should be tested between
notifications. If the error code is set, the DoNotify method should return without calling
further subjects, or it should consume the error and reset the global error code.
z
The undoability of a command should be fixed at command construction (see ICommand::GetUndoability). Its default value is kRegularUndo. If a different undoability is
87
Commands
Extension patterns
required, it should be set in the command’s constructor. See ICommand::Undoability in the
API Reference.
z
A command must have a name (see Command::CreateName), if it has an undoability of
kRegularUndo and it returns kTrue for IsNameRequired. Such commands can appear in
undo and redo menu items. The name must have a translation, so it can be displayed in
undo and redo menu items in a localized form. Commands with an undoability of kRegularUndo that override IsNameRequired to return kFalse must pick up their name from a
subcommand. Some commands create and process other commands (subcommands) to
make their changes and want to pick up the name of the first subcommand called within
their scope instead of providing their own name.
z
Commands with undoability of kAutoUndo do not need a name, because their changes
merge with an existing step in the undo and redo menu items.
z
The destructor of a command normally is empty. Do not change the model within the
destructor.
See also
z
For documentation on commands: Command and ICommand in the API Reference.
z
For documentation on notification: ISubject and IObserver in the API Reference.
z
If you need error codes set by your command to map onto strings that describe the error
condition: “Error string service” on page 88.
z
For sample code: kBPISetDataCmdBoss, BPISetDataCmd, and BPIHelper from the
BasicPersistInterface plug-in.
Error string service
Description
Suppose error conditions can occur in your plug-in, and you need to inform the user of these
errors.
Architecture
Sometimes, error conditions can occur in your plug-in, often within commands or command
sequences. When control is returned to the application by your plug-in, and the global error
code is set to something other than kSuccess, the user is informed and the model is reverted
back to the last consistent state. An error string service provides the ability to map error codes
onto error strings. To handle this, follow these steps:
88
z
You Provide an error string service. This service allows ODFRez resources in your plug-in
to map error codes onto strings that describe the error. Detailed documentation on how to
define the error codes, resources, and implementations involved is the API Reference for
IErrorStringService.
z
On detecting the error condition, call ErrorUtils::PMSetGlobalErrorCode to set your error
and return control to the application. The application informs the user of the error.
Commands
Extension patterns
See also
For sample code: BPIErrorStringService from the BasicPersistInterface plug-in.
Persistent interface
Description
Suppose you need to add custom data to an existing object in the model and have that persist.
For example, you want frames in a document to have a set of custom properties that you control.
Architecture
The state of a persistent interface that is associated with a database that supports undo is captured by the application before it is changed; its state is reverted on undo and restored on redo.
The state is saved in the command history (see “Command history” on page 80) for undo and
redo, and in the database for persistence. Figure 17 shows the typical sequence of calls made to
modify persistent interface, and when that modification subsequently is undone or redone.
Commands
89
Commands
Extension patterns
FIGURE 17
Application core
Sequence of calls to a persistent interface
ICmdHistory
CmdUtils
A Command
A Persistent
Interface
ProcessCommand
DoImmediate
Set
PreDirty
ReadWrite
Push
ReadWrite
Undo
ReadWrite
Redo
ReadWrite
90
Some entity
processes a
command which
results in a
PreDirty persistent interface
being modified. The
persistent interface
calls PreDirty before
mutating the data,
which allows the
application core to
snapshot the state of
the interface for the
command history
before it is changed
(to allow the
modification to be
undone), and to
require it to be
Modify streamed to the
database afterwards.
The application calls on
the persistent interface to
stream its data to the
database.
On undo, the command
history reverts the
persistent interface to the
state it had before the
modification was made.
On redo, the command
history restores the
persistent interface to the
state it had after the
modification was made.
Commands
Extension patterns
To implement a persistent interface, follow these steps:
1. Aggregate your interface on the boss class in the model to which you want to add custom
data that persists. Use an add-in interface (an ODFRez AddIn statement) if you are adding
the interface to an existing boss class. To define a new type of persistent boss, see “Persistent
boss” on page 91.
2. Provide an abstract interface that uses IPMUnknown as a base class.
3. Provide an implementation of this abstract interface.
4. Use CREATE_PERSIST_PMINTERFACE to make your implementation class available to
the application (instead of CREATE_PMINTERFACE).
5. Define a ReadWrite method for your implementation class. It gets called to serialize your
data when needed.
6. Call PreDirty within mutator methods before changing member variables that need to be
serialized. This gives the application the ability to recognize that this interface is about to
change.
7. Call mutator methods to update the state of your interface. The application calls your ReadWrite method to serialize your data when needed.
NOTE:
If your interface persists in a document, you must consider what should happen when
users who do not have your plug-in open documents that contain your plug-in’s data.
See the section on missing plug-ins in the “Persistent Data and Data Conversion”
chapter.
See also
z
“Command” on page 86.
z
For sample code: IBPIData, BPIDataPersist and BPISetDataCmd from the BasicPersistInterface plug-in.
z
For documentation on persistence: the “Persistent Data and Data Conversion” chapter.
Persistent boss
Description
Suppose you need to add a new type of persistent object to the model. For example, you need to
add a list of custom style objects to defaults.
Architecture
To implement a persistent boss that has a UID, follow these steps:
1. Define a new boss class.
2. Aggregate IID_IPMPERSIST, and use the kPMPersistImpl implementation provided by the
API.
Commands
91
Commands
Extension patterns
3. Add persistent interfaces to the boss to store your custom data. See “Persistent interface” on
page 89.
4. Choose a boss class to own the instances of your new persistent boss. Add a persistent interface into the boss class you have chosen that stores the UIDs of the new persistent boss. For
example, to add custom styles to defaults, add this interface into kWorkspaceBoss.
NOTE:
If your boss persists within a database that supports undo, such as a document or
defaults, you must create, modify, and delete the persistent boss using commands.
Implement a create command to allocate a new UID for each new instance of the
peristent boss, and store this UID in an interface on boss that owns it. The delete
command deletes all child UIDs owned by the persistent boss, removes all references to
the persistent boss from the model, then deletes the UID.
See also
z
For sample code: kPstLstDataBoss and PstLstUIDList from the PersistentList plug-in.
z
For documentation on persistence: the “Persistent Data and Data Conversion” chapter.
z
“Command” on page 86.
Snapshot interface
Description
Suppose you have an object containing transient data, which depends on model objects that
persist in a database that supports undo, and:
z
Your object needs to be updated when the model objects are modified initially and on any
subsequent undo or redo.
z
You cannot use lazy notification to update your object, because it must be kept up to date
with changes to the model at all times.
z
Your object must be updated on undo or redo, before observers get called.
For example, the text state (see the ITextState interface) caches the text attributes that are
applied to the “text caret”; the next text insertion uses these attributes. This cache is implemented as a snapshot interface and maintained on undo and redo.
Architecture
The state of an snapshot interface is captured by the application before it is changed; its state is
reverted on undo and restored on redo. The state is saved in the command history (see “Command history” on page 80) for undo and redo. Figure 18 shows a typical sequence of calls for a
snapshot interface when it is modified, and when that modification is subsequently undone or
redone.
92
Commands
Extension patterns
FIGURE 18
Application core
Sequence of calls for a snapshot interface
ICmdHistory
An Observer
A Snapshot
Interface
Update
Set
PreDirty
PreDirty
SnapshotReadWrite
Push
Some change causes an observer to
be called which results in a snapshot
interface being modified. The
snapshot interface calls PreDirty
before mutating the data, which
allows the application core to
capture the state of the interface for
the command history before it is
changed (to allow the modification
to be undone).
Modify
Undo
Undo
SnapshotReadWrite
On undo, the command history is
invoked. This leads to a
SnapshotReadWrite call on the
snapshot interface, to revert the state
of the interface back to the was it
was before the modification.
Redo
Redo
SnapshotReadWrite
Similarly, on redo, the snapshot
interface has the SnapshotReadWrite
method called to restore the state to
the way it was after the modification.
Often, an observer or responder is used to modify a snapshot interface, as shown in Figure 18.
Before changing any data, the snapshot interface implementation calls PreDirty, indicating to
the application that the SnapshotReadWrite method should be called to capture the state of the
interface (this state is associated with the CmdStackItem for the current sequence)). On undo
Commands
93
Commands
Extension patterns
or redo, the SnapshotReadWrite method is invoked with the data associated with that particular sequence on the undo/redo menu.
To implement a snapshot interface, follow these steps:
1. Add your interface to a persistent boss class that is part of the model on whose state your
interface depends. Check that this class aggregates IPMPersist, since the boss must have a
UID to support a snapshot interface. For example, if your object depends on objects in a
document, you might choose kDocBoss.
2. Provide an abstract interface that uses IPMUnknown as a base class.
3. Provide an implementation of this abstract interface.
4. Use CREATE_SNAPSHOT_PMINTERFACE to make your implementation class available
to the application (instead of CREATE_PMINTERFACE).
5. Define a SnapshotReadWrite method for your implementation class. It gets called to serialize snapshots of your data when needed.
6. Call PreDirty within mutator methods before changing member variables you serialize in
SnapshotReadWrite.
7. When the model objects you depend on are changed, call mutator methods to modify the
state of your interface. You might need to track the change to the object using regular notification. The application takes a snapshot of your interface, reverts the state on undo, and
restores it on redo.
NOTE:
Even though this extension pattern is very similar in its implementation to a persistent
interface, its effect is very different. The information in your snapshot interface is
transient and not saved persistently with a document, defaults, or whatever other model
it is associated.
See also
z
For sample code: LnkWtchCache in the LinkWatcher plug-in and GTTxtEdtSnapshotInterface in the GoToLastTextEdit plug-in.
z
The “Notification” chapter.
Inval handler
Description
Suppose you have an object that depends on model objects that persist in a database that supports undo, and:
94
z
Your object must be updated at undo and at redo when the model objects it depends on
change.
z
You cannot use lazy notification (see the “Notification” chapter) to update your object,
because you need it to be kept up to date with changes to the model at all times.
Commands
Extension patterns
z
You cannot use a snapshot interface (see “Snapshot interface” on page 92) because you need
to do more than restore the state of an object within the application.
z
You need to program behavior that runs at undo and redo.
NOTE:
This extension pattern is very rare. This is an advanced pattern and should be used only
if absolutely necessary.
For example, an observer might watch a subject that can be deleted from the model. Typically,
such observers are attached to the subject when it is created and detached just before it gets
deleted. If an object is created, the observer is attached. If there is an undo at this point, the
observer should be detached; a subsequent redo should result in the observer being reattached. Similarly, if an object is deleted, the observer is detached. At this point, an undo
should result in the observer being attached; a subsequent redo should result in the observer
being detached.
Consider an observer that attaches to a spread (see kSpreadBoss). When a spread is created, the
observer gets attached to the new spread. If the creation of the spread is undone, this observer
must be detached before the undo takes place. If the creation of the spread is redone, the
observer must be re-attached after the redo takes place. The inval handler extension pattern
allows for this. The GoToLastTextExit plug-in provides sample code that shows how to use an
inval handler to manage the attachment and detachment of an observer that watches stories in
a document.
Architecture
Inval handlers allow plug-in code to be called at undo and redo. There are two parts to the
mechanism:
z
An inval handler (see IInvalHandler) that registers interest in a database that supports
undo.
z
An inval cookie that gets called on undo and redo. The inval cookie is created by the inval
handler at the end of a transaction that contained changes in which the plug-in was interested.
Inval handlers are used in situations where a plug-in needs to achieve more than the restoration
of the state (through persistent interfaces and snapshot interfaces), and the lazy notification
broadcast occurs too late to meet this need. The plug-in has code that must be called immediately at undo or redo.
To implement an inval handler, follow these steps:
Commands
z
Provide an implementation of an inval handler (see the IInvalHandler class in the API Reference for details. See GTTxtEdtInvalHandler for sample code).
z
Provide an implementation of an inval cookie (see the IInvalCookie class in the API Reference for details. See GTTxtEdtInvalCookie for sample code).
z
Create an instance of the inval handler, and call DBUtils::AttachInvalHandler to associate
this instance with the database that persists the objects in which you are interested. The
scope of this association can be defined by the lifetime of the database or the enabling of
particular functionality. For example, an inval handler might be attached to a database
when a document is opened.
95
Commands
Extension patterns
z
Once an inval handler is attached to a database, call DBUtils::StartCollectingInvals to indicate the inval handler is interested in participating in undo and redo. This usually occurs on
some event (through an observer or some other means). The StartCollectingInvals call
informs the database that the inval handler should be called at the end of the current transaction. If an undoable transaction is ongoing, the inval handler is said to be “collecting
invals.”
z
The plug-in code records the semantics of the change. Where this is recorded is implementation-dependent: it could be in the state associated with the inval handler (recommended),
an inval handler cookie, or elsewhere.
z
Further changes of interest can occur within the same transaction. The plug-in code should
record the semantics of the changes; for example, by adding state to a list within the inval
handler.
z
At the completion of the transaction, the IInvalHandler::EndCollectingInvals method is
called. The inval handler can return an instance of an inval cookie representing the
change(s) of interest that occurred. This inval cookie is kept in the command history for
this database transaction. The inval handler is now said to be “not collecting invals.”
z
The inval cookie is called on undo and redo of the transaction.
z
When the lifetime of an inval handler is over, call DBUtils::DetachInvalHandler to detach
the inval handler from the database. For example, an inval handler might be detached from
a document database when a document is closed.
Figure 19 shows the lifetime of a typical inval handler. Figure 20 shows the typical sequence of
calls made to an inval cookie on undo. On undo, any inval cookies that were previously associated with the CmdStackItem are called. The inval cookie setting the error state aborts the undo
(the model is reset to the state from before the undo).
96
Commands
Extension patterns
FIGURE 19
Typicalinval handler calls (sequence diagram)
Application core
ICmdHistory
Observer (or
other type of
object)
An Inval Handler
An Inval Cookie
Event that causes attach to the model
new
DBUtils::AttachInvalHandler
An inval handler is created and associated
with the database that persists the objects
of interest. For example, an inval handler
might be attached when a document is
opened.
Event in the model of interest for undo / redo
AddInvalInfo(info)
bool16= DBUtils::StartCollectingInvals
StartCollectingInvals is called to indicate
that the inval handler is interested in
participating in undo and redo of a
transaction. This usually occurs on some
event (through an observer, or some other
means).
More events of interest for undo / redo can occur within the same transaction. The inval handler must be called to
record each event. At the end of the transaction, EndCollectingInvals is called. The inval handler returns the cookie
to be called on undo and redo of the transaction. See sequence diagram showing inval cookie calls.
IInvalCookie*= EndCollectingInvals
new
AddInvalInfo(vector of info)
Push
New events in the model of interest for undo / redo cause the above sequence to be repeated and a new inval
cookie to be associated with the transaction.
Event that causes detach from the model
DBUtils::DetachInvalHandler
delete
Commands
When the lifetime of an inval handler is over,
DetachInvalHandler is called to detach the
inval handler from the database. For
example, an inval handler might be
detached when a document is closed.
97
Commands
Extension patterns
FIGURE 20
Application core
Inval cookie calls on undo
ICmdHistory
Model
An Inval Cookie
Undo
ErrorCode= InvalBeforeUndo
Prior to the model state being
reverted, all inval cookies are
called.
ReadWrite / SnapshotReadWrite
After all inval cookies are
called, the model is reverted
(assuming no error).
ErrorCode= InvalAfterUndo
After the model has been
reverted, all inval cookies are
called. Setting error here will
undo the "Undo".
There are times where the lifetime of an inval handler does not match that of the command history. Inval handlers can be attached and detached at any time, asynchronously with anything
else happening in the model. This means there are entries in the command history that do not
have corresponding inval cookies for a particular inval handler. In this situation, the inval handler’s BeforeRevert_InvalAll and AfterRevert_InvalAll methods are used to provide the opportunity to rebuild the required state directly.
There are times when an inval cookie instance can be asked to merge another instance. This
causes the IInvalCookie::Merge method to be called, to merge cookies that were returned by
each call to IInvalHander::EndCollectingInvals within one undoable transaction. For example,
this can happen at the end of an abortable command sequence.
See also
98
z
IInvalHandler, IInvalCookie, DBUtils::AttachInvalHandler, DBUtils::StartCollectingInvals,
and DBUtils::DetachInvalHandler in the API Reference.
z
For sample code: GTTxtEdtInvalHandler in the GoToLastTextEdit plug-in.
z
For documentation on how to detect change in other objects using observers and responders: the “Notification” chapter.
Commands
Extension patterns
Snapshot view interface
Description
Suppose you have a user interface object like a widget, which has data that depends on model
objects that persist in a database that supports undo, and:
z
Your data needs to be updated when the model objects it depends on are modified or when
modifications are undone or redone.
z
You cannot use lazy notification to update your data, because it must be kept up to date with
changes to the model at all times.
NOTE:
Use of this pattern is extremely rare in the application codebase. For example, it is used
by interface ILayoutControlData on the layout widget. The data in this interface is
depended on by many other objects and always must be kept in tight synchronzsation
with the database. This is an advanced pattern and should be used only if absolutely
necessary.
Architecture
A snapshot is taken of the state of a snapshot view interface, before it is changed, reverted on
undo, and restored on redo. It is very similar to a snapshot interface (see “Snapshot interface”
on page 92). The distinction is that the database with which a snapshot interface is associated is
fixed and implicitly defined by the persistent boss on which it is aggregated. A snapshot view
interface, on the other hand, is part of a user interface object and must explicitly identify the
database containing the model objects on which it depends. Also, it is not one specific database. For example, a widget that tracks some state in the front document must change the database on which its snapshot view interface depends, when the front document changes.
To implement a snapshot view interface, follow the steps described under CViewInterface in
the API Reference.
See also
The CViewInterface template class in the API Reference.
Commands
99
Commands
Extension patterns
100
Notification
Concepts
Notification
Notification patterns inform interested parties of certain changes. This chapter describes the
prevalent notification patterns used by products in the InDesign family.
Concepts
Notification is a mechanism by which an interested object can be made aware of changes in
other components or subsystems. This section focusses on two prevalent notification patterns,
the observer and the responder.
Observer pattern
The observer pattern is described in [Design Patterns] by Gamma et al. The intent of the pattern is to “Define a one to many dependency between objects so that when one object changes
state, all its dependents are notified and updated automatically.” Figure 21 shows the structure.
FIGURE 21
Abstract representation of the observer pattern
Subject
Observer
has observers
Attach(Observer) : void
Detach(Observer) : void 0..*
Notify() : void
Update() : void
0..*
Notify()
for all o in observers {
o->Update()
}
The pattern consists of a subject and one or more observers. A subject is an object that needs to
inform other objects that are interested in changes to the subject’s state. An observer is an object
that needs to keep its state consistent with the state of a subject. A subject can have an association with multiple observers. An observer can observe multiple subjects.
The association between observers and subjects is managed through Attach and Detach calls.
Any change to the subject is declared through a call to the Notify method, which in turn calls
Update on any associated observers.
The subject supports three operations:
Notification
z
Attach — Associates a specific observer with a subject.
z
Detach — Detaches a specific observer from a subject.
z
Notify — Notifies all associated observers of some change to the subject.
101
Notification
Observers
The observer supports one operation:
z
Update — Keeps the observer’s state consistent with the subject’s state.
Typically, when a client wants to be notified of changes in a subject, it calls Attach, passing an
observer that will be the recipient of update messages.
Any update to the subject that could be of interest to observers results in a call to the subject’s
Notify method. The subject, in turn, iterates through all attached observers, calling their
Update operations.
At any point, a client can remove the observer from the subject, using Detach. The observer no
longer receives update events from that subject.
For more information, see “Observers” on page 102.
Responder pattern
With the observer pattern, notification is provided when an object is modified. The responder
pattern provides notification on an event. Notification consists of a signal sent from one party
to a responder that can react to the event. The application predefines a set of events and corresponding signals in which responders can register interest. For example, signals are associated
with document events like create, open, save, and close. A responder that registers interest in
the document open event is called each time a document is opened. Figure 22 shows the structure.
FIGURE 22
Responder pattern
has responders
SignalManager
+ SignalResponder() : void
1
0..*
Responder
+ Respond() : void
The pattern consists of a signal manager and responders. The signal manager is the object
responsible for notifying responders of some event. A responder is an object that is called as a
result of that event. A responder registers its interest in a particular event using the service provider extension pattern. A responder is a service provider, although this is not shown explicitly
in Figure 22. For more information, see “Responders” on page 115.
Observers
This section describes the implementation of the observer pattern (see “Observer pattern” on
page 101) within the application. The implementation consists of two interfaces, a subject (see
the ISubject interface) and an observer (see the IObserver interface), as shown in Figure 23.
102
Notification
Observers
FIGURE 23
Subject and observer interfaces (class diagram)
«interface»
ISubject
+
+
+
+
+
+
+
+
0..*
AttachObserver() : void
Change() : void
DetachObserver() : void
ModelChange() : void
«interface»
IObserver
0..*
+
+
+
+
AutoAttach() : void
AutoDetach() : void
LazyUpdate() : void
Update() : void
«realize»
«realize»
CSubject
CObserver
AttachObserver() : void
Change() : void
DetachObserver() : void
ModelChange() : void
A default implementation,
kCSubjectImpl, is provided. It is
unlikely that you would need
to specialise it.
+ AutoAttach() : void
+ AutoDetach() : void
+ LazyUpdate() : void
SomeObserver
+ Update() : void
Subjects
A subject is an object potentially of interest to other objects. Within the application, any object
that maintains state of interest to other objects is a candidate for being a subject. An object with
an ISubject interface is a subject. The interface supports the following key operations:
z
AttachObserver — Associates a particular observer with the subject.
z
DetachObserver — Detached a particular observer from the subject.
z
Change — Notifies observers of a change to the subject. Observers get called on their IObserver::Update method. See “Regular and lazy notification” on page 106.
z
ModelChange — Notifies observers of a change to a subject. Observers get called on their
IObserver::Update method and/or their IObserver::LazyUpdate method. See “Regular and
lazy notification” on page 106.
Change manager
Subjects within the application defer the functionality required to track the subject and
observer dependencies to the change manager (the IChangeManager signature interface). The
ISubject interface forms a facade over the change manager. It is unlikely third-party developers
need to interact with the change manager directly. It is not considered further here.
Notification
103
Notification
Observers
Observers
An observer is an object interested in changes to a subject. Any object within the application
object model can be an observer by aggregating the IObserver interface. The interface supports
the following key operations:
z
AutoAttach and AutoDetach — Provided when the observer knows the subject with which
it should be associated. A client would then delegate the association of subject and observer
to the observer. See “Relating observers to subjects” on page 112 for a description of how to
associate an observer with a subject.
z
Update and LazyUpdate — Receives notification when a subject changes. An observer can
get called via its Update or LazyUpdate method, depending on how it chooses to receive
notifications from the subject. See “Regular and lazy notification” on page 106.
Message protocols
Normally, an observer attaching to a subject indicates a protocol of interest. This protocol
(identified by type PMIID) allows the observer to be called only for a subset of the total set of
messages being broadcast by the subject. Generally, the protocol is synonymous with a data
interface on the subject. Observers interested in changes to a specific data interface on the subject attach using the PMIID of the data interface.
Figure 24 shows a subject, kSomeSubjectBoss, that aggregates a data interface, ISomeData. The
observer, kSomeObserverBoss, is interested in changes to this data interface. The observer
attaches to the subject, specifying IID_ISOMEDATA as the protocol of interest. This corresponds to the PMIID of the data interface. After the data interface is modified, the ISubject::ModelChange or ISubject::Change method is called to broadcast notification. Observers
are called via their Update or LazyUpdate methods, with a message protocol of
IID_ISOMEDATA.
The object that actually modifies the data interface and calls the ISubject interface to broadcast
notification is not shown in Figure 24. If the subject object is part of a document or defaults,
the object that calls the mutator method on the data interface normally is a command (see the
“Commands” chapter). The command calls ISubject::ModelChange if it performs notification.
If the subject object is a user interface object, mutator methods on the data interface itself normally call ISubject::Change to broadcast notification.
104
Notification
Observers
FIGURE 24
A subject and an observer boss class (class diagram)
«boss»
kSomeBoss
«boss»
kSomeObserverBoss
«interface»
ISubject
«interface»
ISomeData
+ Get() : void
+ Set() : void
+
+
+
+
AttachObserver() : void
Change() : void
DetachObserver() : void
ModelChange() : void
0..*
«interface»
IObserver
0..*
+
+
+
+
AutoAttach() : void
AutoDetach() : void
LazyUpdate() : void
Update() : void
«realize»
«realize»
«realize»
SomeData
CSubject
CObserver
+ Get() : void
+ Set() : void
+
+
+
+
AttachObserver() : void
Change() : void
DetachObserver() : void
ModelChange() : void
+ AutoAttach() : void
+ AutoDetach() : void
+ LazyUpdate() : void
SomeObserver
+ Update() : void
Subject and observer types
Within the application, subjects are split into distinct “types.” The type of object a subject is
determines the method on ISubject that is called to perform notification. The types of subject
are as follows:
z
Model — An object that is part of the document model (see kDocBoss), the defaults model
(see kWorkspaceBoss), or another model that supports undo. The ISubject::ModelChange
method is used to broadcast notification when model objects change.
z
User interface — An object that is part of the user interface; for example a checkbox or button This is distinct from an element in the user interface that is interested in model data.
The ISubject::Change method is used to broadcast notification when user interface objects
change.
Within the application, observers are split into distinct “types.” The type of subject object an
observer is watching determines the type of observer. The types of observers are as follows:
Notification
105
Notification
Observers
z
Model — An observer attached to objects that persist in a document (see kDocBoss),
defaults (see kWorkspaceBoss), or another model that supports undo. Notification of model
observers is done through a call to ISubject::ModelChange.
z
User interface — An observer attached to a user interface object; for example a checkbox or
button.
z
Selection — An observer watching an artifact of the current application selection.
z
Context — An observer watching some context, such as the active document.
z
Hybrid — An observer watching multiple types of subjects. A particular observer may
attach to any set of objects; however we do not consider what requirements observers that
attach to multiple types of subjects might have.
Regular and lazy notification
The application supports two types of notification, regular and lazy. If an observer needs to be
called as soon as a subject broadcasts the change message, it should request regular notification.
If the observer can afford to wait until idle time before being notified that a change of interest
occurred, it can register for lazy notification. When an observer attaches to a subject, it defines
whether it wants to attach for regular notification, lazy notification, or both.
Regular notification
Regular notification is passed to an observer via the IObserver::Update method.
With regular notification, an observer is called within the scope of the call to a subject’s ISubject::Change (or ISubject::ModelChange) method. Figure 25 shows the typical sequence of calls
made when a subject notifies an observer using regular notification.
106
Notification
Observers
FIGURE 25
Regular notification
A Client Object
A Subject
A Data Interface
An Observer
Set
Change
Update
Get
A client object is called and modifies some data interface, then broadcasts notification about
the changes it makes. It mutates the state of a data interface and then calls ISubject::ModelChange. Observers get called via IObserver::Update and can retrieve the state of the data interface. The subject and the data interface are most often on the same boss object, but not always.
If a change is undone or redone, the observer is not notified. That is, the Update message is not
re-broadcast. See Figure 27 for the sequence of calls made on undo.
Lazy notification
Lazy notification is passed to observers via the IObserver::LazyUpdate method. Lazy notification is available only to observers of subject objects associated with persistent databases supporting undo.
Lazy notification is used when observers do not need to be in tight synchronization with
changes being made to the subject objects they observe. Rather than participating in each
change that occurs on a subject object, observers using lazy notification are notified after all
updates are made and the application is idle.
When the state of an object in the model is changed (for instance, by a command), and notification is required, the notification always is broadcast by calling ISubject::ModelChange. The
ISubject::ModelChange method queues a message to be broadcast later. Once the application is
idle, observers get called with this message via the IObserver::LazyUpdate method, as shown in
Figure 26.
Notification
107
Notification
Observers
FIGURE 26
Application Idle
Task
Lazy notification
A Command
A Model Subject
A Persistent
Interface
An Observer
Do
Set
DoNotify
ModelChange
Observers using lazy notification are notified later on when the application is idle.
LazyUpdate
Get
NOTE:
The approach above holds for subject objects in the document model, the defaults
model, or, in general, any subject object that persists in a database that supports undo.
Only subject objects which persist in a database that supports undo can use lazy
notification.
Regardless of the number of changes made to the persistent interface, only one LazyUpdate call
is made per sequence for a given message protocol.
An implication of this is that ordering of lazy notification with respect to other lazy notification
messages cannot be guaranteed. Lazy notification can be thought of as a message indicating a
change was applied to an object some time in the past. The observer is responsible for determining the nature of the change.
An observer that registers for lazy notification is not guaranteed to receive all messages. For
example, suppose an observer is attached to a page item and is watching for the page item to
move. If a sequence moves the page item and then deletes it, a lazy update message for the
move is not sent to the observer. If a subject object is deleted, any lazy notification messages
being queued for that object are purged.
We recommend any model observer used to keep some user interface component synchronized with the model should use lazy notification. This eliminates any unnecessary refresh of
108
Notification
Observers
user interface data. An example use of lazy notification is a user interface panel that reports
information about the set of text frames in a document. The user interface panel is associated
with a model observer watching for textual changes or the addition or removal of text frames.
These operations result in the panel being updated. There is no real requirement for the panel
to be updated in real time as these changes are made; the panel should be updated when the
text updates are complete (and the application is idle). Lazy notification enables this by sending
a single IObserver::LazyUpdate message to the observer when the move is complete.
If a change is undone or redone, the observer is notified; that is, the LazyUpdate message is rebroadcast. The sequence of calls made to an observer’s LazyUpdate method on undo is shown
in Figure 27.
FIGURE 27
Application
Core
Do
DoNotify
Sequence of calls made to an observer on undo
ICmdHistory
A Command
A Model
Subject
A Persistent
Interface
An Observer
Set
ModelChange
LazyUpdate
Get
The lazy update on do is discussed elsewhere
Undo
LazyUpdate
Get
On undo (and redo) the LazyUpdate method of the observer is
invoked. Note, the command is no longer in existence when
undo or redo is processed.
Notification
109
Notification
Observers
Both notifications
An observer can register for both regular and lazy notification. Figure 28 shows the sequence of
calls this might involve.
FIGURE 28
Application Idle
Task
Regular and lazy notification
A Command
A Model Subject
A Persistent
Interface
An Observer
Do
Set
DoNotify
ModelChange
Update
Get
Do
Set
DoNotify
ModelChange
Update
Get
LazyUpdate
Get
110
Notification
Observers
While every ISubject::ModelChange message sent to the subject is propagated to the observer
via IObserver::Update, only one IObserver::LazyUpdate call is made, regardless of the number
of changes made in a single transaction. For example, if multiple updates are made on an
object, the update messages being transmitted using a single protocol identifier, and the lazy
update method is called only once.
On undo or redo, only the LazyUpdate messages are re-broadcast.
Registering for both notifications allows an observer to have the rich context of the Update
message and have to re-build their view of the model only on rarer undo/redo events.
Lazy notification data
With regular notification, the observer (IObserver::Update) method is passed a context about
what caused the change (usually, the command that initiated the notification). With lazy notification (IObserver::LazyUpdate), the observer gets notified only that something changed. The
observer must be capable (for instance, by examining the model) of deducing or otherwise
rebuilding the set of information it requires. In some cases, the time this takes causes performance issues; therefore, the API provides lazy notification data as an optimization.
Lazy notification data objects are data-carrying (C++) objects created by the message originator. For example, a command that deletes a document page object (kPageBoss) might create a
lazy notification data object and populate it with the UID of the deleted page. Generally, this
lazy notification data object is passed to observers via IObserver::LazyUpdate. There are situations where the lazy notification data objects associated with a change are purged (for example,
in low memory condition), and the observer is called with a nil-pointer; therefore it is important that observers be designed to deal with this.
The lazy notification data objects used within the application are not documented in the public
API.
NOTE:
Use of lazy notification data objects is not pervasive in the application. They exist only
as an optimization, and the types of data they consist of varies. We recommend you
avoid using them if possible and refresh the observer’s state entirely when
IObserver::LazyUpdate is called.
Observers and undo
Only observers registered for lazy notification are signalled when an object they are associated
with is modified through an undo or redo action. For each undo or redo, the observer receives
one call (per message protocol), regardless of the number of changes made to the subject
object.
If the observer performs model modifications using commands (see “Model modifications” on
page 114), on undo or redo these are automatically undone or re-done as required. For observers watching the model to update some aspect of the user interface, lazy notification should
give the required update. If there are more exotic requirements for notification on undo and
redo, two other patterns exist: see the Snapshot interface and Inval Handler extension patterns
in the “Commands” chapter.
Notification
111
Notification
Observers
Relating observers to subjects
This section discusses how observers are attached to particular subjects.
Determining the subject
Before being able to attach and detach observers, an understanding of both the subject to
attach to and the protocol to listen along is required. The sample code provided as part of the
SDK has examples for many key events of interest. Similarly, Adobe InDesign CS4 Solutions has
advice on detecting when events occur for domains like layout, text, XML, and so on.
If neither of these resources provide the answer, the debug version of the application can be
used. Under the test menu is an item called Test > Spy. It can be configured to log all executing
commands, along with the subjects that are notified, the protocol used for notification, and the
change that occurs. Perform the action of interest, and Spy provides the information required
to observe the action.
Attaching and detaching subjects
In most situations, an observer understands the subjects in which it is interested and how to
attach to them. Such an observer attaches to the subjects in the IObserver::AutoAttach call and
detaches in IObserver::AutoDetach. For example, an observer on a panel can discover the widgets it contains (see the IPanelControlData interface) and attach to their subject interface.
NOTE:
The IObserver::AutoAttach and IObserver::AutoDetach methods still need to be called
by another object.
There are other situations where the observer does not know the subject(s) with which it
should be associated. Given a subject and an observer, a client object can directly attach or
detach the observer using ISubject::AttachObserver and ISubject::DetachObserver calls.
Using a responder to attach and detach an observer
Sometimes, an observer is required based on an action within the application. Responders (see
“Responders” on page 115) are called on different application events and can be used to manage the association between an observer and a subject. For example, a responder can be called
when documents are created, opened, or closed. On create or open, the responder can attach an
observer to the document. On close, the responder can detach the observer.
Using an observer to attach and detach another observer
Sometimes, an observer attaches and detaches a second observer based on notification received
by the first observer. This pattern commonly is seen in user interface objects that view objects
in the front most document. The active context (see kActiveContextBoss, the IActiveContext
interface, and the IID_IACTIVECONTEXT protocol) is the subject that is updated when the
user chooses the document on which to work. The user interface object has a first observer that
watches the active context. This observer attaches and detaches a second observer to the subjects of interest in the front-most document.
112
Notification
Observers
Observing a subject that can be deleted
Sometimes, a client object needs to observe a subject created as part of an undoable action. To
do this, a pattern involving two observers and an inval handler can be used (inval handlers are
described in the “Commands” chapter). The first observer watches for create events for the
subject and attaches the second observer to the newly created subject. To ensure this second
observer gets detached if the create action is undone, the first observer also must set up an
inval handler. The inval handler provides a mechanism to detach the observer on undo and
reattach it on redo.
Observing a subject that can be deleted as part of an undoable action is similar. On undo, the
observer must be reattached; on redo, it must be detached. Again, this requires the use of an
inval handler.
Document notification
The application supports a large object model. Any object is a candidate for becoming a subject
(and thousands of subjects exist). Management of observers interested in the set of messages
broadcast from a subject can be difficult, if many instances of a subject can exist. For example,
consider objects in the page item hierarchy (which can be deleted and recreated, undone and
redone, as a consequence of user actions). It is difficult and undesirable to maintain an
observer on each page item. To solve this problem, as well as the actual subject object broadcasting the change message, the document’s subject also broadcasts the change. This means any
observer associated with the ISubject interface on the document (kDocBoss) subject is notified, without the overhead of managing the association with finer-grained objects within the
document.
Notification of changes to page items are broadcast on the document subject using IPageItemUtils::NotifyDocumentObservers. Some other types of object use the ISubject interface of the
document (see kDocBoss) directly to notify of a change.
Observers and the model
This section details some considerations around the use of model observers within the application.
Observers “watch” model data for two primary reasons:
1. To update a user interface element to reflect the state of the model. (This type of observer is
distinct from user interface observers, which watch for changes in user interface elements
like checkboxes.)
2. To provide a point where existing functionality can be extended. For example, an observer
can be used to set custom data in a story object (kTextStoryBoss), on story creation. There
are side effects for both the error state and model modifications that should be considered
when using observers to extend the model.
Notification
113
Notification
Observers
Error state
Model observers that extend functionality can set the global error state either directly (using
ErrorUtils::PMSetGlobalErrorCode) or indirectly (through processing of commands). Generally, observers are notified during the processing of a command or sequence of commands. If
the application tries to process a command when the global error state indicates anything but
success (kSuccess), a protective shutdown is invoked to protect the integrity of open documents.
An observer setting the global error state either aborts the current change (that is, the current
change is undone) or causes the shutdown of the application. The behavior is determined on a
per-use basis and can be complicated by the use of commands in different scenarios. For example, the creation of a text story object (kTextStoryBoss) can occur through the type text tool or
from an import and place operation. The type text tool is straightforward, and an observer setting global error state causes the change to be undone. For import and place, however, the story
creation occurs as part of a longer sequence of commands. A similar observer setting global
error state causes an application shutdown as further commands are processed.
Unless it is explicitly suggested that it is appropriate to set error state in an observer of a particular subject change, an observer should not set error state. Also, an observer that uses objects
that potentially set error state (such as commands) should test for and consume any errors, taking appropriate remedial action.
Model modifications
A common pattern involves calling commands within an observer to make further modifications to the model. This can lead to problems. If the observer is notified as part of a sequence of
commands, and the commands that are due to be processed after the current notification phase
ends make any assumptions about the state of the model, instability and document corruption
can result.
For example, consider a sequence of two commands. The first inserts text into a text story
object (kTextStoryBoss). The second changes the color of the newly inserted text to blue.
Pseudo-code for this sequence is in Example 15
EXAMPLE 15 Pseudo-code showing dependencies between commands
void MyTextUtils::AddNote(string str, ITextModel* story, TextIndex position)
{
int32 stringLen = str.Length();
BeginSequence("AddNote");
MyTextUtils::AddToStory(story,position,str,stringLen);
MyTextUtils::SetTextColor(story,position,stringLen,BLUE);
EndSequence("AddNote");
}
At some point, a third party decides to extend the text functionality by writing a macro
expander, the purpose of which is to expand known acronyms as text is entered into stories.
The third party uses an observer that is notified when text is added to a story. If the observer
modifies the model, specifically changing the number of characters entered into the story, it
breaks the sequence in Example 15. The stringLen value is no longer valid.
114
Notification
Responders
If there is no risk of side effects from the model change within an observer (for example, the
observer modifies custom data, not the core application data on the object), it is safe to proceed
with the change. If there is a need to perform further modifications on the core application
data, consider factoring the change to execute as a separate task (possibly on application idle).
In Example 15, the observer might mark some custom, persistent data in the story, indicating it
needs to be processed, and schedule a custom command. The custom command examines the
story, determines whether the macro expansion operation is still valid, and updates the story if
required. No assumptions can be made about what might occur between marking the story in
the observer and the follow-on processing (for example, the document could close, further
updates could occur, or the application could terminate).
Commands that delete objects generally pre-notify, so subject observers get an opportunity to
tidy up associated objects before the object is deleted. Observers called in this context should
ensure the command state is CommandState::kNotDone.
Responders
This section describes the implementation of the responder pattern (see “Responder pattern”
on page 102) within the application. The responder pattern provides notification that a specific
event has occurred. For example, a responder can be called when a document is created,
opened, or closed. Figure 29 shows the general structure.
Notification
115
Notification
Responders
FIGURE 29
Pattern used for responders
«boss»
kASignalMgrBoss
«interface»
ISignalMgr
CountResponders() : int32
Init(ServiceID) : void
SignalNextResponder() : ErrorCode
Terminate() : void
«boss»
kAResponderBoss
1
0..*
«interface»
IResponder
Respond(ISignalMgr*) : void
SignalFailed(ISignalMgr*) : void
«realize»
«interface»
IK2ServiceProvider
GetServiceID() : ServiceID
«realize»
«realize»
SignalMgr
CResponder
AResponder
CServiceProvider
AResponderServiceProvider
ServiceID: = kARespServiceID
Respond(ISignalMgr*) : void
ServiceID: = kARespServiceID
The components to the responder pattern shown in Figure 29 are as follows:
z
Signal manager (signature interface ISignalMgr) — The entity responsible for notifying
responders of an event.
z
Service provider (interface IK2ServiceProvider) — The entity that identifies the object as a
responder and indicates the events of interest.
z
Responder (signature interface IResponder) — The entity invoked as a result of some event.
Signal manager
A signal manager is an object that controls the calling of responders. A boss class that has an
ISignalMgr interface is a signal manager. A signal manager is called by some other object that
wants responders to be notified when an event of interest has occurred. The object that calls a
signal manager is often a command (see interface ICommand). For example, the command
that opens a document calls the signal manager to notify responders which have registered an
interest in this event.
Responders are service providers. Each event for which a responder can be notified has a corresponding ServiceID. The application predefines a set of events and corresponding ServiceIDs
in which responders can register interest. For example, the open document event has a ServiceID of kOpenDocSignalResponderService. See the “Notification” chapter of Adobe InDesign
CS4 Solutions.
116
Notification
Responders
When a particular event occurs the signal manager is called to signal all responders that have
registered interest. The signal manager uses the service registry (not shown in Figure 29) to
find the responders that need to be called for the event.
If a responder sets the global error state, the signal manager is responsible for calling previously
signalled responders indicating there was an error condition (allowing the responder to take
whatever action might be necessary) and the action is aborted.
Responder
A responder is a boss class that supports two specific interfaces, the service provider interface
(IK2ServiceProvider) that identifies the events the responder is interested in, and the
responder interface (signature interface IResponder) that is called for a particular event.
Responder registration
Each event for which a responder can be called has a corresponding ServiceID. A responder
registers for one or more of these events by returning the ServiceIDs for the events of interest
via its IK2ServiceProvider implementation. See the “Notification” chapter of Adobe InDesign
CS4 Solutions for details on the predefined events the application makes available. For sample
code, see the DocWchServiceProvider class in the DocWatch plug-in.
Responder implementation
The responder is called via IResponder::Respond on an event of interest. The Respond method
performs whatever functionality is required and returns.
If the global error state is set (ErrorUtils::PMSetGlobalErrorCode) by a responder, any previously signalled responders have their IRespond::SignalFailed method called to indicate failure.
The SignalFailed method performs the function needed to recover. For example, if the
Respond method attached an observer to an object, the SignalFailed method detaches that
observer.
Both methods receive a pointer to a signal manager interface, so they can find out information
about the event. Typically, the Respond and SignalFailed methods query the signal manager
interface for other contextual interfaces. For example, the document signal manager (see kDocumentSignalMgrBoss) provides context in the IDocumentSignalData data interface. Responders that register for document signals like kAfterOpenDocSignalResponderService use this
data interface to access the document being opened.
The Respond and SignalFailed methods often call the service manager using the ISignalMgr::GetServiceID method, to verify the ID of the service that occurred. This allows one
responder implementation to be handle more than one type of event
For sample code, see the DocWchResponder class in the DocWatch plug-in.
Responders and the model
Responder implementations can perform follow-on modifications to the model. For example, a
responder called as the result of a command modifying the model can make further changes to
the model by calling other commands.
Notification
117
Notification
Responders
Take care that these modifications do not interfere with the model data on which the signal is
based. For example, if a responder signalled on a create story signal (kNewStorySignalResponderService) deletes or otherwise modifies the created story, commands, observers, and
responders that might follow the errant responder could make assumptions about the state of
the story. Avoid direct modifications that interfere with the data on which a responder is based.
If such a change is required, consider using the responder to set an associated, independent
state that can be used by an idle task (or similar) to perform the actual modification on the
model. This technique guarantees the model is in a consistent state before modifications are
made.
NOTE:
A responder should check the global error state is kSuccess
(ErrorUtils::PMGetGlobalErrorCode), before modifying the model.
Responders and global error state
Responders are designed to accommodate error; however, in practice, how a particular signal
service deals with error state is implementation dependent.
NOTE:
If you set global error state in the IResponder::Respond method, thoroughly test all areas in
the domain in which you are working.
If a responder sets global error state, the signal manager calls all previously signalled responders via IResponder::SignalFailed and aborts the action. The ability to clean up a failed action is
implementation and context dependent. For example, the creation of a text story object (kTextStoryBoss) can occur as a result of using the type text tool or importing a file. In the case of
importing a file, the responder is called from a much deeper stack than in the case of using the
type text tool. The implementation in the latter case does not handle aborting the import from
a responder.
Another implication of the IResponder::SignalFailed method being called on error is the
method must be able to consume and take remedial action for any errors caused within the
method. There is nothing the application core can do if an IResponder::SignalFailed method
itself fails; the condition is indeterminate.
Ordering of responders
No guarantees can be made about the order in which responders are called to react to an event.
If a client requires a particular responder to be called before another, the implementation of the
second responder should call ISignalMgr::SignalResponder (identifying the first responder) at
the start of its respond method. Even though the signal manager’s iteration of the responders
has been short-circuited by calling one of the responders directly, no responder is called more
than once.
118
Notification
Key client APIs
Responders and undo
Responders for the predefined events made available by the application are not called at undo
or redo of an operation.
If a responder processes commands to make follow-on changes to the model, these changes are
undone automatically as part of an undo operation; the client code need to manage this. The
same is true for redo of an operation.
If a responder is being used to attach an observer to a newly created model object, that
observer must be detached on undo and reattached on redo. Similarly, if a responder is being
used to detach an observer from a model object about to be deleted, that observer must be reattached on undo and detached on redo. To achieve this, the undo or redo actions need to be
tracked using an inval handler. See the “Commands” chapter for a description of the inval handler extension pattern.
If the responder is being used for a non-model based activity (for example, keeping a log of
specific operations), and the undo or redo actions need to be tracked, an inval handler again is
required.
Key client APIs
Table 9 shows key client APIs for observers and responders
TABLE 9 Key client APIs for observers and responders
API
Observers
Responders
Notification
Note
IObserver
Client code mainly calls AutoAttach and
AutoDetach to have an observer attach or detach
automatically to or from the subject(s) of interest.
ISubject
Client code mainly calls AttachObserver and
DetachObserver methods to attach or detach an
observer.
ISignalManager
If you are extending the application and need to
signal new types of events to responders you will call
the signal manager to perform the notification.
Often this notification is performed by a custom
command that you provide. Signal managers in the
application use the default implementation,
kSignalMgrImpl, provided by the API. This default
implementation provides all functionality for the
signal manager, it is unlikely that a third party
developer would need to further specialize it.
119
Notification
Extension patterns
Extension patterns
This section describes the mechanisms a plug-in can use to receive notification.
User-interface widget observer
Description
This observer is attached to the subject object for a user interface artifact and called on a user
interface event, such as selection of a checkbox or activation of a button.
Architecture
To implement a user interface observer, follow these steps:
z
Subclass the widget type you are creating (for example, kDialogBoss for dialogs), as
described in the “User Interfaces” chapter of Adobe InDesign CS4 Solutions.
z
Provide an implementation for IObserver. The AutoAttach and AutoDetach methods
should attach the observer to any sub-widgets (for example, all buttons on a dialog). Partial
implementations exist; for example CDialogObserver for dialogs and CWidgetObserver for
widgets.
z
User interface widget observers register for regular notification. Implement the IObserver::Update method to react to the user interface state change.
See also
z
The “User-Interface Fundamentals” chapter.
z
The “User Interfaces” chapter of Adobe InDesign CS4 Solutions.
z
For sample code: PictureIcon/PicIcoRollOverButtonObserver and WListBoxComposite/WLBCmpListBoxObserver.
Model observer
Description
A model observer provides an extension point for when a persistent object supported by an
application database supporting undo is updated.
Architecture
To implement a model observer, follow these steps:
120
z
Determine the subject that is to be observed, the protocol to listen on, and the change of
interest. See the “Notification” chapter of Adobe InDesign CS4 Solutions.
z
Determine which boss class in the object model will aggregate the observer. Commonly, the
boss class that represents the object being observed is a good candidate (for example, aggregating the observer on the document if you are interested in changes to the document). If
Notification
Extension patterns
the observer relates to multiple subjects (for example, attaching to all open documents),
place the observer somewhere accessible from client code.
z
Provide the implementation of IObserver, deriving from the CObserver partial implementation.
z
Determine whether it is appropriate to implement the AutoAttach and AutoDetach methods. Generally, these methods can be used if the subject object to be observed is known
within the scope of the observer (for example, if the subject is on the same boss object as the
observer). If the AutoAttach and AutoDetach methods are implemented, client code still
needs to call them to create the dependency between subject and observer. If these methods
are not provided, client code needs to manage the dependency between subject and
observer directly (through the ISubject::AttachObserver and ISubject::DetachObserver
calls).
z
Determine whether the subject object is maintained in a database that supports undo. If so,
lazy notification is available.
z
If lazy notification is available, determine the type of notification required. If the subject
object represents a UID-based object from a database that supports undo/redo, lazy notification is available. If the observer does not need to track every change to an object within a
sequence of changes (for example, the observer is used to keep information in a user interface panel up to date, and the panel needs to be updated just once, after all changes are
applied), lazy notification can be used. If the observer needs to be aware of all changes to the
object as they occur, use regular notification. When attaching the observer to the subject
(either in the AutoAttach/AutoDetach methods or through client code), specify the notification type.
z
If lazy notification is not available, register for regular notification.
z
Provide implementations of the IObserver::Update and/or IObserver::LazyUpdate methods, as required.
See also
Notification
z
“Relating observers to subjects” on page 112
z
“Determining the subject” on page 112
z
“Regular and lazy notification” on page 106
z
For sample code showing a model observer using lazy notification: CustomDataLink / CusDtLnkUITreeObserver, LinkWatcher / LnkWtchActiveContextObserver, PanelTreeView /
PnlTrvTreeObserver.
z
For sample code showing a model observer using regular notification: CustomDataLink/
CusDtLnkDocObserver, GoToLastTextEdit/GTTxtEdtNewDeleteStoryObserver, LinkWatcher/LnkWtchCacheManager, and PersistentList/PstLstDocObserver.
121
Notification
Extension patterns
Selection observer
Description
Selection observers provide an opportunity for client code to be called when there is a change
in the active selection. For details, see the “Selection Observers” section in the “Selection” chapter.
See also
z
“Selection Observers” section in the “Selection” chapter.
z
For sample code: TableAttributes/TblAttSelectionObserver, StrokeWeightMutator/StrMutSelectionObserver, BasicPersistInterface/BPISelectionObserver, and Persistent List/PstLstUISelectionObserver.
Document observer
Description
As well as notifying a change on a subject, often the notification is repeated on the document
for a subject. This allows observers to watch for a change of interest on a document (rather
than attach/detach from every object within the document). For example, an observer interested in page items being moved around the layout does not need to attach to each page item
(and detach if the page item is deleted, re-attaching if it is un-deleted, and so on), but attaches
to the document instead.
A document observer is a special example of a model observer, using the document object
(kDocBoss) as the subject.
See also
z
“Model observer” on page 120
z
For sample code: CustomDataLink/CusDtLnkDocObserver, LinkWatcher LnkWtchCacheManager, and PersistentList/PstLstDocObserver.
Active context observer
Description
A context observer is an observer that watches the active context.
Architecture
The active context object (kActiveContextBoss) is session wide and available from the session
object (kSessionBoss) through the ISession::GetActiveContext call. An active context observer
is an observer attached to the subject interface on the active context object. Active context notifications always are broadcast on the IID_IACTIVECONTEXT protocol. Only regular notification is available. The change context provided is another IID, indicating the type of
122
Notification
Extension patterns
contextual change that occurred. The common contextual changes of interest are shown in
Table 10.
TABLE 10 Types of context changes
changedBy
Note
IID_IDOCUMENT
The active document changed.
IID_ICONTROLVIEW
The active layout view changed (for example, when
the document has multiple layout views open).
IID_ISPREAD
The current spread changed.
IID_ISELECTIONMANAGER
The type of selection changed.
To implement a context observer, follow these steps:
z
Aggregate an observer on a boss class accessible from client code.
z
Provide implementations of IObserver::AutoAttach and IObserver::AutoDetach (attaching
to the ISubject interface on the kActiveContextBoss accessible from the GetExecutionContextSession()->GetActiveContext() call).
z
Provide an implementation of the IObserver::Update method. The protocol ID is
IID_IACTIVECONTEXT.
Example 16 shows how to detect when the active document has changed.
EXAMPLE 16 How to detect the active document has changed
void MyObs::Update(const ClassID& theChange, ISubject* iSubj, const PMMID&
protocol, void* changedBy)
{
if (protocol == IID_IACTIVECONTEXT){
InterfacePtr<IActiveContext> context(iSubj, UseDefaultIID());
IActiveContext::ContextInfo* info=
(IActiveContext::ContextInfo*)changedBy;
if(info->Key() == IID_IDOCUMENT){
HandleDocumentChange(context);
}
}
}
See also
z
Notification
For sample code: LinkWatcher/LnkWtchActiveContextObserver.
123
Notification
Extension patterns
Subject
Description
If there is a need for a boss class to be observable, it needs to be a subject.
Architecture
To turn a boss class into a subject, follow these steps:
z
Aggregate the ISubject interface onto the boss class.
z
Provide the implementation for the interface. Generally, re-using the API-supplied implementation (kCSubjectImpl) is sufficient, and there should be no need for further specialization of this interface.
See also
z
For sample code: PersistentList/kPstLstDataBoss.
Responder
Description
A responder provides notification of an event, such as document open.
Architecture
To implement a responder, follow these steps:
z
Determine the event of interest (as described in the “Notification” chapter of Adobe InDesign CS4 Solutions).
z
Aggregate the service provider interface (IK2ServiceProvider) on a boss class. Either re-use
an existing implementation (see the “Notification” chapter of Adobe InDesign CS4 Solutions), or provide an implementation that defines the set of ServiceIDs for the events of
interest.
z
Aggregate the responder (IResponder) on the same boss class.
z
Provide an implementation for the responder, derived from the CResponder partial implementation.
See also
124
z
The “Notification” chapter of Adobe InDesign CS4 Solutions
z
For sample code: DocWatch/DocWchResponder, CustomDataLink/CusDtLnkDocResponder, and BasicPersistInterface/BPIDefaultResponder.
Selection
Concepts
Selection
This chapter describes the InDesign/InCopy selection architecture. The selection architecture
is based on the concept of suites—abstract interfaces that allow client code to interact with an
abstract selection. A suite remains neutral to the underlying format of the objects that are
selected.
Concepts
Selection
A selection is whatever is chosen or picked out. For instance, when a user selects a range of text
with the Text tool, the user chooses a range of characters. In this case, the selection is a choice
of characters from the text model. Highlighting in the user interface is incidental feedback that
helps the user make the intended selection. Usually, making a choice using selection is a precursor to modifying the properties of the model underlying the selection. Selection primarily is
about targeting objects on which to perform some action, and the selection subsystem exists to
mediate changes to those objects.
Selection format and target
It is necessary to distinguish between the format of a selection and its content, the selection target. A selection format describes the model format of a selectable object, like the layout
(frames), text, table cells, or XML structure. Each of these has a distinct arrangement of objects
that represents its model. A selection target specifies a set of selected objects of a given selection format, like particular frames or a specific range of text.
The principal goal of the selection subsystem is to hide all details about formats and targets
from client code. Client code needs ask only, “Can this client code do this operation to the
selection?” If yes, the client code can try to change the properties of the selection target. This
encapsulation is achieved by means of the facade pattern discussed in “Design patterns” on
page 126.
InDesign 2.0 included support for a handful of selection formats, including layout objects
(frames), text, table cells, and XML-structure items. InCopy CS added support for other selection formats, like notes. See Table 11.
Selection
125
Selection
Concepts
TABLE 11 Selection formats and their targets
Selection format
When obtained
Selection target
Application defaults
No documents are open.
Global preferences
Document defaults
Document is open with nothing selected.
Document preferences
Galley text
Insertion position is in text, or text range is chosen
in Galley view.
Range of characters in a text model
(ITextTarget)
Layout
One or more page items are chosen in a layout view.
Layout objects (see ILayoutTarget)
Note
Insertion position or text range is chosen in a note
view.
Range of characters in note text
(ITextTarget)
Story editor text
Insertion position is in text, or text range is chosen
in a story-editor view.
Range of characters in a text model
(ITextTarget)
Table (table cells)
One or more table cells are chosen in a layout view.
Table cells in a table model
(ITableTarget)
Text
Insertion position is in text, or text range is chosen
in a layout view.
Range of characters in a text model
(ITextTarget)
XML structure
One or more nodes are chosen in an XML-structure
view.
Elements/attributes in the XML
structure (IXMLNodeTarget)
Design patterns
A key task for a plug-in developer is writing model code that extends some aspect of the document model. For instance, if your principal requirement is adding private data to the layout
model (see the BasicPersistInterface sample plug-in), the data added extends the layout model
and persists with document contents. The command pattern, used in the application API to
support undo and preserve document integrity, requires you to write further code to change
your model. The API already uses the facade pattern, like ITableCommands and IXMLElementCommands, to hide the complexity of parameterizing and processing low-level commands.
The selection architecture is an instance of the facade pattern, in that it makes client code easier
to write by defining a high-level interface on top of the selection subsystem. The selection
architecture shields client code from requiring detailed knowledge of what was selected in the
selection subsystem. A facade also makes software maintenance easier, because a facade weakens the coupling between a subsystem and its clients. This weak coupling allows the subsystem
components to be varied without necessarily affecting its clients.
126
Selection
Selection architecture
Selection architecture
The most important objective of the selection architecture is to ensure that new selection formats can be added easily.
To support new selection formats without rewriting large parts of the application, client code
must be de-coupled from the selection format. Client code must not need to know the selection’s subject identifier, meaning client code must not need to know which objects are selected
or any boss object associations in the underlying model (for example, the architecture of a text
frame). Client code also must not need to know any commands that manipulate the objects.
To achieve this goal, the selection architecture makes extensive use of suites that are neutral to
selection format. A suite provides a high-level, model-independent API that encapsulates the
details of the boss classes, interfaces, and commands in the underlying model. Clients interact
with suites instead of directly with the selected objects. In programming terms, a suite is a layer
of abstraction that sits on top of the selection and provides a narrow API through which client
code communicates its intent with respect to the portion of the model selected. In design-pattern terms, a suite represents a facade that makes it easier to write client code. Writing code
that sits on the model-manipulation side requires an understanding of the selection architecture and the model.
Figure 30 shows the layers of encapsulation in the selection architecture. Client code communicates through suites on an abstract selection boss (ASB) class. In turn, these suites communicate with concrete selection boss (CSB) suites on one or more CSB classes. The CSB suites deal
with the selection format of only their CSB, and they know how to access and change objects
only in the part of the model the CSB manipulates.
Selection
127
Selection
Selection architecture
FIGURE 30
Layers of encapsulation
Client
Client code uses the selection
manager (ISelectionManager)
to acquire a suite, and then calls
one of its methods.
Abstract selection:
Client code is decoupled
from the model by accessing
attributes of the selection
through an ASB.
Selection subsystem::Abstract selection
Concrete selection
comprises one or more CSBs,
each of which target the part
of the model that underlies its
selection format.
Integrator suite
forwards the method
call to the corresponding
CSB suite method on
enabled CSBs.
Selection subsystem::Concrete selection
Model stores attributes in
interfaces on boss objects. The
attributes targeted by selection
are selection attributes.
CSB suite accesses interfaces
on boss objects in the model
or processes commands to
change the model to do what
this suite method does.
Model
A window can have one or more views that can support selection. Each view that uses selection
instantiates a selection subsystem. In general, client code can remain unaware of what selection
subsystem is in operation. While software developers need to be concerned about selection formats if they are writing code that manipulates the model, there is less need to be concerned
about selection subsystems. For more information about selection subsystems, see “Selection
architecture” on page 127.
To write code that uses selection to choose attributes of the model to be displayed or changed,
it is important to understand the structure that owns the selection format. To illustrate this, a
schematic diagram of the selection subsystem used by the layout view (kLayoutWidgetBoss) to
128
Selection
Selection architecture
edit design documents is shown in Figure 31. It presents the static structure and highlights
important boss classes and interfaces.
FIGURE 31
ASB
Selection subsystem used by layout view
ASB:
abstract selection
IIntegratorTarget
IXxxxSuite
kIntegratorSuiteBoss
Suite:
interface that provides clients with a capability to access or change
attributes of the selection. More generically, a suite acts on or
reports information about the target. In the future, there may be
subclasses of kIntegratorSuiteBoss that are not selection related.
ISelectionManager
kAbstractSelectionBoss
kAbstractLayoutDocumentSelectionBoss
Selection manager:
Manages a selection subsystem.
ITableSelectionSuite
ITextSelectionSuite
ILayoutSelectionSuite
Layout CSB
Text CSB
ILayoutTarget
IXxxxSuite
kLayoutSuiteBoss
IConcreteSelection
kNewLayoutSelectionBoss
Table CSB
ITextTarget
IXxxSuite
kTextSuiteBoss
kDocumentDefaultSuiteBoss
kTableSuiteBoss
IConcreteSelection
kTextSelectionBoss
Defaults
ITableTarget
IXxxxSuite
IConcreteSelection
kTableSelectionBoss
CSBs:
concrete selection boss classes that join this selection subsystem. Each CSB targets
its selection format(layout, text, or tables).
kDocWorkspaceBoss
Defaults:
Defaults are targeted
when nothing is selected.
The ASB provides the suites called by client code. In Figure 31, interface IXxxxSuite illustrates
such a capability. The suite implementation on the ASB is an integrator suite; its responsibilities
are described in detail in “Integrator suites” on page 142. For now, think of these suites as forwarding any calls they receive to corresponding suites on the CSBs. Further implementations
of IXxxxSuite are provided on each CSB that supports the capability. These implementations
are CSB suites; their role is discussed further in “CSB suites” on page 142
In Figure 31, you can see that IXxxxSuite is implemented on the layout CSB and the text CSB.
This means layout and text selections have its capability. Also, when nothing is selected, client
code still has the capability represented by IXxxxSuite, because the ASB then targets defaults
and finds an implementation of the suite on kDocumentDefaultSuiteBoss. For more information on how to use suites to manage preferences, see “Selection architecture” on page 127. Table
selections do not have the capability represented by IXxxxSuite, because there is no implementation of the suite on the table CSB.
Selection
129
Selection
Abstract selection bosses and suites
ISelectionManager is the signature interface that identifies a boss class as an ASB. To obtain a
suite, client code queries an ISelectionManager interface for the suite of interest. If the suite is
available, its interface is returned; otherwise, nil is returned. The ASB is not a normal boss
class; all suite interfaces are aggregated on the ASB, but an interface pointer is returned to client
code only when an active CSB (or the default CSB) that supports the suite exists. The ASB uses
a special implementation of IPMUnknown to control whether a query for an interface returns
the interface or nil.
Suite implementations get added to suite boss classes. The following suite boss classes are in
Figure 31:
z
kDocumentDefaultSuiteBoss for defaults
z
kIntegratorSuiteBoss for the ASB
z
kLayoutSuiteBoss for the layout CSB
z
kTableSuiteBoss for the table CSB
z
kTextSuiteBoss for the text CSB
A suite boss class has a target (like IIntegratorTarget or ILayoutTarget) and a collection of suite
implementations. Figure 31 shows how the selection extends the suite boss classes. Other boss
classes extend the suite boss classes in different ways. For example, scripting uses the suite boss
kTextSuiteBoss to manipulate text content.
Each box in Figure 31 is a self-contained, portable unit. This is one reason why the selection
architecture is not coupled to a view, and CSBs do not have pointers to their ASB. This design
allows re-use of suite boss classes by other features in the future. A catalogue of suite boss
classes is provided in Table 22, in “Suite boss classes” on page 145.
Each selection format has a CSB that extends its suite boss class. IConcreteSelection is the signature interface that identifies a boss class as a CSB. Only those CSBs that join the selection
subsystem used by the layout view are shown in the Figure 31. Other CSBs exist and are used
by other subsystems. For a complete list of all CSBs, see the API reference documentation for
IConcreteSelection. To understand more about the responsibilities of a CSB, see “Concrete
selection bosses” on page 131.
Also, the ASB supports a set of interfaces that allows the selection content to be set programmatically. ISelectionManager provides the ability to select/deselect all objects. Each selection
format provides an interface to allow the selection target to be varied. In Figure 31, for example, ILayoutSelectionSuite allows page items to be selected, and ITextSelectionSuite allows text
to be selected.
Abstract selection bosses and suites
The selection architecture isolates client code from model code by means of an ASB. The ASB
provides suite interfaces, which are collections of capabilities, as well as other interfaces needed
for managing the selection. The main point to understand about the ASB is that it implements
the facade for the selection subsystem. Client code interacts with the facade (a suite on the
130
Selection
Concrete selection bosses
ASB). Writing client code to take advantage of the selection architecture is straightforward; you
need to understand only the facade. To write model code and code to change your model or the
existing document model, however, it is important to understand how to interact fully within
the selection subsystem, and this is less straightforward.
The client code in Example 17 uses a parameter to determine the selection manager (ISelectionManager) to use by calling IActiveContext::GetContextSelection. The client code then queries for the suite of interest, IFooSuite. If the suite is available, the client checks to see if the
capability it wants is enabled by calling IFooSuite::CanDoSomething. If so, the client uses the
capability on the selection by calling IFooSuite::DoSomething. If someone were to create a new
selection format and add support for IFooSuite, this client code would work. It does not need
to know any details of how IFooSuite is implemented on any specific selection format.
EXAMPLE 17 Client code using the selection architecture
void FooActionComponent::DoAction(IActiveContext* ac, ActionID actionID...)
{
...
InterfacePtr<IFooSuite> iFooSuite(ac->GetContextSelection(), UseDefaultIID());
if (iFooSuite != nil && iFooSuite->CanDoSomething()) {
iFooSuite->DoSomething();
}
...
}
In most cases, either client code is passed a parameter that identifies the selection manager to
query for a suite, or the selection manager is implied by the kind of code being written.
Concrete selection bosses
A CDB is a boss class that encapsulates a selection format. It supports the manipulation and
observation of the selected objects in the part of the model where the objects are located.
This section describes the available selection formats and their associated CSB. The key properties of each CSB are tabulated. Table 12 describes the label for each field.
TABLE 12 CSB description legend
Selection
Label
Specifies
CSB
The boss class that represents this concrete selection.
CSB name
The name of the CSB that represents this concrete selection.
Description
A general description of this concrete selection.
131
Selection
Concrete selection bosses
Label
Specifies
Selection
attribute
extensibility
Indicates whether this concrete selection can be extended to notify of
changes to new selection attributes introduced by third-party software
developers. Each CSB is responsible for defining the mechanism used
to call selection extensions when selection attributes change. Some
CSBs allow the mechanism to be extended. See “Selection extensions”
on page 146 and “Selection observers” on page 147.
Selection
format
The format of the objects this concrete selection makes selectable.
Selection
suite
The interface that can specify the set of objects selected for this concrete
selection.
Suite boss
class
The boss class to which suites get added that extend the attributes that
can be manipulated by this concrete selection. A list of all suites
supported by the CSB is in the documentation for this boss class.
Custom suites get added to the suite boss class.
Target
The interface that identifies the set of objects selected for this concrete
selection.
Layout selection
Layout selections—made with the Selection tool or programmatically—are represented by the
layout CSB. For examples of the suites available, see IGeometrySuite, ITransformSuite, and
IGraphicAttributeSuite. For a complete list of available suites, see the API reference documentation for kLayoutSuiteBoss. Also see Figure 32 and Table 13.
FIGURE 32
Layout selection containing a page item
TABLE 13 Layout-selection CSB
132
CSB Name
Layout CSB
Selection format
Layout objects like frames and other kinds of page items.
CSB
kNewLayoutSelectionBoss.
Selection
Concrete selection bosses
Suite boss class
kLayoutSuiteBoss.
Target
ILayoutTarget.
Selection suite
ILayoutSelectionSuite.
Selection
attribute
extensibility
Extensible. The CSB has a shared observer attached to the
document’s subject (kDocBoss). Selection extensions define
CreateObserverProtocolCollection to extend the protocols that are
attached. Selection extensions define SelectionAttributeChanged
to handle change broadcasts.
Table selection
Table selections—made with the Type tool or programmatically—are represented by the table
CSB. For an example of a suite, see ITableSuite. For a complete list of available suites, see the
API reference documentation for kTableSuiteBoss. Also see Figure 33 and Table 14.
FIGURE 33
Table selection containing table cells
TABLE 14 Table-selection CSB
Selection
CSB Name
Table CSB
Selection format
Table cells.
CSB
kTableSelectionBoss.
Suite boss class
kTableSuiteBoss.
Target
ITableTarget.
Selection suite
ITableSelectionSuite.
Selection
attribute
extensibility
Extensible. The CSB has a cell focus (ICellFocus) that monitors
table-attribute changes in the selection. Changes to custom table
attributes also are detected by this CSB. Also, the CSB has a shared
observer attached to the document’s subject (kDocBoss). Selection
extensions define CreateObserverProtocolCollection to extend the
protocols that are attached. Selection extensions define
SelectionAttributeChanged to handle change broadcasts.
133
Selection
Concrete selection bosses
Text selection
Text selections in the Layout view—made with the Type tool or programmatically—are represented by the text CSB. Text selection take one of two principal forms, a selected range of characters in a story (TextRange) or an insertion point within a story (represented by a TextRange
with zero span). Although it may not seem intuitive that an insertion point is considered a text
selection, it is considered a selection from the viewpoint of selection management.
NOTE:
Other views use other CSBs to represent text selection.
For examples of the suites available, see ITextEditSuite and ITextAttributeSuite. For a complete
list of available suites, see the API reference documentation for kTextSuiteBoss. Also see
Figure 34, Figure 35, and Table 15.
FIGURE 34
Text selection containing a tange of characters
FIGURE 35
Text selection containing an insertion point
TABLE 15 Text-selection CSB
134
CSB name
Text CSB.
Selection format
Text.
CSB
kTextSelectionBoss.
Suite boss class
kTextSuiteBoss.
Target
ITextTarget.
Selection suite
ITextSelectionSuite.
Selection
Concrete selection bosses
Selection
attribute
extensibility
Extensible. The CSB has a text focus (ITextFocus) on
kTextSelectionFocusBoss that monitors text-attribute changes; this
picks up changes to custom text attributes. Also, the CSB has a
shared observer attached to the document’s subject (kDocBoss).
Selection extensions define CreateObserverProtocolCollection to
extend the protocols that are attached. Selection extensions define
SelectionAttributeChanged to handle change broadcasts.
Figure 36 shows a text selection in a table cell. Although the selection is inside a table, this is a
text selection and not a table selection. The selected text is indicated by the ITextTarget target
interface. The table implied by this selection is indicated by ITableTarget, a secondary targeting
interface provided by text selection.
FIGURE 36
Text selection in a table cell
Galley text selection
Text selections in the galley view—made with the Type tool or programmatically—are represented by the galley-text CSB. The galley-text CSB represents text selection made in a galley
view window.
Text selection uses many subsystems. For the most part, client code can remain unaware of
what subsystem is in operation, but suite developers need to know, because the CSB that supports text selection in the galley view is different than the one that supports it in the layout
view. If you do not add your suite to the appropriate suite boss class, your suite might not be
available when you want it. In most cases, the same suite implementation can be re-used. For
an example of a suite that is available, see ITextEditSuite. For a complete list of available suites,
see the API reference documentation for kGalleyTextSuiteBoss. Also see Figure 37 and
Table 16.
Selection
135
Selection
Concrete selection bosses
FIGURE 37
Galley text selection containing a range of characters
TABLE 16 Galley text-selection CSB
CSB name
Galley text CSB.
Selection format
Galley text.
CSB
kGalleyTextSelectionBoss.
Suite boss class
kGalleyTextSuiteBoss.
Target
ITextTarget.
Selection suite
ITextSelectionSuite.
Selection attribute
extensibility
Limited extensibility. The CSB has a text focus (ITextFocus) on
kTextSelectionFocusBoss, which monitors text-attribute changes.
Changes to custom text attributes are picked up by this.
Story-editor text selection
The story editor is available in InDesign and InCopy.
Text selections by the story editor are similar to galley text selection. The story-editor CSB
extends the galley-text CSB, and both share the same suite boss class, kGalleyTextSuiteBoss.
For an example of a suite that is available, see ITextEditSuite. For a complete list of available
suites, see the API reference documentation for kGalleyTextSuiteBoss. Also see Figure 38,
Figure 39, and Table 17.
FIGURE 38
136
Story-editor text selection in InCopy
Selection
Concrete selection bosses
FIGURE 39
Story-editor text selection in InDesign
TABLE 17 Story-editor text-selection CSB
CSB name
Story-editor text CSB.
Selection format
Story editor text.
CSB
kStoryEditorSelectionBoss.
Suite boss class
kGalleyTextSuiteBoss.
Target
ITextTarget.
Selection suite
ITextSelectionSuite.
Selection
attribute
extensibility
Limited extensibility. The CSB has a text focus (ITextFocus) on
kTextSelectionFocusBoss, which monitors text attribute changes.
Changes to custom text attributes are also picked up by this.
Note text selection
When text is selected in a note, the CSB involved is not the same as the one used for normal
text selection. The note-text CSB represents text selection made in a note. For an example of a
suite that is available, see INoteSuite. For the complete list of available suites, see the API reference documentation for kNoteTextSuiteBoss. Also see Figure 40 and Table 18.
FIGURE 40
Note selection
TABLE 18 Note text-selection CSB
CSB name
Selection
Note text CSB.
137
Selection
Concrete selection bosses
Selection format
Note text.
CSB
kNoteTextSelectionBoss.
Suite boss class
kNoteTextSuiteBoss.
Target
ITextTarget.
Selection suite
ITextSelectionSuite.
Selection attribute
extensibility
Limited extensibility. The CSB has a text focus (ITextFocus) on
kTextSelectionFocusBoss, which monitors text attribute changes.
Changes to custom text attributes are also picked up by this.
XML selection
XML selections in the structure view (the left panel of Figure 41)—made with the Selection
tool or programmatically—are represented by the XML CSB. There is a selection subsystem
associated with the XML structure view that is distinct from the selection subsystem associated
with the layout view, even though both views are displayed in the same window. Do not confuse XML selection with the XML tags shown in the layout view (the right panel of Figure 41).
For examples of suites that are available, see IXMLStructureSuite and IXMLTagSuite. For a
complete list of available suites, see the documentation for kXMLStructureSuiteBoss. Also see
Figure 41 and Table 19.
FIGURE 41
XML node selections in XML-structure view
TABLE 19 XML-selection CSB
138
CSB name
XML CSB
Selection format
XML node (element, attribute, or both)
CSB
kXMLStructureSelectionBoss
Selection
Concrete selection bosses
Selection
Suite boss class
kXMLStructureSuiteBoss
Target
IXMLNodeTarget
Selection suite
IXMLNodeSelectionSuite
Selection attribute extensibility
Not extensible.
139
Selection
Concrete selection bosses
Document defaults
When a document is open but nothing is selected, the document workspace (kDocWorkspaceBoss) becomes the selection subsystem’s target. For more information, see “Selection architecture” on page 127. Also see Figure 42 and Table 20.
FIGURE 42
Document defaults: open document with no selection
TABLE 20 Document-defaults CSB
140
CSB name
Defaults.
Selection format
Preferences maintained as data interfaces on kDocWorkspaceBoss.
CSB
kDocWorkspaceBoss. (Strictly speaking, this is not a CSB, because
it does not aggregate IConcreteSelection; however,
kDocWorkspaceBoss often is referred to as the defaults CSB.)
Suite boss class
kDocumentDefaultSuiteBoss.
Target
NA
Selection suite
NA
Selection
attribute
extensibility
Extensible. The CSB has a shared observer attached to the
document workspace’s subject (kDocWorkspaceBoss). Selection
extensions define CreateObserverProtocolCollection to extend the
protocols that are attached. Selection extensions define
SelectionAttributeChanged to handle change broadcasts.
Selection
Concrete selection bosses
Application defaults
When no document is open, a special selection subsystem (the workspace selection, kAbstractWorkspaceSelectionBoss) is activated. It targets application-wide defaults on kWorkspaceBoss.
For more information, see “Selection architecture” on page 127. Also see Figure 43 and
Table 21.
FIGURE 43
Application defaults: no document open
TABLE 21 Application-defaults CSB
Selection
CSB name
Defaults.
Selection format
Preferences maintained as data interfaces on kWorkspaceBoss.
CSB
kWorkspaceBoss. (Strictly speaking, kWorkspaceBoss is not a CSB,
because it does not aggregate IConcreteSelection; however,
kWorkspaceBoss often is referred to as the defaults CSB.)
Suite boss class
kApplicationDefaultSuiteBoss.
Target
NA
Selection suite
NA
Selection
attribute
extensibility
Extensible. The CSB has a shared observer attached to the session
workspace’s subject (kWorkspaceBoss). Selection extensions define
CreateObserverProtocolCollection to extend the protocols that are
attached. Selection extensions define SelectionAttributeChanged to
handle change broadcasts.
141
Selection
Integrator suites
Integrator suites
An integrator suite is an implementation of a suite that was added to kIntegratorSuiteBoss,
which is part of an ASB. Consequently, an integrator suite sometimes is known as an ASB suite.
All communication between client code and the selection is performed using the integrator
suite. A properly designed suite API is neutral to selection format, meaning there are no references to a specific selection format. This is essential to avoid tight coupling between the client
code and the model.
The purpose of an integrator suite is to iterate over the concrete-selection formats supported by
the suite or, more formally, to integrate over the integrator target (represented by IIntegratorTarget). Although it may seem that “iterator” may be a more appropriate name, “integrator” is
correct in this context, because it means “to join together.” In the mathematical sense, “integrate” also can refer to summation, and the integrator suite performs a summation over the
capabilities of individual selection-format-specific suites.
The implementation of an integrator suite is very stereotypical; it delegates any call to CSB
suites on CSBs that are active (i.e., on CSBs that have a selection). Integrator-suite implementations should be based on templates provided by the SDK. The templates reduce the amount of
code you need to write.
An example of an integrator suite is the kTableSuiteIntegratorImpl implementation of ITableSuite on kIntegratorSuiteBoss. Both text (kTextSuiteBoss) and table selections (kTableSuiteBoss) provide support for ITableSuite. The client code is shielded from the selection format
that is providing the implementation. Integrator suite calls are forwarded to the selection-format–specific implementations on the CSB suites.
CSB suites
A CSB suite is a suite implementation on a CSB. CSB suites receive calls forwarded from integrator suites when they are called by client code.
CSB suites exist to do model-specific work, like navigating relationships between objects in the
model and processing of commands. When you manipulate the model from within a CSB suite,
you interact with objects in a specific selection format (e.g., by means of target interfaces like
ITextTarget or ILayoutTarget). There are several CSBs that extend suite-boss classes; for example, kTextSelectionBoss subclasses kTextSuiteBoss. For a complete list of CSBs, see the API reference documentation for the IConcreteSelection interface, from which you can examine the
documentation page for each CSB to see which suite boss it extends.
142
Selection
Encapsulation
Encapsulation
A key concept in the selection architecture is that client code should be insulated from code
that accesses and modifies the model (both the document and associated data). Code that
works in terms of UID, UIDRef, or UIDList values that relate to stories (kTextStoryBoss
objects) or layout objects (for example, something that exposes an IGraphicFrameData interface) interacts with a portion of the document model.
A selection of objects that is represented, for example, as a UIDList (as would have been the
case in the old selection architecture) is tightly coupled to the model. Such tight coupling
makes it very difficult to add new selection formats, because client code already has detailed
knowledge about what is in a selection. Also, with such tight coupling, you cannot change the
way in which the model is represented without breaking client code, because the client code
has detailed model knowledge.
To add new selection formats, the client code needs to be de-coupled from the selection formats. This means client code should be neutral with respect to selection format when dealing
with the selection. The client code communicates by means of ASB suites. These suites, in turn,
communicate with CSB suites. The CSB suites deal only with the model format of their CSB;
they do not even know about the ASB. This is necessary because an alternative use for suites is
scripting, which most likely does not have an analog to the ASB.
Suites and the user interface: an example
In the StrokeWeightMutator sample plug-in, the stroke-weight drop-down widget is enabled
when the stroke weight of whatever is selected can be changed. These steps are followed:
1. The widget looks for the API’s stroke attribute suite (IStrokeAttributeSuite).
2. If the suite is available, a CSB (the one with something selected or the defaults, if nothing is
selected) has an implementation of the suite.
3. The widget asks the integrator suite on the ASB whether the stroke-weight attribute is available, by calling IStrokeAttributeSuite::GetStrokeWeightCount.
4. The integrator-suite implementation on the ASB forwards the call to the CSB-suite implementations.
5. The CSB implementation accesses the model for the selection format it supports and
returns the answer.
6. The widget then calls IStrokeAttributeSuite::GetStrokeWeight to retrieve the value if applicable.
When the widget handles a user click to change the stroke weight, the widget calls IStrokeAttributeSuite::ApplyStrokeWeight on the integrator suite to dispatch to the CSB suites, which
change the value stored in the model. The widget synchronizes what it displays with any
Selection
143
Selection
Responsibilities
changes to the selection using a selection observer. It is sent a message when the stroke weights
displayed in the widget need to be refreshed. (See “Selection observers” on page 147.)
Responsibilities
Basic client-code responsibilities
If you are writing client code that modifies the properties of an abstract selection and can use a
pre-existing suite, your responsibilities are minor. Query the selection manager (ISelectionManager) that represents the abstract selection for the suite you want (like ITextAttributeSuite
to change properties of text in a selection) and, if you get the suite, use it. Look for a code fragment showing these responsibilities for the ITextMiscellanySuite suite.
You must understand how to query the abstract selection for a particular capability. To understand what suites are provided by the API, see the API reference documentation and examine
the boss classes that aggregate ITextMiscellanySuite, to see the selection formats that support
this particular capability.
Selection-observer responsibilities
If you are writing code that needs to update when the selection changes, implement a selection
observer (see “Selection observers” on page 147). If there is a pre-existing suite that informs
your client code of the changes you are interested in, your selection observer looks for messages from this suite. For example, ITextAttributeSuite notifies clients when a text-attribute
change happens to the selection. Otherwise, implement a custom suite to inform client code
about the changes of interest (see “Custom suites” on page 145).
Custom-suite responsibilities
If you have any client code that depends on the state of the selection, write a custom suite to
access that state. For example, the BasicMenu sample implements a custom suite because it has
menu items that need to know whether a page item is selected. A suite is required because the
client code needs to access a selection target interface to determine whether a page item is
selected.
If you are writing code that changes the existing document model (for example, you are implementing or processing commands that change this model), and you want to use selection to
target the objects, you must create a suite.
If you are writing code that extends the document model (for example, you are adding a custom data interface to the layout model, like BasicPersistInterface), and you want to use selection to target the manipulation of this data, you must create a suite.
The model also extends to default preferences; changes to the application’s default preferences
and the document default preferences can be mediated by the selection subsystem.
144
Selection
Custom suites
Custom suites
When writing a custom suite that you want to make available to client code in one or more
selection formats, provide two implementations, one relating to abstract selections and another
to concrete selections. The former (see “Integrator suites” on page 142) is added to the kIntegratorSuiteBoss suite boss class, and the latter (see “CSB suites” on page 142) is added to the
selection-format-specific suite boss classes. For details about the selection formats that can be
of interest to client code, see “Concrete selection bosses” on page 131. Table 22 lists suite boss
classes.
TABLE 22 Suite boss classes
Suite boss class
Selection format
Rationale
kApplicationDefaultSuiteBoss
Application defaults
If you want your suite to be available to client
code when no documents are open, add the
suite to this boss class.
kDocumentDefaultSuiteBoss
Document defaults
If you want your suite to be available to client
code when a document is open but nothing is
selected, add the suite to this boss class.
kGalleyTextSuiteBoss
Text in InCopy Galley
or Story Editor views
(ITextTarget)
If you want your suite to be available to client
code during galley- or story-editor-view text
selections, add the suite to this boss class.
kIntegratorSuiteBoss
Abstract
(IIntegratorTarget)
Required for any custom suite.
kLayoutSuiteBoss
Layout (ILayoutTarget)
If you want your suite to be available to client
code during layout selections, add the suite to
this boss class.
kNoteTextSuiteBoss
Text in InCopy Note
(ITextTarget)
If you want your suite to be available to client
code during note selections, add the suite to
this boss class.
kSelectionInterfaceAlwaysActiveBoss
If you want your suite to be available to client
code regardless of the kind of concrete
selection that exists, add the suite to this boss
class.
kStoryEditorSuiteBoss
Text in Story Editor
view (ITextTarget)
(Not supported.) If you want the story editor
suite to be different from that for InCopy
Galley view, add the suite to this boss class.
kTableSuiteBoss
Tables (ITableTarget)
If you want your suite to be available to client
code during table selections, add the suite to
this boss class.
Selection
145
Selection
Selection extensions
Suite boss class
Selection format
Rationale
kTextSuiteBoss
Text (ITextTarget)
If you want your suite to be available to client
code during text selections, add the suite to
this boss class.
kXMLStructureSuiteBoss
XML structure,
(IXMLNodeTarget)
If you want your suite to be available to client
code during XML selections, add the suite to
this boss class.
Selection extensions
A selection extension (ISelectionExtension) is a bridge between a suite implementation with an
unknown interface and the selection architecture. A suite that requires advanced function registers a selection extension with the selection subsystem. The selection subsystem then communicates with the extension, which in turn forwards a message to the suite implementation. If
your suite requires one or more of the services described in this section, you must implement a
selection extension.
Caches
A suite may need to cache data to improve performance. You can add caches to suite implementations. In general, caches are added to CSB suite implementations. Cache validation
should be delayed as long as possible. Rather than updating the cache every time a model
change occurs, declare a Boolean cacheValid flag. When the model changes, set the flag to
kFalse. Before accessing the cache, rebuild it if it is invalid. Thus, if your suite provides data to a
user-interface panel and that panel is not visible, your suite does not needlessly consume CPU
cycles recalculating cache values the user does not need.
Furthermore, caches should be enabled only on CSB boss classes, not suite boss classes.
Because the scripting architecture extends suite boss classes, no caching should be done on
them. If you are implementing a custom suite that needs a cache, add your cache interface only
to the CSB. IConcreteSelection is the signature interface that identifies a boss class as a CSB.
Your custom suite implementation can then perform a run-time check to see if your cache
interface is available. If not, it can handle this situation gracefully and carry on without using
the cache.
Selection change notification
The design of some suites requires notification of selection changes. A selection extension
allows your suite to be called by a CSB whenever the selection changes. A selection extension
allows you to control whether clients of your suite—in the form of selection observers—get
notified. (See section “Selection observers” on page 147.)
146
Selection
Selection observers
There are two kinds of selection change events:
z
An object is added to or removed from the selection (that is, the selection changes); for
example, a user Shift-clicks to add a frame to the selection.
z
An attribute of an object in the selection changes (that is, a selection attribute changes); for
example, the stroke color of selected frames changes.
A suite that has a cache must mark its cache as invalid when notified the selection has changed.
A suite needs to invalidate its cache when notified that a selection attribute changed only if the
specific selection-attribute change affects this suite's cache. Many selection-attribute changes
may not affect a particular cache.
NOTE:
Selection-attribute change notification is not supported on all CSBs. For information on
each CSB, see “Concrete selection bosses” on page 131.
Initialization
Some suites need to perform an action on start-up or shut-down of a selection subsystem. Any
such initialization is done by implementing a selection extension. Do not confuse this kind of
start-up and shut-down with application services, like IStartupShutdownService, which are
called for the application session. Selection subsystems get created and destroyed as the views
that own them open and close.
Communication with Integrator suite
Rarely, a CSB implementation needs to communicate with its integrator. Because there is no
pointer on the CSB back to the integrator, a messaging system was created: IConcreteSelection::BroadcastToIntegratorSuite.
Selection observers
As described in “Selection change notification” on page 146, there are two kinds of selection
change event: selection-changed event and selection-attribute-changed event. A selection
observer (ActiveSelectionObserver) is the abstraction that allows client code to be called when
a selection-changed event occurs. For more information, see the SelectionObserver.h header
file.
