JavaServer Pages

JavaServer Pages

JavaServer Pages

Hans Bergsten

First Edition, December 2000

ISBN: 1-56592-746-X, 572 pages

Preface

What's in This Book

Audience

Organization

About the Examples

Conventions Used in This Book

How to Contact Us

Acknowledgments

I: JSP Application Basics

1. Introducing JavaServer Pages

1.1 What Is JavaServer Pages?

1.2 Why Use JSP?

1.3 What You Need to Get Started

2. HTTP and Servlet Basics

2.1 The HTTP Request/Response Model

2.2 Servlets

2.3 Packaging Java Web Applications

3. JSP Overview

3.1 The Problem with Servlets

3.2 The Anatomy of a JSP Page

3.3 JSP Processing

3.4 JSP Application Design with MVC

4. Setting Up the JSP Environment

4.1 Installing the Java Software Development Kit

4.2 Installing the Tomcat Server

4.3 Testing Tomcat

4.4 Installing the Book Examples

4.5 Example Web Application Overview

II: JSP Application Development

5. Generating Dynamic Content

5.1 What Time Is It?

5.2 Input and Output

6. Using Scripting Elements

6.1 Java Primer

6.2 Implicit JSP Objects

6.3 Conditional Processing

6.4 Displaying Values

6.5 Using an Expression to Set an Attribute

6.6 Declaring Variables and Methods

7. Error Handling and Debugging

7.1 Dealing with Syntax Errors

7.2 Debugging a JSP-Based Application

7.3 Dealing with Runtime Errors

8. Sharing Data Between JSP Pages, Requests, and Users

8.1 Passing Control and Data Between Pages

8.2 Sharing Session and Application Data

8.3 Using Custom Actions

8.4 Online Shopping

8.5 Memory Usage Considerations

9. Database Access

9.1 Accessing a Database from a JSP Page

9.2 Input Validation Without a Bean

9.3 Using Transactions

9.4 Application-Specific Database Actions

10. Authentication and Personalization

10.1 Container-Provided Authentication

10.2 Application-Controlled Authentication

10.3 Other Security Concerns

11. Internationalization

11.1 How Java Supports Internationalization and Localization

11.2 Generating Localized Output

11.3 A Brief History of Bits

11.4 Handling Localized Input

12. Bits and Pieces

12.1 Buffering

12.2 Including Page Fragments

12.3 XML and JSP

12.4 Mixing Client-Side and Server-Side Code

12.5 Precompiling JSP Pages

12.6 Preventing Caching of JSP Pages

12.7 How URLs Are Interpreted

III: JSP in J2EE and JSP Component Development

13. Web Application Models

13.1 The Java 2 Enterprise Edition Model

13.2 The MVC Model

13.3 Scalability

14. Combining Servlets and JSP

14.1 Using a Servlet as the Controller

14.2 A More Modular Design Using Action Objects

14.3 Sharing Data Between Servlets and JSP Pages

14.4 Using a JSP Error Page for All Runtime Errors

15. Developing JavaBeans for JSP

15.1 JavaBeans as JSP Components

15.2 JSP Bean Examples

15.3 Unexpected Behavior

16. Developing JSP Custom Actions

16.1 Tag Extension Basics

16.2 Developing a Simple Action

16.3 Processing the Action Body

16.4 Letting Actions Cooperate

16.5 Creating New Variables Through Actions

16.6 Developing an Iterating Action

16.7 Creating the Tag Library Descriptor

16.8 Validating Syntax

16.9 How Tag Handlers May Be Reused

16.10 Packaging and Installing a Tag Library

17. Developing Database Access Components

17.1 Using Connections and Connection Pools

17.2 Using a Generic Database Bean

17.3 Developing Generic Database Custom Actions

17.4 Developing Application-Specific Database Components

IV: Appendixes

A. JSP Elements Syntax Reference

A.1 Directive Elements

A.2 Scripting Elements

A.3 Action Elements

A.4 Comments

A.5 Escape Characters

B. JSP API Reference

B.1 Implicit Variables

B.2 Servlet Classes Accessible Through Implicit Variables

B.3 Tag Extension Classes

B.4 Other JSP Classes

C. Book Example Custom Actions and Classes Reference

C.1 Generic Custom Actions

C.2 Internationalization Custom Actions

C.3 Database Custom Actions

C.4 Utility Classes

C.5 Database Access Classes

D. Web-Application Structure and Deployment Descriptor Reference

D.1 Web Application File Structure

D.2 Web Application Deployment Descriptor

D.3 Creating a WAR File

E. JSP Resource Reference

E.1 JSP-Related Products

E.2 Web Hosting

E.3 Information and Specifications

Colophon

© 2001, O'Reilly & Associates, Inc.

Preface

JavaServer Pages™ (JSP) is a new technology for web application development that has received a great deal of attention since it was first announced.

Why is JSP so exciting? One reason is that JSP is Java-based, and Java is well-suited for enterprise computing. In fact, JSP is a key part of the Java™ 2 Enterprise Edition (J2EE) platform and can take advantage of the many Java Enterprise libraries, such as JDBC, JNDI, and Enterprise JavaBeans™.

Another reason is that JSP supports a powerful model for developing web applications that separates presentation from processing. Understanding why this is so important requires a bit of a history lesson.

In the early days of the Web, the only tool for developing dynamic web content was the Common

Gateway Interface (CGI). CGI outlined how a web server made user input available to a program, as well as how the program provided the web server with dynamically generated content to send back.

CGI scripts were typically written in Perl. (In fact, CGI Perl scripts still drive numerous dynamic web sites.) However, CGI is not an efficient solution. For every request, the web server has to create a new operating-system process, load a Perl interpreter and the Perl script, execute the script, and then dispose of the entire process when it's done.

To provide a more efficient solution, various alternatives to CGI have been added to programmers' toolboxes over the last few years: FastCGI, for example, runs each CGI program in an external permanent process (or a pool of processes). In addition, mod_ perl for Apache, NSAPI for Netscape, and ISAPI for Microsoft's IIS all run server-side programs in the same process as the web server itself.

While these solutions offer better performance and scalability, each one is supported by only a subset of the popular web servers.

The Java Servlet API, introduced in early 1997, provides a solution to the portability issue. However, all these technologies suffer from a common problem: HTML code embedded inside programs. If you've ever looked at the code for a servlet, you've probably seen endless calls to out.println( ) that contain scores of HTML tags. For the individual developer working on a simple web site this approach may work fine, but it makes it very difficult for people with different skills to work together to develop a web application.

This is becoming a significant problem. As web sites become increasingly complex and are more and more critical to the success of an organization, the appearance and usability of the web interface becomes paramount. New client technologies, such as client-side scripts and DHTML, can develop more responsive and interactive user interfaces, stylesheets can make it easier to globally change fonts and colors, and images can make the interface more appealing. At the same time, server-side code is getting more complex, and demands for reliability, performance, and fault tolerance are increasing. The

growing complexity of web applications requires a development model that allows people with different skills to cooperate efficiently.

JavaServer Pages provides just such a development model, allowing web page authors with skills in graphics, layout, and usability to work in tandem with programmers who are experienced in server-side technologies such as multithreading, resource pooling, databases, and caching. While there are other technologies, such as ASP, PHP, and ColdFusion, that support similar development models, none of them offers all the advantages of JSP.

© 2001, O'Reilly & Associates, Inc.

What's in This Book

This book covers Version 1.1 of the JavaServer Pages specification, which was released in late 1999.

In this book, you will learn how to use all the standard JSP elements and features, including elements for accessing JavaBeans components, separating the processing over multiple pages to increase reusability and simplify maintenance, and sharing information between pages, requests, and users. You will also learn how to use and develop custom components. A rich set of custom components, for tasks such as integration of database data, internationalization, access control, and conditional processing, is described in detail. Many of these components are generic enough that you can reuse them directly in your own applications.

The examples in this book guide you through solutions to common JSP design problems, from basic issues such as retrieving and validating user input, to more advanced areas such as developing a database-driven site, authenticating users, providing personalized content, and implementing internationalization. The last part of the book describes how you can combine JSP with other Java technologies; in particular, I describe the combination of JSP and servlets and provide an overview of how JSP fits into the larger scope of J2EE.

© 2001, O'Reilly & Associates, Inc.

Audience

This book is for anyone interested in using JSP technology to develop web applications. In particular, it is written to help the two types of people commonly involved in the development of a JSP-based application:

Page authors

Page authors primarily develop the web interface to an application. This group uses HTML, stylesheets, and client-side code to develop a rich user interface, and wants to learn how to use

JSP elements in web pages to interact with the server components of the application, such as databases and Enterprise JavaBeans (EJB).

Java programmers

Java programmers are comfortable with the Java programming language and Java servlets. This group is interested in learning how to develop JSP components that page authors can use in web pages, such as JSP custom actions and JavaBeans, and how to combine JSP with other Java server-side technologies, such as servlets and EJB.

This book is structured into three parts, which I describe shortly, to make it easier to find the material you are most interested in.

What You Need to Know

It's always hard to assume how much you, as the reader, already know. For this book, it was even harder, since the material is intended for two audiences: page authors and programmers.

I have assumed that anyone reading this book has experience with HTML; consequently, I will not explain the HTML elements used in the examples. But even if you're an HTML wiz, this may be your first exposure to dynamic web content and web applications. A thorough introduction to the HTTP protocol that drives all web applications, as well as to the concepts and features specific to servlet and

JSP-based web applications, is therefore included. If you want to learn more about HTML, I recommend HTML and XHTML: The Definitive Guide, by Chuck Musciano and Bill Kennedy

(O'Reilly & Associates).

If you're a page author, I have assumed that you don't know anything about programming, although it doesn't hurt if you have played around with client-side scripting languages like VBScript or JavaScript

(ECMAScript). This book contains a brief Java primer with enough information to allow you to use a

modest amount of Java code in JSP pages. As you will see, I recommend that you use Java components developed by a Java programmer instead of putting your own Java code in the pages, so you don't have to know all the intricate details of the Java language to use JSP.

I have assumed that programmers reading this book are familiar with Java programming, objectoriented concepts, and Java servlets. If you plan to develop JSP components for page authors and are not familiar with Java programming, I recommend that you read an introductory Java book, such as

Exploring Java by Patrick Niemeyer and Joshua Peck (O'Reilly). If you need to learn about servlets, I recommend Java Servlet Programming by Jason Hunter and William Crawford (O'Reilly) or another book that covers servlet technology.

The chapters dealing with database access require some knowledge of SQL and databases in general. I will explain all that you need to know to run the examples, but if you're hoping to develop databasedriven applications, you will need to know more about databases than what's in this book.

© 2001, O'Reilly & Associates, Inc.

Organization

This book is structured into three parts. The first part describes the fundamentals of HTTP (the protocol used by all web applications), how servlets and JSP are related, and how to set up a JSP development environment.

The focus of the second part is on developing JSP-based web applications using both standard JSP elements and custom components. Through practical examples, you will learn how to handle common tasks such as validating user input, accessing databases, authenticating users and protecting web pages, localizing your web site, and more. This portion of the book is geared more towards web content designers.

In the third part, you will learn how to develop your own custom actions and JavaBeans, and how to combine JSP with other Java server-side technologies, such as servlets and Enterprise JavaBeans

(EJB). This portion of the book is targeted towards the programming community.

All in all, the book consists of 17 chapters and five appendixes as follows.

Part I, JSP Application Basics

Chapter 1

Explains how JSP fits into the big picture of web applications and how it compares to alternative technologies.

Chapter 2

Describes the fundamental HTTP and servlet concepts you need to know to use JSP to its full potential.

Chapter 3

An overview of the JSP features, as well as the similarities and differences between JSP pages and servlets. Also introduces the Model-View-Controller design model and how it applies to

JSP.

Chapter 4

Describes where to get the JSP reference implementation, Apache Tomcat, and how to set it up on your system. Also explains how to install the book examples.

Part II, JSP Application Development

Chapter 5

Explains how to use JSP to generate dynamic content and how to receive and validate user input.

Chapter 6

A brief introduction to Java programming, followed by descriptions of all the JSP elements that let you embed Java code directly in your JSP pages.

Chapter 7

Describes the kinds of errors you may encounter during development of a JSP-based application, and strategies and JSP features that help you deal with them.

Chapter 8

Explains the JSP features that let you separate different types of processing in different pages to simplify maintenance and further development. Also describes how sessions can be used to build up information over a sequence of requests from the same user, and how information that applies to all users can be shared using the application scope.

Chapter 9

A quick overview of relational databases, JDBC, and SQL basics. Introduces a set of generic custom actions for reading, updating, and deleting database data.

Chapter 10

Describes how authentication and access control can be implemented using container-provided and application-controlled mechanisms, and how to use information about the current user to personalize the web pages.

Chapter 11

Explains internationalization and localization, as well as the Java features available to implement an internationalized application. Describes a set of custom actions used to implement a web site with support for multiple languages.

Chapter 12

Covers various areas not discussed in previous chapters, such as using XML and XSL with JSP, combining JSP with client-side code, reusing JSP fragments by including them in JSP pages, precompiling JSP pages, and more.

Part III, JSP in J2EE and JSP Component Development

Chapter 13

An overview of J2EE and web application architectures using JSP in combination with other

Java technologies.

Chapter 14

Describes in detail how JSP can be combined with servlets.

Chapter 15

Provides details about JavaBeans as they relate to JSP, including threading and synchronization concerns for session and application-scope JavaBeans, as well as how using JavaBeans can make it easier to eventually migrate to an EJB architecture. The beans used in previous chapters are reused as examples.

Chapter 16

Describes the JSP Tag Extension mechanism and how it is used to develop custom actions, reusing many of the custom actions from previous chapters as examples.

Chapter 17

Describes the database-access custom actions used in the previous chapters and how to use them with both connection pools developed in-house and those provided by a third-party vendor. Also explains how you can reuse the database-access beans to develop your own application-specific database custom actions.

Part IV, Appendixes

Appendix A

Contains descriptions of all the standard JSP 1.1 elements.

Appendix B

Contains descriptions of all implicit objects available in a JSP page as defined by the servlet and

JSP APIs, as well as the tag extension mechanism classes and interfaces.

Appendix C

Contains descriptions of the custom actions, beans, and utility classes used in the examples.

Appendix D

Contains descriptions of the standard web-application structure and all elements in the webapplication deployment descriptor.

Appendix E

Contains references to JSP-related products, web-hosting services, and sites where you can learn more about JSP and related technologies.

If you're a page author, I recommend that you focus on the chapters in Part I

and

Part II

. You may want to browse through

Part III to get a feel for how things work behind the scenes, but don't expect to

understand everything if you're not a Java programmer.

If you are a Java programmer,

Part III is where the action is. If you're already familiar with HTTP and

servlets, you may want to move quickly through

Part I . However, this part does include information

about the web application concept introduced in the Servlet 2.2 API that you may not be familiar with, even if you've worked with servlets for some time. I recommend that you read

Part II to learn how JSP

works, but you may want to skip ahead to the chapters in Part III

from time to time to see how the components used in the examples are actually implemented.

© 2001, O'Reilly & Associates, Inc.

Chapter 1. Introducing JavaServer Pages

The Java 2 Enterprise Edition ( J2EE) has taken the once-chaotic task of building an Internet presence and transformed it to the point where developers can use Java to efficiently create multitier, server-side applications. Today, the Java Enterprise APIs have expanded to encompass a number of areas: RMI and CORBA for remote object handling, JDBC for database interaction, JNDI for accessing naming and directory services, Enterprise JavaBeans for creating reusable business components, JMS ( Java

Messaging Service) for message-oriented middleware, and JTA ( Java Transaction API) for performing atomic transactions. In addition, J2EE supports servlets , an extremely popular Java substitute for CGI scripts. The combination of these technologies allows programmers to create distributed business solutions for a variety of tasks.

In late 1999, Sun Microsystems added a new element to the collection of Enterprise Java tools:

JavaServer Pages ( JSP). JavaServer Pages are built on top of Java servlets and are designed to increase the efficiency in which programmers, and even nonprogrammers, can create web content. This book is all about JavaServer Pages.

© 2001, O'Reilly & Associates, Inc.

Chapter 2. HTTP and Servlet Basics

Let's start this chapter by defining the term web application . We've all seen regular client-side applications. But what exactly is a web application? Loosely, we could define it as an application running on a server that a user accesses through a thin, general-purpose client. Today, the most common client is a web browser on a PC or workstation, but soon all kinds of clients will be used, such as wireless PDAs, cellular phones, and other specialized devices.

The lofty goal here is to access all the information and services you need from any type of device you happen to have in front of you. This means that the same simple client program must be able to talk to many different server applications, and the applications must be able to work with many different types of clients. To satisfy this need, the protocol of how a client and a server talk to each other must be defined in detail. That's exactly what the HyperText Transport Protocol (HTTP) is for.

The communication model defined by HTTP forms the foundation for all web application design. You therefore need a basic understanding of HTTP to develop applications that fit within the constraints of the protocol, no matter which server-side technology you use. In this chapter, we look at the most important details of HTTP that you need to be aware of as a web application developer.

One other item. This book is about using JSP as the server-side technology, so that's what we'll primarily focus on. As we saw in

Chapter 1

, JSP is based on the Java servlet technology. Both technologies share a lot of terminology and concepts, so knowing a bit about servlets will help you even when you develop pure JSP applications. And to really understand and use the full power of JSP, you need to know a fair bit about servlets. We will therefore take a quick look at servlet fundamentals in the last section of this chapter, including a programmer's introduction for those of you familiar with

Java.

© 2001, O'Reilly & Associates, Inc.

Chapter 3. JSP Overview

JSP is the latest Java technology for web application development, and is based on the servlet technology introduced in the previous chapter. While servlets are great in many ways, they are generally reserved for programmers. In this chapter, we look at the problems that JSP technology solves, the anatomy of a JSP page, the relationship between servlets and JSP, and how a JSP page is processed by the server.

In any web application, a program on the server processes requests and generates responses. In a simple one-page application, such as an online bulletin board, you don't need to be overly concerned about the design of this piece of code; all logic can be lumped together in a single program. But when the application grows into something bigger (spanning multiple pages, with more options and support for more types of clients) it's a different story. The way your site is designed is critical to how well it can be adapted to new requirements and continue to evolve. The good news is that JSP technology can be used in all kinds of web applications, from the simplest to the most complex. Therefore, this chapter also introduces the primary concepts in the design model recommended for web applications, and the different roles played by JSP and other Java technologies in this model.

© 2001, O'Reilly & Associates, Inc.

Chapter 4. Setting Up the JSP Environment

This book contains plenty of examples to illustrate all the JSP features. All examples were developed and tested with the JSP reference implementation, known as the Apache Tomcat server, which is developed by the Apache Jakarta project. In this chapter you will learn how to install the Tomcat server and add a web application containing all the examples used in this book. You can, of course, use any web server that supports JSP 1.1, but Tomcat is a good server for development and test purposes. You can learn more about the Jakarta project and Tomcat, as well as how you can participate in the development, at the Jakarta web site: http://jakarta.apache.org/ .

© 2001, O'Reilly & Associates, Inc.

Chapter 5. Generating Dynamic Content

JSP is all about generating dynamic content: content that differs based on user input, time of day, the state of an external system, or any other runtime conditions. JSP provides you with lots of tools for generating this content. In this book, you will learn about all of them—standard actions, custom actions, JavaBeans, and scripting elements. Before we do that, however, let's start with a few simple examples to get a feel for how the basic JSP elements work.

In this chapter, we develop a page for displaying the current date and time, and look at the JSP directive element and how to use JavaBeans in a JSP page along the way. Next, we look at how to process user input in your JSP pages and make sure it has the appropriate format. We also look at how you can convert special characters in the output, so they don't confuse the browser.

© 2001, O'Reilly & Associates, Inc.

Chapter 6. Using Scripting Elements

When you develop a JSP-based application, I recommend that you try to place all Java code in

JavaBeans, in custom actions, or in regular Java classes. However, to tie all these components together,

you sometimes need additional code embedded in the JSP pages themselves. Recall from Chapter 3

, that JSP lets you put actual Java code in pages using a set of scripting elements. In this chapter we look at how you can use these scripting elements and when it makes sense to do so.

We start with a brief introduction to the Java language constructs you're likely to use in a JSP page. If you already know Java by heart you can safely skip the first section. But if you have never written a

Java program, or are still a "newbie," you should read it carefully. Don't expect to become a Java guru after reading this introduction, of course. The Java language, combined with the standard libraries, provides many powerful features not covered here. To learn more about Java, I recommend that you read one of the many books dedicated to the language and its libraries, for instance Java in a Nutshell and Java Examples in a Nutshell, both by David Flanagan (O'Reilly).

© 2001, O'Reilly & Associates, Inc.

Chapter 7. Error Handling and Debugging

When you develop any application that's more than a trivial example, errors are inevitable. A JSPbased application is no exception. There are many types of errors you will deal with. Simple syntax errors in the JSP pages are almost a given during the development phase. And even after you have fixed all the syntax errors, you may still have to figure out why the application doesn't work as you intended due to design mistakes. The application must also be designed to deal with problems that can occur when it's deployed for production use. Users can enter invalid values and try to use the application in ways you never imagined. External systems, such as databases, can fail or become unavailable due to network problems.

Since a web application is the face of a company, making sure it behaves well, even when the users misbehave and the world around it falls apart, is extremely important for a positive customer perception. Proper design and testing is the only way to accomplish this goal. Unfortunately, many developers seem to forget the hard-learned lessons from traditional application development when designing web applications. For instance, a survey of 100 e-commerce managers, conducted by

InternetWeek magazine (April 3, 2000 issue), shows that 50% of all web site problems were caused by application coding errors. That's the highest ranking reason in the survey, ahead of poor server performance (38%), poor service provider performance (35%), and poor network performance (22%).

In this chapter, we look at the types of problems you can expect during development, as well as those common in a production system. We see how you can track down JSP syntax and design errors, and how to deal with runtime problems in a graceful manner.

© 2001, O'Reilly & Associates, Inc.

Chapter 8. Sharing Data Between JSP Pages,

Requests, and Users

So far we've covered the JSP basics: how to generate dynamic content and capture user input using JSP standard action elements for working with beans, how to do conditional processing and embed Java code in pages using JSP scripting elements, and how to locate and fix different types of errors in a JSP page. With that out of the way, we can turn our attention to the JSP features and techniques needed to develop real applications.

Any real application consists of more than a single page, and multiple pages often need access to the same information and server-side resources. When multiple pages are used to process the same request, for instance one page that retrieves the data the user asked for and another that displays it, there must be a way to pass data from one page to another. In an application in which the user is asked to provide information in multiple steps, such as an online shopping application, there must be a way to collect the information received with each request and get access to the complete set when the user is ready. Other information and resources need to be shared among multiple pages, requests, and all users. Examples are information about currently logged-in users, database connection pool objects, and cache objects to avoid frequent database lookups.

In this chapter you will learn how scopes in JSP provide access to this type of shared data. You will also see how using multiple pages to process a request leads to an application that's easier to maintain and expand, and learn about a JSP action that lets you pass control between the different pages.

© 2001, O'Reilly & Associates, Inc.

Chapter 9. Database Access

Almost any web application you see on the Internet accesses a database. Databases are used to store customer information, order information, product information, even discussion forum messages—in short, all information that needs to survive a server restart and is too complex to handle in plain text files.

There are many types of databases used in the industry today. However, relational databases are by far the most common. A relational database uses tables to represent the information it handles. A table consists of rows of columns, with each column holding a single value of a predefined datatype.

Examples of these data types are text data, numeric data, dates, and binary data such as images and sound. A specialized language called Structured Query Language (SQL) is used to access the data.

SQL is an ANSI standard and is supported by all major database vendors.

Relational database engines come in all shapes and sizes, from simple one-person databases with limited features to sophisticated databases capable of handling large numbers of concurrent users, with support for transactions distributed over multiple servers and extremely optimized search algorithms.

Even though they all use SQL as the data access language, the API used to execute SQL statements is different for each database engine. To help programmers write code that's portable between database engines, the standard Java libraries include an API called the Java Database Connectivity (JDBC) API.

JDBC defines a set of classes that can be used to execute SQL statements the same way in any relational database.

The complexity of databases varies extensively. A database for an online discussion forum, for instance, requires only one or two tables, while a database for a human resources system may contain hundreds of related tables. In this chapter, we look at a set of generic database custom actions you can use to build any type of database-driven web application. But if the database is complex, you may want to use another approach: hiding the database behind application-specific beans and custom actions, or moving all the database processing to a servlet and using JSP only to show the result. Both these approaches are discussed briefly at the end of this chapter, and in more detail in

Chapter 13

,

Chapter

14 , and

Chapter 17

.

© 2001, O'Reilly & Associates, Inc.

Chapter 13. Web Application Models

Part II of this book described how you can create many different types of applications using only JSP

pages with generic components, such as custom actions and beans, to access databases, present content in different languages, protect pages, and so forth—all without knowing much about Java programming. This approach works fine for many types of web applications, such as employee registers, product catalogs, and conference room reservation systems. But for applications with complicated schemas, intricate business rules, and tricky control flows, the generic components just don't cut it, and you suddenly find that you need a more powerful way to handle the request processing and the business logic.

As I mentioned in

Chapter 3

, JSP pages can also be combined with other Java technologies such as servlets and EJB in more complex applications. In this chapter, we look at how JSP fits into this larger

picture. After this brief description of the most common application models, Chapter 14 , describes the

combination of servlets and JSP pages in detail.

The material presented in this part of the book is geared towards Java programmers. If you're not a programmer, you may still want to browse through this part to get a feel for the possibilities, but don't expect to understand everything. To really appreciate the techniques described in this part of the book, you should have experience with Java programming in general and also be familiar with Java servlets.

© 2001, O'Reilly & Associates, Inc.

Part II: JSP Application Development

The focus of this part of the book is on developing JSP-based web applications using both standard JSP elements and custom components. Through the use of practical examples, you will learn how to handle common tasks such as validating user input, accessing databases, authenticating users and protecting web pages, localizing your web site, and more.

Chapter 5

Chapter 6

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Chapter 11

Chapter 12

© 2001, O'Reilly & Associates, Inc.

Chapter 10. Authentication and Personalization

Authentication means establishing that a user really is who he or she claims to be. Today, it's typically done by asking the user for a username and a matching password, but other options are becoming more and more common. For example, most web servers support client certificates for authentication.

Biometrics, which is the use of unique biological patterns like fingerprints for identification, will likely be another option in the near future. What's important is that an application should not be concerned with the way a user has been authenticated (since the method may change), but only that he or she has passed the test.

Access control, or authorization, is another security mechanism that's strongly related to authentication.

Different users may be allowed different types of access to the content and services a web site offers.

When you have established who the user is through an authentication process, access control mechanisms let you ensure that the user can access only what he or she is allowed to access.

In the end, authentication provides information about who the user is, and that's what is needed to provide personalized content and services. For some types of personalization, the procedures we might think of as authentication may be overkill. If the background colors and type of news listed on the front page are the extent of the personalization, a simple cookie can be used to track the user instead. But if personalization means getting access to information about taxes, medical records, or other confidential personal information, true authentication is definitely needed.

In this chapter we look at different approaches to authentication and access control with JSP, and we use the information about who the user is to provide modest personalization of the application pages.

Security, however, is about more than authentication and access control. So the last section of this chapter presents a brief summary of other areas that need to be covered for applications dealing with sensitive data.

© 2001, O'Reilly & Associates, Inc.

Chapter 11. Internationalization

Taking the term World Wide Web literally means that your web site needs to respect the languages and customs of all visitors, no matter where they come from. More and more, large web sites provide content in several different languages. Just look at a site like Yahoo!, which provides directory services in the local languages of more than 20 countries in Europe, Asia, and North America. Other good examples are CNN, with local news for 8 different countries, and Vitaminic

( http://www.vitaminic.com/ ), a site with MP3 music and artist information customized for different countries. If the site contains only static content, it's fairly easy to support multiple languages: just make a static version of the site for each language. But this approach is not practical for a site with dynamic content. If you develop a separate site for each language, you will have to duplicate the code that generates the dynamic content as well, leading to maintenance problems when errors are discovered or when it's time to add new features. Luckily, Java and JSP provide a number of tools to make it easier to develop one version of a site that can handle multiple languages.

The process of developing an application that caters to the needs of users from different parts of the world includes two phases: internationalization and localization.

Internationalization means preparing the application by identifying everything that will be different for different geographical regions, and providing means to use different versions of all these items instead of hardcoded values. Examples of this are labels and messages, online help texts, graphics, format of dates, times, and numbers, currencies, measurements, and sometimes even the page layouts and colors.

You should note that instead of spelling out the word internationalization, the abbreviation I18N is often used. It stands for "an I followed by 18 characters and an N."

When an application has been internationalized, it can also be localized for different regions. This means providing the messages, the help texts, the graphics and so forth, as well as the rules for formatting dates, times, and numbers, for one or more regions that the internationalized application can use. Localization is sometimes abbreviated L10N, following the same logic as the I18N abbreviation.

The set of localized information for one region is called a locale. Support for new locales can be added without changing the application itself.

In this chapter, we first look at the basic Java classes used for internationalization. If you're not a programmer, you can skim through this section without worrying about the details. (However, you should understand the terminology, and knowing a bit about the inner workings of these classes also makes it easier to understand the rest of the chapter.) We then develop a web application in which visitors can answer a poll question and see statistics over how other visitors have answered, using a set of custom actions that hide the Java classes to make internationalization a lot easier. You can reuse these custom actions in your own application to handle most internationalization needs. The poll site is

localized for three languages. The initial language is based on the user's browser configuration. The user can also explicitly select one of the supported languages.

© 2001, O'Reilly & Associates, Inc.

Chapter 12. Bits and Pieces

In the previous chapters, I have demonstrated the standard JSP features as well as a number of custom actions through practical, complete examples. But some features are hard to fit nicely into these examples without losing focus, so they are described separately in this chapter instead. Things covered here include buffering of the response body, ways to include shared page fragments, using XML and

XSL with JSP, using client-side code to provide a more interactive interface, preventing JSP pages from being cached, and a discussion about the different types of URIs used in JSP pages.

© 2001, O'Reilly & Associates, Inc.

Chapter 14. Combining Servlets and JSP

As I described in the previous chapter, combining servlets and JSP pages lets you clearly separate the application logic from the presentation of the application; in other words, it lets you use the most appropriate component type for the roles of Model, View, and Controller. To illustrate how a servlet can act as the Controller for an application—using beans as the Model and JSP pages as Views—we redesign the Project Billboard application from

Chapter 10 , in this chapter. Along the way, we look at

how servlets and JSP pages can share data, how to deal with URL references between servlets and JSP pages in a flexible manner, and how to handle runtime errors consistently in an application that mixes these two technologies.

Java servlets offer a powerful API that provides access to all the information about the request, the

session, and the application data maintained as servlet context attributes. Chapter 2

, contains a very brief introduction to the servlet API, and

Appendix B

, contains reference material for the main classes and interfaces. To really make use of the techniques described in this chapter, however, you need to know more. If you haven't worked with servlets, I recommend that you read up on them (try Jason

Hunter and William Crawford's Java Servlet Programming from O'Reilly) before you apply the ideas presented here in your own application.

© 2001, O'Reilly & Associates, Inc.

Appendix B. JSP API Reference

Besides the JSP elements described in

Appendix A

, the JSP specification also defines a number of Java classes and interfaces. Instances of some of these classes are assigned to the implicit variables available to scripting elements in a JSP page. Others are used for development of custom actions and to allow

JSP container vendors to encapsulate internal implementations. This appendix describes the classes and interfaces in all these categories.

© 2001, O'Reilly & Associates, Inc.

Appendix A. JSP Elements Syntax Reference

JSP defines three types of elements: directives, scripting elements, and action elements. In addition, you can define your own custom actions. This appendix contains descriptions of all JSP elements as well as the general syntax rules for custom actions.

© 2001, O'Reilly & Associates, Inc.

Chapter 17. Developing Database Access

Components

In this final chapter, we look at more examples of how to develop custom actions, namely the database custom actions introduced in

Chapter 9

.

Before digging into the code for these actions, a number of fundamental Java database access features are discussed. First, we take a look at the JDBC Connection class, and how pooling Connection objects helps solve a number of common problems. We look at two ways to provide connection pooling capabilities to an application: with JDBC 2.0, and by letting a JDBC 1.0 connection pool simulate a JDBC 2.0 pool. The purpose of a connection pool is to be able to share database connections between all components of an application. The approach discussed in this chapter is to use an application initialization servlet that makes the pool available to all servlets and JSP pages.

No matter if you use a servlet or a custom action in a JSP page to access the database, there are a number of things to think about. We look at a generic database access bean and related classes that take care of datatype issues and make the result of a query easy to access. Next, we look at how the bean is used by the database access custom actions described in

Chapter 9

. You can also use the bean directly in servlets, as described in

Chapter 15 , or in your own application-specific database access actions. The

last section contains an example of an application-specific custom action using the bean.

To really appreciate the material in this chapter, you should already be familiar with JDBC. If this is not the case, I recommend that you look at the JDBC documentation online at http://java.sun.com/products/jdbc/ or read a book about JDBC, such as George Reese's Database

Programming with JDBC and Java (O'Reilly).

© 2001, O'Reilly & Associates, Inc.

Chapter 15. Developing JavaBeans for JSP

The JavaBeans specification

[1]

was primarily developed with graphical components in mind. But

JavaBeans represents a design pattern for components that also makes sense for faceless components used to structure a server-side application. The JSP specification provides a number of standard actions to support the use of JavaBeans, as described in the previous chapters.

[1]

This specification is available at http://java.sun.com/beans/docs/spec.html

.

You can use JavaBeans components in a pure JSP application to structure the application and minimize the amount of code needed in the JSP pages. In an application that uses both servlets and JSP pages, beans can be used as a carrier of data between the two domains. By using beans with an eye towards the recommendations in the J2EE application programming model, you can also make it easier to migrate the business logic to Enterprise JavaBeans when warranted by new requirements.

In this chapter, we look at the JavaBeans model and how it applies to the type of faceless JavaBeans you can use in a server-side application, using some of the beans from previous chapters as examples.

© 2001, O'Reilly & Associates, Inc.

Chapter 16. Developing JSP Custom Actions

Custom actions let you encapsulate logic and make it available to page authors in a familiar format.

Throughout this book, a number of generic custom actions are used for such tasks as accessing a database, including localized content, encoding URLs, and much more. Using these actions, the amount of Java code in the JSP pages can be kept to a minimum, making the application easier to debug and maintain. However, for a complex application, the generic actions presented in this book are not enough. Perhaps you want to develop application-specific actions to access the database instead of putting SQL statements in the JSP pages. Or you may want to present complex data as a set of nested

HTML tables with cells formatted differently depending on their values. Instead of using conditional scripting code in the JSP page to generate this table, an application-specific custom action can be used.

Custom actions know about their environment. They automatically get access to all information about the request, the response, and all the variables in the JSP scopes. Another common use for a custom action is as an HTTP-specific adapter to a bean. JavaBeans components are frequently used in a JSP application, and a bean is easier to reuse if it doesn't know about the environment where it's used.

To develop a custom action, you use a set of classes and interfaces referred to in the JSP 1.1 specification as the tag extension mechanism. The simplest custom action implementation is just a class with bean-like accessor methods plus a couple of other well-defined methods. But it's a very powerful mechanism, letting you develop custom actions to do pretty much anything. As always, with increased power comes some amount of complexity. For more advanced actions you need to implement additional methods, and in some cases an extra class. But it's still not rocket science. We'll take it step by step, starting with the most common and simple cases, and then work through some examples of the advanced features in the later sections of this chapter.

© 2001, O'Reilly & Associates, Inc.

Appendix C. Book Example Custom Actions and

Classes Reference

This appendix contains reference material for all custom actions, utility classes, and beans described in this book that can be used as-is in other applications.

Example code in this book that is not intended to be reused directly is not included in this appendix. All source code for the book can, however, be downloaded either from the O'Reilly web site at http://www.oreilly.com/catalog/jserverpages/ or from the web site dedicated to this book at http://www.thejspbook.com/ .

© 2001, O'Reilly & Associates, Inc.

Appendix D. Web-Application Structure and

Deployment Descriptor Reference

A complete web application may consist of several different resources: JSP pages, servlets, applets, static HTML pages, custom tag libraries, and other Java class files. Version 2.2 of the servlet specification defines a portable way to package all these resources together with a deployment descriptor that contains configuration information, such as how all the resources fit together, security requirements, etc. This appendix describes the standard file structure for a web application, and how to use the deployment descriptor to configure the application.

© 2001, O'Reilly & Associates, Inc.

Appendix E. JSP Resource Reference

This appendix contains references to JSP-related products, web hosting services, and sites where you can learn more about JSP and related technologies.

© 2001, O'Reilly & Associates, Inc.

Part I: JSP Application Basics

This part of the book describes the fundamentals of HTTP (the protocol used by all web applications), how servlets and JSP are related, and how to set up a JSP development environment and install the book examples.

Chapter 1

Chapter 2

Chapter 3

Chapter 4

© 2001, O'Reilly & Associates, Inc.

Part III: JSP in J2EE and JSP Component

Development

If you're a programmer, this is the part of the book where the real action is . Here you will learn how to develop your own custom actions and JavaBeans, and how to combine

JSP with other Java server-side technologies, such as servlets and Enterprise JavaBeans

(EJB).

Chapter 13

Chapter 14

Chapter 15

Chapter 16

Chapter 17

© 2001, O'Reilly & Associates, Inc.

About the Examples

This book contains over 50 examples that demonstrate useful techniques for database access, application-controlled authentication and access control, internationalization, XML processing, and more. The examples include complete applications, such as an online shopping site, an employee directory, and a personalized project billboard, as well as numerous smaller examples and page fragments. The included example tag library contains more than 20 custom actions that you can use directly in your application or as a starting point for your own development. The code for all the examples and most of the custom actions is contained within the text; you can also download all code from the O'Reilly web site at http://www.oreilly.com/catalog/jserverpages/ . In addition, you can see all the examples in action at http://www.thejspbook.com/ .

All examples have been tested with the official JSP reference implementation, Apache Tomcat, on

Windows (98 and NT 4.0) and Linux (Red Hat Linux 6.2) using Sun's Java 2 SDK (1.2.2 and 1.3). If you need more information on downloading and installing the Apache Tomcat server for use with the

examples, see Chapter 4 .

© 2001, O'Reilly & Associates, Inc.

Conventions Used in This Book

Italic is used for:

Pathnames, filenames, directories, and program names

New terms where they are defined

Internet addresses, such as domain names and URLs

Boldface is used for:

Particular keys on a computer keyboard

Names of user interface buttons and menus

Constant Width

is used for:

Anything that appears literally in a JSP page or a Java program, including keywords, datatypes, constants, method names, variables, class names, and interface names

Command lines and options that should be typed verbatim on the screen

All JSP and Java code listings

HTML documents, tags, and attributes

Constant Width Italic

is used for:

General placeholders that indicate that an item should be replaced by some actual value in your own program

Constant width bold

is used for:

Text that is typed in code examples by the user

© 2001, O'Reilly & Associates, Inc.

How to Contact Us

We have tested and verified all the information in this book to the best of our abilities, but you may find that features have changed or that we have let errors slip through the production of the book.

Please let us know of any errors that you find, as well as suggestions for future editions, by writing to:

O'Reilly & Associates, Inc.

101 Morris St.

Sebastopol, CA 95472

1-800-998-9938 (in the U.S. or Canada)

1-707-829-0515 (international/local)

1-707-829-0104 (fax)

You can also send messages electronically. To be put on our mailing list or to request a catalog, send email to: [email protected]

To ask technical questions or to comment on the book, send email to: [email protected]

We have a web site for the book, where we'll list examples, errata, and any plans for future editions.

You can access this page at: http://www.oreilly.com/catalog/jserverpages/

For more information about this book and others, see the O'Reilly web site: http://www.oreilly.com/

© 2001, O'Reilly & Associates, Inc.

Acknowledgments

I love to write and have always wanted to write a book someday. After getting a number of articles about Java servlets and a couple of chapters for a server-side Java book published, my confidence was so high that I sent an email to O'Reilly & Associates and asked if they wanted me to write a book about

JSP. Much to my surprise (I guess my confidence was not so high after all), they said, "Yes!" I knew that it would be more work than I could imagine, and it turned out to be even more than that. But here I am, almost a year later, with 17 chapters and 5 appendixes in a nice stack on my desk, written and rewritten countless times. All that remains is to give thanks to everyone who helped me fulfill this dream.

First, I'd like to thank my editors, Paula Ferguson and Bob Eckstein. Paula was the one who accepted my book proposal in the first place, and then helped me through my stumbling steps of writing the first half of the book. Bob came aboard for the second half, and I'm really grateful to him for thoroughly reading everything and giving me helpful advice.

Thanks also to Rob Romano for doing the illustrations, to Christien Shangraw for helping out with the coordination, and to all the production people behind the scenes at O'Reilly who made sure the book got published.

Big thanks also go to the JSP and servlet specification leads, Eduardo Pelegri-Llopart and Danny

Coward, for providing feedback, answering all my questions, and clarifying the vague and ambiguous areas of the specifications. You helped me more than I could ask for. I hope my contributions to the specifications repay my debt to some extent.

Thanks also to all of you who helped me improve the book in other ways: Jason Hunter for letting me borrow his connection pool code and Japanese examples; Craig McClanahan, Larry Riedel, Steve Jung

(Steve dedicates his effort to the memory of his father, Arthur H. Jung, who passed away March 17,

2000), Sean Rohead, Jerry Croce, Steve Piccolo, and Vikram David for reviewing the book and giving me many suggestions for how to make it better; all the Apache Tomcat developers for making a great

JSP reference implementation; and the members of the jsp-interest mailing list for all the ideas about what to cover in this book.

Finally, thanks to everyone who encouraged me and kept my spirits high: Mom, Dad, and my sister, for their support and for teaching me to do what I believe in; all my old friends in Sweden, especially

Janne Ek, Peter Hellström (and his dad, who helped me with the translation of the German example),

Janne Andersson, Roger Bjärevall and Michael Rohdin; Anne Helgren, my writing teacher who convinced me I could do this; and all the guys in and around Vesica Pisces

( http://www.vesicapisces.com/ ), Kelly, Brian, Adam, Bill, and James: I really enjoyed getting away from the writing now and then to hang with you and listen to you play.

-Hans Bergsten, September 2000

© 2001, O'Reilly & Associates, Inc.

1.1 What Is JavaServer Pages?

Put succinctly, JavaServer Pages is a technology for developing web pages that include dynamic content. Unlike a plain HTML page, which contains static content that always remains the same, a JSP page can change its content based on any number of variable items, including the identity of the user, the user's browser type, information provided by the user, and selections made by the user. As you'll see later in the book, functionality such as this can be used to create web applications like shopping carts and employee directories.

A JSP page contains standard markup language elements, such as HTML tags, just like a regular web page. However, a JSP page also contains special JSP elements that allow the server to insert dynamic content in the page. JSP elements can be used for a wide variety of purposes, such as retrieving information from a database or registering user preferences. When a user asks for a JSP page, the server executes the JSP elements, merges the results with the static parts of the page, and sends the dynamically composed page back to the browser, as illustrated in

Figure 1.1

.

Figure 1.1. Generating dynamic content with JSP elements

JSP defines a number of standard elements useful for any web application, such as accessing JavaBeans components, passing control between pages, and sharing information between requests, pages, and users. Programmers can also extend the JSP syntax by implementing application-specific elements that perform tasks such as accessing databases and Enterprise JavaBeans, sending email, and generating

HTML to present application-specific data. The combination of standard elements and custom elements allows for the creation of powerful web applications.

© 2001, O'Reilly & Associates, Inc.

1.2 Why Use JSP?

In the early days of the Web, the Common Gateway Interface (CGI) was the only tool for developing dynamic web content. However, CGI is not an efficient solution. For every request that comes in, the web server has to create a new operating system process, load an interpreter and a script, execute the script, and then tear it all down again. This is very taxing for the server and doesn't scale well when the amount of traffic increases.

Numerous CGI alternatives and enhancements, such as FastCGI, mod_ perl from Apache, NSAPI from Netscape, ISAPI from Microsoft, and Java Servlets from Sun Microsystems, have been created over the years. While these solutions offer better performance and scalability, all of these technologies suffer from a common problem: they generate web pages by embedding HTML directly in programming language code. This pushes the creation of dynamic web pages exclusively into the realm of programmers. JavaServer Pages, however, changes all that.

1.2.1 Embedding Elements in HTML Pages

JSP tackles the problem from the other direction. Instead of embedding HTML in programming code,

JSP lets you embed specialized code (sometimes called scripting code) into HTML pages. Java is the default scripting language of JSP, but the JSP specification allows for other languages as well, such as

JavaScript, Perl, and VBScript. We will begin looking at all the JSP elements in detail later, but at this point let's introduce you to a simple JSP page:

<html>

<body bgcolor="white">

<% java.util.Date clock = new java.util.Date( ); %>

<% if (clock.getHours( ) < 12) { %>

<h1>Good morning!</h1>

<% } else if (clock.getHours( ) < 18) { %>

<h1>Good day!</h1>

<% } else { %>

<h1>Good evening!</h1>

<% } %>

Welcome to our site, open 24 hours a day.

</body>

</html>

This page inserts a different message to the user based on the time of day: "Good morning!" if the local time is before 12:00 P.M., "Good day!" if between 12:00 P.M. and 6:00 P.M., and "Good evening!" if

after 6:00 P.M. When a user asks for this page, the JSP-enabled web server executes all the highlighted

Java code and creates a static page that is sent back to the user's browser. For example, if the current time is 8:53 P.M., the resulting page sent from the server to the browser looks like this:

<html>

<body bgcolor="white">

<h1>Good evening!</h1>

Welcome to our site, open 24 hours a day.

</body>

</html>

A screen shot of this result is shown in Figure 1.2

. If you're not a programmer, don't worry if you didn't

pick up what happened here. Everything will become clear as you progress through this book.

Figure 1.2. The output of a simple JSP page

Of course, embedding too much code in a web page is no better than programming too many HTML tags in server-side code. Fortunately, JSP servers provide a number of reusable action elements that

perform common tasks, as we'll see starting in Chapter 3 . These action elements look similar to HTML

elements, but behind the scenes they are componentized Java programs that the server executes when the page is requested by a user. Action elements are a powerful feature of JSP, as they give web page authors the ability to perform complex tasks without having to do any programming.

In addition to the standard action elements, in-house programmers and third parties can develop custom action elements (known as custom actions or custom tags, and packaged in custom tag libraries) that web page authors can use to handle even more complex and specialized tasks. This book includes a large set of custom actions for conditional processing, database access, internationalization, and more.

Custom tag libraries are also available from various open source organizations and commercial companies.

1.2.2 Using the Right Person for Each Task

As I alluded to earlier, JSP allows you to separate the markup language code, such as HTML, from the programming language code used to process user input, access databases, and perform other application tasks. One way this separation takes place is through the use of the JSP standard and custom action elements: the elements are implemented with programming code and used the same way as page markup elements in regular web pages. Another way is to combine JSP with other Java

Enterprise technologies. For example, Java servlets can handle input processing, Enterprise JavaBeans

(EJB) can take care of the application logic, and JSP pages can provide the user interface.

This separation means that with JSP, a typical business can divide its efforts among two camps that excel in their own areas of expertise, and comprise a JSP web development team with programmers who create the actions for the logic needed by the application, and page authors who craft the specifics of the interface and use the complex actions without having to do any programming. I'll talk more about this benefit as we move through the book, although I should reiterate that the first half of the book is devoted more to those without programming experience, while the second half is for programmers who wish to use the JSP libraries to create their own actions.

1.2.3 Precompilation

Another benefit that is important to mention is that a JSP page is always compiled before it's processed by the server. Remember that older technologies such as CGI/Perl require the server to load an interpreter and the target script each time the page is requested. JSP gets around this problem by compiling each JSP page into executable code the first time it is requested, and invoking the resulting code directly on all subsequent requests. When coupled with a persistent Java virtual machine on a JSPenabled web server, this allows the server to handle JSP pages much faster.

1.2.4 Integration with Enterprise Java APIs

Finally, because JavaServer Pages is built on top of the Java Servlets API, JSP has access to all of the powerful Enterprise Java APIs, including:

JDBC

Remote Method Invocation (RMI) and OMG CORBA support

JNDI ( Java Naming and Directory Interface)

Enterprise JavaBeans (EJB)

JMS ( Java Message Service)

JTA ( Java Transaction API)

This means that you can easily integrate JavaServer Pages with your existing Java Enterprise solutions, or take advantage of many aspects of enterprise computing if you're starting from scratch.

1.2.5 Other Solutions

At this point, let's digress and look at some other solutions for dynamic web content. Some of these solutions are similar to JSP, while others are descendants of older technologies. Many do not have the unique combination of features and portability offered by JavaServer Pages.