Suites that need to communicate extra data to selection observers during the selection-changed
event implement a selection extension. Suites that need to include information about a selection-attribute–changed event also implement a selection extension. (See “Selection extensions”
on page 146.)
When a selection-changed event occurs, a selection observer’s ActiveSelectionObserver::HandleSelectionChanged member method is called. On receiving this call, an observer can take
action. For example, the observer can update some data displayed in a user-interface widget.
Selection
147
Selection
Selection-utility interface (ISelectionUtils)
Optionally, before taking action, the observer can examine the message parameter (ISelectionMessage) and look for extra data placed there by a suite.
When a selection-attribute–changed event occurs, a selection observer’s ActiveSelectionObserver::HandleSelectionAttributeChanged member method is called. On receiving this call, an
observer must call ISelectionMessage::WasSuiteAffected before taking action. There are many
kinds of selection attributes, and they can be given meaning only by the CSB suite. It examines
the broadcast from the selection format that changed (originating from a command notification that the model was updated), then passes a model-independent message about the change
to selection observers. A selection observer may be called when any attribute of the selected
objects changes, so a selection observer can receive messages that do not affect it. The selection
observer must filter these calls, so it does not needlessly take action and waste CPU cycles. For
sample code, see BasicPersistInterface.
Selection-utility interface (ISelectionUtils)
A utility interface for selection (ISelectionUtils) is provided on kUtilsBoss. For detailed information on the member methods, see the API reference documentation for ISelectionUtils.
Some member methods provided by ISelectionUtils allow you to access the active selection.
Most client code does not need to depend on the active selection, because client code normally is
given access to the selection manager to use. If you want the active selection, ISelectionUtils::GetActiveSelection returns the selection manager from the active context. There also are
other member methods, like QueryActiveLayoutSelectionSuite and QueryActiveTextSelectionSuite, that depend on the active selection manager; these should be used only by client
code that needs to work with the active selection.
148
Layout Fundamentals
Terminology
Layout Fundamentals
This chapter describes the features and supporting architecture of the layout subsystem, as well
as how a client can use these features and the layout-related extension patterns in the InDesign
API.
For use cases related to layout, see the “Layout” chapter of Adobe InDesign CS4 Solutions.
Terminology
See the “Glossary” for definitions of terms. Table 23 lists terms used in this chapter and sections that relate to them.
TABLE 23 Terminology
Term
See ...
Bounding box
“Bounding box and IGeometry” on page 190
Child
“Parent and child objects and IHierarchy” on page 155
Content page item
“Page items” on page 173
Current spread
“Current spread and active layer” on page 197
Document
“Documents and the layout hierarchy” on page 152 and
“Spreads and pages” on page 157
Document layer
“Layers” on page 161
Frame
“Frames and paths” on page 173
Front document
“The layout presentation and view” on page 194
Front view
“The layout presentation and view” on page 194
Geometric page item
“Coordinate systems” on page 183
Graphic frame
“Frames and paths” on page 173
Graphic page item
“Graphic page items” on page 177
Group
“Groups” on page 178
Guide
“Guides and grids” on page 180
Layer
“Documents and the layout hierarchy” on page 152 and
“Layers” on page 161
Layout hierarchy
“Spreads and pages” on page 157
Layout view
“Layout view” on page 196
Layout Fundamentals
149
Layout Fundamentals
Concepts
Term
See ...
Layout presentation
“The layout presentation and view” on page 194
Master page, Master
spread
“Documents and the layout hierarchy” on page 152 and
“Master spreads and master pages” on page 168
Page
“Documents and the layout hierarchy” on page 152 and
“Master spreads and master pages” on page 168
Page item
“Documents and the layout hierarchy” on page 152 and
“Page items” on page 173
Pages layer
“Spreads and pages” on page 157
Parent
“Parent and child objects and IHierarchy” on page 155
Path
“Frames and paths” on page 173
Publication
“Spreads and pages” on page 157
Spread
“Documents and the layout hierarchy” on page 152 and
“Spreads and pages” on page 157
Spread layer
“Layers” on page 161
Text frame
“Frames and paths” on page 173
Text page item
“Text page items” on page 177
Transformation matrix
“Transformation matrices” on page 183
Concepts
Figure 44 shows a facing-pages publication with three pages arranged over two spreads. The
document is presented for edit in the layout presentation. The Pages panel is used to edit the
spreads and pages, including the master spreads and master pages. The Layers panel is used to
edit the layers.
150
Layout Fundamentals
Concepts
FIGURE 44
Layout
presentation
Layout concepts
Document
Page item
(empty graphic frame)
Spread
Page item
(text frame)
Page
Pasteboard
Pages
palette
Layers palette
Master spread and
master pages
Layer
Some basic terms are given below. A full layout-related glossary is in “Terminology” on
page 149.
z
Document — An InDesign document, unless otherwise stated.
z
Layer — The abstraction that controls whether objects in a document are displayed and
printed and whether they can be edited. A layer can be shown or hidden, locked or
unlocked, arranged in front-to-back drawing order, etc. A page item is assigned to a layer. If
a layer is shown, its associated page items are drawn; if a layer is hidden, its associated page
items are not drawn. Layers affect an entire document: if you alter a layer, the change applies
across all spreads.
z
Layers panel —The user interface for creating, deleting, and arranging layers.
z
Layout presentation — The user-interface window in which the layout of a document is presented for viewing and editing. The layout presentation contains the layout view and other
widgets that control the presentation of the document within the view.
z
Master page — A page that provides background content for another page. When a page is
based on a master page, the page items that lie on the master page also appear on the page.
A master page eliminates the need for repetitive page formatting and typically contains page
numbers, headers and footers, and background pictures.
Layout Fundamentals
151
Layout Fundamentals
Documents and the layout hierarchy
z
Master spread — A special kind of spread that contains a set of master pages.
z
Page — The object in a spread on which page items are arranged.
z
Page item — Represents content the user creates and edits on a spread, like a path, a group,
or a frame and its content.
z
Pages panel — The user interface for creating, deleting, and arranging pages and masters.
z
Spread — The primary container for layout data. A spread contains a set of pages on which
page items that represent pictures, text, and other content are arranged. The pages can be
kept together to form an island spread, a layout that spans more than one page.
Documents and the layout hierarchy
This section introduces the data model on which the layout subsystem is based. The boss
classes and interfaces that represent spreads, pages, and page items in a document are
described in general.
Architecture
This section shows how the items introduced in “Concepts” on page 150 are represented by
boss classes and interfaces in a document. Figure 45 shows the overall arrangement.
152
Layout Fundamentals
Documents and the layout hierarchy
FIGURE 45
Layout-hierarchy class diagram
1
IMasterSpreadList
«boss»
kDocBoss
ILayerList
1
1
ISpreadList
Represents a
master spread.
Represents a
document.
Represents a
spread.
1..*
«boss»
kMasterPagesBoss
0..1
2..*
1..*
<<extends>>
«boss»
kSpreadBoss
«boss»
kDocumentLayerBoss
1
1
Represents a
layer.
4..*
IHierarchy
ISpreadLayer
«boss»
kSpreadLayerBoss
IMasterPage
IHierarchy
1..10
1
«boss»
kPageBoss
Represents a page or master page. Note:
A page doesn't own the content that lies
upon it; the spread owns the content.
0..*
«boss»
kGuideItemBoss
Represents a
ruler guide.
2
IHierarchy
0..*
«boss»
kGroupItemBoss
Represents a
group.
0..*
«boss»
kSplineItemBoss
Represents a frame or path.
Graphics or text page items can
be nested inside a frame.
A document (kDocBoss) comprises one or more spreads (kSpreadBoss) and master spreads
(kMasterPagesBoss). The ISpreadList interface lists the spreads. The IMasterSpreadList interface lists the master spreads. Each spread contains one or more pages (kPageBoss), ruler guides
(kGuideItemBoss), frames (kSplineItemBoss), paths (kSplineItemBoss), or groups (kGroupItemBoss). These objects are arranged in a spread in a tree represented by the interface IHierarchy. Frames and groups have additional page items nested within them. The objects in a layout
are layered, and this layering is reflected within each spread. A layer is represented by a document layer (kDocumentLayerBoss) that has two corresponding spread layers (kSpreadLayerBoss) in each spread. The document layers are listed by the ILayerList interface; the spread
layers are the immediate children of a spread on the IHierarchy interface. A spread (kSpreadBoss) contains 2n spread layers as children, where n is the number of document layers. All the
immediate children of a spread must be spread layers (kSpreadLayerBoss).
Figure 45 showed a document has at least two layers:
z
The pages layer — Pages (kPageBoss) always are associated with the pages layer. The pages
layer is represented by the document layer (kDocumentLayerBoss) found at index 0 in the
ILayerList interface. It has two corresponding spread layers (kSpreadLayerBoss) in each
Layout Fundamentals
153
Layout Fundamentals
Documents and the layout hierarchy
spread. The spread layer that owns the spread’s pages (kPageBoss) is found at child index 0
in the spread’s hierarchy (IHierarchy); the other spread layer, which is always empty, is
found at child index 1.
z
A layer for content.
The remaining layers—and there always is at least one other layer in the ILayerList interface—
are the layers to which guides (kGuideItemBoss), frames (kSplineItemBoss), paths (kSplineItemBoss), and groups (kGroupItemBoss) can be assigned. Each of theses layers is represented by a document layer that has two corresponding spread layers. When a page item is
assigned to belong to a particular layer, the object becomes owned by the corresponding spread
layer through IHierarchy. Changes to document layers are document-wide, meaning changes
to document layers affect all spreads, including master spreads in the document. For more
information about layers, see “Spreads and pages” on page 157 and “Layers” on page 161.
Figure 45 also shows that master spreads (kMasterPagesBoss) have the same organization as
spreads. A master spread (kMasterPagesBoss) contains master pages (kPageBoss), which provide background content for other pages. Master spreads form their own hierarchy, and each
page connects to its master by the interface IMasterPage. For more information, see “Master
spreads and master pages” on page 168.
The key interfaces involved in the layout data model are summarized in Table 24. For more
information, see the API reference documentation for these interfaces.
TABLE 24 Key interfaces in the layout data model
154
Interface
Note
IDocumentLayer
Signature interface for a document layer
(kDocumentLayerBoss). Stores a document layer’s properties.
Has an associated content-spread layer and guide-spread layer.
IHierarchy
Connects boss objects in a tree that represents a layout hierarchy.
Provides a mechanism to find parent and child objects and
interfaces. See “Parent and child objects and IHierarchy” on
page 155.
ILayerList
List of document layers (kDocumentLayerBoss) in a document.
A layer is represented by a kDocumentLayerBoss object that has
two associated kSpreadLayerBoss objects in each spread.
IMasterPage
Stores the master spread (kMasterPagesBoss) associated with a
page (kPageBoss) and the index of the master page within the
master spread on which this page is based.
IMasterSpread
Signature interface for a master spread (kMasterPagesBoss).
Stores the name of a master spread and contains the incremental
information that defines a master spread. This information is
beyond what is contained in a spread.
IMasterSpreadList
List of the master spreads (kMasterPagesBoss) in a document.
Layout Fundamentals
Documents and the layout hierarchy
Interface
Note
ISpread
Signature interface for a spread (kSpreadBoss). Contains useful
methods for discovering the page items that lie on a page or
spread and for navigating between document layers
(kDocumentLayerBoss) and spread layers (kSpreadLayerBoss).
A document is laid out as a set of spreads in which each spread
contains one or more pages and other page items.
ISpreadLayer
Signature interface for a spread layer (kSpreadLayerBoss). A
spread layer is the container for the page items in a layer through
the IHierarchy interface on the kSpreadLayerBoss. The
ISpreadLayer interface maintains a relationship back to the
corresponding document layer boss (kDocumentLayerBoss).
ISpreadList
List of the spreads (kSpreadBoss) in a document.
The SnpInspectLayoutModel code snippet can inspect the layout hierarchy. Run the snippet in
SnippetRunner to create a textual report for a document. See SnpInspectLayoutModel for sample code that examines the layout hierarchy in various ways.
Parent and child objects and IHierarchy
A page item can contain other page items. The containing object is the parent; the contained
objects are the children. Child index order defines z-order, the order in which objects draw; the
child with index 0 draws first (behind), and the child with index n-1 draws last (in front). This
association between page items is implemented by the IHierarchy interface, as shown in
Figure 46.
FIGURE 46
Page Item parent-child composition (class diagram)
0..*
«boss»
kPageItemBoss
children
1
parent
IHierarchy
kPageItemBoss is the abstract base class for objects that participate in the layout hierarchy. IHierarchy is a required interface. Subclasses of kPageItemBoss include spreads (kSpreadBoss),
spread layers (kSpreadLayerBoss), and frames (kSplineItemBoss). To see all subclasses, see the
API reference documentation for kPageItemBoss. Each subclass of kPageItemBoss may have
different constraints on the number and type (boss class) of children it can contain.
The child objects contained within the parent are said to be nested. Figure 47 shows an example. The object diagram is an example of how the internal representation changes when a circle
is nested inside a square using cut and paste. The parent-child association between boss objects
in a layout is realized by the IHierarchy interface. When the circle is cut, the association with its
Layout Fundamentals
155
Layout Fundamentals
Documents and the layout hierarchy
initial parent, the spread layer, is broken. When the circle is pasted, it is associated with its new
parent, the square. The square becomes a frame, because it now contains the circle. The drawing of the circle is then clipped to its frame.
FIGURE 47
Change in parent-child relationships when a circle is nested in a square
g
Above:
The original circle and square paths.
Below:
The parent-child relationships
between the boss objects that
represent these paths on a
spread.
g
Above:
The circle cut and pasted into
the square.
«boss»
spread1
: kSpreadBoss
Below:
The changed parent-child
relationships. The square is
the parent of the circle.
parent
child
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
circle
: kSplineItemBoss
parent
child
«boss»
layer1-content
: kSpreadLayerBoss
parent
child
«boss»
spread1
: kSpreadBoss
child
«boss»
square
: kSplineItemBoss
parent
child
«boss»
square
: kSplineItemBoss
parent
child
The parent-child associations
are represented by the
IHierarchy interface.
«boss»
circle
: kSplineItemBoss
The low-level commands used to edit the hierarchy programmatically are kAddToHierarchyCmdBoss and kRemoveFromHierarchyCmdBoss. A utility interface, IHierarchyUtils, provides a
facade that processes these commands. For more information, see the API reference documentation.
The boss objects in the hierarchy illustrated above can be visited using a recursive function, as
shown in Example 18:
156
Layout Fundamentals
Spreads and pages
EXAMPLE 18 Recursive function that visits each child boss object in a hierarchy
void VisitChildren(IHierarchy* parent)
{
int32 childCount = parent->GetChildCount();
for (int32 childIndex = 0; childIndex < childCount; childIndex++)
{
InterfacePtr<IHierarchy> child(parent->QueryChild(childIndex));
#ifdef DEBUG
// Trace the boss class name of the child.
DebugClassUtilsBuffer className;
DebugClassUtils::GetIDName(&className, ::GetClass(child));
TRACEFLOW("LayoutFundamentals", "%s\n", className);
#endif
// Add code to examine other interfaces on the boss object.
VisitChildren(child);
}
}
Spreads and pages
A spread is the primary container for layout data. A spread contains a set of pages on which the
page items that represent pictures, text, or content of another format are arranged. A spread
owns all the objects it contains, including its pages, ruler guides, frames, paths, groups, and any
nested content. The objects in a spread are organized into layers, so the user can control
whether the objects assigned to a layer should be displayed or editable. See Figure 48 for the
user’s perspective. The publication shown is a facing-pages document with one spread, one
page, and one layer.
Layout Fundamentals
157
Layout Fundamentals
Spreads and pages
FIGURE 48
Spread
User’s perspective of a sample spread and page
Page
Layer
Figure 49 is an object diagram showing the layout-related boss objects that exist in a publication containing a ruler guide and graphic frame on a page. The sample document (kDocBoss)
contains a spread (kSpreadBoss), page (kPageBoss), graphic frame (kSplineItemBoss), and
ruler guide (kGuideItemBoss). The content of the spread is layered on spread layers (kSpreadLayerBoss) that map onto corresponding document layers (kDocumentLayerBoss). There are
two document layers (kDocumentLayerBoss), the pages layer and layer 1. The page, graphic
frame, and ruler guide are arranged on spread layers (kSpreadLayerBoss) under the spread’s
IHierarchy interface.
158
Layout Fundamentals
Spreads and pages
FIGURE 49
Document containing a frame and ruler guide on a page
«boss»
layout-d
: kDocBoss
ISpreadList::GetNthSpreadUID(0)
ILayerList::QueryLayer(0)
«boss»
spread1
: kSpreadBoss
ILayerList::QueryLayer(1)
«boss»
pages
: kDocumentLayerBoss
«boss»
layer1
: kDocumentLayerBoss
IHierarchy::QueryChild(3)
IHierarchy::QueryChild(0)IHierarchy::QueryChild(1)IHierarchy::QueryChild(2)
«boss»
pages
: kSpreadLayerBoss
IHierarchy::QueryChild(0)
«boss»
page1
: kPageBoss
«boss»
private
: kSpreadLayerBoss
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
layer1-guides
: kSpreadLayerBoss
IHierarchy::QueryChild(0)IHierarchy::QueryChild(0)
«boss»
graphic-frame
: kSplineItemBoss
«boss»
ruler-guide
: kGuideItemBoss
A layer is represented by a kDocumentLayerBoss object with two corresponding kSpreadLayerBoss objects in each spread.
The pages layer is the layer to which pages (kPageBoss) are assigned. In Figure 49, the pages
layer is represented by the kDocumentLayerBoss object named “pages” and the kSpreadLayerBoss objects named “pages” and “private.” The pages document layer (kDocumentLayerBoss)
always is at index 0 in the layer list (ILayerList). The first child on the spread’s hierarchy always
is the associated page’s spread layer and contains the spread’s pages (kPageBoss). The second
child of the spread is a private spread layer that always is empty. Pages are drawn behind all
other objects, because they always are stored on the spread’s hierarchy at child index 0. Note the
pages layer is not shown in the Layers panel.
Layout Fundamentals
159
Layout Fundamentals
Spreads and pages
In Figure 49, Layer 1 is represented by the kDocumentLayerBoss object layer1. It has two corresponding spread layers (kSpreadLayerBoss) on the spread’s hierarchy, one for content
(layer1-content, which stores the graphic frame (kSplineItemBoss)) and one for ruler guides
(layer1-guides, which stores the ruler guide). The order of these spread layers can vary,
depending on whether guides are being drawn behind or in front of content. Figure 49 illustrates the order when guides are drawn in front of content.
NOTE:
Child index order in IHierarchy defines z-order, the order in which objects on those
spread layers (kSpreadLayerBoss) draw. The child with index 0 draws first (behind); the
child with index n-1, last (in front).
The code shown in Example 19 visits each child boss object of the spread shown in Figure 49.
(For the implementation of VisitChildren, see Example 18.) All boss objects on the spread’s
hierarchy are visited by the code in Example 19, including spread layer and page boss objects.
EXAMPLE 19 Code that iterates spreads and visits their children using IHierarchy
void VisitSpreadChildren(UIDRef& documentUIDRef)
{
InterfacePtr<IDocument> document(documentUIDRef, UseDefaultIID());
InterfacePtr<ISpreadList> spreadList(document, UseDefaultIID());
IDataBase* database = documentUIDRef.GetDataBase();
int32 spreadCount = spreadList->GetSpreadCount();
for ( int32 spreadIndex = 0; spreadIndex < spreadCount; spreadIndex++ )
{
UIDRef spreadUIDRef(database, spreadList->GetNthSpreadUID(spreadIndex));
InterfacePtr<IHierarchy> spreadHierarchy(spreadUIDRef, UseDefaultIID());
VisitChildren(spreadHierarchy);
}
}
In contrast, the code shown in Example 20 filters the objects by the pages on which they lie.
The method ISpread::GetItemsOnPage calculates which items lie on a page. Since pages do not
own the items that lie on them (spreads own the items), the calculation of which items lie on a
page is geometrical, comparing the intersection of the bounding box of each item with the
bounding box of the page. (For details, see ISpread::GetItemsOnPage.) If this code is run on the
spread shown in Figure 49, the only object that appears in the list itemsOnPage is the graphic
frame. The ruler guide is not returned as an item, because it is a pasteboard-ruler guide; if the
ruler guide were a page-ruler guide, the code in Example 20 would list it.
160
Layout Fundamentals
Layers
EXAMPLE 20 Code that iterates spreads and filters items by the page on which they lie using
ISpread
void FilterItemsByPage(UIDRef& documentUIDRef)
{
InterfacePtr<IDocument> document(documentUIDRef, UseDefaultIID());
InterfacePtr<ISpreadList> spreadList(document, UseDefaultIID());
IDataBase* database = documentUIDRef.GetDataBase();
int32 spreadCount = spreadList->GetSpreadCount();
for ( int32 spreadIndex = 0; spreadIndex < spreadCount; spreadIndex++ )
{
UIDRef spreadUIDRef(database, spreadList->GetNthSpreadUID(spreadIndex));
InterfacePtr<ISpread> spread(spreadUIDRef, UseDefaultIID());
for (int32 pageIndex = 0; pageIndex < spread->GetNumPages(); pageIndex++)
{
UIDList itemsOnPage(database);
const bool16 bIncludePage = kFalse;
const bool16 bIncludePasteboard = kFalse;
spread->GetItemsOnPage(pageIndex, &itemsOnPage, bIncludePage,
bIncludePasteboard);
// Add code to manipulate itemsOnPage.
}
}
}
The SnpInspectLayoutModel code snippet can inspect the spreads in a document and their
associated hierarchy. Run the snippet in SnippetRunner to create a textual report. For sample
code, see SnpInspectLayoutModel::ReportDocumentByHierarchy.
Layers
Layers are presented to the user and edited with the Layers panel. Layers can be shown or hidden, locked or unlocked. By re-arranging the order of layers, the user can control the front-toback order in which page items assigned to those layers draw. The user also can change the
layer to which a selected page item is assigned, by dragging and dropping in the Layers panel.
Figure 50 shows the user’s perspective of a layer. The document shown is a facing-pages document with one spread, one page, and one layer.
Layout Fundamentals
161
Layout Fundamentals
Layers
FIGURE 50
User’s perspective of a layer
Spread
Page
Layer
Layers in a basic document
Figure 51 is an object diagram showing the layout hierarchy that results when a new layer is
added to the document in Figure 49. The diagram shows a document (kDocBoss) containing
one spread (kSpreadBoss), one page (kPageBoss) and three layers. Each layer is represented by
one kDocumentLayerBoss object that has two corresponding kSpreadLayerBoss objects in
each spread. The pages layer that stores the page (kPageBoss) is represented by the kDocumentLayerBoss object named “pages” and the kSpreadLayerBoss objects named “pages” and
“private.” Layer 1 is represented by the objects layer1, layer1-content, and layer1-guides. Layer 2
is represented by the objects layer2, layer2-content, and layer2-guides. A graphic frame (kSplineItemBoss) and a ruler guide (kGuideItemBoss) are assigned to Layer 1. Layer 2 is empty.
162
Layout Fundamentals
Layers
FIGURE 51
Document with two layers
«boss»
layout-c
: kDocBoss
ISpreadList::GetNthSpreadUID(0)
«boss»
spread1
: kSpreadBoss
ILayerList::QueryLayer(0) ILayerList::QueryLayer(1)
«boss»
pages
: kDocumentLayerBoss
ILayerList::QueryLayer(2)
«boss»
layer1
: kDocumentLayerBoss
«boss»
layer2
: kDocumentLayerBoss
IHierarchy::QueryChild(5)
IHierarchy::QueryChild(4)
IHierarchy::QueryChild(0)IHierarchy::QueryChild(1)IHierarchy::QueryChild(2)IHierarchy::QueryChild(3)
«boss»
pages
: kSpreadLayerBoss
IHierarchy::QueryChild(0)
«boss»
page1
: kPageBoss
«boss»
private
: kSpreadLayerBoss
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
layer2-content
: kSpreadLayerBoss
IHierarchy::QueryChild(0)
«boss»
graphic-frame
: kSplineItemBoss
«boss»
layer1-guides
: kSpreadLayerBoss
«boss»
layer2-guides
: kSpreadLayerBoss
IHierarchy::QueryChild(0)
«boss»
ruler-guide
: kGuideItemBoss
When a new layer is added, three new objects are created:
z
A new document layer is added to the layer list (ILayerList). This is the kDocumentLayerBoss object layer2.
z
Two corresponding spread layers are added to the spread. One new spread layer is added to
contain paths, frames, or groups; the other new spread layer is added to contain ruler
guides. These are the kSpreadLayerBoss objects layer2-content and layer2-guides.
Figure 51 shows that when a document layer is created, two corresponding spread layers are
added to each spread. When a document layer is deleted, its corresponding spread layers are
removed from each spread. Document-layer edits are document-wide; i.e., they affect all
spreads in a document.
Content-spread layers are stored contiguously in the spread’s hierarchy (IHierarchy). Guidespread layers also are stored contiguously. For example, the layer1-content and layer2-content
objects in Figure 51 are contiguous content-spread layers, and layer1-guides and layer2-guides
are contiguous guide-spread layers.
Layout Fundamentals
163
Layout Fundamentals
Layers
The order of content-spread layers and guide-spread layers as child objects of the spread’s hierarchy (IHierarchy) varies, depending on whether guides are being drawn in front or in back of
content. Ruler guides are kept in a distinct spread layer, to allow them to be drawn in front of or
behind other content, or even to turn them off completely, as determined by a preference setting. (See “Guides and grids” on page 180.) Figure 51 shows the order when guides are drawn
in front of content. The corresponding index order is as follows:
index[0]
index[1]
index[2]
index[3]
index[4]
index[5]
pages spread layer
private spread layer
layer1-content spread layer
layer2-content spread layer
layer1-guides spread layer
layer2-guides spread layer
If guides are drawn behind other content, the index order of spread layers in the spread’s hierarchy (IHierarchy) is as follows:
index[0]
index[1]
index[2]
index[3]
index[4]
index[5]
NOTE:
pages spread layer
private spread layer
layer1-guides spread layer
layer2-guides spread layer
layer1-content spread layer
layer2-content spread layer
Child-index order in IHierarchy defines z-order, the order in which objects on those
spread layers (kSpreadLayerBoss) draw. The child with index 0 draws first (behind); the
child wth index n-1, last (in front).
The spread’s ISpread interface has methods that return the spread layer associated with a given
document layer. (See “Navigating spread content using ISpread” on page 166.) The associations
are calculated using the following algorithm.
The indices of the pages-spread layer and private-spread layer are fixed:
z
Pages-spread layer IHierarchy child index = 0.
z
Private-spread layer IHierarchy child index = 1.
Variable definitions:
z
i = index of document layer (kDocumentLayerBoss) in ILayerList whose associated spread
layer (kSpreadlayerBoss) is wanted.
z
c = number of document layers in ILayerList.
If guides are displayed in front of content, the following calculations are used:
z
Content-spread layer IHierarchy child index = i + 1.
z
Guide-spread layer IHierarchy child index = i + c.
If guides are displayed behind content, the following calculations are used:
164
z
Content-spread layer IHierarchy child index = i +c.
z
Guide-spread layer IHierarchy child index = i + 1.
Layout Fundamentals
Layers
Layer options
There are layer options that can be controlled via the Layers Options dialog. User can double
click on a layer name in the Layers palette to bring up the Layer Options dialog, as shown in
Figure 52.
FIGURE 52
Layer Options dialog
The attributes in the Layer Options dialog are managed by the IDocumentLayer interface,
which is aggregated on the kDocumentLayerBoss. Table 25 lists some of the commands that
can be used to modify layer options.
TABLE 25 Layer Options-related commands
Command
Note
kChangeLayerNameCmdBoss
Changes the name of the layer.
kIgnoreHiddenTextWrapCmd
Boss
Sets text-wrap options for a document layer.
kLockGuideLayerCmdBoss
Locks or unlocks a guide layer. A “locked” layer
means nothing in that layer (only guides) can be
selected.
kLockLayerCmdBoss
Locks or unlocks a layer. A “locked” layer means
nothing in that layer can be selected. Layers are
unlocked by default.
kPrintLayerCmdBoss
Sets the printability of the layer. A non-printable
layer does not print or export.
kSetLayerColorCmdBoss
Changes the layer color.
kShowGuideLayerCmdBoss
Sets the guide visibility of the layer. An “invisible”
layer is not displayed on the screen.
kShowLayerCmdBoss
Sets the visibility of the layer. An “invisible” layer is
not displayed on the screen.
Layout Fundamentals
165
Layout Fundamentals
Layers
Use the commands in Table 25 whenever possible to modify layer options.
The SnpPrintDocument.cpp SDK snippet allows the user to select which layer to print. It uses
kPrintLayerCmdBoss to set the printability for each layer. For more information, see the snippet source code.
NOTE:
A similar set of attributes also exists in the ISpreadLayer. “Layers in a basic document”
on page 162 explains the relationship between IDocumentLayer and ISpreadLayer in a
document. The ISpreadLayer interface primarily provides “getter” methods that
forward the request for information to the corresponding IDocumentLayer interface
Navigating spread content using ISpread
“Layers in a basic document” on page 162 shows the content of a spread is stored on its IHierarchy interface. Locating content using IHierarchy can be laborious. Alternately, the ISpread
interface makes it easy to examine the content of a spread, as shown in the object diagram in
Figure 53.
166
Layout Fundamentals
Layers
FIGURE 53
Layout Fundamentals
Spread layer navigation using ISpread
167
Layout Fundamentals
Master spreads and master pages
Figure 53 shows you can use the ISpread interface on a spread (kSpreadBoss) to find the guide
or content spread layer associated with a document layer. (For more information, see
ISpread::QueryLayer.) You also can discover which page items lie on a given page. (For more
information, see ISpread::GetItemsOnPage.)
Using ISpread to discover the content of a spread is much easier than using IHierarchy.
When run on the document shown in Figure 53, the code in Example 21 visits the content of
spread layer layer1-content if parameter wantGuideLayer is kFalse. The code visits the content
of spread layer layer1-guides if parameter wantGuideLayer is kTrue.
EXAMPLE 21 Code that iterates over spreads in a document, then iterates over document layers to
visitiitems on the spread layer associated with each document layer
void FilterItemsByLayer(UIDRef& documentUIDRef, bool16 wantGuideLayer)
{
InterfacePtr<IDocument> document(documentUIDRef, UseDefaultIID());
InterfacePtr<ISpreadList> spreadList(document, UseDefaultIID());
InterfacePtr<ILayerList> layerList(document, UseDefaultIID());
IDataBase* database = documentUIDRef.GetDataBase();
int32 spreadCount = spreadList->GetSpreadCount();
for ( int32 spreadIndex = 0; spreadIndex < spreadCount; spreadIndex++ )
{
UIDRef spreadUIDRef(database, spreadList->GetNthSpreadUID(spreadIndex));
InterfacePtr<ISpread> spread(spreadUIDRef, UseDefaultIID());
int32 layerCount = layerList->GetCount();
// Skip the pages layer (layer 0).
for (int32 layerIndex = 1; layerIndex < layerCount; layerIndex++)
{
IDocumentLayer* documentLayer = layerList->QueryLayer(layerIndex);
int32 pPos;
// Visit the content spread layer associated with the document layer
// if wantGuideLayer is kFalse, the guide spread layer otherwise.
InterfacePtr<ISpreadLayer> contentSpreadLayer
(
spread->QueryLayer(documentLayer, &pPos, wantGuideLayer)
);
InterfacePtr<IHierarchy> hierarchy(contentSpreadLayer, UseDefaultIID());
VisitChildren(hierarchy);
}
}
}
NOTE:
For an implementation of VisitChildren, see Example 18.
Master spreads and master pages
A master spread is a special kind of spread that contains master pages. A master page is a page
designed to contain background content for another page. When a page is based on a master
page, the page items that lie on the master page also appear on the page. A master page elimi-
168
Layout Fundamentals
Master spreads and master pages
nates the need for repetitive page formatting and typically contains page numbers, headers and
footers, and background pictures. The layout presentation is used to edit the content of master
spreads. The Pages panel is used to arrange the pages in a master spread and assign a master
page to a page.
See Figure 54 for the user’s perspective on master spreads and master pages. The publication
shown is a facing-pages document with a master spread containing two master pages.
FIGURE 54
User’s perspective of master spreads and master pages
Master spread
Master page
Master page
Master spreads and master pages in a basic document
A master spread (kMasterPagesBoss) is a special kind of spread; it extends the spread boss class
(kSpreadBoss). The content of a master spread is organized the same way as a spread: each has
a hierarchy (IHierarchy) with spread layers acting as parent objects for content. Compare the
spread shown in Figure 49 to the master spread shown in Figure 55. The document (kDocBoss)
shown contains a master spread (kMasterPagesBoss), with two pages (kPageBoss), and two
paths (kSplineItemBoss) in the form of a square and a circle.
NOTE:
Layout Fundamentals
Pages (kPageBoss) that are part of a master spread are master pages. This name clashes
with the name of the master-spread boss class, kMasterPagesBoss, so the distinction
between the two is worth repeating: a master page is a page (kPageBoss) owned by a
master spread (kMasterPagesBoss).
169
Layout Fundamentals
Master spreads and master pages
FIGURE 55
Facing-pages master spread
«boss»
layout-e
: kDocBoss
IMasterSpreadList
Master spread
«boss»
A-Master
: kMasterPagesBoss
IHierarchy
«boss»
A-Master-pages
: kSpreadLayerBoss
«boss»
A-Master-private
: kSpreadLayerBoss
«boss»
A-Master-layer1-content
: kSpreadLayerBoss
IHierarchy
«boss»
A-Master-layer1-guides
: kSpreadLayerBoss
IHierarchy
Master pages
«boss»
master-page-1
: kPageBoss
«boss»
master-page-2
: kPageBoss
«boss»
square
: kSplineItemBoss
«boss»
circle
: kSplineItemBoss
When a page is based on a master page, the page items that lie on the master page also appear
on the page. When the master spread in Figure 55 is applied to page 1 of the facing-pages document in Figure 49, page items from the master page appear on page 1. The circle in the
screenshot of page 1 in Figure 56 is an object from the master spread.
Figure 56 is an object diagram showing a one-page document that has a master (A-Master)
applied. The circle drawn on page 1 is an object inherited from the associated master page. The
boss objects that represent the spread and the master spread it is based on are shown. The association between a page (kPageBoss) and its master page is implemented by the IMasterPage
interface. The screenshot shows that filtering of the objects on the master spread is being performed. Objects that intersect the bounding box of the master page associated with the page
are drawn, other objects on the master spread are filtered out. The circle is the only object from
170
Layout Fundamentals
Master spreads and master pages
the master spread to draw, because it is the only object that intersects the bounding box of the
master page associated with page 1.
FIGURE 56
Spread based on a master spread
Page item inherited
from master
«boss»
layout-e
: kDocBoss
ISpreadList
Spread
IMasterPage
«boss»
spread1
: kSpreadBoss
«boss»
pages
: kSpreadLayerBoss
«boss»
private
: kSpreadLayerBoss
«boss»
page1
: kPageBoss
IMasterSpreadList
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
layer1-guides
: kSpreadLayerBoss
«boss»
graphic-frame
: kSplineItemBoss
«boss»
ruler-guide
: kGuideItemBoss
«boss»
A-Master-pages
: kSpreadLayerBoss
«boss»
master-page-1
: kPageBoss
Master spread
«boss»
A-Master
: kMasterPagesBoss
«boss»
A-Master-private
: kSpreadLayerBoss
«boss»
master-page-2
: kPageBoss
«boss»
A-Master-layer1-content
: kSpreadLayerBoss
Master
page
«boss»
square
: kSplineItemBoss
«boss»
A-Master-layer1-guides
: kSpreadLayerBoss
«boss»
circle
: kSplineItemBoss
When two new pages based on the master shown in Figure 55 are added to the document, the
layout hierarchy is as shown in Figure 57. The object diagram shows the boss objects that represent three pages in a facing-pages document based on the same master spread. The circle and
square objects drawn on pages 1, 2, and 3 come from the master.
FIGURE 57
Two spreads based on a facing-pages master spread
Page items inherited
from master
«boss»
layout-f
: kDocBoss
ISpreadList
«boss»
spread1
: kSpreadBoss
«boss»
pages
: kSpreadLayerBoss
«boss»
private
: kSpreadLayerBoss
«boss»
page1
: kPageBoss
Layout Fundamentals
IMasterSpreadList
IMasterPage
«boss»
spread2
: kSpreadBoss
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
layer1-guides
: kSpreadLayerBoss
«boss»
graphic-frame
: kSplineItemBoss
«boss»
ruler-guide
: kGuideItemBoss
«boss»
pages
: kSpreadLayerBoss
«boss»
page2
: kPageBoss
«boss»
page3
: kPageBoss
«boss»
private
: kSpreadLayerBoss
«boss»
layer1-guides
: kSpreadLayerBoss
«boss»
layer1-content
: kSpreadLayerBoss
«boss»
A-Master-pages
: kSpreadLayerBoss
«boss»
master-page-1
: kPageBoss
«boss»
master-page-2
: kPageBoss
«boss»
A-Master
: kMasterPagesBoss
«boss»
A-Master-private
: kSpreadLayerBoss
«boss»
A-Master-layer1-content
: kSpreadLayerBoss
«boss»
square
: kSplineItemBoss
«boss»
A-Master-layer1-guides
: kSpreadLayerBoss
«boss»
circle
: kSplineItemBoss
171
Layout Fundamentals
Master spreads and master pages
The master spread (kMasterPagesBoss) associated with a page (kPageBoss) is found by calling
IMasterPage::GetMasterPageUID. The associated master page (kPageBoss) within that master
spread is calculated by calling IMasterPage::GetMasterIndex. The calculation depends on the
index position (0, 1, 2, ...) of the page within its spread and the document set-up. Given the
result returned by IMasterPage::GetMasterIndex, ISpread::GetNthPageUID can be called to get
a reference to the master page (kPageBoss), or ISpread::GetItemsOnPage can be called to calculate the objects on the master spread that intersect the bounding box of the master page.
To base a page or set of pages on a master spread, process the kApplyMasterSpreadCmdBoss
command.
Master-page item overrides
Page items from a master spread can be overridden. For example, to change the fill color of the
circle on page 1 in the Figure 57, a master page item override is created. A record of the override is kept by the page (kPageBoss) in the IMasterOverrideList interface.
Overrides to master-page items are applied programmatically with kOverrideMasterPageItemCmdBoss.
Even when a master-page item is overridden, changes made to its counterpart on the master
still can effect page items. An overridden page item continues to inherit all properties of its
master counterpart that are not overridden. For example, if the only property that was overridden on page 1 is the fill color of the circle, modifying the position of the circle in the master
causes the position of the circle to be updated on the pages based on that master (pages 1 and
3). The application of the initial fill-color override creates a copy of the circle (kSplineItemBoss) object on spread 1; a record of the override is kept in IMasterOverrideList on page 1. The
circle on spread 1 is a controlled page item; the circle on the master spread is the controlling
page item. Master page items maintain a list of controlled page items in the IControlledPageItems interface, which represents the page items for which overrides were made. An overridden
master-page item on a spread maintains a reference to its controlling master-page item, in the
IControllingPageItem interface.
Basing one master page on another
A master page can be based on another master page. The association is maintained by the
IMasterPage interface as it is for a normal page (kPageBoss) owned by a spread (kSpreadBoss).
For example, consider two masters, A and B, with different sets of page items. When you drag
master A onto master B in the Pages panel, master B becomes based on master A. Master page
items on master A appear on master B; however, the page items on master B inherited from
master A remain owned by master A. If you alter the page items on master A, the change is
inherited by master B. Master B can override page items from master A for individual properties, just as the master-page items on a spread can be overridden.
172
Layout Fundamentals
Page items
Page items
A page item is a path, frame, group, picture, text frame, or page-item type that represents content the user creates and edits on a spread. For general information about the organization of
page items in a spread, see “Spreads and pages” on page 157.
Frames and paths
A path represents a straight line, curved line, or closed shape like a rectangle, ellipse, or polygon. A path comprises one or more segments, each of which may be straight or curved. Each
path is open (the default) or closed. Graphic attributes are properties that control how the path
is formatted when drawn, like stroke weight and color. For more information about paths, see
the “Paths” section of the “Graphics Fundamentals” chapter. A frame is simply a path that contains—or is designated to contain—other objects, like a graphic page item, text page item, or
page item that represents another type of content.
Frames and paths are represented by the same boss class, kSplineItemBoss.
A path can become a frame, and a frame can become a path. The difference between a path and
a frame is the state of the interfaces on kSplineItemBoss. A path is a kSplineItemBoss object
that contains no children in its IHierarchy interface. A graphic frame is a kSplineItemBoss
object that contains a graphic page item in its IHierarchy interface. (For more information
about graphic page items, see the “Graphic Page Items” section of the “Graphics Fundamentals”
chapter.) A text frame is a kSplineItemBoss object that contains a text page item (kMultiColumnItemBoss). (For more information on text page items, see the “Text” chapter.)
Figure 58 shows a square path, an empty graphic frame, a graphic frame displaying a picture,
an empty text frame, and a text frame containing text. The object diagram shows a square being
used as a path and as a frame.
Layout Fundamentals
173
Layout Fundamentals
Page items
FIGURE 58
Square used as a path and a frame
Above:
Square.
Above:
Empty graphic frame.
Below:
Representation
as a spline,
a path-based
object. Any
spline can be
used as a frame.
Below:
Representation
as a spline
designated for
graphics content.
IGraphicFrameData::
IsGraphicFrame
returns kTrue.
«boss»
spread1
: kSpreadBoss
Above:
Graphic frame
displaying JPEG
picture.
Below:
Representation
as a spline
containing an
image in its
IHierarchy
interface. The
content of the
frame is clipped
by the frame's path.
Above:
Empty text frame.
Below:
Representation
as a spline
containing a
column controller
in its IHierarchy
interface that in
turn contains a
column. The
story (not shown)
associated with
these objects is empty.
Above:
Text frame
displaying text.
Below:
Representation
is the same as
the empty text
frame. The
difference is
that the associated
story (not shown)
contains text.
«boss»
layer1-content
: kSpreadLayerBoss
IHierarchy
«boss»
square
: kSplineItemBoss
«boss»
empty-graphic-frame
: kSplineItemBoss
«boss»
graphic-frame
: kSplineItemBoss
«boss»
empty-text-frame
: kSplineItemBoss
«boss»
text-frame
: kSplineItemBoss
«boss»
jpeg-picture
: kImageItem
«boss»
column-controller
: kMultiColumnItemBoss
«boss»
column-controller
: kMultiColumnItemBoss
«boss»
column1
: kFrameItemBoss
«boss»
column1
: kFrameItemBoss
The toolbox palette provides the user with several path-drawing tools. The Rectangle, Ellipse,
Polygon, and Line tools create basic shapes. The Rectangle Frame, Ellipse Frame, and Polygon
Frame tools create basic graphic-frame shapes. The Pen and Pencil tools create free-form
shapes. When one of these tools is used to draw a path, a spline item (kSplineItemBoss) is created to represent the path. The IPathGeometry interface on kSplineItemBoss describes the
points in the path. Other interfaces on kSplineItemBoss provide properties like stroke weight
174
Layout Fundamentals
Page items
and stroke color that control how the path is drawn. When a path is used as a frame, the content is clipped by the frame’s path.
IGraphicFrameData is a frame’s signature interface. IGraphicFrameData stores data that distinguishes whether the boss object represents a frame or a path. IFrameType is an alternative signature interface. It provides an easy-to-use interface that indicates whether the object
represents a path or a frame. Major interfaces on kSplineItemBoss, the boss class representing
paths and frames, are shown in the class diagram in Figure 59 and summarized in Table 26. For
more information on the interfaces noted and details of the other interfaces on kSplineItemBoss, see the API reference documentation.
FIGURE 59
kSplineItemBoss
IFrameType
«boss»
kSplineItemBoss
IGeometry
IGraphicFrameData
ITransform
IHierarchy
IPathGeometry
IShape
TABLE 26 Key kSplineItemBoss interfaces
Interface
Note
IFrameType
Indicates whether the boss object is a path or frame.
IGeometry
Stores the bounding box of the boss object in inner
coordinates. See “Bounding box and IGeometry” on
page 190.
IGraphicFrameData
Stores whether the boss object is a graphic frame and
provides other helper methods to determine the content
of a frame.
IHandleShape
Draws selection handles on the boss object’s bounding
box. These selection handles are used to move and resize
the object.
IHandleShape
(IID_IPATHHANDLESH
APE)
Draws selection handles on the points in the boss object’s
path. These selection handles are used to edit the path
points.
Layout Fundamentals
175
Layout Fundamentals
Page items
Interface
Note
IHierarchy
Connects the boss object into a tree that represents a
layout hierarchy. Provides mechanism to find parent and
child objects and interfaces. The objects the boss object
contains are the children of the IHierarchy interface. The
object in which the boss object is contained is its parent.
The child-index order in IHierarchy defines the z-order
of child objects. See “Parent and child objects and
IHierarchy” on page 155.
IPageItemAdornmentList
A list of adornments that the page item object may have.
For details, see the “Graphics Fundamentals” chapter.
IPathGeometry
Stores the points in the boss object’s path in inner
coordinates. The interface can describe multiple paths.
IScrapItem
Creates commands that can copy and paste the boss
object. Different types of page items need to transfer
different kinds of data. For example, a path transfers a
description of its shape, and a text frame transfers text as
well as a description of the frame.
IShape
Draws the boss object.
ITransform
Represents any transformation applied to the boss object
(e.g., translation, rotation, scaling) in a transformation
matrix (PMMatrix). See “Transformation and
ITransform” on page 191.
A frame can contain at most one child in its hierarchy (IHierarchy). The child boss object is the
content of the frame. The child may be one of the following boss classes: kSplineItemBoss,
kGroupItemBoss, a graphics page item (e.g., kImageItem, kSVGItem, kPlacedPDFItemBoss,
and kEPSItem), a text page item (kMultiColumnItemBoss), or another content page item.
NOTE:
The child object can be a group (kGroupItemBoss). To put several objects in a frame,
first group them.
A frame can be owned by one parent using its hierarchy (IHierarchy) interface. Typically, a
frame on a spread is owned by a spread layer (kSpreadLayerBoss), but a frame also can be
owned by other boss objects. For example, a frame that is part of a group is owned by the
kGroupItemBoss, a frame nested inside a frame is owned by a kSplineItemBoss, and an inline
frame is owned by a kInlineBoss.
Paths and frames are created programmatically using IPathUtils. For sample code, see SDKLayoutHelper::CreateRectangleGraphic,
SDKLayoutHelper::CreateSplineGraphic,
SDKLayoutHelper::CreateRectangleFrame, and SDKLayoutHelper::CreateTextFrame. The frame for a
page item can be created when a file is placed.
For sample code that can be used to examine the content of a frame, see Example 18. For sample code that can be used to discover frames in a document, see Example 19 and Example 20.
176
Layout Fundamentals
Page items
The SnpInspectLayoutModel code snippet can inspect the spreads in a document and their
associated hierarchy, including their frames. To create a textual report, run the snippet in SnippetRunner. For sample code, see SnpInspectLayoutModel::ReportDocumentByHierarchy.
Graphic page items
A graphic page item represents a picture. Table 27 shows the boss classes that represent graphics-format files that are imported and placed in a document. Graphic page items are contained
in frames (kSplineItemBoss) when placed on a spread. For more information, see “Graphic
Page Items” in the “Graphics Fundamentals” chapter.
TABLE 27 Graphic page-item boss classes
Boss class
Represented graphics format
kDCSItemBoss
DCS
kEPSItem
EPS
kImageItem
Raster image formats (e.g., TIFF, JPEG, PNG, GIF)
kPICTItem
PICT
kPlacedPDFItemBoss
PDF
kSVGItem
SVG
kWMFItem
WMF
Text page items
A text page item represents a visual container that displays text. Table 28 shows the boss classes
that participate in the display of text. Text page items are contained in frames (kSplineItemBoss) when placed on a spread. For more information, see the “Text Fundamentals” chapter.
TABLE 28 Text page-item boss classes
Boss class
Represents
kFrameItemBoss
Column
kMultiColumnItemBoss
Column controller
Layout Fundamentals
177
Layout Fundamentals
Page items
Interactive page items
An interactive page item represents an item in a multimedia format. Table 29 shows the boss
classes that represent interactive page items. Interactive page items are contained in frames
(kSplineItemBoss) when placed on a spread.
TABLE 29 Interactive item boss classes
Boss class
Represents
kMoviePageItemBoss
Movie
kPushButtonItemBoss
Button
kSoundPageItemBoss
Audio
Groups
Grouping is a way to combine two or more objects so they can be worked with as a single
object. The objects in a group move, transform, copy, and paste as a unit. Grouping boss objects
results in the creation of a kGroupItemBoss to represent the group. The grouped objects
become the children of the group in its IHierarchy interface. Unlike a frame, a group does not
have a path that clips its content when drawn. The geometry and shape of the group is defined
by the objects in the group.
The object diagram in Figure 60 shows an example of how the internal representation changes
when two objects are grouped.
178
Layout Fundamentals
Page items
FIGURE 60
Grouping of two kSplineItemBoss path objects
Before grouping
Above:
The original circle and
square paths.
Below:
The IHierarchy associations
between the boss objects that
represent these paths on a
spread.
After grouping
«boss»
spread1
: kSpreadBoss
«boss»
layer1-content
: kSpreadLayerBoss
Above:
The circle and square
combined in a group.
Below:
The new group boss
object, kGroupItemBoss,
and the changed
IHierarchy associations.
The circle and square
become children of the
group.
«boss»
spread1
: kSpreadBoss
«boss»
layer1-content
: kSpreadLayerBoss
IHierarchy
«boss»
circle
: kSplineItemBoss
«boss»
square
: kSplineItemBoss
«boss»
group1
: kGroupItemBoss
IHierarchy
«boss»
circle
: kSplineItemBoss
«boss»
square
: kSplineItemBoss
A group must contain at least two children in its IHierarchy. The only boss classes that can be
part of a group are kSplineItemBoss and kGroupItemBoss. In other words, only frames, paths,
and groups can be grouped. To put a graphic page item—such as an image (kImageItem)—into
a group, that item must be contained in a frame (kSplineItemBoss).
There is no signature interface on kGroupItemBoss that identifies a boss object uniquely as a
group. To test for a group, call IPageItemTypeUtils::IsGroup.
Objects are grouped using the kGroupCmdBoss command and ungrouped using the kUngroupCmdBoss command. Selected objects can be grouped and ungrouped using IGroupItemSuite.
Abstract page items and kPageItemBoss
The kPageItemBoss boss class is the abstract base class for page items. The kDrawablePageItemBoss boss class is the abstract base class for page items that can be selected and edited on a
Layout Fundamentals
179
Layout Fundamentals
Guides and grids
spread. These abstract classes are subclassed and implemented by the set of page items that can
be created and edited on a spread, such as kSplineItemBoss, kGroupItemBoss, and kImageItem.
In general terms, a page item is any object that can participate in a hierarchy (IHierarchy). This
perspective is valid and includes spreads (kSpreadBoss) and pages (kPageBoss) as page items.
There is an alternative, more advanced definition of page item. To see all subclasses of kPageItemBoss, see the API reference documentation for kPageItemBoss.
Guides and grids
Guides and grids are used to align and position objects. Some, like ruler guides, can be dynamically created and changed. Others, like the margins of a page, are static properties of an object.
Ruler guides
A ruler guide is a horizontal or vertical guide line used to align and position objects on a
spread. Ruler guides are represented by the kGuideItemBoss boss class and are a type of page
item. Ruler guides are represented as page items, because ruler guides share many behaviors of
page items. Ruler guides can be associated with a layer, moved around on a spread, and copied
and pasted. The signature interface that identifies a boss object as a ruler guide is IGuideData,
which stores the properties of the guide. There are two kinds of ruler guides: page guides span
a specific page, and pasteboard guides span the pasteboard of a spread.
Ruler guides are kept in their own spread layer (kSpreadLayerBoss) in the spread hierarchy
(IHierarchy), so ruler guides can be drawn in front or back of other content or completely hidden. For a description of this organization, see “Layers” on page 161.
Ruler guides are created by the kNewGuideCmdBoss command. They can be changed by the
kMoveGuideRelativeCmdBoss, kMoveGuideAbsoluteCmdBoss, kSetGuideViewThresholdCmdBoss, kChangeGuideColorCmdBoss, kSetGuidesBackCmdBoss, kSetGuideOrientationCmdBoss, and kSetGuideFitToPageCmdBoss commands.
Guide preferences are stored in the IGuidePrefs interface on the document workspace (kDocWorkspaceBoss). A distinct set of preferences that are inherited by new documents is maintained on the session workspace (kWorkspaceBoss). The kSetGuidePrefsCmdBoss command is
used to manipulate these preferences.
Margin and column guides
Margin guides represent the margins of a page. Column guides represent columns on a page.
Each page (kPageBoss) has its own margin and column settings stored in IMargins and IColumns interfaces, respectively.
The margins for a page can be changed using kSetPageMarginsCmdBoss. The columns for a
page can be changed using kSetPageColumnsCmdBoss and kSetColumnGutterCmdBoss.
180
Layout Fundamentals
Guides and grids
Document grid
The document grid is a grid (like graph paper) across a spread, on which page items can be
aligned.
Grid preferences are stored in the IGridPrefs interface on the document workspace (kDocWorkspaceBoss). A distinct set of preferences that are inherited by new documents is maintained on the session workspace (kWorkspaceBoss). The kSetGridPrefsCmdBoss command is
used to manipulate these preferences.
Baseline grid
The baseline grid is used to align the baseline of text across multiple columns on a spread.
Baseline-grid preferences are stored in the IBaselineGridPrefs interface on a story (kTextStoryBoss), the document workspace (kDocWorkspaceBoss), and the session workspace (kWorkspaceBoss). When a new story is created, the preferences in the document workspace are
inherited by the story. New documents inherit preferences from the session workspace. The
kSetBaselineGridPrefsCmdBoss command is used to manipulate these preferences. Each story
can override its baseline-grid preference settings.
Snap
If snapping is on, when the user drags an object within a certain distance of a guide or grid, the
object’s position is snapped to the guide or grid. If snapping is off, an object can be moved
freely. Snapping is implemented by interaction between three main objects:
z
The object on which other objects are being moved around; for example, a spread (kSpreadBoss). This object supports snapping by instantiating the ISnapTo interface.
z
An object that can be snapped onto, like a guide or grid. The snapping is performed by a
snap-to service. The signature interface is ISnapToService, and the ServiceID is kSnapToService.
z
The object handling the user interaction. Normally, this is a tracker created by a tool of
some sort. See the API reference documentation for the ITracker interface.
Snap preferences are stored in the ISnapToPrefs interface on the document workspace (kDocWorkspaceBoss). A distinct set of preferences that are inherited by new documents is maintained on the session workspace (kWorkspaceBoss). The kSetSnapToPrefsCmdBoss command
is used to manipulate these preferences.
Layout Fundamentals
181
Layout Fundamentals
Layout-related preferences
Layout-related preferences
The interfaces that store the major preference settings related to layout are summarized in
Table 30. Unless noted otherwise, each interface is present on the session workspace (kWorkspaceBoss) and document workspace (kDocWorkspaceBoss). kWorkspaceBoss contains the
settings inherited when a new document is created. kDocWorkspaceBoss contains the settings
for the document. For more information about the interfaces and commands listed, see the API
reference documentation.
TABLE 30 Layout-related preferences
Interface
Mutator
IAutoTextFramePrefs
kSetAutoTextFramePrefsCmdBoss
IBaselineGridPrefs
kSetBaselineGridPrefsCmdBoss
IColumnPrefs
kSetColumnPrefsCmdBoss
IDocStyleListMgr
182
Note
Stores a list of InDesign document presets
(kDocStyleBoss). The Save Preset button in the
New Document dialog box saves entries into
this list. Present only on the session workspace
(kWorkspaceBoss).
kDocAddStyleCmdBoss,
kDocDeleteStyleCmdBoss,
kDocEditStyleCmdBoss,
kDocSetStyleNameCmdBoss,
kSaveDocumentStyleDataCmdBoss,
kDocSetDefaultStyleNameCmdBoss
IFrameEdgePrefs
kSetFrameEdgePrefsCmdBoss
IGridPrefs
kSetGridPrefsCmdBoss
IGuidePrefs
kSetGuidePrefsCmdBoss
ILayerPrefs
kSetLayerPrefsCmdBoss
IMarginPrefs
kSetMarginPrefsCmdBoss
IPageLayoutPrefs
Stores the default shuffling behavior
preferences for pages (between spreads) when
pages are added, deleted, or moved.
kSetPageLayoutPrefsCmdBoss
IPageSetupPrefs
Stores the default set-up of an InDesign
document (number of pages, page size,
orientation, etc.).
kSetPageSetupPrefsCmdBoss
IPasteboardPrefs
kSetPasteboardPrefsCmdBoss
ISnapToPrefs
kSetSnapToPrefsCmdBoss
IZeroPointPrefs
kSetZeroPointPrefCmdBoss
Layout Fundamentals
Coordinate systems
Coordinate systems
This section describes the coordinate spaces used by layout, and the arrangement of the layoutrelated interfaces that store geometric data.
Coordinate systems define the geometrical location of objects in a document and the canvas on
which drawing occurs. Coordinate systems determine the position, orientation, and size of the
text, graphics, and images that appear on a page.
InDesign uses two-dimensional graphics. Position is defined in terms of coordinates on a twodimensional surface (a Cartesian plane). A coordinate is a pair of real numbers, x and y, that
locate a point horizontally and vertically within a two-dimensional coordinate space. A coordinate space is determined by the location of the origin, the orientation of the x and y axes, and
the lengths of the units along each axis.
InDesign defines several coordinate spaces in which the coordinates that specify objects are
interpreted. The following sections describe these spaces and the relationships among them.
Transformations among coordinate spaces are defined by transformation matrices, which can
specify any linear mapping of two-dimensional coordinates, including translation, scaling,
rotation, reflection, and skewing. Matrices are used to move from one coordinate space to
another.
The approach used by InDesign aligns with PostScript and Adobe PDF. for more information,
see
the
Adobe
PDF
Reference
(http://partners.adobe.com/public/developer/pdf/index_reference.html). Computer Graphics Principles and Practice (Foley, James D., et
al., Addison-Wesley, 1990) also provides useful theory on two-dimensional geometrical transformation.
Transformation matrices
A transformation matrix specifies the relationship between two coordinate spaces as a linear
mapping of one coordinate space to another. By modifying a transformation matrix, objects
can be scaled, rotated, translated, or transformed in other ways.
In InDesign, a transformation matrix is specified by six numbers in the form of a PMMatrix. In
its most general form, this PMMatrix is denoted [a b c d e f]; it can represent any linear transformation from one coordinate system to another.
The PMMatrix patterns that specify the most common transformations are as follows:
z
Translation is specified by [1 0 0 1 Tx Ty], where Tx and Ty are the distances to translate the
origin of the coordinate system in the horizontal and vertical dimensions, respectively.
z
Scaling is specified by [Sx 0 0 Sy 0 0]. This scales the coordinates so one unit in the horizontal and vertical dimensions of the new coordinate system is the same size as Sx and Sy units,
respectively, in the previous coordinate system.
Layout Fundamentals
183
Layout Fundamentals
Coordinate systems
z
Rotation is specified by [cos(A) sin(A) -sin(A) cos(A) 0 0], which has the effect of rotating
the coordinate system axes by an angle A counterclockwise.
z
Skew is specified by [1 tan(A) tan(B) 1 0 0], which skews the x axis by angle A and the y axis
by angle B.
Pasteboard coordinate space
The pasteboard coordinate space is the global coordinate system that encloses all objects in a
document, as shown in Figure 61.
FIGURE 61
Pasteboard coordinate space
Xpasteboard
Ypasteboard
All objects in a layout can have their coordinates expressed in pasteboard coordinates. The origin is the center of the first spread. The x axis increases from left to right; the y axis decreases
from up to down. The length along each axis is measured in the PostScript unit of points. For
example, Figure 62 shows the pasteboard coordinates of the bounding boxes for the spreads
and pages in a basic document. The figure shows the pasteboard coordinates of the bounding
box of each spread and page in a facing-page document with letter-sized pages of width 612
points and height 792 points. All values shown are in PostScript points.
184
Layout Fundamentals
Coordinate systems
FIGURE 62
Pasteboard coordinates of spread and page bounding boxes
-1224,-468
0,-396
612,-396
1224,-468
Xpasteboard
-1224,468
-1224,504
-1224,1440
0,396
-612,576
-612,1368
612,396
0,576
0,1368
1224,468
1224,504
1224,1440
Ypasteboard
The unit in which all measurements are stored in a document is the PostScript point. The unit
in which measurements are shown in the user interface is controlled by a preference (see “Measurement units” on page 192).
Inner coordinate space and parent coordinate space
Each page item has its own coordinate system, known as its inner coordinate space. Each page
item has an associated parent coordinate space. The inner coordinate space and its relationship
with its parent coordinate space are defined by three interfaces:
z
IHierarchy defines the parent association. The parent coordinate space is the first ancestor
boss object on IHierarchy that has an IGeometry interface. This may not be the boss object’s
immediate parent. For example, many objects are owned by a spread layer (kSpreadLayerBoss). Spread layers are not geometrical objects; they do not have an IGeometry interface;
therefore, the parent coordinate space for a boss object owned by a spread layer is the spread
(kSpreadBoss).
z
IGeometry defines the bounding box, which implies the origin and orientation of the x and
y axes.
z
ITransform defines the transformation matrix that maps from the inner coordinate space to
the parent coordinate space.
Figure 63 shows a square containing a rectangular child page item as an example of a parent
page item. The inner coordinate system of the parent as defined by its IGeometry interface is
shown.
Layout Fundamentals
185
Layout Fundamentals
Coordinate systems
FIGURE 63
Parent coordinates
-100,-100
100,-100
Parent
xparent
Child
-100,100
100,100
yparent
B. Inner coordinates of parent
A. Parent and child page items
In the A part of Figure 63, the parent is a square (kSplineItemBoss), 200 points wide. The child
is a rectangle (kSplineItemBoss), 40 points wide and 30 points high. The square is used as a
frame for the rectangle; this association is defined by the IHierarchy interface. The objects are
shown on a document grid with a grid line every 100 points and subdivisions every 10 points.
In the B part of Figure 63, the bounding box stored by the square’s IGeometry interface defines
the origin of the object’s inner coordinate space and the orientation of its x and y axes. The
coordinates of the bounding box of the square in its inner coordinate space are shown, together
with the x and y axes these coordinates imply. For a parent object, this is known as the parent
coordinate space.
The IGeometry and ITransform interfaces on the child object express the coordinate system
relationship to the parent. There is more than one way to arrange this data. In Figure 64, the
child’s coordinate system (the rectangle) is coincident with the parent’s coordinate system (the
square).
FIGURE 64
Coincident inner coordinates and parent coordinates
30,40
70,40
30,70
70,70
yparent
C. Inner coordinates of child
186
xparent
xparent
xchild
ychild
Transformation matrix
[1 0 0 1 0 0]
(the identity matrix)
30,40
70,40
30,70
70,70
yparent
D. After transformation of C to parent
Layout Fundamentals
Coordinate systems
In the C part of Figure 64, the rectangle has its own inner coordinate space. The bounding box
stored by the rectangle’s IGeometry interface is shown. This bounding box defines the origin of
the object’s inner coordinate space and the orientation of its x and y axes, labeled xchild and
ychild. The transformation matrix defined by the rectangle’s ITransform interface maps the
rectangle’s inner coordinate space (the child coordinate space) into the square’s coordinate
space (the parent coordinate space). In the arrangement shown, the parent and child coordinate systems are coincident; therefore, the transformation matrix in ITransform is the identity
matrix.
In the D part of Figure 64, when the bounding box of the child (the rectangle) is transformed
using the child’s transformation matrix to the parent coordinate space (the square), the coordinates are expressed in the parent coordinate space as shown.
In the arrangement shown in Figure 65, the child coordinate space is not coincident with its
parent coordinate space. Instead, the origin of the rectangle’s coordinate space is located at the
rectangle’s center.
FIGURE 65
Non-coincident Inner coordinates and parent coordinates
xparent
xparent
-20,-15
-20,15
20,-15
xchild
20,15
ychild
yparent
E. Inner coordinates of child
Transformation matrix
[1 0 0 1 50 55]
(translation Tx=50, Ty=55)
30,40
70,40
30,70
70,70
yparent
F. After transformation of E to parent
In the E part of Figure 65, the origin of the rectangle’s inner coordinate space is located at its
center. The coordinates of the bounding box defined by the rectangle’s IGeometry interface, are
shown together with the x and y axes these coordinates imply, labeled xchild and ychild. The
transformation matrix defined by the child’s ITransform interface specifies the translation
required to map to the parent coordinate space.
In the F part of Figure 65, when the bounding box of the child (the rectangle) is transformed
using the child’s transformation matrix to the parent coordinate space (the square), the coordinates are expressed in the parent coordinate space as shown.
A comparison of Figure 64 and Figure 65 shows that, after transformation, the bounding box of
the rectangle in its parent’s coordinate space is the same.
NOTE:
Layout Fundamentals
To transform any point from inner coordinate space to any other coordinate space, the
page item’s transformation matrix (ITransform) must be applied. For utilities to help
with the calculations, see TransformUtils.
187
Layout Fundamentals
Coordinate systems
Tools like the Rectangle tool use a tracker (CPathCreationTracker) to create a path (kSplineItemBoss) whose coordinate system is coincident with their parent coordinate system as
shown in Figure 64. Facades like IPathUtils often are used to create a path (kSplineItemBoss)
arranged as shown in Figure 65.
Spread coordinate space
Spread coordinate space is the coordinate system of a spread (kSpreadBoss). Each spread has its
own coordinate space, also known as the inner coordinate space for a spread. The origin of the
spread coordinate space is the center of the spread. The parent coordinate space is pasteboard
coordinate space. Unlike other page items, the parent of a spread is not defined by IHierarchy;
the parent of a spread is fixed as the document (kDocBoss). See Figure 66.
FIGURE 66
Spread coordinate space
Xspread1
Yspread1
Xspread2
Yspread2
In spread coordinates, the bounding box of each spread is identical. It is returned by IGeometry::GetStrokeBoundingBox. The spread’s (kSpreadBoss) implementation of IGeometry determines the bounding box using the IPasteboard interface. The spread’s ITransform interface
stores a transformation matrix that translates the spread coordinate space to pasteboard coordinate space. This matrix is a translation matrix [1 0 0 1 0 Ty], where Ty specifies the spread’s
offset in the vertical dimension.
Figure 67 shows the spread coordinates of the bounding box of each spread and page in a facing-page document with letter-sized pages 612 points wide and 792 points high.
188
Layout Fundamentals
Coordinate systems
FIGURE 67
Spread coordinates of spread and page bounding boxes in a basic document
-1224,-468
0,-396
612,-396
1224,-468
Xspread1
-1224,468
-1224,-468
0,396
Yspread1
-612,-396
0,-396
-612,396
Yspread2
0,396
612,396
1224,468
1224,-468
Xspread2
-1224,468
1224,468
Page coordinate space
Each page has its own coordinate space, also known as the inner coordinate space for a page
(kPageBoss). The parent coordinate space for page coordinate space is spread coordinate space.
The origin of page coordinate space is the top-left corner of the page. See Figure 68.
FIGURE 68
Page coordinate space
Xpage1
Ypage1
Xpage2
Ypage2
Layout Fundamentals
189
Layout Fundamentals
Coordinate systems
In page coordinates, the bounding box of each page (kPageBoss) in a document is identical. It
is returned by IGeometry::GetStrokeBoundingBox. The page’s ITransform interface stores a
transformation matrix that translates the page coordinate space to spread coordinate space.
Figure 69 shows the page coordinates of the bounding box of each page in a facing-page document with letter-sized pages 612 points wide and 792 points high.
FIGURE 69
Page coordinates of page bounding boxes in a basic document
0,0
Xpage1
612,0
Ypage1
0,792
0,0
Xpage2
612,792
612,0
Ypage2
0,792
612,792
Page-item coordinate space
Each page item has its own coordinate space, known as its inner coordinate space. The page
item’s bounding box is stored in interface IGeometry. The geometrical data stored in IGeometry and other data interfaces specific to the page item type are in the inner coordinate space of
the page item. For example, the points that describe a path in interface IPathGeometry are
stored in the inner coordinate space. A transformation matrix that transforms from inner coordinate space to parent coordinate space is stored in the ITransform interface. The parent coordinate system is defined by the IHierarchy interface.
For more information, see “Inner coordinate space and parent coordinate space” on page 185.
Bounding box and IGeometry
The bounding box (IGeometry::GetStrokeBoundingBox) is the smallest rectangle (PMRect)
that encloses a geometric page item:
190
z
The bounding box of an image (kImageItem) encloses the pixels that define its raster data.
z
The bounding box of a group (kGroupItemBoss) is the union of the bounding boxes of the
objects in the group.
Layout Fundamentals
Coordinate systems
z
The bounding box of a path (kSplineItemBoss) encloses all points in the path (IPathGeometry) and includes the effect of the properties that control how the path is stroked, such as
stroke weight and corner style.
z
The bounding box of a page (kPageBoss) or spread (kSpreadBoss) encloses the objects that
lie on it.
For illustrations, see “Spread coordinate space” on page 188 and “Page coordinate space” on
page 189.
A page item stores the data that describes its geometry in its inner coordinate space. For example, the bounding box in IGeometry and the points that describe a path in interface IPathGeometry are stored in inner coordinate space (see “Inner coordinate space and parent coordinate
space” on page 185). This stored data is independent of any transformation that may be in
effect. Transformation (e.g., scaling and rotation) is described by the page item’s ITransform
interface.
The IGeometry::GetPathBoundingBox method gives the path bounding box, the bounding
box of a path excluding the effects of adornments.
NOTE:
Paths (kSplineItemBoss) have path bounding boxes; page items of other types may not.
The IShape::GetPaintedBBox method gives the painted bounding box, the bounding box of the
page item including the effect of adornments.
Table 31 lists some useful geometry-related APIs.
TABLE 31 Useful geometry-related APIs
API
Note
IGeometry
Stores the bounding box of a page item.
IGeometryFacade
Positions and resizes page items.
IGeometrySuite
Positions and resizes objects currently selected.
Transformation and ITransform
A page item has its own coordinate space, its inner coordinate space. (See “Inner coordinate
space and parent coordinate space” on page 185.) The coordinates of the points that describe
the page item are stored in this inner coordinate space. A transformation matrix (PMMatrix)
that maps from inner coordinate space to the parent coordinate space is stored in the ITransform interface. Scaling, rotation, and other transformations that may be in effect are represented in this matrix. Points in inner coordinate space are mapped to the parent coordinate
space using the matrix obtained from ITransform::GetInnerToParentMatrix. There are useful
functions in TransformUtils.h that help to get the transform matrix (InnerToParentMatrix) and
perform transform calculations. For details, see the API reference documentation.
For example, to calculate the bounding box of a page item in its parent coordinate space, get the
bounding box from the page item (IGeometry::GetStrokeBoundingBox), and apply its transformation matrix (ITransform::GetInnerToParentMatrix). The calculated bounding box takes
into account all scaling, rotation, and other transformations that are applied. The parent could
Layout Fundamentals
191
Layout Fundamentals
Coordinate systems
be a group (kGroupItemBoss), frame (kSplineItemBoss), spread (kSpreadBoss), or other kind
of page item.
A transformation matrix can be inverted (PMMatrix::Invert) to create a matrix that performs
the inverse transformation. For example, if you have a matrix that transforms from inner coordinate space to parent coordinate space (InnerToParentMatrix), the inverse matrix can be used
to transform points from parent coordinate space to inner coordinate space.
The matrix that transforms from inner coordinate space to pasteboard coordinate space is
returned by ITransform::GetInnerToRootMatrix. To access it, we recommend you use a helper
function, InnerToPasteboardMatrix. This code walks up the hierarchy (IHierarchy) and concatenates the transformation matrix (ITransform::GetInnerToParentMatrix) of each page item.
The end result is a matrix that maps from the inner coordinates of the page item it started from
into pasteboard coordinates.
To compare the coordinates of two or more page items, the coordinates first should be transformed to a common coordinate space. The pasteboard coordinate space often is used for this
purpose.
When creating a path (kSplineItemBoss), it often is useful to specify its position in pasteboard
coordinates. You often position new page items relative to an existing boss object, like a spread,
page, or frame. For example, to determine a position relative to a page, you first need the IGeometry interface of the page (kPageBoss); then you can call InnerToPasteboard to transform
points on that page into pasteboard coordinates. When you have the points described in pasteboard coordinates, you can use IPathUtils to create the path. Alternately, you can describe the
points in the parent coordinate space as shown by the sample code in SnpCreateFrame and
SDKLayoutHelper.
Table 32 lists some useful transformation-related APIs.
TABLE 32 Useful transformation-related APIs
API
Note
ITransform
Stores the page item’s transformation matrix.
TransformUtils
Utility functions for transform calculations. See methods like
InnerToPasteboard and PasteboardToInner in
TransformUtils.h.
ITransformFacade
Scale, rotate, skew, or move page items.
ITransformSuite
Scale, rotate, skew, or move objects currently selected.
Measurement units
The internal unit of measurement used by the application is PostScript points. All measurements in a document are stored in the internal measurement unit as PMReal values. The user
can choose to work in other measurement units, like centimeters, inches, picas, or ciceros. The
user’s preferred measurement unit is converted to and from the standard internal unit (PostScript points) at the boundary between the user interface and the document model.
192
Layout Fundamentals
Coordinate systems
Unit-of-measure service
The conversion from a specific measurement unit to the internal measurement unit is performed by a unit-of-measure service.
A unit-of-measure service is a boss class that aggregates the IUnitOfMeasure and
IK2ServiceProvider interfaces. IUnitOfMeasure is the interface responsible for converting
between the unit of measure and the internal unit of measure. IUnitOfMeasure also has methods for producing a formatted string representation from units and tokenizing a string to produce a formatted number from the string. The measurement units supported by the application
can be extended. For more information, see “Custom unit-of-measure service” on page 203.
Common unit-of-measure services include kPointsBoss, kInchesBoss, kInchesDecimalBoss,
kMillimetersBoss, kPicasBoss, kCicerosBoss, and kCentimetersBoss.
Measurement-unit preferences
The user’s preferred measurement unit is stored in the IUnitOfMeasureSettings interface. The
session workspace (kWorkspaceBoss) stores the measurement-unit preferences inherited by
new documents. The document workspace (kDocWorkspaceBoss) stores the document’s preferences. The user changes these preferences by choosing Edit > Preferences > Units & Increments and using the resulting dialog box. To change these preferences programmatically, use
the kSetMeasureUnitsCmdBoss command.
Figure 70 is a class diagram of IUnitOfMeasureSettings.
FIGURE 70
Measurement-unit preferences (IUnitOfMeasureSettings)
kWorkspaceBoss
kDocWorkspaceBoss
IUnitOfMeasureSettings
IUnitOfMeasureSettings
The measurement system, IMeasurementSystem, is an interface aggregated on kSessionBoss
that is used to help locate and use a unit-of-measure service. Unit-of-measure services can be
referred to by either their ClassID or an index the measurement system assigns. The index
assigned by the measurement system is used while referring to a unit of measure in memory,
but this index should be converted to a ClassID when writing or reading from a stream.
Figure 71 is a class diagram of IMeasurementSystem.
FIGURE 71
Measurement system (IMeasurementSystem)
kSessionBoss
IMeasurementSystem
Layout Fundamentals
193
Layout Fundamentals
The layout presentation and view
Geometrical data types
The core geometrical data types are as follows:
z
PMPoint — (x, y) coordinates expressed as real (PMReal) numbers.
z
PMRect — Rectangle represented by points that designate the top-left and bottom-right
corners.
z
PMMatrix — Two-dimensional transformation matrix that maps from one coordinate
space to another.
The core geometrical-collection data types are as follows:
z
PMPointList
z
PMRectCollection
z
PMMatrixCollection
The layout presentation and view
This section describes the objects that present the layout to the user and allow the content to be
edited interactively.
Layout presentation
The layout of a document is presented within a layout presentation. A document has a layout
presentation (kLayoutPresentationBoss) for each view of the publication’s layout. A layout presentation is implemented by kLayoutPresentationBoss and its associated widgets. The pages
and spreads are presented by the layout view (kLayoutWidgetBoss). The general arrangement
is shown in the class diagram in Figure 72.
194
Layout Fundamentals
The layout presentation and view
FIGURE 72
Layout presentation
Layout presentation
kLayoutPresentationBoss
Horizontal ruler
kHorzRulerWidgetBoss
Zero point
kZeroPointWidgetBoss
Vertical ruler
kVertRulerWidgetBoss
A document can be presented
in one or more presentation
views. Layout presentations
present pages and spreads.
Layout view
kLayoutWidgetBoss
«boss»
kDocBoss
1
0..*
Note: Presentations returned
by IPresentationList are not
always layout presentations.
Other types of presentation
may be open on a document;
e.g., story windows.
The other widgets provided by the
layout presentation have not been
modelled; e.g., the widgets that
control the magnification of the
IPresentationList view, page navigation, and scrolling.
«boss»
kLayoutPresentationBoss
1
1
IDocumentPresentation
IPanelControlData
«boss»
kLayoutGenericPanelBoss
1
1
IPanelControlData
«boss»
kLayoutPanelBoss
1
0..1
«boss»
kZeroPointWidgetBoss
Layout Fundamentals
0..*
«boss»
kVertRulerWidgetBoss
0..1
«boss»
kHorzRulerWidgetBoss
1
IPanelControlData
«boss»
kLayoutWidgetBoss
195
Layout Fundamentals
The layout presentation and view
The application may have several layout presentations open at once, each displaying different
documents or different views of the same document, as shown in the class diagram in
Figure 73. The figure shows a screenshot of the application with three layout presentations
open. The boss classes and interfaces required to discover the documents open under the current session also are also shown. Each document knows the windows that are open on it by
means of the IPresentationList interface on kDocBoss.
FIGURE 73
Multiple layout presentations
To discover the layout presentations
that are open, iterate the
IPresentationList interface on
each document. Documents
opened by the current session
can be discovered as shown.
Layout
presentation
of a second
document
Layout presentations
showing different views
of the same document.
«boss»
kSessionBoss
1
1
ISession::QueryApplication
«boss»
kAppBoss
1
1
IApplication::QueryDocumentList
«boss»
kDocumentListBoss
1
0..*
IDocumentList::GetNthDoc
«boss»
kDocBoss
IDocument
IPresentationList
A document does not need to have a presentation view. Documents that do not have a presentation view can be edited programmatically. Sometimes this called editing a headless document.
Layout view
The layout view presents the layout of a document and is represented by the kLayoutWidgetBoss boss class. Documents can contain a variety of objects, like spreads, pages, frames, text,
and graphics. The document stores these objects; the layout view displays them and allows
them to be edited interactively. The layout view is part of a layout presentation (see Figure 72).
196
Layout Fundamentals
The layout presentation and view
Only the visible part of a document is presented in layout view. The layout presentation controls which part of a document is visible using magnification and page navigation widgets, etc.
Other panels, like the Pages and Navigation panels, also may adjust the view.
The layout view sometimes is known as the layout widget.
The layout view causes the visible objects to draw by discovering the spreads that are visible
and calling each spread to draw. The general arrangement is shown in the class diagram in
Figure 74.
FIGURE 74
Layout view and its associated document
The layout widget draws the area
within the red highlight. Interface
ILayoutControlData provides access
to the document being presented.
Additionally ILayoutControlData provides
access to the spread, page,
and layer being edited.
Spread
Floor
«boss»
kLayoutWidgetBoss
1
1
ILayoutControlData
«boss»
kDocBoss
1
1..*
ISpreadList
«boss»
kSpreadBoss
The layout view uses its ILayoutControlData interface to get to its associated document. The
ISpreadList interface on the document provides access to any particular spread. The IHierarchy
interface can be used to access children of the spread. For reasons of efficiency, navigation of
the layout hierarchy during a window draw is somewhat more selective than shown in
Figure 74. Portions of spreads that are not visible in a window are not called to draw. The layout
view also discriminates between redrawing an entire window and drawing only the region that
changed. These two strategies mean page items are not guaranteed to be called to draw during
every window draw.
Current spread and active layer
The current spread is the spread (kSpreadBoss) targeted for edit operations. Normally, new
page items created by the tools that handle user actions in the layout view are created in the
current spread. Each view (kLayoutWidgetBoss) stores a reference to its current spread in ILayoutControlData::GetSpreadRef. The current spread is set in the user interface by clicking on a
Layout Fundamentals
197
Layout Fundamentals
The layout presentation and view
spread in the layout view. The tool that handles the click processes the kSetSpreadCmdBoss
command. A reference to the current spread of the view that most recently edited the document (kDocBoss) is stored in the IPersistUIDData interface with PMIID
IID_ICURRENTSPREAD.
The active layer is the layer targeted for edit operations. New page items created by the tools
that handle user actions in the layout view are assigned to the active layer. Each view (kLayoutWidgetBoss) has a reference to its active layer (kDocumentLayerBoss) given by ILayoutControlData::GetActiveDocLayerUID. The active layer is set in the user interface by clicking on a
layer in the Layers panel, which then processes the kSetActiveLayerCmdBoss command. A reference to the active layer of the view that most recently edited the document can be found by
calling ILayerUtils::QueryDocumentActiveLayer.
When creating a page item programmatically, choose its parent boss object. Normally, this is a
spread layer (kSpreadLayerBoss). If you have access to a layout view, ILayoutControlData::QueryActiveLayer returns the hierarchy of the spread layer to use.
NOTE:
The current page is not stored anywhere in a document (in contrast, the current spread
and active layer are stored as described above). ILayoutControlData::GetPage returns
the visible page. This is calculated by means of ILayoutUIUtils::GetVisiblePageUID, by
finding the page whose center point is closest to the center of the layout view.
Layout-presentation and layout-view coordinate spaces
The relationship between layout-presentation coordinates and layout-view coordinates is relatively static and is described by the respective IControlView implementations. When the user
hides the rulers, the layout-view coordinate space is coincident with the layout-presentation
coordinate space. The relationship between pasteboard coordinates and layout-view coordinates is stored in the layout view’s IControlView interface. The layout view’s IControlView::GetContentToWindowTransform method returns the transform that maps from
pasteboard coordinates to layout-view coordinates. See Figure 75.
198
Layout Fundamentals
Key client APIs
FIGURE 75
Layout-presentation and layout-view coordinate spaces
X layoutpresentation
X layoutwidget
Xpasteboard
Ypasteboard
Y layout-widget
Y layout-presentation
As the user zooms and scrolls the view, the relationship between layout-view coordinates and
pasteboard coordinates changes. The layout view (kLayoutWidgetBoss) has an IPanorama
interface to track the scroll and zoom changes. The relationship between the pasteboard and
layout-view coordinate systems is stored in the layout view’s IControlView interface. The layout view’s IControlView::GetContentToWindowTransform method returns the transform that
maps from pasteboard coordinates to window coordinates. By using the known coordinate
relationships in the layout hierarchy, a region described in a page item’s inner coordinate space
can be mapped to the window coordinates.
Global-screen coordinates can be converted to pasteboard coordinates using ILayoutUIUtils::ComputePasteboardPoint. Event handlers often return mouse locations in global-screen
coordinates.
Key client APIs
This section summarizes the APIs provided to manipulate layout-related objects. For more
information, see the API reference documentation.
The layout view (kLayoutWidgetBoss) displays a document for editing and provides some
important data interfaces, summarized in Table 33. The layout view can be obtained in several
ways, one of which is calling ILayoutUIUtils::QueryFrontLayoutData.
Layout Fundamentals
199
Layout Fundamentals
Key client APIs
TABLE 33 Layout user-interface APIs
API
Note
ILayoutControlData
Data interface that refers to the document being viewed
and the spread, page, and layer currently selected
ILayoutControlViewHelper
Helper routines for performing page item hit-testing
Table 34 summarizes the suite interfaces that manipulate layout-related objects that are
selected. To obtain a suite, client code can query a selection manager interface (ISelectionManager) for the suite of interest. If the suite is available, its interface is returned; otherwise, a nil
pointer is returned. For details of how to obtain the selection manager, see the “Selection” chapter of Learning the Adobe InDesign CS4 Architecture. Given a selection manager, you can call
suite methods using code that follows this pattern:
InterfacePtr<IGeometrySuite> suite(selectionManager,UseDefaultIID());
if (suite) suite->MethodName();
TABLE 34 Layout suites
200
API
Description
IAlignAndDistributeSuite
Aligns and distributes objects that are selected.
IArrangeSuite
Brings forward, sends backward, or performs other zorder-related operations on objects that are selected.
IFrameContentSuite
Content-fitting operations and frame conversion
operations on objects that are selected.
IGeometrySuite
Positions and resizes objects that are selected.
IGroupItemSuite
Groups and ungroups objects that are selected.
IGuideDataSuite
Manipulates the properties of guides that are selected.
ILayerSuite
Manipulates document layers.
ILayoutHitTestSuite
Hit-tests objects that are selected.
ILayoutSelectionSuite
Sets the page items that are selected.
IMasterPageSuite
Manipulates master page item overrides.
IPageItemLockSuite
Locks or unlocks objects that are selected.
IPathOperationSuite
Performs path operations of selected page items.
IPathSelectionSuite
Performs path selections.
IReferencePointSuite
Manipulates reference point.
ITransformSuite
Scales, rotates, skews, or moves objects that are currently
selected.
Layout Fundamentals
Key client APIs
There are many command boss classes related to layout, as summarized in “Commands that
manipulate page items” on page 204. When possible, try to avoid to avoid processing low-level
commands when possible; instead, look for a facade or utility, to see if there is a method that
serves your purpose on one of these interfaces. Facades encapsulate parameterizing and processing the low-level commands. For example, to add an item to a page-item hierarchy, you
could process kAddToHierarchyCmdBoss, but instead you should use IHierarchyUtils::AddToHierarchy, which provides a ready-made facade.
The facades and utilities related to layout are listed in Table 35 and Table 36. These interfaces
are aggregated on kUtilsBoss. The Utils smart pointer class makes it straightforward to acquire
and call methods on these interfaces. You can call their methods by writing code that follows
this pattern:
Utils<IPathUtils>()->MethodName(...)
TABLE 35 Layout facades and utilities
API
Description
IGeometryFacade
Positions and resizes page items.
ITransformFacade
Scales, rotates, skews, or moves page items.
IPathUtils
Creates and works with paths (kSplineItemBoss).
IFrameContentUtils
Helps work with text or graphic-frame content.
IHierarchyUtils
Adds and removes page items from a hierarchy
(IHierarchy).
IImageUtils
Helper functions to get image information.
ILayoutUIUtils
Helper functions related to the layout user interface. Most of
these are related to the active or front document.
ILayoutUtils
Helper functions related to layout. Most of these are related
to pages or the page-item hierarchy.
IMasterSpreadUtils
Helper functions for master spreads and master page items.
IPageItemUtils
Page-item helper functions, like cache and change
notifications.
IPageItemTypeUtils
Helps determine the type of a page item (e.g., path, frame,
and image).
IPasteboardUtils
Helps determine the spread by location or gets the location
of page items.
IPathInfoUtils
Helps get path information.
IPathPointUtils
Stores path points during path-point transformation.
IPathUtils
Creates and manipulates path page items.
IRefPointUtils
Utility interface for functions that compute reference points.
Layout Fundamentals
201
Layout Fundamentals
Extension patterns
API
Description
ITransformUpdateUtils
Helps determine values relative to the zero point.
IValidateGeometryUtils
Validates transformation data.
TABLE 36 Layout helper classes
API
Note
Arranger
Brings to front, sends to back, and performs other z-orderrelated operations. See ArrangeUtils.h.
TransformUtils
Utility functions for transform calculations. See methods in
TransformUtils.h, such as such as InnerToPasteboard and
PasteboardToInner.
Extension patterns
This section summarizes the mechanisms a plug-in can use to extend the layout subsystem.
New-page-item responder
A new-page-item responder lets a plug-in receive notification when a new page item (e.g.,
frame, path, group, or image) is created by kNewPageItemCmdBoss. For the complete list of
page item types, see “Page items” on page 173.
A new-page-item responder often is used to detect the creation of a specific type of page item.
For example, to detect the creation of a graphic frame, implement a new-page-item responder
that gets a reference to the new page item (INewPISignalData::GetPageItem) and then checks
whether it is a graphic frame (IPageItemTypeUtils::IsGraphicFrame).