1.2.5.1 Active Server Pages (ASP)

Microsoft's Active Server Pages (ASP) is a popular technology for developing dynamic web sites. Just like JSP, ASP lets a page author include scripting code, such as VBScript and JScript, in regular web pages to generate the dynamic parts. For complex code, COM (ActiveX) components written in a programming language such as C++ can be invoked by the scripting code. The standard distribution includes components for database access and more, and other components are available from third parties. When an ASP page is requested, the code in the page is executed by the server. The result is inserted into the page and the combination of the static and dynamic content is sent to the browser.

ASP+, currently in beta, will add a number of new features to ASP. As an alternative to scripting, dynamic content can be generated by HTML/XML-like elements similar to JSP action elements. For improved performance, ASP+ pages will be compiled instead of interpreted, and compiled languages such as C++, C#, and VisualBasic will be added to the current list of scripting languages that can be embedded in a page.

ASP is bundled with Microsoft's Internet Information Server (IIS). Due to its reliance on native COM code as its component model, it's primarily a solution for the Windows platform. Limited support for other platforms, such as the Apache web server on Unix, is available through third-party products such as Chili!Soft (Chili!Soft), InstantASP (Halcyon Software), and OpenASP (ActiveScripting.org). You can read more about ASP and ASP+ on Microsoft's web site, http://www.microsoft.com/ .

1.2.5.2 PHP

PHP

[1]

is an open source web scripting language. Like JSP and ASP, PHP allows a page author to include scripting code in regular web pages to generate dynamic content. PHP has a C-like syntax with some features borrowed from Perl, C++, and Java. Complex code can be encapsulated in both functions and classes. A large number of predefined functions are available as part of PHP, such as accessing databases, LDAP directories, and mail servers, creating PDF documents and images, and encrypting and decrypting data. A PHP page is always interpreted by the server when it's requested, merging the result of executing the scripts with the static text in the page, before it's returned to the browser. The latest version is PHP 4, which uses compiled pages instead of interpreted pages to improve

performance.

[1]

The precursor to PHP was a tool called Personal Home Page. Today PHP is not an acronym for anything; it's simply the name of the product.

PHP is supported on a wide range of platforms, including all major web servers, on operating systems like Windows, Mac OS, and most Unix flavors, and with interfaces to a large number of database engines. More information about PHP is available at http://www.php.net/ .

1.2.5.3 ColdFusion

Allaire's ColdFusion product is another popular alternative for generating dynamic web content. The dynamic parts of a page are generated by inserting HTML/XML-like elements, known as the

ColdFusion Markup Language (CFML), into web pages. CFML includes a large set of elements for tasks like accessing databases, files, mail servers, and other web servers, as well as conditional processing elements like loops. The latest version of ColdFusion also includes elements for communication with Java servlets and Enterprise JavaBeans. Custom elements can be developed in

C++ or Java to encapsulate application-specific functions, and CFML extensions are available from third parties. ColdFusion did not initially support scripting languages, but in ColdFusion 4.5, JavaScriptlike code can be embedded in the web pages in addition to the CFML tags.

The ColdFusion 4.5 Enterprise Edition is supported on Windows, Solaris, HP/UX, and Linux for all major web servers and databases. For more information, visit Allaire's web site at http://www.allaire.com/ .

1.2.5.4 Java servlet template engines

A Java servlet template engine is another technology for separating presentation from processing.

When servlets became popular, it didn't take long before developers realized how hard it was to maintain the presentation part when the HTML code was embedded directly in the servlet's Java code.

As a result, a number of so-called template engines have been developed as open source products to help get HTML out of the servlets. These template engines are intended to be used together with pure code components (servlets) and use only web pages with scripting code for the presentation part.

Requests are sent to a servlet that processes the request, creates objects that represent the result, and calls on a web page template to generate the HTML to be sent to the browser. The template contains scripting code similar to the alternatives described earlier. The scripting languages used by these engines are less powerful, however, because scripting is intended only for reading data objects and generating HTML code to display their values. All the other products and technologies support generalpurpose languages, which can (for better or for worse) be used to include business logic in the pages.

Two popular template engines are WebMacro ( http://www.webmacro.org/ ) and FreeMarker

( http://freemarker.sourceforge.net/ ).

1.2.6 The JSP Advantage

JSP 1.1 combines the most important features found in the alternatives:

JSP supports both scripting and element-based dynamic content, and allows programmers to develop custom tag libraries to satisfy application-specific needs.

JSP pages are precompiled for efficient server processing.

JSP pages can be used in combination with servlets that handle the business logic, the model supported by Java servlet template engines.

In addition, JSP has a couple of unique advantages that make it stand out from the crowd:

JSP is a specification, not a product. This means vendors can compete with different implementations, leading to better performance and quality.

JSP is an integral part of J2EE, a complete platform for Enterprise class applications.

© 2001, O'Reilly & Associates, Inc.

1.3 What You Need to Get Started

Before we begin, let's quickly look at what you need to run the examples and develop your own applications. You really need only three things:

A PC or workstation with a connection to the Internet, so you can download the software you need

A Java 2-compatible Java Software Development Kit ( Java 2 SDK)

A JSP 1.1-enabled web server, such as Apache Tomcat from the Jakarta Project

The Apache Tomcat server is the reference implementation for JSP 1.1. All the examples in this book

were tested on Tomcat. In Chapter 4 , I'll show you how to download, install, and configure the Tomcat

server, as well as all the examples from this book.

In addition, there are a wide variety of other tools and servers that support JSP, from both open source projects and commercial companies. Close to 30 different server products support JSP to date, and roughly 10 authoring tools with varying degrees of JSP support are listed on Sun's JSP web site

( http://java.sun.com/products/jsp/ ).

Appendix E

, also contains a collection of references to JSP-related products, web hosting services, and sites where you can learn more about JSP and related technologies.

You may want to evaluate some of these products when you're ready to start developing your application, but all you really need to work with the examples in this book are a regular text editor, such as Notepad, vi, or Emacs, and of course the Tomcat server.

So let's get going and take a closer look at what JSP has to offer. We need a solid ground to stand on, though, so in the next chapter we will start with the foundations upon which JSP is built: HTTP and

Java servlets.

© 2001, O'Reilly & Associates, Inc.

2.1 The HTTP Request/Response Model

HTTP and all extended protocols based on HTTP are based on a very simple but powerful communications model. Here's how it works: a client, typically a web browser, sends a request for a

resource to a server, and the server sends back a response corresponding to the requested resource (or a response with an error message if it can't deliver the resource for some reason). A resource can be a simple HTML file, or it can be a program that stores the information sent in a database and generates a dynamic response. This request/response model is illustrated in

Figure 2.1

.

Figure 2.1. HTTP request/response with two resources

This simple model implies three things you need to be aware of:

1. HTTP is a stateless protocol. This means that the server does not keep any information about the client after it sends its response, and therefore cannot recognize that multiple requests from the same client may be related.

2. Web applications cannot easily provide the kind of immediate feedback typically found in standalone GUI applications such as word processors or traditional client-server applications.

Every interaction between the client and the server requires a request/response exchange.

Performing a request/response exchange when a user selects an item in a list box or fills out a form element is usually too taxing on the bandwidth available to most Internet users.

3. There's nothing in the protocol that tells the server how a request is made; consequently, the server cannot distinguish between various methods of triggering the request on the client. For example, the HTTP protocol does not allow a web server to differentiate between an explicit request caused by clicking a link or submitting a form and an implicit request caused by resizing

the browser window or using the browser's Back button. In addition, HTTP does not allow the server to invoke client-specific functions, such as going back in the browser history list or sending the response to a certain frame.

Over the years, people have come up with various tricks to overcome the first problem: HTTP's stateless nature. We'll look at them in general terms later in this chapter. The other two problems are harder to deal with, but some amount of interactivity can be achieved by generating a response that includes client-side code (code executed by the browser), such as JavaScript or a Java applet. This approach is discussed briefly in

Chapter 12

.

2.1.1 Requests in Detail

Let's take a closer look at requests. A user sends a request to the server by clicking a link on a web page, submitting a form, or explicitly typing a web page address in the browser's address field. To send a request, the browser needs to know which server to talk to and which resource to ask for. This information is specified by the Uniform Resource Identifier (URI), also commonly referred to as a

Uniform Resource Locator (URL). URI is the general term, while a URL is the specific type of URI used to completely identify a web resource such as an HTML page. Here is an example of a URL:

http://www.gefionsoftware.com/index.html

The first part of this URL specifies that the HTTP protocol is used to request the resource. This is followed by the name of the server, www.gefionsoftware.com. The web server waits for requests to come in on a special TCP/IP port. Port number 80 is the standard port for HTTP requests. If the web server uses another port, the URL must specify the port number in addition to the server name. For example:

http://www.gefionsoftware.com:8080/index.html

This URL is sent to a server that uses port 8080 instead of 80. The last part of the URL, /index.html, identifies the resource that the client is requesting. This is sometimes called the URI path.

The client browser always makes a request by sending a request message. An HTTP request message consists of three things: a request line, request headers, and sometimes a request body.

The request line starts with the request method name, followed by a resource identifier and the protocol version used by the browser:

GET /index.html HTTP/1.0

The most commonly used request method is named GET. As the name implies, a GET request is used to

retrieve a resource from the server. It's the default request method, so if you type a URL in the browser's address field or click on a link, the request will be sent to the server as a GET request.

The request headers provide additional information the server may need to process the request. The message body is included only in some types of requests, like the POST request discussed later.

Here's an example of a valid HTTP request message:

GET /index.html HTTP/1.0

Host: www.gefionsoftware.com

User-Agent : Mozilla/4.5 [en] (WinNT; I)

Accept: image/gif, image/jpeg, image/pjpeg, image/png, */*

Accept-language : en

Accept-charset : iso-8859-1,*,utf-8

The request line specifies the GET method and asks for the resource /index.html to be returned using the

HTTP/1.0 protocol version. The various headers provide additional information the server can use to fulfill the request.

The Host header tells the server the hostname used in the URL. A server may have multiple names, so this information is used to distinguish between multiple virtual web servers sharing the same web server process.

The User-Agent header contains information about the type of browser making the request. The server can use this to send different types of responses to different types of browsers. For instance, if the server knows whether the request is sent by Internet Explorer or Netscape Navigator, it can send a response that takes advantage of each browser's unique features. It can also tell if a browser other than an HTML browser is used, such as a Wireless Markup Language (WML) browser on a cell phone or a

PDA device, and generate an appropriate response.

The Accept headers provide information about the languages and file formats the browser accepts.

These headers can be used to determine the capabilities of the browser and the user's preferences, and adjust the response to use a supported image format and the preferred language. These are just a few of the headers that can be included in a request message. The HTTP specification describes all of them.

The resource identifier (URI) doesn't necessarily correspond to a static file on the server. It can identify an executable program, a record in a database, or pretty much anything the web server knows about.

That's why the generic term resource is used. In fact, there's no way to tell if the /index.html URI corresponds to a file or to something else; it's just a name that means something to the server. The web server is configured to map these unique names to the real resources.

2.1.2 Responses in Detail

When the web server receives the request, it looks at the URI and decides, based on configuration information, how to handle it. It may handle it internally by simply reading an HTML file from the filesystem, or it may forward the request to some component that is responsible for the resource corresponding to the URI. This might be a program that uses a database to dynamically generate an appropriate response. To the client, it makes no difference how the request is handled; all it cares about is getting a response.

The response message looks similar to the request message. It consists of three things: a status line, response headers, and possibly a response body. Here's an example:

HTTP/1.0 200 OK

Last-Modified: Mon, 20 Dec 1999 23:26:42 GMT

Date: Tue, 11 Jan 2000 20:52:40 GMT

Status: 200

Content-Type: text/html

Servlet-Engine: Tomcat Web Server/3.2

Content-Length: 59

<html>

<body>

<h1>Hello World!</h1>

</body>

</html>

The status line starts with the name of the protocol, followed by a result code and a short description of the result code. Here the result code is 200, meaning the request was executed successfully. The response message has headers just like the request message. In this example, the Last-Modified header gives the date and time that the resource was last modified. The client can use this information as a timestamp in a local cache; the next time the user asks for this resource, the client can ask the server to send it only if it's been updated since the last time it was requested. The Content-Type header tells the client what type of response data the body contains, and the Content-Length header shows how large it is. You can likely figure out what the other headers are for. A blank line separates the headers from the message body. Here, the body is a simple HTML page:

<html>

<body>

<h1>Hello World!</h1>

</body>

</html>

Of course, the body can contain a more complex HTML page or any other type of content. For example, the request may return a page with <img> elements. When the browser reads the first response and finds the <img> elements, it sends a new request for the resource identified by each element, often in parallel. The server returns one response for each request, with a Content-Type header telling what type of image it is (for instance, image/gif) and the body containing the bytes that make up the image. All responses are then combined by the browser to render the complete page.

This interaction is illustrated in Figure 2.2

.

Figure 2.2. Interaction between a web client and a server

2.1.3 Request Parameters

Besides the URI and headers, a request message can contain additional information in the form of

parameters. If the URI identifies a server-side program for displaying weather information, for example, request parameters can provide information about which city the user wants to see a forecast for. In an e-commerce application, the URI may identify a program that processes orders, with the user's customer number and the list of items to be purchased transferred as parameters.

Parameters can be sent in one of two ways: tacked on to the URI in the form of a query string , or sent as part of the request message body. Here is an example of a URI with a query string:

http://www.weather.com/forecast?city=Hermosa+Beach&state=CA

The query string starts with a question mark (?) and consists of name/value pairs separated by ampersands (&). These names and values must be URL encoded , meaning that special characters such as whitespace, question marks, ampersands, and all other nonalphanumeric characters are encoded so that they don't get confused with characters used to separate name/value pairs. In this example, the space between Hermosa and Beach is encoded as a plus sign. Other special characters are encoded as their corresponding hexadecimal ASCII value: for instance, a question mark is encoded as %3F. When parameters are sent as part of the request body, they follow the same syntax: URL-encoded name/value pairs separated by ampersands.

2.1.4 Request Methods

As described earlier, GET is the most commonly used request method, intended to retrieve a resource without causing anything else to happen on the server. The POST method is almost as common as GET.

A POST request is intended to request some kind of processing on the server, for instance, updating a database or processing a purchase order.

The way parameters are transferred is one of the most obvious differences between the GET and POST request methods. A GET request always uses a query string to send parameter values, while a POST request always sends them as part of the body (additionally, it can send some parameters as a query string, just to make life interesting). If you code a link to a URI in an HTML page using an <a> element, clicking on the link results in a GET request being sent to the server. Since the GET request uses a query string to pass parameters, you can include hardcoded parameter values in the link URI:

<a href="/forecast?city=Hermosa+Beach&state=CA">

Hermosa Beach weather forecast

</a>

When you use a form to send user input to the server, you can specify whether to use the GET or POST method with the method attribute, as shown below:

<form action="/forecast" method="POST">

City: <input name="city" type="text">

State: <input name="state" type="text">

<p>

<input type="SUBMIT">

</form>

If the user enters "Hermosa Beach" and "CA" in the form fields and clicks on the Submit button, the browser sends a request message like this to the server:

POST /index.html HTTP/1.0

Host: www.gefionsoftware.com

User-Agent : Mozilla/4.5 [en] (WinNT; I)

Accept: image/gif, image/jpeg, image/pjpeg, image/png, */*

Accept-language : en

Accept-charset : iso-8859-1,*,utf-8 city=Hermosa+Beach&state=CA

Due to the differences in how parameters are sent by GET and POST requests, as well as the differences in their intended purposes, browsers handle the requests in different ways. A GET request, parameters and all, can easily be saved as a bookmark, hardcoded as a link, and the response cached by the browser. Also, the browser knows that no damage is done if it sends a GET request again automatically, for instance if the user clicks the Reload or Back button.

A POST request, on the other hand, can not be bookmarked as easily; the browser would have to save both the URI and the request message body. Since a POST request is intended to perform some possibly irreversible action on the server, the browser must also ask the user if it's okay to send the

request again. You have probably seen this type of confirmation dialog, shown in Figure 2.3

, numerous

times with your browser.

Figure 2.3. Repost confirmation dialog

Besides GET and POST, HTTP specifies the following methods:

OPTIONS

HEAD

The OPTIONS method is used to find out what options (e.g., methods) a server or resource offers.

The HEAD method is used to get a response with all headers that would be generated by a GET request, but without the body. It can be used to make sure a link is valid or to see when a

resource was last modified.

PUT

The PUT method is used to store the message body content on the server as a resource identified by the URI.

DELETE

The DELETE method is used to delete the resource identified by the URI.

TRACE

The TRACE method is used for testing the communication between the client and the server.

The server sends back the request message, exactly as it was received, as the body of the response.

Note that these methods are not normally used in a web application.

2.1.5 State Management

As I touched on earlier, HTTP is a stateless protocol; when the server sends back the response corresponding to the request, it forgets all about the transaction. If a user sends a new request, the server has no way of knowing if it is related to the previous request.

This is fine for static content such as regular HTML files, but it's a problem for web applications where a number of requests may be needed to complete a transaction. Consider a shopping cart application: the server-side application needs to allow the user to select items in multiple steps, check the inventory when the user is ready to make the purchase, and finally process the order. In this scenario, the application needs to keep track of information provided by multiple requests from the same browser. In other words, it needs to remember the client's transaction state.

There are two ways to solve this problem, and both have been used extensively for web applications with a variety of server-side technologies. The server can either return the complete state with each response and let the browser send it back as part of the next request; or, it can save the state somewhere on the server and send back only an identifier that the browser returns with the next request. The identifier is then used to locate the state information saved on the server.

In both cases, the information can be sent to the browser in one of three ways:

As a cookie

Embedded as hidden fields in an HTML form

Encoded in the URIs in the response body, typically as links to other application pages (this is known as URL rewriting)

Figure 2.4

outlines these methods.

Figure 2.4. Client state information transportation methods

A cookie is a name/value pair the server passes to the browser in a response header. The browser stores the cookie for the time specified by the cookie's expiration time attribute. When the browser sends a request to a server, it checks its "cookie jar" and includes all cookies it has received from the same server (that have not yet expired) in the request headers. Cookies used for state management don't have an expiration time, and expire as soon as the user closes the browser. Using cookies is the easiest way to deal with the state issue, but cookies are not supported by all browsers. In addition, a user may disable cookies in a browser that does support them because of privacy concerns. Hence, we cannot rely on cookies alone.

If hidden fields in an HTML form are used to send the state information to the browser, the browser returns the information to the server as regular HTTP parameters when the form is submitted. When the state information is encoded in URIs, it is returned to the server as part of the request URI, for instance when the user clicks on an encoded link.

Sending all state information back and forth between the browser and server is not efficient, so most modern server-side technologies employ the idea of keeping the information on the server and passing only an identifier between the browser and the server. This is called session tracking : all requests from a browser that contain the same identifier (session ID) belong to the same session, and the server keeps track of all information associated with the session. As you will see in the next section, the servlet specification hides the mechanisms used to implement session tracking to a large extent, making life easier for the application developer. You will learn how the JSP specification makes it even easier to

use session tracking in Chapter 8 .

A session is valid until it's explicitly terminated (for instance, when the user logs out) or until it's automatically timed out by the server after a period of user inactivity (typically 30 minutes). Note that there's no way for the server to tell if the user closes the browser, since there's no permanent connection between the browser and the server, and no message is sent to the server when the browser disappears.

Still, closing the browser usually means losing the session ID; the cookie expires or the encoded URIs are no longer available. So when the user opens a browser again, the server is unable to associate the new request with the previous session, and therefore creates a new session. However, all the session data associated with the previous session remains on the server until the session times out.

© 2001, O'Reilly & Associates, Inc.

2.2 Servlets

The JSP specification is based on the Java servlet specification. In fact, JSP pages are often combined with servlets in the same application. So to use JSP effectively, it's important to understand the similarities and the concepts that apply to both technologies. In this section, we first take a brief look at what a servlet is, and then discuss the concepts shared by servlets and JSP pages. In

Chapter 3 , we'll

take a closer look at how JSP pages are actually turned into servlets automatically.

If you're already familiar with servlets, this is old news. You can safely skip the rest of this chapter. If you're not familiar with programming, don't worry about the details. The important thing is that you get familiar with the concepts described in the remainder of this chapter.

2.2.1 Advantages Over Other Server-Side Technologies

In simple terms, a servlet is a piece of code that adds new functionality to a server (typically a web server), just like CGI and proprietary server extensions such as NSAPI and ISAPI. But compared to other technologies, servlets have a number of advantages:

Platform and vendor independence

Servlets are supported by all the major web servers and application servers, so a servlet-based solution doesn't tie you to one specific vendor. And because servlets are written in the Java programming language, they can be used on any operating system with a Java runtime environment.

Integration

Servlets are developed in Java and can therefore take advantage of all the other Java technologies, such as JDBC for database access, JNDI for directory access, RMI for remote resource access, etc. Starting with Version 2.2, the servlet specification is part of the Java 2

Enterprise Edition ( J2EE), making servlets an important ingredient of any large-scale enterprise application, with formalized relationships to other server-side technologies such as Enterprise

JavaBeans (EJB).

Efficiency

Servlets execute in a process that runs until the servlet-based application is shut down. Each servlet request is executed as a separate thread in this permanent process. This is far more efficient than the CGI model, where a new process is created for each request. First of all (and

most obviously), a servlet doesn't have the overhead of creating the process and loading the CGI script and possibly its interpreter. But another timesaver is that between requests, servlets can also access resources that remain loaded in the process memory, such as database connections and client state.

Scalability

By virtue of being written in Java and the broad support for servlets, a servlet-based application is extremely scalable. You can develop and test the application on a Windows 98 PC using the standalone servlet reference implementation, and deploy it on anything from a more powerful server running Linux and Apache to a cluster of high-end servers with an application server that supports loadbalancing and failover.

Robustness and security

Java is a strongly typed programming language. This means that you catch a lot of mistakes in the compilation phase that you would only catch during runtime if you used a scripting language like Perl. Java's error handling is also much more robust than C/C++, where an error like division by zero typically brings down the whole server.

In addition, servlets use specialized interfaces to server resources that are not vulnerable to the traditional security attacks. For instance, a CGI Perl script typically uses shell command strings composed of data received from the client to ask the server to do things like sending email.

People with nothing better to do love to find ways to send data that will cause the server to crash, remove all files on the hard disk, or plant a virus or a backdoor when the server executes the command. A CGI script programmer must be very careful to screen all input to avoid these threats, but these problems are almost non-existent with a servlet since it doesn't communicate with the server in the same insecure way.

As you will see in Chapter 3 , JSP inherits all these advantages by being based on the servlet

specification.

2.2.2 Servlet Life Cycle

If you're already a Java programmer, there are some fundamental points you should know about servlets. A servlet is a Java class that uses the Servlet Application Programming Interface (API). The

Servlet API consists of a number of classes and interfaces that define the methods that make it possible to process HTTP requests in a web server-independent manner.

When a web server receives a request that should be handled by a servlet, it first checks if an instance of the specific servlet class exists. If it doesn't, it creates one. This is referred to as loading the servlet.

It then asks the servlet to process the request. Once a servlet has been loaded, the same servlet instance

(object) is called to process succeeding requests. Eventually the web server needs to shut down the servlet, typically when the web server itself is shut down. It first informs the servlet about the shutdown; this gives the objects a chance to do necessary housekeeping, such as closing a database connection, before shutting down.

These three interactions between the web server and the servlet are defined by methods in the javax.servlet.Servlet

interface, and are referred to as the servlet's life-cycle methods. Here are their formal definitions: public void init(ServletConfig config)

The init( ) method is called when the servlet is loaded so it can initialize its state: for instance, set up references to external resources such as a database and read configuration information.

public void service(ServletRequest req, ServletResponse res)

The service( ) method is called to service a request. It's called zero or more times during the servlet's lifetime, and passes objects representing the request and response messages to the servlet.

public void destroy( )

The destroy( ) method is called just before the servlet is taken out of service. It allows the servlet to release references to any external resources it has acquired during its lifetime.

Figure 2.5

illustrates how the web server uses the life-cycle methods.

Figure 2.5. Servlet life cycle

Most interesting to us is the service( ) method. It gives the servlet access to two objects, which are passed as arguments to the method: a ServletRequest object and a ServletResponse object

(when HTTP is used, specialized objects of type HttpServletRequest and

HttpServletResponse

are used instead). Through methods implemented by the

ServletRequest

object, the servlet can access all information known about the request message: parameter values, header values, authentication information, etc. The servlet uses methods of the

ServletResponse

object to generate the response message. It can set headers, the status code, and the actual response body, which is typically a dynamically generated HTML page.

In Chapter 3

, I discuss how a JSP page is turned into a servlet the first time it's requested, and then loaded, called, and shut down in exactly the same way as a regular servlet.

2.2.3 Servlet Containers

A servlet container is the connection between a web server and the servlets. It provides the runtime environment for all the servlets on the server as defined by the servlet specification, and is responsible for loading and invoking those servlets when the time is right.

There are many different types of servlet containers. Some containers are called add-ons, or plug-ins, and are used to add servlet support to web servers without native servlet support (such as Apache and

IIS). They can run in the same operating-system process as the web server or in a separate process.

Other containers are standalone servers. A standalone server includes web server functionality to

provide full support for HTTP in addition to the servlet runtime environment. Containers can also be embedded in other servers, such as a climate-control system, to offer a web-based interface to the system. A container bundled as part of an application server can distribute the execution of servlets over multiple hosts. The server can balance the load evenly over all containers, and some servers can even provide failover capabilities in case a host crashes.

No matter what type it is, the servlet container is responsible for mapping incoming requests to a servlet registered to handle the resource identified by the URI and passing the request message to that servlet. After the request is processed, it is the container's responsibility to convert the response object created by the servlet into a response message and send it back to the client. This is illustrated in

Figure

2.6

.

Figure 2.6. Request dispatching

2.2.4 Servlet Contexts

A servlet container implementing the Servlet 2.1 API (or later) can group servlets and other resources such as JSP pages, HTML pages, and image files into separate servlet contexts. Each servlet context represents a web application, and is associated with a unique URI path prefix called the context path, as

shown in Figure 2.6

. For instance, your human-resources application can be associated with the context

path /hr and your sales-tracking system with the context path /sales. This allows one servlet container to distinguish between applications and dispatch requests like

/sales/report?month=Jan

to the sales tracking application and /hr/emplist to the humanresources application.

The remaining URI path is then used within the selected context to decide how to process the request by comparing it to path mapping rules. Such rules can be set up to send all requests starting with

/report

to one servlet and with /forecast to another. Another type of rule can be set up to let one servlet handle all requests with paths ending with a specific file extension, such as .jsp.

Figure 2.6

shows how the different parts of the URI paths are used to direct the request processing to the right resource through the container and context.

Each context is self-contained and doesn't know anything about other applications running in the same container. All references between the servlets and JSP pages in the application are relative to the context path, and therefore referred to as context-relative paths. By using context-relative paths within the application, a web application can be deployed using any context path. The servlet specification defines a standard packaging format for web applications that all compliant containers know how to

install and associate with a context. This is described in more detail in Section 2.3

.

A web application can be more than just JSP pages, HTML pages, and images. Therefore, a context can

hold on to objects shared by all components of the application,

[1]

such as database connections and other shared resources needed by multiple servlets and JSP pages. This is represented by the

application scope in JSP, and we'll take a closer look at how to use it in

Chapter 8

. Each context also has its own set of configuration data, discussed in more detail in the last section of this chapter.

[1]

There are special considerations for applications distributed over multiple servers. Chapter 13 , describes this in more

detail.

2.2.5 Sessions

Earlier, I mentioned that the Servlet API hides the mechanisms used to implement session tracking to a large extent. A servlet-based application doesn't need to know if the session ID is passed between the server and the browser as a cookie or encoded in the URIs. Instead, the servlet container looks at the information it receives with each request and decides which mechanism to use. If it receives a session

ID cookie, it uses cookie-based tracking; if it receives an encoded URI, it uses URL rewriting. No matter which mechanism is used, the container gives the servlet access to the state information associated with the browser through the request object it passes to the servlet.

The state information is represented by a session object, which is an instance of a Servlet API class named javax.servlet.http.HttpSession. The session object acts as a container for other objects that make up the session state, with methods for adding, getting, and removing these objects.

For instance, in an e-commerce application, the user picks items to buy from an online catalog. When the servlet receives a request to put an item in the shopping cart, it gets the session object from the request and places a Java object representing the item in the session by calling its setAttribute(

)

method. Later, when the user checks out, another servlet picks up all items from the session using other methods, and processes the order.

Since a JSP page is turned into a servlet, it has access to the session in the same way, but JSP makes it even easier to work with session data through the concept of a session scope. We look at all aspects of

sessions from a JSP perspective in Chapter 8 .

© 2001, O'Reilly & Associates, Inc.

2.3 Packaging Java Web Applications

A complete web application may consist of several different resources: JSP pages, servlets, applets, static HTML pages, custom tag libraries and other Java class files. Until very recently, different servers required an application with all these components to be installed and configured in different ways, making it very hard for web application developers to provide easy-to-use installation instructions and tools.

Version 2.2 of the servlet specification defines a portable way to package all these resources together, along with a deployment descriptor. A deployment descriptor is a file that outlines security requirements and describes how all the resources fit together. All files for the web application are placed in an archive file, called a Web Archive (WAR) file. A WAR file has a .war file extension and can be created with the Java jar command or a ZIP utility program such as WinZip (the same compression scheme is used).

All Servlet 2.2-compliant servers can install a WAR file and associate the application with a servlet context. During installation, a server is free to unpack the contents of the file and store it for runtime use in any way it sees fit, but the application developer needs to deal with only one delivery format.

This standardized deployment format also enables server vendors to develop installation and configuration tools that make it easy to install a new web application.

The internal structure for a WAR file is defined by the JSP specification. During development, however, it's often more convenient to work with the web application files in an open filesystem instead of packaging and repackaging them into a WAR file every time you make a change. As a result, most containers support the WAR structure in an open filesystem as well.

The structure required for both is outlined here:

/index.html

/company/contact.html

/products/list.jsp

/images/banner.gif

/WEB-INF/web.xml

/WEB-INF/lib/bean.jar

/WEB-INF/lib/actions.jar

/WEB-INF/classes/com/mycorp/servlets/PurchaseServlet.class

/WEB-INF/classes/com/mycorp/util/MyUtils.class

/WEB-INF/...

The top-level in this structure is the document root for all web application files, such as HTML pages,

JSP pages, and image files—in other words, all the files requested directly by the browser.

You're probably wondering about the WEB-INF directory. This directory contains the application deployment descriptor (web.xml ) as well as subdirectories for other types of resources, such as Java class files and configuration files. A browser does not have access to the files under this directory, so it's safe to place files that you don't want public here.

The deployment descriptor file, web.xml, is a simple XML file. We will get much more familiar with

the contents of this file as we proceed through the book. ( Appendix D , also contains a complete

reference of this file.) In addition, two WEB-INF subdirectories have special meaning if you're a programmer: lib and classes. The lib directory typically contains Java Archive ( JAR) files

(compressed archives of Java class files). As an alternative, class files can be stored in the classes directory without being compressed, which can be convenient during development. However, class files must be stored in subdirectories of the classes directory that mirror their package structure, and must follow standard Java conventions for how class files are organized in a directory tree.

© 2001, O'Reilly & Associates, Inc.

3.1 The Problem with Servlets

In many Java servlet-based applications, processing the request and generating the response are both handled by a single servlet class. A example servlet looks like this: public class OrderServlet extends HttpServlet {

public void doGet(HttpServletRequest request,

HttpServletResponse response)

throws ServletException, IOException {

response.setContentType("text/html");

PrintWriter out = response.getWriter( );

if (isOrderInfoValid(request)) {

saveOrderInfo(request);

out.println("<html>");

out.println(" <head>");

out.println(" <title>Order Confirmation</title>");

out.println(" </head>");

out.println(" <body>");

out.println(" <h1>Order Confirmation</h1>");

renderOrderInfo(request);

out.println(" </body>");

out.println("</html>");

}

...

If you're not a programmer, don't worry about all the details in this code. The point is that the servlet contains request processing and business logic (implemented by methods such as isOrderInfoValid( )

and saveOrderInfo( )) and also generates the response HTML code, embedded directly in the servlet code using println( ) calls. A more structured servlet application isolates different pieces of the processing in various reusable utility classes, and may also use a separate class library for generating the actual HTML elements in the response. But even so, the pure servlet-based approach still has a few problems:

Detailed Java programming knowledge is needed to develop and maintain all aspects of the application, since the processing code and the HTML elements are lumped together.

Changing the look and feel of the application, or adding support for a new type of client (such as a WML client), requires the servlet code to be updated and recompiled.

It's hard to take advantage of web page development tools when designing the application interface. If such tools are used to develop the web page layout, the generated HTML must then be manually embedded into the servlet code, a process that is time-consuming, error-prone, and extremely boring.

Adding JSP to the puzzle lets you solve these problems by separating the request processing and business logic code from the presentation, as illustrated in

Figure 3.1

. Instead of embedding HTML in the code, you place all static HTML in JSP pages, just as in a regular web page, and add a few JSP elements to generate the dynamic parts of the page. The request processing can remain the domain of servlet programmers, and the business logic can be handled by JavaBeans and Enterprise JavaBeans

(EJB) components.

Figure 3.1. Separation of request processing, business logic, and presentation

As I mentioned before, separating the request processing and business logic from presentation makes it possible to divide the development tasks among people with different skills. Java programmers

implement the request processing and business logic pieces, web page authors implement the user interface, and both groups can use best-of-breed development tools for the task at hand. The result is a much more productive development process. It also makes it possible to change different aspects of the application independently, such as changing the business rules without touching the user interface.

This model has clear benefits even for a web page author without programming skills who is working alone. A page author can develop web applications with many dynamic features, using generic Java components provided by open source projects or commercial companies.

© 2001, O'Reilly & Associates, Inc.

3.2 The Anatomy of a JSP Page

A JSP page is simply a regular web page with JSP elements for generating the parts of the page that

differ for each request, as shown in Figure 3.2

.

Everything in the page that is not a JSP element is called template text . Template text can really be any text: HTML, WML, XML, or even plain text. Since HTML is by far the most common web page language in use today, most of the descriptions and examples in this book are HTML-based, but keep in mind that JSP has no dependency on HTML; it can be used with any markup language. Template text is always passed straight through to the browser.

Figure 3.2. Template text and JSP elements

When a JSP page request is processed, the template text and the dynamic content generated by the JSP elements are merged, and the result is sent as the response to the browser.

3.2.1 JSP Elements

There are three types of elements with JavaServer Pages: directive, action, and scripting elements.

The directive elements, shown in

Table 3.1

, are used to specify information about the page itself that remains the same between page requests, for example, the scripting language used in the page, whether session tracking is required, and the name of a page that should be used to report errors, if any.

Element Description

Table 3.1, Directive Elements

<%@ page ... %>

Defines page-dependent attributes, such as scripting language, error page, and buffering requirements

<%@ include ... %> Includes a file during the translation phase

<%@ taglib ... %>

Declares a tag library, containing custom actions, used in the page

Action elements typically perform some action based on information that is required at the exact time the JSP page is requested by a client. An action element can, for instance, access parameters sent with the request to do a database lookup. It can also dynamically generate HTML, such as a table filled with information retrieved from an external system.

The JSP specification defines a few standard action elements, listed in Table 3.2

, and includes a

framework for developing custom action elements. A custom action element can be developed by a programmer to extend the JSP language. The examples in this book use custom actions for database access, internationalization, access control, and more.

Element

Table 3.2, Standard Action Elements

Description

<jsp:useBean>

Makes a JavaBeans component available in a page

<jsp:getProperty> Gets a property value from a JavaBeans component and adds it to the response

<jsp:setProperty> Sets a JavaBeans property value

<jsp:include>

Includes the response from a servlet or JSP page during the request processing phase

<jsp:forward>

Forwards the processing of a request to a servlet or JSP page

<jsp:param>

<jsp:plugin>

Adds a parameter value to a request handed off to another servlet or JSP page using

<jsp:include>

or <jsp:forward>

Generates HTML that contains the appropriate client browser-dependent elements (OBJECT or EMBED) needed to execute an Applet with the Java Plugin software

Scripting elements, shown in

Table 3.3

, allow you to add small pieces of code to a JSP page, such as an

if

statement to generate different HTML depending on a certain condition. Like actions, they are also executed when the page is requested. You should use scripting elements with extreme care: if you embed too much code in your JSP pages, you will end up with the same kind of maintenance problems as with servlets embedding HTML.

Table 3.3, Scripting Elements

Element Description

<% ... %>

Scriptlet, used to embed scripting code.

<%= ... %> Expression, used to embed Java expressions when the result shall be added to the response. Also used as runtime action attribute values.

<%! ... %> Declaration, used to declare instance variables and methods in the JSP page implementation class.

JSP elements, such as action and scripting elements, are often used to work with JavaBeans . Put succinctly, a JavaBeans component is a Java class that complies with certain coding conventions.

JavaBeans are typically used as containers for information that describes application entities, such as a customer or an order. We'll cover each of these element types, as well as JavaBeans, in the following chapters.

© 2001, O'Reilly & Associates, Inc.

3.3 JSP Processing

A JSP page cannot be sent as-is to the browser; all JSP elements must first be processed by the server.

This is done by turning the JSP page into a servlet, and then executing the servlet.

Just as a web server needs a servlet container to provide an interface to servlets, the server needs a JSP container to process JSP pages. The JSP container is often implemented as a servlet configured to handle all requests for JSP pages. In fact, these two containers—a servlet container and a JSP container—are often combined into one package under the name web container (as it is referred to in the J2EE documentation).

A JSP container is responsible for converting the JSP page into a servlet (known as the JSP page

implementation class ) and compiling the servlet. These two steps form the translation phase . The JSP container automatically initiates the translation phase for a page when the first request for the page is received. The translation phase takes a bit of time, of course, so a user notices a slight delay the first time a JSP page is requested. The translation phase can also be initiated explicitly; this is referred to as

precompilation of a JSP page. Precompiling a JSP page avoids hitting the user with this delay, and is discussed in more detail in

Chapter 12 .

The JSP container is also responsible for invoking the JSP page implementation class to process each request and generate the response. This is called the request processing phase. The two phases are

illustrated in Figure 3.3

.

Figure 3.3. JSP page translation and processing phases

As long as the JSP page remains unchanged, any subsequent processing goes straight to the request processing phase (i.e., it simply executes the class file). When the JSP page is modified, it goes through the translation phase again before entering the request processing phase.

So in a way, a JSP page is really just another way to write a servlet without having to be a Java programming wiz. And, except for the translation phase, a JSP page is handled exactly like a regular servlet: it's loaded once and called repeatedly, until the server is shut down. By virtue of being an automatically generated servlet, a JSP page inherits all of the advantages of servlets described in

Chapter

2

: platform and vendor independence, integration, efficiency, scalability, robustness, and security.

Let's look at a simple example of a servlet. In the tradition of programming books for as far back as anyone cares to remember, we start with an application that just writes Hello World, but this time we will

add a twist: our application will also show the current time on the server. Example 3.1

shows a handcoded servlet with this functionality.

Example 3.1. Hello World Servlet

public class HelloWorldServlet implements Servlet {

public void service(ServletRequest request,

ServletResponse response)

throws ServletException, IOException {

response.setContentType("text/html");

PrintWriter out = response.getWriter( );

out.println("<html>");

out.println(" <head>");

out.println(" <title>Hello World</title>");

out.println(" </head>");

out.println(" <body>");

out.println(" <h1>Hello World</h1>");

out.println(" It's " + (new java.util.Date( ).toString( )) +

" and all is well.");

out.println(" </body>");

out.println("</html>");

}

}

As before, don't worry about the details if you're not a Java programmer. What's important here is that the service( )

method is the method called by the servlet container every time the servlet is requested, as described in

Chapter 2

. The method generates all HTML code, using the println( ) method to send the strings to the browser. Note that there's no way you could use a web development tool to develop this type of embedded HTML, adjust the layout with immediate feedback, verify that links are intact, etc. This

example is so simple that it doesn't really matter, but imagine a complex page with tables, aligned images, forms, some JavaScript code, etc., and you see the problem.

Also note the following lines, which add the current date and time to the response (shown in Figure 3.4

):

out.println(" It's " + (new java.util.Date( ).toString( ))

+ " and all is well.");

Figure 3.4. The output from the Hello World servlet

Example 3.2

shows a JSP page that produces the same result as the Hello World servlet.

Example 3.2. Hello World JSP Page

<html>

<head>

<title>Hello World</title>

</head>

<body>

<h1>Hello World</h1>

It's <%= new java.util.Date( ).toString( ) %> and all is well.

</body>

</html>

This is as simple as it gets. A JSP page is a regular HTML page, except that it may also contain JSP elements like the highlighted element in this example. This element inserts the same Java code in the page as was used in the servlet to add the current date and time. If you compare this JSP page to the corresponding servlet, you see that the JSP page can be developed using any web page editor that allows you to insert extra, non-HTML elements. And the same tool can later be used to easily modify the layout.

This is a great advantage over a servlet with embedded HTML.

The JSP page is automatically turned into a servlet the first time it's requested, as described earlier. The generated servlet looks something like in

Example 3.3

.

Example 3.3. Servlet Generated from JSP Page

import javax.servlet.*; import javax.servlet.http.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import java.io.*; import org.apache.jasper.*; import org.apache.jasper.runtime.*; public class _0005chello_0002ejsphello_jsp_1 extends HttpJspBase {

public void _ jspService(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

JspFactory _ jspxFactory = null;

PageContext pageContext = null;

HttpSession session = null;

ServletContext application = null;

ServletConfig config = null;

JspWriter out = null;

Object page = this;

String _value = null;

try {

_ jspxFactory = JspFactory.getDefaultFactory( );

response.setContentType("text/html");

pageContext = _ jspxFactory.getPageContext(this, request,

response,"", true, 8192, true);

application = pageContext.getServletContext( );

config = pageContext.getServletConfig( );

session = pageContext.getSession( );

out = pageContext.getOut( );

out.write("<HTML>\r\n <HEAD>\r\n <TITLE>" +

"Hello World</TITLE>\r\n </HEAD>\r\n" +

" <BODY>\r\n <H1>Hello World</H1>\r\n" +

" It's ");

out.print( new java.util.Date( ).toString( ) );

out.write(" and all is well.\r\n </BODY>\r\n" +

"</HTML>\r\n");

} catch (Exception ex) {

if (out.getBufferSize( ) != 0)

out.clear( );

pageContext.handlePageException(ex);

} finally {

out.flush( );

_ jspxFactory.releasePageContext(pageContext);

}

}

}

The generated servlet in

Example 3.3

looks a lot more complex than the hand-coded version in

Example

3.1

. That's because a number of objects you can use in a JSP page must always be initialized (the hand-

coded version doesn't need this generic initialization). These details are not important now; programming examples later in the book will show you how to use all objects of interest.

Instead, you should note that the servlet generated from the JSP page is a regular servlet. The

_jspService( )

method corresponds to the service( ) method in the hand-coded servlet; it's called every time the page is requested. The request and response objects are passed as arguments to the method, so the JSP page has access to all the same information as does a regular servlet. This means it can read user input passed as request parameters, adjust the response based on header values (like the ones described in

Chapter 2

), get access to the session state, etc.—just like a regular servlet.

The highlighted code section in

Example 3.3

shows how the static HTML from the JSP page in Example

3.2

has been embedded in the resulting code. Also note that the Java code to retrieve the current date and

time has been inserted in the servlet as-is. By letting the JSP container convert the JSP page into a servlet that combines code for adding HTML to the response with small pieces of Java code for dynamic content, you get the best of both worlds. You can use familiar web page development tools to design the static parts of the web page, drop in JSP elements that generate the dynamic parts, and still enjoy all the benefits of servlets.

Client-Side Versus Server-Side Code

Page authors who have some experience developing client-side scripts using JavaScript

(ECMAScript) or VBScript can sometimes get a bit confused when they start to use a serverside technology like JSP.

Client-side scripts, embedded in <script> elements, execute in the browser. These types of scripts are often linked to a form element such as a selection list. When the user selects an item in the list, the associated script is executed, perhaps populating another selection list with appropriate choices. Since all this code is executed by the browser, the client-side script provides immediate feedback to the user.

Server-side code, like action and scripting elements in a JSP page, executes on the server.

Recall from

Chapter 2

that the browser must make a request to the server to execute a JSP page. The corresponding JSP code is then used to produce a dynamic response.

This brings up an important point: there's no way a client-side script can directly call an individual Java code segment in the JSP page. A client-side script can ask the browser to make a request for the complete page, but it can't process the response and use it to do something such as populate a selection list with the data.

It is possible, although not very efficient, to link a user action to a client-side script, invoking an applet that in turn makes a request to a servlet or JSP page. The applet can then read the response and cause some dynamic action in the web browser. This approach may be reasonable on a fast intranet, but you probably won't be happy with the response times if you tried it on the Internet during peak hours. The reason is that the HTTP request/response model was never intended to be used for this type of incremental user interface update.

Consequently, there's a great deal of overhead involved. If you still want to do this, be careful not to open up a security hole. For instance, if you develop an applet that can send any SQL statement to a servlet and get the query result back, you have made it possible for anyone to access all data in your database (that is accessible to the servlet), not just the data that your applet asks for.

Client-side and server-side code can, however, be combined with good results. You can embed client-side scripts as template text in your JSP pages, or generate it dynamically with actions or scripting elements. But keep in mind that it's still client-side code; the fact that it's generated by a JSP page doesn't change anything. A common use of client-side code is to validate user form input. Doing the validation with client-side code gives the user faster feedback about invalid input and reduces the load on the server. But don't forget that clientside scripting is not supported in all browsers, and even if it is, the user may have disabled the execution of scripts. Therefore, you should always perform input validation on the server as well.

Instead of using client-side scripts, you can of course use a Java applet to provide a more interactive user interface. Ideally the applet is self-contained; in other words, it doesn't have to talk to the server at all in order to present a user-friendly interface. If it needs to communicate with the server, however, it can do so using a far more efficient protocol than HTTP. Java

Servlet Programming by Jason Hunter and William Crawford (O'Reilly) includes a chapter about different applet communication options.

© 2001, O'Reilly & Associates, Inc.

3.4 JSP Application Design with MVC

JSP technology can play a part in everything from the simplest web application, such as an online phone list or an employee vacation planner, to full-fledged enterprise applications, such as a human resource application or a sophisticated online shopping site. How large a part JSP plays differs in each case, of course. In this section, we introduce a design model suitable for both simple and complex applications called Model-View-Controller (MVC).

MVC was first described by Xerox in a number of papers published in the late 1980s. The key point of using MVC is to separate components into three distinct units: the Model, the View, and the Controller.

In a server application, we commonly classify the parts of the application as: business logic, presentation, and request processing. Business logic is the term used for the manipulation of an application's data, i.e., customer, product, and order information. Presentation refers to how the application is displayed to the user, i.e., the position, font, and size. And finally, request processing is what ties the business logic and presentation parts together. In MVC terms, the Model corresponds to business logic and data, the View to the presentation logic, and the Controller to the request processing.

Why use this design with JSP? The answer lies primarily in the first two elements. Remember that an application data structure and logic (the Model) is typically the most stable part of an application, while the presentation of that data (the View) changes fairly often. Just look at all the face-lifts that web sites have gone through to keep up with the latest fashion in web design. Yet, the data they present remains the same. Another common example of why presentation should be separated from the business logic is that you may want to present the data in different languages or present different subsets of the data to internal and external users. Access to the data through new types of devices, such as cell phones and

Personal Digital Assistants (PDAs), is the latest trend. Each client type requires its own presentation format. It should come as no surprise, then, that separating business logic from presentation makes it easier to evolve an application as the requirements change; new presentation interfaces can be developed without touching the business logic.

This MVC model is used for most of the examples in this book. In Part II, JSP pages are used as both the Controller and the View, and JavaBeans components are used as the Model. The examples in

Chapter 5 through

Chapter 7 use a single JSP page that handles everything, while

Chapter 8 through

Chapter 11

show how you can use separate pages for Control and View to make the application easier to maintain. Many types of real-world applications can be developed this way, but what's more important is that this approach allows us to examine all the JSP features without getting distracted by

other technologies. In Part III

, we look at other possible role assignments when JSP is combined with servlets and Enterprise JavaBeans.

© 2001, O'Reilly & Associates, Inc.

4.1 Installing the Java Software Development Kit

Tomcat is a pure Java web server with support for the Servlet 2.2 and JSP 1.1 specifications. To use it, you must first install a Java runtime environment. If you don't already have one, you can download a

Java SDK for Windows, Linux, and Solaris at http://java.sun.com/j2se/ .