Another common use of a new-page-item responder is initializing a custom data interface on a
page item from defaults.
Consider the case where you defined a custom data interface and added this data interface to
the document workspace (kDocWorkspaceBoss) to store default data. You also added this data
interface to kDrawablePageItemBoss to store the settings per page item, and you want the
default data to be inherited when new page items are created. In this case, you can implement a
new-page-item responder that copies the data from defaults into the new page item. For an
implementation of this pattern, see BasicPersistInterface.
202
Layout Fundamentals
Extension patterns
A new-page-item responder is a service provider characterized by the following:
z
The responder interface IResponder — The implementation can expect to be able to acquire
an INewPISignalData interface from the signal manager (ISignalMgr) that it is passed.
z
The signature interface IK2ServiceProvider — The ServiceID must be of type kNewPISignalResponderService. You can re-use the API implementation kNewPISignalRespServiceImpl
for this.
Sample code
BasicPersistInterface copies data from defaults into new page items.
Custom page item
A custom page item lets a plug-in extend the set of objects that can be laid out on a page. A custom page item gives the plug-in control of how the object is drawn and how it behaves when
selected and manipulated by the user. A common use of a custom page item is adding support
for a graphics format for which the application provides no native support. In this case, the custom page item also requires a custom import provider (IImportProvider), to allow files of that
graphics format to be placed. Some custom page items also may require custom tools to be
implemented to create and manipulate the custom page items.
Implementing a custom page item can be complex. Before beginning development of a custom
page item, evaluate whether another type of extension pattern—like a page-item adornment or
draw-event handler—might meet your needs. If you do need to implement a custom page item,
study the subclasses of kPageItemBoss to see if there is an existing page item that is a close
match to what you need. (See kPageItemBoss in the API reference documentation.) If an existing page item comes close to meeting your needs, use this boss class as the base class for your
implementation. The examples provided in the SDK show only the basics of what is involved.
Sample code
z
BasicShape specializes kSplineItemBoss and shows how to implement the interfaces
involved in drawing a page item.
z
CandleChart specializes kSplineItemBoss to show how to draw some data as a chart.
Custom unit-of-measure service
Plug-ins can extend the set of measurement units supported by the application by defining a
unit-of-measure service. The following steps outline what a plug-in needs to do to make a custom unit of measure appear in the Edit > Preferences > Units & Increments dialog box:
1. Define a unit-of-measure service boss class.
2. Implement the C++ code for IUnitOfMeasure, using CUnitOfMeasure as a base.
Layout Fundamentals
203
Layout Fundamentals
Commands that manipulate page items
3. Implement IK2ServiceProvider using the default kUnitOfMeasureService provided by the
API, which registers the custom unit with the application-measurement system.
4. Include a ruler-resource description of the ODFRez type RulerDataType for your custom
unit of measure in the .fr resource file.
Rulers
Rulers are displayed by the ruler widgets (kHorzRulerWidgetBoss and kVertRulerWidgetBoss).
The presentation of the ruler is described by the data defined in the ODFRez resource type,
RulerDataType. This type describes the font, associated unit-of-measure service ClassID, and
tick-mark layout. For more details, see RulerType.fh.
Typically, a ruler is associated with a particular unit-of-measure service. The resource description for the ruler is in the same plug-in as the unit-of-measure service. The unit-of-measure
service returns the resource ID of its ruler description in its implementation of IUnitOfMeasure::GetRulerSpecRsrcSpec.
When a new view is created on a document, the preferred measurement unit is determined.
Then, the associated unit-of-measure service is queried for the resource ID of its ruler description. If one exists, that description is used to create a new ruler as part of the new view of the
document.
Sample code
CustomUnits creates a custom unit. In addition, it provides a resource description of how the
ruler tick markers are specified for the custom units. CstUniRuler.fr from the sample shows a
sample ruler resource.
Commands that manipulate page items
Page-item creation commands
TABLE 37 Commands that create page items
204
Command
Description
kCreateMultiColumnItemCmdB
oss
Creates a new text frame.
kImportAndLoadPlaceGunCmd
Boss
Combination of kImportPIFromFileCmdBoss and kLoadPlaceGunCmdBoss.
kImportAndPlaceCmdBoss
Combination of kImportPIFromFileCmdBoss and kPlacePICmdBoss.
Layout Fundamentals
Commands that manipulate page items
Command
Description
kImportPIFromFileCmdBoss
First, this command imports a file. If what is imported is a story, a
multicolumn item is created. If what is imported is not in a graphic frame, a
new graphic frame is created and the imported file is put into the frame. The
UID of the new item is returned on the command’s item list.
kLoadPlaceGunCmdBoss
Loads the item specified on the command’s item list into the place gun. Only
one item can be loaded by this command.
kNewPageItemCmdBoss
Used to create a new page item. You will need to use this command only if you
are creating a custom page item. Do not use this command to create a pageitem type supplied by the API, a kSplineItemBoss, kImageItem, or
kMultiColumnItemBoss. Each page-item type provides a command to create
it. The newly created page item can be any simple page item that supports the
IGeometry interface. All attributes for the new page item are stored in the
INewPageItemCmdData interface. The UID of the new page item is returned
in the command’s item list. If the parent is supplied, the new page item is
added to the hierarchy.
kPlaceGraphicFrameCmdBoss
Creates a graphic frame and places an image item into it. The item to be placed
can be passed through the command’s item list or placed from the place gun.
To use the place gun, set the usePlaceGunContents parameter on the
IPlacePIData interface to kTrue, and do not set the command's item list. When
passing the UID of the placed item using the command's item list, set the
usePlaceGunContents parameter to kFalse.
kPlaceItemInGraphicFrameCmd
Boss
Places a page item into an existing graphic frame. The page item can be passed
into the command’s item list, or it can be placed from the place gun. The UID
of the graphic frame is returned on the command’s item list.
kPlacePICmdBoss
Places a page item. This command does not create a new page item. It
command takes an existing page item from the command’s item list or the
place gun, then places the page item into the page item hierarchy.
kReplaceCmdBoss
Replaces one page item with another. Both the old page item and the new page
item are specified in the IReplaceCmdData interface.
Page-item update commands
TABLE 38 Commands that add/remove a page item to/from a hierarchy
Command
Description
kAddToHierarchyCmdBoss
Adds the page item in the command’s item list to the parent specified in the
IHierarchyCmdData interface. The index position for the page item also is
specified on the IHierarchyCmdData interface. For example, when a new
page item is created, it calls this command if the new item’s parent was
supplied.
kRemoveFromHierarchyCmdBoss
Removes a page item from its parent when deleting a page item or detaching
the imported content from the imported frame.
Layout Fundamentals
205
Layout Fundamentals
Commands that manipulate page items
TABLE 39 Commands that position page items
Command
Description
kAlignCmdBoss
Aligns the page items in the command’s item list to the alignment type
specified in the IIntData interface. The command itself is a compound
command made up of a set of move-relative commands. Since the move
commands are undoable, the align command also is undoable.
kCenterItemsInViewCmdBoss
Moves a set of page items so they appear centered in the current view. This
command is not undoable.
kFitPageItemInWindowCmdBoss
Used to make the current selected page items fit in a window. This command
is not undoable.
kMoveToLayerCmdBoss
Moves the page items specified on the command’s item list to the document
layer specified on the IUIDData interface. The page items appear in the same
z-order relative to each other as before in front of the new layer. This
command does not move the items that are children of a group or a graphic
frame, nor does it move standoffs.
kMoveToSpreadCmdBoss
Like kMoveToLayerCmdBoss, except kMoveToSpreadCmdBoss moves the
page items from one spread to another.
TABLE 40 Commands for content fitting
Command
Description
kAlignContentInFrameCmdBoss
Aligns the contents in the command’s item list to the specified corners of
their respective frames. The corner, which is the same for all items, is
specified on the IAlignContentInFrameCmdData interface. This command
ignores frames without content. The command’s item list holds the UIDs of
the contents, not the UIDs of the frames that contain the contents. This
command sends out a notification that the content’s subject has changed.
kCenterContentInFrameCmdBoss
Positions the contents at the centers of their respective frames. The content
UIDs are stored in the command’s item list.
kFitContentPropCmdBoss
Takes a list of frame content and modifies the content size and
transformation to fit the frames and maintain the content’s proportions.
kFitFrameToContentCmdBoss
Resizes each frame in the command’s item list to fit its content’s path
bounding box. This command takes a list of page items, filters out the empty
frame and non-graphical frame, and transforms the frame's path geometry.
The text frames are still in the command’s item list, even though they are not
affected by this and other commands.
NOTE:
206
Sometimes, after replace or paste operations, the graphic frame content (e.g., the placed
image, PDF, or EPS item) does not fit into the frame or is not at the position you want
inside the frame. Unfortunately, there is no graphic attribute or helper function that can
determine the fitting mode for a page item. You can compare the bounding boxes for
the content (for example, the image item) and the parent (the frame), to decide whether
Layout Fundamentals
Commands that manipulate page items
the content should be positioned in the center, fit into the frame, or aligned to the
corner of the frame using a content-fitting command. Because these commands are
one-time actions and not attributes, however, these commands must be executed again
if the page items are resized.
TABLE 41 Commands that designate the type of a frame
Command
Description
kConvertFrameToItemCmdBoss
Takes a list of graphic frames and text frames with no content and converts
them to graphic items (unassigned), which are neither text frames nor
graphic frames. This command sets the graphic-frame attribute of the
specified page items to kFalse.
kConvertItemToFrameCmdBoss
The opposite of kConvertFrameToItemCmdBoss.
kConvertItemToFrameCmdBoss sets the graphic-frame attribute of the
specified page items to kTrue and converts unassigned graphic items to
graphic frames. It command also converts empty text frames to empty
graphic frames.
kConvertItemToTextCmdBoss
Converts unassigned page items and empty graphic frames to empty text
frames.
TABLE 42 Commands that copy and paste page items
Command
Description
kCopyCmdBoss
A generic command for copying page items from one document to another. This
command applies to any page item that supports the IScrapItem interface.
kCopyImageItemCmdBoss
Copies the specified image item to the specified target. Both the image item and
the target are specified in the ICopyCmdData interface.
kCopyPageItemCmdBoss
Copies one or more page items to the specified parent object, if any. A UIDList of
the page items created is returned in the command’s item list.
kDuplicateCmdBoss
Duplicates the items specified in the command’s item list. The page items in the
command's item list are duplicated if the command is prepared successfully.
kPasteCmdBoss
Pastes one or more page items to the specified database in the command’s
ICopyCmdData interface. The UIDList of the items to paste are passed to the
command in the ICopyCmdData interface.
kPasteGraphicItemCmdBoss
Used for pasting EPS, image, and PDF items into the destination database. The
difference between kPasteGraphicItemCmdBoss and kCopyPageItemCmdBoss is
that if the specified parent is a graphic frame, the graphic item (EPS, image, PDF)
is created as a child of the frame. If no frame is specified as the parent,
kPasteGraphicItemCmdBoss creates a new rectangular frame as its parent and
applies all its transformation values to the frame, so it appears in the correct
location.
kPasteInsideCmdBoss
Pastes the specified content into a frame and, if necessary, deletes previous content.
The pasted content is aligned to the top left.
Layout Fundamentals
207
Layout Fundamentals
Commands that manipulate page items
TABLE 43 Commands that transform page items
Command
Description
kTransformPageItemsCmd
Boss
Performs various page-item transformations, such as move, rotate, scale, and skew.
The caller needs to pass in appropriate data.
kTransformPathPointsCmd
Boss
Transforms path points of page items. The transformation could be move, rotate,
scale, skew, etc. The caller needs to pass in appropriate data.
NOTE:
We strongly recommend using ITransformFacade and ITransformSuite for your
transformation needs.
Although other transform commands are still available, they may be removed in the
future, so we strongly recommend you use only the two commands in Table 43.
TABLE 44 Commands that group page items
NOTE:
Command
Description
kGroupCmdBoss
Groups selected items in the list and creates a new page item consisting of the group. The
parent for the group is the same as the parent of the last item in the sorted selected-item list.
The UID of the new group page item is returned in the command’s item list. This command
requires that the items to be grouped have valid parents, because only items at the same
level in the hierarchy can form a group. For example, it is invalid to try to group an image
with the spline item that contains it (or any other splines not at the same level as the image
of the document hierarchy).
kUngroupCmdBoss
Ungroups the items specified in the command’s item list. The parent of the group becomes
the parent of all ungrouped items.
TABLE 45 Commands that lock page items
208
Command
Description
kSetLockPositionCmdBoss
Used to lock/unlock the page items. (See the ILockPosition interface.)
Layout Fundamentals
Commands that manipulate page items
Page-item deletion commands
TABLE 46 Commands that delete page items
Command
Description
kDeleteCmdBoss
Deletes the page items specified in the command’s item list. Any groups left empty
by the deletions also are deleted.
kDeleteFrameCmdBoss
Deletes both the frame and its contents. In general, you should avoid directly
calling this command. Instead, use the GetDeleteCmd method of the IScrapItem
interface to delete the frame.
kDeletePageItemCmdBoss
Deletes the page items specified in the command’s item list and removes the items
from the hierarchy. In general, you should avoid directly calling this command.
Instead, use the GetDeleteCmd method of the IScrapItem interface to delete the
page item.
kDeleteImageItemCmdBoss
kDeleteImageItemCmdBoss is a subclass of kDeletePageItemCmdBoss. The
purpose of kDeleteImageItemCmdBoss is to release an image object when the page
item is deleted. This command deletes the image items on the command’s item list.
If the ClassID of any item is not kImageItem, that item is not deleted.
kDeleteLayerCmdBoss
Deletes the layer specified on the command’s item list. The layers in the item list
must be in the same document. After this operation, all page items on the layer list
also are deleted.
kDeletePageCmdBoss
Deletes the pages specified on the command’s item list. Page items on the pages
also are deleted. If the spread is left without pages by this command, the spread also
is deleted. If page reshuffling is specified on the IBoolData interface, the remaining
pages in the affected spread and pages in the spreads that follow are reallocated so
each spread has the number of pages specified in the page set-up preferences for
the document.
kDeleteSpreadCmdBoss
Deletes one or more spreads and all their pages and page items. If the deletion
would leave the document without spreads, the command is aborted and nothing is
deleted. The list of spreads to delete is specified in the command’s item list.
kDeleteUIDsCmdBoss
Deletes the UIDs specified on the command’s item list.
kRemoveInternalCmdBoss
Used when you delete a page item with embedded data. This command removes
the embedded data from the publication.
Layout Fundamentals
209
Layout Fundamentals
Commands that manipulate page items
210
Graphics Fundamentals
Paths
Graphics Fundamentals
This chapter explains how the appearance of page items is specified, how page items are drawn,
and how you can customize how page items are drawn.
The objectives of this chapter are as follows:
z
Show how paths are defined and manipulated.
z
Illustrate graphic page-item structure and how to import and export graphics files.
z
Examine how colors and related properties of graphics are represented.
z
Introduce graphic attributes.
z
Describe stroke-related effects, like custom path strokers.
z
Introduce transparency and transparency effects.
z
Outline how spreads and page items are drawn to the screen and printed.
z
Describe extension patterns to let you customize how page items are drawn.
For definitions of terms, see the “Glossary.”
Paths
This section shows how a path is defined and drawn according to its control points.
Path concepts
Paths
Paths can be created using various tools in InDesign. Shape tools and Frame tools create closed
paths, the Pencil tool creates continuous smooth paths with lots of anchor points, and the Pen
tool allows you create paths with great precision.
A path page item may have one or more paths. For example, a line, a rectangle, or an oval consists of one path; a compound path, two or more paths.
Path points
A path is defined by its path points. A path may have one or more path points. Path points are
represented by PMPathPoint objects. There are three types of path points:
z
kL, line point — Used to form a straight line. These (and kCK path points) are called corner
points.
z
kCS, continuous smooth point — Used to form a continuous smooth curve. These are called
smooth points.
z
kCK, continuous unsmooth point — The point’s left and right tangents are not on the same
line. These (and kL path points) are called corner points.
Graphics Fundamentals
211
Graphics Fundamentals
Paths
A PMPathPoint object stores three PMPoint objects, represented as pairs of (x,y) coordinates.
These PMPoint objects represent the anchor point, the left-direction point, and the right-direction point, respectively (see Figure 77). For a kL path point, the left- and right-direction points
are the same as the anchor point.
Figure 76 shows examples of path points. In part A, the line consists of two kL points. In part B,
the middle kCK point has two direction points that are not on a line with the anchor point. In
part, the middle kCS point’s anchor point and two direction points form a line.
FIGURE 76
Path-point types
Path drawing
A path is controlled by its path points. InDesign draws Bezier curves based on the path points.
Figure 77 shows an open path with six path points.
FIGURE 77
Path with six path points
D
th
ep
of v e
r
cu
212
Graphics Fundamentals
Paths
In the figure, note the following:
z
Every anchor point (1, 2, 3, 4, 5, and 6) is on the path.
z
Points 21, 31, 41, 51, and 61 are left-direction points. Points 22, 32, 42, 52, and 62 are rightdirection points.
z
Direction lines (from anchor point to direction point) always are tangent to (perpendicular
to the radius of) the curve at the anchor points.
z
The angle of each direction line determines the slope of the curve, and the length of each
direction line determines the height, or depth, of the curve.
z
A path is drawn segment by segment. The shape of each segment is determined by its two
end anchor points and their direction points. For example, curve (1,2) is determined by
anchor points 1 and 2 and left-direction point 21; curve (2,3) is determined by anchor
points 2 and 3, right-direction point 22, and left-direction point 31.
If a path is closed, the last segment is controlled by the last and first path points.
For more details, see PMBezierCurve.h.
Winding rule
The winding rule determines whether a given point is inside or outside the area defined by a
path. InDesign supports both even-odd and non-zero winding rules:
z
Even-odd winding rule — If a ray drawn from a point in any direction crosses the path an
odd number of times, the point is inside; otherwise, the point is outside.
z
Non-zero winding rule — The crossing count for a ray is the total number of times the ray
crosses a left-to-right portion of the path minus the total number of times the ray crosses a
right-to-left portion of the path. If a ray drawn from a point in any direction has a crossing
count of zero, the point is outside; otherwise, the point is inside.
For details, see the book PostScript Language Reference, available on the Adobe Web site.
The winding rule of an object is stored as a graphic attribute of the kGraphicStyleEvenOddAttrBoss page item; therefore, you can get or set the winding rules through IGraphicsAttributeUtils the same way as other graphic attributes. For more information, see “Graphic attributes”
on page 256.
Paths data model
Path page-item structure
Paths can exist on any page item that defines an IPathGeometry interface. Figure 78 show the
relationship among page item, path, path point, and point. A path can be viewed as connected
path segments or connected path points.
Graphics Fundamentals
213
Graphics Fundamentals
Paths
FIGURE 78
Path page-item structure
A page item has
at least one path
associated with it.
This package includes all
ancestors of a page item.
See details in "Layout
fundamentals" for information
on navigating to page items.
Document
hierarchy
1owns
1..*
A path consists of
straight or curved
line segments.
IPathGeometry
Path page item
1owns
1..*
owns
Path
A path point defines
an anchor point and
the directions of the
path at the anchor
point.
1
1 owns
0..*
Path segment
1..*
2
1
defined by
Path point
1 owns
A point mathematically
defines coordinates of
a position.
3
A path segment
connects two
path points.
Point
Path geometry
There is no object in the InDesign object model that maps to a single path directly; instead, all
path points of all paths of a page item are stored using the interface defined by the IPathGeometry class. IPathGeometry is the signature interface of a path page item. It has methods to
access path information of a page item, like the number of paths the page item has and whether
a path is open or closed.
Paths are differentiated using a path index. Methods for accessing path points normally require
a path index parameter.
The PMPathPointList type is defined as K2Vector<PMPathPoint>. An item of this type stores a
list of path points; for example, all path points of a path. Unlike with IPathGeometry, path
points stored in an item of type PMPathPointList are not divided into paths by path index.
The most common path page item is the spline item, represented by kSplineItemBoss. Spline
items include paths (lines, curves, frames) like those shown in Figure 76 and Figure 77. There
214
Graphics Fundamentals
Paths
are several other kinds of paths, including image-clipping paths, text-wrap paths, and text outlines. All paths use IPathGeometry to store paths.
Path operations
You can access paths of a page item through the interface defined by the IPathGeometry class.
InDesign also encapsulates some complex manipulations into several high-level path operations, which are provided as commands and selection suites.
Compound paths
You can combine two or more paths, compound paths, grouped page items, text outlines, text
frames, or other shapes that interact with and intercept one another to create a new compound
path. The advantage of a compound path over individual paths is that a compound path is a
single object, so you can apply attributes to the compound path as a whole.
Use kMakeCompoundPathCmdBoss to create a new compound path and kReleasePathsCmdBoss to split a compound path into individual paths.
NOTE:
A compound path does not connect or join any two points of existing paths; it only puts
two paths together in one page item.
Comparing compound paths and groups
Both compound paths and groups involve combining multiple page items as a whole. The main
difference between compound paths and groups is that making compound paths combines all
items together into one page item (all other original page items are deleted), whereas grouping
does not delete the original page items; they are just moved into the new group item. See
Figure 79. The steps in the figure are explained below.
FIGURE 79
Compound path contrasted with group
(1) original red
rectangle and
yellow oval
Graphics Fundamentals
(2) result of
compound path
(3) result of
group with 50%
opacity applied
215
Graphics Fundamentals
Paths
In Figure 79:
1. The original page items are a red rectangle and a yellow oval, represented by kSplineItemBoss object A and kSplineItemBoss object B, respectively.
2. To make a compound path out of these two objects, to take the path points from kSplineItemBoss object B and add them to kSplineItemBoss object A through the IPathGeometry interface, then delete kSplineItemBoss object B. The result is only one kSplineItemBoss
object, A, with new path points. The new item retains the first item’s graphic attributes,
including fill color. All path operations use the even-odd winding rule, resulting in the hole
in the compound path.
3. To group the two page items instead, InDesign first creates a kGroupItemBoss object, then
removes both kSplineItemBoss object A and kSplineItemBoss object B from their current
hierarchical parents and adds them as hierarchical children to kGroupItemBoss. The result
is three different boss objects. The graphic attributes of the original items are retained. You
can apply graphic attributes to a group page item or its children independently.
Path-finder operations
Path-finder operations, also referred to as compound-shape operations, combine shapes
enclosed by the paths. InDesign supports adding, subtracting, intersecting, excluding-overlap,
and minus-back operations. Figure 80 illustrates the results of these path finder operations.
FIGURE 80
Path-finder operations
(1) original red
rectangle and
yellow oval
(4) result of path
finder subtract
operation
(2) result of path
finder add operation
(3) result of path
finder intersect
operation
(5) result of path
finder exclude
overlap operation
(6) result of path
finder minus back
operation
You can use kMergePathCmdBoss or IPathOperationSuite to achieve the effect.
Path-finder operations require exactly two input objects. The attributes of the front object are
retained by the result object, except in the case of the subtract operation.
216
Graphics Fundamentals
Graphic page items
Shape conversion
Shape conversion changes the current shape to a new shape. InDesign can convert to various
shapes, including lines, triangles, rectangles with corner effects, ovals, and polygons. Figure 81
shows the result of converting a rectangle and an oval to inverse-rounded rectangles.
FIGURE 81
Convert shape
(1) original red
rectangle and
yellow oval
(2) result of converting
shape to inverserounded rectangle
The command to achieve this effect is kConvertShapeCmdBoss. It gets the bounding box of the
current shape, then inserts or changes path points in the IPathGeometry object to represent the
desired shape, so the resulting page item maintains the same bounding box as the original
shape.
IConvertShapeSuite takes shape type, corner effects, and several other parameters. You can use
this method to convert a shape to a new shape with a new combination of attributes, like shape
type, corner effects, and number of polygon sides.
NOTE:
If you convert a page item from shape A to shape B, then convert from shape B back to
shape A again, the number of path points and their coordinates may be different than
for your original shape A.
Graphic page items
This section discusses the concepts and data model for graphic page items, as well as graphic
page-item special operations, like setting clipping paths and text-wrap contours. This section
also covers importing and exporting graphics files.
Graphic page-item types
Computer graphics fall into two main categories, vector graphics and raster images.
InDesign can import raster images, vector graphics, or a combination of both, depending on
the graphics-file format.
Graphics Fundamentals
217
Graphics Fundamentals
Graphic page items
Raster images
Raster images, or bitmap images, are the most common electronic medium for such continuous-tone images as photographs or images created in painting programs like Adobe Photoshop®. Raster images are resolution dependent.
InDesign supports almost all popular raster-image formats, including PSD, JPEG, TIFF, GIF,
PNG, BMP, and Scitex, all of which are represented by kImageItem boss, which is discussed in
more detail in “Graphic page-item boss-class inheritances” on page 223.
An alpha channel is an invisible channel that defines transparent areas of a graphic. The alpha
channel is stored in a graphic with the RGB or CMYK channels. A user can create alpha channels using background-removal features in Photoshop.
Vector graphics
Vector graphics use geometrical formulas to represent images. Paths created using the drawing
tools in the InDesign Toolbox are examples of vector graphics.
Vector graphics are more flexible than bitmaps, because vector graphics can be resized without
losing resolution. Another advantage of vector graphics is that their representations often
require less memory than representations of bitmap images.
NOTE:
Since most output devices are raster devices, vector objects must be translated into
bitmaps before being printed or displayed. For example, PostScript printers have a
raster image processor (RIP) that performs the translation within the printer.
Vector-graphic formats can differ dramatically. InDesign supports import and export of most
common vector graphics (see Table 47).
TABLE 47 InDesign-supported vector-graphics formats
File format
Import
Export
Related page-item boss
DCS
Yes
No
kEPSItem
EPS
Yes
Yes
kEPSItem
PDF
Yes
Yes
kPlacedPDFItemBoss
PICT
Yes
No
kPICTItem
SVG
No
Yes
N/A. You can export various content or
documents to SVG file format, but there is no
page item for SVG format.
WMF
Yes
No
kWMFItem
For more information on importing vector graphics, see “Graphics import” on page 238. For
more information on exporting vector graphics, see “Export to graphics file format” on
page 243.
218
Graphics Fundamentals
Graphic page items
Graphic page-item settings
Content fitting
Sometimes the size of the graphic page item is not the same as the size of the parent graphics
frame, causing the graphic to be displayed incorrectly. Usually, end users can move, resize, or
rotate the graphic page item or the graphics frame in the same way as any other page item.
Taking advantage of the relationship between the graphic page item and the graphics frame,
InDesign has several content-fitting options:
z
Fit content to frame — Resizes the graphic page item to the same size as the graphics frame.
z
Fit frame to content — Resizes the graphics frame to the same size as the graphic page item.
z
Center content in frame — Moves the graphic page item to the center of the graphics frame.
No resizing is involved.
z
Fit content proportionally — Resizes the graphic page item to barely fit the frame (i.e., the
graphic page-item edge touches the frame vertically or horizontally), while maintaining the
aspect ratio of the graphic page item.
z
Fill frame proportionally — Resizes the graphic page item while maintaining its aspect ratio
until all white space in the frame is filled.
If there is a selection, we recommend you use IFrameContentSuite or IFrameContentFacade
where possible to perform these operations, though individual commands for these operations
do exist.
NOTE:
These transformations of graphic page items and frames change only the IGeometry or
ITransform objects. These transformations do not alter the structure of the graphic
page item and frame objects.
Clipping paths
A clipping path crops part of the image so only part image appears through the shape you create. The clipping path is stored separately from its graphics frame, so you can freely modify one
without affecting the other. The clipping path does not change the image; the clipping path
affects only how the image is drawn.
Clipping paths can be created in the following ways:
z
Use existing paths or alpha channels. You can add paths and alpha channels in Photoshop
files.
z
Let InDesign generate a clipping path with Detect Edges.
z
Draw a path in the shape you want, and paste the graphic into the path.
When you set or generate a clipping path, the clipping path is attached to the image, resulting
in an image that is drawn clipped by the path and cropped by the frame.
IPathGeometry aggregated on the graphic page-item bosses stores the clipping path. For more
information on how a path is defined, see “Path geometry” on page 214.
Graphics Fundamentals
219
Graphics Fundamentals
Graphic page items
The IClipSettings interface is aggregated on kWorkspaceBoss, kDocWorkspaceBoss, and each
graphic page-item boss, to define preferences or individual graphic page items’ clipping-path
settings. For more details on IClipSettings, see the API reference documentation.
Not all settings in IClipSettings are required to determine a clipping path. For example, if you
choose the kClipEmbeddedPath type, you need to know only the embedded path index. Settings like threshold, tolerance, and inset are required only to detect edges. We highly recommend you use the IClippingPathSuite selection-suite interface.
Text wrap
You can wrap text around the frame of any page-item object.
The text-wrap data model is discussed in detail in the “Text Fundamentals” chapter. The path
the text wraps around is defined by a separate page item, kStandOffPageItemBoss, which is
accessible through IStandOffData, aggregated both on the graphic-frame boss (kSplineItemBoss) and graphic page-item boss (See Figure 83). Whenever the text-wrap object changes, the
path is copied to the IID_ITEXTWRAPPATH interface (same implementation as IPathGeometry) of the graphic page-item boss; this improves the performance of text composition and
screen redrawing.
What makes the graphic page item special in text wrapping is that you can specify the contour
options when the text-wrap mode is set to Wrap Around Object Shape. In addition to setting
the text-wrap contour to the stroke bounding box of the graphics frame that owns a graphic
page item, you also can specify that text wrap around any of the following:
z
The rectangle formed by the graphic's height and width (the stroke bounding box of the
graphic page item).
z
Paths generated by edge detection. You can adjust edge detection manually, using a clipping
path (“Clipping paths” on page 219).
z
The image’s alpha channel.
z
The image’s embedded Photoshop path.
z
The image’s clipping path.
As with the clipping path setting, contour options for text wraps are specified by IStandOffContourWrapSettings aggregated on kWorkspaceBoss, kDocWorkspaceBoss, and individual
graphic page-item bosses. We highly recommend you use ITextWrapFacade to manipulate text
wraps, including setting contour options.
Display performance
Display performance is used to set the balance between graphic display speed and quality. You
can specify the quality of how raster images, vector graphics, and transparencies are drawn to
the screen. A display performance group is a set of values for these categories.
NOTE:
220
Graphics display performance options do not affect output resolution when exporting
or printing images in an InDesign document. When printing to a PostScript device,
packaging for GoLive, or exporting to EPS or PDF, the final image resolution depends
on the output options you choose when you print or export the file.
Graphics Fundamentals
Graphic page items
The following display performance groups are available:
z
Fast — The highest display speed but the lowest display quality. In the standard setting for
this group, InDesign draws a raster image or vector graphic as a gray box.
z
Typical — The default option for graphic page items. In the standard setting for this group,
InDesign draws a low-resolution proxy image appropriate for identifying and positioning a
graphic.
z
High Quality — The highest quality but the lowest display speed. In the standard setting for
this group, InDesign draws a raster image or vector graphic at high resolution (i.e., uses the
file provided by the graphic page item’s link).
NOTE:
Standard settings of display performance groups can be modified through the Display
Performance area of the Preferences dialog box or programmatically. See the “Graphics”
chapter of Adobe InDesign CS4 Solutions.
These display performance groups are session preferences, represented by IDrawOptions,
aggregated on kWorkspaceBoss (see Figure 82). The interface stores an array of DrawOptionsSet (also defined in IDrawOptions.h), corresponding to the fast, typical, and high-quality
groups, respectively. Each group itself, represented by a DrawOptionsSet, has a set identifier
and defines the qualities for each category, including vector, raster, and transparency. (For
details, see the API reference documentation.) The array of DrawOptionsSet defines a twodimensional map that indicates whether the graphic page item should be displayed as a gray
box, proxy image, or a high-resolution graphic, given a display group (such as fast) and a category (such as raster). You can contain the current active group by calling IDrawOptions::GetActiveSetID, and can set groups through the Edit > Preferences > Display
Performance menu.
Display-performance settings are not document-level preferences; instead, they are set as view
preferences. IDrawOptionsSetID, aggregated on kLayoutWidgetBoss (see Figure 82), allows
each layout presentation to have different default settings, even if they are of the same document. IDrawOptionsSetID stores a DrawOptionsSet as default settings when importing
graphic page items. These settings can be set through the View > Display Performance menu.
Individual graphic page items can have their own display-performance overrides. IDrawOptionOverrides, aggregated on kDrawablePageItemBoss—which is inherited by all graphic page
item bosses (see Figure 82)—stores a display-performance group (DrawOptionsSet) for each
graphic page item. Display performance overrides can be set through the Object > Display Performance menu. The IDisplayPerformanceSuite selection-suite interface facilitates the manipulation of display-performance settings.
NOTE:
To let page item overrides take effect, the IDrawOptions::GetIgnoreOverrides
preferences flag must be kFalse. You can set this flag through the View > Display
Performance menu.
The display-performance object model is summarized in the class diagram in Figure 82.
Graphics Fundamentals
221
Graphics Fundamentals
Graphic page items
FIGURE 82
Display-performance object model
The application
preferences
The layout view
The individual graphic
page item
«boss»
kWorkspaceBoss
IDrawOptions
«boss»
kLayoutWidgetBoss
1
IDrawOptionsSetID
«boss»
kImageItem/...
1
IDrawOptionOverrides
IDrawOptionsSetID::GetID
1
IDrawOptions::GetActiveSetID
1..*
IDrawOptionOverrides::GetDrawOptions
1..*
«struct»
DrawOptionsSet
1..*
Represents a display performance group,
which stores settings for raster images, vector
graphics, transparency, etc.
Drawing a graphic page item
Changing display performance from typical to fast does not replace a proxy image with a gray
box directly. In fact, changing display performance does not change the instance of the graphic
page item at all. The effect is achieved by invalidating the layout and letting the page-item
drawing process take care of displaying.
Drawing a layout containing graphic page items follows a generic drawing process detailed in
“Dynamics of drawing” on page 278. A graphic page item is drawn in the following steps:
1. Set the crop rectangle. This involves calculating the graphic page item’s bounding box,
cropped by the frame bounding box and view bounding box.
2. Clip the graphics with their clipping paths.
3. Determine drawing options and draw the graphics.
You can participate in the graphic page-item drawing process by implementing the custom,
drawing event-handler extension pattern, described in “Extension patterns” on page 289.
222
Graphics Fundamentals
Graphic page items
Graphic page-item data model
This section illustrates the data model and common interfaces for graphic page items.
Graphic page-item boss-class inheritances
Although graphic page items are represented by various boss classes (kImageBoss and bosses
for vector graphics in Table 47), they share similar interfaces and demonstrate similar properties.
kImageItem (the boss representing raster images), kPlacedPDFItemBoss, kEPSItem, kPICTItem, and kWMFItem directly or indirectly inherit from kDrawablePageItemBoss, which
aggregates interfaces necessary for page-item drawing. For the complete inheritance graph for
the bosses, see the API reference documentation; search for kDrawablePageItemBoss.
Graphic page-item class hierarchy
Graphic page-item boss classes have the following characteristics:
z
Are wrapped in a spline item (kSplineItemBoss).
z
Have text wrap around their contour or frame (represented as kStandOffPageItemBoss).
z
Store low-resolution proxy images as kImageInDB.
z
Store a link object (ILinkObject) which keeps the UIDRef(s) of link boss(es). A link boss,
like kImportLinkBoss, links the link object (page item) and external graphic files.
z
May embed link resources internally as kPMPersistDataBlobBoss.
Figure 83 is a simplified view of the graphic page items in the class hierarchy of an InDesign
document. ILinkObject provides access to the link object, which can be used to access the
linked graphic resource. If the link is embedded, IStoreInternal provides access to kPMPersistDataBlobBoss, which stores the bitmap of the image. IImageDataAccess provides access to the
proxy image, and the bitmap of the images also are stored as kPMPersistDataBlobBoss, which
is accessible through IPersistUIDData on kImageInDB. Navigation from the document root to
the kSplineItemBoss is wrapped into a package, which is discussed in the “Layout Fundamentals” chapter.
Graphics Fundamentals
223
Graphics Fundamentals
Graphic page items
FIGURE 83
Graphic page-item class hierarchy
This package includes all
ancestors of a spline page item.
See details in the "Layout
Fundamentals" section about how
to navigate to the page item.
Document
hierarchy
Stores the path
geometry of
text wrap.
1 owns
0..*
«boss»
kSplineItemBoss
The graphic frame
wrapping the image
page item.
IPathGeometry
«boss»
kStandOffPageItemBoss
1
The graphic page items such as
kImageItem, kPlacedPDFItemBoss,
kEPSItem, kPICTItem, kWMFItem,
or other graphic page item bosses.
IHierarchy::QueryChild(0)
1
1
«boss»
Any
graphic page item boss
IGeometry
IPathGeometry
IID_ITEXTWRAPPATH(IPathGeometry)
1
1
kImageInDB stores a lowresolution image (of a
high-resolution image)
IImageDataAccess::
GetLowResImageUID()
IGraphicItem
IImageItem(kImageItem only)
IImageAttributes(kImageItem only)
1
ILinkObject::GetLinks
IStoreInternal::GetStoredUID()
0..1
IImageAttributes
0..1
IStandOffData::GetStandOffUID()
0..*
«boss»
kImageInDB
1
0..1
IPersistUIDData::GetUID()
«boss»
kImportLinkBoss
«boss»
kPMPersistDataBlobBoss
0..1
Stores whole image internally
as a blob. Could be either proxy
image (from IPersistUIDData)
or embedded image (from
IStoreInternal).
Stores the link of
an external graphic
file.
Graphic page-item interfaces
Several common interfaces are aggregated on graphic page-item bosses to provide features specific to graphic page items. These interfaces are either aggregated directly on the boss (e.g.,
kPlacedPDFItemBoss), or aggregated by way of their parent bosses (e.g., kImageBaseItem, the
parent of kImageItem, or kDisplayListPageItemBoss, the parent of all other vector graphic page
item bosses). Table 48 summarizes these important interfaces and their uses.
224
Graphics Fundamentals
Graphic page items
TABLE 48 Graphics-interface summary
Interface ID
Interface
Description
IID_IGRAPHICITEM
IIntData
This is the signature interface for all graphic types (raster,
EPS, etc.). The integer data is never used.
IID_IIMAGEITEM
IImageItem
Aggregated onto kImageBaseItem, this interface provides
access to the embedded color profile of the image. More
importantly, testing the presence of this interface verifies that
a page item boss is a raster image.
IID_IGEOMETRY
IGeometry
This interface defines the geometry (position, size, etc.) of a
graphic page item. It is used for calculating cropping during
the graphic page-item drawing process.
IID_IPATHGEOMETRY
IPathGeometry
This interface stores the clipping path of a graphic page item
(see “Clipping paths” on page 219).
IID_ITEXTWRAPPATH
IPathGeometry
This interface stores the text-wrap path; however, this path is
only a cache (or mirror) of the text-wrap path. The real path
is stored at IPathGeometry on kStandOffPageItemBoss. Every
time the real text-wrap path is changed, this path is updated.
(See “Text wrap” on page 220.)
IID_IIMAGEATTRIBUTES
IImageAttributes
This interface stores a set of tags to describe the attributes of a
raster image. Predefined tags are defined as enumerators,
each of which describes the type, length, and data of a raster
image. These tags are crucial in defining image item
properties, such as alpha channel, clipping path, and color
profile.
Proxy images
When a graphic is placed, the contents of the original file are not actually copied into the document. Instead, InDesign creates a link to the original file on the disk and adds a screen-resolution bitmap image (the proxy image) to the layout, so you can view and move the graphic.
When you export or print, InDesign uses the link to retrieve the original graphic, creating the
final, desired, full-resolution output. Although end users can set preferences to tell InDesign to
show high-resolution images (see “Display performance” on page 220), by default InDesign
shows a low-resolution proxy image for performance reasons; the proxy image is created by
processing kCreateLowResImageCmdBoss during the import process.
A proxy image is represented by the boss kImageInDB, which is accessible from the graphic
page item through IImageDataAccess::GetLowResImageUID. The connection is established
during the image-import process. For the data model of the proxy image in relation to the
graphic page items, see “Graphic page-item class hierarchy” on page 224 and “Graphic pageitem examples” on page 226.
One of the most important interfaces on kImageInDB is IPersistUIDData, which holds a UID
pointing to an instance of kPMPersistDataBlobBoss, which stores the bitmap of the proxy
image.
Graphics Fundamentals
225
Graphics Fundamentals
Graphic page items
The proxy image is not always created by asking the image-import filter for the data. For raster
images, if image auto-embedding is allowed in the image-import preferences, and the image is
smaller than a predefined amount (defaults to 48 KB, accessible through ILinkState::GetEmbedSize), the original high-resolution image is embedded directly.
Creating a proxy image is aided by the IImageStreamManager interface, which provides a
mechanism to convert the source-image format to a destination-image format according to the
image attributes.
Graphic page items and links
Every graphic page item that was placed from a file has an associated link. End users can
choose the Embed Link menu in the Links panel’s fly-out menu to embed the link manually.
The file remains in the Links panel, marked with an embedded link icon.
NOTE:
This behavior differs from link for a text file. You can create a link when placing a text
file; however, since text is inheritently part of the InDesign document, you cannot really
“embed” the link. If you create a link when placing a text file, you can choose to unlink
the file from the Links panel. Once it is unlinked, the link is removed from the Links
panel.
As a software developer, you also can use ILinkFacade::EmbedLinks to embed the image link
(ILinkFacade is aggregated on kUtilsBoss). kStoreInternalCmdBoss, processed internally, reads
the original high-resolution file and stores the image data as a data blob, referenced by the
IStoreInternal interface on the graphic page item.
Embedding a link is not the same as embedding the small image to create a proxy image, discussed in “Proxy images” on page 225:
z
The image data is stored in a different place. Embedding a link stores a data blob referenced
by IStoreInternal on the graphic page item. Embedding an image as a proxy stores the data
blob referenced by IPersistUIDData on kImageInDB boss.
z
Embedding a link does not cause a check for the size of the image.Embedding an image as a
proxy requires the image file to be smaller than a predefined size.
Graphic page-item examples
New graphics can be placed into a layout with or without an existing graphics frame. This section illustrates how instances of graphics objects evolve as you place a PDF file, replace it with
an image, move the image around, embed a link, and set clipping-path and text-wrap contour
options.
226
Graphics Fundamentals
Graphic page items
Begin with an empty graphics frame
We start exploring the graphics object model by examining an empty graphics frame. Figure 84
shows a screenshot and the object model when only an empty graphics frame is created. The
graphics frame is represented by kSplineItemBoss, and it does not have any hierarchical children.
FIGURE 84
Empty graphics frame
Above:
Empty links panel.
Right:
Empty graphic frame.
Below:
A graphic frame
represented as
kSplineItemBoss.
No child object.
Document
hierarchy
«boss»
graphic-frame
: kSplineItemBoss
The graphic frame. Its
pasteboard coordinates
are shown.
LeftTop = (36, -360)
RightBottom = (276, 0)
Graphics Fundamentals
227
Graphics Fundamentals
Graphic page items
Place a PDF item in a graphics frame
Next, we place a PDF file into the graphics frame. During import, InDesign creates a kPlacedPDFItemBoss object to represent the PDF item and adds it as a child of the graphics frame.
Figure 85 shows the object model of the graphics frame with a PDF item. A few other objects
are created to represent information of the PDF item. New objects are marked in yellow.
Comparing Figure 85 to Figure 84, notice the Links panel shows a link to the external PDF file.
The link is represented by kImportLinkBoss through ILinkObject, aggregated on kPlacedPDFItemBoss.
A proxy image is created for the PDF item as kImageInDB boss (UID 0xcd), and the proxy
image bitmap is represented by kPMPersistDataBlobBoss (UID 0xce).
By default, a graphic page item is put into the graphics frame with an offset of (0,0) in inner
coordinate space, so the PDF item has the same PMRect::LeftTop position (obtained from IGeometry::GetPathBoundingBox) on the pasteboard as the graphics frame. Although the PDF
item is slightly wider than the frame (PMRect::RightBottom member), InDesign does not draw
that part, because it is cropped by the frame.
228
Graphics Fundamentals
Graphic page items
FIGURE 85
PDF item in graphics frame
Area not
cropped
by frame
Above:
A PDF file appears
in the Links panel.
Right:
PDF item placed in
the graphic frame.
Below:
The graphic frame
now has a PDF item
represented by
kPlacedPDFItemBoss.
Document
hierarchy
The graphic frame
is not changed.
«boss»
graphic-frame
: kSplineItemBoss
LeftTop = (36, -360)
RightBottom = (276, 0)
The PDF item is the only child
of the graphic frame. It is
wider than, but shorter than
the graphic frame.
A proxy image is
created for the
PDF file.
«boss»
PDF
: kPlacedPDFItemBoss
«boss»
proxy image
: kImageInDB
UID = 0xcb
LeftTop = (36, -360)
RightBottom = (298, -200)
UID = 0xcf
UID = 0xcd
«boss»
image blob
: kPMPersistDataBlobBoss
«boss»
link : kImportLinkBoss
A proxy image is
stored as a data
blob.
Link maintains a
connection to
external PDF file.
UID = 0xce
Graphics Fundamentals
229
Graphics Fundamentals
Graphic page items
Replace a PDF item with an image
Next, we import an image file to replace the PDF item and examine the changes to the object
model. Figure 86 shows the screenshot and object model after the change. Again, new objects
are marked in yellow. Note the following:
230
z
The Links panel shows a link to the new image file (PalmTreeCMYK.tif).
z
The overall structure of the object model did not change, but a new boss object kImageItem
replaced kPlacedPDFItemBoss.
z
The UID of the link, proxy image, and proxy image data blob also changed. These objects
were created during the import process—and the old objects were deleted.
Graphics Fundamentals
Graphic page items
FIGURE 86
Imported image
Above:
New image appears
in the Links panel.
Right:
Image replaced PDF item
in the graphic frame.
Below:
The graphic frame
now has an image
represented by
kImageItem.
Document
hierarchy
The graphic frame
is not changed.
«boss»
graphic-frame
: kSplineItemBoss
LeftTop = (36, -360)
RightBottom = (276, 0)
The image is the only child
of the graphic frame. It is
larger than the graphic frame,
thus only a portion is shown.
A proxy image is
created for the
the image file.
«boss»
image
: kImageItem
«boss»
proxy image
: kImageInDB
UID = 0xb3
LeftTop = (36, -360)
RightBottom = (418, 302)
«boss»
link : kImportLinkBoss
UID = 0xb7
UID = 0xb4
«boss»
image blob
: kPMPersistDataBlobBoss
A proxy image is
stored as a data
blob.
Link maintains a
connection to
external image file.
UID = 0xb5
Graphics Fundamentals
231
Graphics Fundamentals
Graphic page items
Move an image within a frame
You may have noticed the palm tree image is large, and we can only see a small part of the tree.
Now we move the image a little bit within the graphics frame, using the Direct Selection tool.
Figure 87 shows the resulting object model. Everything is the same, except the pasteboard
coordinates of the image item. The changed object is marked by a red circle; changed values are
marked in yellow. The coordinates represent the bounding box of the image item.
232
Graphics Fundamentals
Graphic page items
FIGURE 87
Image moved within a frame
Right:
The image is moved to
a new location within
the graphic frame.
The graphic frame
is not changed.
Document
hierarchy
A proxy image is
created for the
the image file.
«boss»
graphic-frame
: kSplineItemBoss
LeftTop = (36, -360)
RightBottom = (276, 0)
«boss»
proxy image
: kImageInDB
UID = 0xb4
«boss»
image blob
: kPMPersistDataBlobBoss
UID = 0xb5
«boss»
image
: kImageItem
After the move, the image has
new pasteboard coordinates.
«boss»
link : kImportLinkBoss
UID = 0xb7
UID = 0xb3
LeftTop = (-84, -528)
RightBottom = (298, 134)
Link maintains a
connection to
external image file.
A proxy image is
stored as a data
blob.
Graphics Fundamentals
233
Graphics Fundamentals
Graphic page items
Embed an image’s datalink
Next, we manually embed the image into the document by choosing Embed Link on the Links
panel fly-out menu. Figure 88 shows the changed object model after the image is embedded.
The new object is within the red circle, marked in yellow. A new kPMPersistDataBlobBoss is
created to store the embedded image information. Other bosses—including the link object
(kImportLinkBoss)—do not change.
There are two kPMPersistDataBlobBoss objects in the object model, but they represent different things. The one accessible through IPersistUIDData on the proxy image:kImageInDB boss
represents the proxy image; the other—accessible through IStoreInternal on the image:kImageItem boss—represents the embedded image data.
234
Graphics Fundamentals
Graphic page items
FIGURE 88
Embedded image
The image is embedded
into the InDesign file. The
look and feel of the image
does not change.
Document
hierarchy
The graphic frame
is not changed.
A proxy image is
created for the
the image file.
«boss»
graphic-frame
: kSplineItemBoss
LeftTop = (36, -360)
RightBottom = (276, 0)
Embedding the image does not
change pasteboard coordinates.
«boss»
proxy image
: kImageInDB
«boss»
image
: kImageItem
UID = 0xb4
UID = 0xb3
LeftTop = (-84, -528)
RightBottom = (298, 134)
1
IPersistUIDData::GetUID()
1
IStoreInternal::GetStoredUID()
1
1
«boss»
image blob
: kPMPersistDataBlobBoss
«boss»
embedded image blob
: kPMPersistDataBlobBoss
UID = 0xb5
UID = 0xd8
A proxy image is
stored as a data
blob.
Graphics Fundamentals
«boss»
link : kImportLinkBoss
UID = 0xb7
Link maintains
a connection to
external image file.
Embedded image blob (orignal
graphic file data), accessible
through IStoreInternal.
235
Graphics Fundamentals
Graphic page items
Setting clipping path and text wrap
Let’s use a different image to set clipping-path and text-wrap options. In this section, we place a
linked image, overlap it on top of a text frame, and examine how the image and text frame
change when you set clipping-path and text-wrap options.
Figure 89 shows examples of clipping-path options. In the left screenshot, no clipping path was
set, so the image overlaps the background text. The center screenshot shows the result of the
clipping-path Detect Edges option (with the Threshold parameter set to 150). Although Detect
Edges can detect most of the contours of the bottle, the result looks even better if you use the
existing Photoshop paths, as shown in the right screenshot.
FIGURE 89
Image clipping
path set to None
Clipping-path options
Clipping path is generated
by detecting edges.
Clipping path set to
existing Photoshop
path.
Reminder: The clipping path is stored with the IPathGeometry interface of the graphic page
item, so no extra object is created. For the graphic page item’s class diagram, see Figure 83; for
descriptions, see “Clipping paths” on page 219.
Now we set the text wrap of the image item, shown in Figure 90. All three screenshots were
taken with the text wrap set to wrap around the graphic page item’s object shape. The left
screenshot shows the result when the text-wrap contour option is set to Same As Clipping, with
clipping-path type set to Detect Edges. The top-right corner of the bottle frame does not have
text, because a small part of the image is there. You can set the contour option to use existing
Photoshop paths independent of the clipping path, shown in the center screenshot. You also
can set both the clipping path and text-wrap contour options to use the Photoshop path, which
produces the best result, as shown in the right screenshot.
Object model changes after setting the text-wrap options are shown in the lower part of
Figure 90. A new kStandOffPageItemBoss object—within the red circle, marked in yellow—is
created to hold the text-wrap path. For information on how this affects text composition, see
the “Text Fundamentals” chapter.
236
Graphics Fundamentals
Graphic page items
FIGURE 90
Text-wrap options
Clipping path is set to
detect edges, and
text wrap is set to
Same As Clipping
The graphic frame
Clipping path is generated
by detecting edges.
Text wrap is set to use
Photoshop path
Both clipping path and
text wrap are set to
use existing Photoshop
path.
«boss»
graphic-frame
: kSplineItemBoss
LeftTop = (156, 123)
RightBottom = (288, 336)
The image item
A proxy image is
created for the
the image file.
«boss»
proxy image
: kImageInDB
UID = 0xd2
«boss»
image
: kImageItem
UID = 0xd1
LeftTop = (156, 123)
RightBottom = (308,350)
«boss»
image blob
: kPMPersistDataBlobBoss
«boss»
text wrap
: kStandOffPageItemBoss
UID = 0xd3
UID = 0xd7
A proxy image is
stored as a data
blob.
Graphics Fundamentals
«boss»
link : kImportLinkBoss
UID = 0xd4
Link maintains
a connection to
external image file.
The standoff object, which
stores the path geometry of
the text wrap.
237
Graphics Fundamentals
Graphic page items
Graphics import
Overview of the import process
When a user chooses to place a graphic into a document, InDesign performs a sequence of
steps to import the graphics file as a graphic page item. Figure 91 shows the steps involved in
importing PDF, EPS, and image files.
You also can import another InDesign document to a document. The import process works
like PDF import. The only difference is importing InDesign document also imports swatches,
inks, fonts and links, etc. and creates secondary links if the document being imported also contains links. You can set InDesign document-import options like other graphic file types; see
“InDesign document-import preferences” on page 242.
238
Graphics Fundamentals
Graphic page items
FIGURE 91
Graphics-import process
Create a link that connects the page item and link resource
Choose import provider based on file type
[Vectors (except PDF)]
[PDF]
[Raster Images]
Choose image read format provider
Create kPlacedPDFItemBoss
New kImageItem
Create EPS, DICT or WMF item
Set PDF import prefs
Set image IO prefs
Set EPS import prefs
Set PDF attributes
Set Image attributes
Set EPS attributes
Generate raster preview
Create kImageInDB
Create TIFF proxy kImageInDB
Stream into kImageInDB
Embed image?
[no]
[yes]
Sample high-resolution image
Read and set proxy image's data blob
Graphics Fundamentals
239
Graphics Fundamentals
Graphic page items
At a high level, the process involves the following steps:
1. Create a kImportLinkBoss that connects the link resource and the page item.
2. Query for an import provider that provides kImportProviderService for the graphic. This
service also is used for importing other types of files, like plain-text files. For graphics files,
one of four providers is called: kPDFPlaceProviderBoss, kImagePlaceProviderBoss,
kEPSPlaceProviderBoss or kNativeImportServiceBoss. See Table 49 for details on file formats supported by these providers.
3. Create the appropriate new page item according to the type of the graphics file. See
Table 49.
4. Import the file and set new page-item attributes, like IPDFAttributes, IEPSAttributes, and
IImageAttributes. This may involve searching for an appropriate image import filter
(“Image-import filters” on page 242) and setting import options (“Import options” on
page 240).
5. Create and set a low-resolution proxy image for the new page item, which is accessible
through IImageAccess on the new page item.
TABLE 49 Supported graphic-file formats for import
Import provider
Filename extension
Page-item boss
kEPSPlaceProviderBoss
AI, EPS, DCS
kEPSItem
kEPSPlaceProviderBoss
PCT, PIC
kPICTItem
kEPSPlaceProviderBoss
WMF, EMF
kWMFItem
kImagePlaceProviderBoss
TIF, TIFF, SCT, PSD, PNG,
PCX, JPEG, JPG, GIF, DIB
kImageItem
kPDFPlaceProviderBoss
PDF
kPlacedPDFItemBoss
kNativeImportServiceBoss
INDD
kInDesignPageItemBoss
Import options
To control how the file should be imported, end users can choose Show Import Options in the
Place dialog box, which allows the user to set import options through an Import Options dialog box. The specific options depend on the file type and graphics data in the file.
PDF-import preferences
The IPDFPlacePrefs (IID_IPDFPLACEPREFS) interface allows you to set various PDF import
options, like how to crop the graphic and which pages to import. If the PDF file contains layers,
you can set up layer information on the dialog using IGraphicLayerInfo.
Both interfaces are aggregated on kWorkspaceBoss. IGraphicLayerInfo can be initialized by the
imported file before bringing up the Import Options dialog. Table 50 lists some default settings
of IPDFPlacePrefs.
240
Graphics Fundamentals
Graphic page items
For details, see the “PDF Import and Export” chapter.
TABLE 50 Selected default PDF-import preferences
Preference
Default value
Crop
kCropToContent
Page number
1
Proxy resolution
72
Show preview
kTrue
Transparent background
kTrue
EPS-import preferences
IEPSPreferences (IID_IEPSPREFERENCES) also is aggregated on kWorkspaceBoss. Some settings are set programmatically; for example, importing an EPS graphic from a file (not from
the clipboard) automatically sets the import mode to “import whole” (kImportWhole).
Table 51 lists some default settings of IEPSPreferences.
TABLE 51 Selected default EPS-import preferences
Preference
Default value
Import mode
kImportWhole
Display resolution
72
Read OPI comments
kDontReadOPIComments
Create frame
kDontCreateFrameFromClipPath
Create proxy
kCreateIfNeeded
NOTE:
“Create frame” corresponds to the Apply Photoshop Clipping Path checkbox, and
kAlwaysCreate corresponds to the Rasterize The PostScript radio button in the EPS
Import Options dialog box.
Image-import preferences
The Raster Image Import Options dialog box uses separate panels to set raster image-specific
import options, like color profile and clipping path. Usually, InDesign uses session preferences
(IImageIOPreferences on kWorkspaceBoss) to import an image file. Table 52 lists some of the
default settings.
Graphics Fundamentals
241
Graphics Fundamentals
Graphic page items
TABLE 52 Selected default image-import preferences
Preference
Default value
Allow auto-embedding
kTrue
Create clip frame
kTrue
Preview resolution
72
When auto-embedding is allowed and the image size is small, InDesign automatically embeds
the high-resolution image as the proxy image, so the data and the interfaces of kImageInDB
reflect the high-resolution image rather than a new, low-resolution image.
InDesign document-import preferences
Placing InDesign document directly into an InDesign document has advantages in the publishing workflow, including automatically maintained links, fonts, swatches, links and so on. The
IImportDocOptions (IID_IIMPORTDOCOPTIONS) interface allows you to set various InDesign document-import options, like how to crop the pages, which pages to import, and where
to display previews. IGraphicLayerInfo let you set up layer information on the dialog to choose
which layer to import.
Both IImportDocOptions and IGraphicLayerInfo are session preferences aggregated on
kWorkspaceBoss. They also store information to pass to kImportDocCmdBoss, if you want to
use the command directly to place InDesign files. Table 53 lists some default settings of IImportDocOptions.
TABLE 53 Selected default InDesign document-import preferences
Preference
Default value
Crop
kCropToPage
Page number
1
Show preview
kTrue
Image-import filters
You may have noticed in Table 49 that the same kImagePlaceProviderBoss is used for importing various kinds of raster images. These raster images have different file formats and, therefore, they need different ways to read data. InDesign provides another level of abstraction: the
image-import provider queries for service providers that support kImageReadFormatService.
This mechanism also serves as another extension pattern software developers may implement
to support new raster-image formats.
InDesign implements a set of providers to read various image formats, to import TIF (TIFF),
SCT, PSD, PNG, PCX, JPEG (JPG), GIF, and DIB files. These providers serve as the imageimport filters. When InDesign is instructed to place an image, it iterates over all the providers
until it finds one that can import the image. See Table 54 for a list of these providers.
242
Graphics Fundamentals
Graphic page items
TABLE 54 Image read-format providers
File type
Image read-format provider
TIF
kTIFFImageReadFormatBoss
SCT
kSCTImageReadFormatBoss
PSD
kPSImageReadFormatBoss
PNG
kPNGImageReadFormatBoss
PCX
kPCXImageReadFormatBoss
JPEG
kJPEGImageReadFormatBoss
GIF
kGIFImageReadFormatBoss
DIB
kDIBImageReadFormatBoss
Export to graphics file format
InDesign supports exporting all or part of a document to PDF, EPS, SVG, and JPEG formats.
All the export processes are alike at a high level: exports are implemented using the IExportProvider service-provider mechanism (see the “Service Providers” chapter of Learning the
Adobe InDesign CS4 Architecture). Each type of export destination format provides export services. Table 55 lists export providers that export to a graphics file format.
TABLE 55 Graphic-export providers
Export destination
Export provider
PDF
kPDFExportBoss
EPS
kEPSExportBoss
SVG
kSVGExportProviderBoss
JPEG
kJPEGExportProviderBoss
For examples of export provider implementations, see SDK samples like CHMLFilter, TextExportFilter, and XDocBookWorkflow.
Each export provider has a corresponding command that exports content to a desired file format in the following two stages:
1. Drawing — Drawing page items to an intermediate entity, such as a viewport.
2. Streaming — Writing the content to the provided file stream.
Export to PDF and EPS are discussed in detail in the “Exporting to EPS and PDF” section of
the “Printing” chapter. PDF import and export also are discussed in the “PDF Import and
Export” chapter.
Graphics Fundamentals
243
Graphics Fundamentals
Graphic page items
Only SVG and JPEG export are discussed in this section. Figure 92 shows high-level activities
of SVG and JPEG export.
FIGURE 92
SVG and JPEG export process
Get export file format
Choose export provider
[SVG]
244
[JPEG]
Set SVG export prefs
Set JPEG export prefs
Create kSVGExportViewPortBoss and set attributes
Create SnapshotUtilsEx
Install draw event handler
Draw page items on SnapshotUtilsEx
Set up SVGWriter
Create output stream
Draw page items
Export to JPEG stream
Graphics Fundamentals
Graphic page items
SVG export
SVG export preferences
Settings for SVG export are stored with the ISVGExportPreferences interface on kWorkspaceBoss. For details, see the API reference documentation.
End users can modify the preferences through the SVG Export Options dialog box. You also
can change the preferences programmatically, with the kSVGExportSetPrefsCommandBoss
command. Table 56 lists some of the default SVG-export preferences.
TABLE 56 Selected default SVG-export preferences
Preference
Description
Default value
Embed fonts
Whether to embed fonts in
the SVG file
kTrue
Embed image
Whether the image should be
embedded in the SVG file
kTrue
Export bitmap sampling
Sampling quality
kHiResSampling
Image format
Image format
kDefaultImageFormat,
which is set to the same as
kPNGImageFormat
JPEG quality
Image quality
kJPEGQualityMed
Page item export
Whether to export the
selected page item only
kFalse
Range format
The range of pages to export
kAllPages
Exporting to SVG format
InDesign exports to SVG format by drawing page items and documents to a special viewport,
kSVGExportViewPortBoss. For more information about viewports, see “Data model for drawing” on page 276. kSVGExportCommandBoss aggregates ISVGExportController, which controls the drawing process and hides much of the complexity involved in SVG export.
kSVGExportViewPortBoss aggregates ISVGWriterAccess, which helps in writing contents to
an SVG file stream.
JPEG export
JPEG export preferences
Settings for JPEG export are stored with the IJPEGExportPreferences interface on kWorkspaceBoss. For details, see the API reference documentation.
End users can modify the preferences through the JPEG Export Options dialog box. You also
can change the preferences programmatically, with the kJPEGExportSetPrefsCommandBoss
command. Table 57 lists some of the default JPEG-export preferences.
Graphics Fundamentals
245
Graphics Fundamentals
Colors and swatches
TABLE 57 Selected default JPEG-export preferences
Preference
Description
Default value
Export bitmap sampling
Sampling quality
kHiResSampling
JPEG quality
Image quality
kJPEGQualityMed
Range format
The range of pages to export
kAllPages
Exporting to JPEG format
kJPEGExportCommandBoss creates a SnapshotUtilsEx object and draws document contents
on it. kJPEGExportCommandBoss then calls SnapshotUtilsEx::ExportImageToJPEG to write
to the file stream. SnapshotUtilsEx also has methods to export to TIFF and GIF formats, so you
can export to TIFF and GIF if you write your own export command, though you may need to
provide your own image-write format boss. In the case of JPEG, the write-format boss provided is kJPEGImageWriteFormatBoss. For more information on SnapshotUtilsEx, see “Snapshots” on page 285. For use of SnapshotUtils (the older version of SnapshotUtilsEx, see the
Snapshot SDK sample plug-in).
NOTE:
Multiple JPEG files are created for documents with multiple pages.
Colors and swatches
This section examines the data model for swatches, colors, and gradients.
Architecture
This section covers the representation of color, a fundamental graphic property, in the InDesign API. When we talk about color in the context of InDesign, we refer to the following:
246
z
A color system refers to how colors and gradients are represented and printed and how to
achieve consistent color throughout different applications and different devices. See “Color
management” on page 252.
z
A rendering attribute of a page item refers to how a color or gradient can be used to render
a page item’s graphic attributes, like stroke and fill. See “Graphic attributes” on page 256.
Graphics Fundamentals
Colors and swatches
Figure 93 shows some key color-related concepts and how they appear in the user interface.
FIGURE 93
Color concepts
Color-related
panels
Linear
gradient
Graphics Fundamentals
Radial
gradient
Color
Tint
Color
applied
Swatches
247
Graphics Fundamentals
Colors and swatches
Figure 94 shows some of the key color-related concepts and the relationships among them.
FIGURE 94
Conceptual model for color
A swatch is a rendering object
that can render document
objects. It can be normal color,
tint, mixed ink, or gradient.
Color could be of type
process or spot.
Gradients are continuous, smooth
color transitions along a vector
from one color to another. It could
be linear or radial type.
swatch
«subtype»
«subtype»
«subtype»
1
gradient
«subtype»
1..*
based on
1
«subtype»
-Composite
«subtype»
color
1
radial
«subtype»
linear
tint
spot
Tint is a percentage
of a color.
process
A mixed ink consists of at
least one process color ink
and at least one spot color.
based on 1..*
11
based on
1..*
mixed ink
Figure 95 is a UML class diagram involving color-related boss classes. It shows the relationship
among color and swatch–related boss classes. The relationships also apply if you replace kDocWorkspaceBoss with kWorkspaceBoss.
248
Graphics Fundamentals
Colors and swatches
FIGURE 95
Color-related boss-class diagram
IID_ISWATCHLIST and IID_IUICOLORLIST share
the same interface ISwatchList, but have different
implementations (kSwatchListImpl and
kUIColorListImpl respectively).
kPMColorBoss from IID_ISWATCHLIST is a swatch,
kPMColorBoss from IID_IUICOLORLIST is a UI color.
The document workspace
«boss»
kDocWorkspaceBoss
1
1
11
1
IID_ISWATCHLIST
«boss»
kGraphicStateNoneRenderingObjectBoss
0..*
1
IID_IINKLIST
IID_IUICOLORLIST,
IID_ISWATCHLIST
«boss»
kAGMBlackBoxRenderingObjectBoss
0..*
1
«boss»
kGradientRenderingObjectBoss
0..*
0..*
*
«boss»
kPMColorBoss
1
1
IGradientFill::GetNthGradientColorUID()
1
*
*
«boss»
kPMInkDataBoss
* IInkResources::GetUsedInks().
**
These four bosses
represent swatches.
Comparing Figure 95 to the to the conceptual diagrams, note the following:
z
The kDocWorkspaceBoss is used here to show where these boss classes belong. These interfaces also exist on kWorkspaceBoss.
z
The boss objects of type kPMColorBoss and kGradientRenderingObjectBoss and other
classes represent swatches directly. IRenderingObject is the signature interface.
z
kPMColorBoss implements solid color, tint, mixed ink, process color, and spot color.
When swatches are created, modified, or deleted, the state of the swatch list changes. For a
detailed discussion of these changes, see “Swatch-list state” on page 296.
Graphics Fundamentals
249
Graphics Fundamentals
Colors and swatches
Swatches
A swatch is an abstraction that can represent a color, gradient, tint or mixed ink.
A rendering object implements the IRenderingObject interface. IRenderingObject represents
information about a color or gradient and exposes capabilities to render this information to a
device. The following are some of the boss classes that expose the IRenderingObject interface:
z
kPMColorBoss represents colors (including tint and mixed ink)
z
kGradientRenderingObjectBoss represents gradient.
z
kGraphicStateNoneRenderingObjectBoss represents the absence of rendering information;
for example, this boss can represent no-fill or no-stroke attributes. This boss is used internally.
z
kAGMBlackBoxRenderingObjectBoss is used for special page items created from Adobe
Illustrator® clipboard format. It is only for internal use.
Some methods of IRenderingObject take an IGraphicsPort item as an argument. IGraphicsPort
is an interface with methods that parallel PostScript operators like setcolorspace. The key
responsibility of the rendering object implementation is to set up the color space, tint, and
color components to draw to the graphics port. IRenderingObjectApplyAction is another
required interface on rendering objects. For more details on IRenderingObject, IRenderingObjectApplyAction, IGraphicsPort, and the other interfaces on rendering object boss classes, see
the API reference documentation.
Solid colors
Color is represented by the kPMColorBoss class, responsible for representing solid colors,
including tints and mixed inks. In addition to the signature interface for a rendering object,
IRenderingObject, kPMColorBoss aggregates several other interfaces, including the following:
z
IColorOverrides — Stores tint and color remarks so the kPMColorBoss also can represent
tint and special color types (e.g., reserved color). For details, see the API reference documentation.
z
IInkData — Stores the ink information used by the color. InkType represents color type
(process or spot).
z
IColorData — Represents color coordinates in a ColorArray vector and a color space.
There are three color spaces of interest:
z
CMYK defines a color with four components: cyan, magenta, yellow, and black. Typically,
these components are represented by percentages.
z
RGB represents colors with three components: red, green, and blue. Typically, these components are represented by values between 0 and 255.
z
LAB (more precisely, L*a*b*) is a device-independent color space
See “Color management” on page 252.
250
Graphics Fundamentals
Colors and swatches
Gradients
Gradients are represented by the kGradientRenderingObjectBoss boss class, which aggregates
IRenderingObject as well as IGradientFill, representing gradient-specific properties. For
details, see the API reference documentation.
Gradients can be linear or radial; see the enum declaration of GradientType in GraphicTypes.h.
A gradient consists of one or more ranges of color, which are smooth interpolations between
gradient stop positions. The color is defined only at the gradient stop positions, with calculated
color values at intermediate positions. For more detail on working with gradients, see “Gradients” in InDesign Help.
Swatch lists
Rendering objects (IRenderingObject) are referenced in a swatch list (ISwatchList). The swatch
list (ISwatchList) is exposed on workspaces (IWorkspace). For examples of the swatch list (and
ink list) when an end user creates and applies swatches to page items in a document, see
“Swatch-list state” on page 296.
If you create a color through the Color panel, or create a new gradient through the Gradient
panel and then apply the gradient to a page item without creating a swatch beforehand, an
unnamed (or local) swatch is created in the swatch list. The unnamed swatches do not appear
in the Swatches panel, but they can be used for the strokes and fills of document objects by client code.
User-interface color list
Colors and color names that appear in the application user interface in places like the Layers
panel and Tags panel are not the same ones used in representing color in the document. Userinterface colors also are stored in an ISwatchList interface on workspaces (IWorkspace), but
with the interface identifier IID_IUICOLORLIST.
User-interface colors are normally RGB only. A key utility interface for working with userinterface colors is IUIColorUtils. User-interface colors are represented by kUIColorDataBoss
rather than a rendering-object boss class.
Inks
Colors are ultimately printed by means of inks; for example, CMYK color separation is performed before printing to a four-color press. Inks are represented by the kPMInkDataBoss boss
class.
The rendering object classes like kPMColorBoss or kGradientRenderingObjectBoss aggregate
the IInkResources interface, with which you can refer to the inks used in a color (or other page
item) by using the IInkResources::GetUsedInks method, then querying for IPMInkDataBoss.
IInkData::GetInkUIDList returns an empty list when the color type is process. IInkData is an
interface on kPMColorBoss.
Graphics Fundamentals
251
Graphics Fundamentals
Color management
There is an IInkList interface on the workspace boss classes that specifies the inks needed for
all swatches in the document and application. The ink list contains process inks and zero or
more spot inks.
The relation between inks and the swatch list is fairly direct: adding a new spot color to the
swatch list and applying it to a document object results in a new entry in the ink list. However,
creating another swatch that is a process color and applying it does not necessarily change the
entries in the ink list, if the process inks needed to print this new color already are in the ink
list.
The ink-manager feature (Ink Manager in the Swatches panel menu) shows the inks associated
with the contents of the swatch list. For example, if the swatches in the document were defined
in terms of four process colors and two spot colors, the ink-manager user interface is similar to
that shown in Figure 96. For information on how to work with the ink manager and iterate
through the ink list, see the “Graphics” chapter of Adobe InDesign CS4 Solutions.
FIGURE 96
Ink Manager dialog box
It is necessary to preflight documents to ensure they print correctly when submitted to the
press. As far as color is concerned, it is necessary to identify the inks that must be loaded in the
press for a job to print correctly.
Color management
The purpose of color management is to get accurate color on all kinds of devices. Color-management modules provide color-conversion calculations from one device’s color space to
another, based on ICC device profiles. For detailed descriptions of the color spaces used by
InDesign, see “Color spaces” on page 300.
252
Graphics Fundamentals
Color management
ICC profiles
An ICC profile is a record of the unique color characteristics of a color input or output device.
An ICC profile contains data for the translation of device-dependent color to L*a*b* color. The
following illustrates the concept of translation from the L*a*b* color space to RGB and CMYK
color spaces using ICC profiles:
z
L*a*b* = RGB + ICC profile
z
L*a*b* = CMYK + ICC profile
With an ICC profile, color-management systems can do the following:
z
Translate a user-defined, device-specific color to a device-independent, L*a*b* color. For
example, if the user creates an RGB color for a specific computer monitor, that color could
be translated to L*a*b*.
z
Translate a L*a*b* color to a device-specific color. For example, when printing to a printer,
the L*a*b* color is translated to a CMYK color.
Color-management workflow
Color management within InDesign is relatively complex because of the following:
z
InDesign supports multiple color spaces per document. For example, a user could assign
text in one paragraph an RGB color, another paragraph a CMYK color, and so on.
z
InDesign supports external links (by means of ILink) to assets that can have different color
settings than those used in the document.
There are three levels of ICC profile sets involved in InDesign:
z
Image ICC profile — This is embedded in the image file when the image is created.
z
Document ICC profile — This is associated with a particular InDesign document.
z
Application ICC profile preference — This is the default profile of the application. The
default profile is assigned to any new document.
InDesign features relating to color management are described comprehensively in “Color Management” in InDesign Help.
Data model for color management
The implementation of color management is somewhat complex. The class diagram in
Figure 97 illustrates some color management-related boss class relationships in InDesign.
Graphics Fundamentals
253
Graphics Fundamentals
Color management
FIGURE 97
Color management
Application
preferences
The document
workspace
<<boss>>
kWorkspaceBoss
<<boss>>
kDocWorkspaceBoss
1
1
The color profile. Note: although
both application-preference and
document-side bosses
(kDocumentPresentationBoss and
kImageBaseItem) can refer to a
profile by UIDRef, they are pointing
to different objects, because they
are at different databases.
1
ICMSProfileList
IColorPresetManager
0..*
IColorPresetSettings
<<boss>>
kCMSProfileBoss
1..*
1
ICMSProofing
<<boss>>
kDocumentPresentationBoss
1..*
1..*
ICMSProfileUID
Color
preset
files
Color preset files are
installed during
InDesign installation
and shared with other
Adobe applications.
1
<<boss>>
kImageBaseBoss
Document window for
proofing
Images may
each have their
own color
profile.
Color presets
Outside the application, predefined Adobe color-management settings are stored in color preset files, which are saved at the following location:
Windows:
Program Files\Common Files\Adobe\Color\Settings\
Mac OS:
Library/Application Support/Adobe/Color/Settings
These presets are shared among all Adobe Creative Suite® applications, to make it easier to
achieve consistent color across the suite. The kWorkspaceBoss boss class aggregates the IColorPresetsManager interface, which is responsible for managing the presets, like loading color pre-
254
Graphics Fundamentals
Color management
set files and saving customized presets. The settings of the current preset are populated into the
IColorPresetsSettings interface on kWorkspaceBoss.
The IColorPresetsSettings interface stores the working color-management settings of the application, like RGB and CMYK, the ICC profile, and color-management policies. In the user
interface, the color-management settings are accessed through the Color Settings dialog box.
Color profile
The kCMSProfileBoss boss class represents a color profile. The ICMSProfile interface on kCMSProfileBoss stores information about the type and category of the profile. This interface has a
method for getting and setting profile data by accessing the IACESpecificProfileData interface
on the same boss.
The ICMSProfileList interface, exposed on workspaces (IWorkspace), indicates which profiles
are used, as either document profiles or image profiles, and which are assigned to specific document categories (e.g., document CMYK and document RGB). A profile can be assigned to
each category of a document through Edit > Assign Profiles.
Image settings
Any imported image can have either an embedded profile or a profile already assigned to it.
(See Object > Image Color Settings.) The ICMSItemProfileSource interface aggregated on the
kImageBaseItem boss class stores the basic profile-assignment type (see the enumeration ICMSItemProfileSource::ProfileSourceType) and the name of the image's embedded profile, if any.
If an image uses an external profile or an embedded profile, an interface IPersistUIDData (with
interface identifier IID_ICMSPROFILEUID) stores the UID of an instance of kCMSProfileBoss, so you will be able to instantiate an interface on kCMSProfileBoss from this UID and get
the profile data needed.
There is no link from a specific profile to the images that use that profile. To find the images
that use a specific profile in the profile list, you need to iterate through all images using the
links manager (ILinksManager).
Rendering intent
The ICMSSettings interface is aggregated on docworkspace (kDocWorkspaceBoss) and the
kImageBaseItem boss class, a superclass for image boss classes, which stores rendering-intent
information for document and image page item. Please use the ICMSUtils interface to set rendering intent, rather than manipulating the ICMSSettings interface directly.
For application defaults, rendering intent are determined by IColorPresetsSettings interface
aggregated on kWorkspaceBoss. You can use GetIntent and SetIntent methods to manipulate
rendering intent.
Proofing
The ICMSProofing interface is aggregated on kWorkspaceBoss, kDocumentPresentationBoss,
and several user-interface panels. The implementation on kWorkspaceBoss is a proofing preference, and the implementation on kDocumentPresentationBoss is responsible for setting up
drawing when users choose to proof their design using the Proof Setup dialog box. This ICMSProofing interface stores the proofing settings as well as a proofing profile.
Graphics Fundamentals
255
Graphics Fundamentals
Graphic attributes
Graphic attributes
A graphic attribute describes an aspect of how a page item should be drawn. There are many
kinds of graphic attribute for properties like stroke color, fill color, stroke type, and stroke
weight. This section introduces the concepts involved and describes specific sets of graphic
attributes for features like page-item rendering.
Graphic-attribute data model
Figure 98 is a conceptual model of the graphic attributes associated with a page item.
FIGURE 98
Graphic-attribute conceptual model
A page item that carries
a various number of
graphic attributes
Page item
1
owns
A graphic attribute controls
one aspect of the drawing of
an object.
0..*
Graphic attributes
«subtype»
«subtype»
«subtype»
«subtype»
Stroke color
Stroke color
controls the
color of stroke.
Stroke style
Stroke style
attribute controls
the style of
stroke (dotted
line, etc.).
...
Other various
attributes
Gradient fill
Gradient fill
controlls fill
of the object.
Transparency
Transparency
controls the
transparecy
effect.
A page item owns a collection of graphic attributes, specifying properties that control how the
page item is drawn, like stroke color, stroke style, and transparency. Each attribute is responsi-
256
Graphics Fundamentals
Graphic attributes
ble for a single aspect of how the page item is drawn. A list of graphic attributes is represented
by IGraphicStyleAttributeBossList.
Figure 99 shows the graphic-attribute UML class model of a drawable page item.
FIGURE 99
Graphic attributes on a page item
IID_IGRAPHICSTYLE_IS_APPLICABLE
can answer whether an attribute is
applicable to the page item.
Drawable page items such as
Spline, Group, and Image have
graphic attributes applied.
1
1
IID_IGRAPHICSTYLE_REFERENCE
IGraphicStyleReference
«boss»
kDrawablePageItemBoss
1
1
IID_IGRAPHICSTYLE_IS_APPLICABLE
IGraphicStateIsApplicable
1
«boss»
kGraphicStyleBoss
IID_IGRAPHICSTYLE_OVERRIDES
IGraphicStyleAttributeBossList
IID_IGRAPHICSTYLE_DESCRIPTOR
IGraphicStyleDescriptor
1
1
IID_IGRAPHICSTYLE_DEFINITION
IGraphicStyleAttributeBossList
«partial-implementation»
IGraphicStyleAttributeBossList
1
1
1
IGraphicStyleAttributeBossList
is a list of individual graphic
attribute boss objects.
Iterating method
1
0..*
Every individual attribute
class inherits from
kGraphicsAttrBoss.
«boss»
Individual kGraphicsAttrBoss
IID_IGRAPHICSTYLE_DESCRIPTOR
can be used to query an attribute
or return a list of attributes applied
to the page item.
There are several important interfaces on a page item responsible for managing graphic
attributes:
z
IGraphicStateIsApplicable.
z
IGraphicStyleAttributeBossList (IID_IGRAPHICSTYLE_OVERRIDES), which stores a list
of graphic attributes that override those stored in a graphic style.
Graphics Fundamentals
257
Graphics Fundamentals
Graphic attributes
z
IGraphicStyleAttributeBossList.
z
IGraphicStyleDescriptor.
z
IGraphicStyleReference.
For details on each interface’s responsibilities, see the API reference documentation.
Representation of graphic attributes
Graphic attributes are represented by boss classes with the signature interface IGraphicAttributeInfo. The graphic attributes in the application appear in the API reference documentation for IGraphicAttributeInfo. Typically, they derive from the kGraphicsAttrBoss boss class.
The IGraphicAttributeInfo interface stores the name and type of the attribute. Some graphic
attributes have corresponding text and table attributes. See “Mapping graphic attributes
between domains” on page 260.
Graphic attributes have additional data interfaces that specify their values. Because there are
many graphic attributes, sometimes it is a challenge to determine which attribute boss corresponds to which attribute. For a description of attributes, see “Catalog of graphic attributes” on
page 301.
Most of the attributes represent a single value, like a boolean, int16, int32, PMPoint, or
PMReal. These simple attributes generally are backed by interfaces like IGraphicAttrBoolean,
IGraphicAttrInt16, IGraphicAttrInt32, IGraphicAttrPoint, and IGraphicAttrRealNumber. For
example, kGraphicStyleEvenOddAttrBoss (IGraphicAttrBoolean) corresponds to the even-odd
fill rule when you have a compound path; set it to kTrue to use the even-odd rule or to kFalse
to use the default InDesign fill rule (non-zero winding fill rule).
Some other attributes have UID values. For example, attributes related to object rendering refer
to a rendering object through the IPersistUIDData interface. For more information, see “Rendering attributes” on page 261.
Some other attributes have ClassID values, which are all stored with IGraphicAttrClassID.
These attributes are used to specify stroke effects of a page item. See “Stroke effects” on
page 262.
Graphic styles
A graphic style (kGraphicStyleBoss, UID-based) is a collection of graphic attributes. The
IGraphicStyleNameTable interface exposed on workspaces (IWorkspace) stores the names and
UIDs of graphic styles.
There are two key interfaces on kGraphicStyleBoss:
z
IGraphicStyleInfo stores the name and UID for the style on which this style is based (the
style from which this style inherits).
z
IGraphicStyleAttributeBossList (IID_IGRAPHICSTYLE_DEFINITION) stores the list of
graphic attributes this graphic style owns.
Only [No Style] is used in the InDesign CS4 and InCopy CS4 implementations. kGraphicStyleNoneBoss implements the default [No Style], which is set up during application start-up (for
258
Graphics Fundamentals
Graphic attributes
the session workspace) and document creation (for document workspace). This style stores
default settings of various graphic attributes, like stroke weight as 1.0 and stroke type as solid
line.
Although there still is a public API to work with user-named graphic styles, these are deprecated since the introduction of object styles. Adobe applications do not define user-named
styles using graphic styles as the container. We recommend you avoid using graphic styles for
this purpose in your plug-ins. User-named styles are supported by the object styles feature.
An object style is a style that defines the appearance of a page-item object. An object style contains settings of a set of attributes, including graphic attributes and text attributes. An object
style is a superset of graphic style.
Graphic state
Graphic state is the representation of the current state of things related to the graphic
attributes. Usually this is a collection of graphic attributes for the current selection or, if there is
no selection, the current defaults. The graphic state is represented by kGraphicStateBoss.
An active graphic state refers to the graphic state at the current context. Active graphic state can
be accessed through IGraphicStateAccessor, aggregated on kDocWorkspaceBoss or kWorkspaceBoss. A reference to an interface of the active graphic state also can be acquired through
IGraphicStateUtils::QueryActiveGraphicState, aggregated on kUtilsBoss.
The kGraphicStateBoss boss class aggregates interfaces like IGraphicStateRenderObjects and
IGraphicStateData.
IGraphicStateRenderObjects
This interface is key to the graphic state. It can be acquired from the IGraphicStateUtils utility
interface by querying active graphic state or the graphic state of a specific database. For sample
code that shows how to acquire the interface, see the “Graphics” chapter of Adobe InDesign CS4
Solutions.
Given this interface, you can query the current IRenderingObject interface for the specified
rendering class or the active rendering class in the graphic state. The active rendering could be
fill or stroke.
Changing the graphic state
Changing any graphic attributes changes the graphic state. Virtually every user-interface
action—like selecting a new swatch, picking a color in the Color Picker panel, switching fill and
stroke, selecting a different stroke style, changing selection, or changing the front document—
causes a change to current active graphic state. You can use methods on IGraphicAttributeUtils
or IGraphicStateUtils to acquire appropriate commands.
kGraphicStateBoss also aggregates an ISubject interface. Changes to the graphic state can be
observed by listening along the IID_IGRAPHICSTATE_RENDEROBJECTS and
IID_IGRAPHICSTATE_DATA protocols.
Graphics Fundamentals
259
Graphics Fundamentals
Graphic attributes
Creating new page items using default attributes
When a new page item is created, default attributes in the graphic-state boss objects are copied
and applied to new page items when you specify INewPageitemCmdData::kDefaultGraphicAttributes as attrType parameter on spline creation methods on IPathUtils.
Mapping graphic attributes between domains
Some graphic attributes also can be applied to text or tables. For example, Figure 100 shows
stroke and fill colors and gradients applied to text, table cells, and a spline item. Although the
intent is the same in each case, the attributes differ in each case. Text and tables have some
attributes corresponding to the graphic attributes.
FIGURE 100
Stroke and fill attributes
kGraphicStyleStroke
RenderingAttrBoss
(blue)
kTextAttrStrokeColorBoss
(yellow)
kCellStrokeAttrDataBoss
(red)
kGraphicStyleFill
RenderingAttrBoss
(Gradient)
kTextAttrColorBoss
(Cyan)
kCellAttrFillColorBoss
(green)
There are independent attributes for each of these domains. For example, an attribute represented by kGraphicStyleFillRenderingAttrBoss on a spline page item is represented by the
kCellAttrFillColorBoss type when the attribute is applied to a table cell, and it is represented by
the kTextAttrColorBoss type when the attribute is applied to a text run.
When a graphic attribute is applied to a text run, there is a conversion process in which a new
text attribute is applied as an override to the run. The same applies when trying to apply a
graphic attribute to a table; the graphic attribute can be converted to a corresponding table
attribute.
Not all graphic attributes apply to text or tables. The IGraphicAttributeInfo::IsTextAttribute
and IGraphicAttributeInfo::IsTableAttribute methods return kTrue if the corresponding
attribute in the respective domain exists. IGraphicAttributeInfo::CreateTextAttribute returns a
260
Graphics Fundamentals
Rendering attributes
text attribute that maps onto the equivalent graphic attribute. Similarly, IGraphicAttributeInfo::CreateTableAttribute returns a table attribute that maps onto the equivalent
graphic attribute. Each of these methods is expected to return a non-zero value if the
Is<XXX>Attribute method returns kTrue.
The selection suites (e.g., IGraphicAttributeSuite and IGradientAttributeSuite) make it
straightforward for client code to change the properties of a selection; the client code needs to
work with only the graphic attributes. There is no need to worry about whether the selection is
text, choosing the specific text attribute, and whether the selection is text, table, or layout.
The mapping between different domains is hard-coded. New graphic attributes added by thirdparty software developers may provide their own mapping by overriding the default implementations of IGraphicAttributeInfo. The mappings between graphic attributes and text
attributes are shown in Table 85. The mappings between graphic attributes and table attributes
are shown in Table 86. Note that the mapping between graphic attributes and table attributes is
many-to-one rather than one-to-one because; for example, the kCellStrokeAttrDataBoss boss
class can represent information about multiple stroke parameters, whereas each of the corresponding graphic attributes refers to one parameter.
Rendering attributes
A rendering attribute refers to a swatch to be used when drawing an aspect of a page item. Rendering attributes are a special kind of graphic attribute that refer to their associated swatch by
UID. The swatch can be a color, gradient, tint, etc.
Color-rendering attributes
A page item’s stroke, fill, or gap can be rendered independently. Table 58 lists color-rendering
attributes.
TABLE 58 Color-rendering attributes
Rendering-attribute class
Description
kGraphicStyleFillRenderingAttrBoss
Fill rendering object
kGraphicStyleFillTintAttrBoss
Fill tint
kGraphicStyleGapRenderingAttrBoss
Gap rendering object
kGraphicStyleGapTintAttrBoss
Gap tint
kGraphicStyleStrokeRenderingAttrBoss
Stroke rendering object
kGraphicStyleStrokeTintAttrBoss
Stroke tint
Graphics Fundamentals
261
Graphics Fundamentals
Stroke effects
Gradient attributes
Page items also can be rendered using gradients; however, multiple graphic attributes are
needed to represent a gradient. For example, you need to specify gradient fill angle, center, and
length. There is no characteristic signature for a gradient attribute other than the boss class
name; for the canonical graphic attributes, the boss-class name is of the following form:
kGraphicStyleGradient<graphic_attribute>AttrBoss.
A list of these boss classes is in the master attributes list in “Catalog of graphic attributes” on
page 301.
You can use the same color-rendering attributes class name to apply gradient attributes to page
items. IRenderingObjectApplyAction::PreGraphicApply automatically forwards the attributes
to appropriate gradient attributes. For example, if you are supplying kGraphicStyleFillRenderingAttrBoss, the method adds attributes including the following:
z
kGraphicStyleGradientFillAngleAttrBoss
z
kGraphicStyleGradientFillGradCenterAttrBoss
z
kGraphicStyleGradientFillHiliteAngleAttrBoss
z
kGraphicStyleGradientFillHiliteLengthAttrBoss
z
kGraphicStyleGradientFillLengthAttrBoss
z
kGraphicStyleGradientFillRadiusAttrBoss
The situation gets more complicated by the possibility of applying gradients to table cells; when
a gradient is applied to a table cell, additional cell-specific attributes are involved. For example,
kCellAttrGradientFillBoss and kCellAttrGradientStrokeBoss are table-cell attributes representing properties of gradients.
Stroke effects
There are several graphic attributes that define the look and feel of the stroke of a page item.
Some of these attributes are straightforward, like kGraphicStyleStrokeWeightAttrBoss, which
defines stroke weight.
Path stroker
A path stroker is responsible for the basic drawing of a path. To find existing path-stroker boss
classes, see the API reference documentation for IPathStroker. For information on how to add
custom path strokers, see “Custom path-stroker effects” on page 289.
IPathStrokerList allows you to access the list of available path strokers. IPathStrokerUtils provides utility methods to do the same thing.
The kGraphicStyleStrokeLineImplAttrBoss boss class specifies the class ID of a path stroker
that applies to a particular path.
262
Graphics Fundamentals
Transparency effects
Path corners
A path-corner effect draws the corners of a stroked path. For a list of path-corner-effect boss
classes, see the API reference documentation for IPathCorner. For information on how to add
custom path corners, see “Custom corner effects” on page 290.
The kGraphicStyleCornerImplAttrBoss boss class specifies the class ID of the corner effect.
Path-end strokers
A path-end stroker (IPathEndStroker) draws the start and end of a stroked path. To find pathend strokers, see the API reference documentation for IPathEndStroker. For information on
how to add custom path ends, see “Custom path-end effects” on page 290.
The kGraphicStyleLineEndStartAttrBoss boss class specifies the class ID of the line-end (for
example, arrow) style at the starting side of a line. The kGraphicStyleLineEndEndAttrBoss boss
class specifies the class ID of the line-end style at the terminating side of a line.
Transparency effects
Concepts
Transparency effects are a set of graphic attributes that define color blending for overlapped
page items and backgrounds. Common transparency effects include basic transparency, drop
shadow, and feather. Figure 101 shows examples of transparency effects supported in InDesign.
Graphics Fundamentals
263
Graphics Fundamentals
Transparency effects
FIGURE 101
No effect
Drop shadow
Inner shadow
Outer glow
Inner glow
Bevel & emboss
Satin
Basic feather
Directional feather
Gradient feather
264
Transparency effects supported in InDesign
Graphics Fundamentals
Transparency effects
These transparency effects can be divided into two classes:
z
Outer effects include drop shadow, outer glow, and outer bevel, and emboss. In this class of
effects, a copy of the artwork is converted to raster; the raster has the effects applied and is
then drawn first, the artwork is then drawn on top of the effect. (The exception is emboss,
in which artwork is dean first then the effect is drawn on top)
z
Inner (clipped) effects include inner shadow, inner glow, inner bevel, satin, and feathers. The
artwork fill and shape is combined with the effect parameters, and the result is a new fill for
the artwork. Because the fill is clipped to the artwork shape, this class of effect is highly
dependent on the type of artwork involved.
Common controlling parameters
Different transparency effects are controlled by different sets of parameters, Some of these
parameters are used commonly in defining different transparency effects.
Opacity
Opacity is the converse of transparency. It represented by the following formula:
opacity = (1.0 - transparency)
Opacity is expressed as a percentage, where 0% is completely clear and 100% is completely
opaque. Page items can have individual opacities assigned to them. Items can be grouped and
have a group opacity assigned, which is combined with the individual opacities to calculate the
resultant color on the backdrop.
Blending mode
Blending mode defines how the background (backdrop) and foreground (source) colors interact. In the physical world, transparency is the result of light passing through translucent materials. Blending modes are somewhat different: they do not try to model the complex physics of
transparency; rather, they are analytic expressions that produce interesting visual effects. The
following blending modes are encountered often: normal, multiply, and screen. There also are
blending modes related to color spaces, like hue, saturation, and color-blending modes.
Blending modes are vector functions that take as input the object source color and backdrop
color and produce a resulting color that is the new color of the backdrop. This is not pixelbased blending; rather, this is in terms of points with no dimension. The colors that are the
input for a blending function need to be specified in a common blending-color space (document CMYK or RGB). This is specified document-wide through the Edit menu.
A blending mode is described by an analytic expression that specifies how to calculate the
resultant backdrop color given the mode, current backdrop color, foreground color, and opacities of the background and foreground. Ordinarily, a blending function is a vector function:
C_r = F(Mode, C_b, C_f, O_b, O_f)
where C_r is the resultant backdrop color, Mode is the blending mode, C_b is the current backdrop color, O_b is the opacity of the backdrop, and O_f is the opacity of the foreground (source
object).
Graphics Fundamentals
265
Graphics Fundamentals
Transparency effects
In general, the blending function gives the resulting color at a point. The two-dimensional
coordinates in the backdrop plane also can be specified in the expression above, to give an
exhaustive expression for a blending function.
Normal blending mode is a trivial blending mode, which specifies that the resulting blend is the
value of the source color. The result of a normal blend is given by the following formula:
B(C_b, C_s) = C_s
where B is the blending function, C_b is the backdrop color, and C_s the source-object color.
Multiply-blending mode is the scalar product of the backdrop color and the source object color.
For an additive color space like RGB, the multiply-blending mode is calculated by taking the
component-wise product of the two vectors. The multiply blending mode is given by the following formula:
B(C_b,C_s) = C_b•C_s
With an additive color space, the resulting blend color is darker than either of the two inputs,
so the multiply-blending mode also is referred to as shadow mode.
For a subtractive color space like CMYK, it is necessary to take the reciprocal of the colors. The
expression for the blending mode is identical to the above, but with the substitution of (1-c) for
the color.
Screen-blending mode often is referred to as highlight mode. The analytic expression for screen
blending mode is given by the following formula:
B(C_b, C_s) = 1 - (1-C_s)•(1-C_b)
This can be thought of as the reciprocal of multiply mode, in the sense that it is multiplying the
reciprocal input colors and using the reciprocal of the result. The resulting components are
larger than the corresponding input components, so the net effect is a highlight of the two colors.
Position
Position determines the location of a transparency effect. It could be specified as an X offset-Y
offset pair or a distance-angle pair. They can be used to calculate each other. InDesign also have
a concept of global light, which is the angle of the light and is maintained the same for every
object of a document.
Size
Size refers to the size of a transparency effect. It is specified in number of points. For example, a
drop shadow with size of 0p5 mains the shadow width is 5 points.
Spread
Adding spread outsets the object's outline before blurring, so the effect seems bolder. Spread is
expressed as a percentage of the blur width. Typically, spread is for outer effects, Spread is the
percentage of the size or blur width that is simply outset rather than tapered off. For example,
for a drop shadow with a blur width of 10 points and a spread of 0%, you get 10 points of blur.
If the spread is 50%, however, you get 5 points of outset and 5 points of blur. If the spread is
100% you get 10 points of outset and no blur.
266
Graphics Fundamentals
Transparency effects
Noise
Random noise can be added to the transparency effect, to give it a rougher appearance.
Choke
Choke is conceptually the same as spread, but for effects that go “inward,” like inner glow. For
example, an inner glow with size 10 points and spread 50%, you get 5 points of inset and 5
points of blur.
Feather width
Feather width is conceptually the same as size, simply the total amount of inset applied. For
example, if you have 10 points of width, the feather extends from the edge of the object to 10
points inside the object. For directional feather, you can specify width in four directions.
Basic transparency
Basic transparency is an effect that lets the viewer see through an object. The Transparency
panel in the user interface allows end users to specify the transparency attributes of selected
page items, including blending mode, opacity, whether these should be isolated, and whether
knockout is required.
Transparency in group items
A group item is a collection of one or more page items. Transparency can be applied to a page
item or a group item. Transparency groups are constructs to work around real-world problems
involving grouped transparent objects. The application allows the assignment of opacity,
knockout, and isolation properties to transparency groups (i.e., collections of page items).
These group properties determine how a group of page items interact transparently with the
backdrop. It also is possible to specify knockout and isolation properties for one page item.
Group attributes combine with those of the objects in the group. For example, if objects have
50% opacity and the group has 50% opacity, the resulting opacity of the composited objects is
equivalent to an object with 25% opacity.
These group properties are independent of, and in addition to, the opacities and blending
modes of the individual objects. The objects interact with each other and the backdrop using
their individual opacity and blending modes. The group opacity and group-blending mode do
not affect the interaction of the individual objects with each other, but they do affect the interaction of these objects with the backdrop.
Knockout groups
A knockout group is a group whose individual elements do not interact with each other but do
interact with the backdrop. The objects within the group paint opaquely over each other, but
compositing occurs with the initial backdrop data. An example in which this is useful is a
group that consists of the fill and the stroke of a path. If the group is not a knockout group,
there is an interaction of the fill and stroke in the region they have in common, which often is
undesirable. Instead, the desired effect is to have the stroke opaquely cover the fill in the common overlap, and to have each interact with the backdrop with respect to their individual opacities and blending modes. Figure 102 shows a comparison of transparency effects with and
without knockout groups.
Graphics Fundamentals
267
Graphics Fundamentals
Transparency effects
FIGURE 102
Knockout groups
Transparency without
knockout group
Transparency with
knockout group
Isolated groups
An isolated group is a group whose individual elements do not blend with the backdrop independently but do blend with the backdrop as a group. An isolated group can be represented by
its group object of color and opacity values. By definition, an isolated group can be represented
by color and opacity values that are independent of the backdrop into which it is to be composited.
Drop shadow
In the real world, drop shadow is generated when light is blocked by an object. As represented
in graphics, drop shadow is a blurred representation of the shape of an object, offset by a userspecified distance and drawn below the object.
The parameters that control a drop shadow include the ones listed in Table 59.
TABLE 59 Drop-shadow parameters
268
Parameter
Description
Blending mode
See “Blending mode” on page 265.
Noise
See “Noise” on page 267.
Object-knockout shadow
Object does not interact with shadow.
Opacity
See “Opacity” on page 265.
Position
See “Position” on page 266.
Shadow honors other
effects
Shadow can be the result together with other transparency
effects.
Graphics Fundamentals
Transparency effects
Parameter
Description
Size
See “Size” on page 266.
Spread
See “Spread” on page 266.
Inner shadow
Inner shadow is very similar to drop shadow. It adds a shadow that falls just inside the edges of
the object, giving the object a recessed appearance. The parameters that control an inner
shadow effect are like those of drop shadows (see Table 60).
TABLE 60 Inner-shadow parameters
Parameter
Description
Blending mode
See “Blending mode” on page 265.
Opacity
See “Opacity” on page 265.
Position
See “Position” on page 266.
Size
See “Size” on page 266.
Noise
See “Noise” on page 267.
Choke
See “Choke” on page 267.
Outer and inner glow
Glow is a graphic effect that makes an object appear to shine as if with intense heat. Outer glow
refers to the glow effect applied to the outside edges of the object; inner glow refers to the glow
effect applied to the inside edges. The parameters that control outer glow are listed in Table 61.
TABLE 61 Outer-glow parameters
Parameter
Description
Blending mode
See “Blending mode” on page 265.
Opacity
See “Opacity” on page 265.
Technique
The method to produce angled corners. You can get a
smooth/rounded effect (softer) or a mitered effect (precise).
Size
See “Size” on page 266.
Noise
See “Noise” on page 267.
Spread
See “Spread” on page 266.
The parameters that control inner glow are listed in Table 62.
Graphics Fundamentals
269
Graphics Fundamentals
Transparency effects
TABLE 62 Inner-glow controlling parameters
Parameter
Description
Blending mode
See “Blending mode” on page 265.
Choke
See “Choke” on page 267.
Noise
See “Noise” on page 267.
Opacity
See “Opacity” on page 265.
Size
See “Size” on page 266.
Source
Indicates the “polarity” of the glow. “Center” means the center of
the object glows and the edges do not. “Edge” means the opposite.
Technique
The method to produce angled corners. You can get a
smooth/rounded effect (softer) or a mitered effect (precise).
Bevel and emboss
The bevel and emboss effect is a kind of simulation of how an object would look if it were
“puffed up” into the third dimension. The parameters that control bevel and emboss are listed
in Table 63.
TABLE 63 Bevel and emboss parameters
270
Parameter
Description
Style
A choice of inner bevel, outer bevel, emboss and pillow emboss.
Direction
Indicates whether the object appears to be “sunken” or “raised.”
Technique
Simulate effect made with smooth, chisel hard and chisel soft.
Soften
The amount of blurring, more or less, that is applied to the
beveled result. This softens the effect so your bevel is not so harsh.
Size
Width of the effect. See “Size” on page 266.
Depth
Determines the intensity or sharpness of the bevel.
Angle
The direction in the page-plane of the light source.
Altitude
The angular distance above the page plane.
Highlightblending mode
See “Blending mode” on page 265.
Highlight opacity
See “Opacity” on page 265.
Shadow-blending
mode
See “Blending mode” on page 265.
Shadow opacity
See “Opacity” on page 265.
Graphics Fundamentals
Transparency effects
Satin
Satin applies interior shading that creates a satiny finish. Table 64 lists its controlling parameters.
TABLE 64 Satin parameters
Parameter
Description
Blending mode
See “Blending mode” on page 265.
Opacity
See “Opacity” on page 265.
Angle
The direction of the light to create effect.
Distance
The distance of the light to satin effect.
Size
See “Size” on page 266.
Feather effects
Feather is a graphic effect that allows designers to create a smooth edge around a frame. InDesign support three types of feathers: basic, directional, and gradient. The parameters that control a basic feather effect are listed in Table 65. Directional-feather effect feathers only on
selected sides of an object and takes additional parameters, listed in Table 66. Gradient effect
creates a linear or radial gradient of opacity around an object and allows parameters to define a
gradient, like gradient stops, gradient type and angle,; see Table 67.
TABLE 65 Basic feather parameters
Parameter
Description
Feather width
See “Feather width” on page 267.
Corners
Diffused, sharp, or rounded.
Choke
See “Choke” on page 267.
Noise
See “Noise” on page 267. Feather noise is added only to the
feather region.
TABLE 66 Directional-feather parameters
Parameter
Description
Angle
The direction of the feather.
Choke
See “Choke” on page 267
Feather widths
See “Feather width” on page 267.
Noise
See “Noise” on page 267.
Shape
First edge only, leading edges, or all edges.
Graphics Fundamentals
271
Graphics Fundamentals
Transparency effects
TABLE 67 Gradient-feather parameters
Parameter
Description
Gradient stops
Defines the gradient
Gradient type
Linear or radial gradient
Angle
The direction of the gradient
Flattening
PostScript and PDF 1.3 have no representation of transparency information. To print or export
spreads with transparent information, it is necessary to generate nontransparent representations of the effects of transparency; this is known as flattening. In some cases, the interactions
may be among areas of solid color, in which case they can be replaced by a solid color in each
region of interaction. If images are involved, a composite image must be calculated that reflects
the contributions of the objects being composited, including any solid-color areas.
Dealing with text and gradients requires the application of additional rules. If a section of the
page is particularly complex, it can be faster to rasterize the area than to try to calculate all the
interactions; the flattener user interface supports this feature.
The flattener is an Adobe core technology. Its implementation comes from the Adobe Graphics
Manager (AGM) library, and it is used in other Adobe applications that support transparency
within vector graphics, such as Illustrator.
The end user has a degree of control over how the flattener is used. The Print dialog box has an
option to parameterize the flattener. The user may choose flattener presets and specify, for
example, that high quality is preferred over high speed. There is a trade-off between precision
of results and resources used. The lower the resolution, the quicker and less memory-intensive
the calculation.
Flattening is a complex and potentially resource-expensive process. Various optimizations are
required to implement transparency effectively, such as caching high-resolution images to file,
to ensure that the memory requirements are minimized. There also are complexities introduced by OPI image substitution, because printing with transparency requires that embedded
files write graphic primitives into the output stream.
Transparency data model
Transparency features are complicated, but the principles of their implementation are not.
Basically, transparency implementation involves the following related aspects:
272
z
Graphic attributes used to specify states and parameters of transparency effects
z
Page-item adornments used to draw transparency effects
z
Service providers and drawing-event handler to participate in the printing process that flattens the transparency before printing
Graphics Fundamentals
Transparency effects
The boss classes responsible for transparency behavior are delivered mainly by the Transparency plug-in. The user interface to the effects is supplied by the TransparencyUI plug-in.
Transparency information is represented by graphic attributes, and transparency is rendered to
a device through page-item adornments.
Defining transparency: graphic-attribute bosses
Transparency effects are defined programmatically as a set of graphic attributes. These
attribute boss classes derive from kGraphicsAttrBoss and aggregate a collection of interfaces
that set and get a variety of graphic-attribute data types. For general information about graphic
attributes, see “Graphic attributes” on page 256. For a summary of the transparency=attribute
boss structure, see “Transparency-attribute boss structure” on page 273.
Transparency-attribute targets
Transparency effects can be applied to the object, stroke, fill, or content. They are called transparency targets. Accordingly, transparency attributes can be categorized into different targets.
Attribute targets are defined as enums in IXPAttributeSuite::AttributeTarget.
Transparency-attribute groups
Each transparency effect has a set of attributes to represent the parameters that control the
effect. Accordingly, transparency attributes can also be categorized into groups. Each group
corresponds to each transparency effect. Transparency-attribute groups are defined as enums
in IXPAttributeSuite::AttributeGroup.
Transparency-attribute data types and values
Different transparency attributes hold different types of data. The IXPAttributeSuite::AttributeDataType enum defines all possible data types a transparency attribute can have.
IXPAttributeSuite::AttributeValue is a class that can hold values for any of the attributes. It
serves as a generic container for get and set functions.
Transparency-attribute boss structure
Transparency-attribute bosses are organized as a three-level inheritance. The most generic level
is the kGraphicsAttrBoss. This implies transparency attributes work and can be manipulated
the same way as any other graphic attributes.
The next level is attribute groups. These attribute group bosses inherit from kGraphicsAttrBoss
and aggregate the IID_IOBJECTSTYLEATTRINFO interface with the kXPAttrInfoImpl implementation. These groups correspond to the transparency effects; see Table 68.
TABLE 68 Transparency-attribute groups
Group-attribute boss
Transparency effect
kDropShadowAttrBoss
Drop shadow
kVignetteAttrBoss
Basic feather
kXPAttrBoss
Basic transparency
Graphics Fundamentals
273
Graphics Fundamentals
Transparency effects
Group-attribute boss
Transparency effect
kXPBevelEmbossAttrBoss
Bevel and emboss
kXPDirectionalFeatherAttrBoss
Directional feather
kXPGradientFeatherAttrBoss
Gradient feather
kXPInnerGlowAttrBoss
Inner glow
kXPInnerShadowAttrBoss
Inner shadow
kXPOuterGlowAttrBoss
Outer glow
kXPSatinAttrBoss
Satin
The bottom level has the real transparency attributes that define every transparency. For example, kXPBasicOpacityAttrBoss, kXPStrokeBlendingOpacityAttrBoss, kXPFillBlendingOpacityAttrBoss, and kXPContentBlendingOpacityAttrBoss inherit from kXPAttrBoss and
aggregate the IID_IGRAPHICATTR_REALNUMBER interface. They define the opacity
attribute of basic transparency for object, stroke, fill, and content targets, respectively.
Transparency-attribute types
There are about 400 transparency attributes. You may still access them in the general graphicattribute way; however, the InDesign API also provides utility functions in IXPAttributeUtils to
manipulate transparency attributes more effectively. Part of that is the concept of transparencyattribute types.
Transparency-attribute types are defined as enums in IXPAttributeSuite::AttributeType. These
enums are divided into segments according to transparency targets and transparency groups.
The BASE_XP_ATTR(t,x) macro generates the base index of the segment. For example, if you
want to know the enum value of the first transparency attribute for controlling inner shadow
on the stroke of an object, use kStrokeInnerShadowBaseID + 1. kStrokeInnerShadowBaseID is
calculated as BASE_XP_ATTR(kTargetStroke, kGroupInnerShadow).
A transparency-attribute type has a one-to-one mapping to transparency-attribute boss. You
can translate between AttributeTypes and attribute-boss ClassIDs using the GetAttributeClassID and GetAttributeFromClassID methods on IXPAttributeUtils.
Drawing transparency: adornment bosses
To create a transparent rendering for an object, add an IAdornmentShape implementation to
the adornment boss class that provides the behavior for the object. The kBeforeShape and kAfterShape flags specify when to draw in response to the IAdornmentShape::GetDrawOrderBits
method. The IAdornmentShape::GetPaintedBBox method specifies the adornment bounding
box. Bound of the outer effects are larger than the object’s bounds.
Table 69 lists the adornment bosses responsible for drawing transparency effects. One adornment boss can draw multiple effects. The primary reason we do not have one adornment for
every effect is to limit the total number of adornments of a page item, for performance reasons.
274
Graphics Fundamentals
Transparency effects
TABLE 69 Boss classes that aggregate IAdornmentShape
Boss-class name
Description
kXPDropShadowAdornmentBoss
Draws knockout drop shadows and outer glow.
kXPPageitemAdornmentBoss
Draws blending and non-knockout drop
shadows.
kXPVignetteAdornmentBoss
Draws basic feather, directional feather, and
gradient feather.
kXPInnerEffectsAdornmentBoss
Draws inner effects, like inner shadow, inner
glow, bevel and emboss, and satin.
Printing transparency: flattening
Before transparency is printed, transparency effects must be flattened. You can choose different
flattener presets to represent transparency using non-transparency representation, then print
the result as normal page items.
Managing flattener presets
Flattener presets are represented by kXPFlattenerStyleBoss. The key interface is IFlattenerSettings, which defines settings of presets, directly corresponding to the Flattener Preset Options
dialog box.
The flattener preset is an application-wide preference. The IFlattenerStyleListMgr interface is
aggregated on kWorkspaceBoss and is responsible for managing all flattener presets. As with
other general presets, like print presets, the IFlattenerStyleListMgr implementation hooks flattener presets into the generic presets dialog and provides an entry point for adding, deleting,
and editing flattener presets with respective commands.
The kSpreadBoss also aggregates IFlattenerSettings, which allows the spread to override the
flattener settings for a specific spread.
Printing-related bosses
Transparency participates in the printing process by iterating spreads for pages containing
transparency effects that require flattening. With these bosses, the print command can examine
transparency use, setting up viewport attributes and so on. Together with drawing in the spread
and page item, InDesign can achieve results that appear as if there are transparency interactions, even though these are accomplished through flattened, opaque, drawing instructions.
See Table 70.
Graphics Fundamentals
275
Graphics Fundamentals
Data model for drawing
TABLE 70 Printing-related boss classes
Boss-class name
Description
kXPGatherProviderBoss
A document iterator provider that gathers
document-wide use of transparency.
kXPPrintSetupServiceBoss
Implements print services to set up global color
profiles.
kXPDrwEvtHandlerBoss
Transparency start-up and shut-down. Hooks
transparency into the draw manager.
Data model for drawing
Presentation views
Drawing documents to the screen really means drawing to an application-layout presentation,
which is a platform-independent construct containing layout widgets that display the contents
of the document. Layout presentation is represented by kLayoutPresentationBoss, with the key
interface IDocumentPresentation. Note that kLayoutPresentationBoss is inherited from kWindowBoss; however, it is very important not to use the IWindow interface for any layout-presentation operation. Instead, always use IDocumentPresentation for operations like minimizing
the view. The IWindow on a kLayoutPresentationBoss is only for internal use. There may be
several layout presentations open at the same time; these can be hosted in one operating-system window as tabbed document presentation views, or each layout presentation can be hosted
inside its own operating-system window.
In addition to the layout presentation, there are other windows/view containers in InDesign.
Dialog boxes and pop-up tips are application windows and are represented by the platformindependent kWindowBoss object. The panel container, often referred to as palette, is represented by a PaletteRef. For more detail on these interfaces and their responsibilities, see the API
reference documentation.
To summarize, the application manages two main categories of views:
z
Document presentations have a document associated with them and display the content of
the document. They also are referred to as layout presentations. A document presentation
may not always correspond to a standalone operating-system window, because an operating-system window can have many document presentations in it. Note, however, that “window” is still being referred to throughout this document, because that is still a commonly
accepted term for the view to a document.
z
User-interface windows/palettes are used for dialog boxes, panels, and other types of windows that do not have documents associated with them.
Most of the detail in the remainder of this section concerns drawing a layout.
276
Graphics Fundamentals
Data model for drawing
Graphics context
Drawing requires a graphics context. When page items are called to draw, they are provided
with a GraphicsData object, which contains a pointer to an InDesign graphics context. Using
the application’s graphics context, page items can perform platform-independent drawing to
several of device contexts: screen, print, or PDF.
The graphics context for application drawing is described by an object derived from IGraphicsContext, an abstract-data container interface. This interface is not aggregated on boss classes,
nor is it derived from IPMUnknown. Instead, implementations of IGraphicsContext are specialized for different drawing devices.
A graphics context is instantiated based on a viewport boss object (see “Viewport” on
page 277), IControlView interface pointer, and update region. Several important pieces of
information are stored in the graphics context object, including the following:
z
The current rectangular clip region for the drawing port.
z
The transform for mapping content to drawing device coordinates.
The transform describes the relationship of the content coordinates to the drawing-device
coordinates. For screen drawing, the drawing device is a window. The transform is based on
the IControlView implementation supplied at the time of the graphics-context instantiation.
During a normal screen-drawing sequence, this is the IControlView implementation on the
kLayoutWidgetBoss that instantiates the graphics context, so the transform is initialized to be
pasteboard to the application-window coordinates. During the drawing sequence, this transform is modified to represent parent-window coordinates. For a description of coordinate systems related to drawing, see the “Layout Fundamentals” chapter.
Screen draws do not paint the entire document and then copy only a portion of it to the window. Instead, only the region of the window marked invalid is redrawn. The clip region stored
in the graphics context is applied like a mask, discarding any drawing outside the clip bounds.
During a drawing sequence, the clip represents the update region in the window and is stored
in window coordinates.
For drawing to the screen, AGM is used to provide the graphics context. An offscreen context
is used for drawing, to facilitate a smooth, flicker-free screen update. The drawing code renders
off-screen, and the result of the completed drawing is subsequently copied to the screen. For
more details, see “Offscreen drawing” on page 284. The use of off-screen contexts is transparent to page items when they draw. The graphics context provided to the page item hides these
implementation details.
Viewport
The application’s platform-independent API for drawing is provided by the viewport. Page
items use the IGraphicsPort interface on the viewport boss object for drawing. The viewport
boss object does not draw directly to the platform window; instead, the viewport boss object
draws through the AGM, which is specialized for the type of drawing device (for example,
PostScript printing or the screen). The specialization is transparent to clients of the graphics
context and the viewport boss object.
Graphics Fundamentals
277
Graphics Fundamentals
Dynamics of drawing
A viewport boss object operates as a wrapper between the drawing client (often a page item)
and the underlying AGM code that defines the graphics context for the output device. There
are several different types of viewport boss classes, each tailored to a particular type of drawing
output. For a list of relevant boss classes and the responsibilities of the interfaces on the boss
classes, see the API reference documentation for IViewPort. For example, kWindowViewPortBoss is used for drawing to application windows.
Dynamics of drawing
Drawing the layout
Layout drawing is the process of drawing page items to the layout presentation. Layout is
described in depth in the “Layout Fundamentals” chapter. For more detail on how individual
page items are drawn, see “Drawing page items” on page 281.
Invalidating a view
InDesign drawing occurs in response to invalidation of a view, or region. Although InDesign
provides platform-independent APIs for invalidating a view, the APIs simply forward the invalidation to the operating system. The invalidation state is maintained by the operating system,
which generates update events to the application.
Invalidation can be caused by changes to the model (persistent data in a document) or direct
invalidation. Both use the same mechanism to invalidate a view.
Changes to the model are broadcast to interested observers with regular or lazy notification.
For example, the application uses an observer on the layout widget to monitor changes to page
items in the layout. When the observer receives notification of a change, it asks the subject of
the change to invalidate a region based on its bounding box.
Views can be directly invalidated by using the ILayoutUtils::InvalidateViews method. The collaboration of boss objects is shown in Figure 103.
278
Graphics Fundamentals
Dynamics of drawing
FIGURE 103
Boss collaboration when invalidating a view
ILayoutUIUtils
*
InvalidateViews
«boss»
kLayoutWidgetBoss
*
*
ILayoutController::InvalidateContent
«boss»
kWindowViewportBoss
*
*
IWindowPort::InvalRgn
Platform invalidation
*
The ILayoutUtils::InvalidateViews method uses a document pointer to locate the views for a
document. For each view, the method gets the ILayoutController interface on the kLayoutWidgetBoss and calls ILayoutController::InvalidateContent, which creates an invalidation region
equal to the view’s window. The invalidation region is passed to the IWindowPort::InvalRgn
method on the IWindowPort interface of the kWindowViewPortBoss corresponding to the
window. This method then passes the invalidation region to the operating system.
The invalidation region is then accumulated by the operating system, which generates a paint
message. The operating system sends an update message (event) to the InDesign event dispatcher, where the message is wrapped in a platform-independent class and routed to the event
handler on the appropriate window.
The window’s event handler calls indirectly the IControlView::Draw method. At that point, the
drawing sequence begins following the pattern for drawing a widget hierarchy; Draw uses the
IPanelControlData interface to locate each of its children and call Draw on the child’s IControlView interface. This means the kLayoutPanelBoss is called to draw, and it iterates each of its
child widgets (scroll bars, the layout widget, rulers, etc.).
The drawing sequence concludes with validating the contents of the window. This completes
the update cycle, from invalidation to redraw.
Layout drawing order
Layout presentations are updated in several steps that allow better performance and the opportunity to interrupt the drawing. Good performance is achieved by specializing the drawing
Graphics Fundamentals
279
Graphics Fundamentals
Dynamics of drawing
operations according to the view’s z-order, whether the invalidation is due to a selection change
or because all objects changed.
At the level of the kLayoutWidgetBoss, several decisions are made behind the scenes for the
sake of drawing performance. The IControlView implementation on the kLayoutWidgetBoss
chooses what to draw based on the window’s z-order, the selection, and the update region.
Foreground and background drawing
If the window is the front view, two views may be drawn: a background view containing all
objects and a foreground view containing only the decorations for the selection. Offscreen
graphics contexts are created and cached for both views (see “Offscreen drawing” on page 284).
If the update region overlaps the window or the current front view was not the last view to have
been drawn, the background view is redrawn. If the selection is not empty, the foreground view
also is redrawn.
Based on the above rules, the kLayoutWidgetBoss IControlView filters the visible spreads and
calls their IShape::Draw methods through the InDesign draw manager. This filtering means
page items are not guaranteed to be called for a screen draw. The background draw then propagates through the layout hierarchy, as described in “Spread-drawing sequence” on page 307.
After the background draw is complete, if the selection is empty, the background image is copied to the window and the draw is complete. If the selection is not empty, the foreground draw
starts by copying the background image to the foreground offscreen context. The selection is
drawn to the foreground, then the foreground image is copied to the window.
The use of foreground and background draws means a page item could be called twice during a
redraw: the first call to IShape::Draw in the order of the layout hierarchy for the background
draw, and the second call to IHandleShape::Draw in the order of the selection list for the foreground draw.
If the window is not the front view, a more abbreviated drawing sequence takes place. Both the
document items and the selection are drawn to a foreground offscreen context, and it is copied
to the screen.
Z-order
Z-order is the order (depth) of page item objects in the direction perpendicular to the screen.
Z-order determines which object or part of an object is visible to the user. InDesign handles
three levels of z-order:
1. Window z-order — InDesign windows may overlap.
2. Layer order — InDesign documents are layered. InDesign draws page layers first, then other
document layers. For more information, see the “Layers” section of the “Layout Fundamentals” chapter.
3. Page item z-order — On the same layer, page items are ordered according to their positions
in the parent object’s hierarchy. The child with index 0 is drawn first; the child with the
greatest index is drawn last. For more information, see the “Documents and the Layout
Hierarchy” section of the “Layout Fundamentals” chapter.
280
Graphics Fundamentals
Dynamics of drawing
NOTE:
Do not confuse z-order with foreground and background drawing. Z-order controls
what is visible to the viewer, whereas foreground and background are just bitmaps for
speeding up rendering.
Draw manager
Conceptually, kLayoutWidgetBoss calls each spread to draw. In reality, kLayoutWidgetBoss
calls through the InDesign draw manager. The draw manager is an interface that exists on each
viewport and is used to set clipping areas and initiate the drawing of any page element and its
children to that viewport. The draw manager can be used for drawing to the screen, PDF, and
print. Drawing in InDesign occurs on a spread-by-spread basis and is hierarchical. Without a
service like the draw manager, it would be hard to change the behavior of drawing in a subhierarchy. Optionally, you can create a kDrawMgrBoss and use the IterateDrawOrder method,
which allows clients to iterate the document in the same manner as drawing, calling a clientprovided callback routine for every page item in the hierarchy, but without a graphics context.
Funneling all drawing activity through the draw manager provides three important features for
changing the behavior of drawing operations:
z
Clipping of areas to be drawn.
z
Filtering of items to be drawn.
z
Interrupting a drawing sequence.
Clipping and filtering provide a means to select items for the draw based on regions. The draw
manager maintains the current values for the clip and filter regions. Interruption of the drawing sequence relies on the broader mechanism of drawing events. As items draw, they broadcast standard messages that announce the beginning and end of discrete drawing operations.
Drawing events are received by special handlers that register interest in particular types of messages, and each drawing-event handler has the opportunity to abort the draw at that point. Clients of the draw manager (like the layout widget) install a default drawing-event handler.
The draw manager provides services for hit testing, iterating the drawing order, and drawing.
Detail about how spreads are drawn is described in “Spread-drawing sequence” on page 307.
Drawing page items
When page items are called to draw, they are passed information about their drawing context,
including a GraphicsData object and IShape drawing flags (e.g., kPatientUser). For more information, see the API reference documentation for these types. Page items use this information
to perform their drawing operations.
The IGraphicsPort interface is the InDesign interface for drawing and is aggregated on all
viewport boss classes. The IGraphicsPort interface is used by all drawing interfaces on a page
item, regardless of the output device.
The viewport settings can be modified using IGraphicsPort methods. The IGraphicsPort::gsave
and IGraphicsPort::grestore methods effectively push the current port settings onto a stack and
pop the old settings when desired. These methods are used to save the current port settings
when setting the port’s transform to a new space. For example, when a page item is called to
Graphics Fundamentals
281
Graphics Fundamentals
Dynamics of drawing
draw, it should save the port settings, set the port to its inner coordinate space, draw, and
restore the port before returning.
The IGraphicsPort implementations provide methods very similar to PostScript graphics operators. These methods support drawing primitives, port-transformation manipulation, and
changing port-clip settings. For a complete list of the methods, see the API reference documentation for IGraphicsPort.h. Line-drawing methods like IGraphicsPort::lineto, IGraphicsPort::moveto, IGraphicsPort::curveto, IGraphicsPort::closepath, IGraphicsPort::fill, and
IGraphicsPort::stroke are available. If a path is defined in the port, it can be discarded using the
IGraphicsPort::newpath method. These methods are demonstrated in the “Graphics” chapter
of Adobe InDesign CS4 Solutions. Control over the color space, color values, gradient, and
blending modes also is available through IGraphicsPort methods.
See Table 71 for a list of interfaces for drawing page items.
TABLE 71 Interfaces for basic page-item drawing
Interface
Function
IGraphicStyleDescriptor
Renders graphic attributes.
IHandleShape
Renders selection handles. Called during foreground
drawing.
IPageItemAdornmentList
Lists adornments for a page item.
IPathPageItem
Renders low-level draws for path, fill, and stroke.
IShape
Renders path, fill, and stroke. Called during
background drawing.
IShape, IPathPageItem, and IHandleShape are directly involved in drawing a page item. IShape
is the main interface for drawing the page item and is called when the entire page item needs to
be rendered. Path-based items like spline bosses also have the IPathPageItem interface to handle the details of drawing their paths. IHandleShape is called when a page item is in the selected
state and is used for drawing the decorations that denote selection. PathHandleShape is
responsible for drawing path-selection handles.
IGraphicStyleDescriptor maintains the graphic attributes for the page item. The drawing process sets graphics port settings based on various graphic attributes.
IPageItemAdornmentList maintains a list of adornments for the page item. Adornments are
drawn with the page item. This interface provides a means to enhance or decorate a page item’s
appearance. For more detail, see “Extension patterns” on page 289.
IShape is responsible for providing the drawing, hit testing, drawing-order iteration, and invalidation functions for page items. This interface is responsible for rendering a page item and is
used during background draws to create the path, fill, and stroke of an object. For details, see
its API reference documentation.
There are partial implementation classes like CShape and CGraphicFrameShape in the public
API. For details, see the API reference documentation.
282
Graphics Fundamentals
Dynamics of drawing
For details on the drawing sequence for a page item, see “Drawing sequence for a page item” on
page 310.
The IHandleShape interface
The IHandleShape interface is responsible for drawing and hit testing for a page item’s selection
handles. This interface is defined in the abstract base class in IHandleShape.h.
As with the IShape class, the IHandleShape drawing sequence includes extensibility points in
the form of adornment calls. For more detail, see “Extension patterns” on page 289. For more
detail on its responsibilities, see the API reference documentation on IHandleShape.
Most page items (such as kSplineItemBoss) that aggregate IHandleShape (with the ID
IID_IHandleShape interface) also have another implementation of IHandleShape which is
based on PathHandleShape (with the ID IID_IPATHHANDLESHAPE interface). The implementation based on PathHandleShape inherits from the same superclass as IHandleShape;
therefore, it has exactly the same public interface. The purpose of this interface is to draw pathselection handles. Observe the difference in the user interface by switching between the Selection and Direct Selection tools.
The IPathPageItem interface
The IPathPageItem interface encapsulates the specialized drawing functions for path-based
page items. This class provides methods for operations like stroke, fill, clip, and copying a path
to the graphics port. The CGraphicFrameShape class delegates these specific draw operations
to the IPathPageItem class.
In this class, graphic attributes are used to control the appearance. The IPathPageItem::Stroke
method first calls IPathPageItem::Setup, which initializes a set of default stroke properties and
then tries to get the stroke graphic attributes, like stroke weight and a ClassID for a path stroker. The path stroker is a service provider that delivers the kPathStrokerService. The
IPathPageitem::Stroke method then uses the stroke graphic attributes and path stroker to actually create the stroke.
Drawing in user-interface widget windows
A user-interface widget is any descendant of kBaseWidgetBoss. Drawing in a user-interface
widget window—such as a palette, dialog box, or panel—is like drawing to the layout presentation. Instead of calling the draw method of the layout-control view, palette drawing calls IControlView::Draw methods of various IControlView implementations.
The first stage in creating an owner-drawn panel is constructing an AGMGraphicsContext
object from the IControlView::Draw parameters and obtaining an IGraphicsPort from the
graphic context.
Next, set colors and draw the items. This can be done through the IInterfaceColors interface,
aggregated on kSessionBoss. There are a few standard interface colors common to drawing
items in user-interface windows. User-interface colors are defined in InterfaceColorDefines.h.
Next, set the color of the graphics port and draw items.
Graphics Fundamentals
283
Graphics Fundamentals
Dynamics of drawing
Each each widget actually is a window, so the PMRect value from IControlView::GetFrame
represents the coordinates of the widget in its parent widget. An IControlView::MoveTo(0,0)
call is convenient for transforming the coordinates.
You can draw lines, boxes, and so on as you normally draw in the layout presentation, using the
methods of IGraphicsPort. For drawing PMString, InDesign provides utility methods on the
StringUtils class (DrawStringUtils.h) that let you measure a PMString and draw strings directly.
For more details, see the API reference documentation.
For an example of an owner-drawn panel, with a custom IControlView::Draw implementation,
see the SDK sample PanelTreeView in <SDK>/source/sdksamples/paneltreeview, and inspect
PnlTrvCustomView.cpp.
Offscreen drawing
Purpose of offscreen drawing
An offscreen drawing is simply a bitmap. The technique of offscreen drawing has been used for
years by various applications to reduce screen flicker and improve performance when drawing
to the screen. Mac OS X actually creates a separate offscreen buffer for each window, such that
when an application tries to draw to a window, it actually is drawing to the offscreen buffer.
The operating system then periodically transfers its offscreen buffer to the actual screen. This
process is done at the operating-system level and decreases the number of cases in which InDesign needs to draw offscreen on Mac OS X.
Drawing layout offscreen
Offscreen drawing is used just about any time the layout presentation needs to update. InDesign maintains an offscreen representation of the layout presentation with everything except
the current selection highlighting. Hence, when the selection changes, InDesign can very
quickly redraw the screen from the offscreen buffer, followed by drawing the new selection.
The biggest area where offscreen drawing is used is the layout-control view. Drawing background and foreground in layout-control view is discussed in “Layout drawing order” on
page 279. What must be emphasized here is that drawing the layout does not really draw to the
screen directly; instead, drawing the layout actually draws offscreen.
Corresponding to background and foreground drawing, there are background offscreen buffers and foreground offscreen buffers. The background offscreen representation in InDesign is
the offscreen buffer maintained by each layout view, containing everything except current
selection highlighting. Each layout view has its own offscreen buffer. The background offscreen
buffer is invalidated by querying for the ILayoutController of the IControlView interface for
the layout and calling ILayoutController::InvalidateContent.
InDesign can be told to update a layout without dirtying the background offscreen buffer, by
querying for the ILayoutController of the IControlView interface for the layout and calling
ILayoutController::InvalidateSelection.
Palettes and dialog boxes are responsible for their own offscreen representations. In general,
most panels do not use offscreen drawing; however, a few user-interface windows, like the Navigator panel, maintain their own offscreen representations. As mentioned above, Mac OS X
284
Graphics Fundamentals
Client APIs
provides offscreen buffers for all windows, so offscreen drawing on Mac OS X is not necessary
if the goal is simply to avoid flicker.
Offscreen data
Drawing offscreen is drawing to the kOffscreenViewPortBoss. Besides other interfaces that are
involved in drawing, one of the most important interfaces on kOffscreenViewPortBoss is IOffscreenPortData. This is the interface that stores a reference to the offscreen bitmap and distinguishes kOffscreenViewPortBoss from other viewports.
An offscreen bitmap is wrapped in the platform-independent IPlatformOffscreen. You can get
and set bitmaps with IOffscreenPortData; however, you are not allowed to create bitmaps
directly. Instead, you may want to aggregate an IOffscreenViewPortCache interface on the control-view boss class, so you can query background and foreground viewport data.
For more information, see the API reference documentation for IPlatformOffscreen, IOffscreenPortData, and IOffscreenViewPortCache.
Snapshots
A snapshot is a record of a part of the document view at a specific instant. SnapshotUtilsEx
draws items into a memory-based graphics context, creating a bitmap (snapshot) that can be
exported to a stream in several formats, like JPEG, TIFF, or GIF.
SnapshotUtilsEx also creates its own offscreen bitmap and uses an IDrawMgr to walk the IHierarchy and draw specific page items to the offscreen buffer; however, the internal implementation is different. Like layout offscreen, the viewport SnapshotUtilsEx interacts with is
kAGMImageViewportBoss, the bitmap it draws to is of type AGMImageRecord defined in
GraphicsExternal.h, and it uses IAGMImageViewPort to manipulate the bitmap. SnapshotUtilsEx does not interact with the layout's offscreen representation in any way.
SnapshotUtilsEx is very useful for creating proxy images of documents, since it uses the same
draw manager to draw page items in memory, a much faster way. Sample code in the Snapshot
sample plug-in and in SnpCreateInddPreview.cpp demonstrates the use.
The sample code in the SDK still uses SnapshotUtils. We recommend using the more recent
SnapshotUtilsEx.
Client APIs
Path-related client APIs
High-level APIs provided to manipulate paths are summarized in Table 72. For details about
them, see the API reference documentation.
Graphics Fundamentals
285
Graphics Fundamentals
Client APIs
TABLE 72 Path and graphic page-item APIs
API
Description
IPathPointScriptUtils
Utility methods related to path point scripting (called by script
providers).
IPageItemTypeUtils
Utilities to get the type of a page item.
IPathUtils
Path manipulations.
ISplineUtils
Utilities for spline-item manipulations.
IPathInfoUtils
Analyzes whether a list of PathInfo object forms a point or
straight line.
IPathPointUtils
Utility methods related to path point transformations.
IPathOperationSuite
Manipulates paths on selected page items, including
compound paths. Includes path-finder operations.
IConvertShapeSuite
Converts selected page items to a new shape.
Graphic page-item client APIs
APIs provided to manipulate graphics are summarized in Table 73. For details about them, see
the API reference documentation.
TABLE 73 Graphic page-item APIs
286
API
Description
IFrameContentSuite
Selection suite interface for content fitting and framecontent conversion.
IFrameContentFacade
Facade for content fitting and frame-content conversion.
IFrameContentUtils
Utility interface for determining various aspects of frame
contents.
IClippingPathSuite
Gets and sets clipping path for selected graphics items.
ITextWrapFacade
Facade for manipulating text wrap.
IDisplayPerformanceSuite
Changes display performance settings for selection.
IImageObjectSuite
Accesses image layers.
IImageFacade
Facade for getting image information.
IJPEGExportSuite
Exports selection to JPEG file format.
ISVGExportSuite
Exports selection to SVG file format.
Graphics Fundamentals
Client APIs
Key color-related client APIs
Interfaces provided to manipulate color, swatch, ink, and color management objects are summarized in “Color-related APIs” on page 287. For details about them, see the API reference
documentation.
TABLE 74 Color-related APIs
API
Description
IGradientAttributeSuite
A selection suite interface. Provides gradient attribute–
related operations that apply to the application defaults,
document defaults, text, layout, and tables.
ISwatchUtils
Utility interface to manipulate swatches. For example, this
interface can be used to acquire the active swatch list,
retrieve or instantiate new, persistent, rendering boss
objects, and modify the properties of the swatch list.
IUIColorUtils
Utility interface for dealing with user-interface colors.
IInkMgrUtils
Utility interface for managing inks. For example, this
interface can be used to acquire the active ink list and
retrieve information about name, type, ink alias, and
corresponding swatches of a spot ink.
IInkMgrUIUtils
Utility interface, mainly for calling the ink-manager dialog
box.
IColorSystemUtils
Utility interface for color-rendering objects. For example,
this interface can be used to get color space and color
components of a given color swatch.
ICMSAttributeSuite
A selection suite interface. Provides color-management
functionality for the selected image page item, like getting
or applying color profile or rendering intent.
ICMSUtils
Utility interface for the color-management system. For
example, this interface can be used to obtain CMSSettings
and ProfileList. Provides wrappers for color-management
commands.
ICMSManager
A manager interface that provides a generic wrapper
around a color-management system (CMS). Access to the
CMS typically is done by coordinating between this
interface and a data interface for the specific CMS
implementation.
IColorPresetsManager
A manager interface that manages a set of color preset files.
For example, load or save a color preset, and get or set
working RGB/CMYK profile and CMS policy.
Graphics Fundamentals
287
Graphics Fundamentals
Client APIs
Since color-related objects like colors, swatches, inks, and color profiles are stored in the document or application database, you will need commands to manipulate them. The utility interfaces described above provide a higher-level abstraction that wrap the required commands. For
a list of commands, see the API reference documentation.
Graphic-attribute client APIs
APIs provided to manipulate graphic attributes are summarized in “Graphic-attribute APIs” on
page 288. For details about them, see the API reference documentation.
TABLE 75 Graphic-attribute APIs
288
API
Description
IGraphicAttributeUtils
Exposes methods that can be used to create lists of
attribute overrides and create commands to apply
attribute overrides. It also exposes methods to read the
graphic attributes of document objects.
IGraphicStateUtils
Exposes methods to create or process commands to apply
overrides, swap stroke and fill, and so on. It also exposes a
key method to acquire a reference to the active graphic
state, which is discussed in detail in the “Graphics”
chapter of Adobe InDesign CS4 Solutions.
IGraphicAttributeSuite
Selection suite that can be queried for through an abstract
selection (kAbstractSelectionBoss). It is aggregated on the
concrete selection boss classes connected with application
defaults, document defaults, layout, tables, and text.
IGraphicAttrProxySuite
Selection suite that switches graphic attributes between
layout objects and text.
IStrokeAttributeSuite
High-level attribute suite interface specifically to get and
set various stroke attributes. It uses
IGraphicAttributeSuite to get and set appropriate data.
IXPAttributeSuite
Accesses or changes transparency attributes on the
selected page items or group items. This header file also
defines all transparency-attribute-related enums such as
attribute types.
IXPAttributeUtils
Transparency-attribute-related utility functions
IXPAttributeFacade
High-level transparency-attribute facade. Defines
methods to get and set all transparency attributes.
IXPUtils
Transparency utility functions. Mainly for flattening.
IXPManager
Transparency helper functions.
Graphics Fundamentals
Extension patterns
There are command boss classes used to manipulate graphic attributes, graphic style, and
graphic state. The utility interfaces described in Table 75 provide a higher-level abstraction that
wraps the common commands. For a complete list of provided commands, see the API reference documentation.
Extension patterns
Custom graphic attributes
You can provide custom graphic attributes that extend InDesign function for end users. As
explained in “Representation of graphic attributes” on page 258, graphic attributes are applied
to page items by adding to a list of attribute boss objects associated with the page item.
To define a customized graphic attribute, follow these steps:
1. Create a new boss class for the attribute. The boss class should inherit from kGraphicsAttrBoss.
2. Provide an implementation of the IGraphicAttributeInfo interface.
3. If the graphic attribute value could be stored in an existing data interface, like IGraphicAttrBoolean or IGraphicAttrClassID, add it to the attribute class; otherwise, implement an
interface of your own.
4. Provide your own code to initialize and set your attribute data when the attribute is applied
to a page item. Code to apply the attribute to a page item also is required.
If you implemented a custom graphic attribute, you are likely to use a page-item adornment to
decorate the page item based on your graphic-attribute value. For an example, see the TransparencyEffect and TransparencyEffectUI sample SDK plug-ins.
Custom path-stroker effects
You can provide custom path strokers with your own stroke implementation through a kPathStrokerService. The service provider should provide an implementation of IPathStroker, with
IStrokeParameters used for storing parameters for the stroke. See “Path stroker” on page 262.
To implement a custom path stroker, follow these steps:
1. Create a new boss class for the service provider (for example k<XXX>PathStrokerBoss).
Add a service provider to the boss class (the application-provided kPathStrokerServiceProviderImpl can be re-used), and define your implementation of IPathStroker.
2. Create your implementation of IPathStroker. The interface basically provides methods on
how to draw the path on the specific graphic port and hit testing (stroke bounding box,
etc.).
Graphics Fundamentals
289
Graphics Fundamentals
Extension patterns
3. If the implementation of the stroker is complicated, you may aggregate additional interfaces
on the boss. InDesign CS4/InCopy CS4 provides a common stroke-parameter implementation of the IStrokeParameters interface. The interface stores information like where a path
segment starts and the length of the segment. For details, see the API reference documentation.
To change only the dash and gap of a stroke, you do not need a custom stroke implementation.
Instead, just apply the appropriate kDashedAttributeValuesBoss attribute.
Custom corner effects
You can provide a custom corner path with your own corner path implementation through a
kPathCornerService. The service provider should provide an implementation of IPathCorner.
See “Path corners” on page 263. To implement a custom corner effect, follow these steps:
1. Create a new boss class for the service provider (for example k<XXX>CornerBoss). Add a
service provider to the boss class (the application-provided kPathCornerServiceProviderImpl can be re-used), and define your own implementation of IPathCorner.
2. Create your implementation of IPathCorner. The interface basically provides drawing
methods from three points on the corner. For detailed definitions of the pointers, see the
API reference documentation of the interface. The three points are calculated by a private
method not included in the SDK from the stroke path and corner size. You need to provide
only the method for creating the corner effect using these points.
3. Determine the size of the corner, and apply kGraphicStyleCornerRadiusAttrBoss to the
path page item. The corner-size attribute is decoupled from the corner-path implementation, so you could set and change corner size independently and use this setting with other
corner implementations.
Custom path-end effects
You can provide a custom path-end effect with your own path-end implementation through a
kPathEndStrokerService. The service provider should provide an implementation of IPathEndStroker. See “Path-end strokers” on page 263. To implement a custom path-end effect, follow
these steps:
1. Create a new boss class for the service provider (for example k<XXX>ArrowHeadBoss).
Add a service provider to the boss class (the application-provided kPathEndStrokerServiceProviderImpl can be re-used), and define your own implementation of IPathEndStroker.
2. Create your implementation of IPathEndStroker. The interface basically provides drawing
and hit testing methods for a stroke end. A page item’s stroke bounding box includes both
the stroke and stroke-end bounding box returned from the method on this interface.
3. Set the implementation ID to the appropriate line-end attribute, and apply the attribute to
the page item.
290
Graphics Fundamentals
Extension patterns
Custom page-item adornments
Page-item adornments customize the appearance of a page item, giving a plug-in the chance to
add drawing when the page item is rendered. For example, if you want to add a drop shadow or
label for a specific page item, a page-item adornment is one way to add these extra drawings.
Examples of page-item adornments in InDesign are the transparency drop shadow and feather
effects, the graphics-frame outline path, and the empty-graphics frame indicator (the X inside
the frame).
You can participate in page-item drawing by creating a custom page-item adornment. For
example, you could provide an adornment to show in magenta type the width and height of
each selected page item, as shown in Figure 104.
FIGURE 104
Adding adornments to selected page items
Before adding adornments
After adding adornments
Architecture
Page-item adornments are represented by boss objects. There are two main types of page-item
adornments:
z
IAdornmentShape implementations, drawn during execution of IShape::Draw.
z
IAdornmentHandleShape implementations, drawn during execution of IHandleShape::Draw.
The Draw methods on their implementations determine the appearance of the adornment.
Page-item adornments are connected to page items through the IPageItemAdornmentList
interface aggregated on the kPageItemBoss boss class. Figure 105 illustrates the basic data
model for adornments. A page item aggregates the IPageItemAdornmentList interface, which
stores a list of adornments the page item has.
Graphics Fundamentals
291
Graphics Fundamentals
Extension patterns
FIGURE 105
Adornment data model
«boss»
kPageItemBoss
1
IPageItemAdornmentList
*
«boss»
kYourAdornmentBoss
IPageItemAdornmentList
The IPageItemAdornmentList interface manages a list of the adornments for a page item. Each
adornment in the list is drawn in an order determined by the value of its AdornmentDrawOrder. For page-item adornments, there are several opportunities within the drawing cycle to
draw.
Each adornment can be drawn at one or more phases within the drawing cycle if you AND
together the flags defined in the AdornmentDrawOrder enumeration (defined in IAdornmentShape.h and IAdornmentHandleShape). These values indicate when the adornment is to be
drawn. Upon calling the command kAddPageitemAdornmentCmdBoss, the adornment is
inserted into the list based on drawing order.
When adding an adornment to a page item, define the adornment as a new boss class that
aggregates the IAdornmentShape or IAdornmentHandleShape interface. In a more complex
case, you can use more than one adornment for a single page item.
IAdornmentShape
When implementing IAdornmentShape, you can choose to provide hit testing and invalidation
of the view, if needed. At a minimum, your implementation must draw the adornment.
When a page item is drawn, each attached adornment is called to draw through the IAdornmentShape interface. A page-item shape tells the drawing manager at which of the following
times it should be drawn:
z
After the frame shape (kAfterShape).
z
Before the text foreground (kBeforeTextForeground).
z
At one of the other events defined in the adornment objects.
The drawing order controls when the adornment is called to draw. Figure 106 shows the drawing sequences for one rectangle spline item. When the adornments’ drawing methods get called
depends on the drawing order. For detailed information on page item drawing, see “Drawing
page items” on page 281.
292
Graphics Fundamentals
Extension patterns
FIGURE 106
Drawing sequences for a rectangle page item
DrawMgr
Draw kBeforeShape adornments.
Set fill path by calling
IPathPageItem::CopyPath.
Fill shape by calling
IPathPageItem::Fill.
Draw kAfterFill adornments.
Set port’s clip to fill path.
Draw children by calling
IShape::Draw for each child.
Draw kBeforeStroke adornments.
Stroke by calling
IPathPageItem::Stroke.
Draw kAfterStroke adornments.
Draw kAfterShape adornments.
Continue other drawings.
IAdornmentHandleShape
Implementing IAdornmentHandleShape lets you decorate the selection handles associated
with a page item. This decoration can be drawn with IAdornmentHandleShape::kBeforeShape
or IAdornmentHandleShape::kAfterShape; that is, before or after IHandleShape calls
DrawHandlesImmediate.
Implementation hints
To implement a custom page-item adornment, you need to do at least the following:
1. Define an adornment boss class. Depending on the nature of your adornment, you need to
implement IAdornmentShape or IAdornmentHandleShape. This determines how and
when your adornment is drawn. In providing the implementation aggregated on the adornment boss class, you specify how the adornment is drawn and how the painted bounding
box is calculated.
2. Determine when you will connect your adornment to the page-item adornment list, most
likely by processing a low-level command.
For a sample showing how to implement a custom page-item adornment, see FrameLabel in
the SDK; also see other samples that provide an implementation of IAdornmentShape, such as
TransparencyEffect and CustomDataLink.
Graphics Fundamentals
293
Graphics Fundamentals
Extension patterns
Custom drawing-event handler
You can use a custom drawing-event handler to take control when a page item is being drawn
at some point in the drawing cycle, to modify how the item draws or cancel it being drawn.
Architecture
Drawing events are messages that are broadcast at specific points in the drawing order. Each
drawing event signals the beginning or end of a phase of drawing. Drawing events provide
extensibility points in that drawing operations can be modified or cancelled in response to the
event.
Drawing-event types are defined in DocumentContextID.h. Some of the generic drawing event
types associated with shape drawing are shown in Table 76.
TABLE 76 Generic drawing events
Drawing event
Use
kAbortCheckMessage
Opportunity to abort drawing of an entire item.
kBeginShapeMessage
Start of shape drawing.
kDrawShapeMessage
Shape drawing is about to begin.
kEndShapeMessage
End of shape drawing.
kFilterCheckMessage
Opportunity to filter drawing based on object type.
In addition, there are other, more specific event types that signal the beginning and end of
spread, layer, and page drawing. Drawing events are generated for each Draw method in the
hierarchy: kBegin<XXX>Message, kEnd<XXX>Message, and kDraw<XXX>Message.
kBegin<XXX>Message is generated just before a call to the next level of the hierarchy (after
any transformation is applied to the object). kEnd<XXX>Message is generated just after execution returns from a lower level of the hierarchy.
Registering and unregistering
There are two ways a plug-in can register or unregister for drawing events: as a service provider
or using direct registration through the IDrwEvtDispatcher interface.
294
z
Drawing-event handlers can be a type of service provider. If so, they are automatically registered at start-up. A boss can register as a service provider using the IK2ServiceProvider
interface and the kDrawEventService service ID. The boss must provide an implementation
of the IDrwEvtHandler interface. When the application starts, the Register method in the
drawing-event handler registers for the drawing events it wants to receive.
z
A second method of registering for events is to instantiate the IDrwEvtDispatcher interface
and call IDrwEvtDispatcher::RegisterHandler directly. This requires parameters including
the event being registered for, a pointer to the event handler, and the priority of the event
handler.
Graphics Fundamentals
Extension patterns
In the PrintSelection SDK sample, the kPrnSelDrawServicesBoss boss class does not provide an
implementation for IK2ServiceProvider. This is because this particular sample uses the direct
registration method; see the AfterPrintUI method in PrnSelPrintSetupProvider.cpp.
Unregistering is similar to registering, except the call to UnRegisterHandler is on the IDrwEvtDispatcher interface.
Drawing event-handling priorities
Drawing event handlers register their prioritized interest in particular types of drawing events.
Priorities are defined in IDrwEvtDispatcher.h. When a handler registers for an event, it must
pass a priority to RegisterHandler. The priorities include the following:
z
kDEHPostProcess
z
kDEHLowestPriority
z
kDEHLowPriority
z
kDEHMediumPriority
z
kDEHHighPriority
z
kDEHInitialization
As each event is processed, handlers are called in order, from the kDEHInitialization priority
down to kDEHPostProcess. If two handlers have the same priority for the same event, the handler registered first is called first. The return code for handlers registered using the kDEHInitialization priority is ignored. For additional information about return codes for the other
priorities, see “Handling drawing events” on page 295.
Handling drawing events
The drawing-event handler’s HandleEvent method takes two parameters: a ClassID that is the
eventID (see DocumentContextID.h) and a void pointer to a class containing the event data.
For drawing events, this pointer must be cast to the DrawEventData class (see IDrwEvtHandler.h) to access the data.
In the HandleEvent method of the drawing-event handler, if the method returns kTrue, it is
assumed the event was properly handled and no other event handlers are called for the event;
therefore, the drawing at that step ceases. If the method returns kFalse, the drawing event is
considered not handled, and the drawing step proceeds. A drawing-event handler that modifies or decorates the drawing of an object but does not replace it returns kFalse. Also,
kEnd<XXX>Message does not look at the return code from the event handler, because the
draw operation already completed.
Implementation
A custom drawing-event handler is an extension pattern to let third-party code participate in
drawing or printing or interrupt the drawing or printing of a page item. The following steps
summarize how to implement a custom drawing-event handler. The signature interface of the
pattern is IDrwEvtHandler.
Graphics Fundamentals
295
Graphics Fundamentals
Swatch-list state
1. Define your own drawing-event handler boss class.
2. Depending on how you want to register and unregister your event handler, you also may
need to implement the IK2ServiceProvider interface, returning a ServiceID of kDrawEventService.
3. Provide an implementation of IDrwEvtHandler.
4. In your IDrwEvtHandler::HandleEvent implementation, you will get an event ID and a void
pointer to a class containing the event data. You should cast to the DrawEventData and
obtain the GraphicsData pointer; then you can do your custom drawing in the same context
of page-item drawing.
For sample code, see BasicDrwEvtHandler and other samples that implement IDrwEvtHandler
in the SDK.
Swatch-list state
Initial state of swatch list and ink list
When the application starts, it initializes a swatch list by creating several default swatches (rendering objects). There are reserved swatches: None, Paper, Black, and Registration. Table 77
lists the initial state of swatch list, and Table 78 illustrates initial ink list. (UIDs are not necessarily the same for different workspaces.)
TABLE 77 Initial state of swatches
296
Index
UID
Rendering object
Swatch name
Color or gradient information
0
11
kPMColorBoss
Black (reserved)
kPMCsCalCMYK(0,0,0,1)
1
18
kPMColorBoss
C=0 M=0 Y=100 K=0
kPMCsCalCMYK(0,0,1,0)
2
19
kPMColorBoss
C=0 M=100 Y=0 K=0
kPMCsCalCMYK(0,1,0,0)
3
20
kPMColorBoss
C=100 M=0 Y=0 K=0
kPMCsCalCMYK(1,0,0,0)
4
21
kPMColorBoss
C=100 M=90 Y=10 K=0
kPMCsCalCMYK(1,0.9,0.1,0)
5
22
kPMColorBoss
C=15 M=100 Y=100 K=0
kPMCsCalCMYK(0.15,1,1,0)
6
23
kPMColorBoss
C=75 M=5 Y=100 K=0
kPMCsCalCMYK(0.75,0.05,1,0)
7
12
kPMColorBoss
Cyan
kPMCsCalCMYK(1,0,0,0)
8
13
kPMColorBoss
Magenta
kPMCsCalCMYK(0,1,0,0)
9
14
kGraphicStateNoneRenderingO
bjectBoss
None (reserved)
N/A
Graphics Fundamentals
Swatch-list state
Index
UID
Rendering object
Swatch name
Color or gradient information
10
15
kPMColorBoss
Paper (reserved)
kPMCsCalCMYK(0,0,0,0)
11
16
kPMColorBoss
Registration (reserved)
kPMCsCalCMYK(1,1,1,1)
12
17
kPMColorBoss
Yellow
kPMCsCalCMYK(0,0,1,0)
13
97
kPMColorBoss
(invisible)
kPMCsCalCMYK(0,0,0,1)
14
99
kPMColorBoss
(invisible)
kPMCsCalCMYK(0,0,0,0)
15
98
kGradientRenderingObjectBoss
(invisible)
Stop 0 UID=99, Stop 1 UID=11
16
100
kAGMBlackBoxRenderingObje
ctBoss
(invisible)
N/A
Notice there are several invisible, unnamed swatches that are created but do not show up in the
Swatches panel.
TABLE 78 Initial ink list
Index
UID
Name
Type
0
7
Process Cyan
Process
1
8
Process Magenta
Process
2
9
Process Yellow
Process
3
10
Process Black
Process
State of swatch list and ink list after adding a custom stop color
Adding a custom color swatch, PANTONE 368 C, causes an additional swatch-list entry to be
created. For information on how to add a color swatch, see the “Graphics” chapter of Adobe
InDesign CS4 Solutions. Table 79 illustrates the new state of the swatch list, and Table 80 illustrates the new state of the ink list. Note the following:
z
The swatch is sorted by swatch name. Since the new swatch’s name is between index 9 and
10, the new color is inserted into position 10. The following row is added: 10, 167, kPMColorBoss, PANTONE 368 C, kPMCsCalCMYK (0.57,0,1,0), Spot.
z
The swatch list is reordered, so entries 10 and after are shifted down. The UIDs of the existing swatches do not change.
z
A new spot ink is added to the ink list.
Graphics Fundamentals
297
Graphics Fundamentals
Swatch-list state
TABLE 79 Swatches after adding a custom color
Index
UID
Rendering object
Swatch name
Color or gradient information
0
11
kPMColorBoss
Black (reserved)
kPMCsCalCMYK(0,0,0,1)
1
18
kPMColorBoss
C=0 M=0 Y=100 K=0
kPMCsCalCMYK(0,0,1,0)
2
19
kPMColorBoss
C=0 M=100 Y=0 K=0
kPMCsCalCMYK(0,1,0,0)
3
20
kPMColorBoss
C=100 M=0 Y=0 K=0
kPMCsCalCMYK(1,0,0,0)
4
21
kPMColorBoss
C=100 M=90 Y=10 K=0
kPMCsCalCMYK(1,0.9,0.1,0)
5
22
kPMColorBoss
C=15 M=100 Y=100 K=0
kPMCsCalCMYK(0.15,1,1,0)
6
23
kPMColorBoss
C=75 M=5 Y=100 K=0
kPMCsCalCMYK(0.75,0.05,1,0)
7
12
kPMColorBoss
Cyan
kPMCsCalCMYK(1,0,0,0)
8
13
kPMColorBoss
Magenta
kPMCsCalCMYK(0,1,0,0)
9
14
kGraphicStateNoneRendering
ObjectBoss
None (reserved)
N/A
10
167
kPMColorBoss
PANTONE 368 C
kPMCsCalCMYK(0.57,0,1,0), Spot
11
15
kPMColorBoss
Paper (reserved)
kPMCsCalCMYK(0,0,0,0)
12
16
kPMColorBoss
Registration (reserved)
kPMCsCalCMYK(1,1,1,1)
13
17
kPMColorBoss
Yellow
kPMCsCalCMYK(0,0,1,0)
14
97
kPMColorBoss
(invisible)
kPMCsCalCMYK(0,0,0,1)
15
99
kPMColorBoss
(invisible)
kPMCsCalCMYK(0,0,0,0)
16
98
kGradientRenderingObjectBo
ss
(invisible)
Stop 0 UID=99, Stop 1UID=11
17
100
kAGMBlackBoxRenderingObj
ectBoss
(invisible)
N/A
TABLE 80 Ink list after a stop color is added to swatch
298
Index
UID
Name
Type
0
7
Process Cyan
Process
1
8
Process Magenta
Process
2
9
Process Yellow
Process
3
10
Process Black
Process
4
166
PANTONE 368 C
Spot
Graphics Fundamentals
Swatch-list state
Swatch list and ink list after adding a gradient swatch
The user also may add a custom gradient. Adding a gradient may add additional colors for a
gradient stop. Similar to what occurs when adding a custom color swatch, new entries are created for these added colors and gradients. For information on adding a gradient swatch, see the
“Graphics” chapter of Adobe InDesign CS4 Solutions. Table 81 lists only the new entries for the
swatch list, and Table 82 illustrates the updated ink list.
TABLE 81 New entries in swatch list after adding a custom gradient
Index
UID
Rendering object
Swatch name
Color or gradient information
13
169
kPMColorBoss
Stop 1
kPMCsCalCMYK(0.2,1,0.5,0)
14
171
kPMColorBoss
Stop 2
kPMCsCalRGB(0,1,0.5)
15
173
kPMColorBoss
Stop 3
kPMCsCalCMYK(0,0.3,0.9,0)
16
174
kGradientRenderingObjectBoss
Tie-dye
Stop 0 UID=169, Stop 1UID=171, Stop 2
UID =173
The new gradient, Tie-dye, has three stops, so these three color swatches (Stop 1, Stop 2, and
Stop 3) also are inserted into the swatch list.
For the existing swatches we did not list here, the swatch data did not change, but their indices
were changed to make room for new swatches. The new swatches are inserted continuously,
because these new swatch names happened to be in continuous position with existing names.
TABLE 82 Ink list after a gradient with three-stop colors is added toswatch
Index
UID
Name
Type
0
7
Process Cyan
Process
1
8
Process Magenta
Process
2
9
Process Yellow
Process
3
10
Process Black
Process
4
166
PANTONE 368 C
Spot
5
168
Stop 1
Spot
6
170
Stop 2
Spot
7
172
Stop 3
Spot
Since the gradient stop colors are spot colors, they also are added to the ink list.
Graphics Fundamentals
299
Graphics Fundamentals
Swatch-list state
Swatch list and ink list after applying an unnamed color to an object
When an instance of a color is chosen with the Color Picker panel and applied to a page item, a
new, unnamed swatch entry is added to the end of the swatch list. See Table 83.
TABLE 83 New swatch entry added after picking a color from color picker
Index
UID
Rendering object
Swatch name
Color or gradient information
22
177
kPMColorBoss
(invisible)
kPMCsCalCMYK(0.4715,0,0.75,0)
Unnamed swatches are not sorted, so a new unnamed swatch is appended at the end. After the
previous step (adding a gradient), there were 22 swatches total in the list, so this step adds the
swatch at index 22.
Unnamed swatches do not appear in the Swatches panel, but they can be used for the strokes
and fills of document objects by client code. When a color is chosen through the Color panel,
the active graphic state changes, and a new swatch may be created. A new swatch is created in
the swatch list only if the new color is applied to a document object; i.e., if there is an active
selection when the color is created or the color is applied to an object like a spline page item.
Because the added unnamed color is a process color, the inks this color uses to print (i.e., Process Cyan, Process Magenta, Process Yellow, and Process Black) already are in the ink list, so
the ink list does not change. Applying a new gradient to a page-item object through the Gradient panel results in a similar state.
Color spaces
There are three common color spaces: RGB, CMYK, and L*a*b*.
RGB color spaces
RGB is a device-dependent color model. It is the native color model of monitors, scanners and
digital cameras.
The kPMCsCalRGB space denotes calibrated RGB. Although the space is device-dependent,
coordinates within this space are not arbitrary. There are several, slightly different color spaces
that use the RGB color model:
300
z
Adobe RGB (1998) — Previously referred to as SMPTE-240M.
z
Apple® RGB — The default color space for Photoshop 3 and 4.
z
ColorMatch RGB — Based on the Radius PressView display. This has a smaller gamut than
Adobe RGB (1998) and other color spaces for print production jobs.
z
sRGB (IEC61966-2.1) — The default color space for Photoshop 5. This reflects the color
properties of the average computer monitor. It was proposed in 1996 by Hewlett-Packard
and Microsoft® for representing color on the Internet, as described at
http://www.w3.org/Graphics/Color/sRGB.html. sRGB also is the color space used to represent color in SVG documents. For more information about SVG, go to
http://www.w3.org/TR/SVG.
Graphics Fundamentals
Catalog of graphic attributes
CMYK
Like RGB, CMYK is a device-dependent color model. It is the native color model of most printers. Many color spaces use the CMYK color model. For example, the InDesign color-management user interface lets the end user specify any of the following:
z
Euroscale Coated v2
z
Euroscale Uncoated v2
z
Japan Standard v2
z
U.S. Sheetfed Coated v2
z
U.S. Sheetfed Uncoated v2
z
U.S. Web Coated (SWOP) v2
z
U.S. Web Uncoated v2
The choice of color space is governed by expectations about the properties of the press on
which the job will be printed.
LAB
L*a*b* (1976 CIE L*a*b* Space) is device-independent and is the basic color model in PostScript. L*a*b* is used for color management as the device-independent model of the ICC
device profiles.
L*a*b* was standardized in 1976 to provide a space that is perceptually uniform. L*a*b* has its
basis in the CIE standard observer, derived by analysis of the physiology of the retina and the
early visual pathways.
The terms L*, a*, and b* refer to coordinate axes within the space. L is a function of luminance,
the physical correlate of brightness. The other coordinates are less easily understood in physical
terms and better regarded as mathematical abstractions.
Catalog of graphic attributes
There are many graphic attributes in the application. Sometimes several attributes collaborate
to implement a feature, like transparency; sometimes the values of some attributes are important, like color and stroke-line implementations.
Figure 107 is a master list of graphic attributes. The attribute boss class, name, and interface
that store the attribute value are listed. Other information is provided in the boolean matrix on
the right of the table.
The list can be obtained from the API reference documentation in two ways:
z
Look for IGraphicAttributeInfo and see which boss classes aggregate this interface.
z
Look at the kGraphicsAttrBoss boss class and examine the list of subclasses of this class.
Note this list is not a complete list of graphic attributes and is not updated for each release.
Graphics Fundamentals
301
Graphics Fundamentals
Catalog of graphic attributes
302
attributeName
value interfaces
affectsPageItemGeometry?
isRequiredGraphicAttribute?
isTableAttribute?
isTextAttribute?
isObservedByGraphicState?
isObservedByTransparencyAttrSuite?
isFormFieldAttribute?
Master list of graphic attributes
className
FIGURE 107
kCheckDefaultCheckedAttrBoss
kCheckExportValueAttrBoss
kChoiceAllowMultiSelAttrBoss
kChoiceEditableAttrBoss
kChoiceListAttrBoss
kChoiceSortAttrBoss
kDashedAttributeValuesBoss
kFormDefaultValueAttrBoss
kFormDescriptionAttrBoss
kFormExportAttrBoss
kFormExportMappingAttrBoss
kFormExportRequiredAttrBoss
kFormFontColorAttrBoss
kFormFontOverprintAttrBoss
kFormFontSizeAttrBoss
kFormFontStrokeColorAttrBoss
kFormFontStrokeOverprintAttrBoss
kFormFontStrokeTintAttrBoss
kFormFontStrokeWeightAttrBoss
kFormFontStyleAttrBoss
kFormFontTintAttrBoss
kFormFontUIDAttrBoss
kFormNameAttrBoss
kFormPrintVisibleAttrBoss
kFormReadOnlyAttrBoss
kFormScreenVisibleAttrBoss
kFormSpellCheckAttrBoss
kFormStyleAttrBoss
kFormTypeAttrBoss
kFormValueAttrBoss
kGraphicStyleCornerImplAttrBoss
kGraphicStyleCornerRadiusAttrBoss
kGraphicStyleEvenOddAttrBoss
kGraphicStyleFillRenderingAttrBoss
Default Is Checked
Export Value
Allow Multiple Selection
Editable
Choice List
Sort Items
Dash settings
Form Field Default Value
Form Field Description
Form Field Export
Form Field Export Mapping Name
Form Field Required For Export
Form Field Font Color
Form Field Font Color Overprint
Form Field Font Size
Form Field Font Stroke Color
Form Field Font Stroke Color Overprint
Form Field Font Stroke Tint
Form Field Font Stroke Weight
Form Field Font Style
Form Field Font Tint
Form Field Font Family
Form Field Name
Form Field Visible When Printed
Form Field Read Only
Form Field Visible On Screen
Spell Check
Form Field Style
Form Field Type
Form Field Value
Effect
Size
Even-Odd
Color
IGraphicAttrBoolean
IStringAttr
IGraphicAttrBoolean
IGraphicAttrBoolean
IChoiceListAttr (private)
IGraphicAttrBoolean
IDashedAttributeValues
IStringAttr
IStringAttr
IGraphicAttrBoolean
IStringAttr
IGraphicAttrBoolean
ITextAttrUID
IGraphicAttrBoolean
IGraphicAttrRealNumber
ITextAttrUID
IGraphicAttrBoolean
IGraphicAttrRealNumber
IGraphicAttrRealNumber
ITextAttrFont
IGraphicAttrRealNumber
ITextAttrUID
IStringAttr
IGraphicAttrBoolean
IGraphicAttrBoolean
IGraphicAttrBoolean
IGraphicAttrBoolean
IStringAttr
IGraphicAttrInt32
IStringAttr
IGraphicAttrClassID
IGraphicAttrRealNumber
IGraphicAttrBoolean
IPersistUIDData
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
no
no
no
no
yes
yes
no
yes
yes
no
no
no
no
no
no
no
no
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
no
no
no
Graphics Fundamentals
Graphics Fundamentals
affectsPageItemGeometry?
isRequiredGraphicAttribute?
isTableAttribute?
isTextAttribute?
isObservedByGraphicState?
isObservedByTransparencyAttrSuite?
isFormFieldAttribute?
kGraphicStyleFillTintAttrBoss
Tint
kGraphicStyleGapRenderingAttrBoss
Gap color
kGraphicStyleGapTintAttrBoss
Gap tint
kGraphicStyleGradientFillAngleAttrBoss
Gradient fill angle
kGraphicStyleGradientFillGradCenterAttrBoss Gradient fill center
kGraphicStyleGradientFillHiliteAngleAttrBoss
Gradient fill hilight angle
kGraphicStyleGradientFillHiliteLengthAttrBoss Gradient fill hilight length
kGraphicStyleGradientFillLengthAttrBoss
Gradient fill length
kGraphicStyleGradientFillRadiusAttrBoss
Gradient fill radius
kGraphicStyleGradientStrokeAngleAttrBoss
Gradient angle
kGraphicStyleGradientStrokeGradCenterAttrBossGradient stroke center
kGraphicStyleGradientStrokeHiliteAngleAttrBoss Gradient stroke hilight angle
kGraphicStyleGradientStrokeHiliteLengthAttrBossGradient stroke hilight length
kGraphicStyleGradientStrokeLengthAttrBoss
Gradient stroke length
kGraphicStyleGradientStrokeRadiusAttrBoss
Gradient stroke radius
kGraphicStyleJoinTypeAttrBoss
Join
kGraphicStyleLineCapAttrBoss
End cap
kGraphicStyleLineEndEndAttrBoss
Line end
kGraphicStyleLineEndStartAttrBoss
Line start
kGraphicStyleMiterLimitAttrBoss
Miter limit
kGraphicStyleNonPrintAttrBoss
Non print
kGraphicStyleOverprintFillAttrBoss
Overprint fill
kGraphicStyleOverprintGapAttrBoss
Overprint gap
kGraphicStyleOverprintStrokeAttrBoss
Overprint stroke
kGraphicStyleStrokeAlignmentAttrBoss
Stroke alignment
kGraphicStyleStrokeLineImplAttrBoss
Stroke type
kGraphicStyleStrokeRenderingAttrBoss
Color
kGraphicStyleStrokeTintAttrBoss
Tint
kGraphicStyleStrokeWeightAttrBoss
Stroke weight
kStrokeParametersBoss
Stroke parameters
kTextAlignmentAttrBoss
TextAlignment (form field)
kTextHasMaxLengthAttrBoss
Has Maximum Field Length
kTextMaxLengthAttrBoss
Maximum Field Length
kTextMultilineAttrBoss
Multiline
IGraphicAttrRealNumber
no
IPersistUIDData
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrPoint
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrPoint
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
no
IGraphicAttrInt32
yes
IGraphicAttrInt32
yes
IGraphicAttrClassID
yes
IGraphicAttrClassID
yes
IGraphicAttrRealNumber
yes
IGraphicAttrBoolean
no
IGraphicAttrBoolean
no
IGraphicAttrBoolean
no
IGraphicAttrBoolean
no
IGraphicAttrInt32
yes
IPersistUIDData IGraphicAttrClassID yes
IPersistUIDData
no
IGraphicAttrRealNumber
no
IGraphicAttrRealNumber
yes
IStrokeParameters
no
IGraphicAttrInt32
no
IGraphicAttrBoolean
no
IGraphicAttrInt32
no
IGraphicAttrBoolean
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
no
yes
no
yes
yes
no
yes
no
no
no
no
no
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
no
yes
no
yes
yes
yes
yes
no
no
no
no
no
yes
no
no
yes
yes
no
no
yes
no
yes
yes
no
no
yes
no
no
no
no
no
no
yes
yes
no
yes
no
no
yes
yes
yes
no
no
no
no
no
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
value interfaces
attributeName
className
Catalog of graphic attributes
303
Graphics Fundamentals
className
attributeName
value interfaces
affectsPageItemGeometry?
isRequiredGraphicAttribute?
isTableAttribute?
isTextAttribute?
isObservedByGraphicState?
isObservedByTransparencyAttrSuite?
isFormFieldAttribute?
Catalog of graphic attributes
kTextPasswordAttrBoss
kTextScrollAttrBoss
kTextUseForFileSelAttrBoss
kXPBasicBlendModeAttrBoss
kXPBasicIsolationGroupAttrBoss
kXPBasicKnockoutGroupAttrBoss
kXPBasicOpacityAttrBoss
kXPDropShadowBlendModeAttrBoss
kXPDropShadowBlurRadiusAttrBoss
kXPDropShadowColorAttrBoss
kXPDropShadowModeAttrBoss
kXPDropShadowNoiseAttrBoss
kXPDropShadowOffsetXAttrBoss
kXPDropShadowOffsetYAttrBoss
kXPDropShadowOpacityAttrBoss
kXPDropShadowSpreadAttrBoss
kXPVignetteCornersAttrBoss
kXPVignetteInnerOpacityAttrBoss
kXPVignetteModeAttrBoss
kXPVignetteNoiseAttrBoss
kXPVignetteOuterOpacityAttrBoss
kXPVignetteWidthAttrBoss
Password
Scroll
Used For File Selection
Mode
Isolate blending
Knockout group
Opacity
Mode
Blur radius
Color
Drop shadow
Noise
X offset
Y offset
Opacity
Spread
Corners
Inner opacity
Feather
Noise
Outer opacity
Feather width
IGraphicAttrBoolean
IGraphicAttrBoolean
IGraphicAttrBoolean
IGraphicAttrInt32
IGraphicAttrBoolean
IGraphicAttrBoolean
IGraphicAttrRealNumber
IGraphicAttrInt32
IGraphicAttrRealNumber
IPersistUIDData
IGraphicAttrInt32
IGraphicAttrRealNumber
IGraphicAttrRealNumber
IGraphicAttrRealNumber
IGraphicAttrRealNumber
IGraphicAttrRealNumber
IGraphicAttrInt32
IGraphicAttrRealNumber
IGraphicAttrInt32
IGraphicAttrRealNumber
IGraphicAttrRealNumber
IGraphicAttrRealNumber
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
no
no
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
bo
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
Some boss classes inherit from kGraphicsAttrBoss and implement IGraphicAttributeInfo. They
are the base classes for graphic attributes, but they do not represent standalone attributes.
These bosses are excluded from the master attributes list and are listed in Table 84 for reference.
304
Graphics Fundamentals
Catalog of graphic attributes
TABLE 84 Parent boss class of attributes
Boss ID
Description
kVignetteAttrBoss
Parent attribute boss for all basic feather
attributes.
kDropShadowAttrBoss
Parent attribute boss for all drop-shadow
attributes.
kXPAttrBoss
Parent attribute boss for all basic-transparency
attributes.
kXPInnerShadowAttrBoss
Parent attribute boss for all inner-shadow
attributes.
kXPOuterGlowAttrBoss
Parent attribute boss for all outer-glow attributes.
kXPInnerGlowAttrBoss
Parent attribute boss for all inner-glow attributes.
kXPBevelEmbossAttrBoss
Parent attribute boss for all bevel and emboss
attributes.
kXPSatinAttrBoss
Parent attribute boss for all satin attributes.
kXPDirectionalFeatherAttrBoss
Parent attribute boss for all directional-feather
attributes.
kXPGradientFeatherAttrBoss
Parent attribute boss for all gradient-feather
attributes.
kDefaultGraphicsAttrBoss
Parent of kGraphicStyleNonPrintAttrBoss.
kStrokeEffectGraphicsAttrBoss
Parent of all stroke-effect attributes (line cap, join
type, etc.).
kCornerEffectGraphicsAttrBoss
Parent of all corner-effect attributes (corner
implementation, size).
kFillGraphicsAttrBoss
Parent of all fill attributes (gradient fill, gradient
fill angle, fill color, overprint, etc.).
kStrokeGraphicsAttrBoss
Parent of all stroke attributes (gradient stroke,
gradient-stroke angle, stroke color, stroke weight,
etc.).
kGradientStrokeGraphicsAttrBoss
Gradient stroke itself is parent of all gradientstroke attributes (length, gradient center, radius,
etc.).
kGradientFillGraphicsAttrBoss
Gradient fill stroke itself is parent of all gradient
fill attributes (length, gradient center, radius,
etc.)
Graphics Fundamentals
305
Graphics Fundamentals
Mappings between attribute domains
Mappings between attribute domains
TABLE 85 Graphic-attribute-to-text-attribute mapping
Graphic-attribute class
Text-attribute class
kGraphicStyleFillRenderingAttrBoss
kTextAttrColorBoss
kGraphicStyleFillTintAttrBoss
kTextAttrStrokeTintBoss
kGraphicStyleGradientFillAngleAttrBoss
kTextAttrGradLengthBoss
kGraphicStyleGradientFillGradCenterAttrBoss
kTextAttrGradCenterBoss
kGraphicStyleGradientStrokeAngleAttrBoss
kTextAttrStrokeGradAngleBoss
kGraphicStyleGradientStrokeGradCenterAttrBoss
kTextAttrStrokeGradCenterBoss
kGraphicStyleGradientStrokeLengthAttrBoss
kTextAttrStrokeGradLengthBoss
kGraphicStyleOverprintFillAttrBoss
kTextAttrOverprintBoss
kGraphicStyleOverprintStrokeAttrBoss
kTextAttrStrokeOverprintBoss
kGraphicStyleStrokeRenderingAttrBoss
kTextAttrStrokeColorBoss
kGraphicStyleStrokeWeightAttrBoss
kTextAttrOutlineBoss
TABLE 86 Graphic-attribute-to-table-attribute mapping
306
Graphic-attribute class
Table-attribute class
kGraphicStyleFillRenderingAttrBoss
kCellAttrFillColorBoss
kGraphicStyleFillTintAttrBoss
kCellAttrFillTintBoss
kGraphicStyleOverprintFillAttrBoss
kCellAttrFillOverprintBoss
kGraphicStyleOverprintStrokeAttrBoss
kCellStrokeAttrDataBoss
kGraphicStyleStrokeLineImplAttrBoss
kCellStrokeAttrDataBoss
kGraphicStyleStrokeRenderingAttrBoss
kCellStrokeAttrDataBoss
kGraphicStyleStrokeWeightAttrBoss
kCellStrokeAttrDataBoss
Graphics Fundamentals
Spread-drawing sequence
Spread-drawing sequence
Once the kLayoutWidgetBoss IControlView implementation obtains a list of the visible
spreads, it has to ask each spread to draw through the InDesign draw manager. This section
describes the overall sequence for drawing the layout hierarchy.