I recommend that you install the Java 2 SDK as opposed to the slimmed-down Runtime Environment (

JRE) distribution. The reason is that JSP requires a Java compiler, which is included in the SDK but not in the JRE. Sun Microsystems has made the javac compiler from the SDK available separately for redistribution by the Apache Software Foundation. So technically, you could use the JRE and download the Java compiler as part of the Tomcat package, but even as I write this chapter, the exact legal conditions for distributing the compiler are changing.

Another alternative is to use the Jikes compiler from IBM

( http://www10.software.ibm.com/developerworks/opensource/jikes/ ). Tomcat can be configured to use

Jikes instead of the javac compiler from Sun; read the Tomcat documentation if you would like to try this. To make things simple, though, I suggest installing the Java 2 SDK from Sun. The examples were developed and tested with Java 2 SDK, Standard Edition, v1.2.2 and v1.3. I recommend that you use the latest version of the SDK available for your platform.

If you need an SDK for a platform other than Windows, Linux, or Solaris, there's a partial list of ports made by other companies at Sun's web site: http://java.sun.com/cgi-bin/java-ports.cgi/

Also check your operating system vendor's web site. Most operating system vendors have their own

SDK implementation available for free.

Installation of the SDK varies depending on platform but is typically easy to do. Just follow the instructions on the web site where you download the SDK.

Before you install and run Tomcat, make sure that the JAVA_HOME environment variable is set to the installation directory of your Java environment, and that the Java bin directory is included in the PATH environment variable. On a Windows system, you can see if an environment variable is set by typing the following command in a Command Prompt window:

C:\> echo %JAVA_HOME%

C:\jdk1.1.2

If JAVA_HOME is not set, you can set it and include the bin directory in the PATH like this on a

Windows system (assuming Java is installed in C:\jdk1.2.2):

C:\> set JAVA_HOME=C:\jdk1.1.2

C:\> set PATH=%JAVA_HOME%\bin;%PATH%

On a Windows 95/98 system, you can add these commands to the C:\AUTOEXEC.BAT file to set them permanently. Just use a text editor, such as Notepad, and add lines with the set commands. The next time you boot the PC, the environment variables will be set automatically. For Windows NT and 2000, you can set them permanently from the Environment tab in the System Properties tool.

If you use Linux or some other Unix platform, the exact commands depend on which shell you use.

With bash , which is commonly the default for Linux, use the following commands (assuming Java is installed in /usr/local/jdk1.2.2):

[[email protected] /] export JAVA_HOME=/usr/local/jdk1.2.2

[[email protected] /] export PATH=$JAVA_HOME/bin:$PATH

[[email protected] /] echo $PATH

/usr/local/jdk1.2.2/bin:/usr/local/bin:/bin:/usr/bin

© 2001, O'Reilly & Associates, Inc.

4.2 Installing the Tomcat Server

You can download the Tomcat Server either in binary format or as source code that you compile yourself.

If you're primarily interested in learning about JSP, I recommend that you use the binary download to run the examples in this book and develop your own applications. If you're a Java programmer and interested in seeing how Tomcat is implemented, feel free to download the source and take a look at the internals.

The binary distribution is available at: http://jakarta.apache.org/downloads/binindex.html

On this page you find three types of builds:

Release builds

Milestone builds

Nightly builds

Release builds are stable releases that have been tested extensively and verified to comply with the servlet and JSP specifications. Milestone builds are created as intermediary steps towards a release build. They often contain new features that are not yet fully tested, but are generally known to work. A nightly build, however, may be very unstable. It's actually a snapshot of the latest source code and may have been tested only by the person who made the latest change. You should use a nightly build only if you're involved in the development of Tomcat.

You should download the latest release build. All examples in this book were developed and tested using the 3.2 (Beta 3) version, but any release later than 3.2 should work fine as well. When you click on the link for the latest release build and select the bin directory, you see a list of archive files in different

formats, similar to Figure 4.1

.

Figure 4.1. Release build packages

Pick a compression format that's appropriate for your platform. For Windows, select jakarta-tomcat.zip and save it to your hard drive, for instance in a directory named C:\Jakarta. You can unpack the package either with a ZIP utility program such as WinZip, or by using the jar command that's included in the Java distribution. Using the Command Prompt window where you set the JAVA_HOME and PATH environment variables earlier, change directory to the directory where you downloaded the ZIP file and unpack it:

C:\> cd Jakarta

C:\Jakarta> jar xvf jakarta-tomcat.zip

For Unix platforms, download the jakarta-tomcat.tar.gz file, for instance to /usr/local, and use these commands to unpack it (assuming you have GNU tar installed):

[[email protected] /] cd /usr/local

[[email protected] /usr/local] tar xzvf jakarta-tomcat.tar.gz

If you don't have GNU tar installed on your system, you can use this command:

[[email protected] /usr/local] gunzip -c jakarta-tomcat.tar.gz | tar xvf -

This creates a directory structure with a top directory named jakarta-tomcat with a number of subdirectories. Like most software packages, the doc subdirectory contains a file named Readme ; do exactly that. Software distributions change and if, for instance, the instructions in this chapter no longer apply when you download the software, the Readme file should contain information about how to get started.

You also need to set the TOMCAT_HOME environment variable. For Windows, use:

C:\Jakarta> set TOMCAT_HOME=C:\Jakarta\jakarta-tomcat

For Unix, use:

[[email protected] /usr/local] export TOMCAT_HOME=/usr/local/jakarta-tomcat

The jakarta-tomcat directory contains a number of subdirectories:

bin

Scripts for starting the Tomcat server.

conf

Tomcat configuration files.

doc

Documents describing how to install and start Tomcat. Other documentation is available as web pages once the server is started.

lib

Binary (platform-dependent) modules for connecting Tomcat to other web servers such as Apache.

src

The source code for all servlet and JSP specification classes and interfaces.

webapps

Default location for web applications served by Tomcat.

No matter what your platform, the bin directory contains both Windows batch files and Unix scripts for starting and stopping the server.

4.2.1 Windows Platforms

The Windows files are named startup.bat, shutdown.bat, and tomcat.bat. The tomcat.bat file is the main script for controlling the server; it's called by the two other scripts startup.bat and shutdown.bat. To start the server in a separate window, change directory to the bin directory and run the startup.bat file:

C:\Jakarta> cd jakarta-tomcat\bin

C:\Jakarta\jakarta-tomcat\bin> startup

A new Command Prompt window pops up and you see startup messages like this:

2000-09-01 09:27:10 - ContextManager: Adding context Ctx( /examples )

2000-09-01 09:27:10 - ContextManager: Adding context Ctx( /admin )

Starting tomcat. Check logs/tomcat.log for error messages

2000-09-01 09:27:10 - ContextManager: Adding context Ctx( )

2000-09-01 09:27:10 - ContextManager: Adding context Ctx( /test )

2000-09-01 09:27:13 - PoolTcpConnector: Starting HttpConnectionHandler

on 8080

2000-09-01 09:27:13 - PoolTcpConnector: Starting Ajp12ConnectionHandler

on 8007

Just leave this window open; this is where the server process is running.

If you're running on a Windows 95 or 98 platform, you may see an error message about "Out of environment space" when you try to start the server. That's because the default amount of space allocated for environment variables is not enough. To change this default, run this command in the Command

Prompt window before you run the startup.bat file again:

C:\Jakarta\jakarta-tomcat\bin> COMMAND.COM /E:4096 /P

This command sets the environment space to 4096 bytes (4 KB). That should be enough for the server.

However, If you still get the same message, use a higher value.

For some installations, this command may not work. If it doesn't work, try this instead:

1. Close the Command Prompt window and open a new one.

2. Click on the MS-DOS icon at the top-left of the window.

3. Select the Properties option.

4. Click on the Memory tab.

5. Change the Initial Environment value from Auto to 4096.

6. Click on OK and try to start the server again.

At this point, the server may not start due to other problems. If so, the extra Command Prompt window may pop up and then disappear before you have a chance to read the error messages. If this happens, you can let the server run in the Command Prompt window with this command instead:

C:\Jakarta\jakarta-tomcat\bin> tomcat run

On Windows NT, first make sure that the Command Prompt window has a large enough screen buffer so that you can scroll back in case the error messages don't fit on one screen. Open the Properties window for the Command Prompt window (right mouse button in the upper-left corner), select Layout, and set the screen buffer size height to a large value (for instance 999). Unfortunately, the Command Prompt screen buffer cannot be enlarged for Windows 95/98, so scrolling back is not an option. If you run into problems on these platforms, double-check that you have installed the Java SDK correctly and that you have set the

JAVA_HOME

and PATH environment variables as described earlier.

4.2.2 Unix Platforms

For Unix, the corresponding scripts are named startup.sh, shutdown.sh, and tomcat.sh. Start the server with this command:

[[email protected] /usr/local/jakarta-tomcat/bin] ./startup.sh

If you want Tomcat to start each time you boot the system, you can add the following commands to your

/etc/rc.d/rc.local (or equivalent) startup script:

export JAVA_HOME=/usr/local/jdk1.2.2

export TOMCAT_HOME=/usr/local/jakarta-tomcat

$TOMCAT_HOME/bin/startup.sh &

Two more subdirectories under the Tomcat home directory are then created the first time you start the server:

logs work

Server log files. If something doesn't work as expected, look at the files in this directory for clues as to what's wrong.

A directory for temporary files that are created by the JSP container and other files. This directory is where the servlets generated from JSP pages are stored.

© 2001, O'Reilly & Associates, Inc.

4.3 Testing Tomcat

To test the server—assuming you're running Tomcat on the same machine as the browser and that you're using the default port for Tomcat (8080)—open a browser and enter the following URL in the

Location/Address field:

http://localhost:8080/

The Tomcat main page is shown in the browser (see

Figure 4.2

), and you can now run all servlet and

JSP examples bundled with Tomcat to make sure everything works.

Figure 4.2. The Tomcat main page

When you're done testing Tomcat, stop the server like this:

C:\Jakarta\jakarta-tomcat\bin> shutdown

You should always stop the server this way, as opposed to killing the Command Prompt window the server is running in. Otherwise, the applications don't get a chance to close down gracefully, and when you start to connect external resources, like a database, various problems may occur.

© 2001, O'Reilly & Associates, Inc.

4.4 Installing the Book Examples

All JSP pages, HTML pages, Java source code, and class files for the book examples can be downloaded directly from the O'Reilly web site: http://www.oreilly.com/catalog/jserverpages/

They can also be downloaded from the book web site: http://www.thejspbook.com/

The file that contains all the examples is called jspbook.zip. Save the file on your hard drive, for instance in C:\JSPBook on a Windows platform, and unpack it:

C:\JSPBook> jar xvf jspbook.zip

You can use the same command on a Unix platform.

Two new directories are created: ora and src. The first directory contains all examples described in this book, and the second contains the Java source files for the JavaBeans, custom actions, and utility classes used in the examples.

The examples' directory structure complies to the standard Java web application format described in

Chapter 2 . You can therefore configure any Servlet 2.2-compliant web container to run the examples. If

you like to use a container other than Tomcat, be sure to read the documentation for that container.

To install the example application for Tomcat, copy the web application directory structure to Tomcat's default directory for applications, called webapps. Use this command on a Windows platform:

C:\JSPBook> xcopy /s /i ora %TOMCAT_HOME%\webapps\ora

On a Unix platform it looks like this:

[[email protected] /usr/local/jspbook] cp -R ora $TOMCAT_HOME/webapps

Recall from Chapter 2 that each web application in a server is associated with a unique URI prefix.

When you install an application in Tomcat's webapps directory, the subdirectory name is automatically assigned as the URI prefix for the application ( /ora in this case).

At this point, you must shut down and restart the Tomcat server. After that, you can point your browser to the ora application with the following URL:

http://localhost:8080/ora/

You should see a start page, as in

Figure 4.3

, that contains links for all examples in this book.

Figure 4.3. JSP book examples start page

© 2001, O'Reilly & Associates, Inc.

4.5 Example Web Application Overview

The examples for this book are packaged as a standard Java web application, as described in Chapter 2 .

This file structure is supported by all Servlet 2.2-compliant servers, so you can use the example application as a guide when you create your own web applications. How a web application is installed is not defined by the specification, however, so it varies between servers. With Tomcat, you simply copy the file structure to the special webapps directory and restart the server. To modify the configuration information for an application, you need to edit the application's WEB-INF/web.xml file using a text editor. Other servers may offer special deployment tools that copy the files to where they belong and let you configure the application using a special tool, such as web-based forms.

If you look in the ora web application directory, you'll see that it contains an index.html file and a number of directories corresponding to chapters in this book. These directories contain all the example

JSP and HTML pages.

There's also a WEB-INF directory with a web.xml file, a lib directory, a classes directory, and a tlds directory:

The web.xml file contains configuration information for the example application in the format defined by the Servlet 2.2 specification. It's too early to look at the contents of this file now; we will return to parts of it when needed.

The lib and classes directories are standard directories, also defined by the Servlet 2.2 specification. A common question asked by people new to servlets and JSP (prior to the standard web application format) was, "Where do I store my class files so that the server can find them?" The answer, unfortunately, differed depending on which implementation was used.

With the standard web application format, however, it's easy to answer this question: if the classes are packaged in a JAR file, store the JAR file in the lib directory; otherwise, use the

classes directory (with subdirectories mirroring the classes' package structure). The server will always look for Java class files in these two directories.

The lib directory for the example application contains five JAR files. The orataglib_1_0.jar file contains all the Java class files for the custom actions and beans used in this book. The

jdbc20_stdext_classes.jar file contains classes that are part of the JDBC 2.0 Standard Extension and are used in the database examples. The xalan.jar, xerces.jar, and xsl.jar contain XML parser classes used for an example in

Chapter 12

.

The classes directory contains the class for a servlet used to display the raw source code for the example JSP pages, so you can see what they look like before they are processed by the server.

It also contains .properties files containing localized text for the example in

Chapter 11

.

The tlds directory is not defined by the Servlet 2.2 specification, but is the name used by convention for Tag Library Descriptor (TLD) files. Don't worry about what this means now. As you read through this book, it will become clear.

If you want to try out some of your own JSP pages, beans, and custom actions while reading this book, simply add the files to the example application structure: JSP pages in any directory except under WEB-

INF, and Java class files in either the classes or the lib directory, depending on if the classes are packaged in a JAR file or not. If you want to use the book's custom actions and beans in another application, copy the files in both the lib and tlds directories to the web application structure for the other application.

© 2001, O'Reilly & Associates, Inc.

5.1 What Time Is It?

Recall from Chapter 3 , that a JSP page is just a regular HTML page with a few special elements. JSP

pages should have the file extension .jsp , which tells the server that the page needs to be processed by the JSP container. Without this clue, the server is unable to distinguish a JSP page from any other type of file and sends it unprocessed to the browser.

When working with JSP pages, you really just need a regular text editor such as Notepad on Windows or

Emacs on Unix.

Appendix E , however, lists a number of tools that may make it easier for you, such as

syntax-aware editors that color-code JSP and HTML elements. Some Interactive Development

Environments (IDEs) include a small web container that allows you to easily execute and debug the page during development. There are also several web page authoring tools—the type of tools often used when developing regular HTML pages—that support JSP. I don't recommend that you use them initially; it's easier to learn how JSP works if you see the raw page elements before you use tools that hide them.

The first example JSP page, named date.jsp , is shown in

Example 5.1

.

Example 5.1. JSP Page Showing the Current Date and Time (date.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" />

The current time at the server is:

<ul>

<li>Date: <jsp:getProperty name="clock" property="date" />

<li>Month: <jsp:getProperty name="clock" property="month" />

<li>Year: <jsp:getProperty name="clock" property="year" />

<li>Hours: <jsp:getProperty name="clock" property="hours" />

<li>Minutes: <jsp:getProperty name="clock" property="minutes" />

</ul>

</body>

</html>

The date.jsp page displays the current date and time. We'll look at all the different pieces soon, but first let's run the example to see how it works. Assuming you have installed all book examples as described

in

Chapter 4

, first start the Tomcat server and load the http://localhost:8080/ora/ URL in a browser. You

can then run Example 5.1

by clicking the "Current Date/Time example" link from the book examples main page, shown in

Figure 5.1

. You should see a result like the one shown in

Figure 5.2

.

Figure 5.1. JSP book examples main page

Figure 5.2. Current Date/Time JSP page example

Notice that the month seems to be off by one and the year is displayed as 100. That's because the java.util.Date

class we use here numbers months from to 11, so January is 0, February is 1, and so on, and it reports year as the current year minus 1900. That's just the way this example works. As you will see later, there are much better ways to display dates.

The page shown in Example 5.1

contains both regular HTML elements and JSP elements. The HTML elements are used as-is, defining the layout of the page. If you use the View Source function in your browser, you notice that none of the JSP elements are visible in the page source. That's because the JSP elements are processed by the server when the page is requested, and only the resulting output is sent to the browser. To see the unprocessed JSP page in a separate window, click on the source link for the

date.jsp file in the book examples main page. The source link uses a special servlet to send the JSP page as-is to the browser instead of letting the server process it. This makes it easier for you to compare the source page and the processed result.

Let's look at each piece of Example 5.1

in detail.

5.1.1 Using JSP Directives

The first line in Example 5.1

is a JSP directive element. Directives are used to specify attributes of the page itself, primarily those that affect how the page is converted into a Java servlet. There are three JSP directives: page, include, and taglib. In this example, we're using only the page directive. We'll see the others later.

JSP pages typically start with a page directive that specifies the scripting language and the content type for the page:

<%@ page language="java" contentType="text/html" %>

A JSP directive element starts with a directive-start identifier (<%@ ) followed by the directive name

(e.g., page) and directive attributes, and ends with %>. A directive contains one or more attribute name/value pairs (e.g., language="java"). Note that JSP element and attribute names are casesensitive, and in most cases the same is true for attribute values. For instance, the language attribute value must be java, not Java. All attribute values must also be enclosed in single or double quotes.

The page directive has many possible attributes. In

Example 5.1

, two of them are used: language and contentType.

The language attribute specifies the scripting language used in the page. The JSP reference

implementation (the Tomcat server) supports only Java as a scripting language.

[1]

java is also the default value for the language attribute, but for clarity you may still want to specify it. Other JSP implementations support other languages besides Java, and hence allow other values for the language attribute. For instance, both JRun ( http://www.allaire.com/ ) and Resin ( http://www.caucho.com/ ) support JavaScript in addition to Java.

[1]

In fact, Java is the only scripting language formally supported in the JSP specification, but the specification leaves room for other languages to be supported.

The contentType attribute specifies the type of content the page produces. The most common values are text/html for HTML content and text/plain for preformatted, plain text. But you can also specify other types, such as text/xml for browsers that support XML or text/vnd.wap.wml for devices like cellular phones and PDAs that have built-in Wireless Markup Language (WML) browsers.

If the content generated by the page includes characters requiring a charset other than ISO-8859-1

(commonly known as Latin-1), you need to specify that charset with the contentType attribute. We'll look at the details of charsets in

Chapter 11 .

5.1.2 Using Template Text

Besides JSP elements, notice that the page shown in Example 5.1

contains mostly regular HTML:

...

<html>

<body bgcolor="white">

...

The current time at the server is:

<ul>

<li>Date: ...

<li>Month: ...

<li>Year: ...

<li>Hours: ...

<li>Minutes: ...

</ul>

</body>

</html>

In JSP parlance, this is called template text. Everything that's not a JSP element, such as a directive, action, or scripting element, is template text. Template text is sent to the browser as-is. This means you can use JSP to generate any type of text-based output, such as XML, WML, or even plain text. The JSP container doesn't care what the template text is.

5.1.3 Using JavaBeans

There is also some dynamic content in this example. Step back a moment and think about the type of dynamic content you see on the Web every day. Common examples might be a list of web sites matching a search criteria on a search engine site, the content of a shopping cart on an e-commerce site, a personalized news page, or messages on a bulletin board. Dynamic content is content generated by some server process, for instance the result of a database query. Before it is sent to the browser, the dynamic content needs to be combined with regular HTML elements into a page with the right layout, navigation bars, the company logo, and so forth. In a JSP page, the regular HTML is the template text described earlier. The result of the server processing—the dynamic content—is commonly represented by a JavaBeans component.

A JavaBeans component, or just a bean for short, is a Java class that follows certain coding conventions, so it can be used by tools as a component in a larger application. In this chapter, we discuss only how to use a bean, not how to develop one. (If you're a programmer and not already familiar with JavaBeans, you may want to skip ahead to

Chapter 15

, to learn about these coding conventions.) A bean is often used in JSP as the container for the dynamic content to be displayed by a web page. Typically, a bean represents something specific, such as a person, a product, or a shopping order. A bean is always created by a server process and given to the JSP page. The page then uses JSP elements to insert the bean's data into the HTML template text.

The type of element used to access a bean in a page is called a JSP action element. JSP action elements are executed when a JSP page is requested (this is called the request processing phase, as you may recall from

Chapter 3

). In other words, JSP actions represent dynamic actions that take place at runtime, as opposed to JSP directives, which are used only during the translation phase (when the JSP page is turned into Java servlet code). JSP defines a number of standard actions and also specifies how you can develop custom actions. For both standard and custom action elements, use the following notation:

<action_name attr1="value1" attr2="value2">

action_body

</action_name>

Action elements, or tags as they are sometimes called,

[2]

are grouped into libraries (known as tag

libraries). The action name is composed of two parts: a library prefix and the name of the action within the library, separated by a colon (i.e., jsp:useBean). All actions in the JSP standard library use the prefix jsp, while custom actions can use any prefix except jsp, jspx, java, javax, servlet, sun

, or sunw. You specify input to the action through attribute/value pairs in the opening tag. The attribute names are case-sensitive, and the values must be enclosed in single or double quotes. For some actions, you can also enter data that the action should process in the action's body. It can be any text value, such as a SQL statement, or even other nested JSP action elements. You will see examples of action elements with a body later.

[2]

An element is actually represented by a start tag and an end tag, but the term "tag" is often used to refer to what's formally known as an element.

Before you use a bean in a page, you must tell the JSP container which type of bean it is and associate it

with a name. The first JSP action in Example 5.1

, <jsp:useBean>, is used for this purpose:

<jsp:useBean id="clock" class="java.util.Date" />

The id attribute is used to give the bean a unique name. It must be a name that is a valid Java variable name: it must start with a letter and cannot contain special characters such as dots, plus signs, etc. The class

attribute contains the fully qualified name of the bean's Java class. Here, the name clock is associated with an instance of the class java.util.Date. Note that we don't specify a body for this action. When you omit the body, you must end the opening tag with />, as in this example. In this case, when the JSP container encounters this directive, there is no bean currently available with the name clock

, so the <jsp:useBean> action creates a bean as an instance of the specified class and makes it available to other actions in the same page. In

Chapter 8

, you will see how <jsp:useBean> can also be used to locate a bean that has already been created.

Incidentally, the <jsp:useBean> action supports three additional attributes: scope, type, and beanName

. The scope attribute is described in detail in Chapter 8 , and the other two attributes are

covered in

Appendix A . We don't need to worry about those attributes here.

5.1.4 Accessing JavaBean Properties

The bean's data is represented by its properties . If you're a page author charged with developing a JSP page to display the content represented by a bean, you first need to know the names of all the bean's properties. This information should be available from the Java programmers on the team or from a thirdparty source. In this example, we use a standard Java class named java.util.Date as a bean with

properties representing date and time information. Table 5.1

describes the properties used in this example. (If you're not a programmer, don't worry about the Java Type and Access columns at this

point.)

Table 5.1, Properties for java.util.Date

Property Name Java Type Access Description

date int read The day of the month as a number between 1 and 31 hours int read The hour as a number between 0 (midnight) and 23 minutes int read The number of minutes past the hour as a number between 0 and 59 month int read The month as a number from 0 to 11 year int read The current year minus 1900

Once you have created a bean and given it a name, you can retrieve the values of the bean's properties in the response page with another JSP standard action, <jsp:getProperty>. This action obtains the current value of a bean property and inserts it directly into the response body.

To include the current date property value in the page, use the following tag:

<jsp:getProperty name="clock" property="date" />

The name attribute, set to clock, refers to the specific bean instance we defined with the

<jsp:useBean>

action previously. This action locates the bean and asks it for the value of the

property specified by the property attribute. As documented in Table 5.1

, the date property contains

the day of the month as a number between 1 and 31. In Example 5.1

, multiple <jsp:getProperty> actions are used to generate a list of all the clock bean's property values. The result is the page shown in

Figure 5.2

.

© 2001, O'Reilly & Associates, Inc.

5.2 Input and Output

User input is a necessity in modern web pages. Most dynamic web sites generate pages based on user input. Unfortunately, users seldom enter information in exactly the format you need, so before you can use such input, you probably want to validate it.

And it's not only the input format that's important. Web browsers are also picky about the format of the

HTML you send them. For instance, when you generate an HTML form with values taken from a database, a name such as O'Reilly can cause problems. The single quote character after the O can fool the browser into believing that it's at the end of the string, so you end up with just an O in your form.

5.2.1 Using JavaBeans to Process Input

As we saw earlier, a bean is often used as a container for data, created by some server process, and used in a JSP page that displays the data. But a bean can also be used to capture user input. The captured data can then be processed by the bean itself or used as input to some other server component (e.g., a component that stores the data in a database or picks an appropriate banner ad to display). The nice thing about using a bean this way is that all information is in one bundle. Say you have a bean that can contain information about a person, and it captures the name, birth date, and email address as entered by the person on a web form. You can then pass this bean to another component, providing all the information about the user in one shot. Now, if you want to add more information about the user, you just add properties to the bean, instead of having to add parameters all over the place in your code. Another benefit of using a bean to capture user input is that the bean can encapsulate all the rules about its properties. Thus, a bean representing a person can make sure the birthDate property is set to a valid date.

Using a bean to capture and validate user input is one aspect of building a web application that's easy to maintain and extend as requirements change. But it's not the only option. If you're a page author and intend to use JSP to develop sites using components from third parties, you may wonder how you can capture and validate input without access to a Java programmer who can develop the beans. Don't

worry; we'll see another alternative in Chapter 9 .

Processing and validating input can also be performed by a servlet instead of a JSP page. If you're a

programmer, you'll find examples of this approach in Chapter 14

. In this part of the book, however, we use JSP pages for all aspects of the applications so we can focus on JSP features. And one JSP feature makes it very easy to capture user input, so let's see how it's done.

5.2.1.1 Setting JavaBeans properties from user input

In this next example, we capture information about web site users. It could be the frontend to a

newsletter subscription site, for instance. In order to send the users information that might interest them, we register the birth date, sex, and lucky number, along with the full name and email address, for each person that signs up for the service.

To capture and validate the user input, the example uses a bean named com.ora.jsp.beans.userinfo.UserInfoBean

, with the properties described in

Table 5.2

. If you're a programmer, you may want to skip ahead to peek at the source code for this bean class in

Chapter 15

.

Table 5.2, Properties for com.ora.jsp.beans.userinfo.UserInfoBean

Property Name Java Type Access Description

userName String read/write The user's full name birthDate String read/write The user's birth date in the format yyyy-mm-dd (e.g., 2000-07-07) emailAddr String read/write The user's email address in the format [email protected] sex String read/write The user's sex (male or female) luckyNumber String read/write The user's lucky number (between 1 and 100) valid boolean read true

if the current values of all properties are valid, false otherwise

As shown in the Access column, all properties except valid are read/write properties. This means that,

in addition to using the bean's properties to generate output (like in Example 5.1

), the property values can be set based on user input.

The HTML form shown in Example 5.2

allows the user to enter information corresponding to the bean

properties.

Example 5.2. An HTML Form that Sends User Input to a JSP Page (userinfo.html)

<html>

<head>

<title>User Info Entry Form</title>

</head>

<body bgcolor="white">

<form action="userinfo1.jsp" method="post">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName" >

</td>

</tr>

<tr>

<td>Birth Date:</td>

<td><input type="text" name="birthDate" >

</td>

<td>(Use format yyyy-mm-dd)</td>

</tr>

<tr>

<td>Email Address:</td>

<td><input type="text" name="emailAddr" >

</td>

<td>(Use format [email protected])</td>

</tr>

<tr>

<td>Sex:</td>

<td><input type="text" name="sex" >

</td>

<td>(Male or female)</td>

</tr>

<tr>

<td>Lucky number:</td>

<td><input type="text" name="luckyNumber" >

</td>

<td>(A number between 1 and 100)</td>

</tr>

<tr>

<td colspan=2><input type="submit"></td>

</tr>

</table>

</form>

</body>

</html>

This is a regular HTML page that presents a form with a number of fields, as shown in Figure 5.3

. There

are a few things worth mentioning here. First, notice that each input field has a name attribute with a value that corresponds to a UserInfoBean property name. Matching the names lets us take advantage of a nice JSP feature that sets property values automatically, as you'll see shortly. Also note that the action

attribute of the form specifies that a JSP page, userinfo1.jsp, is invoked when the user clicks

the Submit button.

Figure 5.3

shows what the form looks like in a browser.

Figure 5.3. User input form

Example 5.3

shows the JSP page that is invoked when the user submits the form.

Example 5.3. A JSP Page that Validates User Input with a Bean (userinfo1.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean

id="userInfo"

class="com.ora.jsp.beans.userinfo.UserInfoBean">

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

The following information was saved:

<ul>

<li>User Name: <jsp:getProperty

name="userInfo" property="userName" />

<li>Birth Date: <jsp:getProperty

name="userInfo" property="birthDate" />

<li>Email Address: <jsp:getProperty

name="userInfo" property="emailAddr" />

<li>Sex: <jsp:getProperty

name="userInfo" property="sex" />

<li>Lucky number: <jsp:getProperty

name="userInfo" property="luckyNumber" />

</ul>

The user input is valid: <jsp:getProperty

name="userInfo" property="valid" />

</body>

</html>

Almost at the top of

Example 5.3

, you see that a <jsp:useBean> action is used to associate a name

with the bean:

<jsp:useBean

id="userInfo"

class="com.ora.jsp.beans.userinfo.UserInfoBean">

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

The <jsp:useBean> action looks similar to the one in Example 5.1

. The id attribute specifies the name for the bean, and the class attribute specifies the full name of the bean class. But here we also use a <jsp:setProperty> action as the body of the <jsp:useBean> action. You must therefore use the complete closing tag (</jsp:useBean>) to tell the JSP container where the action ends, instead of the shorthand notation used in

Example 5.1

. The body of the <jsp:useBean> action is

executed only when a new bean is created. In this example, that's always the case, but as you will learn in

Chapter 8

, there are cases in which the bean already exists and the action is needed only to associate it with a name.

Now let's take a closer look at the <jsp:setProperty> action. As the name implies, this action is used to set the bean's property values. Like the <jsp:getProperty> action, it has a name attribute that must match the id attribute of a <jsp:useBean> action, and a property attribute that specifies which property to set.

When a form is submitted, the form field values are sent as request parameters with the same names as the form field elements. In

Example 5.3

, note that an asterisk (*) is used as the property attribute

value of the <jsp:setProperty> action. This means that all bean properties with names that match request parameters sent to the page are set automatically. That's why it's important that the form element names match the bean property names, as they do here. Automatically setting all matching properties is a great feature; if you define more properties for your bean, you can set them simply by adding new matching fields in the form that invokes the JSP page.

Besides the property attribute, the <jsp:setProperty> action has two more optional attributes: param

and value. If for some reason you can't use the same name for the parameters and the property names, you can use the param attribute to set a bean property to the value of any request parameter:

<jsp:setProperty

name="userInfo"

property="userName"

param="someOtherParam"

/>

Here, the userName property is set to the value of a request parameter named someOtherParam.

You can also explicitly set a bean property to a value that is not sent as a request parameter with the value

attribute:

<jsp:setProperty

name="userInfo"

property="luckyNumber"

value="13"

/>

Here, the luckyNumber property is set to the value 13. You typically use the value attribute only when you set the bean properties based on something other than user input, for instance values collected from a database.

5.2.1.2 Validating user input

Never trust your users, at least not when it comes to entering information in the format you need. Often, you need to make sure the input is valid before you continue to process a request. A date, for instance, can be written in many different formats. If you don't live in the United States, you probably have had to fill out both an I-94 and a customs declaration form to be admitted by an immigration officer. You may have noticed that on one of the forms you need to write your birth date as yyyy/mm/dd and on the other you write it as mm/dd/yyyy. I always get it wrong.

Four of the UserInfoBean's properties require a special format: birthDate, emailAddr, sex, and luckyNumber. A good place to make sure the input is valid is in the bean itself, which is exactly what the UserInfoBean does. With this bean, if you try to set any of the above properties to a value that isn't valid, the bean will leave the property unset. In addition, the bean has a true/false (Boolean) property named valid. This property has the value false unless all other properties have been set to valid values.

Let's see this in action.

Example 5.3

displays the property values using the <jsp:getProperty>

action:

<li>User Name: <jsp:getProperty

name="userInfo" property="userName" />

Since a property is set only if the value is valid, no values are shown for improperly specified properties.

Try it. Click on the "User Info 1 example" link under the

Chapter 5

header in the book examples main page shown in

Figure 5.1

. Enter both valid and invalid values in the form and look at the result produced

by the userinfo1.jsp page when you click Submit. A sample result is shown in Figure 5.4

.

Figure 5.4. Output from userinfo1.jsp

Note that the Birth Date information is missing (at my age, you're not so eager to reveal your birth date), so the input is marked as invalid.

5.2.2 Keep On Doing It 'til You Get It Right

Okay, now you know how to set bean properties and you're aware that beans often validate their values.

It would be nice if this technique could be used to display the same form over and over until all required input is correct. You can do that with just a few changes, as shown in

Example 5.4

, the userinfo2.jsp page.

Example 5.4. A JSP Page that Validates and Redisplays Until Correct (userinfo2.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>User Info Entry Form</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="userInfo"

class="com.ora.jsp.beans.userinfo.UserInfoBean">

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

<%-- Output list of values with invalid format, if any --%>

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

</font>

<%-- Output form with submitted valid values --%>

<form action="userinfo2.jsp" method="post">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<jsp:getProperty

name="userInfo"

property="userName"

/>">

</td>

</tr>

<tr>

<td>Birth Date:</td>

<td><input type="text" name="birthDate"

value="<jsp:getProperty

name="userInfo"

property="birthDate"

/>">

</td>

<td>(Use format yyyy-mm-dd)</td>

</tr>

<tr>

<td>Email Address:</td>

<td><input type="text" name="emailAddr"

value="<jsp:getProperty

name="userInfo"

property="emailAddr"

/>">

</td>

<td>(Use format [email protected])</td>

</tr>

<tr>

<td>Sex:</td>

<td><input type="text" name="sex"

value="<jsp:getProperty

name="userInfo"

property="sex"

/>">

</td>

<td>(Male or female)</td>

</tr>

<tr>

<td>Lucky number:</td>

<td><input type="text" name="luckyNumber"

value="<jsp:getProperty

name="userInfo"

property="luckyNumber"

/>">

</td>

<td>(A number between 1 and 100)</td>

</tr>

<tr>

<td colspan=2><input type="submit"></td>

</tr>

</table>

</form>

</body>

</html>

Instead of using a static HTML page for the input form and a separate JSP page with the validation code, in this example we have combined them into a single JSP page. This page generates the form and provides an appropriate message based on whether or not the input is valid. The page also fills in the form with the valid values that have already been specified (if any) so the user needs to enter values only for missing or incorrect input.

Let's look at Example 5.4

from the top. The first thing to note is that the page generates a message using

the UserInfoBean property named propertyStatusMsg. Here is the corresponding snippet:

<%-- Output list of values with invalid format, if any --%>

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

</font>

The first line here is a JSP comment. Text between <%-- and --%> in a JSP page is treated as a comment and never appears in the result sent to the browser. For complex pages, it's always a good idea to include comments to explain things that are not obvious.

The propertyStatusMsg property can have three different values. If none of the properties have been set, the value is "Please enter values in all fields". If at least one value is missing or invalid, the message states "The following values are missing or invalid" and provides a list of the relevant properties. Finally, if all the values are valid, the propertyStatusMsg is "Thanks for telling us about yourself!"

Next we generate the form, filled out with all valid values. Here's the beginning of the form and the code for the userName property:

<%-- Output form with submitted valid values --%>

<form action="userinfo2.jsp" method="post">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<jsp:getProperty

name="userInfo"

property="userName"

/>">

</td>

</tr>

Most of this is plain HTML, which is treated as template text and passed on untouched to the browser.

But note the use of a <jsp:getProperty> action as the HTML <input> element's value attribute. This is how the userName field in the form is filled in with the current value of the userName

bean property. Also note how the form's action attribute points back to the JSP page itself.

Try this out by clicking on the "User Info 2 example" link on the book examples page. Enter both valid and invalid values in the form and look at the results. In

Chapter 8

, we'll expand on this example and look at how you can move on to another page when all input is valid.

One item may look a bit strange to you: an element (<jsp:getProperty>) is used as the value of another element's attribute (the <input> tag's value attribute). While this is not valid HTML syntax, it is valid JSP syntax. Remember that everything not recognized as a JSP element is treated as template text. Whether the template text is HTML, XML, WML, or just plain text doesn't matter. As far as the

JSP container is concerned, the previous code is as valid as: any old template text <jsp:getProperty

name="userInfo"

property="userName" /> more text

When the JSP page is processed, the action element is replaced with the value of the bean's property.

The resulting HTML sent to the browser is therefore valid.

5.2.3 Formatting HTML Output

If you enter a value containing double quotes in the Name field of the userinfo2.jsp page, it doesn't work right. For example, try "Prince, "the artist"" and you'll see what I mean. Only "Prince," appears in the

Name field, and the Birth Date field is not shown at all. What's going on here?

A look at the HTML code generated by the JSP page using your browser's View Source function reveals what's wrong:

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="Prince, "the artist"">

</td>

</tr>

In the JSP file, double quotes are used to enclose the value of the <input> element's value attribute, so when the value itself includes a double quote, the browser gets confused. The first double quote in the value is interpreted as the end of the value. That's why you see only "Prince," in the field. Even worse, the rest of the value interferes with the interpretation of the rest of the form, causing the next input field to be ignored in most browsers.

One solution to this problem would be to use single quotes around the values instead, since HTML accepts either single quotes or double quotes. But then you would have the same problem if the user enters a value that includes a single quote. Fortunately, there's a better way.

What's needed is special treatment of all characters that can cause HTML interpretation problems when we generate HTML from dynamic strings. One way to handle this is to let the bean take care of the special treatment. The UserInfoBean can do this through another set of properties: userNameFormatted

, birthDateFormatted, emailAddrFormatted, sexFormatted, and luckyNumberFormatted.

These are read-only properties that simply represent formatted versions of the corresponding real property values. The bean is designed so that when you use these property names, all troublesome characters in the real property values—such as single quotes, double quotes, less-than symbols, greaterthan symbols, and ampersands—are converted to their corresponding HTML character entities (i.e.,

&#39;

, &#34;, &lt;, &gt;, and &amp;). The browser handles the converted values with no

problem. If you're curious about the Java code for the formatted properties, it's described in

Chapter 15 .

Example 5.5

shows a JSP page that uses the new properties.

Example 5.5. A JSP Page with Validation and Formatting Using a Bean (userinfo3.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>User Info Entry Form</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="userInfo"

class="com.ora.jsp.beans.userinfo.UserInfoBean">

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

<%-- Output list of values with invalid format, if any --%>

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

</font>

<%-- Output form with submitted valid values --%>

<form action="userinfo2.jsp" method="post">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<jsp:getProperty

name="userInfo"

property="userNameFormatted"

/>">

</td>

</tr>

<tr>

<td>Birth Date:</td>

<td><input type="text" name="birthDate"

value="<jsp:getProperty

name="userInfo"

property="birthDateFormatted"

/>">

</td>

<td>(Use format yyyy-mm-dd)</td>

</tr>

<tr>

<td>Email Address:</td>

<td><input type="text" name="emailAddr"

value="<jsp:getProperty

name="userInfo"

property="emailAddrFormatted"

/>">

</td>

<td>(Use format [email protected])</td>

</tr>

<tr>

<td>Sex:</td>

<td><input type="text" name="sex"

value="<jsp:getProperty

name="userInfo"

property="sexFormatted"

/>">

</td>

<td>(Male or female)</td>

</tr>

<tr>

<td>Lucky number:</td>

<td><input type="text" name="luckyNumber"

value="<jsp:getProperty

name="userInfo"

property="luckyNumberFormatted"

/>">

</td>

<td>(A number between 1 and 100)</td>

</tr>

<tr>

<td colspan=2><input type="submit"></td>

</tr>

</table>

</form>

</body>

</html>

It's not always a good idea to have a bean handle this type of formatting, though. A bean is easier to reuse if it doesn't contain logic that is specific for one type of use, such as generating strings suitable for

HTML. When we look at scripting elements and custom actions, we will revisit the subject of HTML formatting and look at other solutions to this problem.

Try the final version of this example by clicking on the "User Info 3 example" link. Now everything works fine, even if you happen to be Prince, "the artist."

© 2001, O'Reilly & Associates, Inc.

6.1 Java Primer

You don't have to be a Java expert to develop JSP pages, but it helps to have an understanding of the basic concepts. This overview of the Java language and some of the standard classes should be enough to get you started.

6.1.1 Classes and Objects

Java is an object-oriented language. This means that everything in Java is an object, except for a few primitive types such as numbers, characters, and Boolean values. An object is an instance of a class, which serves as a source code template describing how the object should behave. It's helpful to think of

a class as a blueprint from which identical copies (objects) are created. Example 6.1

shows a simple

Java class.

Example 6.1. Simple Java Class

/**

* This is just a simple example of a Java class

* with two instance variables, a constructor, and

* some methods.

*/ public class Rectangle {

// Data

private int width;

private int height;

// Constructor

public Rectangle(int w, int h) {

width = w;

height = h;

}

// Methods

public int getWidth( ) {

return width;

}

public void setWidth(int w) {

width = w;

}

public int getHeight( ) {

return height;

}

public void setHeight(int h) {

height = h;

}

public double getArea( ) {

double area;

area = width * height;

return area;

}

}

It's important to remember that a class always defines two items:

data : a collection of information in an object

methods : a set of functions that act on that data

6.1.1.1 Data

Data, often called variables, can consist of primitive datatypes such as integers, Booleans, and floatingpoint values (both the width and height in this example are integers, represented by the keyword int).

In addition, data can also be objects. The type value that a variable holds must always be declared. The following example declares a variable of the object type String:

String title;

Until you give a variable a value, it contains a default value (0, false, or null). The name of the variable must start with a letter followed by a combination of letters and digits. There are many special characters, such as dots and plus signs, that are not allowed in a variable name. By convention, variable names often start with lowercase letters and do not have spaces:

String titleOfBook;

6.1.1.2 Methods

Methods are functions that take in zero or more primitive datatypes or objects, and perform some task on the object that may or may not result in a return value. If it does, the return value is also a primitive datatype or object. Here is an example of a method: public void setHeight(int h) {

height = h;

}

This method, called setHeight( ), takes in a single integer, uses it to set the object variable height

, and returns nothing (note the void keyword before the method name).

One special method that appears in Example 6.1

is the constructor. The constructor method always shares the same name as the class, and its return type is never declared. The constructor allows the object to initialize itself; it is invoked when the new keyword is used to create an instance of the class:

Rectangle rect1 = new Rectangle(28,72);

Here we create an instance of the class Rectangle and keep a reference to it in a variable called rect1

. The new Rectangle object saves the value of the two integer constructor arguments, 28 and

72, in its internal variables, width and height. Note that Java is a case-sensitive language: height and Height are not the same. The standard naming convention for class names, unlike for variable names, is to capitalize the first letter. For both class and variable names, the first letter in internal words is also capitalized; for instance: aVeryLongNameForAVariable // VARIABLE

ANameForAClass // CLASS

6.1.1.3 Statements

A statement is simply an instruction to do something. For example, the following are statements in

Java: area = width * height; return area;

The first statement takes the value of width and height, multiplies them together, and places the result in the variable area. The second statement uses the variable area as the return value for the current method. Statements almost always appear inside of methods. In addition, all statements and variable declarations must end with a semicolon (;) in Java; this takes after other programming languages, such as C and C++.

6.1.1.4 Inheritance

Java includes a number of standard classes. For a specific application, you can create your own classes, often based on the standard Java classes. Classes can be arranged in a hierarchy, where one class extends the functionality of another class. This is one of the fundamental attributes of an objectoriented language, called inheritance. A class that extends another class, usually called a subclass, inherits methods and variables from the class it extends, usually called the superclass. A subclass can add its own methods, or override the existing methods defined in the superclass by creating identical methods in its own class. A typical example is a class Vehicle with subclasses Car and Boat. The

Vehicle

class has a method isLandBound( ), returning true. The Car class uses the method as-

is, but the Boat class overrides it to return false. This is illustrated in Figure 6.1

.

Figure 6.1. Class inheritance

Inheritance lets you write code that works with objects on different levels of abstraction. Let's say the

Car

class adds a method getDoors( ), returning the number of doors for a specific instance of the class. Code that works with Vehicle objects can call the isLandBound( ) method and be ignorant about if an object is a Car or a Boat, while code dealing exclusively with Car objects can call both the isLandBound( ) and getDoors( ) methods: it inherits the former method from

Vehicle

.

6.1.2 Primitive Types

As we mentioned earlier, Java includes a number of primitive types for efficiency. These primitive types are listed in the following chart:

Type Size Values

int short long byte float double char boolean

4 bytes -2,147,483,648 to 2,147,483,647

2 bytes -32,768 to 32,767

8 bytes -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807

1 byte -128 to 127

4 bytes 32-bit IEEE 754 floating-point numbers

8 bytes 64-bit IEEE 754 floating-point numbers

2 bytes Unicode characters

1 bit true

or false

6.1.3 Comments

Java supports two types of comments ( Example 6.1

includes both types).

A comment that spans multiple lines, known as a block comment , starts with the characters /* and ends with the characters */:

/**

* This is just a simple example of a Java class

* with two instance variables, a constructor, and

* some methods.

*/

Another type of comment is called an end-of-line comment. It starts with two slashes (//) and ends, as the name implies, at the end of the same line:

// Instance variable

6.1.4 Standard Classes Commonly Used in JSP Pages

The Java 2 libraries contain more than 1,500 classes. Don't worry, you don't have to learn about all of them to use JSP. The most commonly used standard classes are introduced here. Other classes that you need on occasion will be introduced when they are used in the examples in this book. If you would like

more information, the full documentation for the Java 2 Standard Edition classes is available online at http://java.sun.com/j2se/1.3/docs.html

.

6.1.4.1 String and StringBuffer

The String class represents an immutable string (a series of letters or numerical characters); its value can never be changed. A StringBuffer , on the other hand, is intended to be used when you build a string dynamically. Note that with the Java language, you can create a String object without using the new keyword; instead, just place the contents of the String in quotation marks.

For example, the following are equivalent:

String str1 = new String("The bright day");

String str1 = "The bright day";

String

is the only object in Java where you can abbreviate its creation this way.

Some of the most commonly used String methods are: public boolean equals(Object anObject)

Returns true if the target and the argument represent exactly the same sequence of characters.

For example:

String hello = new String("hello"); hello.equals("hello"); // Returns true hello.equals("Hello"); // Returns false public boolean equalsIgnoreCase(String anotherString)

Returns true if the target and the argument represent exactly the same sequence of characters, ignoring case. For example:

String hello = new String("hello"); hello.equals("hello"); // Returns true hello.equals("Hello"); // Returns true public String trim( )

Returns a new String with all whitespace from both ends of the target String removed. For

example:

String hello = " Hello ";

String trimmed = hello.trim( ); trimmed.equals("Hello"); // Returns true

Some of the most commonly used StringBuffer methods are: public synchronized StringBuffer append(String str)

Appends the argument to the target. For example:

StringBuffer buffer = new StringBuffer("Hello"); buffer.append(" JSP!"); // New value is "Hello JSP!" public String toString( )

Returns a String representing the data in the target. For example:

StringBuffer buffer = new StringBuffer("Hello"); buffer.append(" JSP!"); // New value is "Hello JSP!"

String hello = buffer.toString( ); hello.equals("Hello JSP!"); // Returns true

6.1.4.2 Arrays

An array is a bundling of a number of variables of the same type (a class or a primitive type). In Java, arrays are also objects. You create a new array like this:

String[] myStrings = new String[4]; int[] myInts = new int[7];

The type of component in the array is the type specified before the empty brackets on the left side of the equals sign (=). The size of an array is fixed and must be specified within the brackets of the type definition following the new keyword.

You can access the individual components of an array by specifying the index of the component. The indexes always begin from and count upward to the defined length minus one: int i = myInts[0];

The number of components in an array is available in a variable of the array object named length.

You can access it like this: int myIntsLength = myInts.length;

6.1.5 Flow Control Statements

Flow control statements are used for situations such as testing conditions and looping. Let's briefly look at some examples.