The collaboration for drawing the layout hierarchy is shown in Figure 108. The IControlView
implementation on the kLayoutWidgetBoss calls the draw manager once for each visible
spread in the window’s view. The draw manager is responsible for creating a GraphicsData
object that describes the graphics context for the spread’s drawing. Specifically, it sets the transformation matrix in the graphics port to support drawing in the spread’s parent coordinate system, which is the pasteboard. The draw manager then calls the spread’s IShape::Draw method
to initiate the sequence of drawing that spread.
Graphics Fundamentals
307
Graphics Fundamentals
Spread-drawing sequence
FIGURE 108
Boss collaboration for drawing the layout
kLayoutWidgetBoss
1. : DrawWithTimedUpdate()
IControlView
ILayoutControlData
kViewPortBoss
IDrawMgr
2 : Draw()
kSpreadBoss
IGeometry
IHierarchy
IShape
2.1 : Draw()
kSpreadLayerBoss
IHierarchy
IShape
kPageBoss
2.1.1 : Draw()
IHierarchy
IShape
kSpreadLayerBoss
2.2 : Draw()
IHierarchy
IShape
kSplineItemBoss
2.2.1 : Draw()
IHierarchy
IShape
kSpreadLayerBoss
2.3 : Draw()
IHierarchy
IShape
kSpreadLayerBoss
2.4 : Draw()
IHierarchy
IShape
kGuideItemBoss
2.4.1 : Draw()
IHierarchy
IShape
When an item’s IShape::Draw method is called, it must draw itself, then its children. This
action is analogous to the mechanism used for the IControlView interface on widgets. The
IShape implementation provides methods for drawing an object and iterating over the object’s
children, asking each of them to draw. Before drawing, the IShape implementation sets the
transformation matrix in the graphics port to the item’s inner coordinates. This establishes the
following conventions:
308
Graphics Fundamentals
Controlling the settings in a graphics port
z
Each drawing object expects to receive the graphics port set to its parent’s coordinate system.
z
Each drawing object sets the graphics port to its inner coordinates before drawing itself or
calling its children to draw.
z
Each drawing object reverts the graphics port to its parent’s coordinate system before
returning to the caller.
Iterating over the object’s children is accomplished by using the IHierarchy interface on each
page item. This interface provides methods for iterating over the layout hierarchy. For more
information about the IHierarchy interface, see the “Layout Fundamentals” chapter.
After being called by the draw manager, spread drawing always follows the layout hierarchy.
Users can control whether the guides appear in front of the content or behind it, and this affects
the hierarchy. Assuming guides are behind the content, page layers are drawn first, followed by
the content layers, then the guide layers. If the guide layers appear in front of the content, the
guides draw before the content. All content on a given layer is drawn before proceeding to the
next layer. Embedded children of a page item are called to draw before that item completes its
drawing. For example, when the kSplineItemBoss is called to draw in step 2.2.1 of Figure 108, if
it had children, they would be called to draw as step 2.2.1.1 and 2.2.1.2. before returning to the
kSpreadLayerBoss.
Controlling the settings in a graphics port
The port’s transform can be modified by translations, scale factors, and rotations, and a new
transform can be concatenated to the existing port. Concatenation is most often used when
page items draw; it is how a page item sets the port to draw to its inner coordinates. Example 22
shows sample code demonstrating this operation. For more information on the application
coordinate systems, see the “Layout Fundamentals” chapter.
EXAMPLE 22 Manipulating the graphics-port settings
// Get the graphics port from gd, a GraphicsData*.
IGraphicsPort* gPort = gd->GetGraphicsPort();
if (gPort == nil) return;
// Save the current port settings.
gPort->gsave();
// Get this page item's ITransform interface.
InterfacePtr<ITransform> xform(this, IID_ITRANSFORM);
// Concatenate the inner to parent matrix to the port.
gPort->concat(xform->CurrentMatrix());
Graphics Fundamentals
309
Graphics Fundamentals
Drawing sequence for a page item
// Draw a rectangle around the item.
InterfacePtr<IGeometry> geo (this, IID_IGEOMETRY);
const PMRect r = geo->GetStrokeBoundingBox();
gPort->rectpath(r);
gPort->stroke();
// Restore the previous port settings.
gPort->grestore();
Drawing sequence for a page item
The drawing sequence for a page item that implements IShape involves more than just the
drawing instructions. Extensibility points are built into the sequence, in the form of drawing
events and adornments. Extensibility points are described in “Extension patterns” on page 289.
Although each IShape implementation may have its own specific drawing sequence, all implementations follow these steps as guidelines:
1. When a page item begins drawing, three drawing events are broadcast (kAbortCheckMessage, kFilterCheckMessage, and kDrawShapeMessage). No IShape drawing activities occur
between the broadcasts.
2. Save the graphics port state by calling IGraphicsPort::gsave. The graphics port is set to draw
in the page item’s inner coordinate system.
3. A kBeginShapeMessage drawing event is broadcast. If any drawing-event handler returns
kTrue, the drawing activity ends.
4. Draw kBeforeShape page item adornments.
5. Draw the page item’s own shape by calling the protected method CShape::DrawShape. This
method varies depending on the nature of the page item. The following is a list of common
tasks the DrawShape method may accomplish:
z
Define the item’s path.
z
Fill the path.
z
Set the port to clip to the path.
z
Draw the item’s children.
z
Stroke the path.
6. Draw kAfterShape page-item adornments.
There are types of page-item adornments other than kBeforeShape and kAfterShape. These
may be called to draw at other points in the sequence. The use of graphics-port save and restore
operations allows the adornments and child drawing routines to modify the port as needed and
return it to a known state for the next drawing step.
310
Graphics Fundamentals
Drawing sequence for a page item
For details of the code responsible for drawing, see the CShape.cpp and CGraphicFrameShape.cpp source code in the SDK, under <SDK>/source/public/pageitems/basicinterfaces. For
examples of how to create your own shapes, also see the BasicShape and CandleChart sample
plug-ins.
Graphics Fundamentals
311
Graphics Fundamentals
Drawing sequence for a page item
312
Text Fundamentals
Concepts
Text Fundamentals
This chapter provides background on the fundamental concepts used in the text architecture
and describes the major subsystems that implement text features. It has the following objectives:
z
Identify the core subsystems that implement the text architecture.
z
Describe the model for text content—how raw text and formatting information is managed.
z
Describe how the presentation of text (i.e., the actual rendered text) is modeled within the
application, including both the glyphs that represent the text and the objects that manage
the placement of glyphs on a spread.
z
Describe text composition, the process of taking the content model and incomplete presentation model to generate the final appearance of text in the presentation model.
z
Describe fonts and how they relate to the InDesign text subsystem.
Table 87 lists resources for more information on relevant topics.
TABLE 87 For more information
For ...
Go to ...
More information on
digital typography topics
http://store.adobe.com/type/topics/main.html
A glossary of typographic
terms
http://store.adobe.com/type/topics/glossary.html
An overview of Adobe
type technology
http://partners.adobe.com/asn/developer/type/main.html
The PostScript language
FAQ list
http://www.postscript.org/FAQs/language/FAQ.html
Concepts
This section introduces the subsystems that implement the text architecture. The core text
architecture can be divided into the following core subsystems: text content, text presentation,
and text composition. Subsystems that extend the core include import and export, text styles,
editors, and text search/replace. See Figure 109.
Text Fundamentals
313
Text Fundamentals
Concepts
FIGURE 109
Overview of core text architecture
Visual container
Visual container
Displays / is displayed by
Text layout
«uses»
Visual container (parcel and text frame)
Wax line
is created by
has
Text
composition
The wax
«uses»
Content (story thread, character and text attributes)
Wax strand
Represents composed text
has
Content
The text
model
Content
The text-content subsystem manages the content of a story. The text-content subsystem stores
the characters and attributes that control the styled appearance of text.
The text-presentation subsystem deals with where the text glyphs appear on the page. Text is
composed into frames that display a story. A frame is a visual container for text. Visually, a
story is displayed through a linked set of one or more frames. The frames have properties—
such as the number of columns, column width, and text inset—that control where text flows.
Text-wrap settings on overlapping frames may affect this flow. Once text is composed, it has a
visual appearance and exhibits many of the same geometry traits as other page items.
The text-composition subsystem manages the process that flows text into a set of specified containers. Fonts provide the text-composition subsystem with the glyph dimensions it needs to
arrange glyphs into lines of a given width.
314
Text Fundamentals
Text content
Text content
This section describes how the raw character and formatting information for a story is maintained within the application. Text content represents the information required to render a set
of characters. This information includes the Unicode character values, along with any formatting information that defines the look of text, contents of footnotes and tables, inline graphics,
and styles that can be applied to text.
Stories
Text content within a document is represented by a story (signature interface ITextModel). A
single, related set of text is maintained within a single story. The document can contain multiple stories. All stories within the document can be accessed through the IStoryList interface on
the kDocBoss class. Figure 110 shows this relationship.
FIGURE 110
Stories: maintained on the kdocboss, accessed by means of IStoryList.
kDocBoss
IStoryList
1
IStoryList::GetNthUserAccessibleStoryUID
1..*
kTextStoryBoss
ITextModel
A document can contain private, feature-dependent stories, which are defined not to be useraccessible. Other features (like find/replace and spelling) ignore these stories. Programmatically, they can be accessed through the IStoryList interface, as with any other story.
Generally, stories are not directly created or deleted. The life-cycle of a story is controlled as a
side effect of another operation. For example, creating a text frame using the Type tool causes a
text story to be created, whereas linking two text frames together causes a story to be deleted. It
is possible to control the lifetime of a story directly, though the story is still subject to the side
effects of manipulating an associated text frame.
A story has a length, which can be accessed using ITextModel::TotalLength. A story has a minimum length of one character: a terminating kTextChar_CR, which is inserted into the story
when the story is created. This character should not be deleted.
Text Fundamentals
315
Text Fundamentals
Text content
The textual content of a story is maintained independently from the visual containers in which
it is contained; i.e., its layout on the page. The text content from a story can be contained within
one frame, flow between columns within a frame, or even flow between multiple frames. For
example, Figure 111 shows three stories, each flowing through different sets of containers. The
model used in maintaining the text content for each is the same. Determining where a particular glyph is rendered is the responsibility of the composer.
FIGURE 111
Three text stories flowing through text frames in a spread
Information related to the text in a story is maintained by a set of strands. Strands represent
parallel sets of information, each of which holds a component of the story. The strands intertwine to give a complete description of the story. Strands are identified using the IStrand signature interface. Some common strands are shown in Figure 112.
316
Text Fundamentals
Text content
FIGURE 112
The text story aggregates the strands representing the content.
kTextStoryBoss
ITextModel
1
1
1
1
ITextModel::QueryStrand(kOwnedItemStrandBoss,...)
ITextModel::QueryStrand(kTextDataStrandBoss,...)
ITextModel::QueryStrand(kCharAttrStrandBoss,...)
ITextModel::QueryStrand(kParaAttrStrandBoss,...)
1
kTextDataStrandBoss
IStrand
Holds Unicode character
code data for the text in the
story.
1
1
1
kParaAttrStrandBoss
IStrand
Maintains formatting
information applicable to
paragraphs.
kCharAttrStrandBoss
IStrand
Maintains formatting
information applicable to a
range of characters.
kOwnedItemStrandBoss
IStrand
Maintains a link to objects
anchored into the text,
such as inline images, table
frames, and footnotes.
Strands
Strands (signature interface IStrand) model the linear behavior of text. Each strand represents a
different aspect of textual information. A story can be thought of as the composition of the set
of strands that contain information for a particular aspect of the story. Figure 112 shows a story
with associated text, paragraph-attribute, and character-attribute strands. It also shows the
owned item strand used to maintain objects within the text, like inline graphics.
Each strand has the same virtual length, equal to the value returned by ITextModel::TotalLength. Each position within a strand refers to the same logical position within the other
strands and the story as a whole. For example, position 10 on the kTextDataStrandBoss relates
to a particular character—the character at position 10 within the kCharAttrStrandBoss refers
to formatting information applicable to this character. Any modification of the text length is
reflected in each strand.
Figure 113 shows some text and how it might be represented on two strands. While the actual
storage mechanism for the strands is of little interest, both strands have the same virtual length.
A story (kTextStoryBoss) comprises a set of strands, each with the same length. The figure
shows only the data (kTextDataStrandBoss) and paragraph-attribute (kParaAttrStrandBoss)
strands.
Text Fundamentals
317
Text Fundamentals
Text content
FIGURE 113
A story and two of its associated strands
:kTextStoryBoss
ITextModel
:kTextDataStrandBoss
ITextModel::TotalLength() == 10
:kParaAttrStrandBoss
Text storage subsystem
I ndex
Dat a
0
s
1
o
2 3
m e
4
5
t
6
e
7
x
8
t
9
#
I ndex
Dat a
0 1 2 3 4 5 6 7 8 9
Format t i ng def i ned by root st yl e
Data strand
Paragraph attribute strand
Strands abstract over the raw
data. Developers should not need
to interact directly with the
storage model below strands.
Formatting is defined in “Text formatting” on page 324.
Runs
Strands are further divided into runs. A run represents something about the content of a particular strand. For the kTextDataStrandBoss, a run is a manageable-sized chunk of text data. For
the kParaAttrStrandBoss, a run exists for each paragraph in the story. For the kCharAttrStrandBoss, a run represents a sequence of characters that share the same formatting information. The semantics of a run is defined by its strand.
Owned items
Owned items (signature interface IOwnedItem) exist to allow some object to be associated with
a particular TextIndex position in the text strand. Generally, features anchor the owned item
into the text by placing a special character into the kTextDataStrandBoss; however, there is no
requirement for associating an owned item with a special character in the data strand.
Figure 114 shows an instance diagram associated with a story that contains an inline image.
318
Text Fundamentals
Text content
The story has the character “a” followed by an inline image, then the character “b.” The owned
item strand (kOwnedItemStrandBoss) maintains the UID of the inline object (kInlineBoss).
The actual image (kImageItem—note the nonstandard name) is associated through the hierarchy (IHierarchy).
FIGURE 114
Instance diagram showing an owned item
:kTextStoryBoss
IItemStrand::GetOwnedUID
for TextIndex 0, 2, and 3
would return kInvalidUID.
:kTextDataStrandBoss
:kOwnedItemStrandBoss
:kInlineBoss
IItemStrand
1
IItemStrand::GetOwnedUID(1)
IOwnedItem
1
+IHierarchy::QueryChild
Text storage subsystem
I ndex 0
1
Data a kTextChar_I nline
Data Strand
IHierarchy
1
2 3
b #
I ndex 0
1
2 3
Data
UI D for kI nlineBoss
#
Owned item strand
+IHierarchy::QueryParent
0..*
:kSplineItemBoss
IHierarchy
Inline graphic
represented by
kImageItem
1
+IHiearchy::QueryChild
+IHierarchy::QueryParent
0..*
:kImageItem
IHierarchy
Anchored-item positioning
The positioning of an inline object relative to its anchor position in the text is controlled
through the IAnchoredObjectData interface. The default behavior is defined by the IAnchoredObjectData on the session and document-workspace boss classes. Object styles also can
Text Fundamentals
319
Text Fundamentals
Text content
define how an inline object is positioned (IAnchoredObjectData on kObjectStyleBoss). Specific inline objects can be modified (IAnchoredObjectData on kInlineBoss).
The placement of inline objects can be specified as any of the following:
z
Within the text flow (with variable offset on the y axis).
z
Above the line, aligned to the center, right, or left of the text or spline. A variable amount of
space before and after the inline object can be specified.
z
A custom position on the page, relative to the anchor point, text frame, page, or spline.
As the content that contains the anchor is manipulated (by either adding text that causes the
anchor to flow out of the initial frame or manipulating the frames in which the text is contained), the position of the inline object is updated automatically.
Story threads
Strands model the linear nature of text; however, not all text within a story is linear. For example, an embedded table can have cells with textual content that flows independently from the
main text in the story. Story threads (signature interface ITextStoryThread) represent these distinct flows of text.
Each story has at least one story thread, the primary story thread, which represents the main
text of the story. Other text elements (like footnotes) contained within a story are anchored off
the primary story thread using owned items, as described in “Owned items” on page 318.
The interface that models story threads (ITextStoryThread) is maintained on the boss class
related to the text it represents. For the primary story thread, the interface is found on the
kTextStoryBoss. For footnotes, the interface is on kFootnoteReferenceBoss. The story thread
maintains text indices that identify the range of text within the story that relates to the particular strands.
Figure 115 shows a story with only text within the primary story thread (i.e., the story uses no
features requiring other story threads). The story instance (kTextStoryBoss) has four associated
strands. There is one story thread (ITextStoryThread) in this story.
320
Text Fundamentals
Text content
FIGURE 115
Story with only primary story thread
:kTextStoryBoss
ITextStoryThread::GetTextStart() == 0
ITextStoryThread::GetTextEnd() == 9
ITextStoryThread
ITextModel
ITextModel::TotalLength() == 10
:kTextDataStrandBoss
:kParaAttrStrandBoss
:kCharAttrStrandBoss
IStrand
IStrand
IStrand
:kOwnedItemStrandBoss
IStrand
Text storage subsystem
I ndex
Dat a
0 1 2 3 4 5 6 7 8 9
Format t i ng def i ned by root st yl e
I ndex
Dat a
0
Paragraph attribute strand
I ndex
Dat a
0
s
1
o
2 3
m e
4
Data strand
5
t
6
e
7
x
8
t
1 2 3 4 5 6 7 8
No f or mat t i ng def i ned
9
Character attribute strand
9
#
I ndex
Dat a
0 1 2 3 4 5 6 7 8 9
No i nl i ne obj ect s i n st r eam
Owned item strand
The primary story thread always begins at TextIndex 0. Other features, like footnotes and
tables, use story threads to maintain the feature text as a distinct entity. Figure 116 shows an
instance diagram of a story with a footnote. In the figure, a text-story thread (ITextStoryThread) maintains a relationship with some part of the text in a story. In this example, the primary
story thread (on kTextStoryBoss) ranges between TextIndex 0 and 3, and the footnote (kFootnoteReferenceBoss) ranges between TextIndex 4 and 8.
Text Fundamentals
321
Text Fundamentals
Text content
FIGURE 116
A story containing a footnote
The footnote number and tab
are inserted into the footnote
automatically. The data strand
indicates where the number
should be placed using
kTextChar_FootnoteMarker.
:kTextStoryBoss
ITextStoryThread
ITextStoryThread::GetTextStart() returns 0
ITextStoryThread::GetTextEnd() returns 3
ITextStoryThread::GetTextSpan() returns 4
:kTextDataStrandBoss
:kOwnedItemStrandBoss
IItemStrand::GetOwnedUID(1)
IItemStrand
Text storage subsystem
:kFootnoteReferenceBoss
I ndex 0 1 2 3 4 5 6 7 8
Data a 0x4 b # 0x4 0x9 y z #
ITextStoryThread
Data strand
I ndex
Data
0
1 2 3 4 5 6 7 8
UI D
Owned item strand
ITextStoryThread::GetTextStart() returns 4
ITextStoryThread::GetTextEnd() returns 8
ITextStoryThread::GetTextSpan() returns 5
0x4 == kTextChar_FootnoteMarker
0x09 == kTextChar_Tab
See TextChar.h.
Story-thread dictionaries and hierarchies
Each distinct instance of a text feature within a story has a distinct text-story thread (ITextStoryThread), including the main story text (the primary story thread). Text-story threads are
associated with a particular TextIndex, known as the anchor, obtainable with ITextStoryThreadDict::GetAnchorTextRange. This does not apply to the primary story thread, which is the
only thread that always returns an anchor position of zero. An anchor for a footnote appears to
the users in the user interface as the footnote-reference character. The anchor for a table
appears as the table itself.
A text-story thread can contain anchors for other text story threads, subject to some restrictions; for example, footnotes can contain tables but not other footnotes. A feature does not
322
Text Fundamentals
Text content
need to associate the story thread with a particular TextIndex, in which case it is associated
with the end of story marker for the primary text-story thread (these are unanchored threads).
A thread block is a contiguous text range (TextRange) that contains the contents for one textstory thread. If a text-story thread contains a set of anchors, the thread blocks associated with
these anchors directly follow the thread block for this story thread. This is depicted in
Figure 117. Note how the thread blocks for the table directly follows the thread block for the
footnote into which it is anchored. In the figure, ITextStoryThreadDict manages the story
threads for a particular use of a feature. There is one story thread for the primary story, one for
each footnote, and two for the embedded table.
FIGURE 117
A story with two footnotes, one containing a two-cell table
The auto-numbering
characters have been
edited out of the
footnotes.
kTextStoryBoss
ITextStoryThreadDict
UID3 :kTableModelBoss
UID1 :
kFootnoteReferenceBoss
ITextStoryThreadDict
ITextStoryThreadDict
QueryStoryThread
QueryStoryThread
UID2 :
kFootnoteReferenceBoss
QueryStoryThread
QueryStoryThread
Text storage subsystem
ITextStoryThreadDict
I ndex
Data
Inlines
0 1 2 3 4 5
a 0x4 b 0x4 c #
UID1 UID2
Primary story thread
Thread block 0
Text Fundamentals
6 7 8
y 0x16 P
UID3
9
#
10 11
w #
12 13
x #
14 15
K #
Footnote 1 thread
Table
thread 1
Table
thread 2
Footnote
2
thread
Thread block 1
Thread block 2
Thread block 3
QueryStoryThread
Thread block 4
On data strand:
0x4 == kTextChar_FootnoteMarker
0x16 == kTextChar_Table
See TextChar.h.
323
Text Fundamentals
Text content
The text-story thread dictionary (ITextStoryThreadDict) maintains the set of story threads
associated with a particular use of a feature. In Figure 117, each feature has one associated story
thread, apart from the table. Each cell in the table has a distinct story thread. ITextStoryThreadDict is responsible for managing the set of threads in the table.
Text story threads have a hierarchical relationship (the root being the primary story thread).
Individual features are responsible for managing a set of one or more story threads using the
story-thread dictionary (ITextStoryThreadDict). In Figure 117, kFootnoteReferenceBoss
(UID1) manages one text-story thread, while kTableModelBoss (UID3) manages two story
threads, one for each cell. The story maintains a dictionary hierarchy (ITextStoryThreadDictHier), which provides the ability to iterate across all objects that contain a dictionary associated
with the story.
Text formatting
This section describes how the style information that controls the look of text is maintained
within the application.
Text attributes
The fundamental component for formatting text is the text attribute (signature interface IAttrReport). This is a lightweight, non-persistent set of boss objects, each of which describes one
text property; for example, the color or point size of text. Attributes are interpreted by the textcomposition subsystem, to define how the text appears when composed.
Attributes target either arbitrary ranges of characters (like the color or point size of a set of
characters) or a paragraph (like applying a drop cap or setting paragraph alignment).
Attributes are managed in an AttributeBossList list, a persistent container class. Consider an
AttributeBossList list to be a hash table of attributes, the key being the ClassID. One implication of this is there can be only one instance of any particular attribute within an AttributeBossList list (for example, no conflicting text-color attribute within an AttributeBossList).
Attributes can be applied directly (within an AttributeBossList list) to the appropriate paragraph or character-attribute strand, or they can be applied to a text style. An example of applying the attribute directly to the strand would be italicizing text directly using the text editor.
Such changes are local to the text being formatted and known as local overrides to the current
style.
An attribute boss class that supports the IAttrImportExport interface can participate in the
import and export of InDesign tagged text files.
Default attributes
The application provides a set of default attributes. These attributes are maintained within an
AttributeBossList on both the session and document workspace. Default attributes define a set
of overrides that are applied to any new stories created. These AttributeBossList lists reflect the
state of the Character and Paragraph panels when no documents are open (session workspace)
or new stories are created (document workspace). When a new document is created, the document workspace inherits the value of the default attributes from the session. The default
324
Text Fundamentals
Text content
attributes can be updated directly through the Paragraph and Character panels or by selecting
styles in the styles panels; the defaults are updated to match the style.
Text styles
A text style (signature interface IStyleInfo) provides a mechanism to name and persist a particular set of text attributes with particular values. The application supports two kinds of styles for
text, paragraph styles and character styles. Each style has an AttributeBossList list that maintains the set of attributes that apply to that style. Access to the attributes in a style is achieved
through the ITextAttributes wrapper interface. Character styles are associated with character
attributes; however, paragraph styles can be associated with both paragraph and character
attributes. Paragraph styles are applied to whole paragraphs; character styles can be applied to
arbitrary ranges of text.
All text must be associated with both a paragraph style and character style. To support this, the
application defines two default styles: the root character style and root paragraph style (known
in the application user interface as [No Paragraph Style]). The root paragraph style contains a
complete set of paragraph-based and character-based attributes. This defines the default look
of text with no further formatting applied. The root character style is empty; it exists purely to
provide a root for character styles.
Styles form a hierarchy rooted at the root style. Each style (except the root style) is based on a
style. The AttributeBossList list for a particular style record only the differences from the style
on which it is based; that is, the set of attributes the particular style overrides. This is shown in
Figure 118. In the figure, the kStyleBoss for the root paragraph style defines all attributes with a
default value. The kStyleBoss for the new style (Yellow Text) contains only the attributes that
differ from the root style. The root character style contains no attributes; it exists to provide a
root to the character-style hierarchy.
Text Fundamentals
325
Text Fundamentals
Text content
FIGURE 118
Paragraph-style example
The application provides two basic styles, for character attributes and paragraph attributes.
These are the default styles applied to new stories (assuming no other style is selected in a styles
panel). The basic styles cannot be deleted, but you can modify them with plug-ins. Initially, the
basic styles are based on the root styles; however, the parent styles can be changed.
326
Text Fundamentals
Text content
Users can create, edit and delete folders, called Groups, in the Character, Paragraph, and Object
Styles palettes. Style group is a collection of styles or groups. The user also can nest groups
inside groups and drag styles within the palette to edit the contents of a group. Styles do not
need to be inside a group and can exist at the root level of the palette. The group concept also is
available in the object style palette.
Character and paragraph styles are not required to have unique names across groups; however,
sibling styles and groups must have unique names within a parent (i.e., at the same level in the
hierarchy). Style names are case-sensitive, so the user can have both “heading” and “Heading”
in the same folder. A group and a style cannot have the same name. Since style names are not
unique, they are displayed in the user interface as full paths.
The alphabetical sort is an action applied to the entire list of styles. Users can rearrange the
styles and groups arbitrarily at any time; to view the list alphabetically, they can resort the list
with the Sort By Name command. Styles are intermixed with groups for sorting. Sorting is
undo-able. The sorting is language specific, based on the user-interface language. Reserved
styles like [Basic Paragraph] or [Basic Graphics Frame] are not reordered by this command.
Styles are accessible through the style group manager (signature interface IStyleGroupManager) on the document (kDocWorkspaceBoss) and session (kWorkspaceBoss) workspace boss
classes. When a document is created, its style group manager inherits the existing set of styles
from the session workspace. There are multiple style group managers supported by the workspaces, each identified by a distinct interface identifier; for example,
IID_IPARASTYLEGROUPMANAGER and IID_ICHARSTYLEGROUPMANAGER. IStyleGroupManager has one automatically created object of type IStyleGroupHierarchy called
“Root Hierarchy,” represented by the kStyleGroupHierarchyBoss. IStyleGroupManager provides access to kStyleGroupHierarchyBoss through its GetRootHierarchy() method.
This Root Hierarchy is created on the first call to the GetRootHierarchy method. The kStyleGroupHierarchyBoss holds all the styles (kStyleBoss) and style groups (kStyleGroupBoss) at
root level. The key interface on the kStyleGroupHierarchyBoss is IStyleGroupHierarchy, which
stores the persistent UID-based style tree hierarchy so you can query information like parent/child node information. All children of this root hierarchy must support the IStyleGroupHierarchy interface; therefore, IStyleGroupHierarchy also is aggregated on the
kStyleBoss and kStyleGroupBoss, so access to the style hierarchy also is available if you have
access to the UID of any style or style group in the hierarchy. With IStyleGroupHierarchy, you
gain access to the style-tree hierarchy the user sees in the style panel. IStyleGroupHierarchy has
a method to traverse the style hierarchy, which should be used to iterate through all styles.
IStyleGroupManager also provides two overloaded FindByName() methods that return the
UID of a style or style group.
Figure 119 shows how to navigate the styles, given a particular workspace. The style group
manager (IStyleGroupManager on document and session workspaces) provides access to all
paragraph and character styles (kStyleBoss) through the GetRootHierarchy() method, which
returns an IStyleGroupHierarchy on the kStyleGroupHierarchyBoss. kStyleGroupHierarchyBoss holds all the styles (kStyleBoss) and style groups (kStyleGroupBoss) at root level. Styles
can be iterated over using IStyleGroupHierarchy::GetDescendents or accessed by name using
IStyleGroupManager::FindByName. Given a particular style, the style it is based on can be
accessed using IStyleInfo::GetBasedOn.
Text Fundamentals
327
Text Fundamentals
Text content
FIGURE 119
Navigating styles in a document
«boss»
kDocWorkspaceBoss
1
The IStyleGroupManager is the access point
for style set management. It gives access
to the root IStyleGroupHierarchy from where
you can get the full style set hierarchy
+IStyleGroupManager
IStyleGroupManager::GetRootStyleUID
1
IStyleGroupManager::GetRootHierarchy
1
IStyleGroupHierarchy is aggregated
on kStyleGroupHierarchyBoss,
kStyleBoss, and kStyleGroupBoss, it is
used to access kStyleBoss/kStyleGroupBoss
current object may have.
+IStyleGroupHierarchy
IStyleGroupHierarchy::QueryChild 1
1
0..*
«boss»
kStyleBoss
0..1
Holds all the styles (kStyleBoss)
and style groups (kStyleGroupBoss)
at root level.
«boss»
kStyleGroupHierarchyBoss
+IStyleGroupHierarchy
+IStyleInfo
IStyleGroupHierarchy can be
used to access its styles and
groups.
1
0..*
IStyleGroupHierarchy::QueryChild
0..*
«boss»
kStyleGroupBoss
+IStyleGroupHierarchy
IStyleGroupHierarchy::QueryChild
IStyleInfo::GetBasedOn
1
Represents a text style
1
1
0..*
IStyleGroupHierarchy::QueryChild
Represents a style group
Sorting the style hierarchy is done through the IStyleGroupHierarchy::Sort() method, when it
is invoked, it sorts its children and calls its children’s Sort method. The call to Sort() on the root
hierarchy recursively sorts all styles. This methods takes a flag to indicate whether to sort in
descending/ascending order, all child or immediate child, etc.
Text formatting and stories
All text has an associated character- and paragraph-attribute style. This style is either the root
style (which defines a default value for the complete set of attributes used in composition) or a
style that records differences from the root style. Styles form a hierarchy in which the values of
attributes in child nodes override those in parent nodes. As well as having the character and
paragraph styles defined on the attribute strands, local overrides can be applied. This occurs
when a formatting change is made without modifying a style; for example, selecting a word and
making it bold.
328
Text Fundamentals
Text content
Attributes describe one of the text’s appearance and are used by the composition engine when
rendering the associated glyphs. IComposeScanner (on the kTextStoryBoss) provides APIs that
allow the value of any attribute to be deduced for any TextIndex. Figure 120 shows an example
of how styles and local attribute overrides combine to describe the exact look of text. The paragraph-attribute strand specifies the text will use the custom style pStyle and two local overrides. The overrides specify the body and last line of the paragraph are to be centered. The
character-attribute strand specifies the root style for the word “Aliens,” a local +bold override
for the word “and,” and the word “Earth” to use the custom style cStyle.
FIGURE 120
Formatting and strands
Story
length = 17
0
kTextDataStrandBoss
1
A l
2
3
4
5
i e n s
kParaAttrStrandBoss
6
7
8
9
10
a n d
Start: 0
End: 16
UID: pStyle
11
12
13
14
15
16
E a r t h #
Overrides: + last: centered + body: centered
kCharAttrStrandBoss
Start: 0
End: 6
UID: root
Start: 7
End: 9
UID: root
Start: 10
End: 16
UID: cStyle
Overrides: None
Overrides: +bold
Overrides: None
kTextAttrFontStyleBoss
+bold
kStyleBoss
name: root
based on: none
AttributeBossList
always empty
kStyleBoss
name: cStyle
kTextAttrAlignBodyBoss
kTextAttrAlignLastBoss
+ body: centered
+ last: centered
based on: root
AttributeBossList
kTextAttrUnderlineBoss
+underline
kStyleBoss
name: root
based on: none
AttributeBossList
all attributes
described here
kTextAttrPointSizeBoss
kStyleBoss
name: pStyle
based on: root
AttributeBossList
+size: 24 pt
Text Fundamentals
329
Text Fundamentals
Text content
Class associations
Figure 121 shows the major classes and associations for text content.
FIGURE 121
Major classes and associations for text content
«interface»
kDocBoss
kTextStoryBoss
IStoryList
GetNthTextModelUID
+ GetNthTextModelUID() : UIDRef
«interface»
ITextModel
+ QueryStrand(PMIID, ClassID) : IPMUnknown *
QueryStrand
QueryStrand
kTextDataStrandBoss
QueryStrand
kParaAttrStrandBoss
kCharAttrStrandBoss
IStrand
IStrand
QueryStrand
kOwnedItemStrandBoss
IStrand
IStrand
ITextStrand
«interface»
«interface»
IAttributeStrand
IItemStrand
+ GetStyleUID() : UID
+ GetNthOwnedUID() : UID
GetStyleUID
kStyleBoss
Other owned item boss
kStyleBoss
kTableFrameBoss
kStyleBoss
kStyleBoss
330
GetNthOwnedUID
kFootnoteReferenceBoss
kInlineBoss
Text Fundamentals
Text presentation
Text presentation
This section describes the presentation model responsible for managing the final look of text.
The layout factors (such as the types of containers with which text can be associated) are
described before the representation of the rendered text (the wax) is presented.
The text-presentation model represents both the information required to place a set of glyphs
on a spread and the management of those glyphs within the application. In Figure 109, this is
represented by the text layout and wax components. Text layout is concerned with the containers for text and the attributes of a container that can affect how text is placed in respect to that
container. The wax models the actual glyphs rendered into the layout.
Text layout
Text layout defines the shape and form of the visual containers in which text is composed and
displayed. The textual content is maintained in the text model of the story associated with the
layout. This content is divided into one or more story threads, each representing an independent flow of text. See “Story threads” on page 320.
In text composition, text flows from the story thread into its associated visual containers to create wax that fits the layout. Text layout, in turn, displays the wax. When the layout is updated,
text is recomposed to reflect the change.
Text frame
The text frame is the fundamental container used to place and display text. Text frames have
properties, like the number of columns, column width, gutter width, and text inset, that control
where text flows. A text frame is represented by several associated boss classes.
Figure 122 shows a text frame containing two columns, with one line of text displayed in each
column. Figure 123 shows how this instance of a text frame is represented internally.
FIGURE 122
Text frame
TEXTFRAME
-POINTINSET
-POINTGUTTER
OUTPORT
INPORT
COLUMN
Text Fundamentals
COLUMN
331
Text Fundamentals
Text presentation
FIGURE 123
Instance diagram of text frame
Text frame
«boss class»
graphic frame : layout::kSplineItemBoss
UID = 176
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
IHierarchy::QueryParent
«boss class»
controller : kMultiColumnItemBoss
UID = 177
IHierarchy::QueryChild(0)
«boss class»
column 1 : kFrameItemBoss
UID = 179
IHierarchy::QueryParent
IHierarchy::QueryChild(1)
«boss class»
column 2 : kFrameItemBoss
UID = 180
As shown in Figure 124, a text frame consists of a graphic frame (kSplineItemBoss) containing
one multicolumn item (kMultiColumnItemBoss). The multicolumn item (kMultiColumnItemBoss) is the controller for the text frame and contains one or more columns
(kFrameItemBoss). The ITextColumnSizer interface stores the major properties of the text
frame, such as the number of columns. IMultiColumnTextFrame provides access to the associated story, frame list, and the range of text displayed.
332
Text Fundamentals
Text presentation
FIGURE 124
Structure of a text frame
Text frame
Represents the graphic
frame
Represents the controller
for the text frame
«boss class»
layout::kSplineItemBoss
1
IHierarchy::QueryParent
1
IHierarchy::QueryChild
«boss class»
kMultiColumnItemBoss
1
1..*
Represents a column
within the text frame
IGraphicFrameData
ITextColumnSizer
The relationship between the graphic
frame and its content is maintained
by the IHierarchy interface.
IMultiColumnTextFrame
IHierarchy::QueryParent
IHierarchy::QueryChild
«boss class»
kFrameItemBoss
IGraphicFrameData::GetTextContentUID
is a helper method that returns the UID
of the associated kMultiColumnItemBoss
or kInvalidUID if the frame does not have
textual content.
Stores the major properties
of a text frame, such as the
number of columns
Provides access to the associated
story, frame list, and the range of
text displayed
ITextFrameColumn
The relationship between the graphic frame (kSplineItemBoss) and its content is maintained by
the IHierarchy interface. Use IHierarchy::QueryChild to navigate from the graphic frame to its
associated multicolumn item, then onto individual columns. Conversely, use IHierarchy::QueryParent to navigate up from a column to the multicolumn item and on up to the graphic
frame.
Frame list
A story (kTextStoryBoss) associates the text frames that display its content through the frame
list (kFrameListBoss). Conversely, a text frame associates the story whose text it displays
through the frame list. Figure 125 and Figure 126shows the story and frame list objects for the
example introduced in Figure 122 and Figure 123. Figure 127 shows the general structure of
the frame list (IFrameList) for a story.
FIGURE 125
Text Fundamentals
Frame list
333
Text Fundamentals
Text presentation
FIGURE 126
Instance diagram of frame list
Text frame
«boss class»
graphic frame : layout::kSplineItemBoss
UID = 176
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
IHierarchy::QueryParent
«boss class»
controller : kMultiColumnItemBoss
UID = 177
IHierarchy::QueryParent
IHierarchy::QueryChild(1)
IHierarchy::QueryChild(0)
«boss class»
column 2 : kFrameItemBoss
UID = 180
«boss class»
column 1 : kFrameItemBoss
UID = 179
IFrameList::QueryNthFrame(1)
IFrameList::QueryNthFrame(0)
«boss class»
frame list : kFrameListBoss
ITextFrameColumn::QueryFrameList UID = 174
ITextFrameColumn::QueryFrameList
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story : The Text Model::kTextStoryBoss
UID = 158
334
Text Fundamentals
Text presentation
FIGURE 127
Structure of the frame list (IFrameList) for a story
Text frame
«boss class»
layout::kSplineItemBoss
IGraphicFrameData
1
IHierarchy
1
«boss class»
kMultiColumnItemBoss
IMultiColumnTextFrame
1
IHierarchy
1..*
«boss class»
kFrameItemBoss
1..*
ITextFrameColumn
IFrameList::QueryNthFrame
ITextFrameColumn::QueryFrameList
1
IMultiColumnTextFrame::QueryFrameList
«boss class»
kFrameListBoss
Associates the story and the
frames that display its text
Represents text content
IFrameList
1
ITextModel::QueryFrameList
1
IFrameList::QueryTextModel
«boss class»
The Text Model::kTextStoryBoss
ITextModel
ITextFrameColumn::QueryTextModel
Provides access to the
story's text content
IMultiColumnTextFrame::QueryTextModel
Text Fundamentals
335
Text Fundamentals
Text presentation
Threading and text frames
The text in a frame can be independent of other frames, or it can flow between threaded
frames. Threading of text is the process of connecting the flow of text between frames. Do not
confuse threading text between frames with text-story threads. The connections can be visualized by selecting a text frame and choosing View > Show Text Threads, as shown in Figure 130.
Threaded frames share a common underlying story (kTextStoryBoss) and can be on the same
spread or different spreads in the same document.
Figure 128 shows two unthreaded text frames. Because each text frame is associated with a distinct story, editing the text in one of the frames does not affect the text in the other frame (see
Figure 129).
FIGURE 128
Unthreaded text frames
INPORTATBEGINNINGOFSTORY
OUTPORTINDICATINGOVERSETTEXT
4WO-COLUMNFRAME
/NE-COLUMNFRAME
INPORTATBEGINNINGOFSTORY
336
OUTPORTATENDOFSTORY
Text Fundamentals
Text presentation
FIGURE 129
Instance diagram of unthreaded text frames
Two-column text frame
One-column text frame
«boss class»
graphic frame 1 : layout::kSplineItemBoss
UID = 176
«boss class»
graphic frame 2 : layout::kSplineItemBoss
UID = 262
IHierarchy
IHierarchy
«boss class»
controller : kMultiColumnItemBoss
UID = 177
«boss class»
controller : kMultiColumnItemBoss
UID = 264
IHierarchy
IHierarchy
IHierarchy
«boss class»
column 1 : kFrameItemBoss
UID = 179
«boss class»
column 2 : kFrameItemBoss
UID = 180
«boss class»
column 1 : kFrameItemBoss
UID = 281
IFrameList::QueryNthFrame(0)
IFrameList::QueryNthFrame(1)
-IFrameList::QueryNthFrame(0)
ITextFrameColumn::QueryFrameList
ITextFrameColumn::QueryFrameList
-ITextFrameColumn::QueryFrameList
«boss class»
frame list : kFrameListBoss
UID = 174
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story 1 : The Text Model::kTextStoryBoss
UID = 158
«boss class»
frame list : kFrameListBoss
UID = 335
-ITextModel::QueryFrameList
-IFrameList::QueryTextModel
«boss class»
story 2 : The Text Model::kTextStoryBoss
UID = 319
When the text frames that display a story cannot display all the text, the unseen text is called
“overset text.” A red plus sign (+) on the outport of the two-column frame indicates the story
has overset text.
If the user clicks the outport of the two-column frame with the Selection tool and then clicks
the inport of the one-column frame, the frames become threaded. In Figure 130, the text from
the two-column frame that was overset now flows through the one-column frame. During this
process, the text from the story underlying the one-column frame is appended to the story
underlying the two-column frame. The frame list (kFrameListBoss) and story (kTextStoryBoss) underlying the one-column frame are deleted. Once threaded, the two frames share the
same underlying story (see Figure 131).
Text Fundamentals
337
Text Fundamentals
Text presentation
FIGURE 130
Threaded text frames
INPORTATBEGINNINGOFSTORY
OUTPORTINDICATINGLINKTONEXTFRAME
4WO-COLUMNFRAME
/NE-COLUMNFRAME
INPORTINDICATINGLINKFROMPREVIOUSFRAME
FIGURE 131
Instance diagram of threaded text frames
Two-column text frame
One-column text frame
«boss class»
graphic frame 1 : layout::kSplineItemBoss
UID = 176
«boss class»
graphic frame 2 : layout::kSplineItemBoss
UID = 262
IHierarchy
IHierarchy
«boss class»
controller : kMultiColumnItemBoss
UID = 177
«boss class»
controller : kMultiColumnItemBoss
UID = 264
IHierarchy
IHierarchy
IHierarchy
«boss class»
column 1 : kFrameItemBoss
UID = 179
«boss class»
column 2 : kFrameItemBoss
UID = 180
«boss class»
column 1 : kFrameItemBoss
UID = 281
IFrameList::QueryNthFrame(0)
IFrameList::QueryNthFrame(1)
ITextFrameColumn::QueryFrameList
ITextFrameColumn::QueryFrameList
«boss class»
frame list : kFrameListBoss
UID = 174
ITextFrameColumn::QueryFrameList
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story 1 : The Text Model::kTextStoryBoss
UID = 158
338
OUTPORTINDICATINGOVERSETTEXT
IFrameList::QueryNthFrame(2)
Text Fundamentals
Text presentation
Parcels
A flow of text content within a story is a story thread, is represented by a boss class with an
ITextStoryThread interface. A parcel is a visual container into which the text of a story thread is
flowed for layout and display. A parcel is represented by a boss class with an IParcel interface.
The text of a story thread (ITextStoryThread) can be composed into a list of parcels (IParcel) in
a parcel list (IParcelList).
Examples of application-provided parcels are as follows:
z
Text frame item (kFrameItemBoss)
z
Table cell parcel (kTextCellParcelBoss)
z
Footnote parcel (kFootnoteParcelBoss)
Figure 132 shows how text frames support story threads and parcels. The text in the primary
story thread is displayed through parcels given by the parcel list on the frame list. These parcels
are the frame items of the text frames that display the story.
Text Fundamentals
339
Text Fundamentals
Text presentation
FIGURE 132
Structure of the parcel implementation for text frames
Text frame
«boss class»
layout::kSplineItemBoss
1
1
«boss class»
kMultiColumnItemBoss
1
Displays a range of
text from the
primary story thread
1..*
IParcel
«boss class»
kFrameItemBoss
1..*
IParcelList::QueryParcel
The range of text displayed
in the parcels can be found
in ITextParcelList.
1
IParcel::QueryParcelList
ITextParcelList
IParcelList
«boss class»
kFrameListBoss
1
ITextStoryThread::QueryParcelList
Text in the primary story thread
flows into parcels in this parcel list.
Effectively, this flows the content
through the text frames.
1
ITextParcelList::QueryStoryThread
«boss class»
The Text Model::kTextStoryBoss
ITextStoryThread
ITextStoryThreadDict
ITextModel
The primary story thread
340
Text Fundamentals
Text presentation
Figure 133 shows how tables support story threads and parcels. The text for the story thread of
a table cell (kTextCellContentBoss) is displayed through the parcels given by its parcel list.
These parcels are text cell parcels (kTextCellParcelBoss). For more information, see the
“Tables” chapter.
FIGURE 133
Structure of the parcel implementation for tables
Text frame
A table frame (kTableFrameBoss) represents the rows
of a table that lie in a frame item (kFrameItemBoss).
The rows of a table can flow across one or more frame items
(kFrameItemBoss).
«boss class»
layout::kSplineItemBoss
1
1
0..*
1
«boss class»
kMultiColumnItemBoss
«boss class»
table::kTableFrameBoss
ITableLayout::frame_iterator::QueryFrame
1
Displays a range of
text from the cell's
story thread
1
1
ITableFrame::QueryFrame
1..*
1..*
IParcel
«boss class»
kFrameItemBoss
1..*
IParcel::GetFrameUID
IParcelList::QueryParcel
IParcel::QueryParcelList
«boss class»
kFrameListBoss
1
ITableFrame::const_parcel_iterator
«boss class»
table::kTextCellParcelBoss
1..*
Each text cell in a table knows the
frame item (kFrameItemBoss)
in which it lies; unless it is overset, in
which case the cell is not displayed.
1
ITableFrame
IOwnedItem
IWaxAnchorPt
ITextParcelList
IParcelList
ITextStoryThread::QueryParcelList
IParcel
IParcelList::QueryParcel
The range of text displayed
in the parcels can be found
in ITextParcelList.
1
IParcel::QueryParcelList
«boss class»
table::kTextCellContentBoss
1..*
ITextParcelList
IParcelList
ITextStoryThread
ITextStoryThreadDict::QueryThread
Each cell has a story thread
whose text flows into parcels
in the cell's parcel list.
1
1
ITableFrame::QueryModel
ITextParcelList::QueryStoryThread
«boss class»
The Text Model::kTextStoryBoss
1
ITextStoryThread::GetDictUID
«boss class»
table::kTableModelBoss
1
ITextStoryThread
ITextStoryThreadDict
ITextModel
lTableModelList::QueryNthModel
ITextStoryThreadDict
ITableTextContainer
ITableModel
0..*
ITableTextContainer::QueryTextModel
A table model exists for
each table embedded in
the story.
Text Fundamentals
341
Text Fundamentals
Text presentation
Span
Each object that displays text is associated with a story thread. After composition, the object
stores the index (TextIndex) into the text model of the first character it displays and the total
number of characters it shows. This text range is the span. The span is available on several
interfaces (see Table 88 and Table 89). Table 88 lists the APIs that indicate the range of text displayed. Table 89 lists the APIs that find the parcel or frame displaying a specific TextIndex.
The ranges returned from these APIs are accurate only if the text in the object is fully
composed. Before relying on these methods, you must check for damage and for the
object to recompose, if necessary.
TABLE 88 APIs that indicate the range of text displayed
NOTE:
API
Description
IFrameList
Gives the range of text displayed by each frame item (kFrameItemBoss).
ITextFrameColumn
Gives the range of text displayed by the frame item (kFrameItemBoss).
IMultiColumnTextFrame
Gives the range of text displayed by the multicolumn item
(kMultiColumnItemBoss).
ITextParcelList
Gives the range of text displayed by a parcel. All objects that display composed text
are a kind of parcel (IParcel).
TABLE 89 APIs that find the parcel or frame displaying a specific TextIndex
API
Description
IFrameList::QueryFrameContaining
Use this method to find the frame item (kFrameItemBoss) that displays
the character at a specific index (TextIndex) in the text model.
ITextModel::QueryTextParcelList
Call this to find the text parcel list associated with a given TextIndex. After
you have the ITextParcelList, use ITextParcelList::GetParcelContaining,
then IParcelList::QueryParcel. All objects that display composed text are a
kind of parcel (IParcel). As a result, this approach finds parcels for text
frames, table cells, or other new kinds of parcels. Note, however, that some
text embedded in the text model may not be composed or displayed in a
parcel.
IWaxIterator
After you know the text model (ITextModel) and the range of text you
want, use this class to get the wax data.
TextIterator
After you know the text model (ITextModel) and the range of text that you
want, use this class to get the character data.
Figure 134 shows a story displayed in a two-column frame. The range of text displayed in both
columns is available in the IMultiColumnTextFrame interface on the multicolumn item (kMultiColumnItemBoss). The range of text available in each frame item (kFrameItemBoss) is available in both ITextFrameColumn and ITextParcelList.
342
Text Fundamentals
Text presentation
FIGURE 134
Finding the range of text displayed in a frame or parcel
K-ULTI#OLUMN)TEM"OSS
3TART3PAN
K&RAME)TEM"OSS
3TART3PAN
K4EXT3TORY"OSS
)4EXT-ODEL
K&RAME)TEM"OSS
3TART3PAN
4HETEXTCONTENTISSTOREDin THEASSOCIATED
STORYgSTEXTMODEL$ETERMINETHERANGE
YOUWANT, andTHENUSETHE4EXT)TERATOR
class TOACCESSTHECHARACTERDATA
4EXT)NDEX
#HARACTER!LIENSAND%ARTHr
Conversely, you can discover the parcel (IParcel) displaying the character at a specific index
(TextIndex) in the text model. It is possible the TextIndex of interest is overset text that is not
displayed. The kind of parcel returned depends on the kind of story thread (ITextStoryThread)
that owns the range of text in which the TextIndex lies.
For example, if the TextIndex lies in the primary story thread, it is in the range 0 to (ITextModel::GetPrimaryStoryThreadSpan - 1). The primary story thread is displayed by the text
frames associated with the story. In this case, the frame item (kFrameItemBoss) that displays
the desired TextIndex is determined. Alternately, the desired TextIndex can be in the range
ITextModel::GetPrimaryStoryThreadSpan to (ITextModel::TotalLength - 1). Text in this range
is associated with a table or other feature that embeds text in stories, like notes, footnotes, or
tracked changes. If the feature displays composed text, the parcel displaying a specific TextIndex can be determined.
Text frames and the wax
After text is composed, each line is represented by a wax line (kWaxLineBoss). The wax for a
story is owned by the wax strand. A text frame is associated with its wax lines by the range of
text displayed, which is obtained from ITextFrameColumn or ITextParcelList. The range of text
displayed by the frame is maintained when text is recomposed. Wax lines are accessed using a
wax iterator, IWaxIterator, which is a C++ helper class created by calling IWaxStrand::NewWaxIterator.
Consider the wax for a text frame with two columns, each displaying one line of text. See
Figure 135. Here, the wax line (kWaxLineBoss) objects do not have UIDs. They are managed
by the wax strand (IWaxStrand on kFrameListBoss) that persists some wax data and regenerates the rest by recomposing the text when required.
Text Fundamentals
343
Text Fundamentals
Text presentation
FIGURE 135
Structure of the wax for a text frame
Text frame
The wax associated with a text frame can be discovered
by using a wax iterator (IWaxIterator) to find the wax lines
(IWaxLine) for the the range of text displayed by the frame
(see ITextFrameColumn).
«boss class»
layout::kSplineItemBoss
IWaxStrand::NewWaxIterator is called to create an
IWaxIterator.
IWaxIterator is a helper class that can access the wax lines
associated with a given TextIndex.
1
IHierarchy::QueryParent
1
IHierarchy::QueryChild
«boss class»
kMultiColumnItemBoss
IWaxIterator is not a descendant of IPMUnknown.
It should not be used in conjunction with InterfacePtr.
1
1..*
IHierarchy::QueryParent
IHierarchy::QueryChild
0..1
0..*
IWaxIterator::GetFirstWaxLine
«boss class»
The Wax::kWaxLineBoss
«boss class»
kFrameItemBoss
IWaxLine::GetParcelKey
1..*
IWaxLine
ITextFrameColumn
IFrameList::QueryNthFrame
ITextFrameColumn::QueryFrameList
1
IWaxIterator::GetFirstWaxLine
0..*
1
IWaxLine::QueryWaxStrand
The wax lines associated with a story
can be found by using a wax iterator
(IWaxIterator) for the range of text
of interest.
A wax line for text in the range
0..(ITextModel::GetPrimaryStoryThreadSpan-1)
will be associated with a frame item
(kFrameItemBoss), unless the wax line
represents overset text (in which case it
is not displayed).
A wax line for text in the range
ITextModel::GetPrimaryStoryThreadSpan..
(ITextModel::TotalLength-1) may be associated
with other kinds of parcel, such as table
text cells (kTextCellParcelBoss). This is not
shown on the diagram.
344
«boss class»
kFrameListBoss
IWaxStrand
IFrameList
1
ITextModel::QueryFrameList
1
IFrameList::QueryTextModel
Manages the wax for a story
«boss class»
The Text Model::kTextStoryBoss
Text Fundamentals
Text presentation
Text-frame options
You manipulate text-frame options from the dialog box invoked by choosing Object > Text
Frame Options. Default text-frame options store the properties inherited when a new text
frame is created. The text-frame options are stored in the ITextOptions interface on the document workspace. A distinct set of defaults are maintained on the session workspace and are
inherited. Figure 136 and Figure 137 show the boss classes and interfaces involved.
FIGURE 136
Interfaces storing text-frame options on a text frame
yPosition
Tile for line 1
Baseline
-15
-5
FIGURE 137
Interfaces storing default text-frame options
Document-specific defaults that are
inherited by new text frames
Session-specific defaults that are
inherited by new documents
IWorkspace
«boss class»
kDocWorkspaceBoss
ITextOptions
ITextOptions
IWorkspace
«boss class»
kWorkspaceBoss
Text-frame geometry
In Figure 138, the bounds of the page items that make up the text frame are tabulated in the
inner coordinates of the spline. In Table 90, the geometries of the multicolumn and column
items are expressed in the coordinate space of their parent, the spline. Effectively, they share a
common coordinate space.
FIGURE 138
Text-frame geometry
-100,-15
0,0
100,15
Text Fundamentals
345
Text Fundamentals
Text presentation
TABLE 90 Text-frame geometry
Item
Left
Top
Right
Bottom
kFrameItemBoss (left column)
-100.00
-15.00
-5.00
15.00
kFrameItemBoss (right column)
5.00
-15.00
100.00
15.00
kMultiColumnItemBoss
-100.00
-15.00
100.00
15.00
kSplineItemBoss
-100.00
-15.00
100.00
15.00
Text Inset
Text inset, a property of a text frame, is an area between the edge of the frame and the area
where text flows. If the text frame is regular in shape, you can specify the inset independently
for each side (see Figure 139). With irregularly shaped frames, the inset follows the contour of
the text frame, and there is only one inset value (see Figure 140).
FIGURE 139
Text Inset
TOPINSET POINTS
LEFTINSET POINTS
RIGHTINSET POINTS
BOTTOMINSET POINTS
FIGURE 140
Irregular text inset
INSET POINTS
346
Text Fundamentals
Text presentation
Figure 141 shows the structure of a text inset. The properties of the text inset are found in the
ITextInset interface on the text frame’s kSplineItemBoss. Even when there is no text inset in
effect (all insets have a value of 0 points), a kTextInsetPageItemBoss is associated with the text
frame and describes the contour of the text inset.
FIGURE 141
Structure of text inset
Text frame
1
«boss class»
layout::kSplineItemBoss
1
ITextInset::QueryInsetGeometry
ITextInset
1
«boss class»
kTextInsetPageItemBoss
IStandOffItemData::QueryPageItem
1
«boss class»
kMultiColumnItemBoss
«boss class»
kFrameItemBoss
Text inset is a special
kind of standoff, which
is the abstraction that
represents text wrap.
IStandOffItemData
IPathGeometry
Represents the contour
of the inset associated
with the text frame
«boss class»
kStandOffPageItemBoss
Text wrap
Text in a text frame is affected by other page items that have text wrap applied. Text wrap is represented by a standoff abstraction. A standoff describes whether there should be text wrap and,
if so, the contour around which the text should be wrapped.
Figure 142 shows an empty graphic frame with text wrap set at a 4-point offset around the
frame’s path. Figure 143 shows that if there is no intersection between the boundary of the
standoff and the boundary of the text frame, then the text in the text frame is not affected.
FIGURE 142
MODE
Form
Text-wrap properties applied to empty frame
EMPTYGRAPHICFRAME
K3PLINE)TEM"OSS
4HEMODEMARGIN,ANDFORMPROPERTIES
ARESTOREDININTERFACE)3TAND/FF$ATA.
MARGIN
Text Fundamentals
STANDOFF
K3TAND/FF0AGE)TEM"OSS
347
Text Fundamentals
Text presentation
FIGURE 143
Text wrap: empty frame with text wrap alongside text frame
Figure 144 shows what happens when the empty graphic frame overlaps the text frame. The
text within that frame is forced to wrap around the empty frame, which is said to repel the text.
By inverting the text wrap, a page item can indicate it is to be used to attract text within its
boundary rather than repel it. Figure 145 shows how this causes text to flow within the set
boundaries.
FIGURE 144
Text wrap: empty frame with text wrap overlapping text frame
FIGURE 145
Text wrap: empty frame with inverted text wrap overlapping text frame
The objects representing the text wrap for the sample are shown in Figure 146. The general
structure of text wrap is shown in Figure 147.
348
Text Fundamentals
Text presentation
FIGURE 146
Instance diagram for text wrap
IStandOffData::QueryStandOffGeometry
«boss class»
empty graphic frame : layout::kSplineItemBoss
UID = 200
Text frame
«boss class»
graphic frame : layout::kSplineItemBoss
UID = 176
«boss class»
stand off : kStandOffPageItemBoss
UID = 201
«boss class»
controller : kMultiColumnItemBoss
UID = 177
IStandOffItemData::QueryPageItem
«boss class»
column 1 : kFrameItemBoss
UID = 179
Empty graphic frame with standoff to represent the text wrap
FIGURE 147
Structure of text wrap
Text wrap is represented by a standoff. A standoff can be associated with various
kinds of page items such as graphic frames (kSplineItemBoss), pictures (kImageItemBoss,
etc.), and groups (kGroupItemBoss). They can each have distinct standoff
properties.
IStandOffData::QueryStandOffGeometry
1
0..1
IStandOffData
«boss class»
layout::kPageItemBoss
1
«boss class»
kStandOffPageItemBoss
IStandOffItemData
IPathGeometry
1
IStandOffItemData::QueryPageItem
Represents the contour
of the text wrap
«boss class»
layout::kDrawablePageItemBoss
Other kinds of page items not
shown on this diagram can also
have a standoff.
«boss class»
graphics::kImageBaseItem
«boss class»
layout::kSplineItemBoss
Text Fundamentals
«boss class»
layout::kGroupItemBoss
«boss class»
graphics::kImageItem
349
Text Fundamentals
Text presentation
Properties of a standoff are found in IStandOffData on the page item with the standoff. Use
IStandOffData::GetMode to obtain the kind of text wrap and IStandOffData::GetMargin to get
the offsets. Each page item has distinct standoff applied, so the page item can be a graphic
frame (kSplineItemBoss), a group (kGroupItemBoss), or another type.
When a text wrap is applied, a standoff boss class kStandOffPageItemBoss is associated with
the owning page item. The standoff boss class is not part of the page item instance hierarchy;
you do not navigate to it using IHierarchy. Instead, use IStandOffData::QueryStandOffGeometry to access the standoff boss class. The kStandOffPageItemBoss boss class has a path geometry that represents the contour of the standoff, and text in a text frame wraps to this contour.
Text on a path
Text on a path allows text to flow along the contour of a spline (kSplineItemBoss). Text on a
path frame has an inport and an outport. As a result, you can thread text in other frames to and
from it. Figure 148, Figure 149, and Figure 150 show the objects involved in text on a path.
FIGURE 148
Text on a path
kSplineItemBoss
Inport
350
kTOPSplineItemBoss
kMultiColumnItemBoss
kTOPFrameItemBoss
Outport
Text Fundamentals
Text presentation
FIGURE 149
Instance diagram for text on a path
Text on a path frame
IMainItemTOPData::GetTOPSplineItemUID
«boss class»
path : layout::kSplineItemBoss
UID = 144
ITOPSplineData::GetMainSplineItemUID
«boss class»
text on a path spline : kTOPSplineItemBoss
UID = 184
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
controller : kMultiColumnItemBoss
UID = 185
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
text on path frame item : kTOPFrameItemBoss
UID = 186
IFrameList::QueryNthFrame(0)
ITextFrameColumn::QueryFrameList
«boss class»
frame list : kFrameListBoss
UID = 182
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story : The Text Model::kTextStoryBoss
UID = 166
Text Fundamentals
351
Text Fundamentals
Text presentation
FIGURE 150
Structure of text on a path frame
IMainItemTOPData::GetTOPSplineItemUID
1
1
«boss class»
layout::kSplineItemBoss
Represents the contour along
which the text flows
IMainItemTOPData
IPathGeometry
Text on a path frame
1
0..1
owns text on path
«boss class»
kTOPSplineItemBoss
ITOPSplineData::GetMainSplineItemUID
1
1
1
ITOPSplineData
IHierarchy::QueryParent
IHierarchy::QueryChild
«boss class»
kMultiColumnItemBoss
1
1
IHierarchy::QueryParent
IHierarchy::QueryChild
-IMainItemTOPData::GetTOPTextContentUID
«boss class»
kTOPFrameItemBoss
1
ITOPFrameData
«boss class»
kFrameItemBoss
1..*
Text on a path frame is a special kind of text frame. It is represented
by three associated boss classes: kTOPSplineItemBoss,
kMultiColumnItemBoss, and kTOPFrameItemBoss.
Note that you will not find these boss classes on the hierarchy
(IHierarchy) that owns the kSplineItemBoss. Instead, they
can be accessed by interface IMainItemTOPData on the owning spline.
IFrameList::QueryNthFrame
ITextFrameColumn::QueryFrame
1
«boss class»
kFrameListBoss
1
1
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
The Text Model::kTextStoryBoss
Text on a path can be associated with any spline (kSplineItemBoss). The structure of text on a
path frame parallels that of a normal text frame. It is represented by three associated boss
classes: kTOPSplineItemBoss, kMultiColumnItemBoss, and kTOPFrameItemBoss. You can
navigate from a spline (kSplineItemBoss) to any associated text on a path (kTOPSplineItemBoss) using the IMainItemTOPData interface. You can navigate back using the ITOPSplineData interface. After you have an interface on the kTOPSplineItemBoss, you can navigate up
and down the text on a path frame using IHierarchy.
Figure 151 and Figure 152 show a text frame that also has text on a path running along its contour. In this example, the text displayed inside the frame and the text that runs its contour have
distinct underlying stories. If they were threaded, they would be associated with the same story.
352
Text Fundamentals
Text presentation
FIGURE 151
Text frame with text on a path
FIGURE 152
Instance diagram for text frame with text on a path
Text frame
Text on a path frame
IMainItemTOPData::GetTOPSplineItemUID
«boss class»
graphic frame : layout::kSplineItemBoss
UID = 183
ITOPSplineData::GetMainSplineItemUID
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
controller : kMultiColumnItemBoss
UID = 184
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
column 1 : kTOPFrameItemBoss
UID = 186
IFrameList::QueryNthFrame(0)
ITextFrameColumn::QueryFrameList
«boss class»
frame list : kFrameListBoss
UID = 181
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story : The Text Model::kTextStoryBoss
UID = 165
Text Fundamentals
«boss class»
text on path spline : kTOPSplineItemBoss
UID = 205
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
controller : kMultiColumnItemBoss
UID = 206
IHierarchy::QueryParent
IHierarchy::QueryChild(0)
«boss class»
text on path frame item : kTOPFrameItemBoss
UID = 207
IFrameList::QueryNthFrame(0)
ITextFrameColumn::QueryFrameList
«boss class»
frame list : kFrameListBoss
UID = 203
ITextModel::QueryFrameList
IFrameList::QueryTextModel
«boss class»
story : The Text Model::kTextStoryBoss
UID = 187
353
Text Fundamentals
Text presentation
The wax
The output of text composition is called the wax. The wax is responsible for drawing and hit
testing the text of a story.
The term “wax” comes from the manual paste-up process conventionally used to create magazine and newspaper pages before the advent of interactive pagination software. In the old process, typeset columns of text (galleys) were received from the typesetter. The galleys were run
through a waxer, cut into sections, and positioned on an art board. Wax was used as an adhesive to stick the sections on the art board.
The art board in InDesign is a spread, and you can think of the galleys as text frames. The typeset columns of text are the wax generated by text composition. The wax adheres—fixes—the
position of a line of text within a frame.
Wax strand, wax line, and wax run
The frame list (kFrameListBoss) manages the wax for a story by means of the wax strand
(IWaxStrand interface). The wax strand owns a collection of wax lines that contain wax runs, as
shown in Figure 153.
FIGURE 153
Class diagram of the wax
kTextStoryBoss
kFrameItemBoss
1
1
0..1
The wax
kFrameListBoss
1
IWaxStrand
0..m
Collection (kWaxLineBoss)
1
0..m
kWaxILGRunBoss
354
1
0..m
1
0..m
kWaxTextRunBoss
0..m
Collection
(TCY, WarichuuSet, etc.)
Text Fundamentals
Text presentation
The wax is organized as a hierarchy of collections and leaf runs. The most common case is a
series of wax lines, each with a collection of wax runs. A wax line is created for each line of text
in a frame. A wax run is created each time the appearance of the text changes within the line.
Examples of this include changes in point size, changes in font, and the interruption of the flow
of text in a line because text wrap causes text to flow around another frame.
The IWaxStrand interface on kFrameListBoss owns all wax lines for a story and is the root of
the wax tree. The IWaxIterator and IWaxRunIterator iterator classes provide access to wax lines
and wax runs.
kWaxLineBoss typically represents the wax for one line of text. (Warichuu is an exception to
this, where the Warichuu set contains multiple lines; all lines in the Warichuu set relate to one
kWaxLineBoss object.) Each line owns its wax runs, collections of wax runs, or combination of
runs and collections of runs. kWaxLineBoss provides access to its children through the IWaxCollection interface.
Any boss that supports the IWaxCollection interface is a parent in the wax hierarchy to boss
objects that support the IWaxRun interface. A boss can support both the IWaxCollection and
IWaxRun interfaces, creating levels in the hierarchy.
kWaxTextRunBoss is the object that represents the wax for ordinary text. This object stores the
information needed to render the glyphs and provides the interfaces that draw, hit test, and
select the text.
kWaxILGRunBoss represents the wax for inline frames. Inline frames allow frames to be
embedded in the text flow. An inline frame behaves as if it were one character of text and
moves along with the text flow when the text is recomposed. kWaxILGRunBoss provides the
drawing, hit testing, and selection behavior for inline frames.
For a complete list of wax-run boss classes, see the API reference documentation for IWaxRun.
Examples of the wax
This section describes the wax generated for some sample text frames. To recreate these examples, use the Text tool to create the frames. If you use another tool, like the place gun or a
graphic frame tool, the inner coordinate spaces for the frames will be different from those illustrated. For simplicity, use a sample page size 200 points wide and 100 points deep, and format
text using 8-point type, the Times Regular font, and 10-point leading. All text is composed by
the Adobe Paragraph Composer in a horizontal text frame.
Single line with no format changes
The wax generated for a single line of text with no format changes is shown in Figure 154. This
example has a frame 200 points wide and 30 points deep, containing one line of 8-point Times
Regular text with 10-point leading. The first baseline offset setting for the frame is set to leading. This forces the baseline of the first line of text in the frame to be offset from the top of the
frame by a distance equal to the leading value for the text (10 points).
Text Fundamentals
355
Text Fundamentals
Text presentation
FIGURE 154
Wax for single line with no format changes
left=-100, top=-15
Leading
-5
Width
WaxLine
X=-100
Y=-5
Width=55.76
Origin=0
Span=17
right=100, bottom= 15
WaxRun
X=0
Y=0
Width=55.76
Origin=0
Span=17
Each node in the wax tree records its position, its width, and a reference to the characters in the
text model it describes. The origin records the index into the text model to the first character in
the node. The span records the number of characters in the node.
Wax lines record their position in the inner coordinate space of the frame in which they are
displayed. Wax runs record their position as an offset relative to the position of the wax line.
Additional information is stored specific to the type of node—wax line or wax run.
Figure 155 exposes some additional data in the wax. A wax line stores the leading for the line.
A wax run stores font name, a font-transformation matrix, glyph information, and other data
necessary to describe how the text is to be drawn. A glyph is an element of a font.
356
Text Fundamentals
Text presentation
FIGURE 155
Information stored in the wax
-100,-15
100,15
WaxLine
X=-100
Y=-5
Width=55.76
Origin=0
Span=17
Leading=10
WaxRun
X=0
Y=0
Width=55.76
Transformation matrix
Origin=0
for 8-point text
Span=17
FontName=Times-Roman
FontMatrix=[8 0 0 8 0 0]
GlyphIDs=[0x24 0x4f 0x4c 0x48 0x51 0x56 0x3 0x44 0x51 0x47 0x3 0x28 0x44 0x55 0x57 0x4b 0xffffffff ]
GlyphWidths=[5.78 2.22 2.22 3.55 4.00 3.11 2.00 3.55 4.00 4.00 2.00 4.89 3.55 2.66 2.22 4.00 2.00]
When a character is composed, its character code is mapped to its corresponding GlyphID
from the font being used to display the character. The font provides the initial width of a glyph
at a particular point size. This width may be adjusted by composition to account for letter spacing, word spacing, and kerning. The GlyphID values and their widths after composition are
stored in the wax run, as shown in Figure 155.
Generation of wax runs
Wax runs are generated for each set of format changes or line breaks that occur in the rendered
text. Figure 156 illustrates the effect of applying text attributes that change text format.
Text Fundamentals
357
Text Fundamentals
Text presentation
FIGURE 156
Wax for single line with format changes
-100,-15
100,15
WaxLine
X=-100
Y=-5
Width=60.33
Origin=0
Span=17
WaxRun
X=0
Y=0
Width=22.89
Origin=0
Span=7
FontName=Times-Roman
FontMatrix=[8 0 0 8 0 0]
WaxRun
X=22.89
Y=0
Width=16.12
Origin=7
Span=3
FontName=Times-Bold
FontMatrix=[10 0 0 10 0 0]
WaxRun
X=39.01
Y=0
Width=21.32
Origin=10
Span=7
FontName=Times-Roman
FontMatrix=[8 0 0 8 0 0]
UnderlineMode=kUnderlineSingle
Figure 156 shows how a wax run is created each time the text attributes specify the appearance
of the text is to change. The arrows show the correspondence between the text drawn in the
frame and the wax run that describes its appearance and location. A distinct wax run would be
required for the following:
z
Each time the appearance of the text changes within the line.
z
Each line of text that appears in the parcel list for a story.
z
Each side of a line that is interrupted, either by using an irregular-shaped frame or by overlapping a page item with wrap turned on within the frame.
Text adornments
Text adornments provide a way for plug-ins to adorn the wax (composed text). Text adornments give plug-ins the opportunity to do additional drawing when the wax in a frame is
drawn. The wax draws the text and calls the text adornments it is aware of to draw. For this to
occur, the adornment must be attached to the wax.
Text adornments are boss objects attached by ClassID to an individual wax run. When the run
is drawn, text adornments are given control using the ITextAdornment interface.
358
Text Fundamentals
Text presentation
Calls to the ITextAdornment::Draw method are ordered by the priority value returned by the
ITextAdornment::GetDrawPriority method. Higher-priority adornments (smaller numbers)
are called before lower-priority adornments (larger numbers).
Text adornments cannot influence how text is composed; for example, they cannot influence
how characters are built into wax lines and wax runs by the text-composition subsystem.
Instead, text adornments augment the appearance of the text. A text adornment controls
whether the adornment is drawn in front of or behind the text.
There are two types of text adornments: local and global.
Local text adornments
A local text adornment is controlled by one or more text attributes. A local text adornment
provides visual feedback of the text the text attributes control. For example, an adornment can
highlight the background on which the text is drawn or strike out text by drawing a line in the
foreground.
The text attribute controls the process by interacting with text composition and associating the
adornment with a drawing style (see IDrawingStyle and kComposeStyleBoss). This causes the
text composer to attach the adornment to a wax run (see IWaxRenderData in “The wax” on
page 354).
Table 91 shows the main text adornments used by the application and the text attributes that
control them.
TABLE 91 Text adornments and associated text attributes
Adornment classID
Attribute classID
Description
kTextAdornmentIndexMarkBoss
kTAIndexMarkBoss
Index mark
kTextAdornmentKentenBoss
kTAKentenSizeBoss, etc.
Adorns text for emphasis
kTextAdornmentRubyBoss
kTARubyPointSizeBoss, etc.
Annotates base text
kTextAdornmentStrikethruBoss
kTextAttrStrikethruBoss
Strikes through text
kTextAdornmentUnderlineBoss
kTextAttrUnderlineBoss
Underlines text
kTextHyperlinkAdornmentBoss
kHyperlinkAttributeBoss
Hyperlink mark
To add a local text adornment, first add a new text attribute. In more complex situations, use
multiple text attributes to control one text adornment. For example, kTAKentenSizeBoss,
kTAKentenRelatedSizeBoss, kTAKentenFontFamilyBoss, and kTAKentenFontStyleBoss collectively control the kenten adornment implemented by the kTextAdornmentKentenBoss.
Local text adornments are attached to a drawing style and subsequently to an individual wax
run. These adornments always are called to draw, although they can choose not to draw anything.
Figure 157 shows the sequence of calls that result in an adornment being associated with a particular wax run.
Text Fundamentals
359
Text Fundamentals
Text presentation
FIGURE 157
Sequence for adding an adornment to a specific wax run
Composition
(abstract)
Creates kComposeStyleBoss
to hold the exact value of all
attributes as they apply to an
individual run of text
IAttrReport
(kSomeTextAttrBoss)
IDrawingStyle
(kComposeStyleBoss)
IWaxRun
(kWaxTextRunBoss)
create
TellComposition(ICompositionStyle* style, IDrawingStyle* draw)
AddTextAdornment(ClassID attrID, ClassID adornmentID)
Adornment ClassID added
to the wax run
AddAdornment(ClassID id)
The composition subsystem
invokes each attribute active for
the run of text, allowing the
attribute to p articipate in the
composition process.
The attribute can specify
adornments to be added to the
drawing style. These adornments
will be given the opportunity to
participate in the drawing process.
The composition subsystem calls the IAttrReport::TellComposition method of each attribute
that applies to a particular range of text. This gives the attribute a chance to interact with the
composition subsystem. This interaction includes the registering of adornments that are to be
associated with the wax and called when the wax is drawn (IDrawingStyle::AddTextAdornment). The adornments are attached to the wax as ClassIDs (IWaxRun::AddAdornment); i.e.,
the adornments are not instantiated at this point.
Figure 158 shows how adornments associated with a wax run are instantiated and called to
draw.
360
Text Fundamentals
Text presentation
FIGURE 158
Draw Subsystem
Sequence for local-adornment drawing
IParcelShape
(kFrameItemBoss)
ITextParcelList
(kFrameListBoss)
IWaxLineShape
(kWaxLineBoss)
IWaxRunText
(kWaxTextRunBoss)
IWaxRun
ITextAdornment
(kWaxTextRunBoss) (kSomeAdornmentBoss)
Draw
Draw
Draw
Draw
GetAdornments
create
Draw
Gets the list of adornments
(classID) held by the wax
run
The Draw signal propagates to the wax run (IWaxRunText on kWaxTextRunBoss). The wax
run (IWaxRun on kWaxTextRunBoss) is interrogated for the set of adornments registered
through composition, as shown in Figure 157. Each adornment (ITextAdornment) is created
on whichever boss class it is represented on, and the Draw method is called.
It is possible to associate a data interface (ITextAdornmentData) during registration (IAttrReport::TellComposition). This interface allows control over whether the adornment is associated
with wax runs.
Global text adornments
Global text adornments are attached to all wax runs, though they never have run-specific
adornment data. Global text adornments provide the ability to draw something without requiring the text to be recomposed. Unlike local text adornments, global text adornments do not
need to be attached to individual wax runs to draw; however, global text adornments are asked
if they are active before they are called to draw.
For example, the Show Hidden Characters feature is implemented using a global text adornment, kTextAdornmentShowInvisiblesBoss. The adornment behaves as if it is attached to all
runs, but is called only when the adornment requests draw.
Text Fundamentals
361
Text Fundamentals
Text presentation
Table 92 lists global text adornments.
TABLE 92 Global text adornments
Adornment boss
Description
kTextAdornmentHJKeepsVBoss
Highlights keeps violations.
kTextAdornmentMissingFontBoss
Highlights missing fonts.
kTextAdornmentMissingGlyphsBoss
Highlights missing glyphs.
kDynamicSpellCheckAdornmentBoss
Squiggle line to mark spell checks.
kTextAdornmentShowInvisiblesBoss
Shows hidden characters.
kPositionMarkerLayoutAdornmentBoss
Draws position marker.
kTextAdornmentShowKinsokuBoss
Shows Japanese character line break.
kTextAdornmentShowCustomCharWidthsBoss
Marks character width as a filled rectangle.
kXMLMarkerAdornmentBoss
XML marker.
Global text adornments are service providers. The ServiceID is kGlobalTextAdornmentService. Global text adornments aggregate the IK2ServiceProvider interface and use the default
implementation kGlobalTextAdornmentServiceImpl.
Global adornments have the opportunity to participate in drawing if they are enabled. Figure
Figure 159 shows the sequence of events that cause IGlobalTextAdornment::Draw to be called.
For brevity, the calls to the adornment to determine whether it is enabled (IGlobalTextAdorbment::GetIsActive) and to access the ink bounds of the adornment (IGlobalTextAdornment::GetInkBounds) are not shown.
362
Text Fundamentals
Text presentation
FIGURE 159
Draw Subsystem
Sequence diagram for drawing of global adornments
IParcelShape
(kFrameItemBoss)
IK2Servi ceReg istry
(kSessionBoss)
ITextParcelList
(kFrameListBoss)
IWaxLineShape
(kWaxLineBoss)
IWaxRunText
(kWaxTextRunBoss)
IGlobalTextAdornment
(kSomeAdornmentBoss)
Draw
GetNthServiceProviderID(kGlobalTextAdornmentServiceID)
Draw
Draw
Draw
Draw
Builds a list of global
adornments
requesting to be
drawn
Invokes Draw on all
adornments on the
list previously
constructed
When a frame (IParcelShape on kFrameItemBoss) is instructed to draw, it queries the service
registry (IK2ServiceRegistry on kSessionBoss) for the set of global text adornments that are
enabled (IGlobalTextAdornment::GetIsActive). The frame calls Draw on the parcel list (ITextParcelList on kFrameListBoss), which causes the wax line (IWaxLineShape on kWaxLineBoss)
and wax run (IWaxRunText on kWaxTextRunBoss) Draw methods to be called. During the
drawing of the wax run, the global adornments built up previously (Draw from IParcelShape
on kFrameItemBoss) have their Draw method called. Compare Figure 159 with Figure 158.
The set of global adornments are global to the document (maintained by the service registry),
whereas local adornments are local to a specific wax run.
Text Fundamentals
363
Text Fundamentals
Text composition
Text composition
This section describes the process of creating the final rendered text. It includes information
for software developers who want to develop custom composition engines.
Text composition is the process of converting text content into its composed representation,
the wax. The text-composition subsystem flows text content from a story thread maintained by
a story into a layout represented by parcels in a parcel list. Story threads are described in “Story
threads” on page 320; parcels and the wax, in “Text presentation” on page 331. Figure 160 is an
overview of text composition.
FIGURE 160
Text composition
Text layout
Frames/
parcels
The wax
Wax lines
Text composition
Fonts
Glyphs
Story thread
characters and text attributes
The text model
Figure 161 shows a sample text frame with no format changes. The figure shows the objects
representing the story, the frame, and the composed text; it does not show the details of how
the text actually gets composed.
364
Text Fundamentals
Text composition
FIGURE 161
Sample text frame with no formatting changes
kTextStoryBoss
ITextModel
ITextStoryThread
kSplineItemBoss
IHierarchy
kFrameListBoss
IFrameList
IParcelList
ITextParcelList
ITextParcelListData
IWaxStrand
kWaxLineBoss
kWaxLineBoss
IWaxLine
kWaxLineBoss
IWaxLine
IWaxLine
kWaxTextRunBoss
kWaxTextRunBoss
kWaxTextRunBoss
IWaxRun
IWaxRun
kMultiColumnItemBoss
IHierarchy
kFrameItemBoss
IHierarchy
IParcel
ITextTiler
ITextGrid
ITextFrameColumn
Text composition creates the wax lines (kWaxLineBoss) and their associated wax runs (kWaxTextRunBoss) using the primary story thread’s content and layout as input.
Wax lines are added to the wax strand (IWaxStrand). Text composition also maintains the
range of characters displayed in each parcel/frame (see “Span” on page 342).
Text Fundamentals
365
Text Fundamentals
Text composition
Phases of text composition
There are two distinct phases of text composition:
z
Damage refers to changes that invalidate the wax for a range of composed text. Inserting
text and modifying frame size are examples of changes that cause damage. This is described
further in “Damage” on page 367.
z
Recomposition is the process that repairs the damage and updates the wax to reflect the
change. This is described further in “Recomposition” on page 369.
Text composition toggles the wax between the states shown in Figure 162. The flow of damage
and recomposition is shown in Figure 163.
FIGURE 162
The wax state
Start
Recomposition is completed.
Damaged.
Do recomposition.
Not damaged.
The text model or text layout is updated.
366
Text Fundamentals
Text composition
FIGURE 163
Plug-in
Damage and recomposition, the phases of text composition
Update
The text model
MarkDamage
Characters
+
text attributes
Update
Recompose
MarkDamage
Fonts
Glyphs
Text
composition
Text layout
Parcels/
frames
MarkDamage
Wax lines
The wax
DetectDamage
Recompose
Background
composition
Figure 163 shows a plug-in can update the text model or text layout, causing the wax (the composed representation of the text) to be damaged.
The text-composition subsystem exposes interfaces that mark the damaged story, frame parcels, and wax. The damage-recording flow is represented by dotted lines in Figure 163. The
text-composition subsystem also exposes interfaces that allow text to be recomposed. In the
figure, the recomposition flow is represented by solid lines.
Background composition (see “Background composition” on page 371) is a separate process
that drives text composition. Background composition runs as an idle task when the application has no higher-priority tasks.
Consider the example in which the plug-in shown in Figure 163 is the text editor. In response
to typing, the plug-in processes a command to insert the typed characters into the text model.
This command calls text-composition interfaces that record the damage caused. Background
composition detects the damage and recomposes the affected text.
Damage
Damage is defined as the changes that invalidate the wax for a range of composed text. Damage
is recorded by marking the affected stories, parcels, frames, and wax. The damage indicates
where recomposition is required and the type of damage that has to be repaired.
The change counter on the frame list is incremented whenever something happens that causes
damage. No notification is broadcast when the change counter is incremented, so the change
Text Fundamentals
367
Text Fundamentals
Text composition
cannot be caught immediately; however IFrameList::GetChangeCounter returns a counter that
is incremented (and eventually rolls over) each time any parcel in any frame in the frame list
goes from undamaged to damaged, indicating some text in the frame list was damaged and
must be recomposed.
Damage drives and optimizes the recomposition process. For example, recomposition can
shuffle existing wax lines between frames to repair damage and avoid recomposing the text
from scratch. Based on the action that caused the damage, there are the following categories of
damage:
z
Insert damage occurs when text is inserted.
z
Delete damage occurs when a range of text is deleted.
z
Change damage occurs when a range of text is replaced or the text attributes that apply to a
range of text are modified.
z
Resize damage occurs when a frame is resized.
z
Rect damage occurs when text wrap or text inset is applied.
z
Move damage occurs when a frame is moved.
z
Destroyed damage occurs when some combination of the above types of damage occurs.
When this happens, the existing wax for the affected range must be recomposed.
z
Keeps damage occurs when a line must be pushed to the start of a new frame to eliminate
orphans and widows. Orphans and widows are words or single lines of text that become
separated from the other lines in a paragraph (see Figure 164). An orphan occurs when the
first line of a paragraph becomes separated from the rest of the paragraph body. A widow
occurs when the last line of a paragraph becomes separated from the rest of the paragraph
body.
FIGURE 164
Orphan (left) and widow (right)
Check for damage before relying on the information stored in the wax or the frame spans (see
“Span” on page 342). If your plug-in detects damage, it can either stop with a suitable error or
recompose the text.
368
Text Fundamentals
Text composition
Damage API
To detect damage, use the interface methods shown in Table 93 to detect damage. The damage
bookkeeping methods used to record the damage are not called directly by third-party plugins; the text subsystem calls these methods as necessary in response to changes that affect text.
TABLE 93 Damage-inquiry interfaces and methods
Interface
Method
IStoryList
CountDamagedStories, GetLastDamagedStory,
GetNthDamagedTextModelUID
IFrameList
GetFirstDamagedFrameIndex
IFrameDamageRecorder
GetCompState
IWaxLine
IsContentDamaged, IsDamaged, IsDestroyed,
IsGeometryDamaged, IsKeepsDamaged
ITextParcelListData
GetFirstDamagedParcel, GetParcelIsDamaged
ITextParcelList
GetFirstDamagedParcel, GetIsDamaged
Recomposition
Recomposition is the process of converting text from a sequence of characters and attributes
into the wax (fully formatted paragraphs ready for display). The components of recomposition
are as follows:
z
Interfaces that control text composition.
z
The wax strand that manages the wax for the story.
z
Paragraph composers that create the wax for a line or paragraph.
Recomposition API
To request recomposition, use the interface methods shown in Table 94.
TABLE 94 Recomposition interfaces and methods
Interface
Method
Description
IFrameList
QueryFramesContaining
Helper method that causes recomposition.
IFrameListComposer
RecomposeThruNthFrame,
RecomposeThruTextIndex
Controls recomposition of frames in the frame
list (kFrameListBoss). The methods delegate to
the parcel list composer
(ITextParcelListComposer).
IFrameComposer
RecomposeThruThisFrame
Controls recomposition of a frame. The
methods delegate to the wax strand
(IWaxStrand), which does the real work.
Text Fundamentals
369
Text Fundamentals
Text composition
Interface
Method
Description
ITextParcelListComposer
RecomposeThruNthParcel,
RecomposeThruTextIndex
Controls recomposition of parcels in a parcel
list. The methods delegate to the wax strand
(IWaxStrand), which does the real work.