To execute a piece of code only when a condition is true, use the if statement: if ( hello.equals("Hello") ) {

// Do something

} else {

// Do something else

}

The code within parentheses following the if keyword is called an expression : it must be evaluated by the Java interpreter to some value, in this case a Boolean true or false value. Here we test if the

String

variable hello equals the string literal "Hello". If it evaluates to true, the block of code between the first set of braces is executed. You can optionally use an else clause, like we do in this example, which is executed if it evaluates to false.

In some situations, you can use the conditional operator (?) as an alternative to the if statement:

String greeting = (clock.getHours( ) < 12) ?

"Good morning" : "Good afternoon";

Here's how it works: first, the code before the ? operator is evaluated. If the result is true, the value immediately after the ? is returned. If it's false, the value after the colon is returned. This type of expression is more compact than an if statement when the only thing you want is to get one of two values depending on a simple condition.

Other common flow control statements are while and for. They are used to execute a block of code repeatedly. Let's look at the while statement first: while (i < myArr.length) {

// Do something

i = i + 1;

}

Here, the expression following the while keyword is evaluated. If it is true, the code block within the braces is executed. After it completes, the expression is evaluated again. If it is still true, the code block is again executed. This looping repeats until the expression evalutes to false.

Here is an example of a for statement: for (int i = 0; i < myArr.length; i++) {

// Do something

}

The for statement has three expressions within the parentheses following the for keyword. The first expression is executed before anything else; it is typically used to initialize a variable used as an index.

The expression in the middle must be a Boolean expression. If it evaluates to true, the block of code within braces is executed. After each code block iteration, the final statement is executed and the middle expression is evaluated again. If it's still true, the body is again executed and the third expression is tested, and so on. Note that the final expression is typically used to increment an index variable until the expression in the middle evaluates to false, at which point the loop will exit.

6.1.6 Operators

Operators are used to assign values to variables, perform numeric operations, and compare values, among other things. Java uses the usual operators for numeric addition, subtraction, multiplication, and division: +, -, *, and /. There are also shortcuts for some common expressions; for instance: i += 3; is the same as: i = i + 3; and: i++; is the same as: i = i + 1;

The + operator can also be used to concatenate strings:

String hello = new String("Hello");

String helloWorld = hello + " World!";

This is not the most efficient way of concatenating strings though, so don't overuse it. Where performance is critical and you need to concatenate many strings together, it's more efficient to use a

StringBuffer

:

StringBuffer helloWorld = new StringBuffer("Hello"); helloWorld.append(" World!");

The operators for equal (==), not equal (!=), greater than (>), greater than or equal (>=), less than (<), and less than or equal (<=) are the same as in many other languages.

if (anInt > 2) {

// Do something if an int variable value is greater than 2

}

Note that you cannot use the == operator to test if two String objects have the same value. That's because the == operator, when used with objects, tests if the variables are referencing the same

String

object, not if the values of two String objects are the same. (== works only when comparing the values of two primitive datatypes.) To compare objects with one another you must instead use the equals( ) method described earlier: if ( oneString.equals(anotherString) ) {

// Do something if the two String values are equal

}

Finally, you can combine comparative expressions using the Boolean operators && (and) and || (or): if ( (i >= 4) && (i <= 10) ) {

// Do something if i is between 4 and 10

}

That's it for our brief introduction to programming in the Java language. You should now know enough

Java programming to follow the remaining examples in this part of the book.

© 2001, O'Reilly & Associates, Inc.

6.2 Implicit JSP Objects

When you use scripting elements in a JSP page, you always have access to a number of objects (listed

in Table 6.1

) that the JSP container makes available. These are called implicit objects. These objects are instances of classes defined by the servlet and JSP specifications.

Appendix B , contains complete

descriptions of all methods for each class, and we will cover them in more detail as we move through the book. However, I want to briefly introduce them here, as they are used in a number of examples throughout this book.

Variable Name Java Type

Table 6.1, Implicit JSP Objects request javax.servlet.http.HttpServletRequest

response pageContext session application out config page javax.servlet.http.HttpServletResponse

javax.servlet.jsp.PageContext

javax.servlet.http.HttpSession

javax.servlet.ServletContext

javax.servlet.jsp.JspWriter

javax.servlet.ServletConfig

java.lang.Object

exception java.lang.Throwable

Here is some more information about each of these implicit objects: request

The request object is an instance of the class named javax.servlet.http.HttpServletRequest

. This object provides methods that let

you access all the information that's available about the current request, such as request parameters, attributes, headers, and cookies. We use the request object in a couple of examples later in this chapter.

response

The response object represents the current response message. It's an instance of the javax.servlet.http.HttpServletResponse

class, with methods for setting headers and the status code and for adding cookies. It also provides methods related to session tracking.

These methods are the response methods you're most likely to use. We'll look at them in detail in

Chapter 8

.

session

The session object allows you to access the client's session data, managed by the server. It's an instance of the javax.servlet.http.HttpSession class. Typically you do not need to directly access this object, since JSP also lets you access the session data through action elements, as you will see in

Chapter 8 . One method you may use, however, is invalidate(

)

, which explicitly terminates a session. An example of this is shown in Chapter 10 .

application

The application object is another object that you typically access indirectly through action elements. It's an instance of the javax.servlet.ServletContext class. This object is used to hold references to other objects that more than one user may require access to, such as a database connection shared by all application users. It also contains log( ) methods that you can use to write messages to the container's log file, as you will see in an example later in this chapter.

out

The out object is an instance of javax.servlet.jsp.JspWriter. You can use the print( )

and println( ) methods provided by this object to add text to the response message body. We look at an example of this later in this chapter. In most cases, however, you will just use template text and JSP action elements instead of explicitly printing to the out object.

exception

The exception object is available only in error pages and contains information about a

runtime error. Chapter 7 , describes in more detail how you can use this object.

The remaining three implicit objects (pageContext , config, and page) are so rarely used in scripting elements that we will not discuss them here. If you're interested, you can read about them in

Appendix B .

All variable names listed in

Table 6.1

are reserved for the implicit objects. If you declare your own variables in a JSP page, as you will soon see how to do, you must not use these reserved names for other variables.

© 2001, O'Reilly & Associates, Inc.

6.3 Conditional Processing

In most web applications, you produce different output based on runtime conditions, such as the state of a bean or the value of a request header such as UserAgent (containing information about the type of client that is accessing the page).

If the differences are not too great, you can use JSP scripting elements to control which parts of the JSP page are sent to the browser, generating alternative outputs from the same JSP page. However, if the outputs are completely different, I recommend using a separate JSP page for each alternative and passing control from one page to another. This chapter contains a number of examples in which one page is used. In the remainder of this book you'll see plenty of examples where multiple pages are used instead.

6.3.1 Using JavaBeans Properties

In Chapter 5

, you saw how to use the <jsp:getProperty> and the <jsp:setProperty> actions to access a bean's properties. However, a bean is just a Java class that follows certain coding conventions, so you can also call its methods directly.

Briefly, a bean is a class with a constructor that doesn't take an argument. This makes it possible for a tool, such as the JSP container, to create an instance of the bean class simply by knowing the class name. The other condition of a bean that we are concerned with is the naming of the methods used to access its properties. The method names for reading and writing a property value, collectively known as the bean's accessor methods, must be composed of the keywords get and set, respectively, plus the name of the property. For instance, you can retrieve the value of a property named month in a bean with the method getMonth( ) and set it with the method setMonth( ). Individually, the accessor method for reading a property value is known as the getter method, and the accessor method for writing a property value is the setter method. A property can be read-only, write-only, or read/write depending on whether a getter method, a setter method, or both methods are provided in the class. The Java type for a property, finally, is the type returned by the getter method and the type of the setter methods argument.

To use a bean's property value in a scripting element, call the accessor method directly. To illustrate

this, let's use one of the properties of the java.util.Date class introduced in Chapter 5 .

Table 6.2

shows a bean property sheet for the hours property. (These tables should be getting clearer now, by the way.) It's a read-only property of the type int.

Table 6.2, java.util.Date hours Property

Property Name

hours

Java Type Access Description

int read The hour as a number between 0 (midnight) and 23

Example 6.2

revisits an example from the first chapter: this page uses the value of this property to greet

the user with an appropriate message depending on the time of day.

Example 6.2. Conditional Greeting Page (greeting.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" />

<% if (clock.getHours( ) < 12) { %>

Good morning!

<% } else if (clock.getHours( ) < 17) { %>

Good day!

<% } else { %>

Good evening!

<% } %>

</body>

</html>

As we discussed before, this page will show a different message depending on when you request it.

What we didn't elaborate on in the first chapter, however, is that this magic is accomplished using a set of JSP scriptlet elements. A scriptlet is a block of code enclosed between a scriptlet-start identifier, <%, and an end identifier, %>.

Let's look in detail at Example 6.2

. The <jsp:useBean> action is first used to create a bean. Besides making the bean available to other actions, such as <jsp:getProperty> and

<jsp:setProperty>

, the <jsp:useBean> action also creates a Java variable that holds a reference to the bean. The name of the variable is the name specified by the id attribute, in this case clock

.

The clock bean is then used in four scriptlets, together forming a complete Java if statement with template text in the if and else blocks:

<% if (clock.getHours( ) < 12) { %>

An if statement, testing if it's before noon, with a block start brace.

<% } else if (clock.getHours( ) < 17) { %>

The if block end brace and an else-if statement, testing if it's before 5:00 P.M., with its block start brace.

<% } else { %>

The else-if block end brace, and a final else block start brace, handling the case when it's after 5:00 P.M.

<% } %>

The final else block end brace.

The JSP container combines the code segment in the four scriptlets with code for writing the template text to the response body. The end result is that when the first if statement is true, "Good morning!" is displayed; when the second if statement is true, "Good day!" is displayed; and if none of the if statements is true, the final else block is used, displaying, "Good evening!"

The tricky part when using scriptlets like this is making sure that all the start and end braces are in place. If you miss even one brace, the code that the JSP container generates is not syntactically correct.

And unfortunately, the error message you get is not always easy to interpret.

6.3.2 Using Request Information

Let's look at another example, in which the implicit object request is used to display different

messages depending on whether the Internet Explorer or Netscape Navigator browser is used. Example

6.3

shows the complete page.

Example 6.3. Browser-Dependent Page (browser.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<% if (request.getHeader("User-Agent").indexOf("MSIE") != -1) { %>

You're using Internet Explorer.

<%

} else

if (request.getHeader("User-Agent").indexOf("Mozilla") != 1) {

%>

You're using Netscape.

<% } else { %>

You're using a browser I don't know about.

<% } %>

</body>

</html>

As in

Example 6.2

, four scriptlets are used to provide the code for the conditional message. The

difference is the actual if statements used to figure out which type of browser is requesting the page.

Let's look at all the objects and methods used here.

The request object is not a bean, since it doesn't follow all the JavaBeans conventions described above, but it does provide a number of methods you can use to get information about the request. For instance, the request object's getHeader( ) method is used to get the value of a specific request header. Recall that each request contains a number of headers that provide detailed information the

server may use to process the request. In this case, the User-Agent header retrieved in Example 6.2

contains a description of the browser making the request.

The getHeader( ) method returns a String object, so we can then use the indexOf( ) method of the String class to look for the piece of a string that identifies the browser. This method returns an integer offset from the beginning of the string if the text is found, or -1 if the text is not found. Note that both Internet Explorer and Netscape send a description that contains the text "Mozilla", but

Internet Explorer also includes the text "MSIE", so we must look for "MSIE" first. If we find it, we know it's an Internet Explorer browser. If we don't find "MSIE" but find "Mozilla", it's most likely a

Netscape browser. The final else block takes care of other browsers, such as Lynx and Opera.

6.3.3 Working with Arrays

Another common use of scriptlets is to loop over an array. In

Example 6.4

, we let the user pick a

number of items from a group of checkboxes, and then use scriptlets to display all the choices.

Example 6.4. Looping Over Parameter Array (loop.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<form action="loop.jsp">

<input type="checkbox" name="fruits" value="Apple">Apple<br>

<input type="checkbox" name="fruits" value="Banana">Banana<br>

<input type="checkbox" name="fruits" value="Orange">Orange<br>

<input type="submit" value="Enter">

</form>

<%

String[] picked = request.getParameterValues("fruits");

if (picked != null && picked.length != 0) {

%>

You picked the following fruits:

<ul>

<%

for (int i = 0; i < picked.length; i++) {

out.println("<li>" + picked[i]);

}

%>

</ul>

<% } %>

</body>

</html>

Figure 6.2

shows an example of the resulting page.

Figure 6.2. Page showing values of multivalue parameter

The first part of the page in

Example 6.4

is a regular HTML form, with three checkboxes and a Submit

button. Again, note that the action of the form is to return to the same page, loop.jsp, but this time with parameter values. Then follows three scriptlets. In the first, we use another request method, getParameter-Values ( )

:

<%

String[] picked = request.getParameterValues("fruits");

if (picked != null && picked.length != 0) {

%>

This method returns a String[] containing all values for the parameter specified as the argument.

Here, it's the name used for all checkboxes in the form. If no checkbox is selected, this parameter has no value; getParameterValues( ) returns null. So before we try to loop through the array, an if

statement is used to verify that we in fact received at least one value.

The next scriptlet contains the actual loop:

<%

for (int i = 0; i < picked.length; i++) {

out.println("<li>" + picked[i]);

}

}

%>

A for statement is used to process each array component. Here you can also see how the implicit out object is used to add content to the response body.

Again, make sure the braces in the three scriptlets form complete code blocks when they are combined with the template text by the JSP container. Note how the start brace for the if block is included in the first scriptlet, and the end brace is in the third scriptlet. These braces must balance, or the JSP server will return an error.

© 2001, O'Reilly & Associates, Inc.

6.4 Displaying Values

Besides using scriptlets for conditional output, one more way to employ scripting elements is by using a

JSP expression element to insert values into the response. A JSP expression element can be used instead of the <jsp:getProperty> action in some places, but it is also useful to insert the value of any Java expression that can be treated as a String.

An expression starts with <%= and ends with %>. Note that the only syntax difference compared to a scriptlet is the equals sign (=) in the start identifier. An example is:

<%= userInfo.getUserName( ) %>

The result of the expression is written to the response body. One thing is important to note: as opposed to statements in a scriptlet, the code in an expression must not end with a semicolon. This is because the JSP container combines the expression code with code for writing the result to the response body. If the expression ends with a semicolon, the combined code will not be syntactically correct.

In the final example of Chapter 5

, we used <jsp:getProperty> actions to fill out the form fields with UserInfoBean values. To recap, it looked like this:

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<jsp:getProperty

name="userInfo"

property="userNameFormatted"

/>">

</td>

</tr>

In this case, the <jsp:getProperty> syntax is distracting, since it's used as the value of an HTML element. You can use expressions to make the user input form page easier to read. The following example shows the same part of the page with the <jsp:getProperty> action replaced with an expression:

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<%= userInfo.getUserNameFormatted( ) %>" >

</td>

</tr>

The result is exactly the same, but this is more compact.

Expressions help you write more compact code, but they can also help you with something even more important. The UserInfoBean provides a set of properties with values formatted for HTML output;

that's what we used in Chapter 5 to avoid confusing the browser with special characters in the bean

property values. However, it's much easier to reuse a bean if it doesn't need to format its property values for a certain type of output. With expressions, we can let the bean be ignorant of how its property values

are used, and use a utility class to do the formatting instead, as in Example 6.5

.

Example 6.5. Formatting HTML Output (userinfo4.jsp)

<%@ page import="com.ora.jsp.util.*" %>

...

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<%= StringFormat.toHTMLString(userInfo.getUserName( )) %>" >

</td>

</tr>

This example shows yet another version of the form we first used in

Chapter 5 . Here a utility class called

com.ora.jsp.util.StringFormat

is used to handle special characters in the property values. I use the term utility class for a class that doesn't represent an entity such as a customer, order, or product.

Instead, it's just a collection of useful methods. In this case, the StringFormat class simply contains

methods for formatting strings. All of its methods are described in Appendix C .

The StringFormat class has a method called toHTMLString( ). It formats its argument for

HTML output the same way as the UserInfoBean does, converting all HTML special characters to the corresponding HTML character entities. Here we pass it the unformatted property value by calling the bean's regular property getter method. Using this utility class, the ties between the UserInfoBean and

HTML can be removed and the formatting can be done where it belongs.

6.4.1 Packages

You may have noticed that the full name of the class in Example 6.5

is

com.ora.jsp.util.StringFormat

, but in the expression it's simply referred to as

StringFormat

. This requires an explanation.

A large application may use many different classes, some of them part of the standard Java libraries, and others developed in-house or by third parties. To organize all these classes, Java provides the notion of a

package. A package is a group of related classes. The fully qualified name of a class is the combination of the package name and the class name. For instance, the fully qualified name of the class used in

Example

6.5

is com.ora.jsp.util.StringFormat. You can always use the fully qualified name in your

Java code, but to save you some typing, you can also import a package and then refer to the class with just

the short class name. If you look at the top of Example 6.5

, you see a page directive with the import

attribute set to the name of the package the StringFormat class belongs to:

<%@ page import="com.ora.jsp.util.*" %>

Importing a package doesn't mean that it's physically included in the page. It only tells Java to look for classes with short names in the named package. You can use multiple page directives with import attributes in the same page, or use one with a comma-separated list of import declarations, if you need to import more than one package. In other words, this directive:

<%@ page import="java.util.*, com.ora.jsp.util.*" %> has the same effect as these two directives:

<%@ page import="java.util.* " %>

<%@ page import="com.ora.jsp.util.*" %>

6.4.2 Checking Off Checkboxes Dynamically

In Example 6.4

, a for statement is used to loop through an array, but arrays can also be used in many

other ways. If the array represents choices the user can make at one time and change at a later time, a form for changing the information can contain a set of checkboxes with the current choices checked off.

An application like this typically gets the current choices from a database, and you will see an example of this in

Chapter 10

. To demonstrate a technique for dynamically checking off checkboxes in a form, however, we keep it simple and use the String[] with fruit choices from

Example 6.4

.

Example 6.6

is a modified version of the loop example.

Example 6.6. Setting Checkbox Values Dynamically (checkbox.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="com.ora.jsp.util.*" %>

<html>

<body bgcolor="white">

<form action="checkbox.jsp">

<input type="checkbox" name="fruits" value="Apple">Apple<br>

<input type="checkbox" name="fruits" value="Banana">Banana<br>

<input type="checkbox" name="fruits" value="Orange">Orange<br>

<input type="submit" value="Enter">

</form>

<%

String[] picked = request.getParameterValues("fruits");

if (picked != null && picked.length != 0) {

%>

You picked the following fruits:

<form>

<input type="checkbox" name="fruits" value="Apple"

<%= ArraySupport.contains(picked, "Apple") ?

"checked" : "" %> >Apple<br>

<input type="checkbox" name="fruits" value="Banana"

<%= ArraySupport.contains(picked, "Banana") ?

"checked" : "" %> >Banana<br>

<input type="checkbox" name="fruits" value="Orange"

<%= ArraySupport.contains(picked, "Orange") ?

"checked" : "" %> >Orange<br>

</form>

<% } %>

</body>

</html>

Here, the loop is replaced with a second form with the same checkboxes as in the form at the top of the page. A browser shows a checkbox as checked off if the HTML <input> element includes the checked

keyword. The trick, then, is to add that keyword for all checkboxes in the second form that the user had checked off in the first form.

The result of this example is shown in

Figure 6.3

.

Figure 6.3. Setting checkboxes dynamically

Another utility class from the com.ora.jsp.util package, ArraySupport, has the method contains( )

. This method takes two arguments: a String[] (string array) and the String that you want to test to see if it's a component in that array. If the second argument is a component in the array, the method returns true.

The contains( ) method is used in JSP expressions in

Example 6.6

, inserted in the middle of each

HTML <input> element. The expressions use the conditional operator (?) described in the Java primer earlier in this chapter. As you may recall, the expression before the operator is first evaluated. If the result is true, the expression returns the value after the ?. If it's false, the value after the colon is returned.

Here the result is that if the array contains the choice represented by the checkbox, the checked attribute is inserted to render the checkbox as checked. Otherwise, an empty string is inserted, leaving the checkbox unchecked.

6.4.3 Using More Request Methods

We have already used one of the methods of the implicit request object, but this object provides a wealth of information you may be interested in. So let's use some more request methods.

Example 6.7

shows a page with a number of JSP expressions, each one displaying a piece of information

about the current request.

Example 6.7. Displaying Request Info (reqinfo.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

The following information was received:

<ul>

<li>Request Method: <%= request.getMethod( ) %>

<li>Request URI: <%= request.getRequestURI( ) %>

<li>Request Protocol: <%= request.getProtocol( ) %>

<li>Servlet Path: <%= request.getServletPath( ) %>

<li>Query String: <%= request.getQueryString( ) %>

<li>Server Name: <%= request.getServerName( ) %>

<li>Server Port: <%= request.getServerPort( ) %>

<li>Remote Address: <%= request.getRemoteAddr( ) %>

<li>Remote Host: <%= request.getRemoteHost( ) %>

<li>Browser Type: <%= request.getHeader("User-Agent") %>

</ul>

</body>

</html>

If you don't remember what some of these things mean, look back at

Chapter 2

. (The methods are also described in

Appendix B .)

Figure 6.4

shows an example of the output from this page.

Figure 6.4. Request information page

© 2001, O'Reilly & Associates, Inc.

6.5 Using an Expression to Set an Attribute

In all our JSP action element examples so far, the attributes are set to literal string values. But in many cases, the value of an attribute is not known when you write the JSP page; instead, the value must be calculated when the JSP page is requested. For situations like this, you can use a JSP expression as an attribute value. This is called a request-time attribute value. Here is an example of how this can be used to set an attribute of a fictitious log entry bean:

<jsp:useBean id="logEntry" class="com.foo.LogEntryBean" />

<jsp:setProperty name="logEntry" property="entryTime"

value="<%= new java.util.Date( ) %>" />

...

This bean has a property named entryTime that holds a timestamp for a log entry, while other properties hold the information to be logged. To set the timestamp to the time when the JSP page is requested, a <jsp:setProperty> action with a request-time attribute value is used.

The attribute value is represented by the same type of JSP expression as in the previous snippet, here an expression that creates a new java.util.Date object (representing the current date and time). The request-time attribute is evaluated when the page is requested, and the corresponding attribute is set to the result of the expression. As you might have guessed, any property you set this way must have a

Java type matching the result of the expression. In this case, the entryDate property must be of type java.util.Date

.

Not all attributes support request-time values. One reason is that some attribute values must be known when the page is converted into a servlet. For instance, the class attribute value in the

<jsp:useBean>

action must be known in the translation phase so that the JSP container can generate valid Java code for the servlet. Request-time attributes also require a bit more processing than static string values, so it's up to the custom action developer to decide if request-time attribute values are supported or not.

Appendix A , shows which attributes for the standard actions accept request-time

attributes, and

Appendix C provides the same information for the custom actions used in this book. If

you're a programmer, you may also want to skip ahead to Chapter 16 , to see how to declare that an

attribute in a custom action accepts request-time attributes.

© 2001, O'Reilly & Associates, Inc.

6.6 Declaring Variables and Methods

We have used two of the three JSP scripting elements in this chapter: scriptlets and expressions. There's one more, called a declaration element, which is used to declare Java variables and methods in a JSP page. My advice is this: don't use it. Let me explain why.

Java variables can be declared either within a method or outside the body of all methods, like this: public class SomeClass {

// Instance variable

private String anInstanceVariable;

// Method

public void doSomething( ) {

String aLocalVariable;

}

}

A variable declared outside the body of all methods is called an instance variable. Its value can be accessed from any method in the class, and it keeps its value even when the method that sets it returns.

A variable declared within the body of a method is called a local variable. A local variable can be accessed only from the method where it's declared. When the method returns, the local variable disappears.

Recall from Chapter 3 that a JSP page is turned into a servlet class when it's first requested, and the JSP

container creates one instance of this class. If the same page is requested by more than one user at a time, the same instance is used for each request. Each user is assigned what is called a thread in the server, and each thread executes the main method in the JSP object. When more than one thread executes the same code, you have to make sure the code is thread-safe. This means that the code must behave the same when many threads are executing as when just one thread executes the code.

Multithreading and thread-safe code strategies are best left to programmers. However, you should know that using a JSP declaration element to declare variables exposes your page to multithreading problems.

That's because a variable declared using a JSP declaration element ends up as an instance variable in the generated servlet, not as a local variable in a method. Since an instance variable keeps its value when the method executed by a thread returns, it is visible to all threads executing code in the same instance. If one thread changes the value of the instance variable, the value is changed for all threads. To put this in

JSP terms, if the instance variable is changed because one user accesses the page, all users accessing the same page will use the new value.

When you declare a variable within a scriptlet element instead of in a JSP declaration block, the variable ends up as a local variable in the generated servlet's request processing method. Each thread has its own copy of a local variable, so local variables can't cause any problems if more than one thread executes the same code. If the value of a local variable is changed, it will not affect the other threads.

That being said, let's look at a simple example. We use two int variables: one declared as an instance variable using a JSP expression, and the other declared as a local variable. We increment them both by

one and display the new values. Example 6.8

shows the test page.

Example 6.8. Using a Declaration Element (counter.jsp)

<%@ page language="java" contentType="text/html" %>

<%!

int globalCounter = 0;

%>

<html>

<head>

<title>A page with a counter</title>

</head>

<body bgcolor="white">

This page has been visited: <%= ++globalCounter %> times.

<p>

<%

int localCounter = 0;

%>

This counter never increases its value: <%= ++localCounter %>

</body>

</html>

The JSP declaration element is right at the beginning of the page, starting with <%! and ending with %>.

Note the exclamation point (!) in the start identifier; that's what makes it a declaration as opposed to a scriptlet. The declaration element declares an instance variable named globalCounter, shared by all requests for the page. In the body section of the page, a JSP expression increments the variable's value.

Next comes a scriptlet, enclosed by <% and %>, that declares a local variable named localCounter.

The last scriptlet increments the value of the local variable.

When you run this example, the globalCounter value increases every time you load the page, but localCounter

stays the same. Again, this is because globalCounter is an instance variable (its value is available to all requests and remains between requests) while localCounter is a local variable (its value is available only to the current request and is dropped when the request ends).

In this example, nothing terribly bad happens if more than one user hits the page at the same time. The

worst that could happen is that you skip a number or show the same globalCounter value twice.

This can happen if two requests come in at the same time, and both requests increment the value before it's inserted in the response. You can imagine the consequences, however, if you use an instance variable to save something more important, such as a customer's credit card number or other sensitive information. So even though it may be tempting to create an instance variable (using a JSP expression) to keep a value such as a counter between requests, I recommend that you stay away from this technique. We'll look at better ways to share information between requests in

Chapter 8

.

A JSP declaration element can also be used to declare a method that can then be used in scriptlets in the same page. The only harm this could cause is that your JSP pages end up containing too much code, making it hard to maintain the application. A far better approach is to use JavaBeans and custom actions.

But to be complete, Example 6.9

shows an example of how it can be done.

Example 6.9. Method Declaration and Use (color.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<%!

String randomColor( ) {

java.util.Random random = new java.util.Random( );

int red = (int) (random.nextFloat( ) * 255);

int green = (int) (random.nextFloat( ) * 255);

int blue = (int) (random.nextFloat( ) * 255);

return "#" +

Integer.toString(red, 16) +

Integer.toString(green, 16) +

Integer.toString(blue, 16);

}

%>

<h1>Random Color</h1>

<table bgcolor="<%= randomColor( ) %>" >

<tr><td width="100" height="100">&nbsp;</td></tr>

</table>

</body>

</html>

The method named randomColor( ), declared between <%! and %>, returns a randomly generated

String

in a format that can be used as an HTML color value. This method is then called from an

expression element to set the background color for a table. Every time you reload this page, you see a single table cell with a randomly selected color.

6.6.1 jspInit( ) and jspDestroy( )

You may remember from Chapter 2 that a servlet has two methods that the container calls when the

servlet is loaded and shut down. These methods are called init( ) and destroy( ), and they allow the servlet to initialize instance variables when it's loaded and clean up when it's shut down, respectively. As you already know, a JSP page is turned into a servlet, so it has the same capability.

However, with JSP, the methods are called jspInit( ) and jspDestroy( ) instead.

Again, I recommend that you do not declare any instance variables for your JSP pages. If you follow this advice, there's also no reason to declare the jspInit( ) and jspDestroy( ) methods. But I know you're curious, so here's an example of how they can be used.

Expanding on

Example 6.8

, the jspInit( ) method can be used to set an instance variable to a

java.util.Date( )

object, which represents the date and time when the page is loaded. This variable can then be used in the page to show when the counter was started:

<%@ page language="java" contentType="text/html" %>

<%@ page import="java.util.Date" %>

<%!

int globalCounter = 0;

Date startDate;

public void jspInit( ) {

startDate = new Date( );

}

public void jspDestroy( ) {

ServletContext context = getServletConfig( ).getServletContext( );

context.log("test.jsp was visited " + globalCounter +

" times between " + startDate + " and " + (new Date( )));

}

%>

<html>

<head>

<title>A page with a counter</title>

</head>

<body bgcolor="white">

This page has been visited: <%= ++globalCounter %> times

since <%= startDate %>.

</body>

</html>

The jspDestroy( ) method retrieves a reference to the ServletContext for the page and writes a message to the container's log file. If you recall that the implicit application variable contains a reference to the ServletContext, you may be wondering why it's not used here. The reason is that the implicit variables are available only in the method that the JSP container generates to process the page requests, not in the methods that you declare yourself.

Why Two Notations?

You may have noticed that two different notations are used for the different JSP elements:

XML-style notation, like <jsp:useBean>, for action elements, and <% %> notation for directives and scripting elements. If you're a purist, you may be wondering why the authors of the JSP specification mixed styles like this. Given that XML seems to be the future for all markup languages, why not use XML notation for all JSP elements?

There are two good reasons for not using the XML notation for directives and scripting elements:

Scripting elements contain scripting code, and many characters used in code are not valid in an XML document. If XML notation were used for the scripting elements, you would have to manually encode characters like < and > so they wouldn't be mistaken for XML control characters. That would be messy and a source of pernicious errors.

The <%@ directive %>, <% code %>, <%= expression %>, and <%!

declaration

%> notations are familiar for many developers since they are also used in Microsoft's Active Server Pages (ASP).

JSP actually defines XML-style equivalents for directives and scripting elements. But the

XML notation for these elements is intended to be used only by tools that generate complete

JSP pages. The tools can handle encoding of special characters automatically, as well as a number of other details needed to make a JSP page a well-formed XML document. The

XML style is also not completely defined in JSP 1.1; therefore, a JSP 1.1 container is not required to support it.

© 2001, O'Reilly & Associates, Inc.

7.1 Dealing with Syntax Errors

The first type of error you will encounter is the one you, or your co-workers, create by simple typos: in other words, syntax error. The JSP container needs every JSP element to be written exactly as it's defined in the specification in order to turn the JSP page into a valid servlet class. When it finds something that's not right, it will tell you. But how easy it is to understand what it tells you depends on the type of error, the JSP container implementation, and sometimes, on how fluent you are in computer gibberish.

7.1.1 Element Syntax Errors

Let's first look at how Tomcat reports some typical syntax errors in JSP directives and action elements.

Example 7.1

shows a version of the date.jsp page from Chapter 5 , with a syntax error.

Example 7.1. Improperly Terminated Directive (error1.jsp)

<%@ page language="java" contentType="text/html" >

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" />

The current time at the server is:

<ul>

<li>Date: <jsp:getProperty name="clock" property="date" />

<li>Month: <jsp:getProperty name="clock" property="month" />

<li>Year: <jsp:getProperty name="clock" property="year" />

<li>Hours: <jsp:getProperty name="clock" property="hours" />

<li>Minutes: <jsp:getProperty name="clock" property="minutes" />

</ul>

</body>

</html>

The syntax error here is that the page directive on the first line is not closed properly with %>; the percent sign is missing.

Figure 7.1

shows what Tomcat has to say about it.

Figure 7.1. Error message about an unterminated JSP directive

Tomcat reports the error by sending an error message to the browser. This is the default behavior for Tomcat, but it's not mandated by the JSP specification. The specification requires only that a response with the HTTP status code for a severe error (500) is returned, but how a JSP container reports the details is vendor-specific. For instance, the error message can be written to a file instead of to the browser. If you use a container other than Tomcat, check the container documentation to see how it reports these types of errors.

The actual error message in

Figure 7.1

is what is called an exception stack trace. When something goes really wrong in

a Java method, it typically throws an exception. An exception is a special Java object, and throwing an exception is the method's way of saying it doesn't know how to handle a problem. Sometimes another part of the program can take care of the problem in a graceful manner, but in many cases the best that can be done is to tell the user about it and move on.

That's what the Tomcat container does when it finds a problem with a JSP page during the translation phase: it sends the exception stack trace to the browser. The stack trace contains a message about what went wrong and where the problem occurred. The message is intended to be informative enough for a user to understand, but the actual trace information is of value only to a programmer. As you can see in

Figure 7.1

, the message is:

D:\ch7\error1.jsp(0,49) Unterminated <%@ tag

The first part of the message is the name of the JSP page. The numbers within parentheses indicate on which line and character position in the file the error was found (both the line and the position are numbered from 0), and then the message states what the problem is. So this message tells us that a directive on the first line (a tag starting with <@) is not terminated as expected at position 49. And in this case it's both the correct diagnosis and the right location.

It's not always this easy to interpret the error message, though. Example 7.2

shows another version of date.jsp with a

different syntax error.

Example 7.2. Improperly Terminated Action (error2.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" >

The current time at the server is:

<ul>

<li>Date: <jsp:getProperty name="clock" property="date" />

<li>Month: <jsp:getProperty name="clock" property="month" />

<li>Year: <jsp:getProperty name="clock" property="year" />

<li>Hours: <jsp:getProperty name="clock" property="hours" />

<li>Minutes: <jsp:getProperty name="clock" property="minutes" />

</ul>

</body>

</html>

The syntax error here is almost the same as the "unterminated tag" in Example 7.1

, but now it's the <jsp:useBean>

action element that's not terminated properly (it's missing the closing slash). The message reported by Tomcat in this case is:

D:\ch7\error2.jsp(16,0) useBean tag must begin and end in the same physical file

This is not as easy to relate to the actual location of the error. The line and position information points to the last character in the file. The reason for this is that since the action element doesn't have a body, the opening tag must end with />, as you may remember from

Chapter 5

. But in Example 7.2

, it's terminated with just >. Since that's valid syntax for a JSP action that contains a body, the JSP container can't tell that it's a syntax error at this point. Instead it treats it as the opening tag for an element with a body, and complains that it can't find the closing tag before the file ends. In this case, it's still not too hard to figure out what the real problem is, but if you have multiple

<jsp:useBean>

elements in the same file, it can be more complicated.

Another common error is a typo in an attribute name, as shown in

Example 7.3

. In the first <jsp:getProperty> action, the name attribute is missing the e.

Example 7.3. Mistyped Attribute Name (error3.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" />

The current time at the server is:

<ul>

<li>Date: <jsp:getProperty nam="clock" property="date" />

<li>Month: <jsp:getProperty name="clock" property="month" />

<li>Year: <jsp:getProperty name="clock" property="year" />

<li>Hours: <jsp:getProperty name="clock" property="hours" />

<li>Minutes: <jsp:getProperty name="clock" property="minutes" />

</ul>

</body>

</html>

Tomcat reports the problem like this:

D:\ch7\error3.jsp(7,10) getProperty: Mandatory attribute name missing

In this case, the typo is in the name of a mandatory attribute, so Tomcat reports it as missing. If the typo is in the name of an optional attribute, Tomcat reports it as an invalid attribute name.

The examples here are the most common ones for JSP actions, and as you can see, Tomcat can give you pretty good information about what's wrong in these cases. But this is still an area where I expect many improvements to be implemented in later versions of Tomcat as well as other JSP containers. The JSP authoring tools that are emerging now also help. By providing GUI-based interfaces that generate the action elements automatically, they can eliminate this type of syntax problem.

7.1.2 Scripting Syntax Errors

So far, so good, right? Not quite. Unfortunately, syntax errors in scripting elements result in error messages that are much harder to interpret. This is because of the way the JSP container deals with scripting code when it converts a JSP page into a servlet. The container reads the JSP page and generates servlet code by replacing all JSP directives and actions with code that produces the appropriate result. To do this, it needs to analyze these types of elements in detail.

If there's a syntax error in a directive or action element, it can easily tell which element is incorrect (as you saw in the previous section).

Scripting elements, on the other hand, are more or less used as-is in the generated servlet code. A syntax error in scripting code is not discovered when the JSP page is read, but instead when the generated servlet is compiled. The compiler reports an error in terms of its location in the generated servlet code (as opposed to the location in the JSP page), with messages that don't always make sense to a JSP page author. Let's look at some examples to illustrate this.

Example 7.4

shows a modified version of the greeting.jsp page from

Chapter 6

. The last scriptlet, with a brace closing the last else block, is missing.

Example 7.4. Missing End Brace (error4.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<jsp:useBean id="clock" class="java.util.Date" />

<% if (clock.getHours( ) < 12) { %>

Good morning!

<% } else if (clock.getHours( ) < 17) { %>

Good day!

<% } else { %>

Good evening!

</body>

</html>

This is the error description Tomcat sends to the browser (with some line breaks added to make it fit the page): org.apache.jasper.JasperException: Unable to compile class for JSP

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00034_0002ejsperror4_jsp_0.java:105:

'catch' without 'try'.

} catch (Exception ex) {

^

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00034_0002ejsperror4_jsp_0.java:114:

'try' without 'catch' or 'finally'.

}

^

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00034_0002ejsperror4_jsp_0.java:114:

'}' expected.

}

^

3 errors

This message probably doesn't make much sense to you. First of all, the filename is not the name of the JSP page, it's the name of the generated servlet, which contains sequence numbers and special encodings to make it a unique filename. Part of the name corresponds to the JSP page name, but a different JSP container than Tomcat may not use the same kind of naming convention, so there's no guarantee that this is true for all containers. Secondly, the line numbers listed are line numbers in the generated servlet source code file, not the line numbers in the JSP file. And lastly, the error message refers to 'catch' without 'try', which doesn't seem to match any code in the JSP page scriptlets. That's because the code with the missing brace is inserted into the block of code that outputs template text, invokes actions, and so forth—so the compiler gets confused about what the real problem is.

How can you find the real problem when you get this type of message? If you know how to program Java, you can look at the generated servlet source file and try to figure out what's really wrong. Most JSP containers can be configured to save the generated source code for you to look at. For Tomcat, it's the default behavior, and the complete name of the file is shown in the error message.

But if you're not a programmer, the only thing you can do is to study all scriptlets in your JSP page carefully and try to figure out what's wrong. That's not always easy, and it's yet another reason to avoid scripting in your JSP pages in the first place. When you have to use scripting, use only extremely simple code and be very careful with the syntax.

Let's look at couple of other common syntax errors so you at least know the types of messages to expect. Example 7.5

shows a version of the browser.jsp file from

Chapter 6

in which a closing parenthesis for the first if statement is missing.

Example 7.5. Missing Closing Parenthesis (error5.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<% if (request.getHeader("User-Agent").indexOf("MSIE") != -1 { %>

You're using Internet Explorer.

<%

} else

if (request.getHeader("User-Agent").indexOf("Mozilla") != 1) {

%>

You're using Netscape.

<% } else { %>

You're using a browser I don't know about.

<% } %>

</body>

</html>

The error message for this type of error is a bit easier to understand: org.apache.jasper.JasperException: Unable to compile class for JSP

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00035_0002ejsperror5_jsp_0.java:61:

')' expected.

if (request.getHeader("User-Agent").indexOf("MSIE") != -1 {

^

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00035_0002ejsperror5_jsp_0.java:68:

'else' without 'if'.

} else

^

2 errors

Here the syntax error doesn't cause any strange side effect errors when the scripting code is combined with other generated code, as the error in

Example 7.4

did. Instead, the message shows the code fragment where the real error is located.

Another typical mistake is shown in Example 7.6

. It's a part of the reqinfo.jsp page from Chapter 6 .

Example 7.6. Scriptlet Instead of Expression (error6.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

The following information was received:

<ul>

<li>Request Method: <%= request.getMethod( ) %>

<li>Request URI: <% request.getRequestURI( ) %>

<li>Request Protocol: <%= request.getProtocol( ) %>

</ul>

</body>

</html>

This is simply a case where the opening tag for a JSP expression (<%=) has been mistakenly written as the opening tag for a JSP scriptlet (<%). It looks like an innocent error, but the error message does not give you much help to find it: org.apache.jasper.JasperException: Unable to compile class for JSP

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00036_0002ejsperror6_jsp_0.java:67:

Invalid type expression.

request.getRequestURI( )

^

D:\tmp\Tomcat\jakarta-tomcat\work\localhost_8080%2Fora\

_0002fch_00037_0002ferror_00036_0002ejsperror6_jsp_0.java:70:

Invalid declaration.

out.write("\r\n

Request Protocol: ");

^

2 errors

Again, the scripting code and the generated code clash, resulting in a message that's hard to understand; but at least you can recognize the code from the JSP page and try to see what's really wrong.

The misleading and confusing error messages reported for scripting syntax errors are, in my opinion, a big problem, and one that's hard to solve completely, even with better JSP container implementations and tools. It can be minimized, for instance by providing information about where in the JSP page the error is introduced, but it's always hard for a container to pinpoint the real problem when scripting code is mixed with other generated code. My only advice at this point is (again) to avoid scripting code as much as possible.

© 2001, O'Reilly & Associates, Inc.

7.2 Debugging a JSP-Based Application

After you have fixed all syntax errors, pat yourself on the back and enjoy the moment. If the application is more than a trivial example, however, this moment will probably be short-lived: you will likely find that one or more things still don't work as you expected. Logic errors, such as not taking care of all possible input combinations, can easily slip into an application during development. Finding and correcting this type of problem is called debugging.

For applications developed in compiled languages such as Java, C, or C++, a tool called a debugger is often used in this phase. A debugger steps through the program line by line or runs until it reaches a break point that you have defined, and lets you inspect the values of all variables in the program. With careful analysis of the program flow in runtime, you can discover why it works the way it does, and not the way you want it to. There are debuggers for JSP as well, such as IBM's Visual Age for Java. This product lets you debug a JSP page exactly the same way as you would a program written in a more traditional programming language.

But a real debugger is often overkill for JSP pages. If your pages are so complex that you feel you need a debugger, you may want to move code from the pages into JavaBeans or custom actions instead.

These components can then be debugged with a standard Java debugger, which can be found in most

Java Interactive Development Environments (IDEs). To debug JSP pages, another time-tested debugging approach is usually sufficient: simply add code to print variable values to the screen.

Example 7.7

shows how you can use this approach to find an error in a modified version of the

browser.jsp page from Chapter 6 .

Example 7.7. Testing Header Values in the Wrong Order (browser.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

<% if (request.getHeader("User-Agent").indexOf("Mozilla") != -1) { %>

You're using Netscape.

<%

} else

if (request.getHeader("User-Agent").indexOf("MSIE") != 1) {

%>

You're using Internet Explorer.

<% } else { %>

You're using a browser I don't know about.

<% } %>

</body>

</html>

If you run this example in a Netscape browser, it responds with "You're using Netscape," as expected.

The problem is that if you run it with Internet Explorer, you get the same response. Clearly there's something wrong with the way the User-Agent header value is tested.

To find out why it doesn't work with Internet Explorer, you can add a one-line JSP expression that includes the value of the User-Agent header in the response:

<%@ page language="java" contentType="text/html" %>

<html>

<body bgcolor="white">

User-Agent header value: <%= request.getHeader("User-Agent") %>

<p>

...

The result is shown in Figure 7.2

.

Figure 7.2. Response with debug output

Now it's clear why it doesn't work: the User-Agent header value set by Internet Explorer also contains the string "Mozilla". So the test in the JSP page must be reversed, first looking for the string

"MSIE" (to identify Internet Explorer) and looking for "Mozilla" only if it's not found.

Adding a couple of JSP expressions to see variable values as part of the response in the browser is the easiest way to debug a JSP page. But sometimes multiple pages are involved in the processing of a

single request, as you will see in Chapter 8

. In this case, it may be better to write the debug output to a file or the command window where you started the server instead. To write to the standard log file for

the server, replace the JSP expression with this scriptlet:

<% application.log("User-Agent header value: " +

request.getHeader("User-Agent")); %>

The application variable is one of the implicit JSP objects described in Chapter 6

, containing a reference to the javax.servlet.ServletContext object for the application. It provides a log( )

method that writes messages to the application log file. The name and location of the file are server-dependent. With a default configuration of Tomcat, it's named logs/servlet.log.

Most servers, including Tomcat, also let you write messages to the window where the server was started, represented by the System.out object:

<% System.out.println("User-Agent header value: " +

request.getHeader("User-Agent")); %>

This approach works fine during development, when you run your own web server started in a command window, but you need to remember to remove these lines in your production code. Writing to the log file is useful when you debug an application that is running in a web server you don't have control over, or if you need to record the debug messages in a file for further analysis later.

To make it easier to generate the most common types of debug output, you can use the DebugBean class that I developed for this book. Its properties represent information that's available in a JSP page, as shown in

Table 7.1

.

Property Name

Table 7.1, Properties of com.ora.jsp.util.DebugBean

Java Type Access Description

pageContext elapsedTime requestInfo javax.servlet.jsp.PageContext write

Must be set in order for the bean to find the value of its other properties

String read

A string with the number of milliseconds elapsed since the bean was created or this property was last read

String read

A string, formatted as a table, with information about the request, such as authentication type, content length and encoding, path information, remote host and user, etc.

headers cookies parameters pageScope

String

String

String

String requestScope String sessionScope String applicationScope String read

A string, formatted as a table, with the names and values of all headers received with the request read

A string, formatted as a table, with the names and values of all cookies received with the request read

A string, formatted as a table, with the names and values of all parameters received with the request read

A string, formatted as a table, with the names and values of all page scope variables read

A string, formatted as a table, with the names and values of all request scope variables read

A string, formatted as a table, with the names and values of all session scope variables read

A string, formatted as a table, with the names and values of all application scope variables

The DebugBean has one write-only property, pageContext, that must be set to the corresponding implicit object to provide the bean access to all the information it can report on. All the other properties are read-only, providing access to different subsets of information of interest when debugging a JSP application.

To control where the information is written, pass a debug parameter with the request for the page with the bean. This parameter must have one or more of the following values (separated by plus signs): resp

Include the debug information in the response as an HTML table stdout

Write the debug information to System.out

log

Write the debug information to the application log file

Let's look at an example. The JSP page shown in

Example 7.8

creates an instance of the DebugBean

using a <jsp:useBean> action. It also sets the mandatory pageContext property using a nested

<jsp:setProperty>

action. The <jsp:setProperty> uses a request-time attribute value to assign the pageContext property a reference to the implicit pageContext variable.

Example 7.8. Page with the DebugBean (debug.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="java.util.*" %>

<html>

<body bgcolor="white">

<jsp:useBean id="debug" class="com.ora.jsp.util.DebugBean" >

<jsp:setProperty name="debug" property="pageContext"

value="<%= pageContext %>"

/>

</jsp:useBean>

<p>

<%-- Add test variables to the request scope --%>

<%

String[] arr = {"a", "b", "c"};

request.setAttribute("arr", arr);

java.util.Date date = new Date( );

request.setAttribute("now", date);

%>

<jsp:getProperty name="debug" property="headers" />

<jsp:getProperty name="debug" property="cookies" />

<jsp:getProperty name="debug" property="parameters" />

<jsp:getProperty name="debug" property="requestScope" />

<jsp:getProperty name="debug" property="elapsedTime" />

</body>

</html>

A scriptlet is used to set two request attributes, referred to in JSP as placing objects in the request scope

. Objects placed in the request scope can be accessed by all JSP pages used to process the same request.

Don't worry about how this works now; you'll learn more about all the JSP scopes in Chapter 8

. Here, it's only used to show you how the DebugBean displays scope information. Next, five

<jsp:getProperty>

actions are used to display the headers, cookies, parameters, requestScope

, and elapsedTime properties.

The DebugBean returns its property values only if the request contains a debug parameter with a valid value. Therefore, you can keep the bean in your pages all the time and activate it only when you need the debug info. If you request the page with the URL:

http://localhost:8080/ora/ch7/debug.jsp?debug=resp+stdout&a=b

you get the response shown in Figure 7.3

.

Figure 7.3. Debug output

Since the debug parameter specifies both resp and stdout, you also get all the debug information in the window where you started Tomcat.

© 2001, O'Reilly & Associates, Inc.

7.3 Dealing with Runtime Errors

Eventually, your application will work as you like. But things can still go wrong due to problems with external systems that your application depends on, such as a database. And even though you have tested and debugged your application, there may be runtime conditions that you didn't anticipate.

Well-behaved components such as JavaBeans or JSP actions (standard and custom) deal with expected error conditions in a graceful manner. For instance, the UserInfo bean used in

Chapter 5

has a valid

attribute that is false unless all properties are set to valid values. Your JSP page can then test the property value and present the user with an appropriate message.