IGlobalRecompose
RecomposeAllStories,
SelectiveRecompose,
ForceRecompositionToComplete
Provides methods that can be used to force all
stories to recompose. This interface marks
damage that forces recomposition to recompose
the damaged element, even though they were
not actually changed.
The interfaces and methods in Table 95 are not normally called by third-party plug-ins but play
a central role in the control of recomposition.
TABLE 95 Recomposition interfaces and key methods
Interface
Method
See
IFrameComposer
RecomposeThisFrame
“Wax strand” on page 370
IRecomposedFrames
BroadcastRecomposition
Complete
“Recomposition notification” on page 372
IParagraphComposer
Recompose,
RebuildLineToFit
“Paragraph composers” on page 371
Wax strand
The wax strand (IWaxStrand) manages the wax for a story and is responsible for updating the
wax to reflect changes in the text model or text layout. The wax strand controls the existing wax
and determines when and where a paragraph composer needs to be used to create new wax.
The wax strand also manages the overall damage-recording process.
The main text-recomposition methods are IFrameComposer::RecomposeThisFrame. The
recomposition methods on interfaces, such as IFrameList and IFrameListComposer, delegate
the actual work to these IWaxStrand methods.
NOTE:
These methods are not called directly by third-party plug-ins.
The lifecycle of the wax is as follows:
1. The wax strand is informed a range in the text model was changed. In response, the wax
strand locates the wax lines that represent this range and marks them damaged.
2. On the next recomposition pass, the wax strand finds the first damaged wax line in the
story and asks a paragraph composer to recompose it.
3. The paragraph composer creates new wax lines and applies them to the wax strand (RecomposeHelper::ApplyComposedLine). RecomposeHelper is a helper class within IParagraphComposer, so any existing wax lines that represent the same range in the text model are
destroyed.
370
Text Fundamentals
Text composition
4. The wax strand repeats this process until it finds no more damaged wax lines or reaches the
end of the frame.
Paragraph composers
A paragraph composer (IParagraphComposer) takes up to a paragraph of text and arranges it
into lines to fit a layout. The paragraph composer creates the wax that represents the composed
text. Plug-ins can introduce new paragraph composers. For more information, see “Implementation notes for paragraph composers” on page 372.
The text-composition architecture is designed for paragraph-based composition. The most significant implication of this design is that any change to text or attributes results in at least one
paragraph’s composition information being re-evaluated. For example, inserting one character
into a paragraph potentially can result in changes to any or all the line breaks in the paragraph,
including those preceding the inserted character.
Shuffling
Avoiding recomposition of text and simply moving existing wax lines up or down is called shuffling. The performance improvement gained by shuffling is significant.
The wax strand can determine if a wax line was damaged because of something that occurred
around the wax line and not because the range of text it represents changed, in which case the
wax strand knows that the net result of recomposing the text is simply to move the wax line up
or down. The content of the wax line after recomposition is identical to its current content.
For example, if you select and delete an entire paragraph of text, the next paragraph has not
changed and does not need to be recomposed. By pulling the following paragraph up to the
position occupied by the deleted paragraph, the wax strand avoids the cost of recomposition.
When shuffling, the wax strand implements some of the behaviors of a paragraph composer,
like dealing with space before and after paragraph attributes.
Vertical justification
Vertical justification moves wax lines up and down within the container that displays them,
according to rules. For example, the lines can be justified to the top, center, or bottom. They
also can be justified to fill the available space. The wax strand is responsible for vertical justification, and this mechanism is not extensible by third-party plug-ins.
Background composition
Background composition looks for damage, then fixes it by calling for recomposition. Background composition runs in the background as an idle task (IID_ICOMPOSITIONTHREAD
on kSessionBoss) and recomposes text. Each time background composition is called, it performs the following actions in priority order:
Text Fundamentals
371
Text Fundamentals
Text composition
1. Recomposes visible frames of damaged stories in the frontmost document.
2. Recomposes other damaged stories in the frontmost document.
3. Recomposes damaged stories in other open documents.
The actions are performed until the time slice allocated for background text composition
expires. Background composition requests recomposition by calling the RecomposeUntil
method on IFrameListComposer.
Recomposition transactional model
Plug-ins use commands to update the input data that drives text composition (the text model
and text layout). The commands cause damage to occur. Recomposition subsequently recomposes the text and generates the wax to reflect the changes. Recomposition does not execute
within the scope of a command; instead, the wax strand begins and ends a database transaction
around the recomposition process. This means recomposition cannot directly be undone; however, the commands used to update the input data are undoable. Undo causes damage, and
recomposition repairs it to ensure the wax (the text displayed) reflects the state of the text
model and text layout.
Recomposition notification
The IRecomposedFrames interface on the frame list (kFrameListBoss) keeps track of which
frames are recomposed, so text frame observers can be notified when recomposition completes. When the BroadcastRecompositionComplete method is called, each affected column
(kFrameItemBoss) receives notification that it was recomposed. Observers attach to the
kFrameItemBoss in which they are interested and then watch for the Update method to be
called with theChange == kRecomposeBoss and protocol==IID_IFRAMECOMPOSER.
Implementation notes for paragraph composers
This section provides more in-depth background for software developers developing a custom
paragraph composer for the application.
Paragraph composers
A paragraph composer (IParagraphComposer) takes up to a paragraph of text, arranges it into
lines to fit a layout, and creates the wax that represents the text it composed. Plug-ins can introduce new paragraph composers.
The user selects the composer to be used for a paragraph from the menu on the Paragraph
panel (see Figure 165) or by using the Paragraph Styles palette. Selecting a paragraph composer
sets the kTextAttrComposerBoss paragraph attribute to indicate which paragraph composer to
use.
372
Text Fundamentals
Text composition
FIGURE 165
Setting the paragraph composer
Hyphenation and justification (HnJ)
The term “hyphenation and justification” (‘HnJ’) often is used in computer systems to refer to
breaking of text into lines and spacing of text so it aligns with the margins. This term is not
used in this application’s API or documentation; however, the role performed by an HnJ system
is similar to the role of a paragraph composer. Hyphenation is intimately associated with line
breaking. In this application, the role is split out and implemented by a hyphenation service
(IHyphenationService). A paragraph composer that needs to hyphenate a word uses a hyphenation service.
A paragraph composer’s environment
Paragraph composers are called by the text-composition subsystem. When text is to be recomposed, the wax strand examines the paragraph attribute kTextAttrComposerBoss that specifies
the paragraph composer to be used, locates it through the service registry, and calls the IParagraphComposer::Recompose method. This call gives the paragraph composer a context stored
in the helper class in which to work.
virtual bool16 Recompose(RecomposeHelper* helper)
The RecomposeHelper class is a wrapper of the story being recomposed. It stores information
like the story’s compose scanner, tiler, starting index, text span, and position. The RecomposeHelper class definition is in IParagraphComposer.h.
Figure 166 shows an example of a paragraph composer called to compose text.
Text Fundamentals
373
Text Fundamentals
Text composition
FIGURE 166
Paragraph-composer environment
-100,-15
100,15
kFrameListBoss
IWaxStrand
helper->ApplyComposedLine
Wax lines and runs
Other parcel list boss
kTextCellContentBoss
kTextStoryBoss
kFrameListBoss
helper->GetComposeScanner
Recompose
Characters and effective attribute values
characters=”Aliens “
fontName=Times-Roman,
pointSize=8 point,
leading= 10 points
IParagraphComposer
Tiles
firstBaselineOffset=leading
tile [0] = [-100, -15, 100, -5]
Paragraph composer boss class
The following steps occur:
1. The paragraph composer composes the text and creates at least one wax line in the helper.
2. The paragraph composer accesses the character-code and text-attribute data for the story,
using the IComposeScanner interface acquired from helper->GetComposeScanner.
3. The paragraph composer determines the areas in a line where glyphs can flow, using IParagraphComposer::Tiler.
4. The wax lines are applied to the IWaxStrand interface by calling helper->ApplyComposedLine.
NOTE:
Both yPosition and tiles are specified in the inner bounds (coordinate system) in
RecomposeHelper.
The scanner and drawing style
The scanner (IComposeScanner) provides access to both character and formatting (text
attributes) information from the text model, as shown in Figure 167. The drawing style (interface IDrawingStyle) represents the effective value of text attributes at a given index in the text
374
Text Fundamentals
Text composition
model. The drawing style cached by interfaces on kComposeStyleBoss considerably simplifies
accessing text-attribute values. These interfaces remove the need to resolve text-attribute values
from their representation in the text model and the text styles it references. A drawing style
applies to at least one character. The range of characters it applies to is a run. For more information on styles and overrides, see “Text formatting” on page 324.
FIGURE 167
The scanner and drawing style
kTextDataStrandBoss
kParaAttrStrandBoss
kCharAttrStrandBoss
kStyleBoss
kTextStoryBoss
kStyleBoss
Text attributes
IComposeScanner
kComposeStyleBoss
IDrawingStyle
ICompositionStyle
IHyphenationStyle
IJustificationStyle
Drawing style (effective attribute values)
Characters
Paragraph composer boss class
IParagraphComposer
Fonts and glyphs
Fonts contain the glyphs that display characters for a typeface at a particular size. Paragraph
composers use the drawing style (IDrawingStyle) to find the font to be used. The font APIs
provide the metrics required to compose the text into lines composed of a run of glyphs for
each stylistic run of text.
A font represents a typeface with a given size and style. A typeface, like Times, is the letters,
numbers, and symbols that make up a design of type. A glyph is a shape in a font that is used to
represent a character code on screen or paper. The most common example of a glyph is a letter,
but the symbols and shapes in a font like ITC Zapf Dingbats also are glyphs. For more information, see Figure 168 and “Fonts” on page 389.
Text Fundamentals
375
Text Fundamentals
Text composition
FIGURE 168
Accessing a font
kComposeStyleBoss
QueryFont
IPMFont
IDrawingStyle
QueryFontInstance
IFontInstance
Tiles
Tiles (IParagraphComposer::Tiler) represent the areas on a line into which text can flow. Normally, there is only one tile on a line; however, text wrap may cause intrusions that break up the
places in the line where text can go. The tiler takes care of this, as well as accounting for the
effect of irregularly shaped text containers.
Intrusions are the counterparts of tiles. Intrusions are not the inverse of tiles: they do not specify the areas in the line where text cannot be flowed. From the perspective of ITextTiler, an
intrusion is a horizontal band within a frame, in which text flow within the bounding box of
the frame may be interrupted. Intrusions can be used to optimize recomposition. An intrusion
is a flag that indicates that text cannot be flowed into the entire width of a line. Intrusions are
caused by text wrap and nonrectangular frames.
The tiles returned for a line depend on the y position where they are requested, their depth,
and their minimum width. The intrusions and tiles for some sample text frames are shown in
Figure 169 through Figure 173.
FIGURE 169
Frame
FIGURE 170
Frame
376
Text frame with no intrusions
Intrusions(none)
Tiles
Effect of image frame with text wrap
Intrusions
Tiles
Text Fundamentals
Text composition
FIGURE 171
Effect of two images with text wrap
Frame
FIGURE 172
Tiles
Intrusions
Tiles
Effect of text inset
Frame
FIGURE 173
Intrusions
Effect of nonrectangular frame
Frame
Intrusions
Tiles
First-baseline offset
The first-baseline offset setting controls the distance between the top of a parcel or frame and
the baseline of the first line of text it displays. Some sample settings are shown in Figure 174.
Text Fundamentals
377
Text Fundamentals
Text composition
FIGURE 174
First-baseline offset examples
The setting controls the distance between
the top inset of the frame and the baseline
of the first line.
Use the largest leading on the line.
Leading
Ascent
Cap height
Use the largest ascent on the line. This
makes the top of the tallest glyph fall
below the top inset of the frame.
Use the largest Cap Height on the line. This
makes the top of upper case letters touch
the top inset of the frame.
Control characters
A paragraph composer can apply an appropriate behavior to correctly implement the semantic
intent of many control characters. For a complete list of control-character codes, see TextChar.h in the API reference documentation.
Inline frames
Inline frames (kInlineBoss) allow frames to be embedded in the text flow. An inline frame
behaves as if it were one text character and moves along with the text flow when the text is
recomposed. The kTextChar_ObjectReplacementCharacter or kTextChar_Inline character is
inserted into the text data strand to mark the position of the inline frame in the text. The inline
frames are owned by the owned item strand (kOwnedItemStrandBoss) in the text model. Paragraph composers obtain the geometry of the frame associated with the kInlineBoss by using
ITextModel to access this strand directly. Inline frames are represented in the wax by their own
type of wax run, kWaxILGRunBoss.
378
Text Fundamentals
Text composition
Table frames
Table frames (kTableFrameBoss) are owned items anchored on a kTextChar_Table character
embedded in the text flow. Paragraph composers can and do compose text displayed in table
cells.; however, they never compose (create wax for) the character codes related to tables
(kTextChar_Table and kTextChar_TableContinued). When a paragraph composer encounters
either of these characters, it should create wax for any buffered text and return control to its
caller.
Creating the wax
During the process of building lines, a paragraph composer normally uses an intermediate representation of the glyphs, so it can evaluate potential line-break points and adjust glyph widths.
The specific arrangement used depends on the overall feature set of the paragraph composer
and is beyond the scope of this document. Once the line-break decisions are made, the paragraph composer must create wax lines and runs.
For details of the mechanics of the creation of wax lines and runs, see “The wax” on page 354.
Once created by a call to RecomposeHelper::QueryNewWaxLine, new wax lines are added to
the wax strand for a story with the following call:
RecomposeHelper::ApplyComposedLine(IWaxLine* newLine, int32 newTextSpan)
Figure 175 shows an example of the wax generated by a paragraph composer.
FIGURE 175
Wax for a single line with format changes
-100,-15
100,15
WaxLine
X=-100
Y=-5
Width=60.33
Origin=0
Span=17
WaxRun
X=0
Y=0
Width=22.89
Origin=0
Span=7
FontName=Times-Roman
FontMatrix=[8 0 0 8 0 0]
Text Fundamentals
WaxRun
X=22.89
Y=0
Width=16.12
Origin=7
Span=3
FontName=Times-Bold
FontMatrix=[10 0 0 10 0 0]
WaxRun
X=39.01
Y=0
Width=21.32
Origin=10
Span=7
FontName=Times-Roman
FontMatrix=[8 0 0 8 0 0]
UnderlineMode=kUnderlineSingle
379
Text Fundamentals
Text composition
Adobe paragraph composers
Figure 176 shows the paragraph composers provided by the application under the Roman feature set. Figure 177 shows the paragraph composers provided under the Japanese feature set.
Paragraph composers are implemented as service providers (IK2ServiceProvider) and can be
located using the service registry. The ServiceID used to identify paragraph composers is kTextEngineService.
FIGURE 176
Roman paragraph composers
kParagraphComposerBoss
IParagraphComposer
IK2ServiceProvider
GetServiceID=kTextEngineService
GetName="HL Composer"
FIGURE 177
kSingleLineComposerBoss
IParagraphComposer
IK2ServiceProvider
GetServiceID=kTextEngineService
GetName="HL Single"
Japanese paragraph composers
kJParagraphComposerBoss
IParagraphComposer
IK2ServiceProvider
GetServiceID=kTextEngineService
GetName="HL Composer J"
kJSingleComposerBoss
IParagraphComposer
IK2ServiceProvider
GetServiceID=kTextEngineService
GetName="HL Single J"
Key APIs
Class diagram
Paragraph composers are boss classes that implement IParagraphComposer, the interface
called by the application to compose text. To be recognized as a text composer, the boss class
also must implement a text-engine service (IK2ServiceProvider) with a ServiceID of kTextEngineService. The service also must provide a name (a translatable string) for the paragraph
composer, which is displayed in the user interface. As output, composers generate wax lines
(IWaxLine) and associated wax runs (IWaxRun) for up to one paragraph of text at a time.
Figure 178 shows the class diagram of a custom paragraph composer.
380
Text Fundamentals
Text composition
FIGURE 178
Custom paragraph composer
kYourParagraphComposerBoss
IParagraphComposer
IK2ServiceProvider
GetServiceID=kTextEngineService
GetName=”Your Composer"
IParagraphComposer
There are two distinct phases to text composition, and IParagraphComposer reflects these
roles:
z
Positioning the line and, if necessary, deciding where text in the line should break. This
information is represented persistently in the wax line. Features that affect line-break decisions belong here. This is the role of the Recompose method.
z
Positioning glyphs that represent the characters in the line and, if necessary, adjusting their
widths. This information is represented by a set of wax runs associated with the wax line.
Wax runs are not persistent and must be rebuilt for display when a wax line is read from
disk. Features that affect glyph position (paragraph alignment, justification, letter spacing,
word spacing, and kerning) belong here. This is the role of the RebuildLineToFit method,
which regenerates composed text data from the minimal data stored on disk.
Recompose is called when text needs to be fully recomposed. Each time this method is called, it
should compose at least one line and at most one paragraph of text into the passed-in helper.
RecomposeHelper actually does all the work, by creating wax starting from the character in the
story indicated by RecomposeHelper::GetStartingTextIndex. The top of the area text flows into
is given by RecomposeHelper::GetStartingYPosition, and composition is done using the supplied scanner and tiler. The wax lines generated as a result are applied to the given wax strand.
NOTE:
The paragraph composer is expected to create at least one wax line each time
Recompose is called, even when overset text is what is being composed.
RebuildLineToFit is called to regenerate the wax runs for a wax line from the minimal data
stored on the disk. No line breaking or line positioning needs to be done here.
IComposeScanner
IComposeScanner is an interface aggregated on kTextStoryBoss. It provides methods that help
access character and text-attribute information from the text model. It is the primary iterator
for reading a story.
Example 23 shows QueryDataAt, which gets a chunk of data from the text model starting from
a given TextIndex. QueryDataAt is a primary method text composition uses to iterate through
the text model. QueryDataAt returns a TextIterator, which has its own reference count on the
chunk of data in the model, and thus makes the TextModel lifecycle independent of the com-
Text Fundamentals
381
Text Fundamentals
Text composition
pose scanner cache. Optionally, QueryDataAt returns the drawing style and number of characters in the returned iterator. We recommend you always ask for the returned number of
characters.
EXAMPLE 23 IComposeScanner::QueryDataAt
virtual TextIterator QueryDataAt(TextIndex position, IDrawingStyle** newstyle,
int32* numChars) = 0;
You also can acquire the drawing style by one of the methods listed in Example 24. For more
information on the methods in IComposeScanner, see the API reference documentation.
EXAMPLE 24 Acquiring IDrawingStyle from IComposeScanner
virtual IDrawingStyle* GetCompleteStyleAt
(TextIndex position, int32* lenleft = nil) = 0;
virtual IDrawingStyle* GetParagraphStyleAt
(TextIndex position, int32* lenleft = nil, TextIndex* paragraphStart = nil) = 0;
QueryAttributeAt allows callers to query the effective value of one or more particular text
attributes at a given position. For more details, see the API reference documentation.
IDrawingStyle, ICompositionStyle, IHyphenationStyle, and IJustificationStyle
IDrawingStyle provides access to text-attribute values like font, point size, and leading. If the
text property you want is not in IDrawingStyle, you will find it in one of the other style interfaces on kComposeStyleBoss.
For details, see the API reference documentation for kComposeStyleBoss, IDrawingStyle,
ICompositionStyle, IHyphenationStyle, and IJustificationStyle.
IPMFont and IFontInstance
IPMFont encapsulates the CoolType font API. It represents a typeface and can generate
instances (IFontInstance) of it at various point sizes. IPMFont is not based on IPMUnknown,
but IPMFont is reference counted and can be used with InterfacePtr. For details, see the API
reference documentation.
IFontInstance encapsulates the CoolType font-instance API. IFontInstance represents an
instance of a typeface at a given size. IFontInstance is not based on IPMUnknown, but IFontInstance is reference counted and can be used with InterfacePtr.
IFontInstance controls the mapping from a character code (UTF32TextChar) into the identifier
of its corresponding glyph (GlyphID). For details, see the API reference documentation.
IParagraphComposer::Tiler
IParagraphComposer::Tiler manages the tiles for a parcel list and is used by paragraph composers to determine the areas on a line into which text can flow (see “Tiles” on page 376). It
controls the content area bounds of each line.
The GetTiles method is used to get the tiles. If GetTiles cannot find large enough tiles in the
current parcel that meet all the requirements, it returns kFalse and should be called again to see
382
Text Fundamentals
Text composition
if the next parcel in the list can meet the request. If no parcel can meet the request, the tiler
returns tiles into which overset text can flow.
GetTiles is a complex call to set up. To fully understand how to use it, see the API reference
documentation and the text-composer samples (SingleLineComposer) provided in the SDK.
The “TOP” abbreviation used in many of the parameter names is short for “top of parcel,”
because when a line falls at the top of a parcel, special rules apply that govern its height (see
“First-baseline offset” on page 377).
For Roman composers, a tile must be wide enough to receive one glyph plus any left and right
line indents that are active. A heuristic often is used that approximates the minimum glyph
width to be the same as the leading. The minimum tile height depends on the metric being
used to compose the line. Normally this is leading, but the first line in a parcel is a special case
in which the metric used may vary depending on the first-baseline offset (ascent, cap height,
and leading) associated with the parcel. The pYOffset value indicates the top of the area for text
to flow into. The pYOffset value actually is a position rather than an offset from the top of the
parcel. The pYOffset value initially is the yPosition stored in RecomposeHelper when the
IParagraphComposer::Recompose method is called. For more details, see the API reference
documentation.
Scenarios
This section illustrates how text gets recomposed. The scenarios presented here demonstrate
basic situations you may come across in developing a paragraph composer. To reduce complexity for these examples, assume the paragraph composer being called generates only one line for
each call (a single-line composer).
Simple text composition
Figure 179 shows a composed text frame 200 points wide and 30 points deep, with its baseline
grid shown. The text is 8-point Times Regular with 10-point leading, and the first-baseline offset for the frame is set to leading.
FIGURE 179
Simple text composition
left = -100, top = -15
Baseline
-5
5
15
right = 100, bottom = 15
When text of the same appearance is flowed into the frame, the distance between the baselines
of succeeding lines of text is set to the leading.
Text Fundamentals
383
Text Fundamentals
Text composition
The wax strand examines the paragraph attribute that specifies which paragraph composer to
use and locates it through the service registry. The paragraph composer is then asked to recompose the story from the beginning, causing text to flow into the frame starting from the top by
calling IParagraphComposer::Recompose(helper).
The yPosition stored in the helper indicates the top of the area into which glyphs can flow. This
is the baseline of the preceding line or the top of the parcel.
The composer examines the drawing style and its associated font and calculates the depth and
width of the tiles it needs. The example uses 8-point text with 10-point leading and no line
indents, so the minimum height and width is 10 points. The tiler is then requested for tiles of
the required depth and minimum width at the given yPosition, and it returns the tile shown in
Figure 180.
FIGURE 180
yPosition
Tiles for the first line
Tile for line 1
Baseline
-15
-5
The following sequence of actions then occurs:
1. The paragraph composer scans the text using IComposeScanner. Text flows into the tile
until the tile is full or an end-of-line character is encountered.
2. The composer chooses where the line is to be broken and creates a wax line for the range of
text the line will display.
3. The wax line is applied to the IWaxStrand and, as a side effect, the composer’s IParagraphComposer::RebuildLineToFit method is called to generate wax runs for the line.
4. The IParagraphComposer::Recompose method is finished with its work and returns control
to its caller.
5. The wax strand successively asks the paragraph composer to recompose the second and
third lines with yPositions of -5 and 5, respectively, until the text is fully composed.
The result is shown in Figure 181. For more details, see the SDK sample code SingleLineCompose.
384
Text Fundamentals
Text composition
FIGURE 181
Tiles for the second and third lines
Tile for line 2
yPosition
Baseline
-5
5
Tile for line 3
5
15
Composed frame with baseline grid displayed
-15
-5
-5
5
5
15
15
Change in text height on line
As a line is composed, attributes that control the height of the line, like point size and leading,
may change. The baseline of the text must be set to reflect the largest value of the metric, such
as leading or ascent, that is being used to compose the text. To achieve this, the composer asks
the tiler for tiles deep enough to take text of a given height. If an increase in text height is
encountered as composition proceeds along a line, the tiler needs to be asked for deeper tiles to
accommodate the text. It is the composer’s responsibility to ensure the parcel can receive the
deeper text.
For example, if the point size for the word “Ask” is changed to 16 points and its leading to 20
points, the text suffers change damage. The resulting sequence of recomposition is shown in
Figure 182.
Text Fundamentals
385
Text Fundamentals
Text composition
FIGURE 182
Changing point size and leading of selected text
The text suffers change damage and the paragraph composer is called.
yPosition
Initial tile for line 2
Baseline
-5
5
Leading increased. Get deeper tile.
-5
15
Recomposed frame
-15
-5
-5
5
5
15
15
Initially, the composer encounters 8-point text on 10-point leading, and it requests tiles of an
appropriate depth. When the composer hits the style change to 16-point text and 20-point
leading, it goes back to the tiler and gets a deeper tile. When the style changes back to the
smaller point size, the composer already has a tile deep enough to take this text, so it fills the
tile with glyphs and breaks the line. Note how the baseline is set to reflect the largest leading
value encountered in the line.
386
Text Fundamentals
Text composition
Text wrap
Depending on the yPosition and depth of the tiles needed, the text-wrap settings on other page
items or a change to the shape of the frame may become significant during recomposition. To
illustrate this, in Figure 183, a graphic with text wrap is added to the simple text-frame arrangement.
FIGURE 183
Simple text wrap
Frame
Tiles
Text frame with 8-point Times regular text, 10-point leading
Add graphic frame with text wrap set to wrap tightly around bounding box.
Figure 183 shows the text frame and the tiles the 8-point text flows into with 10-point leading.
Note that the third line in the frame is affected by the text wrap and flows around the bounding
box of the graphic.
The following sequence illustrates what happens when the point size for “Ask” is changed to 12
points and its leading is changed to 14 points. The initial case is shown in Figure 184.
FIGURE 184
Changing the selection to 12-point text, 14-point leading
The wax strand picks up the change damage and asks the paragraph composer to recompose
the first line of the second paragraph. The baseline of the first line of text in the frame, -5
points, is at the yPosition stored in the RecomposeHelper.
The paragraph composer gets tiles deep enough to take the initial leading value (10 points) it
encounters on the line, as shown in Figure 185.
Text Fundamentals
387
Text Fundamentals
Text composition
FIGURE 185
Composer begins to recompose
yPosition
Baseline
-5
5
The paragraph composer then encounters the run of text with the changed point size and leading. Because the text is deeper than the tiles the paragraph composer obtained, the composer
goes back to the tiler and asks for a new set of tiles to receive the deeper text. This is shown in
Figure 186.
FIGURE 186
Composer asks for deeper tiles
yPosition
Baseline
-5
9
Because of the request for deeper tiles and the text-wrap setting on the graphic, the tiler returns
the two tiles shown in Figure 187, causing text to tightly wrap to the bounding box of the
graphic.
FIGURE 187
yPosition
Deeper tiles
Baseline
-5
9
The paragraph composer flows glyphs into the deeper tiles, generating a wax line with four wax
runs: one for the tile to the left of the graphic and three in the tile to the right of the graphic.
After this change, all text for the story cannot be displayed in the frame, so the story is overset,
as shown in Figure 188.
388
Text Fundamentals
Fonts
FIGURE 188
Fully recomposed frame
yPosition
Baseline
-5
-5
9
9
It is important to note that the tiles returned by the tiler depend on the yPosition within the
frame for which they are requested and the depth asked for by the caller.
Frame too narrow or not deep enough
IParagraphComposer::Recompose must return a wax line each time it is called; however, it is
possible a frame is too narrow to accommodate text, when either the indent settings for a paragraph are large or the text has a large point size. The result is overset text. Call the tiler (IParagraphComposer::Tiler’s GetTiles()) repeatedly until it returns kTrue. The tiler iterates over all
parcels; if none can accommodate the text, the tiler returns a tile into which overset text can
flow.
Fonts
This section explains how fonts fit into the InDesign text architecture. Fonts provide the basic
information required to render glyphs. A font is identified by a font name (e.g., “Minion Pro”)
and a face or style name (e.g., “Bold” or “Heavy”). A rendered font also has a size; for example,
this text is 11 pt (points). The font, along with the point size, is an instance of the font.
All rendered text in the application has an associated font. The operating system provides a set
of font services, though fonts unique to the application can be maintained in the application
fonts folder.
An application installation has a set of fonts available (from either the operating system or the
application fonts folder). The font manager provides these fonts to the user while the document
is being edited. There are times when the font manager gets a request for a font that is not available; for example, the font was removed from the system or the document was produced on a
system with a different set of available fonts. The font manager resolves this situation by replacing the font with one that is available. The font manager warns the user when such a replacement occurs.
Fonts have rights, which determine what can be done with a particular font. For example, a font
can indicate that it should not be embedded in a PDF document or it should not be printed.
The font subsystem provides access to these rights and implements the policies required to
adhere to these rights.
Text Fundamentals
389
Text Fundamentals
Fonts
Font management within the application relies on a core piece of Adobe technology, CoolType.
CoolType wraps the operating-system font services to provide a cross-platform abstraction and
provides further services to clients.
Font-subsystem architecture
The application wraps the CoolType representation of fonts, font groups, and font instances
with application-specific types. This is shown in Figure 189.
FIGURE 189
Accessing CoolType fonts in the application
kSessionBoss
IFontMgr
«interface»
IFontMgr
1
1
QueryFont
1
QueryFontGroup
*
*
IPMFont
*
IFontGroup
IFontInstance
GetCTFontDict
CoolType font sybsystem
CTFontDict
390
GetCTFontGroup
CTFontGroup
GetCTFontInst
CTFontInstance
InDesign wrappers for core
CoolType types. Interfaces do
not inherit from IPMUnknown.
QueryFontInstance
OS font subsystem
Text Fundamentals
Fonts
IFontMgr on kSession provides access to the CoolType font wrappers. IPMFont is the representation of a font, IFontGroup represents a group of related fonts, and IFontInstance represents a particular font at a particular point size.
Font manager
The font manager provides access to all font groups available to the application. Each font
group allows you to access each font in the group. The IFontMgr interface is aggregated on
kSessionBoss.
CoolType
CoolType is the core technology through which an application can access fonts. It automatically loads the fonts made available by the operating system. In addition, CoolType can load
fonts directly from the common Adobe fonts folder and the Fonts folder inside the application’s
folder.
The API provides the IFontGroup, IPMFont, and IFontInstance interfaces as shells through
which the application calls CoolType. These are not standard interfaces; that is, they do not
derive from IPMUnknown. You cannot create them using CreateObject or query them for
other interfaces; however, these interfaces are reference counted, and you can manage their lifetime using InterfacePtr. See Table 96 and Figure 190.
TABLE 96 Interfaces that use CoolType
Interface
Description
How to get the interface
Code snippet with example
of use
IFontGroup
Represents all font groups
available to the application
IFontMgr::QueryFontGroup,
FontGroupIteratorCallBack
SnpPerformFontGroupIterat
or
IPMFont
Provides extensive data
related to a particular font.
IPMFont provides font-name
mapping calls and access to
font metrics at a given point
size.
IFontMgr::QueryFont,
IFontMgr::QueryFontPlatform,
IFontFamily, text attributes
kTextAttrFontUIDBoss and
kTextAttrFontStyleBoss,
IStoryService, IWaxRenderData
SnpInspectFontMgr
IFontInstance
Represents an instance of a
font of a given point size. Of
interest if you need to write a
paragraph composer or map
character codes to GlyphIDs.
IFontMgr, IDrawingStyle,
IStoryService
SnpInsertGlyph
Text Fundamentals
391
Text Fundamentals
Fonts
FIGURE 190
Accessing a font
QueryFont
kComposeStyleBoss
IPMFont
IDrawingStyle
QueryFontInstance
IFontInstance
Fonts are used when composing text, as fonts define the characteristics of individual glyphs.
kComposeStyleBoss is used by the composition engine. The IDrawingStyle interface provides
access to the font (IPMFont) and particular instance (IFontInstance) being used
Fonts within the document
Within a document, fonts are represented by the font name (plus information on font style and
point size). When the actual font is required, the font manager resolves the name-to-font (IPMFont or IFontInstance) mapping.
The document-font manager
The document-font manager (IDocFontMgr) provides access to all fonts that persist in a session or document workspace. The fonts that exist in a document include a set of default fonts,
fonts used in visible stories, and fonts used for features that are modeled using the application
text subsystem but not intended to be viewed or printed as part of a story (for example, text
that exists in notes).
Fonts in the persistent document
Fonts are represented in the object model by kFontGroupBoss (signature interface IFontFamily). Each font used in a story results in an instance of this boss object. A font family along with
a font style identifies a particular font in the font subsystem. The font family represents the
name of the font (more correctly, the set of names that identify each font within a particular
group). This decoupling between font and name in the document allows the font manager to
resolve the font when required and take appropriate measures if the font is not available to the
application. The relationship between font manager and fonts is shown in Figure 191
392
Text Fundamentals
Fonts
FIGURE 191
How fonts persist in a document
IDocFontMgr exists on the session or document
workspace. The set of default fonts inherited by
new documents is maintained in IDocFontMg r
on the session workspace. The set of fonts used
by a document is maintained on the document
workspace.
«interface»
IDocFontMgr
Note that the fonts used by visible text in all
stories are ident ified by IDocFontMgr. Fonts
used in placed assets (EPS or PDF) and fonts used
in non-visible text in stories (such as with notes)
are not maint ained on IDocFontMgr.
1
GetNthFontGroupUID
m..*
kFontGroupBoss
IFontFamily
«interface»
«interface»
IFontFamily
+
IPMFont
QueryFace
FontName: PMString
1
1..*
A font persists as a name, contained within a persistent implementation of IFontFamily. Access
to these persistent objects is provided through the (session or document) workspace. IFontFamily::QueryFace requests the font from the font manager. If the returned IPMFont is nonnil, IPMFont::FontStatus can be used to determine the validity of the font.
A set of text within a story is associated with a particular font using standard text attributes.
The text attributes that refer to the font in which text appears are as follows:
z
kTextAttrFontUIDBoss — ITextAttrUID gives the UID of the font family.
z
kTextAttrFontStyleBoss — ITextAttrFont gives the name of the stylistic variant.
z
kTextAttrPointSizeBoss — ITextAttrRealNumber gives the point size used.
The font family, along with the style of the font, identifies a particular font (IPMFont). The
point size of the text is used to provide an instance of the font (IFontInstance).
Text attributes are introduced and described fully in “Text formatting” on page 324.
The used-font list
The used-font list (IUsedFontList) on a document (kDocBoss) details the fonts associated with
story text. The used-font list does not include the fonts used in features that are not intended to
be viewed or printed as part of a story (for example, text that exists in notes).
Text Fundamentals
393
Text Fundamentals
Fonts
Fonts in placed assets
Placed (EPS or PDF) assets support the IFontNames interface. This interface reports the fonts
used for the particular asset, including whether the asset depends on the presence of fonts on
the system. Placed assets are maintained in the native asset’s format (PDF or EPS), with the
IFontNames interface providing access to information within the asset.
Document font use
A utility interface (IDocumentFontUsage) on the document (kDocBoss) provides a facade over
the two interfaces introduced above (IUsedFontList and IFontNames).
Font naming
Internally, application code uses PostScript font names. For example, IPMFont::AppendFontName returns the PostScript font name. If you know the PostScript name of the font you want,
you can get the associated IPMFont using IFontMgr::QueryFont. Several other names are
available on IPMFont and IFontFamily. For example, if you use IFontMgr::QueryFont to
instantiate the PostScript font AGaramond-Italic, IPMFont gives you the names shown in the
table in Figure 192.
FIGURE 192
Font names
Name
394
Method
Example
PostScript font name
IPMFont::AppendFontName
AGaramond-Italic
Family name
IPMFont::AppendFamilyName
Adobe Garamond
Style name
IPMFont::AppendStyleName
Italic
IPMFont::AppendFullName
Adobe Garamond Italic
IPMFont::AppendFamilyNameNative
Adobe Garamond
IPMFont::AppendStyleNameNative
Italic
IPMFont::AppendFullNameNative
Adobe Garamond Italic
Text Fundamentals
Fonts
Notice the typeface name Adobe Garamond Pro in the panel in Figure 192 does not match the
family name in the table. This is because ITextUtils::GetDisplayFontNames supplies the platform-specific name in the Character panel. If you do not have the PostScript font name but
know the name of the typeface on a given platform (the typeface name in Windows or the family name in Mac OS), IFontMgr::QueryFontPlatform returns the corresponding IPMFont.
Font warnings
The IDocumentFontUsage interface (on kDocBoss) manages used-font and missing-font
information for text in stories of a document. The IDocumentFontUsage interface does not
detail fonts used by features that rely on the application text model but are not rendered as part
of the document (for example, notes).
The IMissingFontSignalData interface provides data sent with the missing-font signal (of
which responders of type kMissingFontSignalResponderService are notified).
The IFontUseWarningManager interface can be called to check a particular font and possibly
warn the user about restrictions. Examples are restrictions based on user preferences or
because a font is a bitmap.
Composite fonts and international-font issues
Some languages too many glyphs to be contained in a font; for example, Japanese contains
about 50,000. To accommodate this, font developers create fonts with subsets of glyphs. Also,
there is a desire within typography to replace some glyphs with others to achieve a certain
visual effect. And users might want to shift the position of Roman glyphs relative to the baseline or enlarge or reduce the size of glyphs. See Figure 193.
FIGURE 193
Japanese and Roman typography metrics
Japanese
Roman
Em-box top
Ascent
ICF box top
Em-box center
Baseline
ICF box bottom
Em-box bottom
Descent
The ICompositeFont interface creates a new font composed of base fonts and provides for size
and positioning adjustments. Font switching is based on Unicode values. The base font must be
a Chinese, Japanese, or Korean font. For examples of using ICompositeFont, see the SnpPerformCompFont code snippet. An application can get this interface using any of the following:
z
IStyleNameTable (from the active workspace or document workspace)
z
InterfacePtr<IStyleNameTable> compFontList(workspace, IID_ICOMPOSITEFONTLIST)
z
InterfacePtr<ICompositeFont> rootTable(docDatabase,
StyleUID(), IID_ICOMPOSITEFONT)
Text Fundamentals
compFontList->
GetRoot-
395
Text Fundamentals
Fonts
Use the ICompositeFontMgr interface to manage composite fonts. For examples of its use, see
the SDK code snippet SnpPerformCompFont.
The IFontMetricsTable interface obtains the correct tsume values (similar to side bearings on
glyphs) around either the top and bottom or left and right of glyphs from the document originator and caches them. Later users (who might have only bitmaps of the fonts used to create
the document) can use the tsume metrics saved in the document through this interface.
396
Tables
Concepts
Tables
This chapter describes concepts and architecture related to the table feature of InDesign.
Table attributes are boss classes with ITableAttrReport implementation; they are discussed in
“Table attributes” on page 407. Table and cell styles define a collection of table attributes that
are appropriate for tables or cells. They can be applied to individual tables and cells. See “Table
and cell styles” on page 409.
For information on how to work with table models and APIs, see the “Tables” chapter of Adobe
InDesign CS4 Solutions.
Concepts
Table structure
Tables in InDesign publications consist of rows, columns, and cells. Cells can span multiple
rows or columns, as in the HTML 4 table model, which evolved from the CALS SGML table
model.
There are helper classes in the InDesign API that are used to specify location, dimension of
cells, and areas within tables. The abstractions used to specify these are shown in Figure 194.
For this example, the resolution of the underlying grid is 4 (rows) by 3 (columns). Cells can
consist of merged elements on this underlying grid.
FIGURE 194
Table-structure example
GridAddress(0,0) GridSpan(1,3)
GridAddress(1,0) GridSpan(1,3)
GridAddress(2,0) GridSpan (1,2)
GridAddress(2,2)
GridSpan(1,1)
GridAddress (3,0) GridAddress(3,1) GridAddress(3,2)
GridSpan(1,1)
GridSpan(1,1)
GridSpan(1,1)
Whole table:
GridArea(0,0,4,3)
Tables
397
Tables
Concepts
Anchor cells versus grid elements
The top-left of a cell is its anchor.
A grid element is a unit on the underlying grid. Grid elements may or may not be anchor cells.
A cell can comprise multiple elements on the underlying grid, but there is at most one anchor
point per cell.
To access the text content of a cell, some API methods require a GridAddress that refers to an
anchor, whereas other methods can work with a grid element that is not necessarily an anchor.
GridAddress
The API helper class GridAddress identifies the location of cells within tables. GridAddress
and the other GridXXX classes are defined in TableTypes.h. In InDesign, GridCoord is an alias
for the primitive type int32. The non-default constructor for GridAddress is as follows:
GridAddress(GridCoord row, GridCoord column)
where row and column are the row and column, respectively, in which the top-left of the cell is
located.
Another key concept is the underlying grid on which measurements are made, consisting of
grid elements.
For example, if a cell is split by a vertical line, this increases the resolution of the grid in the
horizontal direction. The resolution of the grid in a given direction is determined by projecting
all cell boundaries onto an axis in that direction. Adding another vertical line means an additional projection onto the horizontal axis; i.e., an increase in horizontal grid resolution.
The grid lines are not uniformly distributed. The grid is rectilinear.
Determining what row a particular cell is in can become quite complex in a table with many
split and merged cells; that is, tables with many nontrivial GridSpan values. The algorithm to
determine the GridAddress for a particular cell is straightforward:
1. Determine the resolution of the underlying grid. For the horizontal direction, this can be
calculated by projecting all vertical edges to the x-axis (top of the table). Similarly, for the
vertical direction, project all horizontal edges to the y-axis (left of the table).
2. Calculate the coordinates on this grid. The process is shown in Figure 195. The underlying
grid is just fine enough so there are no cell edges that do not lie on an edge in this underlying grid. In the figure, the GridAddress of the red cell is (8, 9). The table has an underlying
grid 12 rows wide and 11 columns high.
398
Tables
Concepts
FIGURE 195
0
Complex table and underlying grid
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8,9
The GridAddress of a cell can change even if its physical location does not change.
GridSpan
GridSpan is a class that represents the dimension of an area within a table, expressed as coverage of elements in the notional underlying grid. GridSpan has a constructor with two arguments:
GridSpan(int32 height, int32 width)
where height and columns are the numbers of rows and columns spanned, respectively (on the
underlying grid).
When a table is created (e.g., using the Insert Table menu command), each cell in the table has
the trivial GridSpan(1,1). The concept of GridSpan is illustrated in Figure 194, which shows the
GridSpan for several cell configurations
The GridSpan of a given cell can change as more cells are added to a table—for example, by a
cell being split vertically in a column that this cell spans—even if the physical dimension of the
cell of interest does not vary. GridSpan for a cell is affected by changes in the columns spanned
by a cell and the rows spanned by a cell.
Merging a pair of cells with trivial GridSpan(1,1) gives one cell with a GridSpan(2,1) if the cells
are merged vertically and GridSpan(1,2) if the cells are merged horizontally.
Cells may be unmerged by selecting any merged cells—even the whole table—and choosing
Unmerge Cells.
Splitting a cell with GridSpan(2,1) in the vertical direction gives two cells with trivial GridSpan(1,1). Splitting the cell horizontally leads to the GridSpan of other cells in the table being
recalculated.
Tables
399
Tables
Concepts
GridArea
The GridArea class is used to specify a region within a table. Like GridSpan and GridAddress,
GridArea is calculated on the underlying grid. GridArea has a constructor with four arguments:
GridArea(GridCoord topRow, GridCoord leftColumn, GridCoord bottomRow, GridCoord
rightColumn)
where:
z
topRow is the row coordinate for the top-left of area.
z
leftColumn is the column coordinate for the top-left of area.
z
bottomRow is one greater than the lowest row contained in the area; i.e., the row immediately below the given area and outside it.
z
rightColumn is one greater than the rightmost column contained in this area; i.e., the column immediately to the right of the given area and outside it.
The choice of using one greater than the last row or column allows for specifying an empty area
in a simple way. That is, an empty GridArea can be specified by writing GridArea(0,0,0,0), for
example.
Alternately, given the definition of GridArea, it is equivalent to the following:
GridArea(top, left, row-span, column-span)
Table and cell selection
You can select a whole table or cells within a table, allowing the selected cells to be deleted or
their properties to be altered as a block. Alternately, you can select text within a table cell,
allowing the text properties to be specified for this text run. This is shown in Figure 196 and
Figure 197.
FIGURE 196
400
Table cells selected
Tables
Concepts
FIGURE 197
Table text selected
The selection manager hides the detail of the particular selection type involved from client
code. The intent is that client code interact with the table model only indirectly, through the
integrator suite interfaces, which present a uniform facade to client code. When the cursor is
within a table cell, InDesign can perform operations for any of the three selection types: text
selection, cell selection, and table selection.
Some operations are possible only with table cells selected; for example, the Merge Cells command is enabled only when multiple table cells are selected.
Table composition
Typically, a table is associated with a text frame. One text frame can contain multiple tables.
The table composer flows the row content of tables between table frames. Once a table becomes
too deep for the containing text frame, rows from the table may be flowed to another text
frame, if there is one linked to the current table’s text frame. Only whole rows are flowed by the
table composer.
Text composition within a table cell uses the same text-composition engines that operates over
normal text (i.e., text outside tables). For more information on text composition, see the “Text
Fundamentals” chapter.
A text frame that contains one or more table rows that are too deep to display is marked as
overset. The overset rows can be flowed to another text frame by linking the frames, exactly as
for overset text. The visual indicator for table cell overset is drawn as a text adornment.
Tables can be wider than the containing text frame; table rows, on the other hand, flow between
linked cells, or the frame is marked as overset.
Tables
401
Tables
Design and architecture
Table region
A table can be viewed as composed of regions: header rows, footer rows, left column, right column, and body rows. When inserting a table, headers and footers can be specified in the Insert
Table dialog box. A table style can set different cell styles for each region.
Design and architecture
Table model versus text model
Text-model extensions for tables
The basic text model API is extended to accommodate tables. kTextStoryBoss aggregates interfaces to support the abstraction of text-story threads, which are spans of characters that represent the text content of one cell. The text-story thread interface (ITextStoryThread) is
aggregated on the kTextStoryBoss class, which is used to represent the main text flow within a
story designated as the primary story thread.
Many higher-level APIs make it possible to manipulate the content of text cells without having
to manipulate the underlying text model directly. The essential API for this is the ITableTextContent interface, which is aggregated on the kTableModelBoss class. This provides a mechanism to get and set table text in chunks.
How tables are connected to text
The kTextStoryBoss boss class is the fundamental class that represents the textual content of
stories. For more information, see the “Text Fundamentals” chapter.
A story can contain zero or more tables. Tables can be nested within tables to an arbitrary
depth. The text-model interface (ITextModel) is the fundamental API to interact with stories;
for every table, there is an embedding story. An ITextModel interface can be used to obtain the
text-cell contents. The UML diagram in Figure 198 is a graphical representation of the relationship between the story and table models. It shows other key abstractions, such as the tablemodel list and the TextStoryThreadDictHier.
402
Tables
Design and architecture
FIGURE 198
Relation between story and table models
KTextStoryBoss
1
1
1
1
1
text::ITextStoryThreadDictHier
table::ITableModelList
Manages collection
Manages collection
text::ITextStoryThreadDict
table::ITableModel
1
1
1
1
kTableModelBoss
1
Text and tables are connected in two ways:
z
Through the table-model list (ITableModelList, aggregated on kTextStoryBoss). This is relatively straightforward to understand, but it may be deprecated in future versions of the API.
z
Through ITextStoryThreadDictHier. This is more complex to understand, but it represents
the new architecture and will be a more dependable API moving forward.
Obtaining a reference to ITableModel through the table model list meant making a series of
API calls by client code to navigate the table architecture. The alternative scheme to obtain a
reference to an ITableModel has a starting point of an ITextModel interface on a kTextStoryBoss object. The difference is that there is no dependence on ITableModelList.
Another API that allows connections between tables and text to be explored is ITableTextContainer. Given a table, ITableTextContainer allows client code to determine the embedding text
model. This interface is aggregated on kTableModelBoss.
Tables
403
Tables
Design and architecture
Table data model
A key abstraction in the table model is the kTableModelBoss boss class. Among other things,
this class encapsulates the data model for a table, allows iteration over the cells in a table, allows
access and modification of the attributes of a table, and provides access to the text chunks (or
other content) in a table. It provides APIs that enable copy, paste, and deletion of content and
access to the geometry of the table. For some of the key interfaces aggregated on the boss class,
see Figure 199.
FIGURE 199
Interfaces on the kTableModelBoss boss class
Access to text in cells
Information about text flow in
which table is anchored
<<interface>>
ITableTextContent
<<interface>>
ITableTextContainer
<<interface>>
ITableGeometry
<<interface>>
ITableLayout
ITableCommands encapsulates change
to the model in a command sequence
and provides undo/redo services to
client code.
kTableModelBoss
<<interface>>
ITableCommands
<<use>
<<use>
<<use>
<<interface>>
ITableModel
<<interface>>
ITableAttrAccessor
Read-only access to
the geometry
ITableLayout is responsible for
mapping the model to the
composed state.
<<interface>>
ITableAttrModifier
Core of the table model subsystem;
table structure, copy/paste, content
boss, and text content manager
access
NOTE:
Plug-in (client) code that uses the API exposed through interfaces aggregated on the
kTableModelBoss class should be written to use the selection suites and facades.
The kTableModelBoss boss class aggregates interfaces like ITableModel. This interface has
methods to obtain table iterators; these can be used to navigate a table or iterate through a specific GridArea within a table. These are STL-like iterators, supporting both forward and reverse
iteration.
The ITableAttrAccessor and ITableAttrModifier interfaces also are aggregated on kTableModelBoss. These provide detailed queries over the set of table attributes and allow applying overrides to table and cell attributes. In practice, most of the capability required for client code to
404
Tables
Design and architecture
query attributes and apply overrides is provided by the more convenient suite interface, ITableSuite. ITableAttrModifier is not likely to be used directly by client code; all the capabilities
exposed by this interface are available through methods on ITableCommands, which wrap
changes to the document object model in a command sequence in a clean and convenient way
for client code to use. Client code should use ITableCommands (rather than ITableAttrModifier). See the API Reference for details of the interfaces aggregated on kTableModelBoss.
The ITableTextContent interface is aggregated on kTableModelBoss; this allows the text contained within cells to be accessed. Do not confuse this interface with ITableTextContainer,
which is a fundamental API that represents the connection between a table and the story
(kTextStoryBoss) within which it is embedded.
The kTableModelBoss class represents the data model for an entire table. A table is composed
into a set of table frames (kTableFrameBoss) by the table composer.
<SDK>/source/sdksamples/tablebasics indicates how to acquire a reference to a table model.
This code is in the context of a suite implementation that was added to a concrete-selection
boss class.
Table layout
ITableLayout on kTableModelBoss provides detailed information about the layout of the table.
ITableLayout centralizes the persistent data into one implementation and contains several subclasses responsible for maintaining table layout information, like Row (representing the composed state of table-model rows), Frame (mapped to one persistent kTableFrameBoss object),
and Parcel (holding information about IParcel). ITableLayout also provides iterators to access
these layout elements. For example, ITableFrame can be obtained from ITableLayout using the
following snippet:
// Assume that iTableModel is a valid table model interface ptr.
InterfacePtr<ITableLayout> tableLayout(iTableModel , UseDefaultIID());
ITableLayout::frame_iterator frameIter = tableLayout->begin_frame_iterator();
ITableLayout::frame_iterator endFrameIter = tableLayout->end_frame_iterator();
while(frameIter != endFrameIter)
{
InterfacePtr<ITableFrame> tableFrame(frameIter->QueryFrame());
...
}
The ITableLayout::Row subclass represents the partial mapping of a model row to a table frame
(represented by the kTableFrameBoss). For each model row, there is one or more corresponding table layout rows. Each row contains information for each parcel owned by a table cell, like
the GridCoord of the model row it maps and the UID of the table frame with which it is associated. Each layout row is associated with one table frame.
ITableLayout is responsible for the lifetime management of table frames. Besides containing the
layout rows, the table frame also is responsible for maintaining the link back to the containing
Parcel and associated ParcelList.
Tables
405
Tables
Design and architecture
Table frames
The rows in a table may not all fit in one text frame. The table composer groups as many rows
as will fit into the first text frame, the second text frame, and so on. A table with many rows
may be spread across multiple, linked, text frames.
These groups of rows are table frames. Each row is part of one and only one table frame.
A table is divided into one or more table frames, spread across one or more text frames. A table
frame (represented by the kTableFrameBoss boss class) is a collection of rows within a text
frame. There can be multiple table frames within a text frame (if the text frame contains multiple tables), so there are at least as many table frames as text frames.
An occurrence of a table frame leads to its UID for the corresponding kTableFrameBoss object
appearing in the owned item strand.
Tables can have an arbitrary number of table frames, split between linked text frames. If rows
cannot be displayed within the last linked frame in the series, the frame is marked as overset.
For each text frame, there is a corresponding table frame that represents the collection of rows
being displayed within the given text frame.
For each row of a table, there is an occurrence of the kTextChar_Table or
kTextChar_TableContinued character in the text data strand. For more information, see the
“Text Fundamentals” chapter.
Cell data model
Cell-content managers
The table architecture supports having different cell-content managers. In practice, InDesign
has only text cell-content manager (with behavior provided by kTextCellContentMgrBoss) and
a dummy cell-content manager. (Future versions of InDesign may add canonical cell-content
managers for content types like images.) You may add your own types of cell-content managers. The main responsibility of a cell-content manager is to implement the ICellContentMgr
interface.
A cell-content manager is created during table creation. You can get the cell-content manager
from the table via ITableModel::QueryContentMgr().
Cell contents
Text cells (the only kind in InDesign) have behavior provided by the kTextCellContentBoss
boss class. Instances of this boss class do not store the content directly but provide a convenient
API on top of the table text model to allow its manipulation. It also is possible to quickly and
conveniently manipulate the text content of a cell through the text model. In this case, API
classes like TextIterator can be used to access a range within the text model.
To set cell text, the ITableCommands::SetCellText API method makes it straightforward to
change the cell text at a given GridAddress. The ITableCommands interface is aggregated on
kTableModelBoss.
406
Tables
Design and architecture
Code snippets like SnpAccessTableContent.cpp show how to work with the APIs to access cell
contents. See also the TableBasics plug-in and the SnpSortTable.cpp code snippet for examples
of accessing and manipulating table text content.
Cell strands
The kCellStrandBoss cell strand provides storage for cell attributes and aggregates an ITableStrand interface. This abstraction is likely to be hidden from client code and should be
treated as an implementation feature. Higher-level APIs, like ITableAttrAccessor and ITableCommands, provide methods to access and modify attributes without having to work at the
lower level of representation of the table strand.
Table attributes
The table architecture supports extensible attributes with inheritance. Attributes can be applied
to an entire table or one or more cells within a table, and there also are a small set of row-specific and column-specific attributes. As with the text subsystem, attributes are represented by
boss classes. Table-attribute boss classes have names that follow the pattern k<targetentity>Attr<property>Boss, where <target-entity> is one of Table, Cell, Row, or Column. Some
text-cell-specific attributes have TextCell in the attribute name.
The ITableAttrReport::AppendDescription method reports whether an attribute is a cell-specific, row-specific, column-specific, or table-specific attribute. To obtain information about the
default value of each attribute, see the API reference documentation.
Table-specific attributes
A table attribute describes one property of an entire table. A table attribute is applied to the
table itself; the attributes collectively define aspects of how the table should appear.
A table attribute is represented by a table attribute boss with a name that conforms to the pattern kTableAttr<property>Boss.
Row attributes
There are only a few row-specific attributes, which apply to a row or collection of rows within
the table. Row attributes are represented by boss classes. For details, see kRowAttrBaseBoss in
the API reference documentation. These are named to follow the pattern kRowAttr<property>Boss.
Column attributes
There are very few column-specific attributes. For details, see kColAttrBaseBoss in the API reference documentation. These are represented by boss classes, which are named to follow the
pattern kColAttr<property>Boss.
Tables
407
Tables
Design and architecture
Cell-specific attributes
There are a wide range of cell-specific attributes that can be overridden. These can be divided
into those that apply to an entire cell and those that are specific to one border (bottom, left,
right, top) of the cell. See Figure 200.
FIGURE 200
Examples of cell-specific attributes
This cell has an override (2p) for the bottom
text inset kCellAttrBottomInsetBoss
This cell has an override for the bottom stroke
color kCellAttrBottomStrokeColorBoss
This cell has an override for the bottom tint
kCellAttrBottomStrokeTintBoss
This cell has an override for the cell fill color
This cell has an override for a diagonal tint
kCellAttrFillColorBoss
kCellAttrDiagnolTintBoss
This is a cell that has had its first line
offset set to be fixed and with contents
clipped to cell kCellAttrClipBoss
This is a cell that has had its first line offset set to be fixed but without contents
clipped to cell kCellAttrClipBoss
This cell has an inline graphic and some
This cell has an inline graphic and some
text that flows around it. It has a diagonal
in front of the cell contents
text that flows around it. It has a diagonal
behind the cell contents
kCellAttrDiagnolsOnTopBoss
kCellAttrDiagnolsOnTopBoss
.
408
kCellAttrRotationBoss
This cell has an override
for the rotation (90 deg.)
kCellAttrRotationBoss
This cell has an override
for the rotation (270 deg.)
This cell has an override (4 pt) for the bottom stroke This cell has no overrides.
weight kCellAttrBottomStrokeWeightBoss
Tables
Design and architecture
Text-cell-specific attributes
Some cell attributes apply only to text cells. Text cells are the only supported kind of cell in
InDesign, but future versions may support other cell-content types, so text cells are not
generic-cell attributes.
Text-cell attributes are represented by boss classes named to follow the pattern kTextCellAttr<property>Boss.
Default table attributes
The ITableAttributes interface is aggregated on the workspace (kWorkspaceBoss class) and the
document
workspace
(kDocWorkspaceBoss
class),
with
interface
identifier
IID_IDEFAULTTABLEATTRIBUTES.
When a new document is created, this root set of table attributes is applied to the document.
This attribute list is further enhanced when the swatch list is available for the new document,
since some of the attributes reference a swatch (black or none) by UID.
When a new table is created (by kNewTableCmdBoss class, an instance of which is created and
executed through the ITableCommands interface on kTableModelBoss), the root table style is
applied to the table.
Table and cell styles
A style can be considered a collection of attributes. Table and cell styles provide a mechanism
to name and persist a particular set of table attributes.
Table style and cell style are analogous to paragraph style and character style in the text. Each
style has an AttributeBossList list that maintains the set of attributes that apply to that style.
Access to the attributes in a style is achieved through the ITableAttributes interface. Cell styles
are associated with cell-specific attributes; however, table styles can be associated with both
table attributes and cell-specific and other attributes.
Styles form a hierarchy rooted at the root style. Each style (except the root style) is based on a
style. The AttributeBossList list for a particular style records only the differences from the style
on which it is based; that is, the set of attributes the particular style overrides.
As with other types of styles, table and cell styles support the Style Group concept. Users can
create, edit and delete folders (called Groups), in the table styles and cell styles palettes. Style
group is a collection of styles or groups. The user also can nest groups inside groups and drag
styles within the palette to edit the contents of a group. Styles do not need to be inside a group
and can exist at the root level of the palette.
Table styles
A table style is represented by a kTableStyleBoss. Its IStyleInfo stores general, style-related
information like style name and parent style. Its ITableAttributes interface stores a list of table
attributes that are different from its parent.
The root table style (known in the application user interface as [No Table Style]) contains a
complete set of default table attributes. This defines the default look of the table with no further
formatting applied.
Tables
409
Tables
Design and architecture
The application also defines a basic table style called [Basic Table]. The basic style cannot be
deleted, but you can change it. Initially, the basic style is based on the root style; however, its
parent style can be changed.
Table styles are accessible through the table-style group manager (signature interface IStyleGroupManager with interface identifier IID_ITABLESTYLEGROUPMANAGER) on the document (kDocWorkspaceBoss) and session (kWorkspaceBoss) workspace boss classes. When a
document is created, its style group manager inherits the existing set of styles from the session
workspace.
THe IStyleGroupManager of the table styles works exactly the same way as that of paragraph
and character styles. For details, see the “Text Fundamentals” chapter.
Cell styles
A cell style is represented by a kCellStyleBoss. Its IStyleInfo store general, style-related information like style name and parent style. Its ITableAttributes interface (with implementation
kCellStyleCellAttributeListImpl) stores a list of cell-specific attributes that are different from its
parent cell style.
The root cell style (known in the application user interface as [None]) is really nothing; in fact,
it defines nothing. It is synonymous with the root character style. The root cell style cannot be
deleted.
Cell styles are accessible through the cell-style group manager (signature interface IStyleGroupManager with interface identifier IID_ICELLSTYLEGROUPMANAGER) on the document (kDocWorkspaceBoss) and session (kWorkspaceBoss) workspace boss classes. When a
document is created, its style group manager inherits the existing set of styles from the session
workspace.
As with table styles, the IStyleGroupManager of the cell styles works the same way as that of
paragraph and character styles. For details, see the “Text Fundamentals” chapter.
Regional cell styles
Each table style can assign different cell styles for different regions: headers rows, footers rows,
first column, last column, and body rows. When a table style is applied to a table, these regional
cell styles are applied to the appropriate regions.
Regional cell styles are considered as table attributes of a table style. Except for body rows, the
values of two related attributes, “cell style” and “use body,” determine a regional style. See
Table 97.
410
Tables
Design and architecture
TABLE 97 Regions and their controlling attributes
Region
Cell-style attribute boss
Use body attribute boss
Body rows
kTableAttrBodyCellStyleBoss
kInvalidClass
Header rows
kTableAttrHeaderCellStyleBoss
kTableAttrHeaderUseBodyCellStyleBoss
Footer rows
kTableAttrFooterCellStyleBoss
kTableAttrFooterUseBodyCellStyleBoss
Left column
kTableAttrLeftColCellStyleBoss
kTableAttrLeftColUseBodyCellStyleBoss
Right column
kTableAttrRightColCellStyleBoss
kTableAttrRightColUseBodyCellStyleBoss
To set a regional cell style, you need to create and apply an appropriate “cell style” attribute as
well as create and apply an appropriate “use body” attribute (except for body rows). To get a
regional cell style, you need to look for these attributes at the complete list of attributes. For
more information, see the “Tables” chapter of Adobe InDesign CS4 Solutions.
Table and cell style in the data model
Figure 201 illustrates a simplified class diagram of table styles, cell styles, and their relationships to application/document workspaces and the table model. The descriptions of the boss
classes give hints to navigate through these related classes.
Tables
411
Tables
Design and architecture
FIGURE 201
Table and cell style
The table.
Table's own ITableAttributes
stores local overrides of table
attribute.
The application or document
preferences.
The workspace stores table
styles and cell styles. They
are accessible through
IStyleGroupManager interface
using interface identifier
IID_ITABLESTYLEGROUPMANAGER
and
IID_ITABLESTYLEGROUPMANAGER
respectviely.
It's ITableAttrModifier interface
has access to the table style
applied.
Its ITableAttrAccess encapsulate
the access to applied cell style
for individual cell.
Represents
individual cell.
Its ITableStrand
stores cell attribute
overrides.
«boss»
kWorkspaceBoss
kDocWorkspaceBoss
ITableModelStorage::QueryNthStrand()
1
1
«boss»
kTableModelBoss
ITableAttributes
1
1
1
ITableAttrAccessor::GetCellStyle
ITableAttrModifier::GetTableStyle()
1..*
IStyleGroupManager::FindByName()
«boss»
kCellStrandBoss
1
1..*
«struct» ITableAttributes::QueryByClassID()
kTableStyleBoss
1
1..*
IStyleGroupManager::FindByName()
1..*
«struct»
kCellStyleBoss
ITableAttributes
1..*
The table style.
It can define different cell
style for different region.
They are defined as attributes
of the table style, thus need
to query appropriate Boss
class to get the regional style.
412
The cell style.
Its ITableAttribute interface
stores the cell specific attribute
of the cell style, includes the
default paragraph style.
ITableStrand
Tables
Design and architecture
Formatting tables, cells, and table text
Formatting a table
Tables are formatted according to table-specific attributes. Every table is assigned a table style.
ITableAttributes, aggregated on kTableModelBoss, provides coarse-grained access to the locally
overridden attribute list for a particular table. For a given attribute (such as table stroke), InDesign follows these steps to format a table:
1. If there is a local override, format the table according to the override.
2. Otherwise, check the attributes in the table style of the table. If the attribute is among those
attributes defined in ITableAttributes of kTableStyleBoss, format the table according to the
value of the attribute defined in the table style.
3. Otherwise, check the table style’s parent style, until either the attribute is found or the root
table style (which defines all default attributes) is reached. Format the table according the
value of the attribute.
Formatting a cell
Cells are formatted according to cell-specific attributes, similar to tables. For a given cell
attribute, InDesign follows these steps to format a cell:
1. If there is a local override, format the cell according to the override (which is defined in ITableStrand on kCellStrandBoss).
2. Otherwise, check the attribute against the attribute list of the cell style applied to the cell. To
determine the cell style, check if a cell-style override is applied. If so, use the cell style; otherwise, use the regional cell style defined in the table style of the table.
3. As with table styles, if the attribute is among the attributes defined in ITableAttributes of
kCellStyleBoss, format the cell according to the value of the attribute defined in the cell
style. Otherwise, check the cell style’s parent style, until either the attribute is found or the
root cell style (where nothing is defined) is reached. Format the cell according to the value
of the attribute or leave it as the default format.
Formatting text in a cell
It is helpful to view table attributes as equivalent to paragraph-level attributes for normal text,
and cell attributes as equivalent to character-level attributes. The analogy holds insofar as the
effective attributes are calculated by applying the following in the order listed below:
Tables
z
Table-style attributes
z
Table-level overrides
z
Row or column cell style
z
Cell overrides
z
The paragraph style defined by the cell
z
Paragraph override
413
Tables
Essential APIs
z
Character style
z
Character overrides
To format a character in a table cell, InDesign looks at the reverse of the order above. An override always takes priority than a style; a smaller range style always takes priority over a coarse
one.
TABLE 98 Text-formatting priorities
Overrides
Styles
Character overrides
Character styles
Paragraph overrides
Paragraph styles
Cell overrides
Cell style
Table overrides
Table style
A cell style can have a paragraph style used in the paragraphs of a cell to which the cell style is
applied. If a cell style does not have a paragraph style defined, any cell that has that cell style
applied to it will have all paragraph-style formatting removed.
Essential APIs
Table commands
ITableCommands is an interface that is aggregated by kTableModelBoss and provides methods
to create and execute table-related commands, instead of having to manipulate the table model
at a lower level. You can use ITableCommands to affect the table model on any specific table
area. Some functions in ITableCommands also are provided by ITableSuite, which instead
operates on the currently selected area of the table. For more details, see ITableCommands in
the API reference documentation.
ITableSuite
ITableSuite is a key API for manipulating tables in client code. It provides much of the required
capability for plug-ins, with the exception of insertion and retrieval of cell content.
The ITableSuite interface is aggregated on the integrator suite boss class (kIntegratorSuiteBoss),
which makes it available through the abstract selection. Implementations of this interface are
provided on various concrete-selection boss classes, which are hidden from client code
through the facade of the abstract selection. To obtain ITableSuite, client code should query the
selection manager (ISelectionManager) for the interface.
414
Tables
Essential APIs
The integrator suites expose methods that determine whether a capability is present on the current selection; this always should be tested before trying to exercise a capability. For more detail
on the integrator suites, see the “Selection” chapter.
Suite interfaces typically use this pattern of checking for a service (capability) and, if the
abstract selection supports the capability, calling the method called. For details, see ITableSuite
in the API reference documentation.
Use is illustrated in the following code:
bool16 canDeleteTable = iTableSuite->CanDeleteTable();
if(canDeleteTable)
{
iTableSuite->DeleteTable();
}
ITableStyleSuite and ITableStylesFacade
ITableStyleSuite is the selection suite used to manipulate table styles, such as creating and editing table styles of a selected table; applying a table style to the current selection; and getting and
setting local overrides of table attributes. As with other selection suites, the interface can be
acquired through ISelectionManager.
Aggregated on kUtilsBoss, ITablesStylesFacade is used to manipulate table styles on a table or
table style directly. It is the counterpart of ITableStyleSuite.
The ITableAttrAccessor and ITableAttrModifier interfaces aggregated on the kTableModelBoss
provide access to and mutation of table attributes; however, we recommend you use suites and
facades whenever possible.
ICellStyleSuite and ICellStylesFacade
ICellStyleSuite is the selection suite used to manipulate cell styles, like creating and editing the
cell styles of selected cells; applying a cell style to the current cell selection; and getting and setting local overrides of cell attributes.
The facade counterpart of ICellStyleSuite is ICellStylesFacade, aggregated on kUtilsBoss.
By using suites and facade, implementation details of cell strands and cell styles storage are
encapsulated. We recommend you use only suites and facades.
Tables
415
Tables
Essential APIs
416
Printing
Concepts
Printing
This chapter provides information about the concepts surrounding the process of printing a
publication, including the printing data model and commands, the printing user interface, key
interfaces, and how plug-ins can participate in the printing process.
Concepts
Printing is simply drawing to the printer
The implementation and user interface for printing from InDesign and InCopy are similar to
those for printing from other Adobe applications, like Photoshop and Illustrator. All these
applications use a common component, the Adobe Graphics Manager (AGM), to handle the
core printing tasks. Printing from the application entails drawing part or all of a document to a
printer rather than to the screen. Adobe applications use Display PostScript, supplied by AGM,
to draw graphics primitives to both the screen and the printer.
For more information on how document content is drawn to an output device, see the “Graphics Fundamentals” chapter.
Control can be shared
The tasks of printing from InDesign can be shared among the following:
z
AGM, which is responsible for PostScript generation.
z
The InDesign Print plug-in, which is responsible for the user interface, initialization of
print settings, and drawing pages. (See “The print action sequence” on page 421.)
z
Plug-ins that implement various print-related extension patterns, which provide the ability
to participate in the printing process carried out by the Print plug-in
A plug-in can use printing commands, interfaces, structures, and event-handling mechanisms
to print with the C++ API. By understanding the flow of operations in the application’s Print
plug-in, you can design a plug-in to participate in the print process and the print user interface
by hooking onto the various extensibility points.
Third-party plug-in developers do not have direct access to the AGM API. Instead, the application provides the user interface, commands, interfaces, and data structures necessary to drive
and participate in the printing process. For details, see “Printing data model” on page 419.
Inks and colors
When printing a document to any output device, the colors used in a printable item on each
page need to be rendered. Depending on the output device and print settings, each color can be
separated into multiple inks. This is done so a system like a four-color (CMYK) printing press
Printing
417
Printing
Concepts
can render each printable item. In some cases, printable items can contain spot colors, which
generally correspond to a specific ink.
Colors, gradients, and spot colors generally are referred to as swatches. A swatch is identified by
the IRenderingObject interface. Inks are identified by the IPMInkBossData interface.
For more details on inks and colors, see the “Graphics Fundamentals” chapter.
Overprinting
You can specify the way in which inks are printed on top of one another. Depending on how
you set up overprinting, the colors of overlapping items of different colors may appear differently. For more details on overprinting, see InDesign Help.
Trapping
Trapping allows an output device to compensate for printing problems due to paper misregistration. When the print medium (e.g., paper) is not registered exactly in a multi-color press
system, the output can show unintended gaps between inks. An example of a case in which this
is especially important is when a printable item has a thin, dark border around a lighter-colored shape. By specifying how trapping is set up, you can compensate for media-registration
errors.
Color management and proofing
What you see on the screen may not always be what you see printed on print media. This is due
to variances in how each device used in the design and print process renders colors. For example, your computer monitor may show a bright red apple, but when that is printed on paper
using a specific set of inks on a specific output device, the apple may appear to be a different
color or brightness. Colors also may shift in the input device, such as digital scanners and cameras. It can be hard to make sure the colors appear as you want on the print media. InDesign
and InCopy offer features to manage color throughout the design process.
For more details on color management and proofing, see “Color Management” in InDesign
Help and the “Graphics Fundamentals” chapter in this document.
Preflight and packaging
Preflight is a way to verify before you send a publication to the output device that you have all
associated files, fonts, assets (e.g., placed images and PDF files), printer settings, trapping
styles, and so on. For example, if you placed an image as a low-resolution proxy but do not have
the high-resolution original image accessible on your hard disk (or workgroup server), that
may result in an error during the printing process. Preflight checks for this sort of problem.
For details, see the “Implementing Preflight Rules” chapter.
418
Printing
Printing data model
Exporting to EPS and PDF
Although the user interface for exporting a publication to the EPS and PDF file formats is different from that for printing, the output processes incorporate the same types of drawing components (AGM, inks and color swatches, color management, etc.). Some publishing workflows
may incorporate exporting to EPS or PDF as part of the proofing and output steps. Though this
document does not go into the details of extending the EPS and PDF export capabilities of the
application, you can refer to “Exporting to EPS and PDF” on page 450 for some highlights of
how to drive the EPS and PDF export features.
Printing data model
This section describes the printing data model, commands, interfaces, and structures available
in the SDK. InDesign and InCopy can print documents and books to a registered output
device. The registered output device can be a printer, PostScript file, or external file format like
EPS or PDF.
For more information on the PostScript language, see the PostScript Language Reference at
http://www.adobe.com/products/postscript/pdfs/PLRM.pdf.
Print settings
This section describes the data interfaces that make up the print-settings data model. For information on bosses that aggregate these interfaces, see the API reference documentation. The
print settings generally are set by the user in either the Print or Print Presets dialog box. (For
more information on the Print user interface, see “Print user interface” on page 423.)
IPrintData
IPrintData is the main data interface that keeps most print-settings data. It also is the interface
passed among the print commands during a print operation or in the Print or Print Presets dialog box.
Other interfaces that store print settings
Printing
z
IPrintDeviceInfo stores data about printing devices.
z
IPrintJobData stores data specific to a print job.
z
IPrintContentPrefs specifies which content (e.g., text, page items, Japanese layout grids, or
frame grids) should be printed or omitted from printing.
z
IOutputPages stores data about pages or spreads to print.
z
ITrapStyle stores a set of trap settings with a specified name.
z
IPrintGlyphThresholdPref specifies whether all glyphs of a font are to be downloaded to the
output device completely or subsetted.
419
Printing
Utility APIs
Print preset styles
A print preset style is a group of print settings with a specified name. Of the print-settings data
model interfaces, the IPrintData and IPrintDeviceInfo data interfaces are stored as part of a
print-preset style. A print-preset style is represented by kPrStStyleBoss, and a list of defined
print-preset-style boss instances is managed in the kWorkspaceBoss using the IPrStStyleListMgr interface. kPrStStyleBoss also aggregates IGenStyleLockInfo, which specifies whether the
style is locked or can be deleted.
When a list of print-preset settings is exported to a file (database), the root of that database is
an instance of kPrStExportRootBoss, which also aggregates IPrStStyleListMgr. Using this
IPrStStyleListMgr interface, you can iterate through the print-preset styles that are stored in the
exported file.
Trap styles
A trap style is a set of trap settings with a specified name. A trap style is represented by kTrapStyleBoss and aggregates the ITrapStyle interface. kTrapStyleBoss also aggregates IGenStyleLockInfo, which specifies whether the style is locked or can be deleted. A trap style is
associated with an individual page in a document.
A list of trap styles is accessible from ITrapStyleListMgr on kDocWorkspaceBoss (the trap
styles used on a document), kWorkspaceBoss (the workspace default trap styles), and kBookBoss (the trap styles used in a book).
NOTE:
ITrapStyleListMgr does not inherit from IGenStlEdtListMgr like IPrStStyleListMgr
does; therefore, the behavior of some ITrapStyleListMgr methods differs from that of
any interfaces derived from IGenStlEdtListMgr, and the general style-editing family of
APIs (commands like kGenStlEdtExportStylesCmdBoss and interfaces like
IGenericSettings) do not apply to trap styles managed in ITrapStyleListMgr.
Utility APIs
The following utility interfaces are available to facilitate the management and manipulation of
print data settings:
z
IPrintUtils (kUtilsBoss) contains various methods for checking settings on IPrintData.
IPrintUtils also has InitializeOutputPages, a method that fills a list of pages to output, given
IPrintData and a document.
z
ITrapStyleUtils (kUtilsBoss) contains methods to facilitate manipulation of trap styles.
For details, see the API reference documentation.
420
Printing
The print action sequence
The print action sequence
When you print a document or book in InDesign or InCopy, the application processes a hierarchy of commands. This is the core feature provided by the Print plug-in. This section discusses
the sequence of events that take place during the print process and explains the various extensibility points for third-party plug-ins.
The primary actions taken by the Print plug-in are as follows:
1. Client code (usually in the form of a menu action or script event) receives two pieces of
information: IDocument for the document to be printed and a user-interface options flag
indicating which parts of the print user interface should be displayed. The client code creates and executes a specific print-action command, which is one of the following:
z
kPrintActionCmdBoss — for printing a document in InDesign.
z
kBookPrintActionCmdBoss — for printing a book in InDesign.
z
kInCopyPrintActionCmdBoss — for printing a document in InCopy.
NOTE:
For brief descriptions of these commands, see “Print-action and supporting
commands” on page 448.
2. From within the print-action command, a series of supporting print commands is processed. All subsequent steps originate from inside this print-action command.
3. A print-command data interface (IPrintCmdData) used by all print commands gets the
IDocument supplied by the client code, along with the UIDRef for the print style (if any),
the document’s ink list, and trap-style manager.
4. The print-action command polls print-setup providers (service providers of kPrintSetupService) to determine whether any plug-in wants to participate in the print process by calling
IPrintSetupProvider::StartPrintPub (for document printing) or IPrintSetupProvider::StartPrintBook (for book printing). A print-setup provider can either take over the process at
this point (including aborting the printing process by means of a flag passed in the parameter list) or perform an action and return control to the Print plug-in.
5. If control returns to the Print plug-in, a non-persistent instance of a print-settings data boss
(generally kPrintDataBoss) is created. This is passed around during the rest of the printing
process in each supporting print command.
6. The print-action command polls print-setup providers and calls IPrintSetupProvider::BeforePrintUI. A print-setup provider can determine whether to display the print
user interface before returning control. After all print-setup providers are polled, the userinterface options flag is examined. If the print user interface is to be shown, the kPrintDialogCmdBoss is created and processed. If the dialog box is opened, the user can modify the
settings for this particular print operation. The settings are stored back in the instance of
the print-settings data boss created in the previous step.
Printing
421
Printing
The print action sequence
7. The print-action command polls print-setup providers and calls IPrintSetupProvider::AfterPrintUI. A print-setup provider can either take over the process at this point or
perform an action (for example, set the flag indicating whether to show the Save dialog box)
and return control to the Print plug-in. If the Save dialog box is to be shown, the kSaveFileDialogBoss is processed.
8. If control returns to the Print plug-in, the print settings data (IPrintData) is saved to the
document by means of processing kPrintSavePrintDataCmdBoss (for documents) or
kBookSavePrintDataCmdBoss (for books).
9. Data is gathered for the print job. kPrintGatherDataCmdBoss is created but not yet processed. An instance of IOutputPages is created.
10. The print-action command polls print-setup providers and calls IPrintSetupProvider::BeforePrintGatherCmd. A print-setup provider can determine whether to allow processing of the kPrintGatherDataCmdBoss created earlier and whether to return control.
11. If control returns, the print-action command polls print-setup providers and calls IPrintSetupProvider::AfterPrintGatherCmd. A print-setup provider can determine whether to
quit here or return control to the Print plug-in.
12. If control returns, a core print command (kNewPrintCmdBoss if running in InDesign or
kInCopyNewPrintCmdBoss if running in InCopy) is created and processed. This is where
the document or book data actually is sent to the output device.
13. If an error occurs during the core print command, it polls print-setup providers and calls
IPrintSetupProvider::PrintErrorEvent. A print-setup provider can handle the print event
and determine whether the error message should be displayed to the user. The global error
code is set, and the core print command is aborted.
14. If no errors occurred in the core print command, the print-action command saves the print
data again (in case it changed during printing).
15. The print process ends. The print-action command polls print-setup providers and calls
IPrintSetupProvider::EndPrint. This is the final opportunity for print-setup providers to
participate in the print process.
Common print interfaces
The following are the data interfaces used during the print-action sequence and supporting
commands:
422
z
IPrintCmdData stores most print settings. IPrintCmdData is used commonly among the
print commands.
z
IPrintData stores most print settings data. (See “IPrintData” on page 419.)
z
IOutputPages stores data about pages or spreads to print.
z
IBookPrintData stores book-printing options for the kBookSavePrintDataCmdBoss.
z
IPrintJobData stores data specific to a print job.
Printing
Print user interface
z
IPrintDialogCmdData stores data for the kPrintDialogCmdBoss.
z
IPrintContentPrefs specifies which content (e.g., text, page items, Japanese layout grids, and
frame grids) should be printed or omitted from printing.
z
IInCopyGalleySettingData stores settings used to construct the Galley panel and contains
information about the constructed Galley panel for printing and PDF export in InCopy.
Print user interface
The print user interface lets an end user change print settings during a print process or in print
preset styles. This section describes the design of the Print and Print Presets dialog boxes and
how they are invoked, and it presents an overview of how to extend these dialog boxes.
Print dialog box
The Print dialog box shows the current print settings for the frontmost document. This dialog
box opens when the user chooses File > Print in InDesign or chooses one of the print preset
styles from the File > Print Presets menu. The Print dialog box also opens when the user
chooses Print from the Book panel. This dialog box is invoked programmatically in InDesign
by processing the kPrintDialogCmdBoss command.
NOTE:
An InCopy version of the Print dialog box is invoked by processessing the
kInCopyPrintDialogCmdBoss box. This section focuses on the InDesign version of the
Print dialog box.
The Print dialog box is a selectable dialog box, which contains a list of panels that are selectable
by a list (located on the left-hand side of the dialog box). The Print dialog box contains the following panels:
z
General
z
Setup
z
Marks And Bleed
z
Output
z
Graphics
z
Color Management
z
Advanced
z
Summary
The panels are shown in order in the following sections. Below a screen shot of each panel is a
list of methods on the print data-model interfaces (e.g., IPrintData) corresponding to each
user-interface element on the panel.
For details on these settings, see InDesign CS4 Help.
Printing
423
Printing
Print user interface
Print dialog box: General panel
FIGURE 202
Print dialog box: General panel
The three user-interface elements above the selectable panel region of the dialog box are common in all Print dialog box screenshots and are mapped to the following API methods:
z
Style Name — IPrintData::SetStyleName
z
Printer — IPrintDeviceInfo::UpdatePrinterInfo (which calls IPrintData::SetPrinter internally. Get with IPrintData::GetPrinter.)
z
PPD — IPrintDeviceInfo::UpdatePrinterInfo (which calls IPrintData::SetPPDFile and
IPrintData::SetPPDName internally. Get with IPrintData::GetPPDFile and IPrintData::GetPPDName.)
NOTE:
Generally, there is a corresponding Get method for each Set method.
The user-interface elements on the General panel are mapped to the API methods listed in
Table 99.
TABLE 99 Print dialog box: General panel
424
User-interface element
API method
Copies
IPrintData::SetCopies
Collate
IPrintData::SetCollate
Printing
Print user interface
User-interface element
API method
Reverse Order
IPrintData::SetReverseOrder
Pages: All
IPrintData::SetWhichPages(IPrintData::kAllPages)
Pages: Range
IPrintData::SetWhichPages(IPrintData::kPageRange) (Range
text edit box): IPrintData::SetPageRange
Sequence
IPrintData::SetPrintOption, using IPrintData::kBothPages
(for All Pages), IPrintData::kEvenPagesOnly, or
IPrintData::kOddPagesOnly
Spreads
IPrintData::SetSpreads
Print Master Pages
IPrintData::SetScope, using IPrintData::kScopeMaster or
IPrintData::kScopeDocument
Print Non-printing
Objects
IPrintData::SetPrintNonPrintingObjects
Print Blank Pages
IPrintData::SetPrintBlankPages
Print Visible Guides
and Baseline Grids
IPrintData::SetPrintWYSIWYGGridsGuides
Print dialog box: Setup panel
FIGURE 203
Printing
Print dialog box: Setup panel
425
Printing
Print user interface
The user-interface elements on the Setup selectable panel are mapped to the API methods
listed in Table 100.
TABLE 100 Print dialog box: Setup panel
426
User-interface element
API method
Paper Size
IPrintData::SetPaperSizeSelection if defined by the user
(IPrintData::kPaperSizeDefinedByUser), printer driver
(IPrintData::kPaperSizeDefinedByDriver), or
IPrintData::SetPaperSizeName
Paper Size: Width
IPrintData::SetCustomPaperWidth only if Paper Size is
defined by the user; otherwise (if defined by driver or papersize name),
IPrintData::SetCustomPaperWidth(IPrintData::kCustomPap
erSizeAuto)
Paper Size: Height
IPrintData::SetCustomPaperHeight only if Paper Size is
defined by user; otherwise (if defined by driver or paper-size
name),
IPrintData::SetCustomPaperHeight(IPrintData::kCustomPap
erSizeAuto)
Offset
IPrintData::SetCustomPaperOffset only if Paper Size is
defined by the user
Gap
IPrintData::SetCustomPaperGap only if Paper Size is defined
by the user
Transverse
IPrintData::SetPaperOrientation, using
IPrintData::kTransverse or IPrintData::kNormal
Orientation
IPrintData::SetPageOrientation, using IPrintData::kPortrait,
IPrintData::kLandscape, IPrintData::kReversePortrait, or
IPrintData::kReverseLandscape
Scale (radio button)
IPrintData::SetScaleMode using IPrintData::kScaleXAndY
Scale: Width
IPrintData::SetXScale
Scale: Height
IPrintData::SetYScale
Scale: Constrain
Proportions
IPrintData::SetProportional
Scale to Fit (radio
button)
IPrintData::SetScaleMode using IPrintData::kScaleToFit
Page Position
IPrintData::SetPagePosition, using
IPrintData::kPagePositionUpperLeft,
IPrintData::kPagePositionCenterHorizontally,
IPrintData::kPagePositionCenterVertically, or
IPrintData::kPagePositionCentered
Printing
Print user interface
User-interface element
API method
Print Layers
IPrintData::SetPrintLayers using kPrintAllLayers,
kPrintVisibleLayers, or kPrintVisiblePrintableLayers.
Thumbnails
IPrintData::SetTileThumbMode, using
IPrintData::kThumbnails or IPrintData::kTileThumbOff
Per Page
IPrintData::SetNumberOfThumbsPerPage. Calculate the
number of thumbnails on the page (e.g., 4x4 = 16).
Tile
IPrintData::SetTileThumbMode, using IPrintData::kTiling or
IPrintData::kTileThumbOff
Overlap
IPrintData::SetTilingOverlap
Print dialog box: Marks and Bleed panel
FIGURE 204
Print dialog box: Marks And Bleed panel (Roman feature set)
The user-interface elements on the Marks And Bleed selectable panel are mapped to the API
methods listed in Table 101.
Printing
427
Printing
Print user interface
TABLE 101 Print dialog box: Marks and Bleed panel
428
User-interface element
API method
All Printer’s Marks
Checking this sets the state of the five check boxes below it.
Crop Marks
IPrintData::SetCropMarks
Bleed Marks
IPrintData::SetBleedMarks
Registration Marks
PrintData::SetRegistrationMarks
Color Bars
IPrintData::SetColorBars
Page Information
PrintData::SetPageInformation
Type
IPrintData::SetPageMarkFile. For the Roman and Japanese
feature sets, you get the Default setting; in this case, the
PMString passed into SetPageMarkFile is blank. There are
extra details when using the Japanese feature set; see
“Japanese page-mark files” on page 449.
Weight
IPrintData::SetMarkLineWeight, using the enumerations
ranging from IPrintData::kMarkLineWeight125pt to
IPrintData::kMarkLineWeight30mm (see IPrintData.h.)
Offset
IPrintData::SetPageMarkOffset
Use Document Bleed
Settings
IPrintData::SetUseDocumentBleed
Bleed (chain button)
IPrintData::SetBleedChain
Bleed: Top
IPrintData::SetBleedTop
Bleed: Bottom
IPrintData::SetBleedBottom
Bleed: Inside
IPrintData::SetBleedInside
Bleed: Outside
IPrintData::SetBleedOutside
Include Slug Area
IPrintData::SetIncludeSlug
Printing
Print user interface
Print dialog box: Output panel
FIGURE 205
Print dialog box: Output panel
The user-interface elements on the Output selectable panel are mapped to the API methods
listed in Table 102.