But if something happens that makes it impossible for the component to do its job, it needs to tell the user about the problem. The standard way Java does this is to throw an exception. That's what the JSP container does when it finds a problem with a JSP page during the translation phase, as I described in the first section of this chapter. Components, such as JavaBeans and JSP actions, and the code in JSP scripting elements, can also throw exceptions when something goes really wrong. By default, the JSP container catches the exception and displays its message and stack trace in the browser, similar to

what's shown in Figure 7.1

. But that's hardly the type of error message you want the application users

to see. It's better to tell the JSP container to use a customized error page instead.

Example 7.9

shows a JSP page with a page directive that defines an error page.

Example 7.9. Page with an Error Page Definition (calc.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page errorPage="errorpage.jsp?debug=log" %>

<% request.setAttribute("sourcePage", request.getRequestURI( )); %>

<html>

<body bgcolor="white">

<jsp:useBean id="calc" class="com.ora.jsp.beans.calc.CalcBean">

<jsp:setProperty name="calc" property="*" />

</jsp:useBean>

<%-- Calculate the new numbers and state info --%>

<% String currentNumber = calc.getCurrentNumber( ); %>

<form action="calc.jsp" method="post">

<table border=1>

<tr>

<td colspan="4" align="right">

<%= currentNumber.equals("") ? "&nbsp;" :

currentNumber %>

<input type="hidden" name="currentNumber"

value="<%= currentNumber %>">

<input type="hidden" name="previousNumber"

value="<%= calc.getPreviousNumber( ) %>">

<input type="hidden" name="currentOperation"

value="<%= calc.getCurrentOperation( ) %>">

<input type="hidden" name="previousOperation"

value="<%= calc.getPreviousOperation( ) %>">

<input type="hidden" name="reset"

value="<%= calc.getReset( ) %>">

</td>

</tr>

<tr>

<td><input type="submit" name="digit" value=" 7 "></td>

<td><input type="submit" name="digit" value=" 8 "></td>

<td><input type="submit" name="digit" value=" 9 "></td>

<td><input type="submit" name="oper" value=" / "></td>

</tr>

<tr>

<td><input type="submit" name="digit" value=" 4 "></td>

<td><input type="submit" name="digit" value=" 5 "></td>

<td><input type="submit" name="digit" value=" 6 "></td>

<td><input type="submit" name="oper" value=" * "></td>

</tr>

<tr>

<td><input type="submit" name="digit" value=" 1 "></td>

<td><input type="submit" name="digit" value=" 2 "></td>

<td><input type="submit" name="digit" value=" 3 "></td>

<td><input type="submit" name="oper" value=" - "></td>

</tr>

<tr>

<td><input type="submit" name="digit" value=" 0 "></td>

<td>&nbsp;</td>

<td><input type="submit" name="dot" value=" . "></td>

<td><input type="submit" name="oper" value=" + "></td>

</tr>

<tr>

<td>&nbsp;</td>

<td>&nbsp;</td>

<td><input type="submit" name="clear" value=" C "></td>

<td><input type="submit" name="oper" value=" = "></td>

</table>

</form>

</body>

</html>

The errorPage attribute in the page directive specifies the URL path for the page displayed if an exception is thrown by any component or scripting code. In JSP 1.1, you cannot specify a regular

HTML page as the error page; it must be another JSP page. When the path is specified as in Example

7.9

, the error page must be located in the same directory as the page that references it. However, if it

starts with a slash (/), it's interpreted as relative to the application's context path. The context path is simply the root for all HTML, image, and JSP pages in an application, such as C:\Jakarta\jakarta-

tomcat\webapps\ora for the application containing all book examples (if you installed it in the directory

specified in Chapter 4

). This means you can define a common error page for all the JSP pages in an application, even if you place them in multiple subdirectories, by using a path like

/errorpage.jsp

.

Also note that the error page URL in Example 7.9

includes the debug request parameter, and that a scriptlet is used to set a request attribute:

<%@ page errorPage="errorpage.jsp?debug=log" %>

<% request.setAttribute("sourcePage", request.getRequestURI( )); %>

The debug parameter lets you use the DebugBean to log information about what went wrong in the error page. The sourcePage attribute, set to the URL for the current page, is also used in the error page, as you will see shortly.

The rest of the page in Example 7.9

implements a simple calculator, shown in

Figure 7.4

. It's intended

only to illustrate how the error page handling works, so I will not describe it in detail. When you're done reading this book, it might be a good exercise to figure it out yourself by looking at the source code.

Figure 7.4. Calculator page

If a user tries to divide a number by zero, the CalcBean used in this page to implement the calculator throws an exception. This triggers the error page shown in

Example 7.10

to be invoked.

Example 7.10. Error Page (errorpage.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page isErrorPage="true" %>

<html>

<body bgcolor="white">

We're sorry but the request could not be processed. The processing

error message is:

<blockquote>

<%= exception.getMessage( ) %>

</blockquote>

The message has been logged together with more detailed information

about the error so we can analyze it further. Please try again, and

<a href="mailto:[email protected]">let us know</a> if the

problem persists.

<%

application.log((String) request.getAttribute("sourcePage"),

exception);

%>

<jsp:useBean id="debug" class="com.ora.jsp.util.DebugBean">

<jsp:setProperty name="debug" property="pageContext"

value="<%= pageContext %>" />

</jsp:useBean>

<jsp:getProperty name="debug" property="parameters" />

</body>

</html>

At the top of the page is a page directive with the attribute isErrorPage set to true. This tells the

JSP container that a special implicit JSP object named exception should be initialized with a reference to the exception that caused the page to be invoked. The type of the exception object is java.lang.Throwable

. This class provides a method named getMessage( ) that returns a string with a message about what went wrong. A JSP expression is used to display this message to the

user, as shown in Figure 7.5

.

Figure 7.5. Customized error page

Next follows a scriptlet and a couple of actions to create and use the DebugBean introduced in the previous section. The scriptlet writes the value of the sourcePage request attribute plus the exception itself to the application log file, and the DebugBean writes all parameters received with the request to the same file. In this way, information about which page caused the problem, the exception that was thrown, and all parameter values that were received with the request causing the problem is logged in the application log file when something unexpected happens. You can therefore look at the log file from time to time and see what kinds of problems occur frequently, and hopefully fine-tune the application to avoid them or at least provide more specific error messages.

Dealing with syntax errors and bugs is part of the application development process. In this chapter, we have looked at some of the ways you can ease the pain. To minimize syntax errors, you can use the types of JSP development tools listed in

Appendix E , that provide JSP syntax highlighting. You can

also minimize the scripting code you put in your JSP pages by using beans and custom actions instead.

The DebugBean presented in this chapter helps you to see what's going on at runtime, and you can use it in a customized error page to log information about unexpected errors.

© 2001, O'Reilly & Associates, Inc.

8.1 Passing Control and Data Between Pages

As discussed in

Chapter 3

, one of the most fundamental features of JSP technology is that it allows for separation of request processing, business logic, and presentation, using what's known as the Model-View-

Controller (MVC) model. As you may recall, the roles of Model, View, and Controller can be assigned to different types of server-side components. In this part of the book, JSP pages are used for both the

Controller and View roles, and the Model role is played by either a bean or a JSP page. This is not necessarily the best approach, but it lets us focus on JSP features instead of getting into Java programming. If you're a programmer and interested in other role assignments, you may want to take a peek at

Chapter 13

, and Chapter 14

. These chapters describe other alternatives and focus on using a servlet as the Controller.

Using different JSP pages as Controller and View means that more than one page is used to process a request. To make this happen, you need to be able to do two things:

1. Pass control from one page to another.

2. Pass data from one page to another.

In this section, we look at a concrete example of how to separate the different aspects of an application and how JSP supports the two requirements above.

Let's revisit the User Info example developed in Chapter 5

, to describe how the different aspects of an application can be separated. In this example, the business logic piece is trivial. However, it sets the stage for more advanced application examples in the next section and the remaining chapters in this part of the book, all using the model introduced here.

We can categorize the different aspects of the User Info example like this:

Display the form for user input (presentation).

Validate the input (request processing and business logic).

Display the result of the validation (presentation).

Let's use a separate JSP page for each aspect. The restructured application contains three JSP pages, as shown in

Figure 8.1

.

Figure 8.1. User Info application pages

Here's how it works. The userinfoinput.jsp page displays an input form. The user submits this form to

userinputvalidate.jsp to validate the input. This page processes the request using the UserInfoBean and passes control (forwards) to either the userinfoinput.jsp page (if the input is invalid) or the

userinfovalid.jsp page (if the input is valid). If valid, the userinfovalid.jsp page displays a "thank you" message. In this example, the UserInfoBean represents the Model, the userinputvalidate.jsp page the

Controller, and userinfoinput.jsp and userinfovalid.jsp represent the Views.

This gives you the flexibility and maintainability discussed in Chapter 3 . If the validation rules change, a

Java programmer can change the UserInfoBean implementation without touching any other part of the application. If the customer wants a different look, a page author can modify the View JSP pages without touching the request processing or business logic code.

8.1.1 Passing Control from One Page to Another

Before digging into the modified example pages, let's go through the basic mechanisms. As shown in

Figure 8.1

, the userinfovalidate.jsp page passes control to one of two other pages based on the result of the input validation. JSP supports this through the <jsp:forward> action:

<jsp:forward page="userinfovalid.jsp" />

This action stops processing one page and starts processing the page specified by the page attribute instead, called the target page. The control never returns to the original page.

The target page has access to all information about the request, including all request parameters. You can also add additional request parameters when you pass control to another page by using one or more nested

<jsp:param>

action elements:

<jsp:forward page="userinfovalid.jsp" >

<jsp:param name="msg" value="Invalid email address" />

</jsp:forward>

Parameters specified with <jsp:param> elements are added to the parameters received with the original request. The target page, therefore, has access to both the original parameters and the new ones, and can access both types in the same way. If a parameter is added to the request using the name of a parameter that already exists, the new value is added to the list of values for the existing parameter.

The page attribute is interpreted relative to the location of the current page if it doesn't start with /. This is called a page-relative path. If the source and target page are located in the same directory, just use the name of the target page as the page attribute value, as in the previous example. You can also refer to a file in a different directory using notation like ../foo/bar.jsp or /foo/bar.jsp. When the page reference starts with /, it's interpreted relative to the top directory for the application's web page files. This is called a

context-relative path.

Let's look at some concrete examples to make this clear. If the application's top directory is

C:\Tomcat\webapps\myapp, page references in a JSP page located in

C:\Tomcat\webapps\myapp\registration\userinfo are interpreted like this: page="bar.jsp"

C:\Tomcat\webapps\myapp\registration\userinfo\ bar.jsp

page="../foo/bar.jsp"

C:\Tomcat\webapps\myapp\registration\foo\ bar.jsp

page="/foo/bar.jsp"

C:\Tomcat\webapps\myapp\foo\ bar.jsp

8.1.2 Passing Data from One Page to Another

JSP provides different scopes for sharing data objects between pages, requests, and users. The scope defines for how long the object is available and whether it's available only to one user or to all application users. The following scopes are defined:

Page

Request

Session

Application

Figure 8.2. Lifetime of objects in different scopes

Objects placed in the default scope, the page scope , are available only to actions and scriptlets within one page. That's the scope used in all examples you have seen so far. The request scope is for objects that need to be available to all pages processing the same request. The session scope is for objects shared by multiple requests by the same user, and the application scope is for objects shared by all users of the

application. See Figure 8.2

.

The <jsp:useBean> action has a scope attribute that you use to specify in what scope the bean should be placed. Here is an example:

<jsp:useBean id="userInfo" scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" />

The <jsp:useBean> action looks for a bean with the name specified by the id attribute in the specified scope. If one already exists, it uses that one. If it cannot find one, it creates a new instance of the class specified by the class attribute and makes it available with the specified name within the specified scope. If you would like to perform an action only when the bean is created, place the elements in the body of the <jsp:useBean> action:

<jsp:useBean id="userInfo" scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" >

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

In this example, the nested <jsp:setProperty> action sets all properties when the bean is created. If the bean already exists, the <jsp:useBean> action associates it with the name specified by the id attribute so it can be accessed by other actions and scripting code. In this case, the

<jsp:setProperty>

action is not executed.

8.1.3 All Together Now

At this point, you have seen the two mechanisms needed to let multiple pages process the same request: passing control and passing data. These mechanisms allow you to employ the MVC design, using one page for request processing and business logic, and another for presentation. The <jsp:forward> action can be used to pass control between the pages, and information placed in the request scope is available to all pages processing the same request.

Let's apply this to the User Info example. In Chapter 5

, different output was produced depending on whether or not the user input was valid. This was done by using a UserInfoBean property called propertyStatusMsg

to display either a success or failure message. Yet the input form was always shown, even when the input was valid.

No more of that. When we split up the different aspects of the application into separate JSP pages as shown in

Figure 8.1

, we will also change the example so that the form is shown only when something needs to be corrected. When all input is valid a confirmation page is shown instead.

Example 8.1

shows the top part of the userinfoinput.jsp page.

Example 8.1. Page for Displaying Entry Form (userinfoinput.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>User Info Entry Form</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="userInfo"

scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" />

<%-- Output list of values with invalid format, if any --%>

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

</font>

<%-- Output form with submitted valid values --%>

<form action="userinfovalidate.jsp" method="post">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<%= StringFormat.toHTMLString(userInfo.getUserName( )) %>" >

</td>

</tr>

...

The rest of the example is the same as before. If you compare Example 8.1

with the JSP page used in

Chapter 5

, the only differences are that the userInfo bean is placed in the request scope (the scope attribute is set to request) and the form's action attribute is set to the URL for the validation page instead of pointing back to the same page.

The validation page, userinfovalidate.jsp , is given in

Example 8.2

.

Example 8.2. Input Validation Page (userinfovalidate.jsp)

<%@ page language="java" %>

<jsp:useBean

id="userInfo"

scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" >

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

<% if (userInfo.isValid( )) { %>

<jsp:forward page="userinfovalid.jsp" />

<% } else { %>

<jsp:forward page="userinfoinput.jsp" />

<% } %>

This is the request processing page, using the bean to perform the business logic. Note that there's no

HTML at all in this page, only a page directive specifying the scripting language, action elements, and scriptlets. This is typical of a request processing page: it doesn't produce a visible response message, it simply takes care of business and passes control to the appropriate presentation page.

This example is relatively simple. We first create a new userInfo Bean named userInfo in the request scope and set its properties from the request parameters of the previous form. (Note that we don't obtain the data from an already existing userInfo Bean in that scope; we'll see why shortly.) A scriptlet calls the bean's isValid( ) method to validate the properties and uses the <jsp:forward> action to pass control to the appropriate View page.

If the input is invalid, the userinfoinput.jsp page is used again. This time the <jsp:useBean> action finds the existing userInfo bean in the request scope, and its properties are used to show an error message and fill out the fields that were entered correctly, if any. If all input is valid, the control is passed

to the userinfovalid.jsp page shown in Example 8.3

to present the "thank you" message.

Example 8.3. Valid Input Message Page (userinfovalid.jsp)

<html>

<head>

<title>User Info Validated</title>

</head>

<body bgcolor="white">

<font color=green size=+3>

Thanks for entering valid information!

</font>

</body>

</html>

This page tells the user all input was correct. It consists only of template text, so this could have been a regular HTML file. Making it a JSP page allows you to add dynamic content later without changing the

referring page, however. The results are shown in Figure 8.3

.

Figure 8.3. The valid input message page

Let's review how placing the bean in the request scope lets you access the same bean in all pages. The

user first requests the userinfoinput.jsp page ( Example 8.1

). A new instance of the userInfo bean is created in the request scope and used to generate the "enter all fields" status message. The user fills out

the form and submits it as a new request to the userinfovalidate.jsp page ( Example 8.2

). The previous

bean is then out of scope, so this page creates a new userInfo bean in the request scope and sets all bean properties based on the form field values. If the input is invalid, the <jsp:forward> action passes the control back to the userinfoinput.jsp page. Note that we're still processing the same request that initially created the bean and set all the property values. Since the bean is saved in the request scope, the

<jsp:useBean>

action finds it and uses it to generate an appropriate error message and fill out the form with any valid values already entered.

© 2001, O'Reilly & Associates, Inc.

8.2 Sharing Session and Application Data

As described in Chapter 2 , HTTP is a stateless, request-response protocol. This means that the browser

sends a request for a web resource, and the web server processes the request and returns a response.

The server then forgets this transaction ever happened. So when the same browser sends a new request, the web server has no idea that this request is related to the previous one. This is fine if you're dealing with static files, but it's a problem in an interactive web application. In a travel agency application, for instance, it's important to remember the dates and destination entered to book the flight so the customer doesn't have to enter the same information again when it's time to make hotel and rental car reservations.

The way to solve this problem is to let the server send a piece of information to the browser that the browser then includes in all subsequent requests. This piece of information, called a session ID, is used by the server to recognize a set of requests from the same browser as related: in other words, as part of the same session. A session starts when the browser makes the first request for a JSP page in a particular application. The session can be ended explicitly by the application, or the JSP container can end it after a period of user inactivity (the default value is typically 30 minutes after the last request).

Thanks to the session ID, the server knows that all requests from the same browser are related.

Information can therefore be saved on the server while processing one request and accessed later when another request is processed. The server uses the session ID to associate the requests with a session

object, a temporary in-memory storage area where servlets and JSP pages can store information.

The session ID can be transferred between the server and browser in a few different ways. The Servlet

2.2 API, which is the foundation for the JSP 1.1 specification, identifies three methods: using cookies, using encoded URLs, and using the session mechanism built into the Secure Socket Layer (SSL), the encryption technology used by HTTPS. SSL-based session tracking is currently not supported by any of the major servlet containers, but all of them support the cookie and URL rewriting techniques. JSP hides most of the details about how the session ID is transferred and how the session object is created and accessed, providing you with the session scope to handle session data at a convenient level of abstraction. Information saved in the session scope is available to all pages requested by the same browser during the lifetime of the session.

However, some information is needed by multiple pages independent of who the current user is. JSP supports access to this type of shared information through another scope, the application scope.

Information saved in the application scope by one page can later be accessed by another page, even if the two pages were requested by different users. Examples of information typically shared through the application scope are database connection pool objects, information about currently logged-in users, and cache objects to avoid frequent database lookups.

Figure 8.4

shows how the server provides access to the two scopes for different clients.

Figure 8.4. Session and application scopes

The upcoming examples in this chapter will help you see how the session and application scopes can be used.

8.2.1 Counting Page Hits

A simple page counter bean can be used to illustrate how the scope affects the lifetime and reach of shared information. The difference between the two scopes becomes apparent when you place the bean in both the session and application scopes. Consider the page shown in

Example 8.4

.

Example 8.4. Page with Counter Beans (counter1.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>Counter page</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="sessionCounter"

scope="session"

class="com.ora.jsp.beans.counter.CounterBean"

/>

<jsp:useBean

id="applCounter"

scope="application"

class="com.ora.jsp.beans.counter.CounterBean"

/>

<% String uri = request.getRequestURI( ); %>

<h1>Counter page</h1>

This page has been visited <b>

<%= sessionCounter.getNextValue(uri) %>

</b> times by the current user in the current session, and <b>

<%= applCounter.getNextValue(uri) %>

</b> times by all users since the application was started.

</body>

</html>

The bean used in this example, the com.ora.jsp.beans.counter.CounterBean, keeps a separate counter for each page where it's used. It's a class with just one method: public int getNextValue(String uri);

The method increments the counter for the page identified by the uri argument and returns the new value.

In Example 8.4

, two <jsp:useBean> actions are used to create one bean each for the session and application scopes. The bean placed in the session scope is found every time the same browser requests this page, and therefore counts hits per browser. The bean in the application scope, on the other hand, is shared by all users, so it counts the total number of hits for this page.

A scriptlet is used to ask the request object for the URI of the current page. The URI is then passed as an argument to the bean's getNextValue( ) method. A page is uniquely identified by its URI,

so the bean uses the URI as a unique identifier to represent the counter it manages for each page. If you run this example, you should see a page similar to

Figure 8.5

.

Figure 8.5. A page with session and application page hit counters

As long as you use the same browser, the session and application counters stay in sync. But if you exit your browser and restart it, a new session is created when you access the first page. The session counter starts from 1 again but the application counter takes off from where it was at the end of the first session.

Note that the bean described here keeps the counter values in memory only, so if you restart the server, both will start from 0 again.

Sessions and Multiple Windows

Even though session tracking lets an application recognize related requests, there's still one problem. This problem is related to the server's lack of knowledge of the client, and does not become obvious until you start testing an application that depends on session information. Consider what happens if you open two browser windows and start accessing the same web application. Will each window be associated with its own session, or will they share the same session? Unfortunately, the answer is not well-defined. And it doesn't matter if the server-side logic is implemented as servlets, JSP, ASP, CGI, or any other server-side technology.

The most commonly used browsers, Netscape Navigator and Microsoft Internet Explorer, both let you open multiple windows that are actually controlled by the same operating system process. Internet Explorer can be configured so that each window is instead controlled by a separate process. In the latter case, it's easy to answer the question: each window is associated with its own session. It's only when one process controls multiple windows that it gets a bit tricky; in this case, the answer depends on whether URL rewriting or cookies are used for session tracking.

When URL rewriting is used, the first request to the application from one window doesn't include a session ID, since no response with the session ID has been received yet. The server sends back the new session ID encoded in all URLs in the page. If a request is then submitted from the other window, the same thing happens: the server sends back a response with a new session ID. Hence, in this scenario each window is associated with a separate session.

If cookies are used to pass the session ID, the reverse is true. The first request submitted from one window doesn't contain a session ID, so the server generates a new ID and sends it back as a cookie. Cookies are shared by all windows controlled by the same process. When a request is then made from the other window, it contains the session ID cookie received as a result of the first request. The server recognizes the session ID and therefore assumes that the request belongs to the same session as the first request; both windows share the same session.

There's not much you can do about this. If you want each window to have its own session, most servers can be configured to always use the URL rewriting method for session tracking. But this is not foolproof. The user can open a new window by using the mouse pop-up menu for a link (with the session ID encoded in the URI) and ask to see the linked page in a new window. Now there are two windows with the same session ID anyway. The only way to handle this, unfortunately, is to educate your users.

8.2.1.1 Thread-safe beans

You probably noticed that the CounterBean doesn't have the type of property getter method used in the examples in previous chapters. Instead, it has a method named getNextValue( ) that takes the

URI for the page as an argument. The <jsp:getProperty> action can be used only to display properties when the bean implements the standard no-arguments getter methods, so in

Example 8.4

,

JSP expressions are used to display the counter values instead.

The use of a method with an argument instead of a typical no-argument JavaBeans getter method is by design, and highlights a very important consideration for beans used in the session and application scopes. In

Chapter 6 , we discussed how multiple requests are handled by the server in parallel by

separate threads. The server assigns one execution thread to each request, and all threads have access to the same variable values for variables declared with JSP declaration elements. The result is that if such a variable is changed while processing one page request, the new values are used by all other requests as well. Beans in the session and application scopes are open to the same kind of problem, since they are also shared by multiple threads.

To illustrate this, let's look at what could happen if we used a bean with the traditional setter and getter

methods to implement the counter instead. Such a bean could have the properties shown in Table 8.1

.

Property Name

Table 8.1, Traditional Counter Bean Properties

Java Type Access Description

uri String write The unique URI used to identify the page counter nextValue int read The counter's value incremented by one read-only property

Part of the JSP page would then look like this:

<jsp:useBean

id="applCounter"

scope="application"

class="com.ora.jsp.beans.counter.CounterBean"

/>

<jsp:setProperty

name="applCounter"

property="uri"

value="<%= request.getRequestURI( ) %>"

/>

<h1>Counter page</h1>

This page has been visited <b>

<jsp:getProperty

name="applCounter"

property="nextValue"

/>

</b> times by all users since the application was started.

The first time you access this page, the <jsp:useBean> action creates the bean and saves it in the application

scope. The <jsp:setProperty> action sets the uri property to the unique URI for this page. Then the <jsp:getProperty> action calls the nextValue property getter method.

This method uses the uri property value to locate the counter for the page and increments it.

As long as you use this bean in only one page, this works fine. But if you want to keep track of all hits for two pages, each with its unique URI, and the two pages are requested at almost the same time, you're in trouble. Say one user requests the first page, called /ora/pageOne. This page sets the uri property to /ora/pageOne. But before it gets to the <jsp:getProperty> action, another user

requests the second page, say /ora/pageTwo. Now the second page finds the same bean

(applCounter) and sets the uri property to its URI (/ora/pageTwo). So when the first page eventually executes the <jsp:getProperty> action, it increments and displays the counter for the second page instead of its own counter.

Here I used a bean in the application scope as an example, because it's easy to understand that beans shared by all users can be accessed by more than one request at a time. But session scope beans can also be accessed at roughly the same time, by the same user. A good example of when two requests are made at nearly the same time by the same user is when an HTML frame set is used and each frame contains a JSP page.

You can solve this problem by using a bean with a regular method that takes the URI as an argument, instead of setting the URI with one setter method and incrementing the corresponding counter value with another method that depends on the URI value set by the first. This way, the method gets all the information it needs in one shot, and there's no risk of interference by other threads.

8.2.2 URL Rewriting

As I mentioned earlier, the session ID needed to keep track of requests within the same session can be transferred between the server and the browser in a number of different ways. One way is to encode it in the URLs created by the JSP pages, called URL rewriting . This approach works even if the browser doesn't support cookies (perhaps because the user has disabled them). A URL with a session ID looks like this: counter3.jsp;jsessionid=be8d691ddb4128be093fdbde4d5be54e00

When the user clicks on a link with an encoded URL, the server extracts the session ID from the request URI and associates the request with the correct session. The JSP page can then access the session data in the same fashion as when cookies are used to keep track of the session ID, so you don't have to worry about how it's handled. What you do need to do, however, is to call a method that lets the

JSP container encode the URL when needed. To see how it's done, let's create two pages that reference each other using a regular HTML link. A CounterBean in the session scope is used to increment a

counter for each page. Example 8.5

shows one of the pages. The other page is identical, except for the title and the link at the bottom.

Example 8.5. Page with an Encoded Reference to Another Page (counter2.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>Counter page 1</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="sessionCounter"

scope="session"

class="com.ora.jsp.beans.counter.CounterBean"

/>

<% String uri = request.getRequestURI( ); %>

<h1>Counter page 1</h1>

This page has been visited <b>

<%= sessionCounter.getNextValue(uri) %>

</b> times by the current user in the current session.

<p>

Click here to get to

<a href="<%= response.encodeURL("counter3.jsp") %>">

Counter page 2</a>.

</body>

</html>

The only differences compared to Example 8.4

are that only the session counter is used, and the link to the other page has been added.

The <a> element's href attribute value is converted using the encodeURL( ) method of the

implicit JSP response object, described in Chapter 6

. If a cookie is used to transfer the session ID between the browser and server, the encodeURL( ) method just returns the URL untouched. But if the browser doesn't support cookies, or cookie support is disabled, this method returns the URL with the session ID encoded as a part of the URL, as shown earlier.

If you want to provide session tracking for browsers that don't support cookies, you must use the encodeURL( )

method to rewrite all URL references in your application: in <a> tags, <form> tags, and <frameset> tags. This means all pages in your application (or at least all pages with references to other pages) must be JSP pages, so that all references can be dynamically encoded. If you miss one single URL, the server will lose track of the session.

I recommend that you take the time to add encodeURL( ) calls for all references up front, even if you know that all your current users have browsers that support cookies. One day you may want to extend the user base and lose control over the browsers they use. It's also common that users disable cookies in fear of Big Brother watching. Yet another reason to prepare for URL rewriting from the beginning is to support new types of clients that are becoming more and more common, such as PDAs and cell phones. Cookie support in these small devices is not a given.

© 2001, O'Reilly & Associates, Inc.

8.3 Using Custom Actions

You might be wondering if we are stretching the bean model too far in the previous example. Perhaps.

The CounterBean does more than hold information; it also has a non-conforming method for incrementing the counter. If we stray away from the purely bean model and use methods with arguments, this may force us to use scriptlets instead of the standard actions. That's not necessarily bad, but in this case we can do better using a custom action instead of a bean and the standard actions.

A custom action is just like the standard actions we've used so far. It has a start tag, which may contain attributes, and an end tag. It can also have a body. Here's what a custom action looks like:

<ora:incrementCounter scope="session"/>

The JSP specification defines how the standard set of actions can be extended with custom actions developed by Java programmers in the team or by a third party. A custom action is used in a JSP page in exactly the same way as the standard JSP actions you have seen in previous examples, such as

<jsp:getProperty>

. This makes them easier to use than beans with methods that must be invoked with scripting code, since you don't have to worry about missing braces and semicolons and other syntax details. A custom action can do pretty much anything: it has access to all information about the request and can add content to the response body as well as set response headers.

If you're a programmer, you should know that a custom action is basically a JavaBeans class, with property setter methods corresponding to the action's attributes, plus a few extra methods used by the

JSP container to invoke the action. You can read all about how to develop your own custom actions in

Chapter 16

.

As is often the case in software development, it's hard to say exactly whether a bean or a custom action is the preferred component type. My rule of thumb is that a bean is a great carrier of information, and a custom action is great for processing information. Custom actions can use beans as input and output. For instance, an action can be used to save the properties of a bean in a database, or to get information from a database and make it available to the page as a bean.

If you're a page author, you don't have to worry about the implementation details. All you need to know right now is how to use the custom actions you have available. You'll find many custom actions in this book that you can use, and more are available from open source projects and commercial companies listed in

Appendix E .

Custom actions are grouped together in a tag library. Consequently, you often see custom actions referred to as custom tags, even though that is not strictly correct. A tag library consists of a Tag Library

Descriptor (TLD) and the Java classes used to implement the custom actions. The TLD contains information about the action names and attributes. It's used by the JSP container during the translation phase to verify that all actions are used correctly in the page, for instance that all mandatory attributes are specified. Typically, the TLD and all classes are packaged in a Java Archive (JAR) file. You install such a library by placing the JAR file in the WEB-INF/lib subdirectory for the application in which it's used. If you look at the files in your Tomcat installation for the ora application (containing all the book examples), you see the JAR file in WEB-INF/lib/orataglib_1_0.jar and the TLD in WEB-

INF/tlds/orataglib_1_0.tld.

When you use custom actions in a JSP page, you must identify the library using the taglib directive:

<%@ taglib uri="/orataglib" prefix="ora" %>

The uri attribute value identifies the library. Depending on how the library is installed, different types of values are used: a symbolic name, the path to the JAR file, or the path to the TLD file. My recommendation is to use a symbolic name, as shown in the example. The symbolic name must then be mapped to the location of the library in the WEB-INF/web.xml file for the application:

<web-app>

...

<taglib>

<taglib-uri>

/orataglib

</taglib-uri>

<taglib-location>

/WEB-INF/tlds/orataglib_1_0.tld

</taglib-location>

</taglib>

...

</web-app>

The <taglib-uri> element contains the symbolic name, and the <taglib-location> element contains the path to either the JAR file or the TLD file. The path typically starts with a slash (/) and is then interpreted as a context-relative path, in other words, relative to the top directory for the application. This indirection—using a symbolic name that's mapped to the real location—is especially helpful as it allows you to change the name of the tag library file for all JSP pages in one place, for instance when you upgrade to a later version of the library.

For a simple application, you may feel that the indirection is overkill. If so, you can use the path to the

JAR file explicitly as the uri attribute value:

<%@ taglib uri="/WEB-INF/lib/orataglib_1_0.jar" prefix="ora" %>

All JSP 1.1-compliant containers should be able to find the TLD file in the JAR file, but this is a recent clarification of the specification. If the container you use doesn't support this yet (such as Tomcat 3.1), you must use the path to the TLD file instead of the path to the JAR file:

<%@ taglib uri="/WEB-INF/tlds/orataglib_1_0.tld" prefix="ora" %>

In both cases, the path may start with a slash and is then interpreted as a context-relative path. Without a starting slash, the path is interpreted as relative to the JSP page.

The prefix attribute defines a prefix used for the actions in this library. This prefix is used as part of the custom action names, as you will soon see. If you use more than one library in a page, each must have a unique prefix. You can use any prefix you like except jsp, jspx, java, javax, servlet, sun

, and sunw, which are reserved. The ora prefix is used for all custom actions in the examples in this book.

As I mentioned earlier, a custom action is used in a JSP page just like the standard actions we've used so far. In other words, it has a start tag, which may contain attributes, and an end tag. It can also have a body. Let's revisit our example from earlier:

<ora:incrementCounter scope="session"/>

The name consists of the prefix you specified with the taglib directive, and a unique name within the library, separated by a colon (:). As with standard actions, all attribute names are case-sensitive, and the value must be enclosed in single or double quotes.

Now let's see how we can use two custom tags to improve the counter example. The attributes for the custom actions are described in

Table 8.2

and

Table 8.3

.

Table 8.2, Attributes for <ora:incrementCounter>

Attribute Name Java Type Request-Time Value Accepted Description

scope String

No

Specifies the scope for the counter. Valid values are page

, request, session, and application.

Default is page.

The <ora:incrementCounter> action increments a unique counter for the page where it's used.

The counter can be placed in any of the standard JSP scopes. For instance, it can be placed in the session scope to count hits by different clients, or the application scope to count hits by all clients. The first time the action is used for a specific scope, the counter is created and set to 1.

Table 8.3, Attributes for <ora:showCounter>

Attribute Name Java Type Request-Time Value Accepted Description

scope String

No

Specifies the scope for the counter. Valid values are page

, request, session, and application.

Default is page.

The <ora:showCounter> action inserts the value of the page counter for the specified scope in the response. If a counter has not been created using the <ora:incrementCounter> action, the value -

1 is displayed.

These two actions are generic, so you can use them in your own pages if you want to keep track of the number of hits. The type of information shown in Tables 8-2 and 8-3 is what you should expect (or even demand!) from the custom action developer, whether it's developed in-house or by a third party.

Example 8.6

shows how our custom actions are used.

Example 8.6. Page with Counter Custom Actions ( counter4.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Counter page 1</title>

</head>

<body bgcolor="white">

<ora:incrementCounter scope="session"/>

<ora:incrementCounter scope="application"/>

<h1>Counter page 1</h1>

This page has been visited <b>

<ora:showCounter scope="session"/>

</b> times by the current user in the current session, and <b>

<ora:showCounter scope="application"/>

</b> times by all users since the counter was reset.

<p>

To see that a unique counter is maintained per page,

take a look at

<a href="<ora:encodeURL url="counter5.jsp" />">Counter page 2</a>.

</body>

</html>

As described in Tables 8-2 and 8-3, both actions have a scope attribute, supporting the same scopes as the JSP standard actions: page, request, session, and application. The

<ora:incrementCounter>

action finds or creates a counter for the current page in the specified scope and increments it by one, while <ora:showCounter> displays the current value of the counter. Notice that you don't have to tell the actions about the URI as you did with the beans in

Example 8.5

. That's because the JSP container makes all the implicit objects, such as the request object, available to a custom action automatically. The action can therefore figure out the current URI all by itself.

Another custom action, <ora:encodeURL>, is used to take care of the URL encoding of the link to the next page. It's described in

Table 8.4

.

Table 8.4, Attributes for <ora:encodeURL>

Attribute Name Java Type Request-Time Value Accepted Description

url String

Yes Mandatory. Specifies the URL to encode.

You can use this action element as an alternative to the scripting code used for URL encoding in

Example 8.5

. This action performs the same session ID encoding as the scripting code. Also, it encodes

the parameters defined by nested <ora:param> actions (see Table 8.5

) according to the syntax rules for HTTP parameters:

<ora:encodeURL url="product.jsp">

<ora:param name="id" value="<%= product.getId( )%>" />

<ora:param name="customer" value="Hans Bergsten" />

</ora:encodeURL>

Recall that all special characters, such as whitespace, quotes, etc., in a parameter value must be encoded.

For instance, all spaces in a parameter value must be replaced with plus signs. When you use the

<ora:encodeURL>

action, it takes care of all this encoding. The encoded URL created by the action for this example looks something like this: product.jsp;jsessionid=be8d691ddb4128be0?id=3&customer=Hans+Bergsten

Here, the session ID and the request parameters are added, and encoded if needed (the space between

"Hans" and "Bergsten" is replaced with a plus sign).

Attribute Name

Table 8.5, Attributes for <ora:param>

Java Type Request-Time Value Accepted Description

name String

Yes Mandatory. The parameter name.

value String

Yes Mandatory. The parameter value.

As illustrated by the counter example, custom actions allow you to write cleaner pages, avoiding most

(if not all) scripting code. Since pages without code are easier to develop and maintain, plenty of custom actions are used in the remainder of the examples in this book. Many are generic, so you can use them in your own applications as well. How to implement most of them is described in

Chapter 16

and

Chapter

17

, and you'll find the source code for all actions included in the example code package for this book.

You may be wondering why it's necessary to develop custom actions for generic things such as looping and URL encoding, as well as for common functions such as accessing a database. The reason is that the specification writers only defined a small set of standard actions in JSP 1.1. This was primarily motivated by time constraints; it was important to get the JSP 1.1 specification released as soon as possible. But perhaps more importantly, before specifying a larger set of actions, the specification group wanted feedback on the type of actions users needed. At the time this book is being written, a specification of more standard actions is being prepared. It will likely contain many actions similar to the custom actions you find in this book to be rolled into a future version of the JSP specification.

© 2001, O'Reilly & Associates, Inc.

8.4 Online Shopping

Now let's look at a more useful example: an online shopping site. Besides showing you how the session and application scopes can be used effectively in a larger application, this example also introduces many other useful tools. You'll see a number of generic custom actions you can use in your own applications, and learn how to use the java.text.NumberFormat class to format numbers.

The application consists of three pages. The main page lists all available products. Each product is linked to a product description page, where the product can be added to the shopping cart. A product is added to the shopping cart by a request processing page. The main page with the product list is then displayed again, but now with the current contents of the shopping cart as well, as shown in

Figure 8.6

.

Figure 8.6. The product list and the contents of the shopping cart

Two beans are used to keep track of the products: the com.ora.jsp.beans.shopping.CatalogBean

contains all available products, and the com.ora.jsp.beans.shopping.CartBean

represents one user's shopping cart. Each product in the catalog is represented by a ProductBean .

Tables

Table 8.6

,

Table 8.7

, and Table 8.8

show all the properties for the beans.

Table 8.6, Properties for com.ora.jsp.beans.shopping.CatalogBean

Property Name

productList

Java Type

com.ora.jsp.beans.

shopping.ProductBean[]

Access Description

read A list of all products in the catalog

Table 8.7, Properties for com.ora.jsp.beans.shopping.CartBean

Property Name Java Type Access Description

empty boolean read true

if the cart is empty, false otherwise productList com.ora.jsp.beans.

shopping.ProductBean[] read A list of all products in the cart product total com.ora.jsp.beans.

shopping.ProductBean

float write Adds a product to the cart read The total price for all products in the cart

Property Name

Table 8.8, Properties for com.ora.jsp.beans.shopping.ProductBean

Java Type Access Description

name String read The product name price float read The product price id String read The unique product ID descr String read A description of the product

The ProductBean objects are created by the CatalogBean when it's created.

Figure 8.7

shows how the beans are related.

Figure 8.7. Application and session scope beans

The CatalogBean and the ProductBean objects are placed in the application scope, since all users have access to the same product catalog. A unique CartBean is needed for each user to keep track of individual purchases, so each user has an instance of this bean in the session scope. When a user picks a product from the catalog, a reference to the corresponding ProductBean is added to the user's

CartBean

.

The main page for this application is shown in Example 8.7

.

Example 8.7. Page with a List of Products (catalog.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="java.text.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Product Catalog</title>

</head>

<body bgcolor="white">

<h1>Product Catalog</h1>

Please select a book from our catalog to read more about it and

decide if you would like to purchase a copy:

<jsp:useBean

id="catalog"

scope="application"

class="com.ora.jsp.beans.shopping.CatalogBean"

/>

<%--

Generate a list of all products with links to the product page.

--%>

<ul>

<ora:loop name="catalog"

property="productList"

loopId="product"

className="com.ora.jsp.beans.shopping.ProductBean">

<li>

<a href="<ora:encodeURL url="product.jsp">

<ora:param name="id"

value="<%= product.getId( )%>"/>

</ora:encodeURL>"><%= product.getName( ) %></a>

</ora:loop>

</ul>

<jsp:useBean

id="cart"

scope="session"

class="com.ora.jsp.beans.shopping.CartBean"

/>

<%-- Show the contents of the shopping cart, if any --%>

<%

if (!cart.isEmpty( )) {

NumberFormat numFormat = NumberFormat.getCurrencyInstance( );

%>

Your shopping cart contains the following items:

<p>

<table border=0>

<ora:loop name="cart"

property="productList"

loopId="product"

className="com.ora.jsp.beans.shopping.ProductBean">

<tr>

<td><%= product.getName( ) %></td>

<td><%= numFormat.format(product.getPrice( )) %></td>

</tr>

</ora:loop>

<tr><td colspan=2><hr></td></tr>

<tr>

<td><b>Total:</b></td>

<td><%= numFormat.format(cart.getTotal( )) %></td></tr>

</table>

<% } %>

</body>

</html>

The <jsp:useBean> action near the top of Example 8.7

creates an instance of the CatalogBean the first time a user requests the page. Since the bean is placed in the application scope, all users will then share this single instance.

The CatalogBean has a property that contains a list of all the products in the catalog, named productList

. Its value is an array of ProductBean objects. A custom action called <ora:loop>

, described in

Table 8.9

, is used to loop through the list and generate an HTML list item element for each product.

Table 8.9, Attributes for <ora:loop>

Attribute Name Java Type Request-Time Value Accepted Description

name property

String

No

String

No

Mandatory. The name of a data structure object or bean.

The object must be of type Object[], Vector,

Dictionary

, or Enumeration, or be a bean with a property of one of these types. The object or bean can be located in any JSP scope.

Optional. The name of a bean property. The property must be of type Object[], Vector, Dictionary, or

Enumeration

.

loopId String

No className String

No

Mandatory. The name of the variable that holds a reference to the current element when the action's body is evaluated.

Mandatory. The class name for the elements of the bean or property.

The <ora:loop> action iterates through the elements of an object, or the elements represented by a property, and evaluates the body once for each element, making the element available to other actions and scripting elements in the body through the variable name specified by loopId. The implementation of the loop action is described in

Chapter 16 .

In

Example 8.7

, the name attribute specifies the cart bean. The cart bean has an indexed

(multivalue) property named productList. That's the one we ask the <ora:loop> action to loop over, by naming it in the property attribute. Finally, we set the loopId attribute to product, so we can use product as a variable name in the action element body, and specify the class name for the

ProductBean

with the className attribute.

The body of the <ora:loop> action is evaluated once per element. The action body can contain a mixture of template text, scripting elements, and other actions. Here the body contains the HTML for a list item with a reference to another page, using the product name as the link text. Let's look at how the link is generated:

<a href="<ora:encodeURL url="product1.jsp">

<ora:param name="id"

value="<%= product.getId( )%>"/>

</ora:encodeURL>"><%= product.getName( ) %></a>

Within the body, the <ora:encodeURL> custom action described earlier is used to generate the <a> element's href attribute value. A nested <ora:param> action adds a parameter named id with the value set to the product ID for the current product. It's done by using a JSP expression (a request-time attribute value, described in

Chapter 6 ) that calls the ProductBean property getter method getID(

)

. A similar expression is used to set the link text to the name of the current product.

After the code for generating the product list in Example 8.7

, you see almost identical code for generating a list of the current contents of the shopping cart. First, the <jsp:useBean> action places the cart bean in the session scope, as opposed to the catalog bean, which is placed in the application scope. This means that each user gets a unique shopping cart that remains on the server for the duration of the session, while all users share the same catalog.

8.4.1 Number Formatting

Unless the shopping cart is empty, the second <ora:loop> action generates a list of the contents as an

HTML table with the name and price of each product. Note the java.text.NumberFormat object created in the same scriptlet as the if statement:

<%

if (!cart.isEmpty( )) {

NumberFormat numFormat = NumberFormat.getCurrencyInstance( );

%>

The NumberFormat class is a Java standard class used to format numbers. You can set up rules for the number of decimals to show, where to put number grouping characters, prefix and suffix, etc. Even more

important, the number is formatted according to the number format rules for the specific geographical, political, or cultural region where the server is located (by default). A collection of rules for a region is called a locale. It defines things such as which characters to use as a decimal separator, thousand grouping, and currency symbol. You can read more about the NumberFormat class in the standard

Java API documentation. We will discuss locales in detail in Chapter 11 , but to give you an idea of how

formatting varies between regions, here's an example of the number 10,000.00 formatted as currency for

USA, Sweden, and Italy:

USA: $10,000.00

Sweden: 10 000,00 kr

Italy: L. 10 000

We get a reference to the default formatter for currency information, using the getCurrencyInstance( )

method, and assign it to a variable named numFormat It's then used in the <ora:loop> body to format the price information for each product and for everything in the cart.

8.4.2 Using Request Parameters

As discussed earlier, a link to a description page for each product is generated using the <ora:loop> action in the main page, shown in

Example 8.7

. The link includes the request parameter id, specifying

the product to display information about. When the user clicks on one of the links, the page shown in

Example 8.8

is invoked.

Example 8.8. Product Description Page (product.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Product Description</title>

</head>

<body bgcolor="white">

<jsp:useBean

id="catalog"

scope="application"

className="com.ora.jsp.beans.shopping.CatalogBean"

/>

<%-- Get the ProductBean from the catalog --%>

<ora:useProperty id="product" name="catalog" property="product"

arg="<%= request.getParameter(\"id\") %>"

className="com.ora.jsp.beans.shopping.ProductBean" />

<h1>

<jsp:getProperty name="product" property="name" />

</h1>

<jsp:getProperty name="product" property="descr" />

<p>

<a href="<ora:encodeURL url="addtocart.jsp">

<ora:param name="id" value="<%= product.getId( ) %>"/>

</ora:encodeURL>">Add this book to the shopping cart</a>

</body>

</html>

The value of a request parameter can be retrieved from the implicit request object using the getParameter( )

method. As described in Chapter 6

, the request object is an instance of the class HttpServletRequest, and provides methods to find out everything the server knows about the request. The results are shown in

Figure 8.8

.

Figure 8.8. The product description page

In

Example 8.8

, the getParameter( ) method is used as a request-time attribute value to set the

arg

attribute for the <ora:useProperty> custom action, described in

Table 8.10

.

Table 8.10, Attributes for <ora:useProperty>

Attribute Name Java Type Request-Time Value Accepted Description

id String

No

Mandatory. The name of the variable to hold the retrieved bean. The bean is placed in the page scope.

name property arg className

String

String

String

String

No

No

Yes

No

Mandatory. The name of the object with the bean to retrieve. The object must be available in one of the standard scopes.

Mandatory. The name of the property holding the bean.

Optional. The argument value used to identify one specific bean.

Mandatory. The class name for the retrieved bean.

The <ora:useProperty> action is similar to the <jsp:useBean> action in that it associates a bean with a variable name. But instead of trying to locate the bean in a specified scope and create it if it isn't found, the <ora:useProperty> action gets the bean from another object (available in any of the standard scopes). It does this by calling the getter method for the specified property with the argument specified by the arg attribute if present. In

Example 8.8

, the <ora:useProperty> action is used to get the ProductBean that matches the ID passed as a parameter to the page from the catalog.

Note how the double quotes in the getParameter( ) method argument are preceded with a backslash (\): arg="<%= request.getParameter(\"id\") %>"

Whenever you use the same type of quote within an attribute value as you use to enclose the attribute value, you must escape it with a backslash. If you forget this, the JSP container is unable to figure out where the attribute value ends, and the page will not be converted to a valid servlet. Instead you will get a syntax error message when you access the page the first time. An alternative is to use one type of quote within the value and another to enclose the value. For instance, here you could use single quotes to enclose the value instead: arg='<%= request.getParameter("id") %>'

The remainder of

Example 8.8

uses actions we have already discussed to generate the product

information and the link to the business logic page that adds the product to the shopping cart. The

request processing page is shown in Example 8.9

.

Example 8.9. Adding a Product to the Shopping Cart (addtocart.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib " prefix="ora" %>

<jsp:useBean

id="catalog"

scope="application"

class="com.ora.jsp.beans.shopping.CatalogBean"

/>

<jsp:useBean

id="cart"

scope="session"

class="com.ora.jsp.beans.shopping.CartBean"

/>

<%-- Get the ProductBean from the catalog and save it in the cart --%>

<ora:useProperty id="product" name="catalog" property="product"

arg='<%= request.getParameter("id") %>'

className="com.ora.jsp.beans.shopping.ProductBean" />

<jsp:setProperty name="cart" property="product"

value="<%= product %>" />

<%-- Redirect back to the catalog page --%>

<ora:redirect page="catalog.jsp" />

Since this is a request processing page, it doesn't contain any HTML. The <jsp:useBean> actions locate the CatalogBean and CartBean and associate them with the variables catalog and cart, respectively. Next, the <ora:useProperty> action gets the ProductBean corresponding to the id

request parameter value and associates it with the variable named product. A standard

<jsp:setProperty>

action adds a reference to the product in the cart bean. Once all this is done, the application needs to redisplay the catalog page.

8.4.3 Redirect Versus Forward

There are two ways you can invoke another page: redirecting or forwarding. Forwarding is used in

Example 8.2

to display an appropriate page depending on the result of the user input validation. In

Example 8.9

, redirection is used to display the main page for the application after adding a new product

to the cart, using the <ora:redirect> custom action described in Table 8.11

.

Table 8.11, Attributes for <ora:redirect>

Attribute Name Java Type Request-Time Value Accepted Description

page String

Yes

Mandatory. The URL of the page to redirect to, relative to the current page or, if it starts with a /, relative to the context path.

The <ora:redirect> action sends a redirect response to the client with the new location defined by the page attribute. If URL rewriting is used for session tracking, the URL is encoded with the session

ID. If the body of this action contains <ora:param> actions, described in Table 8.5

, each parameter is

added to the URL as a query string parameter, encoded according to rules in the HTTP specification.

There's an important difference between a forward and a redirect. When you forward, the target page is invoked through an internal method call by the JSP container; the new page continues to process the same request and the browser is not aware that more than one page is involved. A redirect, on the other hand, means that the first page tells the browser to make a new request to the target page. The URL shown in the browser therefore changes to the URL of the new page when you redirect, but stays unchanged when you use forward. A redirect is slower than a forward, since the browser has to make a new request. Also, because it results in a new request, request scope objects are no longer available after a redirect.

So how do you decide if you should use forward or redirect? To a large extent it's a matter of preference.

I look at it like this: forwarding is always faster, so that's the first choice. But since the URL in the browser refers to the start page even after the forward, I ask myself what happens if the user decides to reload the page (or even just resize the window; this often reloads the page automatically). In this example, the start page is the page that adds an item to the shopping cart. I don't want it to be invoked again on a reload, so I redirect to the page that displays the catalog and shopping cart content instead.

© 2001, O'Reilly & Associates, Inc.

8.5 Memory Usage Considerations

You should be aware that all objects you save in the application and session scopes take up memory in the server process. It's easy to calculate how much memory is used for application objects since you have full control over the number of objects you place there. But the total number of objects in the session scope depends on the number of concurrent sessions, so in addition to the size of each object, you also need to know how many concurrent users you have and how long a session lasts. Let's look at an example.

The CartBean used in this chapter is small. It stores only references to ProductBean instances, not copies of the beans. An object reference in Java is 8 bytes, so with three products in the cart we need 24 bytes. The java.util.Vector object used to hold the references adds some overhead, say

32 bytes. All in all, we need 56 bytes per shopping cart bean with three products.

If this site has a modest number of customers, you may have 10 users shopping per hour. The default timeout for a session is 30 minutes, so let's say that at any given moment, you have 10 active users and another 10 sessions that are not active but have not timed out yet. This gives a total of 20 sessions times

56 bytes per session, a total of 1,120 bytes. In other words, a bit more than 1 KB. That's nothing to worry about.

Now let's say your site becomes extremely popular, with 2,000 customers per hour. Using the same method to calculate the number of concurrent sessions, you now have 4,000 sessions at 56 bytes, a total of roughly 220 KB—still nothing to worry about. However, if you store larger objects in each session, for instance the results of a database search, with an average of 10 KB per active session, that corresponds to roughly 40 MB for 4,000 sessions. A lot more, but still not extreme, at least not for a site intended to handle this amount of traffic. However, it should become apparent that with that many users, you have to be a bit more careful with how you use the session scope.

Here are some things you can do to keep the memory requirements under control:

Place only those objects that really need to be unique for each session in the session scope. In the shopping cart example, for instance, each cart contains references only to the shared product beans, and the catalog bean is shared by all users.

Set the timeout period for sessions to a lower value than the default. If you know it's rare that your users leave the site for 30 minutes and then return, use a shorter period. You can change the timeout for all sessions in an application through the application's Deployment Descriptor

(see

Appendix D

), or call session.setMax-InactiveInterval( ) (see Appendix B

) to change it for an individual session.

Provide a way to end the session explicitly. A good example is a logout function. Another possibility is to invalidate the session when something is completed (such as submitting the order form). You can use the session.invalidate( ) method to invalidate a session and make all objects available for garbage collection (the term used when the Java runtime is allowed to remove unused objects to conserve memory). You will see an example of this in

Chapter 10

.

We have covered a lot of ground in this chapter, so let's recap the key points.

The scope concept gives you full control over the lifetime and reach of shared information at a convenient abstraction level. However, be careful about designing your beans for thread safety if they are to be used in the session and application scope, and resist the temptation to keep too much information around in the session scope.

Action elements for passing control between pages, such as the standard <jsp:forward> action and the custom <ora:redirect> action, allow you to allocate different roles to different pages. Other actions, such as the <ora:loop> and <ora:encodeURL> custom actions, can be used to minimize the amount of scripting code needed in the JSP pages. The scope abstraction and the actions together make it possible to develop JSP-based applications that are easy to maintain and extend.

© 2001, O'Reilly & Associates, Inc.

9.1 Accessing a Database from a JSP Page

First, the bad news: JSP 1.1 doesn't specify a standard way to access databases from a JSP page. As I mentioned in

Chapter 8

, work is underway to define a larger set of standard JSP action elements, and actions for database access are high on the priority list.

The good news is that the JDBC API allows Java applications to access databases in a vendorindependent way. You could use JDBC directly in your JSP pages, embedding code in scriptlet elements. But this quickly gets out of hand, leading to too much code in the pages, minimal amount of reuse, and, in general, a web application that's hard to maintain. A better approach is to develop a set of custom action elements based on JDBC. That's what I have done here, and in this chapter we look at how to use them in an employee register application. If you're a programmer and interested in how they

are implemented, skip ahead and glance at Chapter 16 and

Chapter 17

.

Chapter 16 , describes how to

develop custom actions in general, and Chapter 17

, describes the actual database access custom actions.

The database access custom actions developed for this book provide the following features:

Using a connection pool for better performance and scalability

Supporting queries, updates, and inserts

Handling the most common datatype conversions

Supporting a combination of database operations in one transaction

These custom actions are generic, so you can use them to develop your own database-driven web application. Each action is introduced as it is used in the examples in this chapter. In addition, you can

find a complete description of all the actions in Appendix C

.

9.1.1 Example Application Architecture

In this chapter, we build an employee register application. This application contains functions for adding and changing employee information, as well as for searching for employees. The employee information is stored in a relational database and accessed through the database access custom actions.

The employee registration part of the application contains the pages shown in Figure 9.1

.

Figure 9.1. Employee registration pages

This example looks similar (but not identical) to our example from the previous chapter. The enter.jsp page presents a form where the user enters information about an employee. When the form is submitted, it invokes the validate.jsp page, where all input is validated. If the input is invalid, the request is forwarded back to the enter.jsp page to display an error message and the form with all the values the user previously entered. The user can then correct the invalid values and submit the form again. When all input is valid, the validate.jsp page forwards the request to the store.jsp page, where the information is stored in the database. Finally, the store.jsp page redirects to the confirmation.jsp page, which displays the information actually stored in the database as a confirmation to the user.

Figure 9.2

shows the pages used to implement the employee search function.

Figure 9.2. Employee search pages

The search.html page is a regular HTML page with a form for entering the search criteria. The user can enter a partial first name, last name, and department name. Submitting the form invokes the find.jsp page. Here the database is searched for employees matching the criteria specified by the user, and the result is kept in the request scope. The find.jsp page forwards to the list.jsp page, where the result is displayed. For each employee listed, the list.jsp page adds a Delete button. Clicking on the Delete button invokes the delete.jsp page, removing the employee information from the database. The

delete.jsp then redirects to the find.jsp page to get an updated collection of employees matching the search criteria, and the find.jsp forwards to list.jsp as before, to show the result after deleting the employee.

9.1.2 Example Tables

If you develop a database-driven web application from scratch, you must first develop a database schema. The database schema shows how all persistent information in the application is modeled as a set of related tables. For a large application this is a great deal of work, and it's extremely important to find the right balance between flexibility and performance of frequent queries. How database schemas are developed is beyond the scope of this book, but there are plenty of other books available on this subject. Examples are C. J. Date's classic, very academic An Introduction to Database Systems

(Addison Wesley), and a book that's easier to read, Database Design for Mere Mortals: A Hands-On

Guide to Relational Database Design by Michael J. Hernandez (Addison Wesley). In the event that you're developing a web interface to an existing database, you're probably relieved of the schema

development, but you should study the schema anyway to make sure you understand how all the tables fit together.

The schema for the examples in this chapter is simple. To store the employee information, we need only the information described in

Table 9.1

.

Column Name

Table 9.1, Employee Database Table

SQL Datatype Primary Key

UserName

CHAR (Text) Yes

Password

CHAR (Text) No

FirstName

CHAR (Text) No

LastName

CHAR (Text) No

Dept

CHAR (Text) No

EmpDate

DATE (Date/Time) No

EmailAddr

CHAR (Text) No

ModDate

TIMESTAMP (Date/Time) No

In a relational database, one column (or a combination of columns) can be marked as a primary key.

The primary key uniquely identifies one specific row in the table; no two rows can have the same primary key. Here we use a column called UserName as the unique primary key for the table. Each employee must therefore be assigned a unique username, just like the username used to log into an operating system. As you will see in

Chapter 10 , the username, combined with the password you also

find in the Employee table, can be used for application-controlled authentication. Assigning unique usernames can, however, be a problem in a web application available to anyone on the Internet.

Therefore, some applications use a numeric code as the unique identifier instead, such as social security number or a generated sequence number. The tables above are only intended as an example of how to work with databases in JSP, so we allow ourselves to keep it simple.

The SQL datatype name within parentheses in Table 9.1

is the name used in the Microsoft Access

product, to help you create the tables in this commonly used database. This is by no means an endorsement of the Access database for a database-driven web site. (In fact, I recommend that you

don't use Access for a real application. It's a product that's intended as a single-user database, and it doesn't work well with the number of accesses typical for a web application.) For a real site, you should use a more robust multiuser database such as Oracle, Sybase, DB2, or Microsoft SQL Server. The only reason I use Access in this book when I refer to a specific product is that it's a database that you may already have installed. It's also easy to use during development of an application. If you don't have a database installed and you're not ready to spend big bucks for one of the products just listed, there are plenty of free or inexpensive databases you can use. Two examples are Lutris Technologies' InstantDB, a pure Java database available at http://www.lutris.com/products/instantDBNews.html

, and T.c.X's mySQL, a popular database that you can use free of charge for most purposes, available at http://www.mysql.com/ .

To run the examples described in this chapter, you must first create the table outlined in

Table 9.1

in your database. How to do this varies between database engines, so consult the documentation for the database engine you use.

9.1.3 Reading and Storing Information in a Database

The first page the user loads to register an employee in the example application is enter.jsp. This page contains a form for entering all information about an employee, shown in

Figure 9.3

.

Figure 9.3. Employee information entry form

The input is validated by the validate.jsp page when the form is submitted. The enter.jsp and

validate.jsp pages are similar to the pages discussed in detail in Chapter 8

and don't access the database. Instead of going through these pages now, let's jump directly to the store.jsp page, where the database access takes place. We'll return to the enter.jsp and validate.jsp pages at the end of this chapter, as they contain some interesting things not related to database access.

Example 9.1

shows the complete store.jsp page. This page first searches the database for information

about an employee with the specified username. If one is found, the database is updated with all the other information about the employee the user entered. Otherwise, a new employee entry is stored in the database. Then all database information about the employee is collected and the request is forwarded to the confirmation.jsp page. Let's look at the complete page first and then discuss the different pieces in detail.

Example 9.1. Database Access Page (store.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<%--

See if the employee is already defined. If not, insert the

info, else update it.

--%>

<ora:sqlQuery id="empDbInfo" dataSource="example">

SELECT * FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

<% if (empDbInfo.size( ) == 0) { %>

<ora:sqlUpdate dataSource="example">

INSERT INTO Employee

(UserName, Password, FirstName, LastName, Dept,

EmpDate, EmailAddr, ModDate)

VALUES(?, ?, ?, ?, ?, ?, ?, ?)

<ora:sqlStringValue param="userName" />

<ora:sqlStringValue param="password" />

<ora:sqlStringValue param="firstName" />

<ora:sqlStringValue param="lastName" />

<ora:sqlStringValue param="dept" />

<ora:sqlDateValue param="empDate" pattern="yyyy-MM-dd" />

<ora:sqlStringValue param="emailAddr" />

<ora:sqlTimestampValue value="<%= new java.util.Date( ) %>" />

</ora:sqlUpdate>

<% } else { %>

<ora:sqlUpdate dataSource="example">

UPDATE Employee

SET Password = ?,

FirstName = ?,

LastName = ?,

Dept = ?,

EmpDate = ?,

EmailAddr = ?,

ModDate = ?

WHERE UserName = ?

<ora:sqlStringValue param="password" />

<ora:sqlStringValue param="firstName" />

<ora:sqlStringValue param="lastName" />

<ora:sqlStringValue param="dept" />

<ora:sqlDateValue param="empDate" pattern="yyyy-MM-dd" />

<ora:sqlStringValue param="emailAddr" />

<ora:sqlTimestampValue value="<%= new java.util.Date( ) %>" />

<ora:sqlStringValue param="userName" />

</ora:sqlUpdate>

<% } %>

<%-- Get the new or updated data from the database --%>

<ora:sqlQuery id="newEmpDbInfo" dataSource="example" scope="session">

SELECT * FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

<%-- Redirect to the confirmation page --%>

<ora:redirect page="confirmation.jsp" />

At the top of the page in

Example 9.1

you find the taglib directive for the custom action library, as in the previous examples. Then follows a number of database custom actions.

9.1.3.1 JDBC drivers and the DataSource class

The first database custom action you see in

Example 9.1

is the <ora:useDataSource> action

(described in

Table 9.2

):

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver" url="jdbc:odbc:example" />

Table 9.2, Attributes for <ora:useDataSource>

Attribute Name Java Type Request-Time Value Accepted Description

id className String

No url user

String

No

String

No

String

No

Mandatory. The name used to reference the data source from other actions.

Mandatory. The name of the JDBC driver class used to access the database.

Mandatory. The JDBC URL for the database.

Optional. The database user account name.

pw String

No

Optional. The password for the database user account name.

This action looks for a javax.sql.DataSource object with the name specified by the id attribute in the application scope. If it doesn't find it, it creates one for the JDBC driver specified by the class attribute, associates it with the JDBC URL specified by the url attribute, then saves it in the application scope where it will be found the next time around. Before we continue with the rest of the page and the other database custom actions, let's review the DataSource, JDBC driver, and JDBC

URL in more detail.

The DataSource class is defined by the JDBC 2.0 Standard Extension. It represents a data source that can be accessed through the JDBC API. The JDBC API is a set of classes and interfaces that allows a Java application to send SQL statements to a database in a vendor-independent way. For each type of database, an implementation of the interfaces defined by the JDBC API is needed. This is called a JDBC driver. Using different drivers that all provide the same interface allows you to develop your application on one platform (for instance, a PC with an Access database), and then deploy the application on another platform (for instance, a Solaris server with an Oracle database).

At least in theory it does. SQL is unfortunately one of these standards that leave a few things open, eagerly filled by different vendors' proprietary solutions. Examples are how to handle embedded quotes

in a string value, how to deal with the input and output of date and time values, semantics for certain datatypes, and creation of unique numbers. The custom actions used in this book take care of some of this, such as string quoting and date/time string format, so if you use these actions and stick to ANSI

SQL you should be able to migrate from one database to another without too much tweaking. However, you should always read your database documentation carefully and try to stay away from proprietary features. And be prepared to spend at least some time in transition when you need to move the application to another database.

All other database custom actions in the example tag library use the DataSource to get a database connection for executing the SQL statement. One nice thing with a DataSource is that it can represent something called a connection pool . This is described in more detail in

Chapter 17 , but a

connection pool is exactly what it sounds like: a pool of database connections that can be shared by multiple clients. Opening a database connection is very time-consuming. With a connection pool, a connection to the database is opened once and stays open until the application is shut down. When a database custom action needs a connection, it gets it from the pool, through the DataSource object, and uses it to execute one or more SQL statements. When the action closes the connection, the connection is returned to the pool where it can be picked up by the next action that needs it. The

DataSource

created by the <ora:useDataSource> action implements a basic connection pool.

Back to our custom action. The <ora:useDataSource> action has three mandatory attributes: id, className

, and url. Optionally you can also specify user and pw attributes, required to connect to some databases. The id attribute defines the name used for the DataSource object in the application scope. The className and url attributes require a bit more explanation.

As I mentioned earlier, what makes it possible to access databases from different vendors through the standard JDBC API is that JDBC relies on drivers, written for each specific database engine. A driver converts the JDBC API methods to the proprietary equivalents for a specific database engine. You can find JDBC drivers for most database engines on the market, both commercial and open source. If you can't get one from your vendor, Sun has a list of JDBC drivers from third parties at http://industry.java.sun.com/products/jdbc/drivers/ .

The class attribute is used to specify the JDBC driver classname, for instance sun.jdbc.odbc.JdbcOdbcDriver

. It must be specified as a fully qualified classname, i.e., it must include the package name. In this example we use the JDBC-ODBC bridge driver included in the

Java SDK. This driver can be used to access databases that provide an ODBC interface but have no direct JDBC driver interface, as is the case for Microsoft Access. Sun doesn't recommend that you use the JDBC-ODBC driver for a production application, but for development it works fine. When you deploy your application, you should use production-quality drivers, available from the database vendor or a third party.

A database is identified by a JDBC URL. Different JDBC drivers use different URL syntax. All JDBC

URLs start with jdbc:, followed by a JDBC driver identifier, such as odbc:, for the JDBC-ODBC bridge driver. The rest of the URL is used to identify the database instance. For the JDBC-ODBC bridge driver, it's an ODBC Data Source Name (DSN). If you use an Access database, you need to create a System DSN using the ODBC control in the Windows Control Panel to run this example, as

shown in Figure 9.4

. Note that you must create a System DSN as opposed to a User DSN. The reason

for this is that the web server where your JSP pages are executed usually runs as a different user account than the account you use for development. If you specify a User DSN with your development account, the web server's servlet container will not be able to find it.

If you use a different JDBC driver than the JDBC-ODBC bridge driver or use a different ODBC DSN, modify the <ora:useDataSource> attributes in store.jsp accordingly before you try to run the example.

Figure 9.4. System DSN definition window

The <ora:useDataSource> action is intended only for simple examples or during the prototyping phase in a real project. From a maintenance standpoint, it's not a good idea to have the JDBC URL and driver classname in multiple pages. Also, if you need to specify a username and password, a JSP page is not a secure place to put this information. Another reason is that in JSP 1.1, there's no way for a custom action to know when an application is being shut down. This means a custom tag can't gracefully shut down the connections in the pool, potentially leading to problems with database resources not being released as they should.

Instead of using the <ora:useDataSource> action, you should use a servlet that's loaded when the application is started and notified when it's being shut down. This solves all of these problems, and we'll look at such a servlet in

Chapter 17 . As more and more database and JDBC driver vendors add

support for JDBC 2.0 SE and implement their own connection pools, the servlet approach also lets you use a connection pool that's potentially more efficient than the one created by the custom action. No matter how the data source is created, other database custom actions described in this chapter work the same, since they just need the name the DataSource is saved under in the application scope.

9.1.3.2 Reading database information

Now that we've connected to a data source, we can begin to send queries to it. The first SQL custom

action that accesses the database in Example 9.1

is the <ora:sqlQuery> action, described in

Table

9.3

.

Table 9.3, Attributes for <ora:sqlQuery>

Attribute Name Java Type

Request-Time Value

Accepted

Description

id String

No Mandatory. The name of the bean to hold the result.

dataSource String

No scope String

No

Mandatory, unless used with <ora:sqlTransaction>. The name of the data source.

Optional. The scope for the result, one of page, request, session

, or application. Default is page.

The <ora:sqlQuery> action is used to read information from a database using the SQL SELECT statement specified in the element's body. A SELECT statement selects data from the database. It does this by specifying various clauses that identify the table to search, the columns to return, the search criteria, and other options. If you're not familiar with the SELECT statement, you can read up on it in the documentation for your database. The SELECT statement in

Example 9.1

gets all columns in the

Employee table for every row where the UserName column has the value specified in the userName field in the entry form. Since the username is unique in our application, either 0 or 1 row is returned.

The <ora:sqlQuery> action gets a connection from the data source identified by the dataSource

attribute. It then executes the SQL SELECT statement in the action's body, and saves the result as a java.util.Vector with com.ora.jsp.sql.Row objects in the scope specified by the scope attribute, using the name specified by the id attribute. If no scope is specified, as in this example, the result is saved in the page scope. The dataSource attribute value must be the name of a DataSource available in the application scope. Note how it matches the id attribute of the

<ora:useDataSource>

action in Example 9.1

.

Besides the SQL statement, the action element body also contains a <ora:sqlStringValue> action, described in

Table 9.4

.

Table 9.4, Attributes for <ora:sqlStringValue>

Attribute Name Java Type Request-Time Value Accepted Description

value param name property String

No prefix suffix

String

Yes

String

Yes

String

No

String

Yes

String

Yes

Optional. The value to use for a placeholder in the enclosing database action.

Optional. The name of the request parameter holding the value.

Optional. The name of the bean with a property holding the value.

Mandatory if name is specified. The name of the bean property holding the value.

Optional. A string that should be concatenated to the beginning of the value.

Optional. A string that should be concatenated to the end of the value.

The <ora:sqlStringValue> action replaces a placeholder, marked with a question mark (?), in the SQL statement with a value. The value can be specified in one of three ways:

Using the value attribute to specify the value as a literal string or as a request-time attribute that returns a String:

<ora:sqlStringValue value='<%= anObject.getString("Arg") %>' />

Using the param attribute to specify the name of a request parameter that holds the String value:

<ora:sqlStringValue param="aParameterName" />

Using the name and property attributes to specify a bean property that holds the String

value:

<ora:sqlStringValue name="aBeanName" property="aPropertyName" />

In Example 9.1

, the param attribute is used to get the userName request parameter value, corresponding to the form field with the same name in the enter.jsp page:

<ora:sqlQuery id="empDbInfo" dataSource="example">

SELECT * FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

You could use a JSP expression in the body instead to insert the username directly into the SQL statement, like this:

<ora:sqlQuery id=" empDbInfo " dataSource="example">

SELECT * FROM Employee

WHERE UserName = '<%= request.getParameter("userName") %>'

</ora:sqlQuery>

But then you run into the problem of string quoting in SQL. Most database engines require a string literal to be enclosed in single quotes in a SQL statement. That's easy to handle by just putting single quotes around the JSP expression, like I've done in this example. What's not so easy is how to handle quotes within the string value. Different database engines employ different rules for how to encode embedded quotes. Most require a single quote in a string literal to be duplicated, while others use a backslash as an escape character or let you enclose the string literal with double quotes if the value includes single quotes. When you use the <ora:sqlStringValue> action, you don't have to worry about this type of formatting at all; the value is encoded according to the rules for the database you're currently accessing.

The <ora:sqlQuery> element body can contain multiple placeholders and

<ora:sqlStringValue>

actions. The first <ora:sqlStringValue> action replaces the first question mark in the SQL statement with its value, formatted correctly for the database engine you are currently using, the second replaces the second question mark, and so on. Only one dynamic value is

needed in the query in Example 9.1

. If you need more, just add question marks in the SQL statements

and <ora:sqlStringValue> actions in the <ora:sqlQuery> body in the same order.

Back to the result generated by the <ora:sqlQuery> action. As I mentioned earlier, it's a java.util.Vector

with com.ora.jsp.sql.Row objects. A Vector is like a dynamic array that provides methods for accessing its elements: one by one with the elementAt( ) ,

firstElement( )

, and lastElement( ) methods, or as a java.util.Enumeration of all elements with the elements( ) method. The Java API documents contain a complete list of the

Vector

methods. We look at the Row class later in this chapter.

The Vector class also provides a method used in Example 9.1

to see if the query returned any rows at all: the size( ) method. The SELECT statement searches the database for information about the employee entered in the form. If the employee is already registered, the query will return one row. To figure out whether or not the database already contains information about the employee, you can use the size( ) method to test on the number of rows using a simple scriptlet with an if statement:

<% if (empDbInfo.size( ) == 0) { %>

insert

<% } else {

update

<% } %>

9.1.3.3 Inserting database information

Now that you know how to retrieve data from the database, we can move on to inserting information.

You can insert information into a database with a SQL INSERT statement. To execute an INSERT statement, use the <ora:sqlUpdate> custom action, described in

Table 9.5

.

Table 9.5, Attributes for <ora:sqlUpdate>

Attribute Name Java Type

Request-Time Value

Description

Accepted

id dataSource String

No scope

String

No

String

No

Optional. The name of an Integer object to hold the number of rows affected by the statement.

Mandatory, unless used with <ora:sqlTransaction>. The name of the data source.

Optional. The scope for the result, one of page, request, session

, or application. Default is page.

The <ora:sqlUpdate> action executes any SQL statement that doesn't return rows: INSERT,

UPDATE

, DELETE, and even so-called Data Definition Language (DDL) statements such as CREATE

TABLE

. These statements do exactly what they sound like they do: insert, update, and delete

information, and create a new table, respectively. (Refer to your database documentation for details about the syntax.) For INSERT, UPDATE, and DELETE, the <ora:sqlUpdate> action can optionally save an Integer object, telling how many rows were affected by the statement. The

Integer

is saved in the scope specified by the scope attribute using the name specified by the id attribute. This feature is not used in

Example 9.1

, but in some applications it can be used as feedback to

the user or to decide what to do next.

The only mandatory attribute for <ora:sqlUpdate> is dataSource. It must be the name of a

DataSource

available in the application scope, in the same way as for the <ora:sqlQuery> action. The <ora:sqlStringValue> action can be used in the body of an <ora:sqlUpdate>

element as well. Multiple <ora:sqlStringValue> actions are used in Example 9.1

, each one

setting the value for a placeholder to the value of a request parameter:

<ora:sqlUpdate dataSource="example">

INSERT INTO Employee

(UserName, Password, FirstName, LastName, Dept,

EmpDate, EmailAddr, ModDate)

VALUES(?, ?, ?, ?, ?, ?, ?, ?)

<ora:sqlStringValue param="userName" />

<ora:sqlStringValue param="password" />

<ora:sqlStringValue param="firstName" />

<ora:sqlStringValue param="lastName" />

<ora:sqlStringValue param="dept" />

<ora:sqlDateValue param="empDate" pattern="yyyy-MM-dd" />

<ora:sqlStringValue param="emailAddr" />

<ora:sqlTimestampValue value="<%= new java.util.Date( ) %>" />

</ora:sqlUpdate>

Besides <ora:sqlStringValue> actions, an <ora:sqlDateValue> action is used to set the

EmpDate

column value. This action is described in

Table 9.6

.

Attribute Name Java Type

Table 9.6, Attributes for <ora:sqlDateValue>

Request-Time Value

Accepted

Description

value java.util.Date Yes stringValue String

Yes

Optional. The value to use for a placeholder in the enclosing database action.

Optional. The String to use as the value.

param name

String

String property String pattern String

Yes

No

No

Yes

Optional. The name of the request parameter holding the value.

Optional. The name of the bean with a property holding the value.

Mandatory if name is specified. The name of the bean property holding the value.

The pattern used to interpret a String specified by the stringValue, param, or name/property

attributes.

Databases are picky about the format for date and time datatypes. In Example 9.1

we get the date from

the form as a string in the format yyyy-MM-dd (e.g., 2000-03-31), but the EmpDate column is

declared as a DATE column, as you can see in Table 9.1

. Some databases accept a string in the format

used here as a value for a DATE column, but others do not. To be on the safe side, it's best to convert the string into its native date format, a java.sql.Date object, before sending it to the database. The

<ora:sqlDateValue>

action does this for you when you specify the value using the stringValue

, param, or name and property attributes (if the property is of type String).

The pattern attribute contains a pattern that describes the order and format of the year, month, and

day parts in the string representation of the date. Appendix C

shows all the types of patterns you can use—it's very flexible. Say, for instance, that you want the user to enter dates in a format like "Friday

July 14, 2000." You then use the pattern EEEE MMMM dd, yyyy instead of the yyyy-MM-dd pattern used in this example to tell the <ora:sqlDateValue> action how to interpret the date string.

The Employee table also has a column named ModDate, to hold the date and time the information was last modified. It is declared as a TIMESTAMP column. To set its value, an instance of the java.util.Date

class is used as the value attribute of a <ora:sqlTimestampValue> action

in Example 9.1

:

<ora:sqlTimestampValue value="<%= new java.util.Date( ) %>" />

Besides the SQL value actions for dates and timestamps described here, the custom library contains similar value actions for columns declared as TIME as well as all numeric datatypes. They all support the same attributes as the <ora:sqlDateValue> action and perform the conversion from a string value to the appropriate native type.

Appendix C contains information about all of these actions.

9.1.3.4 Updating database information

Once you know how to insert information in a database, updating it is a piece of cake. You just use the

<ora:sqlUpdate>

action with a SQL UPDATE statement instead of an INSERT statement:

<ora:sqlUpdate dataSource="example">

UPDATE Employee

SET Password = ?,

FirstName = ?,

LastName = ?,

Dept = ?,

EmpDate = ?,

EmailAddr = ?,

ModDate = ?

WHERE UserName = ?

<ora:sqlStringValue param="password" />

<ora:sqlStringValue param="firstName" />

<ora:sqlStringValue param="lastName" />

<ora:sqlStringValue param="dept" />

<ora:sqlDateValue param="empDate" pattern="yyyy-MM-dd" />

<ora:sqlStringValue param="emailAddr" />

<ora:sqlTimestampValue value="<%= new java.util.Date( ) %>" />

<ora:sqlStringValue param="userName" />

</ora:sqlUpdate>

No surprises here. The only difference from how you insert information is the SQL statement. The

UPDATE

statement sets all the specified values for rows matching the WHERE clause, in this case the single row for the specified employee.

9.1.4 Generating HTML from a Query Result

Just before the page in

Example 9.1

redirects to the confirmation page, there's one more

<ora:sqlQuery>

action that retrieves the employee information that was just stored in the database:

<ora:sqlQuery id="newEmpDbInfo" dataSource="example" scope="session">

SELECT * FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

The intention here is to present the information actually stored in the database to the user on the final

page in this application, shown in Figure 9.5

, as a confirmation that the operation was successful.

Figure 9.5. Employee registration confirmation page

Since we redirect to the confirmation page, ending the processing of the current request, the result is placed in the session scope. The redirect response tells the browser to automatically make a new request for the confirmation page. Because the new request is part of the same session, it finds the result saved away by the previous page.

Example 9.2

shows the code for the confirmation.jsp page.

Example 9.2. Page Displaying Query Result (confirmation.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="com.ora.jsp.sql.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Employee Info Stored</title>

</head>

<body bgcolor="white">

This is the information stored in the employee database:

<table>

<ora:loop name="newEmpDbInfo" loopId="row" className="Row" >

<ora:loop name="row" property="columns" loopId="col"

className="Column" >

<tr>

<td align=right>

<b><jsp:getProperty name="col" property="name" />:</b>

</td>

<td>

<jsp:getProperty name="col" property="string" />

</td>

</tr>

</ora:loop>

</ora:loop>

</table>

</body>

</html>

At the top of the page is the same JSP directive for using the custom action tag library as in

Example

9.1

, plus a page directive that imports the com.ora.jsp.sql package. This package contains two

classes used to represent rows and columns in the query result, named Row and Column, respectively.

You'll see how they are used in a moment.

Then comes a loop to create an HTML table with cells for all columns in the single row retrieved from the Employee table. The newEmpDbInfo variable contains the Vector of Row objects returned by

the <ora:sqlQuery> action in Example 9.1

. So far, we have used the <ora:loop> action only to loop through arrays, but it can also be used to loop through multiple values contained by a java.util.Vector

, a java.util.Enumeration, or a java.util.Dictionary. Arrays and these three classes are all commonly used in Java to represent a set of values. You can read more about them in the Java API specification ( http://java.sun.com/docs/index.html

).

Inside the <ora:loop> action used to loop through the Vector of Row objects, another nested

<ora:loop>

action is used to loop through the columns of each Row. In the outer loop, the loopId attribute is set to row, which means that for each pass through the loop, a variable named row contains the current Row element of the Vector. In the inner loop, the row variable is used as the name attribute, and the property attribute is set to columns. The columns property of a Row object is an Enumeration of Column objects. A Column object, in turn, has properties name and string .

The name property is a String containing the column's name, and the string property is a

String

representation of the value independent of its real datatype. The Column has other properties

that let you access the value in its native form as well; see Appendix C

for a complete list. In the body of the inner loop, the standard <jsp:getProperty> action is used to create cells with the column name and value.

9.1.5 Searching for Rows Based on Partial Information

Let's move on to the other part of the application, where a user can search for an employee based on a partial first name, last name, and department name. The first page, search.html, contains a form for

entering the search criteria, shown in Figure 9.6

.

Figure 9.6. Search criteria form

The three fields in the search.html page are firstName, lastName, and dept, and when the user clicks the Search button, the find.jsp page is invoked with the information the user entered in the

corresponding request parameters. Example 9.3

shows the complete find.jsp page.

Example 9.3. Search Based on Partial Information (find.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<ora:sqlQuery id="empList" dataSource="example" scope="request">

SELECT * FROM Employee

WHERE FirstName LIKE ?

AND LastName LIKE ?

AND Dept LIKE ?

ORDER BY LastName

<ora:sqlStringValue param="firstName" prefix="%" suffix="%" />

<ora:sqlStringValue param="lastName" prefix="%" suffix="%" />

<ora:sqlStringValue param="dept" prefix="%" suffix="%" />

</ora:sqlQuery>

<jsp:forward page="list.jsp" />

As you probably expected, the <ora:sqlQuery> action is used to search for the matching employees. But here, the SELECT statement uses the LIKE operator to find rows matching a pattern instead of an exact match condition. LIKE is a standard SQL operator. It must be followed by a string consisting of fixed text plus wildcard characters. There are two standard wildcard characters you can use: an underscore (_), which matches exactly one character, and a percent sign (%), which matches zero or more characters. In this example, we want to search for all rows that contain the values specified in the form somewhere in the corresponding column values. The form field values must therefore be enclosed with percent signs. This can be accomplished by using the

<ora:sqlStringValue>

action's prefix and suffix attributes, described in

Table 9.4

. The

values of these attributes—in this case, percent signs—are added at the beginning and the end of the value, respectively. If you want to find values that start with any sequence of characters but end with the string entered by the user, use only the prefix attribute. If you use only the suffix attribute, you get the reverse result: values that start with the specified string but end with any characters.

The three LIKE conditions are combined with AND operators in

Example 9.3

. This means that the

SELECT

statement finds only rows where all three columns contain the corresponding values entered by the user.

9.1.6 Deleting Database Information

The find.jsp page forwards the request to the list.jsp page to display the result of the search. It generates

an HTML table with one row per employee, as shown in Example 9.4

.

Example 9.4. Displaying the Search Result (list.jsp)

...

<table border=1>

<th>Last Name</th>

<th>First Name</th>

<th>Department</th>

<th>Email Address</th>

<th>Modified</th>

<ora:loop name="empList" loopId="row"

className="com.ora.jsp.sql.Row" >

<tr>

<td><%= row.getString("LastName") %></td>

<td><%= row.getString("FirstName") %></td>

<td><%= row.getString("Dept") %></td>

<td><%= row.getString("EmailAddr") %></td>

<td><%= row.getString("ModDate") %></td>

<td>

<form action="delete.jsp" method="post">

<input type="hidden" name="userName"

value='<%= StringFormat.toHTMLString(

row.getString("UserName")) %>'>

<input type="hidden" name="firstName"

value='<%= StringFormat.toHTMLString(

request.getParameter("firstName")) %>'>

<input type="hidden" name="lastName"

value='<%= StringFormat.toHTMLString(

request.getParameter("lastName")) %>'>

<input type="hidden" name="dept"

value='<%= StringFormat.toHTMLString(

request.getParameter("dept")) %>'>

<input type="submit" value="Delete">

</form>

</td>

</tr>

</ora:loop>

</table>

...

The result is shown in Figure 9.7

.

Figure 9.7. Displaying the search result

An <ora:loop> action is used to loop over all rows returned by the query in Example 9.3

. The

loopId

attribute is set to row, so for each pass through the action body, a variable named row holds a reference to the current Row object. For each row, a number of table cells are generated. The value of the cell is retrieved from the Row using the getString( ) method: public String getString(String columnName)

The argument to the method is the name of a column. The method returns the column's value as a

String

. Another version of this method takes an index number instead of a column name. The first column has index 1: public String getString(int columnIndex)

The Row class also provides methods to get the column value in its native form, for instance as a

Date

: public java.sql.Date getDate(String columnName) public java.sql.Date getDate(int columnIndex)

You may want to use these methods if you use one query result as input to another query. Again, see

Appendix C for a description of all the methods.

The last generated cell contains a simple HTML form with a Delete button and a number of hidden fields. The action for the form is set to invoke the delete.jsp page. The hidden fields hold the value of

UserName

for the current row, plus all the parameters used to perform the search. All hidden field values are encoded using the same StringFormat toHTMLString( ) method we used in

Chapter 6 , to make sure that quotes in the value don't cause syntax errors in the generated HTML.

Example 9.5

shows how all these parameters are used in the delete.jsp page.

Example 9.5. Deleting a Row (delete.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<ora:sqlUpdate dataSource="example">

DELETE FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlUpdate>

<ora:redirect page="find.jsp">

<ora:param name="firstName"

value='<%= request.getParameter("firstName") %>' />

<ora:param name="lastName"

value='<%= request.getParameter("lastName") %>' />

<ora:param name="dept"

value='<%= request.getParameter("dept") %>' />

</ora:redirect>

The userName request parameter value is used to uniquely identify the row to remove. The SQL

DELETE

statement supports the same type of WHERE clause condition you have seen used in SELECT and UPDATE statements previously. Here, the condition is used to make sure only the row for the right employee is deleted. And like the INSERT and UPDATE statements, a DELETE statement is executed with the help of the <ora:sqlUpdate> action.

The other parameters passed from the list.jsp page are used in the redirect call to the find.jsp page. This way, the find.jsp page uses the same search criteria as when it was called directly from the search.html file, so the new result is consistent with the first. The only difference is that the employee who was just deleted doesn't show up in the list.

© 2001, O'Reilly & Associates, Inc.

9.2 Input Validation Without a Bean

Before we look at the two remaining database sections, let's go back and take a look at the two application pages we skipped earlier, namely the enter.jsp and validate.jsp pages used for input to the employee registration.

In Chapter 5

, I introduced you to validation of user input using an application-specific bean. The bean contains all validation code and provides an isValid( ) method that can be used in a JSP page to decide how to proceed. This is the approach I recommend, but if you're developing a JSP-based application and there isn't a Java programmer around, there's another way to do the validation. I'll describe this alternative here.

The validate.jsp page uses the StringFormat utility class to validate the input format without a bean. If an input parameter is not valid, an error message is saved in a Vector object and the request is forwarded back to the enter.jsp page. The enter.jsp page loops through all error messages in the

Vector

and adds them to the response, so to the user, the result is identical to that of the bean-based

validation approach you saw in Chapter 5

.

Let's look at validate.jsp first, shown in Example 9.6

.

Example 9.6. Validation Without Application Beans (validate.jsp)

<%@ page language="java" %>

<%@ page import="com.ora.jsp.util.*" %>

<jsp:useBean id="errorMessages" scope="request"

class="java.util.Vector" />

<%

if (request.getParameter("userName").length( ) == 0) {

errorMessages.addElement("User Name missing");

}

if (request.getParameter("password").length( ) == 0) {

errorMessages.addElement("Password missing");

}

if (request.getParameter("firstName").length( ) == 0) {

errorMessages.addElement("First Name missing");

}

if (request.getParameter("lastName").length( ) == 0) {

errorMessages.addElement("Last Name missing");

}

if (request.getParameter("dept").length( ) == 0) {

errorMessages.addElement("Department missing");

}

if (!StringFormat.isValidDate(request.getParameter("empDate"),

"yyyy-MM-dd")) {

errorMessages.addElement("Invalid Employment Date");

}

if (!StringFormat.isValidEmailAddr(

request.getParameter("emailAddr"))) {

errorMessages.addElement("Invalid Email Address");

}

if (errorMessages.size( ) > 0) {

%>

<jsp:forward page="enter.jsp" />

<% } else { %>

<jsp:forward page="store.jsp" />

<% } %>

At the top of Example 9.6

, a <jsp:useBean> action creates a Vector instance in the request scope

to hold possible error messages. Even though the Vector class doesn't provide bean getter and setter methods, an instance of this class can be created by the <jsp:useBean> action (because it has a noargument constructor).

Next comes a scriptlet with an if statement for each input parameter that needs to be validated. For most of them, it's enough to verify that the parameter has a value. This is done by using the length(

)

method. If the result is 0, the parameter doesn't have a value, and the body of the if block adds an appropriate error message to the errorMessages Vector .

Two parameters require a more careful validation: the empDate parameter must contain a valid date string, and the emailAddr a valid email address: if (!StringFormat.isValidDate(request.getParameter("empDate"),

"yyyy-MM-dd")) {

errorMessages.addElement("Invalid Employment Date");

}

if (!StringFormat.isValidEmailAddr(

request.getParameter("emailAddr"))) {

errorMessages.addElement("Invalid Email Address");

}

The empDate parameter is validated with the StringFormat isValidDate( ) method. This method takes a string representation of a date, here retrieved from the empDate request parameter, and a date format pattern. The date format pattern is the same as for the <ora:sqlDateValue> action discussed earlier. If the date string forms a valid date when interpreted according to the pattern, this method returns true.

The emailAddr parameter is validated with another StringFormat method, called isValidEmailAddr( )

. This method returns true if the string looks like a valid email address, that is, if it has the form name@company.topdomain, for instance, [email protected]

.

After the validation scriptlets, there's another set of scriptlets. The if statement tests the size of the errorMessages

Vector. If it's greater than 0, at least one parameter value is invalid, so the request is forwarded to the enter.jsp page again. Otherwise, the processing continues on the store.jsp page, as discussed in the first section of this chapter.

If the request is forwarded to the enter.jsp page, the error messages are displayed and all the values the user entered are used as the default values for all form fields.

Example 9.7

shows how the error

messages are handled.

Example 9.7. Displaying Error Messages (enter.jsp)

...

<jsp:useBean id="errorMessages" scope="request"

class="java.util.Vector" />

<%-- Output list of values with invalid format, if any --%>

<ul>

<font color="red">

<ora:loop name="errorMessages" loopId="msg" class="String" >

<li> <%= msg %>

</ora:loop>

</font>

</ul>

...

The <jsp:useBean> action finds the Vector created by the validate.jsp page and makes it available through the errorMessages variable. The <ora:loop> action is then used to loop

through the Vector and display each message as a list item. The results are shown in Figure 9.8

.

Figure 9.8. The enter.jsp page

Example 9.8

shows how the form field values are filled out with the values submitted by the user.

Example 9.8. Filling Form Field Values with Request Data

...

<form action="validate.jsp" method="post">

<table>

<tr>

<td>User Name:</td>

<td><input type="text" name="userName"

value='<%= StringFormat.toHTMLString(

request.getParameter("userName")) %>' >

</td>

</tr>

...

This is very similar to how it was done with a bean in Chapter 5

. The difference is that the value is retrieved from the implicit request object instead of a bean. Remember that since the validate.jsp page forwards to the enter.jsp page, the enter.jsp page is processing the same request as validate.jsp; therefore, all request parameters are still available. The request parameter value is formatted with the toHTMLString( )

method as before, to make sure possible quotes in the value don't confuse the browser.

© 2001, O'Reilly & Associates, Inc.

9.3 Using Transactions

There's one important database feature we have not discussed yet. In the examples in this chapter, only one SQL statement is needed to complete all database modifications required for a function. This statement either succeeds or fails. However, it's very common that you need to execute two or more

SQL statements in sequence to update the database. A typical example is transferring money between two accounts: one statement removes some amount from the first account and another statement adds the same amount to the second account. If the first statement is successful but the second fails, you have performed a disappearing act that your customers are not likely to applaud.

The solution to this problem is to group all related SQL statements into what is called a transaction. A transaction is an atomic operation, so if one statement fails, they all fail. Otherwise, they all succeed.

This is referred to as committing (if it succeeds) or rolling back (if it fails) the transaction. If there's a problem in the middle of a money transfer, for instance, the database makes sure the money is returned to the first account by rolling back the transaction. If no problems are encountered, the transaction is committed, permanently storing the changes in the database.

There's a custom action in the book tag library to handle transactions, described in

Table 9.7

.

Table 9.7, Attributes for <ora:sqlTransaction>

Attribute Name Java Type Request-Time Value Accepted Description

dataSource String

No

Mandatory. The name of the data source to use for all nested database access actions.

We will use it for real in Chapter 10 , but let's now take a quick look at how it could be used in this

fictitious example:

<ora:sqlTransaction dataSource="example">

<ora:sqlUpdate>

UPDATE Account SET Balance = Balance - 1000

WHERE AccountNumber = 1234

</ora:sqlUpdate>

<ora:sqlUpdate>

UPDATE Account SET Balance = Balance + 1000

WHERE AccountNumber = 5678

</ora:sqlUpdate>

</ora:sqlTransaction>

All SQL actions that make up a transaction are placed in the body of an <ora:sqlTransaction> action. Note that instead of specifying the dataSource attribute for all database actions, it's specified only in the enclosing <ora:sqlTransaction> action.

The <ora:sqlTransaction> action gets a connection from the data source and makes it available to all database actions within its body. If one of the actions fails, the transaction is rolled back.

Otherwise the transaction is committed at the end of the <ora:sqlTransaction> body.

© 2001, O'Reilly & Associates, Inc.

9.4 Application-Specific Database Actions

You can use the database actions described in this chapter to develop many types of interesting web applications, such as product catalog interfaces, employee directories, or online billboards, without being a Java programmer. These types of applications account for a high percentage of the web applications developed today. But at some level of complexity, putting SQL statements directly in the web pages can become a maintenance problem. The SQL statements represent business logic, and for more complex applications, business logic is better developed as separate Java classes.

Therefore, for a complex application, it may be better to use application-specific custom actions instead of the generic database actions described in this chapter. For example, all of the generic database actions in

Example 9.1

, which SELECT and then INSERT or UPDATE the database, can be replaced

with one application-specific action, like this:

<myLib:saveEmployeeInfo dataSource="example" />

Chapter 16

and Chapter 17 describe how you can develop this type of custom action. Besides making it

easier for the page designer to deal with, the beauty of using an application-specific custom action is that it lets you evolve the application behind the scene. Initially, this action can be implemented so it uses JDBC to access the database directly, similar to how the generic actions work. But at some point it may make sense to migrate the application to an Enterprise JavaBeans architecture, perhaps to support types of clients other than web browsers. The action can then be modified to interact with an Enterprise

JavaBeans component instead of directly accessing the database. From the JSP page developer's point of view, it doesn't matter; the custom action is still used in exactly the same way.

Another approach is to use a servlet for all database processing, and use JSP pages only to show the result. You will find an example of this approach in

Chapter 14

.

© 2001, O'Reilly & Associates, Inc.

10.1 Container-Provided Authentication

A JSP page is always executing in a runtime environment provided by a container. Consequently, all authentication and access control can be handled by the container, relieving the application developer from the important task of implementing an appropriate security level. Security is hard to get right, so your first choice should always be to use the time-tested mechanisms provided by the container.

10.1.1 Authenticating Users

The Servlet 2.2 specification, which is the base for JSP 1.1, describes three authentication mechanisms supported by most web clients and web servers that a container can utilize:

HTTP basic authentication

HTTP digest authentication

HTTPS client authentication

In addition, it specifies one additional mechanism that should be implemented by a compliant servlet container:

Form-based authentication

HTTP basic authentication has been part of the HTTP protocol since the beginning. It's a very simple and not very secure authentication scheme. When a client requests access to a protected resource, the server sends back a response asking for the user's credentials (username and password). The web client prompts the user for this information and sends the same request again, but this time with the user credentials in one of the request headers so the server can authenticate the user. But the username and password are not encrypted, only slightly obfuscated by the well-known Base64 encoding. This means it can easily be reversed by anyone who grabs it as it's passed over the network. This problem can be resolved by using an encrypted connection between the client and the server, such as the Secure Sockets

Layer (SSL) protocol. We talk more about this in the last section of this chapter.