TABLE 102 Print dialog box: Output panel
Printing
User-interface element
API method
Color
IPrintData::SetOutputMode, using
IPrintData::kCompositeLeaveUnchanged,
IPrintData::kCompositeGray, IPrintData::kCompositeRGB,
IPrintData::kCompositeCMYK,
IPrintData::kSeparationBuiltIn, or
IPrintData::kSeparationInRIP
Text as Black
IPrintData::SetPrintColorsInBlack
Trapping
IPrintData::SetTrappingMode, using
IPrintData::kTrappingNone, IPrintData::kTrappingBuiltIn,
or IPrintData::kTrappingInRIP
429
Printing
Print user interface
430
User-interface element
API method
Flip
IPrintData::SetFlipMode, using IPrintData::kFlipOff,
IPrintData::kFlipHorizontal, IPrintData::kFlipVertical, or
IPrintData::kFlipBoth
Negative
IPrintData::SetNegative
Screening
When the IPrintData::GetOutputMode is a separation mode,
IPrintData::SetSeparationScreenText; when the
IPrintData::GetOutputMode is a composite mode,
IPrintData::SetCompositeScreenText.
Frequency
If output mode is kCompositeGray and the composite-screen
mode is string key “kCustom” (translates differently in each
locale), IPrintData::SetCompositeFrequency.
Angle
If output mode is kCompositeGray and the composite-screen
mode is string key “kCustom” (translates differently in each
locale), IPrintData::SetCompositeAngle.
Simulate Overprint
IPrintData::SetSpotOverPrint, using
IPrintData::kSimulatePress (if the radio button is checked) or
IPrintData::kLegacy (if the radio button is not checked).
Inks
Stored temporarily in
IPrintDialogData::SetNthInkScreening. When the user clicks
the OK button in the Ind Manager dialog box, the
kChangeInkCmdBoss command is processed for each ink
listed.
Printing
Print user interface
Print dialog box: Graphics panel
FIGURE 206
Print dialog box: Graphics panel
The user-interface elements on the Graphics selectable panel are mapped to the API methods
listed in Table 103.
TABLE 103 Print dialog box: Graphics panel
Printing
User-interface element
API method
Images: Send Data
IPrintData::SetImageData, using IPrintData::kImageDataAll,
IPrintData::kImageDataOptimized,
IPrintData::kImageDataLoRez, or
IPrintData::kImageDataProofPrint
Fonts: Download
IPrintData::SetFontDownload,
usingIPrintData::kFontDownloadNone,
IPrintData::kFontDownloadComplete,
IPrintData::kFontDownloadSubset, or
IPrintData::kFontDownloadSubsetLrg
Download PPD Fonts
IPrintData::SetDownloadPPDFonts
PostScript
IPrintData::SetPSLangLevel, using
IPrintData::kPSLangLevel_2 or IPrintData::kPSLangLevel_3
431
Printing
Print user interface
User-interface element
API method
Data Form at
IPrintData::SetImageDataFormat, using
IPrintData::kImageDataBinary or
IPrintData::kImageDataASCII
Print dialog box: Color Management panel
FIGURE 207
Print dialog box: Color Management panel
The user-interface elements on the Color Management selectable panel are mapped to the API
methods listed in Table 104.
TABLE 104 Print dialog box: Color Management panel
432
User-interface element
API method
Print: Document
IPrintData::SetSourceSpace(IPrintData::kDocumentSourceS
pace)
Print: Proof
IPrintData::SetSourceSpace(IPrintData::kProofSourceSpace)
Color Handling
IPrintData::SetProfileType, using
IPrintData::kUseDocumentProfile,
IPrintData::kUsePostScriptCMS, or
IPrintData::kUseNoCMS
Printing
Print user interface
User-interface element
API method
Printer Profile
IPrintData::SetProfileType, using
IPrintData::kUseDocumentProfile or
IPrintData::kUseWorkingProfile. If this is not one of the
predefined profiles, use IPrintData::kUseSpecificProfile and
then call IPrintData::SetProfileName, specifying the name of
the profile as displayed in the drop-down list.
Preserve CMYK Colors
IPrintData::SetPreserveColorNumbers
Simulate Paper Color
IPrintData::SetIntent, using
IPrintData::kRelativeColorimetric (if the radio button is
unselected) or IPrintData::kAbsoluteColorimetric (if the
radio button is selected).
Print dialog box: Advanced panel
FIGURE 208
Print dialog box: Advanced panel
The user-interface elements on the Advanced selectable panel are mapped to the API methods
listed in Table 104.
Printing
433
Printing
Print user interface
TABLE 105 Print dialog box: Advanced panel
User-interface element
API method
OPI Image
Replacement
IPrintData::SetOPIReplacement
Omit for OPI: EPS
IPrintData::SetOmitEPS
Omit for OPI: PDF
IPrintData::SetOmitPDF
Omit for OPI: Images
IPrintData::SetOmitImages
Transparency
Flattener: Preset
IPrintData::SetFlattenerStyleName
Transparency
Flattener: Ignore
Spread Overrides
IPrintData::SetIgnoreSpreadOverrides
Print dialog box: Summary panel
FIGURE 209
Print dialog box: Summary panel
When you click the Save Summary button, the Save File dialog box opens, asking for the path
of a text file to which to save the summary. If you call IPrStStyleListMgr::GetNthStyleDescrip-
434
Printing
Printing extension patterns
tion for a selected printer style, you get the same text as shown in the Summary multi-line text
widget.
Print Presets dialog box
The main Print Presets dialog box is opened when the user chooses File > Print Presets >
Define.
In the main Print Presets dialog box, the user can see the currently defined print presets and a
text summary of the currently selected preset. The main dialog box also has buttons for creating a new print preset (New), editing an existing print preset (Edit), deleting the selected print
preset (Delete), loading from a print presets file (Load), and saving the print presets to a file
(Save).
When the user clicks the New or Edit button, a selectable dialog box similar to the Print dialog
box opens. The design of the Print Presets selectable dialog box is mostly the same as the Print
dialog box, with a few minor differences:
z
The title of the dialog box differs based on what the end user is doing.
z
The thumbnail image on the bottom-left corner of the dialog box is shown only in the Print
dialog box and is disabled in the Print Presets dialog box.
z
The Print dialog box has a Save Preset button, which allows the user to save the current
print settings as a print-preset style in the workspace. The Print Presets dialog box does not
have that button, since the Print Presets dialog box is where you edit the print-preset style.
Extending the Print dialog box or the Print Presets selectable dialog box
The Print dialog box and the Print Presets selectable dialog boxes are extensible by third-party
plug-ins by means of custom, selectable, dialog panel. By implementing one selectable dialog
panel, you can add your own panel into both the Print dialog box and the Print Presets selectable dialog box. For a discussion on SDK sample plug-ins that implement selectable dialog
panels, see “Adding your own panel to the Print and Print Presets dialog boxes” on page 446.
Printing extension patterns
Print-setup provider
A plug-in can register a print-setup service boss by providing an IK2ServiceProvider implementation that supports the kPrintSetupService service ID. The service is called at various
points during the print-action sequence (see “The print action sequence” on page 421). The
implementation for IPrintSetupProvider provides methods to set up or change print parameters before and during the printing process. See Table 106 for implementation details.
Printing
435
Printing
Printing extension patterns
TABLE 106 Implementation recipe for a print-setup provider
Purpose
Participate in various phases in the print process.
ServiceID
kPrintSetupService
Required
companion
interface
IPrintSetupProvider
Boss class
IID_IK2SERVICEPROVIDER with implementation ID
kPrintSetupServiceImpl (see PrintID.h) and
IID_IPRINTSETUPPROVIDER with your implementation ID
When
called
Methods of IPrintSetupProvider are called at various phases of the print
action command. See “The print action sequence” on page 421.
How called
All providers of this ServiceID get called.
Sample
code
See “Participating in the stages of the print-action sequence” on
page 443.
Print-insert-PostScript proc provider
A plug-in can inject PostScript statements into the print-output stream by registering an
IK2ServiceProvider implementation supporting the kPrintInsertPSProcService serviceID and
providing an implementation for IPrintInsertPSProcProvider. See Table 107 for implementation details.
TABLE 107 Implementation recipe for a print-insert-PostScript proc provider
436
Purpose
Inject PostScript comments at predetermined phases of the printing
process.
ServiceID
kPrintInsertPSProcService
Required
companion
interface
IPrintInsertPSProcProvider
Boss class
IID_IK2SERVICEPROVIDER with implementation ID
kInsertPSProcServiceImpl (see PrintID.h) and
IID_IPRINTINSERTPSPROCPROVIDER with your implementation ID
Printing
Printing extension patterns
When
called
Methods of IPrintInsertPSProcProvider are called at various phases of
the core print command. See “The print action sequence” on page 421.
The Setup method is called first, to give the provider a chance to store
print settings. Then the GetInsertPSProcName method is called, so the
provider can return a name. (The following steps occur only if your
GetInsertPSProcName implementation returns a non-empty string.) The
GetClientData method gets called to obtain the provider’s custom client
data (if any), then the PrintInsertPSProc method is called at various
phases of the printing process. (For a list of phases, see the enum
IPrintInsertPSProcProvider::DocumentSection.)
How called
All providers of this ServiceID get called.
Sample
code
See “Injecting PostScript comments or extra data into the print stream
during the print action sequence” on page 444.
Print-data helper-strategy provider
A plug-in can control whether to override the current locked state and relevant state of print
data items in the Print dialog box by registering an implementation of IK2ServiceProvider supporting the kPrintDataHelperStrategyService serviceID. The IPrintDataHelperStrategy interface provides two methods: IsLocked and IsRelevant.
IsLocked allows a plug-in to lock the Print and Print Presets dialog-box user-interface elements. Although an item's locked state can be partially controlled using this method, the application print components are still free to change an item's value as necessary to maintain the
print data in a consistent and valid context.
IsRelevant allows a plug-in to disable specific Print and Print Presets dialog-box user-interface
elements. Although an item's relevant state can be partially controlled using this interface, the
application print components are still free to change an item's value as necessary to maintain
the print data in a consistent and valid context.
The most restrictive interface takes precedence. After an ID is set to be locked (the most
restrictive setting), other implementations are not called. After an ID’s relevance is set to kFalse
(the most restrictive setting), other implementations are not called.
See Table 108 for implementation details.
TABLE 108 Implementation recipe for a print-data helper-strategy provider
Printing
Purpose
Influence the Print and Print Presets dialog boxes through the ability to
suppress and lock print user-interface elements.
ServiceID
kPrintDataHelperStrategyService
Required
companion
interface
IPrintDataHelperStrategy
437
Printing
Printing solutions
Boss class
IID_IK2SERVICEPROVIDER with implementation ID
kDataHelperStrategyServiceImpl(see PrintID.h) and
IID_IPRINTDATAHELPERSTRATEGY with your implementation ID
When
called
Whenever IPrintData::IsLocked or IPrintData::IsRelevant is called.
Generally this happens when the Print or Print Presets dialog box is
being prepared for display.
How called
All providers of this ServiceID get called.
Related
sample
code
See “Specifying which parts of the Print and Print Presets dialog boxes
are relevant or locked” on page 447.
Draw-event handlers
A plug-in can register a draw-event handler (IDrwEvtHandler) to participate in various draw
events that happen during the printing process. For more details on draw-event handlers,
including implementation details, see the “Graphics Fundamentals” chapter. For samples that
implement this extension pattern for printing purposes, see “Adding a custom watermark during the printing process” on page 444 and “Injecting PostScript comments or extra data into
the print stream during the print action sequence” on page 444.
Printing solutions
Getting started
Printing a document
Execute (not process) one of the following commands:
z
kPrintActionCmdBoss, to print a document in InDesign
z
kInCopyPrintActionCmdBoss, to print a document in InCopy
The minimal settings required to print a document with kPrintActionCmdBoss in InDesign
are as follows:
z
The document you want to print.
z
The range of pages in the document you want to print.
z
Print user-interface options.
The print-action command is executed (not processed), because there is no undo capability
provided. For a sample that uses either kPrintActionCmdBoss or kInCopyPrintActionCmdBoss, see the SnpPrintDocument.cpp sample code snippet, in particular SnpPrintDocument::DoPrintDocument.
438
Printing
Printing solutions
Printing a Book
Execute the kBookPrintActionCmdBoss command.
Working with print-preset styles
Getting information about print-preset styles
Query for IPrStStyleListMgr on kWorkspaceBoss, then call the Get methods to get information
about the print-preset styles that are registered. IPrStStyleListMgr::GetNumStyles reports the
number of print-preset styles registered. IPrStStyleListMgr::GetNthStyleName reports the
name of the print-preset style at index n. IPrStStyleListMgr::GetNthStyleRef returns the
UIDRef of the print-preset style at index n. Using this UIDRef, you can query IPrintData, and
get further information about the print-preset style. There are other useful methods on IPrStStyleListMgr.
NOTE:
IPrStStyleListMgr.h does not declare any methods; however, it inherits
IGenStlEdtListMgr. See IGenStlEdtListMgr.h for details.
For more details, see the SnpManipulatePrintStyles.cpp code snippet, in particular SnpManipulatePrintStyles::InspectPrintStyle.
Adding a print-preset style
Query for IPrStStyleListMgr on kWorkspaceBoss, then call IPrStStyleListMgr::AddStyle. When
the new style is created, it is appended to the end of the list of print-preset styles.
For details, see the SnpManipulatePrintStyles.cpp code snippet, in particular SnpManipulatePrintStyles::AddPrintStyle.
Duplicating a print-preset style
Query for IPrStStyleListMgr on kWorkspaceBoss, then call IPrStStyleListMgr::CopyNthStyle.
When the new style is created, it is appended to the end of the list of print-preset styles.
Modifying the name of a print-preset style
First, query for IPrStStyleListMgr on kWorkspaceBoss.
Then, call IPrStStyleListMgr::SetNthStyleName. When you do this, however, the name of the
print-preset style stored in IPrintData is not updated, so you must query for IPrintData and call
IPrintData::SetStyleName, using the same name. For details, see the SnpManipulatePrintStyles.cpp code snippet, in particular SnpManipulatePrintStyles::ModifyPrintStyleName.
Modifying the settings of a print-preset style
First, query for IPrStStyleListMgr on kWorkspaceBoss.
Then, call IPrStStyleListMgr::EditNthStyle, which invokes the Print Presets selectable dialog
box. This procedure requires a user to use the dialog box to change settings. For details, see the
SnpManipulatePrintStyles.cpp code snippet, in particular SnpManipulatePrintStyles::ModifyPrintStyle.
Printing
439
Printing
Printing solutions
Deleting a print-preset style
Query for IPrStStyleListMgr on kWorkspaceBoss, then call IPrStStyleListMgr::DeleteNthStyle.
For details, see the SnpManipulatePrintStyles.cpp code snippet, in particular SnpManipulatePrintStyles::DeletePrintStyle.
Exporting a set of print-preset styles to a file
Process the kGenStlEdtExportStylesCmdBoss command. In the IGenStlEdtCmdData data
interface, call SetListMgrIID to set the type of style list to be the print style list, by specifying
the IPrStStyleListMgr::kDefaultIID (IID_IPRSTSTYLELISTMGR), then call SetTargetFile to
specify the full path of the file to save. Optionally, specify a set of style indices you want to
export from IPrStStyleListMgr on kWorkspaceBoss. If you do not specify a set of style indices,
all styles stored in IPrStStyleListMgr are exported.
Importing a set of print-preset styles from a file
Process the kGenStlEdtImportStylesCmdBoss command. In the IGenStlEdtCmdData data
interface, call SetListMgrIID to set the type of style list to be the print style list, by specifying
the IPrStStyleListMgr::kDefaultIID (IID_IPRSTSTYLELISTMGR), then call SetTargetFile to
specify the full path of the file to import.
Getting notified when a print-preset style is imported to/exported from a
file
Implement a signal-responder extension pattern that responds one or more of the following
signals:
z
kBeforeExportStyleSignalResponderService
z
kAfterExportStyleSignalResponderService
z
kBeforeImportStyleSignalResponderService
z
kAfterImportStyleSignalResponderService
These IDs are defined in GenericSettingsID.h. In your implementation of IResponder::Respond, you can query for the IStyleSignalData interface using ISignalMgr::QueryInterface (or InterfacePtr). IStyleSignalData gives you a reference to the file that is the target of
the export or import operation.
NOTE:
440
Any preset style that is managed by the generic-settings framework and can be exported
or imported—such as PDF-export styles and document-preset styles—can be
monitored in the same way.
Printing
Printing solutions
Working with trap styles
Getting information about trap styles
First, query for ITrapStyleListMgr on one of the following bosses:
z
kWorkspaceBoss, for trap styles registered in the workspace (also used as document
defaults when a new document is created)
z
kDocWorkspaceBoss, for trap styles registered in a document
z
kBookBoss, for trap styles registered in a book
The utility interface ITrapStyleUtils (on kUtilsBoss) has a QueryTrapStyleListMgr method to
simplify this process. There are two overloaded methods:
z
The one that takes IDocument* returns the ITrapStyleListMgr on kDocWorkspaceBoss
(i.e., the workspace related to the document).
z
The one that takes a UIDRef returns the ITrapStyleListMgr on the same boss as the UIDRef,
or the ITrapStyleListMgr on kWorkspaceBoss if the UID in the UIDRef is kInvalidUID.
Next, call the Get methods to get various information about the trap styles that are registered.
ITrapStyleListMgr::GetNumStyles reports the number of trap styles registered. ITrapStyleListMgr::GetNthStyleName reports the name of the trap style at index n. ITrapStyleListMgr::GetNthStyleRef returns the UIDRef of the trap style at index n. Using this UIDRef, you can query
ITrapStyle and get further information about this trap style. There are other useful methods on
ITrapStyleListMgr.
NOTE:
ITrapStyleListMgr does not inherit from IGenStlEdtListMgr the way IPrStStyleListMgr
does.
For details, see the SnpManipulateTrapStyles.cpp code snippet, in particular SnpManipulateTrapStyles::InspectTrapStyle.
Adding a trap style
Query for ITrapStyleListMgr, then call ITrapStyleListMgr::AddStyle. When the new style is
created, it is appended to the end of the list of trap styles.
For details, see the SnpManipulateTrapStyles.cpp code snippet, in particular SnpManipulateTrapStyles::AddTrapStyle.
Duplicating a trap style
Query for ITrapStyleListMgr, then call ITrapStyleListMgr::CopyNthStyle. When the new style
is created, it i appended to the end of the list of trap styles.
Modifying a trap style
Query for ITrapStyleListMgr, then call ITrapStyleListMgr::EditNthStyle. You are required to
pass in the trap-style data via the parameter list as ITrapStyle. You can create a non-persistent
instance of kTrapStyleBoss to store the settings.
For details, see the SnpManipulateTrapStyles.cpp code snippet, in particular SnpManipulateTrapStyles::ModifyTrapStyle.
Printing
441
Printing
Printing solutions
Deleting a trap style
Query for ITrapStyleListMgr, then call ITrapStyleListMgr::DeleteNthStyle.
For details, see the SnpManipulateTrapStyles.cpp code snippet, in particular SnpManipulateTrapStyles::DeleteTrapStyle.
Exporting a set of trap styles to another trap-style list
First, create a command sequence, as the following procedure processes multiple commands.
Next, query for ITrapStyleListMgr, then call ITrapStyleListMgr::ExportStyles. The first parameter is a UIDRef for the destination trap style list (note the source trap style list is identified by
this particular instance of ITrapStyleListMgr), and that UIDRef must refer to a boss that aggregates ITrapStyleListMgr. If you want to export the trap styles to a file, you can create a new
database using DBUtils::CreateDataBase and IDataBase::New, then put a new root UID by calling IDataBase::NewUID. The class for the new root should be kTrapStyleExportRootBoss. Create a UIDRef for this new root, and pass it into ITrapStyleListMgr::ExportStyles. Once that
completes, save the database by calling IDataBase::SaveAs, and close the database by deleting
the IDataBase pointer.
Importing a set of trap styles from another trap-style list
First, create a command sequence, as the following procedure processes multiple commands.
Next, query for ITrapStyleListMgr, then call ITrapStyleListMgr::ImportStyles. The first parameter is a UIDRef for the source trap style list (note the destination trap style list is identified by
this particular instance of ITrapStyleListMgr), and that UIDRef must refer to a boss that aggregates ITrapStyleListMgr. If you want to import the trap styles from a file, you can open a database using DBUtils::CreateDataBase and IDataBase::Open, then get the root UID by calling
IDataBase::GetRootUID. The class for the new root should be kTrapStyleExportRootBoss. Create a UIDRef for this root, and pass it into ITrapStyleListMgr::ImportStyles. Once that completes, close the database by deleting the IDataBase pointer.
Determining which trap style is associated with a page on a document
Given a specific page (kPageBoss) on a document (which you can query using IPageList on
kDocBoss), query for IPersistUIDData with the specific IID of IID_ITRAPSTYLEUIDDATA.
Get the UID from IPersistUIDData. That UID refers to the trap style registered in ITrapStyleListMgr on the document workspace (kDocWorkspaceBoss). To find out the name of this trap
style, first call ITrapStyleListMgr::GetStyleIndexByUID and then call ITrapStyleListMgr::GetNthStyleName.
Alternately, you can collect a list of pages that use a specific trap style. Call ITrapStyleUtils::GetDocumentTrapStylePageList. For details, see the SnpManipulateTrapStyles.cpp code snippet,
specifically SnpManipulateTrapStyles::InspectTrapStyle.
Associating a trap style with a page on a document
Create a UIDList of pages (kPageBoss) to which you want to associate a trap style, and the
UIDRef of the trap style. Then call ITrapStyleUtils::AssignStyleToPageList. For details, see the
SnpManipulateTrapStyles.cpp code snippet, specifically SnpManipulateTrapStyles::AssignTrapStyleToPages.
442
Printing
Printing solutions
Participating in the print process
Participating in the stages of the print-action sequence
There are various ways to participate in the print-action sequence. The main way is to implement a print-setup provider (see “Print-setup provider” on page 435). By implementing IPrintSetupProvider, your plug-in can be notified at the stages of the print-action sequence described
in “The print action sequence” on page 421.
Other ways to participate in the print-action sequence are noted in the following list. Some of
these require extra hook-ups from within a print setup-provider implementation.
z
Implement a print-insert PostScript proc provider to inject PostScript statements into the
print-output stream. (See “Print-insert-PostScript proc provider” on page 436 and “Injecting PostScript comments or extra data into the print stream during the print action
sequence” on page 444.)
z
Implement a print-data-helper strategy provider to influence the display of the Print and
Print Presets dialog boxes. (See “Print-data helper-strategy provider” on page 437 and
“Specifying which parts of the Print and Print Presets dialog boxes are relevant or locked”
on page 447.)
z
Implement a draw-event handler that handles print-related drawing events. (See “Drawevent handlers” on page 438, “Adding a custom watermark during the printing process” on
page 444, “Specifying which page items should be printed” on page 443, and “Injecting
PostScript comments or extra data into the print stream during the print action sequence”
on page 444).
z
Implement a custom write stream so the print-action sequence writes to it. The custom
stream must be specified from one of the methods in your print-setup provider. (See “Writing printing data to a custom stream” on page 447.)
The following sample plug-ins contain an implementation of a print-setup provider:
z
PrintMemoryStream prints document content to a custom memory stream.
z
PrintSelection adds a flag to the document to print only selected page items.
For details, see the API reference documentation for each sample plug-in.
Specifying which page items should be printed
You can implement a print-setup provider (see “Participating in the stages of the print-action
sequence” on page 443) and a custom draw-event handler to specify which pages items should
be printed. From the print-setup provider (in any method that gets called before the core print
command is executed, the last chance being AfterPrintGatherCmd), you register the custom
draw-event handler that handles the print event only when it encounters document content
you want to print. When the printing is done, you de-register the custom draw-event handler
in your print-setup provider’s EndPrint implementation.
The PrintSelection sample plug-in shows how this is done. Here are the highlights of what the
PrintSelection plug-in does:
Printing
443
Printing
Printing solutions
z
In PrnSelPrintSetupProvider::AfterPrintUI, the currently selected page items are gathered
and the custom draw-event handler is registered for the draw event message kDrawShapeMessage.
z
In PrnSelPrintSetupProvider::BeforePrintGatherCmd, the set of pages to output is modified
based on which pages contain the selected page items.
z
The PrnSelDrawHandler::HandleEvent determines whether the current printable item
should be printed. This draw event handler does not do any drawing; the actual drawing of
the printable item is delegated to other draw-event handlers.
z
In PrnSelPrintSetupProvider::EndPrint, the custom draw-event handler is unregistered.
For details, see the API reference documentation for the PrintSelection plug-in.
Specifying which layer(s) of a document should be printed
In the Setup panel of the Print dialog box, there is a setting called “Print Layers,” which is a
drop-down list that allows the user to choose printing with visible and printable layers, visible
layers, or all layers. This option can be get/set through the IPrintData::GetPrintLayers/SetPrintLayers methods. Each layer’s visibility and printability, however, are managed by layer
options as described in the “Layer Options” section of the “Layout Fundamentals” chapter. To
set the visibility of a layer, use the kShowLayerCmdBoss command. To set the printability of a
layer, use the kPrintLayerCmdBoss command. The SnpPrintDocument.cpp SDK snippet
shows how to use a kPrintLayerCmdBoss. Before processing a kPrintLayerCmdBoss, make
sure the printability of the layer is different from the new state you are going to set. If the new
printability state is the same as the old one, kPrintLayerCmdBoss asserts, although the assert is
benign.
Adding a custom watermark during the printing process
You can implement a custom draw-event handler that draws the watermark on a page or a page
item. This is done in the BasicDrwEvtHandler sample plug-in, a canonical example of a drawevent handler. For details, see the API reference documentation for the BasicDrwEvtHandler
plug-in.
Another way to add custom watermarks to page items is to implement a page-item adornment
that draws when the draw flag has the IShape::kPrinting bit set. While the FrameLabel sample
plug-in does not support printing (i.e., if flags contains IShape::kPrinting, the Draw method
breaks out), you can use the FrameLabel sample as a basis for implementing custom watermarks and extra persistent data on page items. For details, see the API reference documentation for the FrameLabel plug-in.
Injecting PostScript comments or extra data into the print stream during
the print action sequence
You can implement a print-insert PostScript proc provider (see “Print-insert-PostScript proc
provider” on page 436) to inject PostScript comments during predetermined phases of the
print-output process as driven by the core print command. The PrintMemoryStream sample
plug-in contains an implementation of a print-insert PostScript proc provider. The print-insert
PostScript proc provider implementation in this sample is a canonical example that demonstrates when each method in IPrintInsertPSProcProvider gets called by writing a trace mes-
444
Printing
Printing solutions
sage. This plug-in also demonstrates how to manage custom print settings throughout the
print-action sequence.
There are other ways to inject PostScript comments into the print stream during the print
action sequence:
z
Implement a custom draw-event handler that calls the IGraphicsPort::AddComment
method. For a sample of a draw-event handler that responds to the kDrawShapeMessage for
printing, see PrnSelDrawHandler::HandleEvent in the PrintSelection sample plug-in. To
obtain IGraphicsPort, you can first get a pointer to the GraphicsData from DrawEventData::gd, then call GraphicsData::GetGraphicsPort.
z
Add custom registration marks on each page, if you are using the Japanese feature set of
InDesign or InCopy. See “Japanese page-mark files” on page 449.
Adding custom print settings so they are managed like other
print settings
Old method
First, write your own interface that allows you to manipulate your custom print-settings data
(e.g., IMyPrintData). Then, the simplest way to persist your custom print settings is to add a
persistent implementation of this interface (e.g., kMyPrintDataImpl) on a boss class that also is
persisted with the document, like kDocWorkspaceBoss or kDocBoss. By adding your implementation only in the specified boss classes, however, you will not be able to manage your custom print settings together with print-preset styles.
In addition, there are several places where you must copy your custom print-setting data, as
new instances of the aforementioned bosses are created (during processes like the print-action
sequence and the Print Presets dialog) for the purpose of adding a print preset style or keeping
a temporary instance for use during the print-action sequence. For print-data settings stored in
IPrintData, the application calls IPrintData::CopyData to copy only the data managed within
the IPrintData implementation to these new or temporary instances of the print data boss;
however, your custom print settings are not copied at that time. Furthermore, the commands
that copy the IPrintData to the appropriate instance print data boss do not notify any subjects.
To solve these problems, you can implement a special observer. This observer observes changes
that notify on the IID_ICOMMANDMGR protocol on a command’s target database. By knowing which commands copy IPrintData and when the data is copied, you can manage your custom print settings as the same time.
New method
Beginning in CS4, a new kPrintCopyCustomDataService is available; any plug-in that registers
for this service gets a chance to manage its own print data during the print process. This is the
preferred method for managing your own custom print data, instead of using the commandobserver technique discussed above. Note the service provider is called after the application’s
CopyData is done. You should provide the implementation for IPrintCopyCustomDataProvider and aggregate it on the service-provider boss that registered as kPrintCopyCustomDataService. Then your IPrintCopyCustomDataProvider will get called whenever
IPrintData::CopyData is called.
Printing
445
Printing
Printing solutions
Adding your own panel to the Print and Print Presets dialog boxes
You can implement a panel that adds itself to the dialog box identified by kPrintSelectableDialogService. The panel should be 400 pixels wide and 345 pixels high. The ODFRez widget type
should inherit from PrimaryResourcePanelWidget, and the boss class that contains the necessary implementations (see below) should inherit from kPrimaryResourcePanelWidgetBoss.
The required implementations in this boss class are as follows:
z
IK2ServiceProvider (IID_IK2SERVICEPROVIDER), to register the ServiceID values specified in the IPanelCreator::GetServiceIDs method. You do not need to implement this yourself; you can use the implementation ID provided by the application:
kDialogPanelServiceImpl (defined in WidgetID.h).
z
IPanelCreator (IID_IPANELCREATOR), to specify the resource IDs in your ODFRez
resource file (.fr) that declares the selectable dialog ServiceID and the panel resource IDs
provided by the plug-in. Using this implementation, the IK2ServiceProvider in this boss
(kDialogPanelServiceImpl) can get the necessary data to register the panel into the appropriate selectable dialog box. The implementation of this interface must inherit from CPanelCreator.
z
IDialogController (IID_IDIALOGCONTROLLER), to initialize the panel, validate the settings on the panel when the OK button is clicked, apply the settings on the panel when the
OK button is clicked, and perform any clean-up if the user cancels the dialog. The implementation of this interface may inherit from CDialogController.
z
IObserver (IID_IOBSERVER), to handle user-interface actions for widgets on the panel, as
well as widgets on the parent selectable dialog. (To get the widget IDs of the widgets on the
parent selectable dialog box, choose QA > Panel Edit Mode in the debug build of the application.) The implementation of this interface must inherit from AbstractDialogObserver
(as opposed to CSelectableDialogObserver, which is used in the BasicSelectableDialog sample plug-in). See the API reference documentation for AbstractDialogObserver.
In addition to the panel user-interface definition and the supporting boss class, you must
include the following resources:
z
IDList — Specifies the selectable-dialog service ID, which in this case is kPrintSelectableDialogService.
z
IDListPair — Specifies the selectable-dialog service ID (again, kPrintSelectableDialogService), the resource ID of the panel you want to add to the selectable dialog box, and the
PluginID that owns the panel. You can specify multiple panels in this resource.
Both these resources should have the resource ID you specified in your IPanelCreator::GetPanelRsrcID implementation. For example, if the resource ID you specify in IPanelCreator::GetPanelRsrcID is kSDKDefIDListPairResourceID, the resource ID of the panel is
kSDKDefPanelResourceID, and the ID of the plug-in that provides this panel is
kPrtHokUIPluginID, the resources would be written like Example 25.
446
Printing
Printing solutions
EXAMPLE 25 An IDList and IDListPair to support the kPrintSelectableDialogService
resource IDList (kSDKDefIDListPairResourceID)
{
{
kPrintSelectableDialogService,
},
};
resource IDListPair (kSDKDefIDListPairResourceID)
{
{
kPrintSelectableDialogService, kSDKDefPanelResourceID, kPrtHokUIPluginID,
},
};
To find out whether the panel is being opened from the Print or Print Presets dialog box, you
can query the IPrintDialogData interface from the parent selectable dialog (by using the help of
IWidgetParent), and call IPrintDialogData::GetFlags. See Example 26.
EXAMPLE 26 Getting IPrintDialogData flags from a print-dialog selectable-panel implementation
// 'this' maybe a dialog controller or observer
InterfacePtr<IWidgetParent> widgetParent(this, UseDefaultIID());
InterfacePtr<IPrintDialogData> printDialogData
((IPrintDialogData*)widgetParent->QueryParentFor(IID_IPRINTDIALOGDATA));
printFlags = printDialogData->GetFlags();
If the returned value has the IPrintDialogData::kWorkingOnStyle bit set (test by doing a logical
AND on the value with IPrintDialogData::kWorkingOnStyle), the panel is in the Print Presets
dialog box.
A canonical implementation of a selectable dialog box (along with children panels) is provided
in the BasicSelectableDialog sample plug-in. You can use BasicSelectableDialog to understand
the interactions between the parent selectable dialog box and its child panels. For more details,
see the API reference documentation associated with the BasicSelectableDialog sample plugin.
Specifying which parts of the Print and Print Presets dialog boxes are
relevant or locked
You can implement a print-data helper strategy provider (see “Print-data helper-strategy provider” on page 437). Print-data helper strategy provider also are called when the Print or Print
Presets dialog box gets its summary text. (That is, if a setting is not relevant, it is not included in
the summary text.) For a sample implementation of IPrintDataHelperStrategy, see the PrintMemoryStream sample plug-in.
Writing printing data to a custom stream
You can write a class that implements IPMStream (and inherits CStreamWrite), along with a
class that inherits IXferBytes, to copy the data to whatever your stream targets. Your stream
may target things like a file, a memory buffer, or even a database. This custom stream must be
reported to the print-action sequence by calling IOutputPages::SetOutputStream. You can do
this from one of the methods in your print-setup provider (see “Print-setup provider” on
Printing
447
Printing
Bosses that aggregate IPrintData
page 435 and “Participating in the stages of the print-action sequence” on page 443), like
IPrintSetupProvider::BeforePrintGatherCmd. For a sample implementation, see the PrintMemoryStream sample plug-in.
Bosses that aggregate IPrintData
z
kBookBoss stores the Custom print settings for a book.
z
kBookPrintDataBoss stores the print settings when printing a book with the print commands.
z
kDocWorkspaceBoss stores the Custom print settings for a document.
z
kInCopyTempPrintDataBoss stores the print settings for temporary use only (InCopy
only).
z
kPrintDataBoss stores the print settings when printing a document with the print commands. This is the most common boss class for storing print settings during the print process.
z
kPrintDataOnlyBoss stores the print settings for temporary use only.
z
kPrStStyleBoss stores the print settings for a particular style in the list of defined print preset styles. (See “Print preset styles” on page 420.)
z
kStylePrintDataBoss stores the print settings for temporary use when generating the
human-readable text summary of a print-preset style. This text is displayed in the Print Presets dialog box.
For details on these bosses, such as other aggregated interfaces or boss hierarchy, see the API
reference documentation.
Print-action and supporting commands
Table 109 lists the main command bosses used in the print process.
TABLE 109 Command bosses used in the print process
448
Command boss
Description
kBookPrintActionCmdBoss
Top-level print-action command for printing a book. For most clients, this is
the command to execute to print a book in InDesign. It processes the
following supporting commands: kPrintDialogCmdBoss,
kBookSavePrintDataCmdBoss, kPrintGatherDataCmdBoss, and
kNewPrintCmdBoss.
kBookSavePrintDataCmdBoss
Saves the print data into a book.
Printing
Japanese page-mark files
Command boss
Description
kCreatePrintGalleyViewCmdBoss
Creates a galley window view to print.
kInCopyNewPrintCmdBoss
Performs the actual output of a document in InCopy.
kInCopyPrintActionCmdBoss
Top-level print-action command for printing in InCopy. For most clients,
this is the command to execute to print a document in InCopy. It processes
the following supporting commands: kCreatePrintGalleyViewCmdBoss,
kInCopyPrintDialogCmdBoss, kInCopyNewPrintCmdBoss, and
kPrintGatherDataCmdBoss.
kInCopyPrintDialogCmdBoss
Displays the print dialog box in InCopy.
kNewPrintCmdBoss
Performs the actual output of a document or a book to the output device in
InDesign.
kPrintActionCmdBoss
Top-level print-action command for printing a document. For most clients,
this is the command to execute to print a document in InDesign. It processes
the following supporting commands: kPrintDialogCmdBoss,
kPrintSavePrintDataCmdBoss, kPrintGatherDataCmdBoss, and
kNewPrintCmdBoss.
kPrintDialogCmdBoss
Displays the print dialog box in InDesign.
kPrintGatherDataCmdBoss
Gathers print data.
kPrintSavePrintDataCmdBoss
Saves the print data into a document.
Japanese page-mark files
The Japanese feature set provides two extra Type options in the Marks And Bleed panel in the
Print dialog box and Print Presets selectable dialog box:
z
Maru-tsuki Sentaa-tombo — The untranslated string key “kJMarksWithCircle” is passed
into IPrintData::SetPageMarkFile.
z
Maru-nashi Sentaa-tombo — The untranslated string key “kJMarksWithoutCircle” is passed
into IPrintData::SetPageMarkFile.
In addition (with either the Roman or Japanese feature set), you can specify a filename (without path or extension) of a file that contains a custom page mark written using PostScript. The
contents of this PostScript file are injected during printing. The actual file should have an .mrk
extension and be in one of the following locations:
Printing
(Windows)
C:\Program Files\Common Files\Adobe\PrintSpt
(Mac OS)
/Library/Application Support/Adobe/PrintSpt
449
Printing
Exporting to EPS and PDF
Exporting to EPS and PDF
The process of exporting to EPS and PDF file formats is somewhat similar to the process of
printing, in that similar components of the application are employed. The way you drive the
export process and the data model behind the EPS and PDF export features are somewhat different. This section provides highlights of how to drive the EPS and PDF export features in the
application.
Exporting to EPS
EPS export preferences
Settings for EPS export are stored in the IEPSExportPreferences interface on kWorkspaceBoss.
For details, see the API reference documentation.
To get the preferences, query for IEPSExportPreferences on kWorkspaceBoss and call its Get
methods.
To modify the preferences, process the kSetEPSExportPrefsCmdBoss command. This command has a data interface, IEPSExportPrefsCmdData, with which you specify the new values
of the EPS preferences. Before you call any Set methods, you can call IEPSExportPrefsCmdData::CopyPrefs to copy the current preference settings in IEPSExportPreferences.
User interface
When you choose File > Export and specify the file to export and the export format to be EPS,
you will see a selectable dialog box with two panels, General and Advanced. The user-interface
elements on these panels are mapped to the methods on IEPSExportPreferences, as described
below.
450
Printing
Exporting to EPS and PDF
Export EPS dialog box: General panel
FIGURE 210
Printing
Export EPS dialog box: General panel
z
All Pages — IEPSExportPreferences::SetEPSExPageOption using IEPSExportPreferences::kExportAllPages
z
Ranges (radio button) — IEPSExportPreferences::SetEPSExPageOption using IEPSExportPreferences::kExportRanges
z
Ranges (text-edit box) — IEPSExportPreferences::SetEPSExPageRange
z
Spreads — IEPSExportPreferences::SetEPSExReaderSpread using IEPSExportPreferences::kExportReaderSpreadOFF or IEPSExportPreferences::kExportReaderSpreadON
z
PostScript — IEPSExportPreferences::SetEPSExPSLevel using
ences::kExportPSLevel2 or IEPSExportPreferences::kExportPSLevel3
z
Color — IEPSExportPreferences::SetEPSExColorSpace, using IEPSExportPreferences::kExportPSColorSpaceLeaveUnchanged, IEPSExportPreferences::kExportPSColorSpaceDIC,
IEPSExportPreferences::kExportPSColorSpaceCMYK, IEPSExportPreferences::kExportPSColorSpaceGray, or IEPSExportPreferences::kExportPSColorSpaceRGB
z
Preview — IEPSExportPreferences::SetEPSExPreview using IEPSExportPreferences::kExportPreviewNone, IEPSExportPreferences::kExportPreviewTIFF, or IEPSExportPreferences::kExportPreviewPICT (Mac OS only)
z
Embed Fonts — IEPSExportPreferences::SetEPSExIncludeFonts using IEPSExportPreferences::kExportIncludeFontsNone,
IEPSExportPreferences::kExportIncludeFontsWhole,
IEPSExportPreferences::kExportIncludeFontsSubset, or IEPSExportPreferences::kExportIncludeFontsSubsetLarge
IEPSExportPrefer-
451
Printing
Exporting to EPS and PDF
z
Data Format — IEPSExportPreferences::SetEPSExDataFormat, using IEPSExportPreferences::kExportASCIIData or IEPSExportPreferences::kExportBinaryData
z
Bleed: Top — IEPSExportPreferences::SetEPSExBleedTop
z
Bleed: Bottom — IEPSExportPreferences::SetEPSExBleedBottom
z
Bleed: Inside — IEPSExportPreferences::SetEPSExBleedInside
z
Bleed: Outside — IEPSExportPreferences::SetEPSExBleedOutside
z
If any Bleed settings are over 0 — IEPSExportPreferences::SetEPSExBleedOnOff using
either IEPSExportPreferences::kExportBleedON, or IEPSExportPreferences::kExportBleedOFF
Export EPS dialog box: Advanced panel
FIGURE 211
452
Export EPS dialog box: Advanced panel
z
Send Data — IEPSExportPreferences::SetEPSExBitmapSampling, using IEPSExportPreferences::kExportBMSampleNormal or kExportBMSampleLowRes
z
OPI Image Replacement — IEPSExportPreferences::SetEPSExOPIReplace, using IEPSExportPreferences::kExportOPIReplaceON or IEPSExportPreferences::kExportOPIReplaceOFF
z
Omit for OPI: EPS — IEPSExportPreferences::SetEPSExOmitEPS, using IEPSExportPreferences::kExportOmitEPSON or IEPSExportPreferences::kExportOmitEPSOFF
z
Omit for OPI: PDF — IEPSExportPreferences::SetEPSExOmitPDF, using IEPSExportPreferences::kExportOmitPDFON or IEPSExportPreferences::kExportOmitPDFOFF
Printing
Exporting to EPS and PDF
z
Omit for OPI: BitMap Images — IEPSExportPreferences::SetEPSExOmitBitmapImages,
using IEPSExportPreferences::kExportOmitBitmapImagesON or IEPSExportPreferences::kExportOmitBitmapImagesOFF
z
Preset — IEPSExportPreferences::SetEPSExFlattenerStyle, using a UID of a kXPFlattenerStyleBoss object (use IFlattenerStyleListMgr to find one)
z
Ignore Spread Overrides — IEPSExportPreferences::SetEPSExIgnoreFlattenerSpreadOverrides, using IEPSExportPreferences::kExportIgnoreSpreadOverridesON or IEPSExportPreferences::kExportIgnoreSpreadOverridesOFF
Exporting
Using IK2ServiceRegistry, query for service providers supporting kExportProviderService.
Find an export provider that can support the format name “EPS,” by iterating over all export
providers and calling IExportProvider::CountFormats and IExportProvider::GetNthFormatName, or IExportProvider::CanExportThisFormat. Once you find the export provider for EPS,
do the following:
z
Call IExportProvider::CanExportToFile to check whether the document and current selection target can be exported.
z
Optionally, query for IBoolData, set it to kTrue to set IPrintContentPrefs, and query for
IPrintContentPrefs and call its Set methods.
z
Call IExportProvider::CanExportToFile or IExportProvider::ExportToStream to do the
export.
Exporting to PDF
The structure of the PDF-export architecture is very similar to that of EPS export; however, you
can specify a greater level of detail when exporting a document to PDF. For details of exporting
to PDF, see the “PDF Import and Export” chapter.
Printing
453
Printing
Exporting to EPS and PDF
454
PDF Import and Export
PDF import
PDF Import and Export
PDF import
This section focuses on how you can control PDF import preferences during the PDF import
process.
Importing a file into InDesign requires a series of commands to be processed in a certain
sequence. Before a PDF file can be imported, a page item needs to be created to hold the PDF
content. Also, a proper datalink needs to be set up, so the association between the page item
and the external file can be established.
Figure 212 is a high-level illustration of the import process in InDesign for a standalone desktop PDF file (not a workgroup file) to be loaded into the place gun. The details of a complete
import process is beyond the scope of this document. We recommend you always try to import
(or place) a file using the method illustrated in the SnpPlaceFile.cpp snippet sample, which
uses kImportAndPlaceCmdBoss to place the file on the active spread. You also can use kImportAndLoadPlaceGunCmdBoss to load the file into the place gun. Both commands check
with all the import providers and pick the appropriate provider for the type of file to import.
The PDF import provider (kPDFPlaceProviderBoss) is the default InDesign import provider
for PDF files. kPDFPlaceProviderBoss processes kPDFPlaceCmdBoss, which is the central
command that does the real work of importing a PDF file.
PDF Import and Export
455
PDF Import and Export
PDF import
FIGURE 212
High-level view of importing a desktop PDF file
fail
succeed
Receive an IDFile to import
Import PDF
Create a data link
from the IDFile
fail
Open an IPMStream
from the data link for import
Delete the page item
succeed
Update user's PDF import pref
based on the pref just used
Add the data link to the page item
Display PDF import option if necessary
Release the datalink
Release the datalink
Create a new page item
Make sure the page item
is in a graphic frame
Load the new item in the place gun
When you import a PDF file through the InDesign Place dialog box, if you choose Show
Import Options, the Place PDF dialog opens, as shown in Figure 213. The following options
offer flexible PDF import:
z
456
Placing multipage PDF pages (General tab) — You can place a range of PDF pages from a
multipage PDF file. Under Pages, select All or enter a value for Range. The place gun is
loaded as before, but the pointer changes to show that multiple pages exist. Each time you
click, a PDF page is placed on the page. If you hold down the Alt/Option keys, the pointer
changes to the cascade-place pointer, and clicking places all remaining pages on the page,
cascading down from the click point.
PDF Import and Export
PDF import
z
Optional content groups (OCG) layers support (Layers tab) — OCG groups content for
selective viewing and printing, supports complex mapping of objects to groups, and allows
special use cases like language and zoom factors. By recognizing OCG constructs in placed
PDF files, InDesign offers users the ability to selectively include those pieces of content.
FIGURE 213
PDF import options
General PDF import options are stored in the IPDFPlacePrefs interface, and layers options are
stored in IGraphicLayerInfo; both interfaces are added to the kWorkspaceBoss workspace.
Normally, these options are used unless the PDF import is performed through a link update; in
that case, the IGraphicLayerInfo on the page item (kPlacedPDFItemBoss) is used instead, and
some attributes from the IPDFAttributes aggregated on the kPlacedPDFItemBoss are used. The
following attributes override the defaults under the condition of a link update:
z
IPDFAttributes::SetPage — Overrides PDF page number of the placed page.
z
IPDFAttributes::SetTransparentBackground — Overrides Transparency background.
z
IPDFAttributes::SetPreserveScreens — Overrides Preserve Halftone Screens.
z
IPDFAttributes::SetUserPassword — Overrides Set User Password.
z
IPDFAttributes::SetCropTo — Overrides Set Crop To.
Figure 214 shows how kPDFPlaceProviderBoss sets up the data interfaces for kPDFPlaceCmdBoss in different cases. If the import is done with the full user interface, the user can change
settings through the Place PDF dialog (see Figure 213). InDesign uses kPDFPlaceProviderBoss
to do its PDF import, so by changing import options you control InDesign’s PDF import
behavior.
Figure 214 has a “Minimal UI” decision point. If the user deselects Show Import Options when
importing a file, a minimal user interface is used: the user sees only a progress bar, not the
Place PDF dialog.
Figure 214 also has an “Is the PDF import for link update?” decision point. Link update is performed when the user selects Relink or Update from the Links panel. Doing this triggers a PDF
import if the link is a PDF item.
PDF Import and Export
457
PDF Import and Export
PDF import
FIGURE 214
kPDFPlaceProviderBoss import option use
Is the import from clipboard?
No
Yes
Get session prefs: IPDFPlacePrefs
from IID_IPDFPLACEPREFS
Get session prefs: IPDFPlacePrefs
from IID_IPDFCLIPBOARDPLACEPREFS
Process
kPDFPlaceCmdBoss
Get the session preference:
IGraphicLayerInfo
If the import was successful,
and we are NOT importing from
clipboard, and we are NOT in
active INX context, then update
the user’s PDF place preferences
based on what they just chose using
kSetPDFPlacePrefsCmdBoss
Copy session preference data to
kPDFPlaceCmdBoss data interface
Is the PDF import
for link update?
Yes
No
Does the linked page
item have IPDFAttributes
Yes
No
Suppress UI?
Yes
No
Override some session data
with linked page item's
IPDFAttributes
Override the PDF password
setting from the source PDF
Minimal UI?
Yes
No
Bring up PDF Import
Dialog if available
458
PDF Import and Export
PDF export
The main interfaces to control the PDF import options are IPDFPlacePrefs, IGraphicLayerInfo,
and IPDFAttributes. To modify the data stored in these interfaces, use the commands shown in
Table 110. Note there is no command to modify the IPDFPlacePrefs for clipboard import
(IID_IPDFCLIPBOARDPLACEPREFS). If you process your own kPDFPlaceCmdBoss, use the
IPDFPlacePrefs aggregated on the command boss to set up the import option. The IGraphicLayerInfo on the page item passed to the kPDFPlaceCmdBoss is used for layer options.
TABLE 110 Commands to Change PDF Import Options
To change...
Use
IPDFPlacePrefs on the workspace
kSetPDFPlacePrefsCmdBoss
IGraphicLayerInfo on the workspace
kSetPDFPlacePrefsCmdBoss
IPDFAttributes on the page item
kSetPDFAttributesCmdBoss
IGraphicLayerInfo on the page item
kSetGraphicLayerInfoCmdBoss
PDF export
InDesign/InCopy document export
The easiest way to export an InDesign or InCopy document is to use the PDF export service
provider. Example 27 is a snippet showing how to get the PDF export provider and call the
export method.
EXAMPLE 27 Getting the PDF export provider
PMString PDFFormat("Adobe PDF");
PDFFormat.SetTranslatable(kFalse);
InterfacePtr<ISelectionManager> selection(SelectionUtils::QueryActiveSelection());
IDocument *frontDoc = ::GetFrontDocument();
InterfacePtr<IK2ServiceRegistry> k2ServiceRegistry(gSession, UseDefaultIID());
// Look for all service providers with kExportProviderService.
int32 exportProviderCount = k2ServiceRegistry>GetServiceProviderCount(kExportProviderService);
// Iterate through them.
bool found = kFalse;
for (int32 exportProviderIndex = 0 ; exportProviderIndex < exportProviderCount ;
exportProviderIndex++)
PDF Import and Export
459
PDF Import and Export
PDF export
{
// get the service provider boss class
InterfacePtr<IK2ServiceProvider> k2ServiceProvider
(k2ServiceRegistry->QueryNthServiceProvider(kExportProviderService,
exportProviderIndex));
// Get the export provider implementation itself.
InterfacePtr<IExportProvider> exportProvider(k2ServiceProvider,
IID_IEXPORTPROVIDER);
// Check to see if the current selection specifier can be exported by this
provider.
bool16 canExportByTarget = exportProvider->CanExportThisFormat(frontDoc,
selection, PDFFormat);
if (canExportByTarget)
{
found = kTrue;
// assume idFile is a valid IDFile to hold the soon to be created PDF
exportProvider->ExportToFile(idFile, frontDoc, selection, PDFFormat, kFullUI);
}
if (found)
break;
}
In the example, note the following:
z
The format name used to get the PDF export provider is Adobe PDF, defined in the beginning of the example.
z
The example tells the PDF export provider to use the full user interface during the export,
meaning the PDF export settings dialog is brought up for the user to set export options. The
settings dialog box is provided by kPDFExportDialogBoss in InDesign and by kInCopyPDFExptDialogBoss in InCopy. They are selectable dialog service providers you can retrieve
through the IK2ServiceRegistry::GetServiceProviderIndex method, although it is most
practical to specify kFullUI and let the export provider do the work. If kSuppressUI is used
instead, no user interface is presented, and you can change the preference as described
below.
z
The example omits the steps that produce a valid IDFile to be passed into the IExportProvider::ExportToFile method. Normally, the IDFile is obtained through a standard Save File
dialog box. You also can construct an IDFile from scratch. For more information on IDFile,
see the “Using Adobe File Library” chapter.
z
The example is for exporting an InDesign/InCopy document. To export an InDesign book,
see “InDesign book export” on page 468.
The PDF export provider for InDesign/InCopy documents is represented by kPDFExportBoss.
There is an IBoolData in kPDFExportBoss, which indicates whether print content preferences
(IPrintContentPrefs) should be considered. When you export from the InDesign menu, the
default value for the IBoolData is set to false, so no print content preference are used during the
default PDF export. Also, the PDF export provider allows you to use an export style (kPDFExportStyleBoss) instead of dealing with the settings one by one. This is achieved by passing the
style UID in the IUIDData aggregated on the kPDFExportBoss.
The kPDFExportBoss processes kPDFExportCmdBoss. Before the command is processed,
kExportValidationCmdBoss is used to verify that the attributes set for the command will pro-
460
PDF Import and Export
PDF export
duce a valid output. Table 111 summarizes the command interfaces kPDFExportBoss needs to
set up before it processes kPDFExportCmdBoss.
TABLE 111 Critical data interfaces for kPDFExportCmdBoss
Interface
Purpose
IPDFExportPrefs
Holds most of the PDF export settings as seen in the PDF
export options dialog.
IPDFSecurityPrefs
Holds the PDF security preference.
ISysFileData
Holds the destination IDFile for the PDF document.
IBoolData
Indicates whether the IPrintContentPrefs should be used.
IPrintContentPrefs
Allows certain contents to be removed from the output.
IBoolData (IID_IBOOKEXPORT)
Indicates whether the command is used to export an
InDesign book.
IUIFlagData
Tells the command to use kFullUI, kMinimalUI, or
kSuppressUI.
IOutputPages
Holds the UIDs of the pages to output in layout mode.
IInCopyGalleySettingData
Data for galley export.
IBoolData
(IID_IINCOPYPDFNOTEANNOTATIONDATA)
Layout export note annotation flag.
From InDesign layout view
If a style is passed to the export provider, the IPDFExportPrefs on the kPDFExportStyleBoss is
copied to the IPDFExportPrefs on the kPDFExportCmdBos; otherwise, the IPDFExportPrefs
on the kWorkspaceBoss is used. You can query the session’s PDF export preference as follows:
InterfacePtr<IPDFExportPrefs>
appExportPrefs((IPDFExportPrefs*)::QuerySessionPreferences(IID_IPDFEXPORTPREFS));
To change the preferences, process kSetPDFExportPrefsCmdBoss and use the Set*** methods
of IPDFExportPrefs on the command boss to change the settings.
IPDFSecurityPrefs is set up using the preferences saved in the session (i.e., IPDFSecurityPrefs
on the kWorkspaceBoss). You can query the preferences as follows:
InterfacePtr<IPDFSecurityPrefs>
appSecurityPrefs((IPDFSecurityPrefs*)::QuerySessionPreferences(IID_IPDFSECURITYPREF
S));
To change the security preferences, process kSetPDFSecurityPrefsCmdBoss and use the Set***
methods of IPDFSecurityPrefs on the command boss to change the settings.
Another critical interface in kPDFExportCmdBoss that needs to be set up is IOutputPages. It
holds the UIDs of the pages from the document for export. The PDF export provider uses
IPageRange on the kWorkspaceBoss to initialize the IOutputPages on kPDFExportCmdBoss.
kPDFExportDialogBoss uses the same data to initialize the export settings dialog user inter-
PDF Import and Export
461
PDF Import and Export
PDF export
face, and it updates the session data when the user changes it from the dialog box. Example 28
shows how to set up IOutputPages from IPageRange on kWorkspaceBoss.
EXAMPLE 28 Setting up kPDFExportCmdBoss’s IOutputPages from IPageRange
InterfacePtr<IPageRange>
myPageRange((IPageRange*)::QuerySessionPreferences(IID_IPAGERANGE));
IPageRange::RangeFormat pageRangeFormat = myPageRange->GetPageRangeFormat();
// Assume theDoc is a valid IDocument* for export.
UIDList pageUIDs = UIDList(::GetDataBase(theDoc));
InterfacePtr<IPageList> iPageList((IPMUnknown*)theDoc, IID_IPAGELIST);
if (pageRangeFormat != IPageRange::kAllPages)
{
PMString pageRange;
pageRange = myPageRange->GetPageRange();
pageRange.SetTranslatable(kFalse);
iPageList->PageRangeStringToUIDList(pageRange, &pageUIDs);
}
else
{
int32 cPages = iPageList->GetPageCount();
for (int32 iPage = 0; iPage < cPages; iPage++ )
{
UID uidPage = iPageList->GetNthPageUID(iPage);
pageUIDs.Append( uidPage );
}
}
// Assume pdfExportCmd is the valid ICommand* from kPDFExportCmdBoss.
InterfacePtr<IOutputPages> iExportPages(pdfExportCmd, IID_IOUTPUTPAGES);
// Assume exportPrefs is the previously set IPDFExportPrefs from kPDFExportCmdBoss.
iExportPages->InitializeFrom(pageUIDs, (exportPrefs->GetPDFExReaderSpreads() ==
IPDFExportPrefs::kExportReaderSpreadsON));
PMString name;
theDoc->GetName(name);
iExportPages->SetName(name);
From InCopy layout view
There is an IInCopyPDFExptLayoutData interface on kWorkspaceBoss, which is used to store
some InCopy layout export settings. During an InCopy layout export, IPDFExportPrefs is set
up as described in “From InDesign layout view” on page 461. Then the IInCopyPDFExptLayoutData on the kWorkspaceBoss is used to override some of the data on the command’s IPDFExportPrefs.
Table 112
summarizes
the
overrides.
To
change
the
IInCopyPDFExptLayoutData on the kWorkspaceBoss, use kSaveInCopyPDFExptLayoutDataCmdBoss.
462
PDF Import and Export
PDF export
TABLE 112 InCopy Layout-specific export settings
Attributes
Override
Launch Adobe
Acrobat
Changed by SetPDFExLaunchAcrobat. If
IInCopyPDFExptLayoutData::GetPDFExLaunchAcrobat is equal to kTrue, this is set to
kExportLaunchAcrobatON; otherwise, kExportLaunchAcrobatOFF.
Compatibility
Level
Changed by SetPDFExAcrobatCompatibilityLevel. Set to the value returned by
IInCopyPDFExptLayoutData::GetPDFExAcrobatCompatibilityLevel.
Compression Type
Changed by SetCompressionType. If
IInCopyPDFExptLayoutData::GetPDFExAcrobatCompatibilityLevel >= kPDFVersion15,
this is set to kCompressObjects; otherwise, kCompressNone.
Embed Page
Thumbnails
Changed by SetPDFExThumbnails. If IInCopyPDFExptLayoutData::GetPDFExThumbnails
is equal to kTrue, this is set to kExportThumbnailsON; otherwise, kExportThumbnailsOFF.
Optimize for Fast
Web View
Changed by SetPDFExLinearized. If IInCopyPDFExptLayoutData::GetPDFExLinearized is
equal to kTrue, this is set to kExportLinearizedON; otherwise, kExportLinearizedOFF.
Include Page
Information
Changed by SetPDFExPageInfo. If IInCopyPDFExptLayoutData::GetPDFExPageInfo is
equal to kTrue, this is set to kExportPageInfoON; otherwise, kExportPageInfoOFF.
Subset Fonts
Threshold
Changed by SetPDFExSubsetFontsThreshold. Set to the value returned by
IInCopyPDFExptLayoutData::GetPDFExSubsetFontsThreshold.
Export Reader
Spreads
Changed by SetPDFExReaderSpreads. If
IInCopyPDFExptLayoutData::GetPDFExReaderSpreads is equal to kTrue, this is set to
kExportReaderSpreadsON; otherwise, kExportReaderSpreadsOFF.
Interactive
Elements
Changed by SetPDFExAddInteractiveElements. Set to the value returned by
IInCopyPDFExptLayoutData::GetPDFExAddInteractiveElements.
Multimedia
Content to Embed
Changed by SetContentToEmbed. Set to the value returned by
IInCopyPDFExptLayoutData::GetContentToEmbed.
IOutputPages for kPDFExportCmdBoss is set up like kPDFExportCmdBoss IOutputPages
from IPageRange, except IInCopyPDFExptLayoutData::GetPDFExPageRangeFormat is used
to check the range format and IInCopyPDFExptLayoutData::GetPDFExPageRange is used to
get the range if the range format is not IInCopyPDFExptLayoutData::kAllPages.
kPDFExportCmdBoss also aggregates an interface, IInCopyGalleySettingData. During InCopy
layout export, IInCopyGalleySettingData::SetGalleySetting(kFalse) is called, so no IInCopyGalleySettingData is used.
The IBoolData (IID_IINCOPYPDFNOTEANNOTATIONDATA) on the kPDFFxportCmdBoss is set to the value from IInCopyPDFExptLayoutData::GetPDFExAnnotationNotes.
From InCopy story view or galley view
Export from galley view is different than export from layout view, because the galley and story
view on-screen may not be the export (print) view. Galley export allows the user to output the
galley story with multiple columns and a selected range of story lines. When the user selects
PDF Import and Export
463
PDF Import and Export
PDF export
export, a final view to the document is not yet available. The PDF export provider internally
executes kCreatePrintGalleyViewCmdBoss to generate an invisible galley view for PDF port
printing based on the user’s preferences.
IInCopyPDFExptGalleyData on the kWorkspaceBoss is the default galley-specific setting data
used during PDF export. It is used to set up some of the data in IPDFExportPrefs and IInCopyGalleySettingData; both are aggregated on kPDFExportCmdBoss. The PDF export provider
does not take the data from IInCopyPDFExptGalleyData on the kWorkspaceBoss as is: if the
galley document contains any story and the session’s IInCopyPDFExptGalleyData is default
(checked through IInCopyPDFExptGalleyData::GetIsDefaultValues), the data shown in
Table 113 is overridden.
TABLE 113 IInCopyPDFExptGalleyData overrides when it is default
464
Attributes
Override
Font Leading
Changed by SetPDFExFontLeading. Set to the string value returned from
IGalleySettingsOverwrite::GetDisplayFontLeading.
Font Name
Changed by SetPDFExFont. Set to the value returned by Utils<IGalleyUtils>()>GetFontFamilyAndStyle.
Font Size
Changed by SetPDFExFontSize. Set to the string value returned from
IGalleySettingsOverwrite::GetDisplayFontSize.
Font Type
Changed by SetPDFExFontType. Set to the value returned by Utils<IGalleyUtils>()>GetFontFamilyAndStyle.
Include Accurate Line
Endings
Changed by SetPDFExALE. If ITextLine::GetLinesType is equal to
ITextLines::kLayoutLineEnds, set to kTrue; otherwise, kFalse.
Include Inline Notes
Changed by SetPDFExInlineNotes. If there is at least one note anchor in the document,
set to kTrue; otherwise, kFalse.
Include Line
Numbers
Changed by SetPDFExLineNumber. Set to the value returned by Utils<IGalleyUtils>()>InGalley().
Include Paragraph
Styles
Changed by SetPDFExStyle. Set to the value returned from
IGalleySettingsOverwrite::GetShowParagraphStyleNames.
Include Track
Changes
Changed by SetPDFExTrackChanges. Set to the value returned from
IGalleySettingsOverwrite::GetShowTrackChanges.
Notes Type
Changed by SetPDFExNotesType. Set to kVisible.
Show Notes
Backgrounds in Color
Changed by SetPDFExNotesBackground. Set to kFalse.
Show Track Changes
Backgrounds in Color
Changed by SetPDFExTrackChangesBackground. Set to kFalse.
Track Changes Type
Changed by SetPDFExTrackChangesType. Set to kVisible.
PDF Import and Export
PDF export
NOTE:
ITextLine and IGalleySettingsOverwrite should be from the galley view
(kWritingModeWidgetBoss) of the exported document.
In galley view mode, the PDF export provider initializes IPDFExportPrefs and IPDFExportPrefs in much the same way as described in “From InCopy layout view” on page 462, but the
attributes from IPDFExportPrefs shown in Table 114 are overridden by IInCopyPDFExptGalleyData, as described above.
TABLE 114 Galley-specific settings for IPDFExportPrefs
Attributes
Values
Acrobat Compatibility
Level
Set by SetPDFExAcrobatCompatibilityLevel. Set to the value returned from
IInCopyPDFExptGalleyData::GetPDFExAcrobatCompatibilityLevel.
Include Page
Information
Set by SetPDFExPageInfo. If IInCopyPDFExptGalleyData::GetPDFExPageInfo is equal
to kTrue, this is set to IPDFExportPrefs::kExportPageInfoON; otherwise,
IPDFExportPrefs::kExportPageInfoOFF.
Subset Fonts Threshold
Set by SetPDFExSubsetFontsThreshold. Set to the value returned from
IInCopyPDFExptGalleyData::GetPDFExSubsetFontsThreshold.
As described in “From InCopy layout view” on page 462, IOutputPages is a critical interface
that holds the UIDs of pages to be exported for kPDFExportCmdBoss in layout mode. For
InCopy galley and story mode, however, IOutputPages cannot provide enough data for kPDFExportCmdBoss to draw a range of lines from galley view or rearrange the number of columns in the output PDF that are allowed in the galley export options dialog box. Therefore, the
interface IInCopyGalleySettingData is aggregated on kPDFExportCmdBoss to hold the data
used to create a galley window for PDF export. In addition to settings specific to galley mode,
IInCopyGalleySettingData also holds the views (IControlView* of the galley writing widgets,
paragraph information panel, etc.) that represent the real output pages based on the user’s galley export settings. kPDFExportCmdBoss passes these views to the PDF drawing port for output instead of using IOutputPages. The user of kPDFExportCmdBoss is responsible for setting
up IInCopyGalleySettingData properly before the command is processed. The PDF export provider executes kCreatePrintGalleyViewCmdBoss to get an invisible output view based on the
data from the current galley document and IInCopyPDFExptGalleyData. It then uses the view
and IInCopyPDFExptGalleyData to set up IInCopyGalleySettingData for kPDFExportCmdBoss. The PDF export provider sets up IOutputPages of kPDFExportCmdBoss with IOutputPages returned from kCreatePrintGalleyViewCmdBoss, which has the values shown in
Table 115.
PDF Import and Export
465
PDF Import and Export
PDF export
TABLE 115 IOutput pages from kCreatePrintGalleyViewCmdBoss
Attributes
Values
ContiguousPages
kTrue
Name
Name of the galley document
IsSpreads
kFalse
MasterDataBase
Database for the window created by the
kCreatePrintGalleyViewCmdBoss
UIDs
Placeholder. UIDs number from 0 to total pages-1. UIDs
contained in IOutputPages are not used by kPDFExportCmdBoss
in InCopy galley and story mode.
The following steps summarize how the PDF export provider sets up IInCopyGalleySettingData for kPDFExportCmdBoss.
1. A new kInCopyGalleySettingDataBoss is created, with default value.
2. IInCopyPDFExptGalleyData is used to override the IInCopyGalleySettingData of the new
kInCopyGalleySettingDataBoss.
3. IInCopyGalleySettingData is used to execute the kCreatePrintGalleyViewCmdBoss command.
4. A copy of IInCopyGalleySettingData returned from kCreatePrintGalleyViewCmdBoss in
Step 3 is used to set up kPDFExportCmdBoss.
Table 116 shows how attributes from IInCopyGalleySettingData get their values in different
stages. In the table, “PEGD” stands for IInCopyPDFExptGalleyData from Step 2 above.
TABLE 116 Setting up IInCopyGalleySettingData for kPDFExportCmdBoss
466
Attribute and description
Default value
Overridden value
Final value after kCreatePrintGalleyViewCmdBoss
Galley/story control view
nil
nil
Calculated by
kCreatePrintGalleyViewCmd
Boss
Paragraph panel control view
(GetInfoColumnView)
nil
nil
Calculated by
kCreatePrintGalleyViewCmd
Boss
Line number panel control view
(GetLineNumberView)
N/A, not used
N/A, not used
N/A, not used
Splitter control view
(GetInfoSplitterView)
nil
nil
Calculated by
kCreatePrintGalleyViewCmd
Boss
PDF Import and Export
PDF export
Attribute and description
Default value
Overridden value
Final value after kCreatePrintGalleyViewCmdBoss
Total content height of the
galley (GetTotalHeight)
0
0
Calculated by
kCreatePrintGalleyViewCmd
Boss
Start Line Number
(GetStartLineNumber)
N/A, not used
N/A, not used
N/A, not used
End Line Number
(GetEndLineNumber)
N/A, not used
N/A, not used
N/A, not used
Column width
(GetColumnWidth)
0
Calculated based on PEGD
and current unit settings
Calculated by
kCreatePrintGalleyViewCmd
Boss
GalleySetting
kTrue
kTrue
kTrue
Document UIDRef
(GetDocUIDRef)
N/A
UIDRef of galley document
UIDRef of galley document
To print with paragraph style
info (GetParaStyle)
kFalse
from PEGD
from PEGD
To print with line number
(GetLineNumber)
kFalse
from PEGD
from PEGD
To print with accurate line
ending (GetALE)
kFalse
from PEGD
from PEGD
To print with notes displayed
(GetNotes)
kFalse
from PEGD
from PEGD
To print with tracked changes
displayed (GetTrackChange)
kFalse
from PEGD
from PEGD
Notes displayed type
(GetNotesType)
kVisible
from PEGD
from PEGD
Track change displayed type
(GetTrackChangesType)
kVisible
from PEGD
from PEGD
Font name (GetFontName)
“”
from PEGD
from PEGD
Font type (GetFontType)
“”
from PEGD
from PEGD
Font size (GetFontSize)
“”
from PEGD
from PEGD
Font leading (GetFontLeading)
“”
from PEGD
from PEGD
PDF Import and Export
467
PDF Import and Export
PDF export
Attribute and description
Default value
Overridden value
Final value after kCreatePrintGalleyViewCmdBoss
To print with line range scope
(GetWhich)
IIncopyGalley
PrintData::kAll
If PEGD’s
GetPDFExlineRangeFormat
is
IInCopyPDFExptGalleyDat
a::kAllLine, set to kAllLines;
otherwise, kUseRange
If PEGD’s
GetPDFExlineRangeFormat
is
IInCopyPDFExptGalleyData:
:kAllLine, set to kAllLines;
otherwise kUseRange.
To print the line range
(GetRange)
“”
from PEGD
from PEGD
Galley frame size
(GetFrameSize)
Rect(0, 0, 0, 0)
Set frame to the page
bounds from the first page
of the galley document.
Set frame to the page bounds
from the first page of the
galley document.
To print with content filled with
the page (GetFill)
kFalse
If GetALE returns true, then
set to the return value from
PEGD’s GetPDFExFill;
otherwise, kFalse.
If GetALE returns true, then
set to the return value from
PEGD’s GetPDFExFill;
otherwise, kFalse.
Story range (GetScope)
IInCopyGalley
PrintData::kAll
from PEGD
from PEGD
To print with story information
(GetStoryInfo)
kFalse
from PEGD
from PEGD
To print with notes background
in color
(GetNotesBackgroundInColor)
kTrue
from PEGD
from PEGD
To print with track changes
background in color
(GetTrackChangesBackgroundI
nColor)
kTrue
from PEGD
from PEGD
To print with page information
(GetPagesInfo)
kFalse
kFalse
kFalse
(kPDFExportCmdBoss takes
this info from
IPDFExportPrefs)
To print with number of
columns (GetColumns)
1
1
Calculated by
kCreatePrintGalleyViewCmd
Boss
InDesign book export
You can use the high-level kBookExportActionCmdBoss to easily export a book. For details,
see SnpExportBookAsPDF.cpp.
kBookExportActionCmdBoss uses kPDFExportBookBoss (with service ID kExportBookService), which uses the same PDF export service provider (kPDFExportProviderImpl) as the
468
PDF Import and Export
PDF export
kPDFExportBoss; however, kPDFExportBookBoss aggregates an IOutputPages interface that
lists document pages in the book to export. The regular document kPDFExportBoss does not
aggregate IOutputPages; it sets up IOutputPages for the kPDFExportCmdBoss based on the
IPageRange, as shown in Example 28.
Selected page-items export
There is no direct user interface to allow export of selected page items in InDesign; however,
you can copy selected page items to the pasteboard in PDF format. Then, when you paste into
an external application that supports PDF import, the pasted object is in PDF format. This is
done through kPDFExportItemsCmdBoss. The snippet in Example 29 shows how to process
the command.
EXAMPLE 29 How to process a kPDFExportItemsCmdBoss command
// Assume stream is the destination PDF file write stream.
// Assume realPageItems contains no guide item, which can
// cause trouble in PDF export.
InterfacePtr<ICommand> command(CmdUtils::CreateCommand(kPDFExportItemsCmdBoss));
if (command)
{
command->SetItemList(realPageItems);
// Initialize the command data from the session.
InterfacePtr<IPDFExportPrefs> appExportPrefs((IPDFExportPrefs *)
::QuerySessionPreferences(IID_IPDFCLIPBOARDEXPORTPREFS));
InterfacePtr<IPDFExportPrefs> exportPrefs(command, IID_IPDFEXPORTPREFS);
InterfacePtr<IPDFSecurityPrefs> appSecurityPrefs((IPDFSecurityPrefs *)
::QuerySessionPreferences(IID_IPDFSECURITYPREFS));
InterfacePtr<IPDFSecurityPrefs>
exportSecurityPrefs(command,IID_IPDFSECURITYPREFS);
exportPrefs->CopyPrefs(appExportPrefs);
exportSecurityPrefs->CopyPrefs(appSecurityPrefs);
// no progress bar
InterfacePtr<IBoolData> useProgressBar(command, IID_IUSEPROGRESSINDICATOR);
useProgressBar->Set(kFalse);
InterfacePtr<IUIFlagData> uiFlagData(command, IID_IUIFLAGDATA);
uiFlagData->Set(kSuppressUI);
// Assume no destination color profile (i.e., ignore IUIDData).
// (IID_IPDFDESTCMSPROFILE)
// Put the stream pointer in the IIntData.
InterfacePtr<IIntData> exportCmdStreamData(command, IID_IINTDATA);
exportCmdStreamData->Set((long) stream);
// process the command
CmdUtils::ProcessCommand(command);
success = ErrorUtils::PMGetGlobalErrorCode();
}
PDF Import and Export
469
PDF Import and Export
PDF-style import and export
PDF-style import and export
When the PDF export provider brings up the Export Adobe PDF options dialog box, the dialog
box is initialized according to the following rules:
1. A preset style UID may be passed in. When it is available, use it.
2. Otherwise, if the last preset used is named “[Custom]” or the style name ends with “(modified),” use the application preferences of the workspace.
3. Otherwise, if the last preset used is not empty and is valid, use it.
4. Otherwise, use the default preferences. This is the first time the export dialog is run after
deleting Save Data.
PDF export style is represented by kPDFExportStyleBoss, which aggregates an IPDFExportPrefs interface that stores the settings. Usually, the IPDFExportPrefs -> CopyPrefs method is
used to copy the setting out of the style boss to another IPDFExportPrefs you are using. The
last preset style’s name is saved in the IPDFExportStyleLastUsed on the kWorkspace. You can
use the snippet in Example 30 to get the name in PMString:
EXAMPLE 30 Getting the last-used style name
InterfacePtr<IPDFExportStyleLastUsed>
iStyleLast((IPDFExportStyleLastUsed*)::QuerySessionPreferences
(IID_IPDFEXPORTSTYLELASTUSED));
PMString lastPreset;
if (iStyleLast)
lastPreset = iStyleLast->GetString();
lastPreset.SetTranslatable(false);
Given a style name, you can use IPDFExptStyleListMgr (aggregated on kWorkspace) to obtain
the UID style object, Continuing from Example 30, Example 31 shows how to get to a last used
style object from its name:
EXAMPLE 31 Getting a style object using name
InterfacePtr<IPDFExptStyleListMgr>
styleMgr((IPDFExptStyleListMgr*)::QuerySessionPreferences(IID_IPDFEXPORTSTYLELISTMG
R));
int32 nStyle = styleMgr->GetStyleIndexByName(lastPreset);
if (nStyle != -1)
{
UIDRef styleRef = styleMgr->GetNthStyleRef(nStyle);
InterfacePtr<IPDFExportPrefs> pStylePrefs(styleRef, UseDefaultIID());
if( pStylePrefs )
{
// assume myExportPrefs is an IPDFExportPrefs I am trying to set up
myExportPrefs->CopyPrefs(pStylePrefs);
};
}
470
PDF Import and Export
PDF-style import and export
Adding, deleting, and editing styles
To add a style, use kPDFExportAddStyleCmdBoss. Example 32 shows how to process the command:
EXAMPLE 32 Adding a PDF-export style
InterfacePtr<ICommand> addCmd(CmdUtils::CreateCommand(kPDFExportAddStyleCmdBoss));
InterfacePtr<IExportStyleCmdData> cmdData(addCmd, IID_IEXPORTSTYLECMDDATA);
InterfacePtr<IWorkspace> workspace(gSession->QueryWorkspace());
UIDRef workspaceUIDRef = ::GetUIDRef(workspace);
cmdData->SetSrcList(workspaceUIDRef);
cmdData->SetDstList(workspaceUIDRef);
// Assume presetName is a PMString containing the new style name.
cmdData->SetNewName(presetName);
cmdData->SetStyleIndex(-2); // -2 means get the PDF preferences from the command.
InterfacePtr<IPDFExportPrefs> commandPrefs(cmdData, UseDefaultIID());
// Assume pSrcPref is my source IPDFExportPrefs to be made into the new style.
commandPrefs->CopyPrefs(pSrcPrefs);
commandPrefs->SetUIName(presetName);
ErrorCode rc = CmdUtils::ProcessCommand(addCmd);
The command boss aggregates an IExportStyleCmdData that has a method, SetStyleIndex, that
sets up a style index for the command. The style index tells the command where the source
style comes from, according to the following rules:
1. If there is valid index (index >= 0), that style's PDF preferences (queried from an IPDFExptStyleListMgr returned by IExportStyleCmdData::GetSrcList) are used.
2. If the index is -2, the PDF preferences come from the command.
3. If the index is -3, the PDF preferences come from the command, but the data is not written
to disk.
4. Otherwise, the PDF preferences come from the current application preferences.
5. In all cases except -3, the data is written to disk.
To delete a style, use kPDFExportDeleteStyleCmdBoss. Example 33 shows how to delete a style
stored in the application’s workspace:
EXAMPLE 33 Deleting a PDF-export style
InterfacePtr<IWorkspace> workspace(gSession->QueryWorkspace());
UIDRef workspaceUIDRef = ::GetUIDRef(workspace);
InterfacePtr<ICommand>
deleteCmd(CmdUtils::CreateCommand(kPDFExportDeleteStyleCmdBoss));
InterfacePtr<IExportStyleCmdData> commandData(deleteCmd, IID_IEXPORTSTYLECMDDATA);
// Assume i is the (to-be-deleted) style index in the style list.
commandData->SetStyleIndex(i);
commandData->SetDstList(workspaceUIDRef);
CmdUtils::ProcessCommand(deleteCmd);
PDF Import and Export
471
PDF Import and Export
Frequently asked questions
To edit a style, use kPDFEditStyleCmdBoss. Example 34 shows how to edit a style stored in the
application’s workspace:
EXAMPLE 34 Editing a PDF style
InterfacePtr<ICommand> editStyleCmd(CmdUtils::CreateCommand(kPDFEditStyleCmdBoss));
InterfacePtr<IIntData> data(editStyleCmd, IID_IINTDATA);
// Assume index is the index of the style to be edited.
data->Set(index);
InterfacePtr<IPDFExportPrefs> newPrefs(editStyleCmd, IID_IPDFEXPORTPREFS );
// Assume myPrefs is the edited copy of IPDFExportPrefs.
newPrefs->CopyPrefs(myPrefs);
CmdUtils::ProcessCommand(editStyleCmd);
Frequently asked questions
How does the PDF export provider determine whether it should start the
viewer after the export?
In InDesign, IPDFPostProcessPrefs is used to maintain persistent PDF preference data not part
of the standard IPDFExportPrefs. IPDFPostProcessPrefs is aggregated on kWorkspaceBoss and
can be queried through QuerySessionPreferences. Use IPDFPostProcessPrefs::GetViewAfterExport to determine whether the viewer should be launched.
In InCopy, IInCopyPDFExptGalleyData::GetPDFExLaunchAcrobat (for galley) and IInCopyPDFExptLayoutData::GetPDFExLaunchAcrobat (for layout) are used to determine whether the
viewer should be started.
How do I set the PDF clipboard setting as seen in the File Handling
preferences?
Those settings (e.g., Prefer PDF when Pasting) are kept in the IPDFClipboardPrefs aggregated
on the kWorkspaceBoss. To modify the settings, use kSetPDFCBPrefsCmdBoss.
How do I control which layer of a document should be exported?
IPDFExportPrefs::SetExportLayers can be used to control the function of layers for visibility
and printability. Use the enums kExportAllLayers, kExportVisibleLayers, and kExportVisiblePrintableLayers to set the export layer’s preference. Each layer’s visibility and printability are
controlled by each layer’s option, managed by IDocumentLayer. For more information about
how to set each layer’s visibility and printability, see the “Layer Options” section of the “Layout
Fundamentals” chapter.
472
PDF Import and Export
Frequently asked questions
How do I make the two-page spreads in my document export as two
separate PDF pages?
Set the Boolean control for reader spreads in the IPDFExportPrefs to false. The IOutputPages
on the kPDFExportCmdBoss usually is initialized as follows:
InterfacePtr<IOutputPages> iExportPages(cmd, IID_IOUTPUTPAGES);
iExportPages->InitializeFrom(pageUIDs, (exportPrefs->GetPDFExReaderSpreads() ==
IPDFExportPrefs::kExportReaderSpreadsON));
If the preference for the reader spreads is set to false and you set up IOutputPages as above, the
two pages spreads are exported as separated pages in the PDF. See “Setting up kPDFExportCmdBoss’s IOutputPages from IPageRange” on page 462.
Why does kPDFExportCmdBoss give me an assert after the command is
processed (ASSERT 'db != nil' in PDFExportController.cpp)?
Make sure you put the UIDRef of the pages in the IOutputPage interface. Also, set the item list
with the pages’ UIDRef values (in a UIDList). kPDFExportCmdBoss is used for exporting both
books (multiple .indd files) and multiple pages from a document. kPDFExportCmdBoss first
checks to see whether it has a valid IOutputPages and uses the database from the first UIDRef
of the output pages. If the command cannot find a valid database from IOutputPages, it looks
for it from the command item list.
How do I set up line ranges for output in InCopy Galley or Story mode?
Use IInCopyPDFExptGalleyData::SetPDFExLineRange. Pass in a PMString with a format like
“3-10” (i.e., the beginning page number, followed by a dash, followed by the ending page number). IInCopyPDFExptGalleyData is aggregated on the kWorkspaceBoss.
Is it possible to export only selected text from an InDesign document?
No. When printing or exporting to PDF, the decision to draw or not draw is made at the pageitem level. This is evidenced by the “Nonprinting” attribute, which can be applied to a text
frame but not the text itself. Similarly, draw event handlers work on the page-item level, such as
the text frame.
PDF Import and Export
473
PDF Import and Export
Frequently asked questions
474
Implementing Preflight Rules
Introduction
Implementing Preflight Rules
Introduction
This chapter provides high-level information for developers who want to build their own rules
or set of rules. It complements the low-level information found in the SDK.
About preflight in InDesign CS4
InDesign CS4 has a new feature for profile-based preflight. There was a preflight feature before
CS4, but it was designed primarily for improving the reliability of the package operation. CS4
includes a fully rule-driven, parameterized, continuous preflight that can run in the background as you work.
The preflight model comprises several major areas, most of which are fully extensible by external developers. Very little preflight code is written using private interfaces.
z
The preflight object model is a parallel view of the InDesign document and things that do
not appear in the document. It has references to both standard InDesign objects like page
items and documents and to non-persistent entities like marking operations. The object
model is fully extensible to support new objects.
z
Preflight rule services are where objects are checked for compliance with a set of parameters.
A rule provides discovery, visitation, and reporting functions to the preflight engine.
z
Preflight processes are idle-task-driven in-memory objects which coordinate all interactions
between the object model and the rule services. Processes also can be driven synchronously
if desired, but normally they run in the background. The processes are not extensible, but
developers have full access to their status and database.
This chapter is concerned only with the rule services, which are the most common type of
extension.
About rules
In the preflight model, rules are the entities that check conditions in the document and generate errors. An example of a rule is “Missing and Modified Graphics,” which looks for graphics
that are missing or outdated.
Implementing Preflight Rules
475
Implementing Preflight Rules
Rule IDs
To write a rule, you need two things:
z
The rule boss encapsulates both the intelligence and data for a rule.
z
The rule service tells InDesign that your rule exists and lets it create a boss for a rule. A rule
service can “host” as many rules as desired; the native InDesign rule service hosts dozens of
rules.
Rule IDs
Each unique rule (for example, missing fonts or image resolution) is identified by a rule ID. A
rule ID is simply a WideString that is never exposed in the user interface but serves to distinguish it from every other rule out there. For each rule you implement, you need to devise a
unique rule ID. For example, you could prefix your rule ID with your company name or use a
GUID converted to a string. Rule IDs are exposed to scripting.
Rule service
A rule service is implemented as a standard InDesign service with the service ID kPreflightRuleService. On the service boss, you need the IID_IK2SERVICEPROVIDER (you can use
kPreflightRuleSPImpl for the implementation) and an IPreflightRuleService. The latter interface has two methods, shown in Table 117.
TABLE 117 Methods in your IPreflightRuleService implementation
Method
Purpose
GetAllRules
Returns a vector of rule IDs that your services hosts.
CreateRule
Given a rule ID that you host, creates a rule boss and returns it.
In other words, this interface provides a map and boss factory for rule bosses. Most of the complexity is in the rule bosses themselves.
CreateRule can be called on when InDesign is creating new preflight profiles or, in some cases,
for purely temporary purposes. For example, it might create a rule just to get some information
about it, then destroy it. It takes a database pointer to indicate whether it wants the boss to be
created in a document (or preferences) or in memory for temporary purposes.
476
Implementing Preflight Rules
Rule service
IPreflightRuleService example
To illustrate a typical rule-service implementation, here are some sample method implementations. In this example, there are two rules, Maximum Text Size and Maximum Rectangle Size.
IPreflightRuleService::GetAllRules
The first step is to create rule IDs for these rules; typically, we’ll also need rule bosses. If XYZCo
company is developing the rules, we might use the mapping shown in Table 118.
TABLE 118 Mappings used in our PreflightRuleService example
Rule
Rule ID
Boss
Maximum text size
XYZCo_MaxTextSize
kMaxTextSizeRuleBoss
Maximum rectangle size
XYZCo_MaxRectSize
kMaxRectSizeRuleBoss
The GetAllRules method must return a vector of rule IDs. This is not complicated:
const PreflightRuleID kMaxTextSizeRuleID("XYZCo_MaxTextSize");
const PreflightRuleID kMaxRectSizeRuleID("XYZCo_MaxRectSize");
virtual PreflightRuleIDVector GetAllRules() const
{
PreflightRuleIDVector rules;
rules.push_back(kMaxTextSizeRuleID);
rules.push_back(kMaxRectSizeRuleID);
return rules;
}
This method usually will be called only once per session, by the rule manager, which then
holds onto the information in its internal maps.