HTTP 1.1 introduced HTTP digest authentication. As with basic authentication, the server sends a response back to the client when it receives a request for a protected resource. But with the response, it also sends a string called a nonce . The nonce is a unique string generated by the server, typically composed of a timestamp, information about the requested resource, and a server identifier. The client creates an MD5 checksum, also known as a message digest, of the username, the password, the given nonce value, the HTTP method, and the requested URI, and sends it back to the server in a new request.

The use of an MD5 message digest means the password cannot easily be extracted from information

recorded from the network. Additionally, using information such as timestamps and resource information in the nonce minimizes the risk of "replay" attacks. The digest authentication is a great improvement over basic authentication. The only problem is that it's not broadly supported in today's web clients and servers.

HTTPS client authentication is the most secure authentication method supported today. This mechanism requires the user to possess a Public Key Certificate (PKC). The certificate is passed to the server when the connection between the client and server is established, using a very secure challenge-response handshake process, and is used by the server to uniquely identify the user. As opposed to the mechanisms described above, the server keeps the information about the client's identity as long as the connection remains open. When the client requests a protected resource, the server uses this information to grant or refuse access.

The three mechanisms described here are defined by Internet standards. They are used for all sorts of web applications, servlet-based or not, and are usually implemented by the web server itself as opposed to the servlet container. The servlet specification only defines how an application can gain access to information about a user authenticated with one of them, as you will see soon.

The final mechanism, form-based authentication, is unique to the servlet specification and is implemented by the servlet container itself. Form-based authentication is about as insecure as basic authentication for the same reason: the user's credentials are sent as clear text over the network. To protect access to sensitive resources, it should be combined with encryption such as SSL.

Unlike basic and digest authentication, form-based authentication lets you control the appearance of the login screen. The login itself is a form containing two mandatory input fields named j_username and j_ password

, and the action attribute is set to the string j_security_check:

<form method="POST" action="j_security_check">

<input type="text" name="j_username">

<input type="password" name="j_password">

</form>

From the user's point of view, it works just like basic and digest authentication. When the user requests a protected resource, the login form is shown, prompting the user to enter a username and password. The j_security_check

action attribute value is a special URI that is recognized by the container. When the user submits the form, the container authenticates the user using the j_username and j _ password

parameter values. If the authentication is successful, it redirects the client to the requested resource. Otherwise an error page is returned. We'll get to how you specify the login page and error page shortly.

10.1.2 Controlling Access to Web Resources

All of the authentication mechanisms described so far rely on two pieces of information: user definitions and information about the type of access control needed for the web application resources.

How users and groups of users are defined depends on the server you're using. Some web servers, such as Microsoft's Internet Information Server (IIS), can use the operating system's user and group definitions. Others, such as the iPlanet Web Server (formerly Netscape Enterprise Server), let you use its own user directory or an external LDAP server. The security mechanism defined by the servlet specification describes how to specify the access control constraints for a web application, but access is granted to a role instead of directly to a user or a group. A role is an abstract grouping of users that needs to be mapped to real user and group names for a particular server. How the mapping is done also depends on the server, so you need to consult your web server and servlet container documentation if you use a different server than Tomcat.

The Tomcat server uses a simple XML file to define users and assign them roles at the same time. The file is named tomcat-users.xml and is located in the conf directory. To run the examples in this chapter you need to define at least two users and assign one of them the role admin and the other the role user

, like this:

<tomcat-users>

<user name="paula" password="boss" roles="admin" />

<user name="hans" password="secret" roles="user" />

</tomcat-users>

Here the user paula is assigned the admin role and hans is assigned the user role.

The type of access control that should be enforced for a web application resource, such as a JSP page or all files in a directory, is defined in the web application deployment descriptor. The deployment descriptor format is defined by the servlet specification, so all compliant servlet containers support this type of configuration file. It's an XML file named web.xml, located in the WEB-INF directory for the application. Details about how to package and deploy web applications, as well the complete syntax of the web.xml file, are described in

Appendix D

.

Let's look at how you can define the security constraints for the example we developed in Chapter 9 . To

restrict access to all pages dealing with employee registration, it's best to place them in a separate directory. The directory with all examples for

Chapter 10 , has a subdirectory named admin where all

these pages are stored. The part of the web.xml page that protects this resulting directory looks like this:

<security-constraint>

<web-resource-collection>

<web-resource-name>admin</web-resource-name>

<url-pattern>/ch10/admin/*</url-pattern>

</web-resource-collection>

<auth-constraint>

<role-name>admin</role-name>

</auth-constraint>

</security-constraint>

<login-config>

<auth-method>BASIC</auth-method>

<realm-name>ORA Examples</realm-name>

</login-config>

<security-role>admin</security-role>

The <security-constraint> element contains a <web-resource-collection> element that defines the resources to be protected, and an <auth-constraint> element that defines who has access to the protected resources. Within the <web-resource-collection> element, the URL pattern for the protected resource is specified with the <url-pattern> element; here it is set to the directory with all the registration pages: /ch10/admin/*. The <role-name> element within the

<auth-constraint>

element says that only users with the role admin can access the protected resources.

You define the type of authentication to use and a name associated with the protected parts of the application, known as the realm , with the <login-config> element. The <auth-method> element accepts the values BASIC, DIGEST, FORM, and CLIENT-CERT, corresponding to the authentication methods described earlier. Any text can be used as the value of the <realm-name> element. The text is shown as part of the message in the dialog the browser displays when it prompts the user for the credentials.

If you use form-based authentication, you must specify the names of your login form and error page in the <login-config> element as well:

<login-config>

<auth-method>FORM</auth-method>

<form-login-config>

<form-login-page>/login/login.html</form-login-page>

<form-error-page>/login/error.html</form-error-page>

</form-login-config>

</login-config>

<security-role>

elements are used to declare all role names that must be mapped to users and groups in the container's security domain. This information can be used by an application deployment tool to help the deployer with this task. Some containers, like Tomcat, work fine without this element,

but to make sure your application can be deployed in any compliant container, you should still declare all roles with <security-role> elements.

With these security requirement declarations in the WEB-INF/web.xml file, the web server and servlet container take care of all authentication and access control for you. But you may still need to know who the current user is, for instance to personalize the content. Or, if you configure your server to let different types of users access the same pages, you may need to know what type of user is actually accessing a page right now. The Servlet 2.2 API contains methods you can use to get access to this information in your JSP pages, using the implicit request object. Let's add another security constraint in the WEB-INF/web.xml file for the search pages from

Chapter 9

:

<security-constraint>

<web-resource-collection>

<web-resource-name>search</web-resource-name>

<url-pattern>/ch10/search/*</url-pattern>

</web-resource-collection>

<auth-constraint>

<role-name>admin</role-name>

<role-name>user</role-name>

</auth-constraint>

</security-constraint>

With this constraint, the server will allow only authenticated users with the roles admin and user to access the pages in the /ch10/search directory. You can then use information about who the user is to

provide different responses. Example 10.1

shows a fragment of a modified version of the list.jsp page

from

Chapter 9

.

Example 10.1. Generating the Response Based on the Current User (list.jsp)

...

<ora:loop name="empList" loopId="row" class="Row" >

<tr>

<td><%= row.getString("LastName") %></td>

<td><%= row.getString("FirstName") %></td>

<td><%= row.getString("Dept") %></td>

<td><%= row.getString("EmailAddr") %></td>

<td><%= row.getString("ModDate") %></td>

<% if (request.isUserInRole("admin") ||

row.getString("UserName").equals(request.getRemoteUser( ))) { %>

<td><%= row.getString("UserName") %></td>

<td><%= row.getString("Password") %></td>

<% } else { %>

<td>****</td>

<td>****</td>

<% } %>

<% if (request.isUserInRole("admin")) { %>

<td>

<form action="delete.jsp" method="post">

<input type="hidden" name="userName"

value='<%= StringFormat.toHTMLString(

row.getString("UserName")) %>'>

<input type="hidden" name="firstName"

value='<%= StringFormat.toHTMLString(

request.getParameter("firstName")) %>'>

<input type="hidden" name="lastName"

value='<%= StringFormat.toHTMLString(

request.getParameter("lastName")) %>'>

<input type="hidden" name="dept"

value='<%= StringFormat.toHTMLString(

request.getParameter("dept")) %>'>

<input type="submit" value="Delete">

</form>

</td>

<% } %>

</tr>

</ora:loop>

...

The first scriptlet uses the isUserInRole( ) method to see if the authenticated user is an admin, and the getRemoteUser( ) method to see if the employee information to be displayed is information about the authenticated user or someone else. If it turns out that the authenticated user is either an admin or is displaying information about him- or herself, the username and password information is displayed. Otherwise, the username and password fields are filled with dummy values.

The second scriptlet uses the isUserInRole( ) method again. Here, it is used to add the form with the Delete button only if the user is an admin.

© 2001, O'Reilly & Associates, Inc.

10.2 Application-Controlled Authentication

Using one of the container-provided mechanisms described in the previous section should be your first choice for authentication. But by definition, being container-provided means the application cannot dynamically add new users and roles to control who is granted access, at least not through a standard API defined by the servlet and JSP specifications.

For some types of applications, it's critical to have a very dynamic authentication model, one that doesn't require an administrator to define access rules before a new user can join the party. I'm sure you have seen countless sites where you can sign up for access to restricted content simply by filling out a form. One example is project management sites, where registered users can access document archives, discussion groups, calendars, and other tools for distributed cooperation. Another example is personalized news sites that you can customize to show news only about things you care about.

Unless you can define new users programmatically in the database used by an external authentication mechanism, you need to roll your own authentication and access control system for these types of applications. In this section, we'll look at the principles for how to do this. Note that this approach sends the user's password as clear text, so it has the same security issues as the container-provided basic and formbased authentication methods.

Application-controlled authentication and access control require the following:

1. User registration

2. A login page

3. The authentication mechanism, invoked by the login page

4. User information saved in the session scope to serve as proof of successful authentication

5. Validation of the session information in all JSP pages requiring restricted access

We'll reuse the example from

Chapter 9

for user registration: this allows us to focus on the parts of the application that require access control. The application is a simple billboard service, where employees can post messages related to different projects they are involved with. An employee can customize the application to show messages only about the projects he or she is interested in.

Figure 10.1

shows all the

pages and how they are related.

Figure 10.1. Application with authentication and access control

Let's go over it step by step. The login.jsp page is our login page. It contains a form that invokes the

authenticate.jsp page, where the username and password are compared to the information in the employee information database created in

Chapter 9

. If a matching user is found, the autheticate.jsp page creates an

EmployeeBean

object and saves it in the session scope. This bean serves as proof of authentication. It then redirects the client to a true application page. The page the client is redirected to depends on if the user loaded the login.jsp page or tried to directly access an application page without first logging in. All application pages, that is main.jsp, entermsg.jsp, storemsg.jsp, and updateprofile.jsp, contain a custom action that looks for the EmployeeBean object and forwards to the login.jsp page if it's not found, so the user is forced to log in. When the login.jsp page is loaded this way, we keep track of the page that the user tried to access so we can automatically display it after successful authentication. Finally, there's the logout.jsp page; this page can be loaded from a link in the main.jsp page. It simply terminates the session and redirects to the

login.jsp page again.

10.2.1 A Table for Personalization Information

Since the sample application in this chapter lets the user personalize the content of the billboard, we need a database table to store information about each employee's choices. The new table is shown in

Table 10.1

.

Column Name

Table 10.1, EmployeeProjects Database Table

SQL Datatype Primary Key?

UserName

CHAR (Text) Yes

ProjectName

CHAR (Text) Yes

The table holds one row per unique user-project combination. You need to create this table in your database before you can run the example.

10.2.2 Logging In

The login page contains an HTML form with fields for entering the user credentials: a username and a

password. This is why the information was included in the Employee table in Chapter 9

. Example 10.2

shows the complete login.jsp page.

Example 10.2. Login Page (login.jsp)

<%@ page import="com.ora.jsp.util.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Project Billboard</title>

</head>

<body bgcolor="white">

<h1>Welcome to the Project Billboard</h1>

Your personalized project news web site.

<p>

<font color="red">

<% String errorMsg = request.getParameter("errorMsg"); %>

<%= errorMsg == null ? "" : errorMsg %>

</font>

<form action="authenticate.jsp" method="post">

<% String origURL = request.getParameter("origURL"); %>

<input type="hidden" name="origURL"

value="<%= origURL == null ? "" : origURL %>">

Please enter your User Name and Password, and click Enter.

<p>

Name:

<input name="userName"

value="<ora:getCookieValue name="userName" />" size="10">

Password:

<input type="password" name="password"

value="<ora:getCookieValue name="password" />" size="10">

<input type="submit" value="Enter">

<p>

Remember my name and password:

<input type="checkbox" name="remember"

<%= CookieUtils.isCookieSet("userName", request) ?

"checked" : "" %> >

<br>

(This feature requires cookies to be enabled in your browser.)

</form>

</body>

</html>

The form contains the fields for the username and password, and the action attribute is set to the

authenticate.jsp page as expected. However, it also contains scripting elements that need an explanation.

The following fragment is used to display a message that gives the user a hint about why the login page is shown after an error:

<font color="red">

<% String errorMsg = request.getParameter("errorMsg"); %>

<%= errorMsg == null ? "" : errorMsg %>

</font>

First, the errorMsg variable is set to the value of the errorMsg request parameter, using the implicit request

object and accessed using the getParameter( ) method. This method returns the value of a parameter sent with the request, or null if the specified parameter isn't present. The errMsg parameter is set by the other pages when they forward to the login page, as you will soon see. When the user loads the

login.jsp directly, the parameter is not available in the request, so the value of the errorMsg variable is set to null. The message, or an empty string if it's null, is displayed by a JSP expression using the conditional operator described in

Chapter 6

. Figure 10.2

shows an example of the login page with an error message.

Figure 10.2. Login page with error message

Within the form, you find similar scripting elements:

<% String origURL = request.getParameter("origURL"); %>

<input type="hidden" name="origURL"

value="<%= origURL == null ? "" : origURL %>">

Here, a hidden field is set to the value of the originally requested URL. The field is passed as a parameter to the login page when another page forwards to it. This is how we keep track of which page the user wasn't allowed access to because he or she wasn't authenticated yet. Later you'll see how this information is used to load the originally requested page after authentication.

10.2.2.1 Using cookies to remember the username and password

The more web applications with restricted access a web surfer uses, the more usernames and passwords he or she needs to remember. After a while, it may be tempting to resort to the greatest security sin of all: writing down all usernames and passwords in a file such as mypasswords.txt. This invites anyone with access to the user's computer to roam around in all the secret data.

It can be a big problem keeping track of all accounts. Some sites therefore offer to keep track of the username and password using cookies. Cookies are small pieces of text that a server sends to the browser. A cookie with an expiration date is saved on the hard disk and returned to the server every time the user visits the same site until the cookie expires. So is this feature a good thing? Not really, as it amounts to the same security risk as writing down the username and password in a file. Even greater, since anyone with access to the user's computer doesn't even have to find the mypasswords.txt file; the browser takes care of sending the credentials automatically. But for sites that use authentication mainly to provide personalization and that don't contain sensitive data, using cookies can be an appreciated tool.

This example shows you how it can be done. If you decide to use it, be sure to make it optional so the user can opt out. We use a custom action called <ora:getCookieValue> to set the value of the input fields for the username and password:

Name:

<input name="userName"

value="<ora:getCookieValue name="userName" />" size="10">

Password:

<input type="password" name="password"

value="<ora:getCookieValue name="password" />" size="10">

The action has just one attribute: name. It's set to the name of the cookie that you're looking for. The action writes the value of the cookie to the response. If the specified cookie is not available, an empty string is returned. Here, the <ora:getCookieValue> action is used to set the default values for the username and password fields if the corresponding cookies are received with the request. You'll see how to send a cookie to the browser later.

The last part of the form creates a checkbox where the user can tell if this feature should be used or not:

Remember my name and password:

<input type="checkbox" name="remember"

<%= CookieUtils.isCookieSet("userName", request) ?

"checked" : "" %> >

To set the checked attribute, a utility method called isCookieSet( ) in the com.ora.jsp.util.CookieUtils

class is used. It takes two arguments: the cookie name and the implicit request object. If the cookie is found, the method returns true; otherwise, it returns false.

Here, the method is used with a conditional operator to set the checked attribute only if the userName cookie is received from the browser.

10.2.3 Authentication Using a Database

To authenticate a user, you need access to information about the registered users. For this chapter's examples, we keep all user information in a database. There are other options, including flat files and LDAP directories. When a user fills out the login page form and clicks Enter, the authentication page shown in

Example 10.3

is processed. This is a large page, so each part is discussed in detail after the complete page.

Example 10.3. Authentication Page (authenticate.jsp)

<%@ page language="java" %>

<%@ page import="com.ora.jsp.sql.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%-- Remove the validUser session bean, if any --%>

<% session.removeValue("validUser"); %>

<%

if (request.getParameter("userName").length( ) == 0 ||

request.getParameter("password").length( ) == 0) {

%>

<ora:redirect page="login.jsp" >

<ora:param name="errorMsg"

value="You must enter a User Name and Password." />

</ora:redirect>

<% } %>

<%--

See if the user name and password combination is valid. If not,

redirect back to the login page with a message.

--%>

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<ora:sqlQuery id="empInfo" dataSource="example">

SELECT * FROM Employee

WHERE UserName = ? AND Password = ?

<ora:sqlStringValue param="userName" />

<ora:sqlStringValue param="password" />

</ora:sqlQuery>

<% if (empInfo.size( ) != 1) { %>

<ora:redirect page="login.jsp" >

<ora:param name="errorMsg"

value="The User Name and Password you entered are not valid." />

</ora:redirect>

<% } else { %>

<%--

Create an EmployeeBean and save it in

the session scope and redirect to the appropriate page.

--%>

<% Row oneRow = (Row) empInfo.firstElement( ); %>

<jsp:useBean id="validUser" scope="session"

class="com.ora.jsp.beans.emp.EmployeeBean" >

<jsp:setProperty name="validUser" property="userName"

value='<%= oneRow.getString("UserName") %>' />

<jsp:setProperty name="validUser" property="firstName"

value='<%= oneRow.getString("FirstName") %>' />

<jsp:setProperty name="validUser" property="lastName"

value='<%= oneRow.getString("LastName") %>' />

<jsp:setProperty name="validUser" property="dept"

value='<%= oneRow.getString("Dept") %>' />

<jsp:setProperty name="validUser" property="empDate"

value='<%= oneRow.getString("EmpDate") %>' />

<jsp:setProperty name="validUser" property="emailAddr"

value='<%= oneRow.getString("EmailAddr") %>' />

</jsp:useBean>

<%-- Add the projects --%>

<ora:sqlQuery id="empProjects" dataSource="example">

SELECT * FROM EmployeeProjects

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

<%

String[] projects = new String[empProjects.size( )];

int i = 0;

%>

<ora:loop name="empProjects" loopId="row"

className="com.ora.jsp.sql.Row" >

<% projects[i++] = row.getString("ProjectName"); %>

</ora:loop>

<jsp:setProperty name="validUser" property="projects"

value="<%= projects %>" />

<% if (request.getParameter("remember") != null) { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="2592000" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="2592000" />

<% } else { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="0" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="0" />

<% } %>

<%--

Redirect to the main page or to the original URL, if

invoked as a result of an access attempt to a protected

page.

--%>

<% if (request.getParameter("origURL").length( ) != 0) { %>

<ora:redirect

page='<%= request.getParameter("origURL") %>' />

<% } else { %>

<ora:redirect page="main.jsp" />

<% } %>

<% } %>

The first thing that happens in

Example 10.3

is that a session scope object named validUser is removed if

it exists. As you will see later, validUser is the name we use for the EmployeeBean object, and its presence in the session scope indicates that the corresponding user has successfully logged in. If an

EmployeeBean

object is already saved in the session scope, it may represent a user that forgot to log out, so we must make sure it's removed when a new login attempt is made.

Next, a scriptlet is used to ensure that both the username and the password are passed as parameters. The

same getParameter( ) method used in Example 10.2

is used here to retrieve the parameter values. If one or both parameters are missing, the <ora:redirect> action redirects back to the login page again.

Here you see how the errorMsg parameter used in the login.page gets its value.

If the request contains both parameters, one of the database actions introduced in

Chapter 9

is used to see if there's a user with the specified name and password in the database:

<ora:sqlQuery id="empInfo" dataSource="example">

SELECT * FROM Employee

WHERE UserName = ? AND Password = ?

<ora:sqlStringValue param="userName" />

<ora:sqlStringValue param="password" />

</ora:sqlQuery>

<% if (empInfo.size( ) == 0) { %>

<ora:redirect page="login.jsp" >

<ora:param name="errorMsg"

value="The User Name and Password you entered are not valid." />

</ora:redirect>

If the query doesn't match a registered user (i.e., empInfo.size( ) returns 0), an <ora:redirect> action redirects back to the login page with an appropriate error message. Otherwise, the processing continues.

10.2.3.1 Creating the validation object

If a user is found, the single row from the query result is extracted and the column values are used to populate the single value properties of an EmployeeBean object. An EmployeeBean has the properties

shown in Table 10.2

.

Table 10.2, Properties for com.ora.jsp.beans.emp.EmployeeBean

Property Name Java Type Access Description

userName firstName lastName dept empDate emailAddr projects

String

String read/write The employee's unique username read/write The employee's first name

String

String read/write The employee's last name read/write The employee's department name

String read/write

The employee's employment date in the format yyyy-MM-dd

String read/write The employee's email address

String[] read/write A list of all projects the employee is involved in

The bean is named validUser and placed in the session scope using the standard <jsp:useBean> action. All properties are set to the values returned from the database using <jsp:setProperty> actions:

<% Row oneRow = (Row) empInfo.firstElement( ); %>

<jsp:useBean id="validUser" scope="session"

class="com.ora.jsp.beans.emp.EmployeeBean" >

<jsp:setProperty name="validUser" property="userName"

value='<%= oneRow.getString("UserName") %>' />

<jsp:setProperty name="validUser" property="firstName"

value='<%= oneRow.getString("FirstName") %>' />

<jsp:setProperty name="validUser" property="lastName"

value='<%= oneRow.getString("LastName") %>' />

<jsp:setProperty name="validUser" property="dept"

value='<%= oneRow.getString("Dept") %>' />

<jsp:setProperty name="validUser" property="empDate"

value='<%= oneRow.getString("EmpDate") %>' />

<jsp:setProperty name="validUser" property="emailAddr"

value='<%= oneRow.getString("EmailAddr") %>' />

</jsp:useBean>

As I mentioned earlier, this application lets the user select the projects he or she is interested in, so that only messages related to these projects are shown on the main page. The user's choices are stored in the

EmployeeProjects

database table described in

Table 10.1

. Next, we retrieve all projects from

EmployeeProjects

for the current user and set the value of the corresponding property in the bean to the complete list:

<%-- Add the projects --%>

<ora:sqlQuery id="empProjects" dataSource="example">

SELECT * FROM EmployeeProjects

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

<%

String[] projects = new String[empProjects.size( )];

int i = 0;

%>

<ora:loop name="empProjects" loopId="row" class="Row" >

<% projects[i++] = row.getString("ProjectName"); %>

</ora:loop>

<jsp:setProperty name="validUser" property="projects"

value="<%= projects %>" />

The value of the EmployeeBean projects property must be set as a String array. A scriptlet combined with an <ora:loop> action is used to first create a String array with the result from the database. A <jsp:setProperty> action is then used to set it as the projects property value of the validUser

bean.

10.2.3.2 Setting and deleting cookies

If the user asked for the user credentials to be remembered, we need to send the corresponding cookies to the browser. The checkbox value is sent to the authentication page as a parameter named remember:

<% if (request.getParameter("remember") != null) { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="2592000" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="2592000" />

<% } else { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="0" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="0" />

<% } %>

The <ora:addCookie> custom action is used to send cookies to the browser. If the parameter is set, the cookies are sent with a maximum age value representing 30 days, expressed in seconds (2592000). As long as the user returns to this site within this time frame, the cookies are sent with the request and the login page uses the values to automatically fill out the form fields. If the user decides not to use this feature and unchecks the box, we still send the cookies, but with a maximum age of 0. This means the cookies expire immediately and will never be sent to this server again. If you want to send a cookie to a browser that should be valid only until the user closes the browser, set the maximum age to a negative number (i.e., -1).

10.2.3.3 Redirect to the application page

The only thing left is to redirect the browser to the appropriate page. If the authentication process was started as a result of the user requesting a protected page without being logged in, the original URL is sent by the login page as the value of the origURL parameter:

<% if (request.getParameter("origURL").length( ) != 0) { %>

<ora:redirect

page='<%= request.getParameter("origURL") %>' />

<% } else { %>

<ora:redirect page="main.jsp" />

<% } %>

If this parameter has a value, the browser is redirected to the originally requested page; otherwise, it is redirected to the main entry page for the application.

10.2.4 Checking for a Valid Session

Authentication is only half of the solution. We must also add access control to each page in the application.

Example 10.4

shows the main.jsp page as an example of a protected page. This page shows all messages for the projects of the user's choice. It also has a form where the user can change the list of projects, and links to a page for posting new messages and to log out.

Example 10.4. Protected JSP Page (main.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="com.ora.jsp.beans.news.*" %>

<%@ page import="com.ora.jsp.util.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Project Billboard</title>

</head>

<body bgcolor="white">

<%-- Verify that the user is logged in --%>

<ora:validateSession name="validUser"

loginPage="login.jsp" errorMsg="Please log in first." />

<jsp:useBean id="validUser" scope="session"

class="com.ora.jsp.beans.emp.EmployeeBean" />

<h1>Welcome <%= validUser.getFirstName( ) %></h1>

Your profile currently shows you like information about the

following checked-off projects. If you would like to update your

profile, make the appropriate changes below and click

Update Profile.

<form action="updateprofile.jsp" method="post">

<input type="checkbox" name="projects" value="JSP"

<%= ArraySupport.contains(validUser.getProjects( ), "JSP") ?

"checked" : "" %> >JSP<br>

<input type="checkbox" name="projects" value="Servlet"

<%= ArraySupport.contains(validUser.getProjects( ), "Servlet") ?

"checked" : "" %> >Servlet<br>

<input type="checkbox" name="projects" value="EJB"

<%= ArraySupport.contains(validUser.getProjects( ), "EJB") ?

"checked" : "" %> >EJB<br>

<input type="submit" value="Update Profile">

</form>

<hr>

When you're done reading the news, please <a href="logout.jsp">

log out</a>.

<hr>

<a href=entermsg.jsp>Post a new message</a>

<p>

<jsp:useBean id="news" scope="application"

class="com.ora.jsp.beans.news.NewsBean" />

<%

NewsItemBean[] newsItems =

news.getNewsItems(validUser.getProjects( ));

pageContext.setAttribute("newsItems", newsItems);

%>

<table>

<ora:loop name="newsItems" loopId="newsItem"

className=" com.ora.jsp.beans.news.NewsItemBean" >

<tr>

<td colspan="2">

Project:

<jsp:getProperty name="newsItem" property="category" />

</td>

</tr>

<tr>

<td>

<jsp:getProperty name="newsItem" property="postedBy" />

</td>

<td>

<jsp:getProperty name="newsItem" property="postedDate" />

</td>

</tr>

<tr>

<td colspan="2">

<jsp:getProperty name="newsItem" property="msg" />

</td>

</tr>

<tr>

<td colspan="2"><hr></td>

</tr>

</ora:loop>

</table>

</body>

</html>

The most interesting piece of the example, from an access control point of view, is this:

<%-- Verify that the user is logged in --%>

<ora:validateSession name="validUser"

loginPage="login.jsp" errorMsg="Please log in first." />

The <ora:validateSession> custom action must be placed at the beginning of all protected pages in the application. It has three mandatory attributes. The name attribute specifies the name of the session scope object used to indicate that the session belongs to an authenticated user. Here we specify a name for the

EmployeeBean

object created by the authentication page. If the specified object is not found in the session, it means the page is being requested by a user that has not been authenticated. The custom action then forwards to the URL specified by the loginPage attribute, adding an errorMsg parameter with the

value specified by the errorMsg attribute. As in Example 10.2

, the errorMsg parameter is used to add a message on the login page to let the user know why a different page than the requested one is displayed. As with a regular forward, the conditional forward function implemented by the <ora:validateSession> action aborts the processing of the rest of the page.

10.2.4.1 Providing personalized content

The rest of the page shown in

Example 10.4

produces a personalized page for the authenticated user.

Figure

10.3

shows what it might look like.

Figure 10.3. Personalized application page

First, the validUser bean properties are used to welcome the user to the site by name. Next comes a form with checkboxes for all projects. The same technique used in

Chapter 6 is also used here to set the checked

attribute for the projects listed in the user's profile. The user can modify the list of projects and click Update

Profile to invoke the updateprofile.jsp page. This page modifies the profile information in the database. We'll take a look at how it's done later.

A NewsBean containing NewsItemBean objects is then used to display news items for all projects matching the user's profile. The implementations of these beans are intended only as examples. Initially, the

NewsBean

contains one hard-coded message for each news category, and the news items are kept in memory only. A real implementation would likely store all news items permanently in a database.

Example 10.4

also contains a link to a page where a news item can be posted. If you look at the source for the entermsg.jsp file, you can see that it's just a JSP page with the <ora:validateSession> action at the top and a regular HTML form that invokes the storemsg.jsp file with a POST request. The POST method is appropriate here, since the form fields are used to update information on the server (the in-memory

NewsBean

database).

The complete storemsg.jsp page is shown in

Example 10.5

.

Example 10.5. POST Page with Restricted Access (storemsg.jsp)

<%@ page language="java" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%-- Verify that the user is logged in --%>

<ora:validateSession name="validUser"

loginPage="login.jsp" errorMsg="Please log in first." />

<%-- Verify that it's a POST method --%>

<% if (!request.getMethod( ).equals("POST")) { %>

<ora:redirect page="main.jsp" />

<% } %>

<jsp:useBean id="validUser" scope="session"

class="com.ora.jsp.beans.emp.EmployeeBean" />

<%-- Create a new news item bean with the submitted info --%>

<jsp:useBean id="newsItem"

class="com.ora.jsp.beans.news.NewsItemBean" >

<jsp:setProperty name="newsItem" property="*" />

<jsp:setProperty name="newsItem" property="postedBy"

value='<%= validUser.getFirstName( ) + " " +

validUser.getLastName( ) %>' />

</jsp:useBean>

<%-- Make the news bean available to this page --%>

<jsp:useBean id="news" scope="application"

class="com.ora.jsp.beans.news.NewsBean" />

<%-- Add the new news item bean to the list --%>

<jsp:setProperty name="news" property="newsItem"

value="<%= newsItem %>" />

<ora:redirect page="main.jsp" />

This page creates a new NewsItemBean, sets all properties based on field parameters passed from the

entermsg.jsp page, plus the postedBy property using the firstName and lastName properties of the validUser

bean. It then adds the new news item to the NewsBean and redirects to the main page, where the new item is shown as a confirmation.

Let's focus on the access control aspects. At the top of the page, you find the same

<ora:validateSession>

action as with all other protected application pages. If a user fills out the form in entermsg.jsp and walks away from the computer without submitting the form, the user's session may time out. When the user then returns and clicks Submit, the <ora:validateSession> action doesn't find the validUser bean in the session and therefore forwards to the login page, setting the origURL parameter to the URL of the storemsg.jsp. After successful authentication, the authentication page redirects to the original URL, the storemsg.jsp. However, a redirect is always a GET request.

[1]

All the parameters sent with the original POST request for storemsg.jsp are lost; a POST request carries the parameter values in the message body, instead of in the URL (as a query string) as a GET request does, so the original URL we saved doesn't include the parameters. If we don't take care of this special case, an empty NewsItemBean will be added to the list.

[1]

The HTTP specification (RFC 2616) states that a browser is not allowed to change the method for the request when it receives a redirect response (status code 302). But, as acknowledged by the HTTP specification, all major browsers available today change a

POST request into a GET anyway.

There are at least two ways to deal with this. In Example 10.5

, the <ora:validateSession> action is followed by a scriptlet checking that the request for this page is a POST request. If not, it redirects to the main page without processing the request. This is the easiest way to deal with the problem, but it also means that the user will have to retype the message again. The chance that a session times out before a form is submitted is small, so in most cases this is not a big deal. It's therefore the solution I recommend.

If you absolutely must find a way to not lose the POST parameters when a session times out, here is a brief outline of a solution:

1. Modify the <ora:validateSession> action to send a URL in the origURL parameter suitable for use as a forward URL, as opposed to a redirect URL, if the page is invoked with a POST request.

A forward URL must be relative to the servlet context path, while a redirect URL should be absolute.

2. Use a scriptlet in the login page to save all POST parameter values as hidden fields in the form, along with a hidden field that tells if the original request was a GET or a POST request.

3. In the authentication page, forward to the originally requested URL if the method was a POST and

redirect only if it was a GET. The authentication page is always invoked as a POST request. A forward is just a way to let another page continue to process the same request, so the originally requested page is invoked with a POST request as expected, along with all the originally submitted parameters saved as hidden fields in the login page.

Depending on your application, you may also need to save session data as hidden fields in the page that

submits the POST request, so that the requested page doesn't have to rely on session information. But this leads to another problem. What if someone other than the user who filled out the form comes along and submits it? Information will then be updated on the server with information submitted by a user that's no longer logged in. One way out of this is to save information about the current user as a hidden field in the form that sends the POST request, and let the authentication page compare this information with the new user's information. If they don't match, the client can be redirected to the main application page instead of forwarded to the originally requested URL.

As you can see, there are a number of things to think about. Whether or not it makes sense to address all the issues depends on the application. My general advice is to keep it simple and stick to the first solution unless your application warrants a more complex approach.

10.2.5 Updating the User Profile

The updateprofile.jsp page, used if the user makes new project selections in the main page and clicks Update

Profile, is also invoked through the POST method. It follows the same approach as the storemsg.jsp page,

and is shown in Example 10.6

. But what's interesting about the updateprofile.jsp page is that it shows how to replace multirow data for a user, and is an instance of when you need to care about transactions.

Example 10.6. Updating Multiple Database Rows (updateprofile.jsp)

<%@ page language="java" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%-- Verify that the user is logged in --%>

<ora:validateSession name="validUser"

loginPage="login.jsp" errorMsg="Please log in first." />

<%-- Verify that it's a POST method --%>

<% if (!request.getMethod( ).equals("POST")) { %>

<ora:redirect page="main.jsp" />

<% } %>

<%-- Make the bean available for scripting elements --%>

<jsp:useBean id="validUser" scope="session"

class="com.ora.jsp.beans.emp.EmployeeBean" />

<%-- Update the project list in the bean --%>

<jsp:setProperty name="validUser" property="projects"

value='<%= request.getParameterValues("projects") %>' />

<ora:useDataSource id="example"

className="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<ora:sqlTransaction dataSource="example">

<%-- Delete the old project (if any) and insert the new ones --%>

<ora:sqlUpdate>

DELETE FROM EmployeeProjects

WHERE UserName = ?

<ora:sqlStringValue name="validUser" property="userName" />

</ora:sqlUpdate>

<ora:loop name="validUser" property="projects" loopId="project"

className="String" >

<ora:sqlUpdate>

INSERT INTO EmployeeProjects

(UserName, ProjectName) VALUES(?, ?)

<ora:sqlStringValue name="validUser" property="userName" />

<ora:sqlStringValue value="<%= project %>" />

</ora:sqlUpdate>

</ora:loop>

</ora:sqlTransaction>

<%-- Redirect to main page --%>

<ora:redirect page="main.jsp" />

The list of new projects selected by the user is sent to the updateprofile.jsp page in the projects request parameter. The projects bean property can therefore be updated using a <jsp:setProperty> action, setting the value to the result of the getParameterValues( ) method. As you may remember from

Chapter 6 , this method returns a String[] with all values for a parameter, and that's also the data type

defined for the projects property in the bean.

One important item to note here. If the user deselects all checkboxes in the main.jsp page ( Example 10.4

), all projects should be removed from the bean as well. The problem here is that if no checkbox is selected, the projects

request parameter is not sent at all. You must therefore use the type of request-time attribute

value shown in Example 10.6

, as opposed to using the param property, for the <jsp:setProperty> action. The <jsp:setProperty> action calls a property setter method only if it can find a corresponding parameter in the request. With no checkbox selected, the project's property setter is not called and the previous value is not cleared. When you use the getParameterValues( ) method as a request-time attribute value, however, it works as it should: if no checkbox is selected the method returns null, clearing the property value; otherwise, it returns a String[] with the currently selected values, setting the property to the current list.

The EmployeeProjects table ( Table 10.1

) contains one row per project for a user, with the username in the

UserName

column and the project name in the ProjectName column. The easiest way to update the database information is to first delete all existing rows, if any, and then insert rows for the new projects

selected by the user. Since this requires execution of multiple SQL statements and all must either succeed or fail, the <ora:sqlUpdate> actions are placed within the body of an <ora:sqlTransaction> action.

If the first <ora:sqlUpdate> action is successful but one of the others fails, the database information deleted by the first is restored so the database correctly reflects the state before the change.

To delete the rows in the database, use the <ora:sqlUpdate> action with a SQL DELETE statement. A

WHERE

clause is used so that only the rows for the current user are deleted. Then the <ora:loop> action is used to loop through all projects for the validUser bean. The body of the <ora:loop> action contains an <ora:sqlUpdate> action that executes an INSERT statement for each project:

<ora:loop name="validUser" property="projects" loopId="project"

className="String" >

<ora:sqlUpdate>

INSERT INTO EmployeeProjects

(UserName, ProjectName) VALUES(?, ?)

<ora:sqlStringValue name="validUser" property="userName" />

<ora:sqlStringValue value="<%= project %>" />

</ora:sqlUpdate>

</ora:loop>

The variable name to use for the component within the loop body is specified by the loopId attribute. Here it's set to project. Within the body, the loop variable is then used with the <ora:sqlStringValue> action to set the value for the ProjectName column, so a new value is used for each pass through the projects

property array. The UserName column has the same value in each row, so the validUser bean's userName property is used as the value for the corresponding <ora:sqlStringValue> action.

10.2.6 Logging Out

Since the proof of authentication is kept in the session scope, the user will be automatically logged out when the session times out. But even so, an application that requires authentication should always provide a way for the user to explicitly log out. This way a user can be sure that if he or she leaves the desk, no one else can come by and use the application.

The main page in the example application contains a link to the logout page, shown in

Example 10.7

.

Example 10.7. Logout Page (logout.jsp)

<%@ page language="java" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%--

Terminate the session and redirect to the login page.

--%>

<% session.invalidate( ); %>

<ora:redirect page="login.jsp" />

This page explicitly terminates the session by calling the invalidate( ) method of the session object in a scriptlet, and then redirects back to the login page. All objects kept in the session are removed and the session is marked as invalid. The next time someone logs in, a new session is created.

If you want to test the examples described in this chapter, you first must create at least one user with the

application we developed in Chapter 9

. To see how the automatic redirect to the originally requested page works, you can open two browser windows and log in from both. They both share the same session, so if you log out using one window and then try to load the "post a new message" page with the other, you will first be redirected to the login page. After you've entered your username and password, you're redirected to the page for posting a message.

© 2001, O'Reilly & Associates, Inc.

10.3 Other Security Concerns

In this chapter we have discussed only authentication and access control, but there's a lot more to web application security. You also need to ensure that no one listening on the network can read the data. In addition, you need to consider ways to verify that the data has not been modified. The common terms for these concepts (also used in the Servlet 2.2 specification) are confidentiality and data privacy for the first, and integrity checking for the second.

On an intranet, users can usually be trusted not to use network listeners to get to data they shouldn't see.

But on the Internet, you can make no assumptions. If you provide access to sensitive data, you have to make sure it's protected appropriately. Network security is a huge subject area, and clearly not within the scope of this book. Therefore I will touch on only the most common way to take care of both confidentiality and integrity checking: the Secure Socket Layer (SSL) protocol.

SSL is a protocol based on public key cryptography: it relies on a public key and a private key pair.

Messages sent by someone, or something (such as a server), are encoded using the private key, and can be decoded by the receiver only by using the corresponding public key. Besides confidentiality and integrity checking, public key cryptography also provides the means for very secure authentication: if a message can be decoded with a certain public key, you know it was encoded with the corresponding private key. The keys are issued, in the form of certificates together with user identity information, by a trusted organization such as VeriSign ( http://www.verisign.com/ ).

Both the client and the server can have certificates. However, the most common scenario today is that only the server has a certificate, and can thereby positively identify itself to the client. The SSL protocol takes care of this server authentication during the handshaking phase of setting up the connection. If the server certificate doesn't match the server's hostname, the user is warned or the connection is refused. If the client also has a certificate, it can be used to authenticate the client to the server in a more secure fashion than basic and digest authentication.

Even if only a server certificate is used, however, the communication between the client and the server is still encrypted. This means that the issue of sending passwords as clear text for basic authentication and form-based authentication, as well as the application-controlled authentication we developed in this chapter, is nullfied.

Most web servers today support server certificates and SSL. When you use HTTP over SSL (HTTPS), the URLs start with https instead of http. Not all applications need the kind of tight security offered by HTTPS, but you should be aware of all security threats, and carefully evaluate if the risks of not using it are acceptable for your application.

© 2001, O'Reilly & Associates, Inc.

11.1 How Java Supports Internationalization and Localization

Java was designed with internationalization in mind and includes a number of classes to make the effort as painless as possible. The primary class used for internationalization represents a specific geographical region. Instances of this class are used by other classes to format dates and numbers, as well as including localized strings and other objects in an application. There are also classes for dealing with different character encodings, which we will see later in the chapter.

11.1.1 The Locale Class

All Java classes that provide localization support use a class named java.util.Locale. An instance of this class represents a particular geographical, political, or cultural region, as specified by a combination of a language code and a country code. Java classes that perform tasks that differ depending on a user's language and local customs, called locale-sensitive operations, use a Locale instance to decide how to operate. Examples of locale-sensitive operations are the interpretation of date strings and formatting numeric values.

You create a Locale instance using a constructor that takes the country code and language code as arguments: java.util.Locale usLocale = new Locale("en", "US");

Here, a Locale for U.S. English is created. George Bernard Shaw (a famous Irish playwright) once observed, "England and America are two countries divided by a common language," so it's no surprise that both a language code and a country code are needed to describe some locales completely. The language code, a lowercase two-letter combination, is defined by the ISO 639 standard, available at http://www.ics.uci.edu/pub/ietf/http/related/iso639.txt

. The country code, an uppercase two-letter combination, is defined by the ISO 3166 standard, available at http://www.chemie.fuberlin.de/diverse/doc/ISO_3166.html

.

Table 11.1

and Table 11.2

show some of these codes.

Language Code

Table 11.1, ISO-639 Language Codes

Language

af

Afrikaans da

Danish

Country Code

DK

DE

GR

MX

NZ

ZA

GB ru sv zh es fr ja pl de el en

Table 11.2, ISO-3166 Country Codes

Country

Denmark

Germany

Greece

Mexico

New Zealand

South Africa

United Kingdom

German

Greek

English

Spanish

French

Japanese

Polish

Russian

Swedish

Chinese

US

United States

As luck would have it, these two standards are also used to define language and country codes in

HTTP. As you may remember from

Chapter 2 , a browser can send an Accept-Language header

with a request for a web resource such as a JSP page. The value of this header contains one or more codes for languages that the user prefers, based on how the browser is configured. If you use a

Netscape 4 browser, you can specify your preferred languages in the Edit->Preferences dialog, under the Languages tab. In Internet Explorer 4, you find the same thing in View->Internet

Options

when you click the Language button under the General tab. If you specify more than one language, they are included in the header as a comma-separated list:

Accept-Language: en-US, en, sv

The languages are listed in order of preference, with each language represented either by just the language code or by the language code and country code separated by a dash (-). This example header specifies the first choice as U.S. English, followed by any type of English, and finally Swedish. The

HTTP specification allows an alternative to listing the codes in order of preference, namely adding a socalled q-value to each code. The q-value is a value between 0.0 and 1.0 indicating the relative preference between the codes. Very few browsers, if any, use this alternative today, however.

The Accept-Language header helps you localize your application. You could write code that reads this header and creates the corresponding Locale instances. The good news is you don't have to do this yourself; the servlet container takes care of it for you and makes the locale information available through two methods on the implicit request object: java.util.Locale preferredLocale = request.getLocale( ); java.util.Enumeration allLocales = request.getLocales( );

The getLocale( ) method returns the Locale with the highest preference ranking, and the getLocales( )

method returns an Enumeration of all locales in order of preference. All you have to do is match the preferred locales to the ones that your web application supports. The easiest way to do this is to loop through the preferred locales and stop when you find a match. As you will see later, the custom actions developed for this book relieve you of all of this, but now you know how it's done.

11.1.2 Formatting Numbers and Dates

Let's look at how a locale can be used. One thing that we who live on this planet have a hard time agreeing upon is how to write dates and numbers. The order of the month, the day, and the year; if the

numeric value or the name should be used for the month; what character to use to separate the fractional part of a number: all of these details differ between countries, even between countries that speak the same language. And even though these details may seem picky, using the wrong format can cause a great deal of confusion. For instance, if you ask for something to be done by 5/2, an American thinks you mean May 2 while a Swede believes that it's due by February 5.

Java provides two main classes to deal with formatting of numbers and dates for a specific locale, appropriately named java.text.NumberFormat and java.text.DateFormat, respectively.

The NumberFormat class was used in Chapter 9

, to format the price information for items in a shopping cart according to the customs of the country where the server is located. By default, the

NumberFormat

class uses the locale of the underlying operating system. If used on a server configured to use a U.S. English locale, it formats numbers according to American customs; on a server configured with an Italian locale, it formats them according to Italian customs, and so forth. But you can also explicitly specify the locale to format numbers according to the rules for locales other than the one used by the operating system: java.util.Locale locale = request.getLocale( ); java.text.NumberFormat nf =

java.text.NumberFormat.getNumberInstance(locale);

String localNumber = nf.format(10000.00);

This piece of code creates a String with the number 10000.00 formatted according to the locale that corresponds to the preferred language specified by the Accept-Language header in a request.

Besides the getNumberInstance( ) method, you can use the getPercentInstance( ) and the getCurrency-Instance( ) to format a decimal number as a percentage string or any number as a currency string.

The DateFormat class works basically the same way, but how dates are written differs a lot more between locales than numbers do, since the day and month names are sometimes spelled out in the local language. Besides the locale, a formatting style is also specified as one of DEFAULT, SHORT,

MEDIUM

, LONG, or FULL: java.util.Locale locale = request.getLocale( ); java.text.DateFormat df =

java.text.DateFormat.getDateInstance(df.SHORT, locale);

String localDate = df.format(new java.util.Date( ));

If the current date is May 2, 2000, this code formats the date as 5/2/00 with an American locale and as 2000-05-02 with a Swedish locale. If you use the FULL formatting style, the results are

Tuesday

, May 2, 2000 and den 2 maj 2000 instead.

As with the NumberFormat class, there are other specialized date formatters besides the one used here. You can use the getDateTimeInstance( ) and getTimeInstance( ) methods to produce strings including both the date and time or just the time.

11.1.3 Using Localized Text

Automatic translation of numbers and dates into the local language is a great help. But until automatic translation software is a lot smarter than it is today, you have to translate all the text used in the application yourself. A set of Java classes then helps you pick the right version for a specific locale.

The main class for dealing with localized resources (such as text, images, and sounds) is named java.util.ResourceBundle

. This class is actually the abstract superclass for the two subclasses that do the real work, ListResourceBundle and PropertyResourceBundle, but it provides methods that let you get an appropriate subclass instance, hiding the details about which subclass actually provides the resources. Details about the difference between these two subclasses are beyond the scope of this book. It suffices to say, however, that the ListResourceBundle is overkill for our needs when developing web applications, so we will be using an instance of the

PropertyResourceBundle

. To learn more about these classes, I suggest glancing at the Java API documentation.

A PropertyResourceBundle instance is associated with a named set of localized text resources, where each resource is identified by a key. The keys and their corresponding text strings are stored in a regular text file as key-value pairs: site_name=The Big Corporation Inc.

company_logo=/images/logo_en.gif

welcome_msg=Hello!

Here, three keys, site_name, company_logo, and welcome_msg, have been assigned string values. The key is a string, without spaces or other special characters, and the value is any text. If the value spans more than one line, the line break must be escaped with a backslash character (\): multi_line_msg=This text value\ continues on the next line.

The file must use the extension .properties, for instance sitetext.properties, and be located in the class path used by the Java Virtual Machine. In the case of web applications, you should store the file in the application's WEB-INF/classes directory, since this directory is always included in the class path.