IPreflightRuleService::CreateRule
This method creates the appropriate boss from a passed-in rule ID and fills in defaults. In our
example, we would do something like the following. (For ease of reading, most error handling
is omitted.)
virtual IPreflightRuleInfo* CreateRule
(
// Rule to create
PreflightRuleID ruleID,
// Database in which to create it (nil = in memory)
IDataBase* db
) const
{
ClassID bossID = 0;
PMString desc;
if (ruleID == kMaxTextSizeRuleID)
{
bossID = kMaxTextSizeRuleBoss;
desc = PMString("Maximum text size");
}
Implementing Preflight Rules
477
Implementing Preflight Rules
Rule bosses
else if (ruleID == kMaxTextSizeRuleID)
{
bossID = kMaxRectSizeRuleBoss;
desc = PMString("Maximum rectangle size");
}
else return nil;
IPreflightRuleInfo* iRule = (IPreflightRuleInfo*)CreateObject
(db, bossID, IID_IPREFLIGHTRULEINFO);
iRule->SetRuleID(ruleID);
iRule->SetRuleDescription(desc);
iRule->SetPluginDescription("XYZCo Rules Plugin");
// Set up default parameter values.
InterfacePtr<IPreflightRuleUtilities> iUtils(iRule, UseDefaultIID());
iUtils->UpdateRuleData();
return iRule;
}
It is up to you what strategy to use for initialization, mapping rule IDs to bosses and descriptions, and so on. If you have only a few rules, the method above should suffice.
Rule bosses
A rule boss is where all the heavy lifting goes on in the preflight world. A rule boss corresponds
to a particular rule and encapsulates all the behaviors of that rule, including the following:
z
Parameter data — For example, if your rule looks for things larger than X, and the user can
specify X, X is a rule parameter.
z
User interface — This is the interface used to modify your rule parameters.
z
Evaluation — Your rule is a visitor. The preflight engine “carries it around” to objects it
wants to visit, and the rule must determine if the object is alright or has an error.
z
Aggregation — Once your rule finishes gathering errors, it must determine how to present
the errors to the user.
z
Utilities — If your rule has any special requirements when copied or deleted, it needs to provide those implementations.
As shown in the previous example, each rule your plug-in supports typically has its own boss.
Each rule boss must have the interfaces shown in Table 119. You can add whatever other interfaces you want on your rule bosses.
478
Implementing Preflight Rules
Rule bosses
TABLE 119 Required interfaces on rule bosses
Interface
Purpose
IPreflightRuleData
Stores the rule parameter data as a dictionary (key-value
pairs). Use kPreflightRuleDataImpl unless you have a really
good reason not to do so.
IPreflightRuleVisitor
Provides all the evaluation and aggregation functions. This
interface tells InDesign what kinds of objects the rule wants
to visit (page items, styles, artwork, etc.), is called back when
it is time to visit one of those objects, and is called back when
it is time to aggregate (create final report of findings). Also, it
provides some parameter initialization and validation
methods.
IPreflightRuleUtilities
Provides rule-specific utilities for your rule, including
copying the rule, any special deletion requirements, and
determining equality with other rules. In most cases, you can
use the default kPreflightRuleUtilitiesImpl implementation.
These interfaces are documented in the headers and doc++ comments, but the IPreflightRuleVisitor interface is critical, so we expand on it below.
IPreflightRuleVisitor interface
The IPreflightRuleVisitor interface has three methods related to rule evaluation, shown in
Table 120.
TABLE 120 Rule evaluation methods on IPreflightRuleVisitor
Method
Description
GetClassesToVisit
Preflight calls this method, generally on start-up, when it builds
its maps relating rules to objects. It must return a vector of
preflight object class IDs that your rule wants to visit. For
example, if you want to visit all images, you would return a vector
containing kPreflightOM_Image.
Visit
Preflight calls this method when it is time to evaluate a given
object, either because it was not visited before or something was
changed such that the result might be different. This method is
passed an abstract utility class that your method uses to obtain
information and report its findings.
AggregateResults
Preflight calls this method after the first pass (visitation) is
complete and the Visit method found something it might want to
report. This method takes the “raw” results found during
visitation into final presentation, in the form of nodes in the
results tree. This method typically consolidates multiple errors
into a single node, builds strings, and so on.
Implementing Preflight Rules
479
Implementing Preflight Rules
IPreflightRuleVisitor method examples
The IPreflightRuleVisitor interface also has two methods related to rule-parameter initialization and validation, shown in Table 121. (These could be on another interface but since
every rule needs to implement them, there is less overhead to include them on this interface,
since every rule needs its own IPreflightRuleVisitor.)
TABLE 121 Parameter methods on IPreflightRuleVisitor
Method
Description
UpdateRuleData
Preflight calls this method at start-up (for application profiles)
and document open time (for document embedded profiles), to
give the rule an opportunity to initialize and/or update the rule
parameters.
ValidateRuleData
Preflight calls this method when a parameter is set via scripting, to
ensure it is a legitimate value.
IPreflightRuleVisitor method examples
The methods in this interface concentrate on the discovery, visitation, and aggregation phases
of the preflight process. This is the trickiest interface to write.
IPreflightRuleVisitor::GetClassesToVisit
This method simply tells InDesign what object classes your rule wants to visit. For example the
missing-fonts rule looks at text runs (to see directly which fonts are used and where), as well as
placed content that has required-font data, namely placed EPS, EPS, and INDD files. Its implementation looks like the following:
virtual PreflightObjectClassIDVector GetClassesToVisit() const
{
PreflightObjectClassIDVector classes;
classes.push_back(kPreflightOM_WaxRun);
classes.push_back(kPreflightOM_EPS);
classes.push_back(kPreflightOM_PDF);
classes.push_back(kPreflightOM_INDD);
return classes;
}
Do not declare classes in which you are not actually interested. The preflight engine expands
only those portions of the model required to satisfy the rules in the profile, If you ask for artwork, for instance, you will take a big hit in processing cycles; do this only if it is needed.
480
Implementing Preflight Rules
IPreflightRuleVisitor method examples
IPreflightRuleVisitor::Visit
This is the method in which your rule inspects objects it asked to visit. Typically, this method
consists of comparing attributes of the object against rule parameters.
The argument for this method is a IPreflightVisitInfo interface, which provides all the information you should need about the object being visited. It also provides the mechanism for reporting the errors. Table 122 lists the methods in this interface.
TABLE 122 IPreflightVisitInfo methods
Method
Purpose
QueryOptions
Gets the IPreflightOptions interface for the current process. Not
typically used, but sometimes rules want to know the options.
GetObjectID
Gets the PreflightObjectID of the thing being visited. Typically,
this is used to get the ClassID of that object, if your rule looks at
multiple object types.
QueryObject
Obtains the IPreflightObject interface for the object being
visited. This is used in nearly every rule, because it lets you ask
questions about the object.
CreateResultRecord
Creates a result record; that is, records a problem (or potential
problem) with this node. Usually, this is an error indication, but
not necessarily; you might use this to count instances of
something and have an error only if it finds more than a certain
number.
Here is an example of a Visit implementation, the page-count rule. (To enhance readability,
most error checking, like nil interfaces, is omitted.)
virtual void Visit(IPreflightVisitInfo* iVisit) override
{
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
InterfacePtr<IDocument> iDoc(iObj->GetModelRef(), UseDefaultIID());
InterfacePtr<IPageList> iDocPages(iDoc, UseDefaultIID());
int32 count = iDocPages->GetPageCount();
PreflightRuleDataHelper dh(this);
ComparisonType comp = (ComparisonType)dh.GetIntParam
(kParamCountComparisonType);
int32 value = dh.GetIntParam(kParamCountComparisonValue);
int32 value2 = dh.GetIntParam(kParamCountComparisonValue2);
bool32 fails = kFalse;
Implementing Preflight Rules
481
Implementing Preflight Rules
IPreflightRuleVisitor method examples
switch(comp)
{
case kComparison_EqualTo:
fails = (count != value);
break;
case kComparison_Minimum:
fails = (count < value);
break;
case kComparison_Maximum:
fails = (count > value);
break;
case kComparison_MultipleOf:
fails = value > 1 && (count % value != 0);
break;
case kComparison_Between:
fails = (count < value || count > value2);
break;
}
if (fails)
{
InterfacePtr<IPreflightResultRecord> iRec
(iVisit->CreateResultRecord());
iRec->SetCriteria(kPreflightRC_Default);
iRec->AddValue(count);
}
}
In this case, the rule is looking only at the document object, so it does not need to check for the
class ID of the object that is passed in. The first thing it does is as follows:
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
InterfacePtr<IDocument> iDoc(iObj->GetModelRef(), UseDefaultIID());
Since the object in question is the document, it has a model mapping, so iObj->GetModelRef()
produces the UIDRef of the document. Not all objects have a model mapping. You need to do
things that are unique to the kind of object you are looking at, even if in general those things
fall into general categories.
The rest of the implementation simply compares the number of pages against the rule parameter, using the rule’s comparison type (also a rule parameter). If it fails, then it does the following:
InterfacePtr<IPreflightResultRecord> iRec
(iVisit->CreateResultRecord());
iRec->SetCriteria(kPreflightRC_Default);
iRec->AddValue(count);
This code creates a result record, which is simply a small data structure that records what the
rule tells it to, and attaches it in the process database (a non-persistent database) to the node
representing the object (in this case, the document). This record can be produced later (along
with all other records the rule created as it inspected objects), in the results-aggregation
method. In this example, there is only one object, the document, but in a case where you are
inspecting page items or marking operations, there could be dozens or hundreds.
482
Implementing Preflight Rules
IPreflightRuleVisitor method examples
After the result record is created, the rule sets the criteria to kPreflightRC_Default. This is the
value used by rules that do not need to break down errors into different categories (or at least
not at the time it is finding the errors).
Finally, consider the case where the rule is adding a value to the record. The meaning of this
value and the use of the value vector are completely up to the rule; however, when the rule is
inspecting the object, it may make complex decisions or get calculated values that are inconvenient to repeat when results are aggregated. For example, a marking operation that looks at
stroke width will record the stroke width it found, because it does not want to have to look it up
again later. More examples of this are later in this section.
Table 123 shows the contents of a results record.
TABLE 123 Contents of a result record
Field
Uses
Object
ID
The PreflightObjectID of the entity to which the record refers. This always
is the object to which the record was attached.
Subparts
A vector of subpart IDs. Normally, this is set up by one of the aggregation
utilities during the aggregation phase, but a rule also may specify a subpart,
if that is useful.
Values
A vector of PMReals. Normally, this is a width, height, or other quantity
you want to record for this instance. For example, you can record the width
of a stroke you found. The aggregation utilities also use this information
when combining multiple records together.
Criteria
This can be used to differentiate different kinds of failure. This field also is
used by the aggregation utilities to simplify the process of generating
results. An example is the image-resolution rule, which has different
criteria for each image type and min/max failure. This means that the
aggregation method does not have to determine what the image type was
again.
Aux
String
This is just string data that the rule can use as it wishes. For example, the
font-type rule records the font name here. This allows it to build results
easily, because the aggregation utilities can sort by this value.
Process
Node
If this record was created by IPreflightVisitInfo::CreateResultRecord, it has
the process node ID recorded here. This is useful if you want to look at the
result relative to other objects in the tree. The aggregation utilities use this
to determine the relationship between artwork and containing objects.
IPreflightRuleVisitor::AggregateResults
This method is where your rule takes the results (specifically, the result records) found in the
Visit phase and produces nodes in the aggregated results tree. In other words, the input is the
array of result records; for example:
Implementing Preflight Rules
483
Implementing Preflight Rules
IPreflightRuleVisitor method examples
Stroke on object X is 0.1 pt
Stroke on object Y is 0.1 pt
The output is a hierarchy of preflight aggregated result nodes; for example:
Strokes are too small
Object X
Object Y
This example is simple. In the general case, especially with artwork, the raw results can be
numerous, with many grouped under the same object (e.g., the same PDF).
The tree contains several different kinds of nodes:
z
Category nodes are added by the preflight engine and represent the rule categories; for
example, Document, Links, and Text. These nodes are created automatically by preflight,
based on the categories the rule services declare and place the rules in.
z
Criteria nodes are used to group failing objects together under the common failure type. For
example, “Strokes are too small” is a criteria, under which you would put the violations of
that rule (Object X and Object Y, in the above example).
z
Violation nodes are individual failures of a rule. In the example above, “Object X” and
“Object Y” are violation nodes.
z
Criteria/violation nodes are nodes that are both criteria and violation, because there is only a
single object or single failure for that rule. Typically, this means it is a document-level error,
like number of pages.
Each node has a short string that appears in the tree-view widget in the preflight panel and
long-form strings that appear in the Info section. In both cases, these strings also appear in the
preflight report.
Mapping the raw preflight result records to the tree hierarchy can be complex, because each
rule usually has its own strategy for reducing the complexity of the results presentation. The
preflight API allows the rules to do this any way they like, but it supports those mechanisms
with utilities that reduce the line count. All these utilities are in IPreflightAggregatedResultsUtils.
The recommended sequence of events in the AggregateResults method is as follows. The code
assumes a declaration with the following arguments:
virtual void AggregateResults
(
const IPreflightProcess* iProcess,
const IPreflightProcess::NodeIDVector& resultNodes,
IPreflightAggregatedResults* iResults,
IPreflightAggregatedResults::NodeID parentID
) const
The first two parameters can be thought of as the inputs to the aggregation process; i.e., the
nodes in the process database corresponding to the raw results found during the Visit. The last
two parameters can be thought of as the outputs: iResults is the artwork-tree interface you
would use to add nodes, and parentID is the node in that tree under which you should add
your nodes as children.
484
Implementing Preflight Rules
IPreflightRuleVisitor method examples
All utilities in the IPreflightAggregatedResultsUtils work on preflight result records, collecting
them in tables (IPreflightResultRecordTable) for most operations. A IPreflightResultRecordTable is simply a vector of ref-counted record pointers, so dividing a table into subtables is a
fairly quick operation; therefore, most of the utilities transform one table of results into
another, at which point those results are copied into the aggregated results tree.
Typical steps in aggregation are as follows:
1. Get the raw table of results — All the rule’s result records are extracted from the process
database, to get an initial, raw, unprocessed table.
2. Apply standard aggregations — Records are combined in the table using standard combination techniques, such as localizing artwork to the page items that contain it and combining
contiguous text runs. Improves the usability of the results and can dramatically reduce the
size of the results.
3. Combine records into buckets — Once the record list is reduced by standard aggregations,
there are various ways of grouping results together. In some cases, multiple records map to
one result node; in others, result nodes are grouped together.
4. Generate result nodes — To make a result node, all required strings are generated and
added to the result node; then, the result node is inserted into the tree.
Get the raw table of results
The first step in aggregation is getting the initial table of results by collecting them from the
process-database result nodes. Unless you have a good reason otherwise, use IPreflightAggregatedResultsUtils::CreateTableFromNodes for this step:
Utils<IPreflightAggregatedResultsUtils> iUtils;
InterfacePtr<const IPreflightResultRecordTable> iRawTable(
iUtils->CreateTableFromNodes(iProcess, resultNodes));
The raw table now contains all the results added by your iVisit->CreateResultRecord() calls.
Apply standard aggregations
The usual next step is aggregating all results together using “standard” aggregations, using
IPreflightAggregatedResultsUtils::ApplyAllStandardAggregations:
InterfacePtr<const IPreflightResultRecordTable> iTable(
iUtils->ApplyAllStandardAggregations(iProcess, iRawTable));
The standard aggregations utilities do the following:
z
All records corresponding to marking operations are aggregated into the containing object
(shape, text run, table cell, etc.) that contains them. The subpart is determined automatically. Only those records with the same subpart and criteria are aggregated together. Thus,
you are left with “high level” objects that can be selected, rather than artwork, which cannot.
z
Text runs that share the same criteria, subpart, and value are aggregated into a larger text
range.
Implementing Preflight Rules
485
Implementing Preflight Rules
IPreflightRuleVisitor method examples
z
Records that share the same criteria and values, but different subparts, are aggregated
together. This is done last.
You do not need to call ApplyAllStandardAggregations. You do not need to make this call if
your rule looks at only one object or does not look at objects that would ever be aggregated, or
you simply do not want to do any standard aggregation. You also can use any of the “lesser”
aggregation utilities, which do only one kind of aggregation, not all of them.
Combine records into buckets
After standard aggregation, the nodes are more or less ready to go into the tree; however,
depending on your rule’s desired presentation, you may want to put them in buckets. The
native InDesign rules put records into buckets in two ways
z
Single node represents multiple violation records. For example, the image-resolution rule
can fail in multiple ways on the same object; in particular, when the images come from a
PDF, EPS, or INDD file. Rather than creating a node for each failure, there is a single node
for the PDF/EPS/INDD, and all failures are enumerated within that node’s information text.
z
Nodes grouped under a rule or criteria — If you just add violation nodes under parentID,
typically they appear as children of a category, which usually is not desirable. (The exception is nodes that can involve only one object, like the document object; these should share
both the criteria and violation in a single node.) Rather, you want a criteria node, which
describes the error, with the individual violations of that error added as children of the criteria.
Several methods in IPreflightAggregatedResultsUtils are useful for bucketizing. Some of these
are shown in Table 124.
TABLE 124 Bucketizing IPreflightAggregatedResultsUtils methods
486
Method
Purpose/use
CreateSubTable
Use this when you have manually determined how to
divide entries and want to create a new table containing
just those entries (for example, for passing to your shared
node-building subroutine).
CreateSubTableByCriteria
Use this to put all records under a common criteria node.
This method creates a table with only the matching
criteria, so you know all entries share that criteria.
CreateTablesByCriteriaCr
eateTablesByAuxString
Use this to group entries under the same criteria or aux
strings, and might have quite a few different criteria (or in
the case of strings, you probably don’t know ahead of time
how many strings to expect). This method gives you a
vector of tables, each sharing entries of the same criteria or
string. You can then iterate over the vector and pass each
to a common formatting utility.
Implementing Preflight Rules
IPreflightRuleVisitor method examples
Method
Purpose/use
CreateTablesByObject
Use this to ensure that an object appears in only one node.
If there are multiple criteria/subparts under that object,
you just want to add more information strings. For
example, this how the stroke-width rule works.
You can use several of these together. For example, you could create tables by criteria, then for
each of those tables, create tables by object. Remember, tables are lightweight and ref counted,
so there is no record copying when creating subtables. Thus, you can use these utilities to your
advantage, without worrying too much about overhead.
As an example, the colorspace rule’s aggregation method looks like this:
// We only want one entry per object, so divide the table into multiple tables,
// one per common object.
IPreflightAggregatedResultsUtils::VectorOfTables byObject;
iUtils->CreateTablesByObject(iTable, byObject);
int32 i, n;
for(i = 0; i < byObject.size(); i++)
{
// Create a violation node for this object; then for each record just
// add strings corresponding to that record.
IPreflightResultRecordTable* iObjTable = byObject[i];
InterfacePtr<const IPreflightResultRecord> iRec(
iObjTable->QueryNthRecord(0));
InterfacePtr<IPreflightResultNodeInfo> iNode(
iUtils->CreateViolationNode(iRec->GetObjectID()));
InterfacePtr<IPreflightResultNodeData> iData(iNode, UseDefaultIID());
iData->AddInfoString(IPreflightResultNodeData::kFieldRequired, sRequirement);
// For each subpart create its own Problem line.
for(n = 0; n < iObjTable->GetRecordCount(); n++)
{
InterfacePtr<const IPreflightResultRecord> iRec(
ObjTable->QueryNthRecord(n));
Data->AddInfoString(IPreflightResultNodeData::kFieldProblem,
MyMakeProblemString(iRec));
}
}
In the above example, some of the details are omitted. The point of the example is to show how
the original, raw table, having been reduced through standard aggregation, is divided and conquered via CreateTablesByObject.
Implementing Preflight Rules
487
Implementing Preflight Rules
IPreflightRuleVisitor method examples
Generating result nodes
The final step is generating the nodes that appear in the results tree. There are three substeps:
z
Create the node — You can create four kinds of nodes through the existing utilities: Generic,
Criteria, Violation, and Criteria/Violation.
z
Fill in the node’s short-form string (i.e., the one that appears in the tree view panel) — Violation nodes are self-describing through the object model, but the others require you to specify the string you want to appear. These strings also appear in the reports.
z
Fill in the node’s Info strings — These are label-value pairs, like “Required:” and “Text must
be either blue or green.” These strings appear in the Info pane below the tree view. They also
appear in the reports.
To create the nodes, the IPreflightAggregatedResultsUtils methods in Table 125 are handy.
TABLE 125 Methods for node creation
Method
Purpose
CreateGenericNode
Creates a generic node. Typically, this is a child of
a criteria node and is used to group together like
items by something other than a criteria. The
native rules use this when they want to group by
font name, for example.
AddCriteriaNode
Creates a criteria node and immediately adds it to
the tree. You specify the name at the time of
creation. Criteria nodes typically have no info
text, which is why this does both operations.
CreateViolationNode
Creates a violation node; this is a node that
describes a particular problem with a particular
object. Thus, the name is generated automatically
from the passed-in object ID, and all other
behaviors are inherited automatically (page
number reporting, selection, etc).
CreateCriteriaAndViolationNode
Creates a combination criteria and violation node.
Typically, this is used for rules that check only the
document or other singleton. There is not much
point in having a criteria node with a violation
child; that is unnecessary hierarchy.
In most cases, you take the result of these methods and start adding info strings. Unless you are
going to build your own implementation of IPreflightResultNodeInfo, you do this via code like
the following:
488
Implementing Preflight Rules
IPreflightRuleVisitor method examples
InterfacePtr<IPreflightResultNodeInfo> iNode(
iUtils->CreateViolationNode(iRec->GetObjectID()));
InterfacePtr<IPreflightResultNodeData> iData(iNode, UseDefaultIID());
iData->AddInfoString(IPreflightResultNodeData::kFieldProblem,
"The object has the wrong color.");
iData->AddInfoString(IPreflightResultNodeData::kFieldFix,
"Set the color to something else.")
How you fill in the strings is up to you, but there are some utility methods that help with the
task. Some of them are explained in Table 126. Normally, these are used along with StringUtils::ReplaceStringParameters() to handle localization properly.
TABLE 126 IPreflightAggregatedResultsUtils methods that help with formatting
Method
Purpose/application
IsPlacedContent
Determines whether a given record corresponds to a placed
element or a subpart of a placed element. In many cases, the
recommended fix string depends on the distinction.
FormatXMeasureForma
tYMeasureFormatLine
MeasureFormatTextSize
MeasureFormatResoluti
onFormatAsInteger
Returns pre-translated strings ready for substitution via
StringUtils::ReplaceStringParameters(). These take
PMReals and return PMStrings. See below for an example.
GetSubpartsDescription
Returns a description of the subpart (or subparts) for a
record. For example, if the subparts are
kPreflightOM_NativeStroke and kPreflightOSP_NativeFill,
this returns the string “Stroke, Fill”. This leverages the
object model.
Here is an example of how you might use the above:
// Set up requirement string
PreflightRuleDataHelper dh(this);
PMString sRequired("Min size is ^1", PMString::kTranslateDuringCall);
StringUtils::ReplaceStringParameters(&sRequired,
iUtils->FormatXMeasure(dh.GetRealParam("min_size")));
iData->AddInfoString(IPreflightResultNodeData::kFieldRequired, sRequired);
// Set up problem string
PMString sProblem("Actual size is ^1", PMString::kTranslateDuringCall);
StringUtils::ReplaceStringParameters(&sProblem,
iUtils->FormatXMeasure(iRec->GetMinValue()));
if (!iRec->HasCommonValue()) sProblem += PMString(
" (smallest)", PMString::kTranslateDuringCall);
iData->AddInfoString(IPreflightResultNodeData::kFieldProblem, sProblem);
Implementing Preflight Rules
489
Implementing Preflight Rules
IPreflightRuleVisitor method examples
// Set up fix string
if (iUtils->IsPlacedContent(iRec))
{
iData->AddInfoString(IPreflightResultNodeData::kFieldFix,
"Fix it in the original file.");
}
else
{
iData->AddInfoString(IPreflightResultNodeData::kFieldFix ,
"Use Object > Size to change the size.");
}
The above example is almost literally the code found in the InDesign native rules. It takes a
number of lines of code to write the string generation, but this is unavoidable; the presentation
depends entirely on the nature of your rule and how verbose you want the results to be. You do
not need to be this fancy, but users expect more in-depth information to be presented where
possible.
IPreflightRuleVisitor::UpdateRuleData
This method is called when a document is opened, at start-up, or when a new rule is created. It
examines the rule parameters, establishes defaults, and removes any parameters that are no
longer needed. This is analogous to the class constructor and converters used for stream-based
data in other interface implementations.
Typically, this method leverages the utilities in PreflightRuleDataHelper, which makes it easier
to set defaults.
The following example is from the colorspace rule:
virtual void UpdateRuleData() override
{
PreflightRuleDataHelper dh(this);
dh.SetBoolParamDefault(kParamNoRGB, kFalse);
dh.SetBoolParamDefault(kParamNoCMYK, kFalse);
dh.SetBoolParamDefault(kParamNoGray, kFalse);
dh.SetBoolParamDefault(kParamNoLAB, kFalse);
dh.SetBoolParamDefault(kParamNoSpot, kFalse);}
By using SetBoolParamDefault, there will be no change if the parameter was already initialized,
so this implementation works for both updating and initializing.
IPreflightRuleVisitor::ValidateRuleData
Preflight calls this method whenever a script sets a parameter on a rule. Since there is no range
metadata in the rule data, the rule must perform any required checking. If the rule exposes no
parameters with range restrictions, it can simply return kSuccess.
490
Implementing Preflight Rules
More on specific objects
The following example is from the bleed/trim hazard rule:
virtual ErrorCode ValidateRuleData
(
const IPreflightRuleData::Key& key,
const ScriptData& proposedValue
) const
{
PreflightRuleDataHelper dh(this);
if (!dh.DataExists (key.GetPlatformString ().c_str ()))
return kInvalidParameterError;
PMReal value = -1;
if (key == kParamLiveAreaL || key == kParamLiveAreaR ||
key == kParamLiveAreaT || key == kParamLiveAreaB)
{
if (proposedValue.GetType () == ScriptData::s_double)
{
proposedValue.GetPMReal (&value);
if (value < 0 || value > kMaxValue)
return kOutOfRangeError;
}
else
return kInvalidParameterError;
}
return kSuccess;
}
More on specific objects
This section describes how to work with particular objects in the model in your Visitor.
Native, UID-based objects
These are the simplest to work with because their preflight object boss does not have any additional interfaces. You simply use IPreflightObject::GetModelRef to get an interface on a native
(persistent) InDesign boss in the document database. There is no point in preflight providing
additional interfaces on the preflight object in this case, because you can get this information
from the UID and all existing InDesign model interfaces.
For example, the Scaled Graphics rule always looks at graphic page items, so its Visit code looks
like the following:
virtual void Visit(IPreflightVisitInfo* iVisit) override
{
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
InterfacePtr<IShape> iShape(iObj->GetModelRef(), UseDefaultIID());
InterfacePtr<ITransform> iTransform(iShape, UseDefaultIID());
// etc
Implementing Preflight Rules
491
Implementing Preflight Rules
More on specific objects
This version has error checking removed, for readability. Initially, you may want to put in several asserts in your rules, to satisfy yourself that the visitor is looking at the object it thinks it is.
Artwork
A rule can ask to look at marking operations. This is a powerful capability because you can
look at exactly what marking is done for a particular page item or graphic element, regardless
of whether you “own” the implementation of that object. Preflight renders the spread into a
port, then builds a tree data structure based on what it finds. That tree consists of the marking
operations, any logical artwork groups (for example, transparency groups), and groups corresponding to native InDesign elements like wax runs and page items.
Unlike page items, artwork objects are not UID based, so all the information exists on the preflight object boss in interfaces that you query.
Table 127 shows the artwork object types and their related interfaces.
TABLE 127 Artwork preflight object types
Preflight object
class ID
Description
kPreflightOM_
ArtworkMark
The most commonly inspected object type for artwork rules. The boss
that the service creates for this object type has the following interfaces:
kPreflightOM_
ArtworkGroup
492
z
IPreflightArtworkMarkInfo — Provides information about the marking operation, such as the
geometry, colorspace, opacity, and text versus path.
z
IPreflightArtworkContext — Relates the marking
operation to any context objects of which it is a
child. This is useful when you need to know
whether the mark is, say, drawn as part of rendering
a text run or graphic.
Placeholder for future implementation.
Implementing Preflight Rules
More on specific objects
Preflight object
class ID
Description
kPreflightOM_
ArtworkContext
“Metadata” contexts that enclose artwork. In other words, the marking
operation tells you that a stroke of a specified width and color exists; the
context tells you what that stroke is associated with.Preflight supports the
following contexts:
z
Shape — The IShape-supporting page item containing the artwork. This also records the “subpart” of
the shape: stroke, fill, adornments, and so on. Shape
information is obtained via IPreflightArtworkShapeContext.
z
Text — The text with which the artwork is associated, if any. This is defined by the text model, index,
and range. Text information is obtained via
IPreflightArtworkTextContext.
z
Table — The associated table and cells, if any. This
information is available from IPreflightArtworkTableContext.
z
OPI — If the artwork is being drawn as part of an
OPI reference, this context is present. It has no data
and only indicates that there is an OPI context.
To determine what contexts an artwork mark is associated with, you start
with the IPreflightArtworkContext interface on the mark boss; that
interface allows you to search for the nearest, or any, context of a given
type. You also can walk the context tree manually.You can ask to visit these
context types directly, although this would be uncommon. For example,
the OPI rule looks at OPI contexts, because it cares about only their
existence.
Regarding artwork-based rules: if a rule wants to see any of these classes (or any other class that
lists these classes as parents), preflight forces artwork expansion, which is fairly expensive compared to looking at the attributes of native objects. This is one reason why the out-of-the-box
default profile (Basic) does not have any artwork rules. This is not to say you should not have
such rules; many native rules (like colorspace and CMY marking) are artwork rules. If you can
get the information you need by inspecting attributes instead of looking at artwork, however,
you probably should use attributes, which has much lower overhead.
An artwork-rule visitor usually looks something like that shown below. This is actual code
from the colorspace rule.
virtual void Visit(IPreflightVisitInfo* iVisit) override
{
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
InterfacePtr<IPreflightArtworkContext> iContext(iObj, UseDefaultIID());
if (iContext && iContext->GetShapeContextDepth(
kPreflightOSP_GraphicProxy) > 0)
{
Implementing Preflight Rules
493
Implementing Preflight Rules
More on specific objects
// Artwork is part of a graphic proxy (missing graphic); ignore
}
else
{
InterfacePtr<IPreflightArtworkMarkInfo> iMark
(iObj, UseDefaultIID());
InterfacePtr<IPreflightArtworkPaintInfo> iColor
(iMark->QueryColorPaintInfo());
if (!iColor) return;
InterfacePtr<IPreflightArtworkCSInfo> iCS
(iColor->QueryColorSpace());
// etc
Note the use of IPreflightArtworkContext, which looks for an enclosing shape context with a
subpart of kPreflightOSP_GraphicProxy. This identifies the marking operation as proxyrelated. Your rule may or may not want to exclude proxy artwork in this way.
Once the rule determines that the artwork is not part of a proxy, it gathers the marking data,
including the paint characteristics. All these secondary interfaces (e.g., IPreflightArtworkPaintInfo) are obtained via methods in IPreflightArtworkMarkInfo. In the case of the colorspace rule, it simply calls various IPreflightArtworkCSInfo methods to determine whether the
marking operation violates its rule parameters.
Text runs and ranges
Text runs are not persistent objects, so IPreflightObject::GetModelRef will not give you a UID
for one of these. Thus, a text run or range object has all its data on the preflight object boss
itself.
To obtain information about text objects, query for the following interfaces:
z
IPreflightTextRangeInfo — This interface is available on all text-range objects
(kPreflightOM_TextRange, kPreflightOM_TextCharacter, kPreflightOM_WaxRun, and
kPreflightOM_TextParcel). It provides the text model, thread, parcel, index, and span
information for the text object. To obtain further information, you use the InDesign text
model.
z
IPreflightWaxInfo — This interface is available only on a kPreflightOM_WaxRun object,
not the other text types, since they do not necessarily correspond to a wax-run boundary.
This provides the number of glyphs in the run, as well as a method that provides an
IWaxRun.
An example of the use of text interfaces follows (from the missing-fonts rule):
virtual void Visit(IPreflightVisitInfo* iVisit) override
{
const PreflightObjectID& objID = iVisit->GetObjectID();
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
if (objID.GetClassID() == kPreflightOM_WaxRun)
{
InterfacePtr<IPreflightWaxInfo> iWax(iObj, UseDefaultIID());
if (!iWax) return;
494
Implementing Preflight Rules
More on specific objects
InterfacePtr<const IWaxRun> iRun(iWax->QueryRun());
if (!iRun) return;
InterfacePtr<const IWaxRenderData> iRender(iRun, UseDefaultIID());
if (!iRender) return;
if (iRender->FontFaceMissing())
{
InterfacePtr<IPreflightResultRecord> iRec(
iVisit->CreateResultRecord());
iRec->SetCriteria(kPreflightRC_Default);
PMString fontName = iRender->GetFontName();
fontName.SetTranslatable(kFalse);
iRec->SetAuxString(fontName);
}
}
// etc
Tables, rows, columns, and cells
Many table elements in a document are composed of elements and may or may not have an
associated UID. Thus, some of the table objects are UID based (e.g., use GetModelRef) and
some are not. Those that are not have additional data on the preflight object boss.
For non-UID based elements, the IPreflightTableCellInfo interface on the preflight object boss
provides the necessary data, from which you can navigate to the InDesign model. The objects
that support this interface are kPreflightOM_TableCell, kPreflightOM_TableArea,
kPreflightOM_TableFrameCell, and kPreflightOM_TableFrameArea.
The following is an abbreviated example from the overprint rule, which looks for any overprint
attributes that are set. In this case, the rule is interested in page items, text, and tables, all of
which have overprint attributes.
virtual void Visit(IPreflightVisitInfo* iVisit) override
{
PreflightObjectID objID = iVisit->GetObjectID();
InterfacePtr<IPreflightObject> iObj(iVisit->QueryObject());
if (objID.GetClassID() == kPreflightOM_PageItem)
{
InterfacePtr<IGraphicStyleDescriptor> iStyle(iObj->GetModelRef(),
UseDefaultIID());
// Check graphic attributes..
}
else if (objID.GetClassID() == kPreflightOM_WaxRun)
{
InterfacePtr<IPreflightTextRangeInfo> iRange(iObj, UseDefaultIID());
if (!iRange) return;
// Check text attributes via the text model...
}
Implementing Preflight Rules
495
Implementing Preflight Rules
More on specific objects
else if (objID.GetClassID() == kPreflightOM_TableFrame)
{
InterfacePtr<ITableFrame> iFrame(iObj->GetModelRef(), UseDefaultIID());
if (!iFrame) return;
// Check the table attributes...
}
else if (objID.GetClassID() == kPreflightOM_TableFrameCell)
{
InterfacePtr<IPreflightTableCellInfo> iCellInfo(iObj, UseDefaultIID());
if (!iCellInfo) return;
// Check the cell attributes..
}
}
Note how the rule must use a different strategy for each object type. For page items and tables
there is a UID to work with, so it queries the model directly. For text and cells, which are composed entities, the rule uses the interfaces hosted on the preflight object directly: IPreflightTextRangeInfo for text and IPreflightTableCellInfo for table cells.
For example, the first thing it does when looking for cell attributes is to look up the table
attribute’s accessor interface from the table model, using IPreflightTableCellInfo::GetTableModelRef:
InterfacePtr<ITableAttrAccessor> iTableAttrs(
iCellInfo->GetTableModelRef(), UseDefaultIID());
if (!iTableAttrs) return;
496
XML Fundamentals
Introduction
XML Fundamentals
This chapter describes features of the InDesign XML subsystem relevant to plug-in developers,
and the architecture that supports them. It explains how client code can use these features and
take advantage of the XML-related extension patterns in the InDesign API.
For use cases, see the “XML” chapter of Adobe InDesign CS4 Solutions.
Introduction
XML-based workflow
There are several benefits of using an XML workflow in publishing:
z
Separation of concerns; for example, maintaining content and presentation information
independently.
z
Working with standards that are open, relatively stable, and defined by experts in a domain.
This avoids being locked into proprietary, unpublished file formats and helps reduce implementation effort. For example, having standards available like NITF (http://www.nitf.org)
and NewsML (http://www.newsml.org) prevents you from having to reinvent a language to
store articles, transmit news feeds, and support a newspaper workflow.
z
Relatively easy processing, because you have a textual representation of XML documents.
z
Relatively easy re-purposing of content for other media, like HTML and Mobile SVG, in
addition to print publishing and PDF delivery. The main advantage is that the content can
be single-sourced but published to multiple media with relatively little effort.
z
Relatively easy manipulation of XML documents, thanks to the availability of many XMLaware tools, software developer toolkits, and standardized APIs.
z
Existence of many databases that can produce and consume XML data, some natively.
Disadvantages to an XML workflow can include the following:
z
Journalists and graphic designers are unlikely to want to learn about XML. A system
designed around XML may need to shield end users from having to create mark-up manually. For example, templates can be pre-tagged.
z
Some XML technologies are not straightforward to use and may require some effort to
understand (for example, XSLT/XPath), even for capable software developers.
XML Fundamentals
497
XML Fundamentals
Terminology
Using XML with InDesign
Using XML-based data and XML templates in InDesign has some benefits to a programmer,
which include the following:
z
There are mature XML toolkits and technologies to manipulate XML data (see
http://xml.coverpages.org/software.html), which make it relatively straightforward to
manipulate XML data before import or after export. For example, using XSLT to manipulate XML data is particularly convenient in the context of the InDesign import architecture.
You can develop and debug your XSL stylesheets independently of InDesign, using existing
XSLT-aware tools, but deploy them as part of a workflow involving InDesign if you use its
features like the XML transformer. See “XML transformer” on page 554.
z
Importing graphics into tagged placeholders is a convenient way to import and place graphics without writing much (if any) code. See “Tagged graphic placeholder” on page 534.
z
The XML API of InDesign is well documented. The command facades and suite interfaces
in the XML API make it relatively straightforward to write client code. See “Key client API”
on page 552.
Terminology
See the “Glossary” for definitions of terms. Table 128 lists terms used in this chapter and sections that relate to them.
TABLE 128 Terminology
498
Term
See ...
Acquirer; acquirer filter
“Extension patterns” on page 553
Backing store
“Backing store” on page 509
Comment, XML
“Processing instructions and comments” on page 547
Content handler
“Extension patterns” on page 553
Content item
“Content items” on page 506
Document element
“Document element and root element” on page 508
Document Type Declaration
(DTD)
“DTD” on page 545
Entity
“SAX-entity resolver” on page 557
Logical structure
“Native document model and logical structure” on
page 505 and “Elements and content” on page 534
Placed element; placed
content
“Unplaced content versus placed content” on page 518
XML Fundamentals
XML features at a glance
Term
See ...
Processing instruction (PI)
“Processing instructions and comments” on page 547
Repeating element
“Importing repeating elements” on page 519
Root element
“Document element and root element” on page 508
SBOS (small boss object store)
“Persistence and the backing store” on page 510
Tag
“Tags” on page 529
Unplaced element; unplaced
content
“Unplaced content versus placed content” on page 518
XML element
“Elements and content” on page 534
XML features at a glance
XML extension patterns
Several XML extension patterns depend on the re-factored XML import architecture. You can
implement these extension patterns to customize how XML data is imported and exported and
other XML-related features. See “Extension patterns” on page 553.
Tagging in tables and inline graphics
InDesign supports tagging of tables and table cells. See “Tagged tables” on page 543. Inline
graphics can be tagged, and placeholders for inline graphics can be created in XML templates.
See “Tagged inline graphics” on page 542.
Throw away unmatched existing (right)
It may be desirable to filter out elements in your XML template that are not matched by elements in the incoming XML data. When enabled, this features discards unmatched existing
elements in an XML template. See “Throwing away unmatched existing elements on import
(delete unmatched right)” on page 520. The feature is implemented by an import-matchmaker
service, an extension pattern described in “XMl-import matchmaker” on page 555.
XML Fundamentals
499
XML Fundamentals
XML features at a glance
Throw away unmatched incoming (left)
It may be desirable to filter out incoming XML elements that have no corresponding match in
your XML template. When enabled, this feature discards inbound elements in the XML-based
data being imported that have no match in the XML template into which the data is being
imported. In API terms, the feature is called “throw away unmatched left.” See “Throwing away
unmatched incoming elements on XML import” on page 522. The feature is implemented by
an import-matchmaker service, described in “XMl-import matchmaker” on page 555.
Importing repeating elements
This feature allows text styling from elements in an XML template to be preserved (within one
story) when repeated elements are imported. It is described in “Importing repeating elements”
on page 519.
Importing CALS table
This feature allows CALS table to be imported as InDesign table. It is described in “Importing a
CALS table as an InDesign table” on page 524.
Support for DOM core level 2
This specification is implemented by the features related to matching and transforming on
import. Import matchmakers and XML transformers that operate on the DOM (Document
Object Model) are described in “Extension patterns” on page 553. For the DOM Level 2 Core
specification, see http://www.w3c.org/TR/2000/REC-DOM-Level-2-Core-20001113.
Support table- and cell-styles import
This feature allows table and cell styles to be applied to InDesign table when is imported. It is
described in “Support table and cell styles when importing an InDesign table” on page 524.
Support XML-rules processing
The scripting-based XML-rules feature allows an XML DOM to be altered after an XML file is
imported. XML rules also can be triggered by application events, such as open, place, and close.
The feature is described in the “XML Rules” chapter of Adobe InDesign CS4 Scripting Guide.
Snippets
The snippet architecture is described in the “Snippet Fundamentals” chapter. The snippet
architecture depends on the XML subsystem, as well as on INX (InDesign Interchange) or
IDML (InDesign Markup Language) import and export. Snippets are XML-based representations of parts of a document in INX or IDML file format.
500
XML Fundamentals
The user interface for XML
The user interface for XML
Structure view
Figure 215 shows the structure view, a representation of the logical structure of a document
that lets you view and interact with its logical structure. You can make selections in this view
and create new elements as children of the selection, delete elements, create or delete attributes,
and so on. Figure 215 shows the structure view with text snippets visible, which provides context to see what content items in the document are associated with the elements in the logical
structure.
FIGURE 215
Structure view
Depending on your workflow, it may be desirable to control the visibility and other properties
of the Structure pane. These properties are controlled by preferences. See “XML-related preferences” on page 549.
When you tag a graphic frame as a placeholder, the structure view draws a cross through the
icon representing the element to indicate its placeholder status. There are also other icons that
represent whether an element is associated with text content or image-based content.
Tags in layout view and story view
Figure 216 shows content that has been tagged, rendered in the layout view of the application.
In layout view, tags are shown as brackets with a width of zero; that is, they are intended to have
no effect on text composition. Start tags are indicated by brackets pointing right ([), and end
tags by brackets pointing left (]). The tag name associated with the tag markers is encoded by
color, corresponding to the color displayed in the Tags panel. Multiple tags can exist at the
same position in layout view.
XML Fundamentals
501
XML Fundamentals
The user interface for XML
FIGURE 216
Tags in layout view
Story view makes it somewhat easier to understand the logical structure of a story and edit its
structured content. Figure 217 shows structured content rendered in story view. In this view,
the names of tags are shown explicitly.
FIGURE 217
Tags in story view
Tags panel
Figure 218 shows part of the Tags panel, which has a list of tags that can be applied to document content. Each tag has a name and associated color. In this example, some tags were created when a DTD was imported, shown with icons at the right-hand end of the list elements.
FIGURE 218
502
Tags panel
XML Fundamentals
The user interface for XML
To open the Tags panel, choose Window > Tags. By default, the list of tags shown is sorted
alphabetically by tag name; the list has no inherent logical structure. The complete set of tags in
an imported DTD or imported tag list is shown, regardless of what is selected in the structure
view. To populate the tag list, end users can load a set of tags with the Load Tags menu command in the Tags-panel menu.
You can create a new tag through the Tags panel; this adds an entry to the tag list. Creating a
new tag means it is available to mark up content or tag placeholder graphic frames for content.
You can delete a tag through the Tags panel. You can export the tag with the Save Tags menu
command in the Tags-panel menu.
Mapping between tags and styles
You can choose to apply a mapping between element names and styles (character and paragraph styles). Once the mapping is applied to a document, any marked-up text is styled as specified by the style on the right-hand side of this mapping. See Figure 219.
FIGURE 219
User interface for mapping tags to styles
An important use case for mapping tags to styles is preparing a document for XML import.
Assuming XML-based content does not carry style information, the tag-to-style map is a simple mechanism to apply styling to inbound XML-based content; more complex mechanisms
include styling based on XSLT or a proprietary mechanism like an Adobe FrameMaker Element Definition Document (EDD). The FrameMaker EDD effectively is a style sheet containing rules specifying how to map tags to styles, but in a context-dependent way.
You can apply structure to an unstructured InDesign document by choosing to map styles to
tags. The effect of this is to mark up the text ranges that have the styles on the left-hand side of
the style-to-tag mapping. See Figure 220.
XML Fundamentals
503
XML Fundamentals
The user interface for XML
FIGURE 220
User interface for mapping styles to tags
Validation window
The XML-validation-errors pane appears below the Structure pane; see Figure 221. The XMLvalidation-errors pane contains hyperlinks. If the end user clicks on the hyperlinks, InDesign
tries to fix the validation errors. This may not always have the desired effect; for example, it
may delete content that was tagged incorrectly.
FIGURE 221
XML-validation-errors pane
Other
Other components of the user interface for XML—like context-sensitive menus for tagging the
selection—are beyond the scope of this chapter.
504
XML Fundamentals
XML model
XML model
Native document model and logical structure
The InDesign native document model specifies how end-user documents are represented by
InDesign. The InDesign native document model consists of both the description of the document in terms of boss classes and the scripting DOM. End users directly manipulate spreads,
pages, and the items they contain; these abstractions are represented by boss classes in the
native document model or objects in the scripting DOM.
For example, a spread is represented by kSpreadBoss in the low-level model or boss DOM.
Spread contents are organized into a hierarchy (IHierarchy). In the scripting DOM, a spread is
represented by a Spread object, with properties like Pages or PageItems to represent the object
hierarchy at a higher level of abstraction.
The logical structure of an InDesign document is specified by the user; for example, by tagging
content items in the layout or importing their XML data. An XML template, consisting of
tagged placeholders for text and graphics, provides the user with a way to define how userdomain data is mapped into the native model.
The logical structure of an InDesign document is organized into a hierarchy of XML elements
(see IIDXMLElement) representing logical relationships between content. The IIDXMLElement interface plays the same role in the logical structure that IHierarchy plays in the spread
hierarchy. The key operation that modifies logical structure is tagging document objects, like
graphic frames, stories, and text ranges. Tagging establishes associations between objects in the
native document model and the logical structure. See “Elements and content” on page 534.
Importing XML-based data into an InDesign document creates new elements in the logical
structure (see “Importing XML” on page 511). If there are tagged placeholders, content can be
flowed into these placeholders. The logical structure of an InDesign document also can be
exported as an XML file (see “Exporting XML” on page 524). You also can create new elements
by mapping styles to tags (see “Tags” on page 529).
Elements and attributes
Elements represent user-defined logical relationships between content items in a document.
Elements are created by InDesign when XML data is imported or content is tagged. The implementation of elements in InDesign is closely related to the XML specification
(http://www.w3.org/TR/REC-xml), which defines an element as follows:
“Each XML document contains one or more elements, the boundaries of which are either
delimited by start-tags and end-tags, or, for empty elements, by an empty-element tag. Each
element has a type, identified by name, sometimes called its “generic identifier” (GI), and may
have a set of attribute specifications. Each attribute specification has a name and a value.”
You can create a new element in the logical structure with the New Element menu command
on the Structure-pane menu. The behavior of this feature depends on whether the element in
the structure view already is placed and, if placed, what is the associated content item. You can
XML Fundamentals
505
XML Fundamentals
XML model
choose a tag for the element or create a new one. The tag stores the element’s name and color in
the user interface.
Figure 222 shows what happens when the root element is selected in the structure view and the
user chooses the New Element menu command. The application adds a child element to the
logical structure’s root element. By default, the child element is added at the end of the collection of child elements of the root element. In this example, the root element <book> is selected
and a new element <article> is added. The new element is inserted into the logical structure as
the last child of the root element and is shown as unplaced content, since the top-level element
was not placed; only its child elements are placed content.
FIGURE 222
Inserting a new element
You can add an element to the logical structure in the structure view, in which case the element
is not directly associated with document content. You can indirectly add one or more elements
in another view, like layout view, by tagging (e.g., tagging a text range or graphic placeholder
frame).
Attributes are properties of elements. Attributes can be added by end users through the structure view, when a node is selected that supports added attributes. Some elements, like comments, processing instructions, and the DTD element, do not support the addition of
attributes.
Elements are modeled by boss classes that have the IIDXMLElement interface. Tags are modelled by the kXMLTagBoss boss class. See “Tags” on page 529. Attributes are modeled as properties of elements (IIDXMLElement). There are facades that make it relatively easy to change
the properties of elements, tags, and attributes (IXMLElementCommands, IXMLTagCommands, and IXMLAttributeCommands).
Content items
Elements (IIDXMLElement) are created when content items are tagged. Elements also may be
created on import of XML data. New content items can be created when XML content is flowed
into an XML template. The main point is that elements are defined in the user domain and represent logical relationships between content items—or between elements, in the case of structural elements.
Content items are objects in the layout or text domain that can be tagged to create an association with an element in the logical structure of an InDesign document. Many content items can
be tagged:
506
XML Fundamentals
XML model
z
Placeholders for graphics (kPlaceHolderItemBoss)
z
Placeholders for inline graphics (kILPlaceHolderPageItemBoss)
z
Images (kImageItem, any other descendants of kImageBaseItem)
z
Stories (kTextStoryBoss)
z
Text ranges
z
Tables (kTableModelBoss) and cells within them
z
Text within table cells
There is no easy way to identify a content item. Many but not all the objects that can be tagged
have the IXMLReferenceData interface. If you inspect the API reference documentation for
IXMLReferenceData and see the boss classes in the native document model that aggregate this
interface, you get some indication of the variety of document objects that can be tagged.
References to elements and content items
All elements in the logical structure of an InDesign document expose the IIDXMLElement
interface; therefore, one type of reference to an element is an interface pointer of type IIDXMLElement. A more convenient type for programmers to work with is XMLReference, an
instance of which can be obtained from IIDXMLElement::GetXMLReference. This XMLReference is just another way to refer to the same element.
Conversely, given an XMLReference, you can obtain an IIDXMLElement interface pointer by
using XMLReference::Instantiate. XMLReference is the XML subsystem’s equivalent of
UIDRef; note that UIDRef applies only to UID-based objects, and elements in the logical structure are not UID-based.
XMLContentReference refers to a chunk of content rather than an element. Do not mistake the
XMLContentReference type for XMLReference: XMLContentReference is a reference to a content item, which might be a UID-based like a graphic frame or a non-UID-based object like a
table cell. An instance of XMLContentReference can be obtained from IIDXMLElement::GetContentReference. This generalizes and replaces IIDXMLElement::GetContentItem, which is
deprecated.
An instance of XMLContentReference can tell you the type of object with which it is associated, through XMLContentReference::GetContentType. For example, kContentType_Normal
is used for elements associated with a UID-based object in the layout; this could be something
like a tagged graphic placeholder.
For more information on methods and use of these key types, see the API reference documentation for XMLReference and XMLContentReference. The main difference between XMLReference and XMLContentReference is that XMLReference refers to an element
(IIDXMLElement), whereas XMLContentReference refers to a content item.
XML Fundamentals
507
XML Fundamentals
XML model
Document element and root element
The IXMLReferenceData interface on kDocBoss stores a reference to the document element
(instance of kXMLDocumentBoss). There also is an interface on the document (kDocBoss),
with a different PMIID referring to the DTD element (kXMLDTDBoss), if one is associated
with the document. The document element associates with the root element both through
IXMLReferenceData (as shown in Figure 223) and through IIDXMLElement::GetNthChild,
because the document element is the parent of the root element. The class diagram in
Figure 223 shows associations involving the document element (kXMLDocumentBoss) and
other classes.
FIGURE 223
Document element, root element, and DTD
1
«boss»
kDocBoss
IXMLReferenceData::GetReference
IXMLUtils::QueryRootElement
1
IXMLUtils::QueryDocElement is
another way to navigate this
association.
1
1
IXMLReferenceData::GetReference
«boss»
kXMLDocumentBoss
1
«boss»
kTextXMLElementBoss
1
1
IXMLReferenceData::GetReference
0..1
Interface with PMIID of
IID_IXMLDTDXMLREFERENCEDATA
on kXMLDocumentBoss refers to
the DTD object, if one exists.
«boss»
kXMLDTDBoss
The document element is responsible for managing document-level information (like the
DTD) and comments and processing instructions that are peers of the root element.
The root element (kTextXMLElementBoss) appears as the root of the element hierarchy in the
structure view, although the root element is a child of the document element, which is not
shown in the structure view. By default, the root element for a new InDesign document has the
tag Root; you can change this to whatever the content model for your XML vocabulary
requires.
508
XML Fundamentals
XML model
The document element and root element—along with the DTD element and any top-level
comments (kXMLCommentBoss) and processing instructions (kXMLPIBoss)—are held in the
backing store.
Backing store
Each InDesign document contains one backing store, which stores XML elements that are
unplaced and elements that are siblings of the root element, like the document element, root
element, DTD element, comments, and processing instructions. The backing store is a story
(kTextStoryBoss) that is not accessible to the user and is present in every InDesign document.
The architecture of the backing store is described in “Persistence and the backing store” on
page 510.
The object diagram in Figure 224 shows the relationship between the document, backing store,
and elements stored in the backing store.
FIGURE 224
Backing store
There is exactly one backing store for
a document, an instance of a story
(kTextStoryBoss) that is not user-accessible.
It stores XML elements that are not on the page
and other elements, such as the
document element.
«boss»
document : kDocBoss
Backing store (opaque)
«boss»
doc-elem : kXMLDocumentBoss
XML Fundamentals
«boss»
article : kTextXMLElementBoss
«boss»
title : kTextXMLElementBoss
509
XML Fundamentals
XML model
The backing store is opaque to client code. Use facade methods like IXMLUtils::QueryDocElement and IXMLUtils::QueryRootElement to acquire objects like the document element
(kXMLDocumentBoss) or the root element (shown as the article object in Figure 224). To
acquire a reference to the backing store when you need it (e.g., for notification), use IXMLUtils::GetBackingStore.
For more information, see “Backing store and notification of changes in logical structure” on
page 558.
Persistence and the backing store
The main persistence mechanism for the native document model is UID-based; that is, each
instance of the classes shown in the model, like kSpreadBoss, has a record number or UID
(unique identifier) associated with it in the document’s database. The UID is unique only
within the context of a particular database, not globally.
UID-based objects also can store persistent boss objects without a UID; for example, this
model is used for attributes and widely within the table and XML subsystems. This as an example of container-managed persistence, to distinguish it from the conventional, UID-based persistence mechanism. The implementation of container-managed persistence in InDesign is
very efficient compared to UID-based persistence, which played a large part in the decision to
use it for both tables and XML. For more information, see the “Persistent Data and Data Conversion” chapter.
This object diagram in Figure 225 is an elaborated version of Figure 224. The BaseUID shown
for each object is the UID of the kTextStoryBoss implementing the backing store. The LSID
(logical structure ID) starts at 1 for the document element (kXMLDocumentBoss). See the API
reference documentation for XMLReference. The root element, with LSID of 2 in this diagram,
is an instance of kTextXMLElementBoss; it has a reference to a tag (TagUID=126, name=
“Root”) but not to a content item. The image element has a reference to a tag (TagUID=159,
name= “image”) and to a content item (UID=161, an instance of kPlaceHolderItemBoss).
510
XML Fundamentals
Importing XML
FIGURE 225
Backing store, with more detail
«boss»
document : kDocBoss
Backing store (opaque)
«boss»
doc-elem : kXMLDocumentBoss
«boss»
Root : kTextXMLElementBoss
«boss»
image : kTextXMLElementBoss
BaseUID = 111
LSID = 1
BaseUID = 111
LSID = 2
ContentItemUID = 0
TagUID = 126
BaseUID = 111
LSID = 3
ContentItemUID = 161
TagUID = 159
The backing store is an instance of a small-boss object store (SBOS), a storage container for
small boss objects. The container has a UID, but the objects it stores do not have UIDs. The
objects managed by the store have logical IDs (LSIDs) as keys. See the API reference documentation for XMLReference::GetLogicalID and related methods.
Importing XML
The application supports import of XML data encoded as any of the following:
z
UTF-8 (see http://www.ietf.org/rfc/rfc2279.txt)
z
UTF-16 (see http://www.ietf.org/rfc/rfc2781.txt)
z
Shift-JIS, a two-byte format for Japanese (see http://www.w3.org/TR/2000/NOTE-japanesexml-20000414)
XML Fundamentals
511
XML Fundamentals
Importing XML
There are two main workflows for importing XML data:
z
Importing XML data into the backing store (see “Backing store” on page 509), then manually placing it onto the layout to set up the associations between elements in the logical
structure and document object.
z
Creating an InDesign document as an XML template that contains placeholders, which are
tagged with names of the elements in the XML data to import. When the XML data is
imported, the contents of the XML file are flowed into the placeholders. The order in which
the elements appear in the logical structure determines how the content is flowed into the
tagged placeholders.
Import architecture
The import process is controlled by the XML importer (kXMLImporterBoss), which is the
main delegate for the low-level command that performs the import (kImportXMLFileCmdBoss). Unlike other forms of import—like EPS (import provider kEPSPlaceProviderBoss) and
PDF (import provider kPDFPlaceProviderBoss) import—importing XML data does not use
the standard import provider (IImportProvider) architecture. The importer controls the
import process by dispatching messages to extension patterns responsible for different parts of
the import sequence, as shown in Figure 226.
The sequence diagram shows some of the messages sent during XML import. The importer
object shown consists of the low-level import XML command (kImportXMLFileCmdBoss)
and its delegate (kXMLImporterBoss). The main points to note are the order in which the different extension patterns are sent messages and, for the transformer (IXMLTransformer), how
the stream transform precedes the DOM transform message.
512
XML Fundamentals
Importing XML
FIGURE 226
Importer
XML-import sequence
Acquirer
Transformer
Matchmaker
Serializer handler
Post-import iterator
IXMLAcquirerFilter::CreateStreamAndResolver
IXMLTransformer::TransformStream
IXMLTransformer::TransformDOM
IXMLImportMatchMaker::MatchDocument
ISAXDOMSerializerHandler::StartDocument
IXMLPostImportIteration::BeforeIteration
A low-level command (kImportXMLFileCmdBoss) is processed to import an XML file. Much
of the work is done by this command’s delegate, an instance of the kXMLImporterBoss boss
class, in its implementation of IXMLImporter::DoImport. The following operations are performed while importing an XML file:
1. The import parameters are specified through an instance of kImportXMLDataBoss. In particular, see IImportXMLData. This specifies either an IDFile (based on a path, which may
be a file path or an arbitrary URL-like string).
2. An acquirer service (IXMLAcquirerFilter) is located, which knows how to turn the import
specification (in IImportXMLData) into an XML stream, perhaps with a SAX entity
resolver. See “XML acquirer” on page 553 and “SAX-entity resolver” on page 557.
3. The SAX parser service (kXMLParserServiceBoss, ISAXServices) parses the inbound XML
from the stream.
4. XML transformer services (kXMLImporterTransformerService) are called to transform the
stream. See “XML transformer” on page 554.
5. At this point, the stream has been turned into an XML DOM. XML transformer services are
called to transform the DOM. See “XML transformer” on page 554.
6. The XML-import match driver calls XML-import matchmaker services (kXMLImportMatchMakerSignalService) to participate in matching the input XML data against the XML
XML Fundamentals
513
XML Fundamentals
Importing XML
template (the logical structure of the document into which content is being imported). See
“XMl-import matchmaker” on page 555.
7. Matches are registered by the XML match recorder (IXMLImportMatchRecorder) of the
importer (kXMLImporterBoss).
8. SAX DOM serializer handlers (ISAXDOMSerializerHandler) are called to create content
based on the DOM. See “SAX DOM serializer handler” on page 557. The parsing of the
final input XML data is controlled by a top-level SAX DOM serializer handler (kSAXDocumentHandlerBoss). Dependent SAX DOM serializer handlers turn parts of the DOM into
InDesign content.
9. Post-import responders are called to iterate the DOM and take appropriate action. See
IXMLPostImportIteration and “Post-import responder” on page 555.
When XML data is imported, the XML subsystem creates elements (IIDXMLElement) in the
logical structure. See “Elements and content” on page 534. Import of XML data results in content items (IXMLReferenceData) being populated with the imported content, if the content
items are tagged before import. After import, unplaced elements are held in the document’s
backing store. See “Backing store” on page 509.
The UML diagrams in Figure 227 show some interfaces involved in storing options for import
and export of XML data at the session level (kWorkspaceBoss) or document level (kDocWorkspaceBoss). There are other interfaces specific to individual components in the XML import
architecture (e.g., IXMLImportPreferences, which can be found on service boss classes), which
are not shown here.
FIGURE 227
Some preferences related to import and export
IXMLExportSessionOptions
«boss»
kWorkspaceBoss
IXMLImportOptions
IXMLExportOptions
«boss»
kDocBoss
IDocument
«boss»
kDocWorkspaceBoss
514
IXMLImportOptions
IXMLExportOptions
XML Fundamentals
Importing XML
Importing a minimal XML file
A new, empty document has a logical structure consisting of one Root element. To construct a
slightly less trivial example, you can create a new document and import a minimal XML file
into the backing store. The content is as shown in Example 35.
EXAMPLE 35 Minimal XML data to be imported
<article><title>Hello World</title>
</article>
The scenario to perform the import is as follows:
1. Save the XML data above to a text file, hello-world.xml.
2. Create a new InDesign document.
3. Show the structure view. Rename the Root tag to be “article.”
4. Create a new tag named “title.”
5. Import the XML data from the hello-world.xml file. The structure view should look like
Figure 228.
The objects created and instances of associations between them are shown in Figure 228. The
UML object diagram shows the objects involved in representing the logical structure when
minimal XML content is imported into the backing store of a new InDesign document. The
logical structure of the InDesign document represents the tree structure of the original XML
document, which has one title element as a child of the root-article element. Instances of the
associations with the objects that represent tags are shown (kXMLTagBoss).
XML Fundamentals
515
XML Fundamentals
Importing XML
FIGURE 228
Objects representing elements in logical structure and tags
«boss»
doc-elem : kXMLDocumentBoss
IIDXMLElement
«boss»
article : kTextXMLElementBoss
IXMLTag
«boss»
article-tag : kXMLTagBoss
IIDXMLElement
«boss»
title : kTextXMLElementBoss
IXMLTag
«boss»
title-tag : kXMLTagBoss
Given a reference to the document (IDocument) or a document database (IDataBase), you can
acquire a reference to the document element or root element in the logical structure, using the
IXMLUtils facade.
The document element has no tag, but the root element always has a tag; in this example, it is
named “article.” The IIDXMLElement::GetTagString method discovers the tag name of an element.
The root element is one starting point for traversing the logical structure. For the minimal
example, you can start from the root element and look at its children.
516
XML Fundamentals
Importing XML
XML elements have the signature interface IIDXMLElement. Given an IIDXMLElement interface, you can obtain an XMLReference, another way to refer to an XML element. See IIDXMLElement::GetXMLReference. Conversely, given an XMLReference, you can get an
IIDXMLElement interface through XMLReference::Instantiate. XML elements are described in
more detail in “Elements and content” on page 534.
The UML object diagram in Figure 229 shows instances of associations between a document
(kDocBoss), document workspace (kDocWorkspaceBoss), and objects representing tags
(kXMLTagBoss) and their colors within the Tags panel (kUIColorDataBoss).
XML Fundamentals
517
XML Fundamentals
Importing XML
FIGURE 229
Objects representing tags and colors
«boss»
document : kDocBoss
IDocument
«boss»
doc-workspace : kDocWorkspaceBoss
IXMLTagList
IXMLTagList
«boss»
article-tag : kXMLTagBoss
«boss»
title-tag : kXMLTagBoss
IPersistUIDData
IPersistUIDData
«boss»
red-col : kUIColorDataBoss
«boss»
green-col : kUIColorDataBoss
Interface IPersistUIDData that
stores UID of color object has
the identifer (PMIID) of
IID_ICOLORPERSISTUIDDATA
on kXMLTagBoss.
Unplaced content versus placed content
When XML content is imported with default options and there are no tagged placeholders for
the incoming content, the content is held in the backing store. See “Backing store” on page 509.
518
XML Fundamentals
Importing XML
XML template
An XML template is an InDesign document with tagged placeholders. These can be any of the
following:
z
Stories
z
Text ranges
z
Graphic frames
z
Tables and table cells
In addition, for the template to be useful, it is likely to already have some styles established and
a mapping from tags to styles. See “Tag-to-style mapping” on page 531.
Matching against an XML template
When XML data is imported into an XML template that matches elements in the imported
XML data, some or all of the content becomes placed. Determining what constitutes a match is
not a straightforward process.
Architecture
Matching against an XML template is performed by XML-import matchmaker services (IXMLImportMatchMaker), which implement the matching logic. For a description of this extension pattern, see “XMl-import matchmaker” on page 555. There are several existing import
matchmaker services that support the import features; for example, for importing structured
tables or repeating elements. For a more complete listing, see the API reference documentation
for IXMLImportMatchMaker and the boss classes that aggregate this interface.
If you implement your own import matchmaker, you can customize how the matching against
the template is done. For example, you might want to add new pages to the document when
you encounter certain repeating elements in the incoming XML data or perform some other
operation not provided by the application. For details, see “XMl-import matchmaker” on
page 555.
Importing repeating elements
Suppose you have XML-based content that consists of many very similar elements, like classified advertisements for used automobiles. You want an XML template that can handle repeating elements in the incoming XML data, rather than having to specify up front how many
advertisements could be flowed into the template. In designing an XML template, it often is
desirable to design only one instance of a repeating substructure and, during import, have all
incoming instances of the substructure pick up the design of that instance.
Often, the number of instances in the incoming XML data is either unknown or variable, so
having the ability to handle repeating elements adds flexibility. During XML import, InDesign
examines the incoming XML data's structure, compares it to the existing structure in the XML
template, and makes duplicates of the existing template element as necessary when it detects
the incoming elements are repetitions of the existing template element.
XML Fundamentals
519
XML Fundamentals
Importing XML
Suppose you are going to import this hypothetical XML data related to classified advertisements for motor vehicles. The fields in the XML template tagged with the names of the different elements may be styled differently. The number of items in the incoming XML data is
unknown. You may set up a template structure like that in Example 36.
EXAMPLE 36 Sample XML template
<classifieds>
<advert>
<vehicle>
<year></year>
<make></make>
<model></model>
<mileage></mileage>
<price></price>
...
</vehicle>
<contact>
<name></name>
<phone></phone>
</contact>
</advert>
</classifieds>
In this case, the “advert” element is the repeating element. You would set up styling for the
fields in each advert element. During XML import, InDesign duplicates the repeatable advert
element as many times as there are elements in the incoming XML.
When the data is imported into the XML template, the fields under advert retain the original
styling throughout the duplication process, so the final outcome is that there are as many
advert elements as there advert elements in the incoming XML.
Architecture
The feature is implemented by an import matchmaker service (IXMLImportMatchMaker)
whose ClassID is kXMLRepeatTextElementsMatchMakerServiceBoss, which is responsible for
duplicating elements in the XML template to accommodate the incoming XML elements.
The feature is turned on or off by an XML import preference (IXMLImportPreferences); see
“Service-level XML preferences” on page 551.
Throwing away unmatched existing elements on import (delete unmatched
right)
You can choose to throw away (delete) unmatched existing elements on import, to filter out
XML template elements that do not have a match in the incoming XML data. This might occur
if, for example, your XML template contains placeholders for all possible elements in the
incoming XML, including optional elements, and the incoming XML does not contain some of
the optional elements. If an element (like a byline element) in your XML template is not
matched in the actual content, you would not want to leave this unnecessary element in the
logical structure of the document. This feature removes any unmatched optional elements (and
520
XML Fundamentals
Importing XML
their content) in the document after XML import. The user sets this import option in the XML
Import Options dialog box.
Suppose you have an XML template with the logical structure and existing mark-up shown in
Figure 230. If the feature to throw away unmatched existing elements is enabled before importing XML, and the inbound XML does not have elements that match a copyright element in the
template, the copyright element in the template is deleted on import.
FIGURE 230
XML template
For example, suppose the incoming XML contains only the elements shown in Example 37,
rather than the full set of elements in the XML template document in Figure 230.
EXAMPLE 37 Sample XML to import
<article><headline>Another Disastrous Night for Scotland</headline>
<byline>Chick O'Young</byline>
<dateline>22 July 2004</dateline>
<body><para>Scotland's World Cup qualification hopes were dampened by the news
that no-one in the current squad is fit to play for Scotland.</para></body>
</article>
The net result of importing the sample XML with the option to throw away unmatched existing
elements is that only the elements in the incoming XML data explicitly matched in the template
are retained, and the template elements and associated content are deleted from the document
on import. The result for this sample XML is shown in Figure 231.
XML Fundamentals
521
XML Fundamentals
Importing XML
FIGURE 231
After throwing away unmatched existing elements on import
Architecture
This feature is implemented by a matchmaking service (kXMLThrowAwayUnmatchedRightMatchMakerServiceBoss); see “XMl-import matchmaker” on page 555. The feature is turned
on or off by a service-specific import preference (IXMLImportPreferences); see “Service-level
XML preferences” on page 551. For more information on controlling the feature, see the “XML”
chapter of Adobe InDesign CS4 Solutions.
Throwing away unmatched incoming elements on XML import
You can choose to throw away (delete) unmatched incoming elements, to filter out elements in
the incoming XML data that do not have a match in the XML template into which they are
being imported. You can use this feature when you know there are optional elements in the
incoming XML that you do not want to display, and you do not want to use XSLT to explicitly
filter them out.
Architecture
This feature is implemented by a matchmaking service (kXMLThrowAwayUnmatchedRightMatchMakerServiceBoss); see “XMl-import matchmaker” on page 555. This feature is turned
on or off by a service-specific import preference (IXMLImportPreferences); see “Service-level
XML preferences” on page 551. For more information on controlling the feature, see the “XML”
chapter of Adobe InDesign CS4 Solutions.
Attribute-style mapping
Suppose you want to import XML that has (or is transformed to have) attributes that specify
the character or paragraph style to apply to text content in the inbound XML. You can use the
attribute-style mapping feature, which looks for attributes pstyle and cstyle in the inbound
XML, then tries to match these to paragraph and character styles in the document.
Architecture
This feature is implemented by a post-import responder (kXMLImporterPostImportMappingBoss); see “Post-import responder” on page 555. This responder also implements tag-to-style
522
XML Fundamentals
Importing XML
mapping on import; see “Tag-to-style mapping” on page 531. For information on how this feature is turned on or off, see “Service-level XML preferences” on page 551.
Creating links on XML import
When XML is imported into an XML template, the default behavior is not to create a link to the
imported XML file; however, it is possible to turn on a feature that supports creating a link to
the imported XML file.
XML linking is a feature that enables InDesign to keep track of imported XML files. Each
imported XML file has a corresponding link, which is shown in the Links panel along with status. Operations that can be performed on XML links include the following:
z
Create — When an XML file is imported, a link (kXMLImportLinkBoss) is created and
associated with the root element of the imported content. Only one link can be associated
with one element. Importing into an element that already has a link replaces the old link
with the new one.
z
Delete — Deleting an element deletes its associated link, if it has one.
z
Copy, paste, duplicate — When an XML element associated with a link is copied, pasted, or
duplicated, a copy of the link is created and associated with the new element.
z
Place — Placing an element into the layout or unplacing it does not affect its XML link.
Both placed and unplaced elements can be associated with XML links.
z
Check status — An XML link's status (e.g., up-to-date, modified, or missing) is shown as an
icon in the Links panel.
Architecture
This feature is controlled by an XML-related preference. For details on how this feature is
turned on or off, see “Service-level XML preferences” on page 551.
Sparse import
If this feature is turned on during XML import, incoming elements that have no content (or
only whitespace for content) do not replace content in the matching existing XML element in
the XML template. By default, this feature is turned off.
Architecture
The feature is controlled by an XML-related preference (kXMLSparseImportOptionsServiceBoss). The top-level SAX handler (kSAXDocumentHandlerBoss) acts on this preference,
rather than the service (kXMLSparseImportOptionsServiceBoss) being responsible for providing the behavior. This explains why kXMLSparseImportOptionsServiceBoss has no service
interface: it is not implementing a particular extension pattern other than for XML import
preferences. For details on how to turn sparse import on and off, see “Service-level XML preferences” on page 551.
XML Fundamentals
523
XML Fundamentals
Exporting XML
Importing a CALS table as an InDesign table
You can choose whether to import a CALS table as an InDesign table. You can use this feature
when you have a table specified in CALS table format and you want to manipulate it the same
way as InDesign native tables.
Architecture
This feature is implemented by a matchmaking service (kXMLTableMatchMakerServiceBoss)
and DOM serializer handler service; see “XMl-import matchmaker” on page 555 and “SAX
DOM serializer handler” on page 557. During the matchmaker phase, a special attribute, kTableFormatAttr (“tformat”), with value kAttrValueCALS (“CALS”), is inserted into the table element. During the DOM serializer phase, contents in the CALS table are altered to fit the
InDesign table format. This feature can be turned on or off by a service-specific import preference (IXMLImportPreferences); see “Service-level XML preferences” on page 551. For more
information on controlling the feature, see the “XML” chapter of Adobe InDesign CS4 Solutions.
Support table and cell styles when importing an InDesign table
You can specify table and cell styles when importing an InDesign table. A table-style attribute
applies only to a table element. Cell-style attributes apply to either a table element or a cell element; they are ignored for other types of elements. If the style name specified in the attribute
value does not exist, a new style with that name is created and then applied to the table or cell.
Architecture
This feature is implemented by a post-import responder (kXMLImporterPostImportMappingBoss), the same way as character style and paragraph style; see “Attribute-style mapping” on
page 522. The only difference is that the attribute names are tablestyle and cellstyle.
Exporting XML
Exporting XML data from InDesign is considerably more straightforward than importing
XML data into InDesign. The downside to this is that there are fewer opportunities for your
plug-in to customize the export side.
You can export the logical structure of an InDesign document as an XML document. The
export takes place in document order, meaning the logical structures of the InDesign document and exported XML data are the same.
An XML file can be exported from any element in the logical structure. For example, if
exported from the root element, the exported XML data represents the logical structure of the
entire document. If another element—for example, E—is chosen, the exported XML data represents the logical structure of the subtree with E at its root.
Even an empty InDesign document has some (albeit trivial) logical structure: it has one root
element with no dependents. If exported as XML data, one element (<Root></Root>) is writ-
524
XML Fundamentals
Exporting XML
ten to the output file. Figure 232 shows the logical structure of an new, empty document and
the XML data exported from InDesign with the default settings. Note the default encoding
(UTF-8).
FIGURE 232
Structure of empty document and XML exported
"[POYHUVLRQ HQFRGLQJ 87)VWDQGDORQH \HV"!
Root></Root>
There are two points to note about the relationship between XML data exported from InDesign
and the logical structure:
z
XML content can be exported from a document even if it is not associated with placed content in the document. Exported XML data is based on the logical structure, irrespective of
whether the elements are associated with content in the document.
z
The order in which the elements appear in the XML output is based on the logical structure
and has no direct relation with the hierarchy of the native document model. The exception
to this is a story with mark-up, in which case the order of the elements in the logical structure within the subtree corresponding to the story corresponds to the story-reading order.
Export architecture
An XML generator (IXMLGenerator) is an abstraction responsible for generating the XML
content during export. The IXMLGenerator interface is aggregated by classes representing
XML elements (IIDXMLElement) in the logical structure.
The mechanism that supports XML export is a standard export provider (kXMLExportProviderBoss, IExportProvider). This delegates to an instance of kXMLExportHandlerBoss to perform the actual export. The implementation of IXMLHandler iterates elements in the logical
structure, calling methods through the IXMLGenerator interface of the elements.
This sequence diagram in Figure 233 shows some of the messages sent during a typical XML
export.
XML Fundamentals
525
XML Fundamentals
Exporting XML
FIGURE 233
export-provider
XML-export sequence
xmlUtils
xmlAccess
xmlHandler
xmlGenerator
IXMLUtils::ExportElement
IXMLAccess::VisitAllElements
IXMLHandler::HandleElementBefore
IXMLGenerator::GenerateBeginTag
IXMLHandler::HandleAfterNthChild
IXMLGenerator::GenerateContent
IXMLHandler::HandleElementAfter
IXMLGenerator::GenerateEndTag
The export provider (kXMLExportProviderBoss), shown as export-provider in Figure 233,
starts the export. The parser service (kXMLParserServiceBoss) has an IXMLAccess interface,
shown as the xmlAccess object. An instance of kXMLExportHandlerBoss, shown as xmlHandler, with signature interface IXMLHandler, iterates elements in the logical structure. The
implementation of the IXMLGenerator interface on these elements generates XML content to
the output stream. There are several implementations of this interface, generating different
content types on export.
Contributions to XML export come from instances of boss classes that expose the IXMLGenerator interface. These classes represent elements in the logical structure; e.g., kTextXMLElementBoss (tagged stories, tagged text ranges, and tagged graphics) and kXMLCommentBoss
(XML comments).
You can participate in the export process by providing a custom kXMLExportHandlerSignalService. The providers are called when the XML handler iterates elements. For more information, see the extension pattern in “XML-export handler” on page 558.
Document order
The logical structure of an InDesign document is exported as XML in document order, as
defined in the XPath specification (http://www.w3.org/TR/xpath#dt-document-order). For
convenience, part of the definition is repeated here:
526
XML Fundamentals
Exporting XML
“There is an ordering, document order, defined on all the nodes in the document corresponding to the order in which the first character of the XML representation of each node occurs in
the XML representation of the document after expansion of general entities. Thus, the root
node will be the first node. Element nodes occur before their children. Thus, document order
orders element nodes in order of the occurrence of their start-tag in the XML (after expansion
of entities). The attribute nodes and namespace nodes of an element occur before the children
of the element. ... Reverse document order is the reverse of document order.”
Document order has no direct relationship with the order in which a document would be read
or other ordering schemes you might devise (e.g., based on page numbering).
Tagged graphic placeholder, exported
Consider the case, shown in Figure 234, of a simple but nontrivial logical structure in which a
tagged placeholder is intended as the destination for an image.
FIGURE 234
Tagged graphic placeholder
Exporting the logical structure with default export options results in the XML data shown in
Figure 235. To vary the options, like the encoding, see “XML-related preferences” on page 549.
FIGURE 235
Exported XML from document with tagged-graphic placeholder
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
- <Root>
<image />
</Root>
The abstraction responsible for generating the XML data related to the tagged graphic is the
element (IIDXMLElement) that represents the graphic. It exposes the IXMLGenerator interface, which is used during the export process.
XML Fundamentals
527
XML Fundamentals
Exporting XML
Tagged text range, exported
Consider the example in which you create a text frame, tag the story running through it with a
tag named “story,” then tag a range within that story with a different tag named “emphasis.” The
XML story element represents the instance of the tagged story. The XML emphasis element
represents the tagged content item with value “Hello.” This is shown in Figure 236. The screenshot shows logical structure, tags, and layout view for a document with one text frame, with a
tagged text range. The story is tagged “story” and represented by the story element in the logical structure. The text range is tagged “emphasis” and represented by the emphasis element.
FIGURE 236
Tagged text range before export
If you export the logical structure as XML data from the root element, you get the XML data
shown in Figure 237. This time, the export option to export with UTF-16 encoding is set, to
make it easier to inspect the Unicode character that represents the hard carriage return in the
story. This XML data was exported by InDesign from the document shown in Figure 236.
FIGURE 237
XML exported for tagged text range
<?xml version="1.0" encoding="UTF-16" standalone="yes" ?>
- <Root>
- <story>
<emphasis>Hello</emphasis>
world
</story>
</Root>
The XML content of the exported story requires some study. The intent of the InDesign function is that stories should be represented in exported XML data as closely as possible to how
InDesign represents them internally. The end-of-line character corresponding to the hard car-
528
XML Fundamentals
Tags
riage return after the word “world” is represented by the Unicode character 0x2029. If you open
the XML file in a binary editor on Windows—which is little-endian—the Unicode character
0x2029 is represented as the octet 29 and then the octet 20. Soft carriage returns are represented by 0x2028. See IXMLOutStream. This is shown in Figure 238; the 0x2029 (hard carriage
return) is highlighted.
FIGURE 238
Story exported from InDesign as XML in UTF-16 encoding
Tags
Tags can be loaded from several different sources, including the following:
z
An InDesign tag file, an XML document containing only tag-specific information.
z
Any XML document that is an instance of the document type being worked with.
z
By associating a DTD with the document. Doing this means the DTD is parsed for element
tag names, and tags are created correspondingly. Elements defined within entities are
ignored.
Example 38 is a sample tag list.
EXAMPLE 38 Sample tag list
<?xml version="1.0" encoding="UTF-16" standalone="yes"?>
<article colorindex="4">
<articleinfo colorindex="6"/>
... (other elements omitted)
<ulink colorindex="19"/>
</article>
It also is possible to specify the color of the tag using RGB coordinates, which are used for any
tags defined with a custom color. The coordinates are hex-encoded; for example, one tag with a
custom color (0,0,255) in decimal RGB coordinates appears below:
<link rgb="0 0,0 0,3ff00000 0"/>
In addition, the custom-tag service-extension pattern lets you customize what tags are created
when an XML document is being parsed. See “Custom-tag service” on page 557.
XML Fundamentals
529
XML Fundamentals
Tags
Architecture
The kXMLTagBoss boss class represents an individual entry in a tag list. The signature interface of this boss class is IXMLTag; this stores properties like the name and color in the user
interface. The tag list is represented by IXMLTagList.
IXMLTagList is aggregated on the session workspace (kWorkspaceBoss), representing a default
set of tags for any new document.
IXMLTagList is aggregated on the document workspace (kDocWorkspaceBoss), representing
tags in a given document.
The session workspace (kWorkspaceBoss) stores zero or more tags. A document workspace
(kDocWorkspaceBoss) stores one or more tags; the minimal tag set for a new document consists of the Root tag.
Tagging relationships are represented by associations between classes representing elements in
the logical structure and those representing tags. For example, if a graphic frame is tagged for
use as a placeholder, an association is created between the placeholder boss object (an instance
of the kPlaceHolderItemBoss class) and an instance of the kTextXMLElementBoss class. An
association is created between an element in the logical structure (IIDXMLElement) and an
instance of kXMLTagBoss, to represent the tagging.
Tags (kXMLTagBoss) that can be used to mark up content items in the native document model
are held in the tag list (IXMLTagList) of a workspace (kDocWorkspaceBoss, kWorkspaceBoss).
Tags are rendered in the Tags panel (Figure 218) and shown in views like layout view
(Figure 216) or story view (Figure 217). XML elements that have tag names are associated with
tags (kXMLTagBoss) that store the tag names.
Given a reference to an XML element, you can find the tag string through a method on
IIDXMLElement. You also can use IIDXMLElement::GetTagUID to acquire a reference to an
instance of kXMLTagBoss representing the tag.
You can acquire a reference to a tag (kXMLTagBoss) in a document through the tag list (IXMLTagList) on the document workspace (kDocWorkspaceBoss). The tagUIDRef variable refers to
an instance of kXMLTagBoss.
The class diagram in Figure 239 shows associations between tags and workspaces and the document (kDocBoss).
530
XML Fundamentals
Tags
FIGURE 239
Tags
«boss»
kDocBoss
1
IDocument::GetDocWorkspace
1
«boss»
kWorkspaceBoss
«boss»
kDocWorkspaceBoss
1
1
IXMLTagList::GetTag
«boss»
kXMLTagBoss
1..* IXMLTagList::GetTag
0..*
Tag-to-style mapping
Suppose you have XML content with a headline tag and a p-head-1 paragraph style in your
XML template. You can create an association between tags and styles in the document, to
enable incoming XML to be styled automatically.
When a tag-to-style mapping is applied during the XML import, inbound XML content has the
styles applied when elements with tag names matching the tags in the tag-to-style map are
encountered. For example, if you create a mapping from the headline tag to the p-head-1 paragraph style, textual content in a headline element has the p-head-1 style applied.
The user interface for this feature is shown in Figure 219. This establishes a one-to-one mapping from tag names to character or paragraph styles, which is sufficient when context-sensitive styling is not required. If context-sensitive styling for incoming XML is required, you can
use a combination of a stream-based XML transformer and attribute-style mapping to achieve
this. See “XML transformer” on page 554 and “Attribute-style mapping” on page 522.
Architecture
Figure 240 is a class diagram for tag-to-style mappings (IXMLTagToStyleMap). The session
workspace (kWorkspaceBoss) stores the default mapping inherited by new documents. The
document workspace (kDocWorkspaceBoss) stores the mapping applied to a particular document (kDocBoss).
XML Fundamentals
531
XML Fundamentals
Tags
FIGURE 240
Tag-to-style mapping
IXMLTagToStyleMap::GetStyleMappedToTag
«boss»
kStyleBoss
0..* IXMLTagToStyleMap::GetStyleMappedToTag
0..*
1
1
«boss»
kWorkspaceBoss
«boss»
kDocWorkspaceBoss
1
1
0..*
IXMLTagToStyleMap::GetTagAt
«boss»
kXMLTagBoss
IXMLTagToStyleMap::GetTagAt
0..*
Style-to-tag mapping
Figure 241 is a screenshot of an unstructured but systematically styled document, which as an
associated DTD. Figure 242 shows the result of applying a style-to-tag mapping.
532
FIGURE 241
Before mapping styles to tags
FIGURE 242
After mapping styles to tags
XML Fundamentals
Tags
Once a style-to-tag mapping is defined and applied, ranges of styled text with the styles
mapped to given tags end up tagged, creating new elements in the logical structure. There are
some default commitments in terms of the tags applied; for example, the Story tag is used for a
story, even if it is not in the tag list of the document when the style-to-tag mapping is applied.
Architecture
The class diagram in Figure 243 shows the associations between workspaces (IWorkspace),
tags (kXMLTagBoss), and styles (kStyleBoss), mediated by the style-to-tag map (IXMLStyleToTagMap). The session workspace (kWorkspaceBoss) stores the default style-to-tag map for
new documents. The document workspace (kDocWorkspaceBoss) stores the style-to-tag map
for a given document (kDocBoss).
FIGURE 243
Style-to-tag mapping
0..*
IXMLStyleToTagMap::GetStyleAt
«boss»
kStyleBoss
IXMLStyleToTagMap::GetStyleAt
0..*
1
1
«boss»
kWorkspaceBoss
«boss»
kDocWorkspaceBoss
1
1
IXMLStyleToTagMap::GetTagMappedToStyle
0..*
«boss»
kXMLTagBoss
IXMLStyleToTagMap::GetTagMappedToStyle
0..*
Text styles are represented by kStyleBoss. Collections of styles are held in the style-name table
(IStyleNameTable) and stored in a workspace (IWorkspace). The session workspace (kWorkspaceBoss) stores text styles that are defaults for new documents. The document workspace
(kDocWorkspaceBoss) stores text styles that can be used in the associated document. For more
information on text styles, see the “Text Fundamentals” chapter and the API reference documentation for kStyleBoss.
The workspace (IWorkspace) maintains an associative map between text styles (kStyleBoss)
and tags (kXMLTagBoss). This map is stored in the persistent interface IXMLStyleToTagMap.
The workspace style list (IStyleNameTable) may contain styles that are not referenced in the
style-to-tag map. Similarly, there may be tags in the tag list (IXMLTagList) of the workspace
that are not involved in the style-to-tag map.
XML Fundamentals
533
XML Fundamentals
Elements and content
Elements and content
Tagged graphic placeholder
Suppose you are building an XML template and you want to create placeholders into which
images will be placed when an XML data file is imported. To create a placeholder graphic for an
image, follow these steps:
1. Create a new document.
1. Create a tag named “image” through the Tags panel (Window > Tags).
2. Create a rectangle page item with the Rectangle tool.
3. Leave the page item selected, and click on the image tag in the Tags panel.
4. Show the structure view. It should look like Figure 244.
Architecture
The UML object diagram in Figure 244 shows the boss objects representing a tagged graphic
placeholder and some of the relationships between them. The placeholder object (kPlaceHolderItemBoss) is a child of a graphic frame (kSplineItemBoss); the latter is omitted from this diagram, in the interest of simplicity. Content items like kPlaceHolderItemBoss have an
IXMLReferenceData interface, which lets them refer to an element in