When you have created a properties file, you can obtain the text corresponding to a key like this: java.util.Locale locale = request.getLocale( ); java.util.ResourceBundle bundle =

java.util.ResourceBundle.getBundle("sitetext", locale);

String msg = bundle.getString("welcome_msg");

Note that the getBundle( ) method takes two arguments: a Locale argument, the same as the methods for getting number and date formatters; and a bundle name. These arguments are used like this: the method gets the language and country codes from the Locale object and starts looking for a file with a name composed of both the bundle name and the language and country codes. If you pass it a locale for Mexican Spanish, for example, it first looks for a file named sitetext_es_MX.properties, where es is the language code for Spanish and MX is the country code for Mexico. If it can't find a file with this name, it looks for sitetext_es.properties, ignoring the country code. If there's still no such file, it uses the file with just the bundle name, sitetext.properties.

As you can see, this makes it possible for you to create multiple properties files, each with the text values translated into a specific language for a specific country. In other words, you can create one file for each supported locale. The ResourceBundle ensures that when you ask for a bundle, you get the one that most closely matches the specified locale, or the default bundle if there is no match. We'll look at an example in detail in the next section.

Besides the ResourceBundle class, there's a class named java.text.MessageFormat that you can use for messages composed of fixed text plus variable values, such as, "An earthquake measuring 6.7 on the Richter scale hit Northridge, CA, on January 17, 1994." Here, each underlined word represents a variable value. Another class related to localization is the java.text.Collator class, used for localized string comparison and sorting. These classes are less commonly used, so they are not covered in detail here. You can read more about them in the Java API documentation.

© 2001, O'Reilly & Associates, Inc.

11.2 Generating Localized Output

Now that you have an understanding of the type of internationalization support Java provides, let's look at a concrete example. But instead of using the internationalization classes directly in the pages, let's use a set of custom actions based on these classes. Using custom actions minimizes the need for Java code in the JSP pages, making it easier for page authors to develop an internationalized site.

The example application, briefly described in the introduction to this chapter, lets visitors voice their opinions by selecting one of the answers to a question, as well as seeing how others have answered. The

text, numbers, and dates are available in three different languages. Figure 11.1

shows all pages used in this application and how they are related.

Figure 11.1. Localized poll application pages

The first page the user sees is the poll.jsp page, shown in Figure 11.2

. The language used to display the

contents the first time this page is displayed is based on the Accept-Language header value in the request. The top part of the page contains radio buttons for the three supported languages and a Submit button. If the user wants the application to be presented in another language, he or she selects the corresponding radio button and clicks Submit, causing the page to be requested again, this time with a language parameter included in the request. The value of the language parameter is then used to display the page in the selected language. Information about the selected language is saved as session data, so it's available to all the other application pages.

Figure 11.2. The language selection and question page

The poll.jsp page also includes a question, linked to a page with background information for the question, and a group of radio buttons representing the different answers, as well as a Submit button. Clicking on the

Submit button invokes the calculate.jsp page, where the vote is validated. If it's valid, it's added to the global poll result. The request is then forwarded to the result.jsp page, which displays the poll statistics with all numbers formatted according to the selected locale. If it's not valid, the request is forwarded back to the poll.jsp page.

Both the poll.jsp page and the result.jsp page are designed to show text, numbers, and dates according to the selected locale using custom actions based on the Java classes described in the previous section. This approach is perfect when the amount of text is small; only one page has to be maintained. But if a page needs to contain a great deal of text, typing it into a properties file and escaping all line breaks may not be the best approach. Some pages also need to use different layouts, colors, images, and general appearances based on the locale. In this case, it's easier to use a separate page per locale. This approach is illustrated by the pages providing more detailed information about the question in this example. The link on the poll.jsp page leads to different JSP pages depending on the selected language, named according to the same naming convention as ResourceBundle properties files: details.jsp, details_de.jsp, and details_sv.jsp for

English (the default), German, and Swedish pages, respectively. Let's look at the one-page and the multipage approaches separately.

11.2.1 Using One Page for Multiple Locales

Example 11.1

shows the poll.jsp page. That's where the magic of locale selection happens, and the selection is then used to produce text in the corresponding language throughout the page.

Example 11.1. Language Selection and Vote Page (poll.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="poll"

supportedLangs="en, sv, de" />

<html>

<head>

<title>

<ora:getLocalText name="locale" key="poll.title" />

</title>

</head>

<body bgcolor="white">

<h1>

<ora:getLocalText name="locale" key="poll.title" />

</h1>

<ora:getLocalText name="locale" key="poll.select_language" />:

<form action="poll.jsp">

<p>

<input type="radio" name="language" value="en"

<%= locale.getLanguage( ).equals("en") ? "checked" : "" %>>

<%= locale.getText("poll.english") %><br>

<input type="radio" name="language" value="sv"

<%= locale.getLanguage( ).equals("sv") ? "checked" : "" %>>

<%= locale.getText("poll.swedish") %><br>

<input type="radio" name="language" value="de"

<%= locale.getLanguage( ).equals("de") ? "checked" : "" %>>

<%= locale.getText("poll.german") %><br>

<p>

<input type="submit"

value="<ora:getLocalText name="locale"

key="poll.new_language" />">

</form>

<a href="<ora:getLocalPageName

name="locale" pageName="details.jsp" />">

<ora:getLocalText name="locale" key="poll.question" />

</a>

<form action="calculate.jsp" method="post">

<input type="radio" name="answerId" value="1">

<ora:getLocalText name="locale" key="poll.answer1" />

<br>

<input type="radio" name="answerId" value="2">

<ora:getLocalText name="locale" key="poll.answer2" />

<br>

<input type="radio" name="answerId" value="3">

<ora:getLocalText name="locale" key="poll.answer3" />

<p>

<input type="submit"

value="<ora:getLocalText name="locale"

key="poll.submit" />">

</form>

</body>

</html>

At the top of the page, the taglib directive is used to identify the library containing all custom actions, as in previous chapters. Then follows the first custom action for localization: <ora:useLocaleBundle> .

It's described in Table 11.3

.

Attribute Name

Table 11.3, Attributes for <ora:useLocaleBundle>

Java Type Request-Time Value Accepted Description

id String

No

Mandatory. The name used to reference the

LocaleBean

instance.

bundleName String

Yes supportedLangs String

Yes

Mandatory. The base name for text resource properties files.

Mandatory. A comma-separated list of language/country codes. The first code is used as the default language.

This innocent-looking action does a number of things. First, it looks for a LocaleBean with the name specified by the id attribute in the session scope, and creates one if it doesn't exist. The LocaleBean,

described in Table 11.4

, handles all localization tasks. It can be used as-is, as you will soon see, but in most

cases it's used indirectly by other custom actions in the set of localization actions. The action then asks the implicit request object for the list of locales specified by the Accept-Language header, and uses that list to set the bean's requestLocales property. It also looks for a request parameter named language, and if it's present, it uses the value to set the corresponding language property in the bean. Finally, it sets the bean's supportedLangs property to the value of the action attribute with the same name. Both the language property and the supportedLangs property take a value that's either just a language code, or a

language code plus a country code separated by a dash (e.g., Es-MX). As shown in Example 11.1

, you can

specify a number of supported languages as a comma-separated list of codes. The final attribute is called bundleName

; this attribute is the base name for a set of ResourceBundle properties files, as described in the first section of this chapter.

Property Name

bundleName charset language locale

Table 11.4, Properties for com.ora.jsp.bean.locale.LocaleBean

Java Type Access Description

String write The base name for the properties files

String write The charset used to decode parameters

String read/write The language code for the selected locale java.util.Locale

read The locale, selected based on other properties requestLocales java.util.Locale[] write The locales received with the request supportedLangs String write A comma-separated list of language codes

With all these properties set, the bean can decide which locale to use for the user requesting the page. The first time the page is requested, the language property is not set, so it compares the language specified by each locale in the requestLocales property to the set of languages in the supportedLangs property, and selects the first locale that is supported. Since the request locales are ordered by preference, the locale with the highest ranking that represents a supported language is selected. As you will soon see, the user can also request this page with a specific language specified by the language parameter. In this case, the action sets the corresponding bean property and the bean uses this value to select the locale, assuming it's one of the supported languages. If neither the request locales nor the explicitly specified language is supported, the bean selects a locale that represents the first language listed in the supportedLanguages

property.

The next action in

Example 11.1

is <ora:getLocalText> . This is described in

Table 11.5

.

Table 11.5, Attributes for <ora:getLocalText>

Attribute Name Java Type Request-Time Value Accepted Description

name String

No

Mandatory. The name of the LocaleBean instance.

key String

Yes

Mandatory. The name of a property in the text resource properties files.

The <ora:getLocalText> action is used to get the page title and header. The name attribute specifies

the name of the LocaleBean created by the <ora:useLocaleBundle> action, and the key attribute specifies one of the properties in the files with localized strings. These files are named exactly like the files used by the ResourceBundle described in the previous section. In other words, you need one file with the same name as the base name (specified as the bundleName for the <ora:useLocaleBundle> action) for the default locale, and one file with a name that combines the base name and a language code for all other locales. In this example, then, you need the files poll.properties, poll_de.properties, and

poll_sv.properties. If you want to add support for another language, say Italian, just create a

poll_it.properties file and add it (the language code for Italian) to the list of supported languages for the

<ora:useLocaleBundle>

action. All properties files must be placed in the WEB-INF/classes directory for the web application so that the ResourceBundle can find them. Here's what the

poll.properties file looks like: poll.title=Industry Trends poll.select_language=Select your preferred language poll.new_language=New Language poll.english=English poll.swedish=Swedish poll.german=German poll.question=What's the longest development time you dare to plan with?

poll.answer1=One year poll.answer2=Six months poll.answer3=Less than six months poll.submit=Vote poll.number_of_votes=Total number of votes poll.result=Poll result

The value of the poll.title key, used by the first two <ora:getLocalText> actions, is set to

"Industry Trends"; that's what will appear as the title and header of the page when the default locale is selected. If a Swedish locale was selected instead, the text "Industri Trender" would be used, which is how it is listed in the poll_sv.properties file. The <ora:getLocalText> action is used with different keys for all text content in the page. Internally, it uses one of the bean's regular methods: public String getText(String propertyName)

This method returns the specified property (the action element uses the key attribute value) from the properties file that most closely matches the selected locale. The bean provides similar methods for date

and numeric values, as you can see in Appendix C .

To let the user pick another language than the one selected based on the Accept-Language header, the page contains a form with a set of radio buttons and a Submit button. Every time the page is displayed, the radio button group must reflect the currently selected language. This is done by calling the bean's language

property access method and comparing the return value with the language code represented by each radio button:

...

<input type="radio" name="language" value="en"

<%= locale.getLanguage( ).equals("en") ? "checked" : "" %>>

<%= locale.getText("poll.english") %><br>

...

You probably recognize the type of JSP expression used to set the checked attribute for the radio button from previous chapters. The getLanguage( ) method returns the language code for the selected locale as a String. The equals( ) method compares the return value to its argument and returns true if they are the same. If they are, the first string after the question mark is returned as the value of the expression. If not, the second string is used. You also may have noticed that you can use the bean's getText( )

method directly, as an alternative to the <ora:getLocalText> action. Which alternative to use is largely a matter of preference. I used the method here because it's more compact and less intrusive when the text is used as part of an HTML element.

All radio button elements have the name language, which means that they form a group where only one of them can be selected. When the user clicks on the Submit button, the same page is requested with the value of the selected radio button included as a request parameter named language. As described above, this triggers the <ora:useLocaleBundle> action to switch to the selected language.

Next comes another form with radio buttons representing the three possible answers to the poll question. As you can see, both the question and the answers are displayed in the selected language. When the user selects an answer and clicks on the button to submit a vote, the calculate.jsp page shown in

Example 11.2

is invoked.

Example 11.2. Validation and Calculation of Votes (calculate.jsp)

<jsp:useBean id="pollResult" scope="application"

class="com.ora.jsp.beans.poll.PollBean" />

<jsp:useBean id="answer" class="com.ora.jsp.beans.poll.AnswerBean" >

<jsp:setProperty name="answer" property="*" />

</jsp:useBean>

<% if (answer.isValid( )) { %>

<jsp:setProperty name="pollResult" property="answer"

value="<%= answer %>" />

<jsp:forward page="result.jsp" />

<% } else { %>

<jsp:forward page="poll.jsp" />

<% } %>

As with all pure logic pages, this page contains only actions and a few simple scriptlets; no response text is generated. A PollBean in the application scope is used to keep track of the answers from all visitors, and an AnswerBean in the page scope captures and validates a single answer. The AnswerBean has one property named answer, which is set to the value of the corresponding request parameter using the

<jsp:setProperty>

action. It also has an isValid( ) method, used in a scriptlet to test if the answer is valid or not. In this example, it returns true if the answer ID is valid (1, 2, or 3). However, in a real application you may want to include other validation rules. For instance, if the poll information was stored in a database, you could use cookies or a username to make sure each user answers only once. If the answer is valid, a <jsp:setProperty> action is used to set the answer property of the PollBean to

the valid answer, and the request is forwarded to the result.jsp page to display the poll statistics. Figure

11.3

shows a sample of the result page with the Swedish locale.

Figure 11.3. The result page using the Swedish locale

The result.jsp page, shown in Example 11.3

, uses a couple of custom actions we haven't covered yet to

display the localized date and numbers.

Example 11.3. Showing the Result (result.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="java.util.Date" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="poll"

supportedLangs="en, sv, de" />

<html>

<head>

<title>

<ora:getLocalText name="locale" key="poll.title" />

</title>

</head>

<body bgcolor="white">

<jsp:useBean id="pollResult" scope="application"

class="com.ora.jsp.beans.poll.PollBean" />

<h1>

<ora:getLocalText name="locale" key="poll.result" />:

<ora:getLocalDate name="locale" date="<%= new Date( ) %>" />

</h1>

<ora:getLocalText name="locale" key="poll.question" />

<p>

<ora:getLocalText name="locale" key="poll.number_of_votes" />:

<ora:getLocalNumber name="locale"

value="<%= pollResult.getTotal( ) %>" />

<table width="70%">

<tr>

<td width="30%">

<ora:getLocalText name="locale" key="poll.answer1" />:

<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer1Percent( ) %>" />%

(<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer1( ) %>" />)

</td>

<td>

<table width="<%= pollResult.getAnswer1Percent( ) %>%"

bgcolor="lightgreen">

<tr>

<td>&nbsp;</td>

</tr>

</table>

</td>

</tr>

<tr>

<td width="30%">

<ora:getLocalText name="locale" key="poll.answer2" />:

<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer2Percent( ) %>" />%

(<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer2( ) %>" />)

</td>

<td>

<table width="<%= pollResult.getAnswer2Percent( ) %>%"

bgcolor="lightblue">

<tr>

<td>&nbsp;</td>

</tr>

</table>

</td>

</tr>

<tr>

<td width="30%">

<ora:getLocalText name="locale" key="poll.answer3" />:

<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer3Percent( ) %>" />%

(<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer3( ) %>" />)

</td>

<td>

<table width="<%= pollResult.getAnswer3Percent( ) %>%"

bgcolor="orange">

<tr>

<td>&nbsp;</td>

</tr>

</table>

</td>

</tr>

</table>

</body>

</html>

This page starts with the <ora:useLocaleBundle> action, just like the poll.jsp page, to make the

LocaleBean

available to the other actions and scriptlets on the page. It also uses a number of

<ora:getLocalText>

actions to produce text in the selected language.

The first new action is the <ora:getLocalDate> action, described in Table 11.6

.

Attribute Name Java Type

Table 11.6, Attributes for <ora:getLocalDate>

Request-Time Value Accepted Description

name String

No

Mandatory. The name of the LocaleBean instance.

date java.util.Date Yes

Mandatory. The date to format according to the selected locale.

The <ora:getLocalDate> action is used to add today's date to the header. As with all other localization actions, it has a name attribute to specify the name of the bean. The date to format (as dictated

by the selected locale) is specified by the date attribute. In Example 11.3

, a JSP expression that creates a

new Date object representing the current date is used as the attribute value. When you play around with this application, you see how the date format changes depending on the language you select.

The other new action is the <ora:getLocalNumber> action, used to generate numeric values

formatted according to the selected locale. It's described in Table 11.7

.

Table 11.7, Attributes for <ora:getLocalNumber>

Attribute Name Java Type Request-Time Value Accepted Description

name String

No

Mandatory. The name of the LocaleBean instance.

value double

Yes

Mandatory. The number to format according to the selected locale.

The first occurrence of the <ora:getLocalNumber> action is used to display the total number of votes, just before the table that shows the distribution of the votes. Besides the name attribute, it has an attribute named value that specifies the number to be formatted. In

Example 11.3

, it calls the poll bean's getTotal( )

method to set the value.

The table with details about the distribution comes next. Here I have used a trick with nested tables to generate a simple bar chart:

...

<table width="70%">

<tr>

<td width="30%">

<ora:getLocalText name="locale" key="poll.answer1" />:

<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer1Percent( ) %>" />%

(<ora:getLocalNumber name="locale"

value="<%= pollResult.getAnswer1( ) %>" />)

</td>

<td>

<table width="<%= pollResult.getAnswer1Percent( ) %>%"

bgcolor="lightgreen">

<tr>

<td>&nbsp;</td>

</tr>

</table>

</td>

</tr>

...

The main table contains a row with two cells for each poll answer. The first cell is just a regular cell, containing the answer text, the percentage of votes with this answer, and the absolute number of votes with this answer. The values are generated by the <ora:getLocalText> and <ora:getLocalNumber> actions. The next cell, however, is more interesting. It contains a nested table, and the width of the table is set to the same percentage value as the percentage of votes with this answer. By specifying a required space

(using the &nbsp; HTML code) as the value of the single cell and a unique background color, the result is a simple dynamic bar chart. As the percentage values of the answers change, the width of each nested table changes as well, as shown in

Figure 11.3

. Pretty neat!

11.2.2 Using a Separate Page for Each Locale

The <ora:getLocalText> action, as well as the other localization actions, makes it easy to use the same page for all locales. But as described earlier, sometimes it's better to use a separate page for each locale. The poll example uses this approach for the detailed description of the question.

As shown in

Example 11.1

, the poll.jsp page uses the <ora:getLocalPageName> action to insert the name of a localized page in an HTML link:

<a href="<ora:getLocalPageName

name="locale" pageName="details.jsp" />">

<ora:getLocalText name="locale" key="poll.question" />

</a>

This action, described in

Table 11.8

, generates filenames based on the same naming convention as for localized property files.

Table 11.8, Attributes for <ora:getLocalPageName>

Attribute Name Java Type Request-Time Value Accepted Description

name String

No

Mandatory. The name of the LocaleBean instance.

pageName String

Yes Mandatory. The page base name.

The pageName attribute value represents the page base name. From this base name, the action inserts the language code and the country code (if any) of the selected locale, unless the selected locale represents the default language. The default language is the first language listed in the supportedLang attribute for the

<ora:useLocaleBundle>

action. For the languages supported in this example, you therefore need the

details.jsp file for the English locale (default), the details_de.jsp file for the German locale, and the

details_sv.jsp file for the Swedish locale. Note that the <ora:getLocalPageName> action doesn't verify that the localized page exists; it just generates the name of the localized page, based on the currently

selected locale. Example 11.4

shows the Swedish page.

Example 11.4. Swedish Details Page ( details_sv.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="poll"

supportedLangs="en, sv, de" />

<html>

<head>

<title>

<ora:getLocalText name="locale" key="poll.title" />

</title>

</head>

<body bgcolor="yellow">

<h1>

<font color="blue">

<ora:getLocalText name="locale" key="poll.question" />

</font>

</h1>

<font color="blue">

Idag introduceras nya teknologier och affärsideer mycket

snabbt. Produkter som såg ut som givna vinstbringare

igår är idag så vanliga att det inte går att tjäna

pengar på dem, med flera versioner tillgängliga gratis

som Open Source. En affärsplan baserad på inkomst från

annonser på en populär web site, eller att lägga till

".com" till företagsnamnet, väcker inte samma intresse

hos investerare idag som det gjorde för bara några månader

sedan.

<p>

I en industri som rör sig så här snabbt, hur lång tid

törs du allokera till utveckling av en ny produkt eller

tjänst, utan att riskera att den är ointressant när den

väl är färdig?

</font>

</body>

</html>

As you can see, most of this page consists of Swedish text. The colors of the Swedish flag (yellow and blue) are also used as the background, header, and text colors. The detail pages for the other locales follow the same pattern. When the amount of text is large and other details of the page differ, like the colors in this example, it's often convenient to use a separate page for each locale instead of the one-page approach described earlier.

© 2001, O'Reilly & Associates, Inc.

11.3 A Brief History of Bits

Before we discuss the different charsets, let's shift gears a little. Once upon a time, not so long ago, bits were very expensive. Hard disks for storing bits, memory for loading bits, communication equipment for sending bits over the wire; all the resources needed to handle bits were costly. To save on these expensive resources, characters were initially represented by only seven bits. This was enough to represent all letters in the English alphabet, the numbers through 9, punctuation characters, and some control characters. And that was all that was really needed in the early days of computing, since most computers were kept busy doing number crunching.

But as computers were given new tasks, often dealing with human-readable text, seven bits didn't cut it.

Adding one bit made it possible to represent all letters used in the western European languages. But there are other languages besides the western European languages, even though companies based in

English-speaking countries often seem to ignore them. And eight bits is not enough to represent all characters used around the world. At first, this problem was partially solved by defining a number of standards for how eight bits should be used to represent different character subsets. Each of the ten ISO-

8859 standards defines what is called a charset: a mapping between eight bits (a byte) and a character.

For instance, ISO-8859-1, also known as Latin-1, defines the subset used for western European languages, such as English, French, Italian, Spanish, German, and Swedish. This is the default charset for HTTP. Other standards in the same series are ISO-8859-2, covering central and eastern European languages such as Hungarian, Polish, and Romanian; and ISO-8859-5, with Cyrillic letters used in

Russian, Bulgarian, and Macedonian. You can find information about all ten charsets in the ISO-8859 series at http://czyborra.com/charsets/iso8859.html

.

Some languages such as Chinese and Japanese contain thousands of characters, but with eight bits you can only represent 256. A set of multibyte charsets have therefore been defined to handle these languages, such as Big5 for Chinese, Shift_JIS for Japanese, and EUC-KR for Korean.

As you can imagine, all these different standards make it hard to exchange information encoded in different ways. To simplify life, the Unicode standard was defined by the Unicode Consortium, which was founded in 1991 by large companies such as Apple, IBM, Microsoft, Novell, Sun, and Xerox.

Unicode uses two bytes (16 bits) to define unique codes for 49,194 characters in Version 3.0. Java uses

Unicode for its internal representation of characters, and Unicode is also supported by many new technologies such as XML and LDAP. Support for Unicode is included in all modern browsers, such as

Netscape and Internet Explorer since Version 4. If you would like to learn more about Unicode, visit http://www.unicode.org/ .

What does all of this mean to you as a web application developer? Well, since Latin-1 is the default charset for HTTP, you don't have to worry about this at all when you work with western European languages. But if you provide content in another language, such as Japanese or Russian, you need to

tell the browser which charset you're using so it can interpret and render the characters correctly. In addition, the browser must be configured with a font that can display the characters. You find information about fonts for Netscape at http://home.netscape.com/eng/intl/ , and for Internet Explorer at http://www.microsoft.com/ie/intlhome.htm

.

You can specify a charset in a JSP page using the page directive and the contentType attribute, as

shown in Example 11.5

. The charset you specify is used for multiple purposes. First, it tells the JSP

container the charset used to encode the bytes in the JSP page file itself, so the container can translate the bytes correctly to Unicode for internal processing. It's also used to convert the Unicode characters used internally to the specified charset encoding when the response is sent to the browser, and to set the charset attribute in the Content-Type header to let the browser know how to interpret the response.

You may think it's a waste of time to first convert from one charset to Unicode, and then from Unicode back to the same charset. But using Unicode as an intermediary format makes it possible to store the page in one charset, say Shift_JIS, and send it to the browser as another, for instance UTF-8 (an efficient charset that encodes Unicode characters as one, two, or three bytes, as needed). This is not possible in JSP 1.1, but it's being discussed for a future version.

Enough theory. Figure 11.4

shows a simple JSP page that sends the text "Hello World" in Japanese to the browser. The Japanese characters are copied with permission from Jason Hunter's Java Servlet

Programming (O'Reilly).

Figure 11.4. Japanese JSP page ( japanese.jsp)

To create a file in Japanese or another non-western language, you obviously need a text editor that can

handle multibyte characters. The JSP page in Figure 11.4

was created with WordPad on a Windows NT system, using a Japanese font called MS Gothic and saved as a file encoded with the Shift_ JIS charset.

Shift_ JIS is therefore the charset specified by the contentType attribute, using the charset attribute. Note that the page directive that defines the charset must appear as early as possible in the

JSP page, before any characters that can be interpreted only when the charset is known. I recommend that you insert it as the first line in the file to avoid problems.

© 2001, O'Reilly & Associates, Inc.

11.4 Handling Localized Input

So far we have discussed only how to generate pages in different languages, but most applications also need to deal with localized input. As long as you're supporting only western European languages, the only thing you typically need to worry about is how to interpret dates and numbers. The LocaleBean introduced in the previous section can help with this.

Example 11.5

shows a JSP page with the same form for selecting a language as you saw in Example 11.1

, plus a form with one field for a date and another for a number.

Example 11.5. Date and Number Input Form (input.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="java.util.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="input"

supportedLangs="en, sv, de" />

<html>

<head>

<title>

<ora:getLocalText name="locale" key="input.title" />

</title>

</head>

<body bgcolor="white">

<h1>

<ora:getLocalText name="locale" key="input.title" />

</h1>

<ora:getLocalText name="locale" key="input.select_language" />:

<form action="input.jsp">

<p>

<input type="radio" name="language" value="en"

<%= locale.getLanguage( ).equals("en") ? "checked" : "" %>>

<%= locale.getText("input.english") %><br>

<input type="radio" name="language" value="sv"

<%= locale.getLanguage( ).equals("sv") ? "checked" : "" %>>

<%= locale.getText("input.swedish") %><br>

<input type="radio" name="language" value="de"

<%= locale.getLanguage( ).equals("de") ? "checked" : "" %>>

<%= locale.getText("input.german") %><br>

<p>

<input type="submit"

value="<ora:getLocalText name="locale"

key="input.new_language" />">

</form>

<form action="store.jsp" method="post">

<ora:getLocalText name="locale" key="input.date" />

<br>

<input type="text" name="date">

(<ora:getLocalDate name="locale" date="<%= new Date( ) %>" />)

<p>

<ora:getLocalText name="locale" key="input.number" />

<br>

<input type="text" name="number">

(<ora:getLocalNumber name="locale" value="<%= 10000.11 %>" />)

<p>

<input type="submit"

value="<ora:getLocalText name="locale"

key="input.submit" />">

</form>

</body>

</html>

As in

Example 11.1

, custom actions are used to display various text labels in the selected language. In the

date and number entry form, the <ora:getLocalDate> and <ora:getLocalNumber> actions are used as before to generate samples for the date and number format, respectively.

Now the interesting part. Example 11.6

shows the JSP page that is requested when the form is submitted.

Example 11.6. Processing Localized Input (store.jsp)

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="input"

supportedLangs="en, sv, de" />

<ora:useDataSource id="example"

class="sun.jdbc.odbc.JdbcOdbcDriver"

url="jdbc:odbc:example" />

<ora:sqlUpdate dataSource="example">

INSERT INTO InputTest VALUES(?, ?)

<ora:sqlDateValue

value='<%= locale.getDate(request.getParameter("date")) %>' />

<ora:sqlDoubleValue

value='<%= locale.getDouble(request.getParameter("number")) %>'/>

</ora:sqlUpdate>

<jsp:forward page="input.jsp" />

This page stores the values in a database. However, in order to do that, the date and number strings must be interpreted and turned into the corresponding Java object. The LocaleBean provides methods to handle the conversion, with a little bit of help from the DateFormat and NumberFormat classes described earlier: public Date getDate(String date)

throws ParseException public double getDouble(String number)

throws ParseException

These two methods use the format classes, initialized with the currently selected locale, to convert the

String

argument to the appropriate return type. With the strings converted to the corresponding Java

type, the custom actions introduced in Chapter 9 are used to store the values in a database. To run this

example, you must first create a table named InputTest with a DATE and a NUMBER column in your database.

11.4.1 Dealing with Non-Western European Input

An HTML form can be used for input in languages other than western European languages, but the charset discussed in the previous section comes into play here as well. When you create a page with a form for entering non-western characters, you must define the charset with the contentType attribute of the page directive, the same as for any page with non-western content, as shown in

Chapter 11

. The user can then enter values with the characters of the corresponding language (e.g., Japanese characters).

There's something else to be aware of here. Parameter values sent from a form are encoded according to a special format. Characters other than a-z, A-Z, and 0-9 are converted to byte values in a hexadecimal format, preceded by a percent sign (%). For instance, the characters for "Hello World" in Japanese (shown

in Figure 11.4

) are sent like this:

%8D%A1%93%FA%82%CD%90%A2%8AE

This code represents the byte codes for the five Japanese characters. In order to process this information, the target JSP page must know which charset was used by the browser to encode it. The problem is that today's browser versions don't provide this information. You must therefore provide this information

yourself, and convert the bytes in the parameter values accordingly. Let's see how that can be done.

Example 11.7

shows a JSP page with a form for entering a date and a text value in Japanese.

Example 11.7. Japanese Input Page ( input_ja.jsp)

<%@ page language="java" contentType="text/html;charset=Shift_JIS" %>

<%@ page import="java.util.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:useLocaleBundle id="locale" bundleName="input"

supportedLangs="ja" />

<html>

<head>

<title>

Japanese Input Test

</title>

</head>

<body bgcolor="white">

<h1>

Japanese Input Test

</h1>

<form action="process_ja.jsp" method="post">

Enter a date:

<br>

<input type="text" name="date">

(<ora:getLocalDate name="locale" date="<%= new Date( ) %>" />)

<p>

Enter some text:

<br>

<input type="text" name="text">

<p>

<input type="submit" value="Send" >

<input type="hidden" name="charset" value="Shift_JIS">

</form>

</body>

</html>

This page sets the charset to Shift_JIS and creates a LocaleBean for the Japanese locale through the

<ora:useLocaleBundle>

action with just one supported language: ja, the language code for

Japanese. In the form, the <ora:getLocalDate> action is used to generate an example of how the date must be entered. The most important part of this page, however, is the hidden charset field, set to

the same encoding value as is used for the page. This field value is sent to the target JSP page,

process_ja.jsp , together with the other field values when the form is submitted.

Example 11.8

shows the

process_ja.jsp page.

Example 11.8. Processing Japanese Input (process_ja.jsp)

<%@ page language="java" contentType="text/html;charset=Shift_JIS" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Processing Japanese Input</title>

</head>

<body>

<h1>Processing Japanese Input</h1>

<ora:useLocaleBundle id="locale" bundleName="input"

supportedLangs="ja" />

Text string converted to a Java Unicode string:

<%= locale.getParameter("text") %>

<p>

Date string converted to the internal Java Date type:

<%= locale.getDate(locale.getParameter("date")) %>

</body>

</html>

The LocaleBean, initialized by the <ora:useLocaleBundle> action, takes care of all conversion for you. The action element reads the value of the charset parameter from the hidden field and sets the corresponding bean property. You can then use the following bean method to get the decoded values of all the other request parameters: public String getParameter(String parameter)

throws UnsupportedEncodingException

This method uses the specified charset value to decode the value for the parameter you ask for and returns it as a regular Java Unicode string. The string can then be used with all the other bean methods introduced

in Example 11.6

. For instance, the value of the date parameter can be converted to a Java Date object with the getDate( ) method, as shown in

Example 11.8

.

Note that if you use the getParameter( ) method provided by the implicit request object instead of the bean's method, you get a corrupt string. The reason for this is that the request object doesn't know how the parameter values were encoded, so it tries to interpret the values as if they were encoded based on

the Latin-1 charset.

The result of the processing by the page in

Example 11.8

is shown in

Figure 11.5

.

Figure 11.5. Processed Japanese input

In this example, we simply display the processed values. In a real-world application you can do anything you like with the values, such as storing them in a database.

© 2001, O'Reilly & Associates, Inc.

12.1 Buffering

There's one important thing about how a JSP page is processed that has not been covered in any

example so far: buffering of the response body. As you may recall from Chapter 2

, an HTTP response message contains both headers and a body. The headers tell the browser things like what type of data the body contains (HTML text, an image), the size of the body, if the body can be cached, and so forth.

Headers are also used to set cookies and to tell the browser to automatically get another page (a redirect). All response headers must be sent to the browser before the body is sent.

As soon as a JSP page writes something to the body of the message, the JSP container may start sending the response to the browser. It is then too late to set headers, since they have to be sent first. In a servlet, you have full control over when something is written to the response body, so you can make sure that you set all the headers you need before you generate the body. In a JSP page, however, it's not that easy. Everything you put in a JSP page that is not a JSP element is written to the response body automatically by the JSP container. Here's the top part of the autheticate.jsp page from

Chapter 10

:

<%@ page language="java" %>

<%@ page import="com.ora.jsp.sql.*" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%-- Remove the validUser session bean, if any --%>

<% session.removeValue("validUser"); %>

...

It doesn't contain any HTML, so you may think that this does not add anything to the response body.

But actually it does. This fragment contains six lines: five lines with JSP elements and one blank line.

The JSP elements themselves are evaluated by the JSP container and never show up in the response, but the linefeed character at the end of each line is not a JSP element, so it's added to the response body.

Later in the same page, custom actions are used to set cookies, or in other words, set response headers:

<% if (request.getParameter("remember") != null) { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="2592000" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="2592000" />

<% } else { %>

<ora:addCookie name="userName"

value='<%= request.getParameter("userName") %>'

maxAge="0" />

<ora:addCookie name="password"

value='<%= request.getParameter("password") %>'

maxAge="0" />

<% } %>

This does not work if the linefeed characters added to the body have caused the response to be sent to the browser (if the response has been committed, as it's called in the servlet specification). Besides not being able to set headers after the response has been committed, the servlet specification also prohibits a request to be forwarded when data has already been written to the response body. This is because when you forward to another JSP page or servlet, the target servlet should have full control over the request. If the originating page has already started to generate the response body, the target is no longer in charge.

Buffering solves this problem. Instead of sending the response to the browser as soon as something is written to the response body, the JSP container writes everything that's not a JSP element and all dynamic content generated by JSP elements to a buffer. At some point, such as when the buffer is full or the end of the page is reached, the container sends all headers that have been set, followed by the buffered body content. So in this example, all linefeed characters end up in the buffer, and the cookie headers are set. When the whole page has been processed, the JSP container sends all headers first and then the contents of the buffer. Works like a charm.

You can control the size of the buffer and what to do when the buffer is full with two page directive attributes:

<%@ page buffer="12kb" autoFlush="false" %>

Note that the buffer attribute accepts a value that specifies the minimum size of the buffer; the container may choose to use a bigger buffer than specified. The value must be the number of kilobytes followed by kb. A buffer that holds at least 8 KB is used by default. The keyword none is also accepted. If you use this keyword, the JSP container will not perform any buffering of the response body.

The autoFlush attribute can be set to true or false, with true being the default. It specifies what to do when the buffer is full. If the value is true, the headers currently set and the buffered content is sent (flushed) to the browser when the buffer is full, and the rest of the page gets buffered until the buffer is full again. If you specify the value false, the JSP container throws an exception

when the buffer is full, ending the processing of the page.

In most cases, you want to use the default values. If you have an extremely large page where you set headers at the end of the page, you may need to increase the buffer size. 8 KB, however, is enough for most pages. Disabling buffering may make sense if you have a page that generates the result slowly and you want to send what's ready to the browser as soon as possible. But even if you disable the JSP buffering, the servlet container may still do some buffering of the result, so there's no guarantee that it will be sent immediately. No matter what value you use for the buffer attribute, however, you can force the buffer to be flushed with a scriptlet like this:

<% out.flush( ); %>

Setting the autoFlush attribute to false is rare. A possible use for this is if you have no control over the size of the dynamic content you generate and you want to ensure that the processing is aborted if you reach a certain limit.

© 2001, O'Reilly & Associates, Inc.

12.2 Including Page Fragments

You can use a JSP directive and an action to include page fragments in a JSP page. This is a useful technique when parts of all pages in an application are the same, such as headers, footers, and navigation bars.

The JSP include directive reads the content of the specified page in the translation phase (when the

JSP page is converted into a servlet) and merges it with the original page:

<%@ include file="header.html" %>

The file attribute is a relative URL. If it starts with a slash, it's a context-relative path, interpreted relative to the URI prefix assigned for the application. If it doesn't start with a slash, it's a page-relative path, interpreted relative to the path for the page that includes the file.

The included file can contain only static content (such as HTML) or it can be a regular JSP page. Its contents are merged with the page that includes it, and the resulting page is converted into a servlet, as described in

Chapter 3

. This means that all scripting variables declared in JSP declarations, scriptlets, or actions, such as <jsp:useBean> or custom actions that introduce scripting variables, are shared by the main page and all included pages. If the main page declares a variable and the same name is used to declare another variable in an included page, it will result in a translation phase error, since the combined page cannot be compiled.

What happens when the file you include using the include directive is changed actually isn't specified by the JSP specification. With Tomcat, you must change the modification date for the main page, for example using the touch command on a Unix system, before the changes take effect. An alternative is to delete the class file (the compiled version of the page) for the page. Other JSP containers may detect changes in included files automatically and go through the translation phase just like when you modify the main JSP page.

The <jsp:include> action is an alternative to the include directive, used to include another resource at runtime:

<jsp:include page="navigation.jsp" flush="true" />

The action is executed in the request processing phase instead of in the translation phase. The page attribute value is interpreted as a relative URI, the same way as the include directive's file attribute. The <jsp:include> action does not include the actual contents of the specified page: it

includes the response produced by executing the page. This means you can specify any type of web resource (e.g., a servlet or a JSP page) that produces a response of the same content type as the JSP page. The JSP container executes the specified resource by an internal function call. Hence, the included resource helps to process the original request, and therefore has access to all objects in the request scope as well as all original request parameters.

Since the page is not included until the main page is requested, you can use a request-time attribute value for the page attribute to decide which page to include depending on a runtime condition, and add request parameters that can be read by the included page:

<jsp:include page="<%= pageSelectedAtRuntime %>" flush="true">

<jsp:param name="aNewParamer" value="aStaticValue" />

<jsp:param name="anotherParameter" value="<%= aDynamicValue %>" />

</jsp:include>

If you change the included JSP page, the new version is used immediately. This is because the included page is treated in the same way as a JSP page invoked directly by a browser: the container detects the modification and goes through the translation phase for the new version of the page.

The flush attribute requires an explanation. It specifies whether the response body should be flushed

(sent to the browser) before the page is included or not. Due to limitations in the Servlet 2.2 API, this value must be set to true in JSP 1.1, meaning that the response body is always flushed before the page is included. The consequence is that the included page cannot set headers, such as cookies or redirect headers, or forward to another page. It also means that the main page cannot set headers or forward to another page after the <jsp:include> action element is executed. Work is in progress to remove the flushing requirement for a future version of the JSP specification.

Table 12.1

outlines the differences between the include directive and the <jsp:include> action.

Syntax

Table 12.1, Differences Between the include Directive and the <jsp:include>Action

When What

<%@ include file="relativeURI" %> Translation phase

Static text (HTML, JSP) merged with the

JSP page before it's converted to a servlet

<jsp:include page="relativeURI"

flush="true" />

Request processing phase

The response text generated by executing the page or servlet

Let's look at a concrete example of how you can use the two methods for including pages. Example

12.1

shows a page that includes three other pages.

Example 12.1. Including Pages (page1.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ include file="header.html" %>

<table width="90%">

<tr>

<td valign="top" align="center" bgcolor="lightblue">

<jsp:include page="navigation.jsp" flush="true" />

</td>

<td valign="middle" align="center" width="80%">

This is page 1

</td>

</tr>

</table>

<%@ include file="footer.html" %>

The application here contains two more main pages, page2.jsp and page3.jsp, that differ from

page1.jsp only in the HTML they contain (i.e., "This is page 2", "This is page 3"). The common header and footer for all pages in the example application consist of static HTML, shown in

Example 12.2

and

Example 12.3

. The include directive is used to include the header and footer files in each main page.

Example 12.2. Header (header.html)

<html>

<head>

<title>Welcome to My Site</title>

</head>

<body bgcolor="white">

<h1>My Site</h1>

Note that the header.html file is not a complete HTML page. It contains only the start tags for the

<html>

and <body> elements.

Example 12.3. Footer (footer.html)

<hr>

Copyright &copy; 2000 My Company

</body>

</html>

The end tags for the <body> and <html> tags are included in the footer.html file. Merging

header.html, one of the main pages, and footer.html results in a complete HTML page.

Each page in the application also has a navigation bar, with links to all pages in the application. The page names in the navigation bar are links to the corresponding pages, except for the current page, which is just written as plain text as shown in

Figure 12.1

.

Figure 12.1. A page composed by including other pages

The JSP code for the navigation bar is separated out into its own file, shown in Example 12.4

, and

included in each page with the <jsp:include> action as shown in

Example 12.1

.

Example 12.4. Navigation Bar with Scriptlets (navigation_script.jsp)

<%@ page language="java" %>

<% String uri = request.getServletPath( ); %>

<table bgcolor="lightblue">

<tr>

<td>

<% if (uri.equals("/ch12/page1.jsp")) { %>

<b>Page 1</b>

<% } else { %>

<a href="page1.jsp">Page 1</a>

<% } %>

</td>

</tr>

<tr>

<td>

<% if (uri.equals("/ch12/page2.jsp")) { %>

<b>Page 2</b>

<% } else { %>

<a href="page2.jsp">Page 2</a>

<% } %>

</td>

</tr>

<tr>

<td>

<% if (uri.equals("/ch12/page3.jsp")) { %>

<b>Page 3</b>

<% } else { %>

<a href="page3.jsp">Page 3</a>

<% } %>

</td>

</tr>

</table>

The navigation bar page first gets the context-relative path for the current page by calling the getServletPath( )

method on the implicit request object. This works because the request object reflects the information about the page that includes the navigation bar page, not about the included page. An HTML table is then built with one cell for each main page in the application. In each cell, a scriptlet is used to test if the cell represents the current page or not. If it does, the page name is written as bold text; otherwise, it's written as an HTML link.

To be honest, Example 12.4

contains too much scripting code for my taste. An alternative is to use a

custom action that does all the testing and generates the appropriate HTML, as shown in Example 12.5

.

Example 12.5. Navigation Bar with Custom Action (navigation.jsp)

<%@ page language="java" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<table bgcolor="lightblue">

<tr>

<td>

<ora:menuItem page="page1.jsp">

<b>Page 1</b>

</ora:menuItem>

</td>

</tr>

<tr>

<td>

<ora:menuItem page="page2.jsp">

<b>Page 2</b>

</ora:menuItem>

</td>

</tr>

<tr>

<td>

<ora:menuItem page="page3.jsp">

<b>Page 3</b>

</ora:menuItem>

</td>

</tr>

</table>

The <ora:menuItem> action inserts the HTML found in its body into the page. If the page specified by the page attribute is not the current page, the HTML is inserted as is. Otherwise, it's embedded as an HTML link element, the same way as with the scriptlets in

Example 12.4

. But unlike the scriptlet version of this page, the <ora:menuItem> action also performs URL rewriting on the HTML link

URI if needed (this includes the session ID in the URI).

You may wonder why I use the include directive for the header and footer and the

<jsp:include>

action for the navigation bar. Either one will do for all files in this example, but I chose the action for the navigation bar because this page needs to be updated as new pages are added to the application. Using the action guarantees that the new version of the file is used immediately. I picked the directive for the header and footer pages because there's a slight performance penalty when using the action (the container must make a function call at request time). In this example, I assumed that both the header and footer contain stable information. In the rare event that they change, I'm willing to force the JSP container to go through the translation phase by deleting the class files corresponding to each main page or by changing the modification date for each page as described earlier.

If the included file sets headers or forwards to another page, you must use the include directive, since the <jsp:include> action flushes the buffer and commits the response before including the page. Same thing if you need to set headers or forward in the main page after including another page.

On the other hand, if you can't decide which page to include until runtime, you must use the

<jsp:include>

action.

© 2001, O'Reilly & Associates, Inc.

12.3 XML and JSP

If you're developing web sites for a living, you've surely encountered the Extensible Markup Language

(XML). XML is a set of syntax rules for how to represent structured data using markup elements represented by a start tag (optionally with attributes), a body, and an end tag:

<employee id="123">

<first-name>Hans</first-name>

<last-name>Bergsten</last-name>

<telephone>310-555-1212</telephone>

</employee>

This XML example contains four elements: <employee>, <first-name>, <last-name>, and

<telephone>

.

By selecting sensible element names, an XML file may be understandable to a human, but to make sense to a program it must use only a restricted set of elements in which each element has a welldefined meaning. This is known as an XML application (the XML syntax applied to a certain application domain). A couple of examples are the Wireless Markup Language (WML), used for browsers in cellular phones and other small devices, and XHTML, which is HTML 4.0 reformulated as an XML application. Other examples are JSP action elements and the Web Application Deployment

Descriptor elements introduced in

Chapter 2

.

12.3.1 Generating an XML Document

As we discussed in Chapter 3

and Chapter 5

, everything in a JSP page that is not a JSP element is template text. In all the examples so far, we have used HTML as the template text. But we can use any

text, such as XML elements. Example 12.6

shows a JSP page that sends a simple phone book to a

wireless device, using the XML elements defined by the WML specification as the template text.

Example 12.6. WML Phone Book JSP Page (phone_wml.jsp)

<?xml version="1.0"?>

<!DOCTYPE wml PUBLIC "-//WAPFORUM//DTD WML 1.1//EN"

"http://www.wapforum.org/DTD/wml_1.1.xml">

<%@ page contentType="text/vnd.wap.wml" %>

<wml>

<card id="list" title="Phone List" newcontext="true">

<p>

<anchor>Bergsten, Hans

<go href="#Bergsten_Hans"/>

</anchor>

<br/>

<anchor>Eckstein, Robert

<go href="#Eckstein_Robert"/>

</anchor>

<br/>

<anchor>Ferguson, Paula

<go href="#Ferguson_Paula"/>

</anchor>

</p>

</card>

<card id="Bergsten_Hans" title="Bergsten, Hans">

<p>

Phone: 310-555-1212

<do type="prev" label="Back">

<prev/>

</do>

</p>

</card>

<card id="Eckstein_Robert" title="Eckstein, Robert">

<p>

Phone: 512-555-5678

<do type="prev" label="Back">

<prev/>

</do>

</p>

</card>

<card id="Ferguson_Paula" title="Ferguson, Paula">

<p>

Phone: 213-555-1234

<do type="prev" label="Back">

<prev/>

</do>

</p>

</card>

</wml>

A discussion of the WML elements is outside the scope of this book, but let's look at some important

details of the JSP page. The first line in Example 12.6

is an XML declaration , telling which version of

XML the document conforms to. Some WML browsers are very picky that this is the very first thing in

an XML document, and even whitespaces—regular spaces, linefeed characters, and tab characters—before the declaration can throw them off. In all examples you have seen so far, the JSP page

directive has been on the first line. Here, I have moved it down, so that the linefeed character that ends the directive line doesn't cause any problems.

The second and third lines in

Example 12.6

contain an XML document type declaration. This identifies

the so-called Document Type Definition (DTD) for the document, basically the definition of all XML elements that a conforming document of this type can contain. Here, it's the DTD for the WML elements.

The JSP page directive on the fourth line is important. The content type for a JSP page is html/text

by default. For a WML document, you must specify the content type text/vnd.wap.wml

using the contentType attribute. Otherwise, the WML browser doesn't accept the document.

The rest of the page in Example 12.6

is just static WML code. To run this example, you need a WML

browser. You can use the WML browser included in Nokia's WAP Toolkit, available at http://www.forum.nokia.com/ .

Figure 12.2

shows what the phone list menu card and a details card look like in Nokia's WML browser. The toolkit also includes WML documentation, in case you want to learn more about how to serve content to devices like cellular phones and PDAs.

Figure 12.2. Phone list in WML browser

12.3.2 Transforming XML into HTML

You may also have heard about the Extensible Stylesheet Language (XSL). XSL defines one set of

XML elements used to transform an XML document into some other type of document, and another set of elements used to produce a formatted version of an XML document suitable for display. The formatting part of XSL is used by browsers and other programs that need to render an XML document, using different styles for different elements, such as a bold large font for a header and a regular font for paragraph text. The transformation part of XSL is referred to as XSLT. XSLT can be used to turn a source XML document, such as a document representing an order, into different forms by using different stylesheets. This is useful in business-to-business (B2B) applications, where different partners often require the same information in slightly different formats. You can read more about XSL and

XSLT at http://www.w3.org/TR/xsl/ .

XSLT can also be used to transform structured XML data into HTML. Example 12.7

shows an

example in which the same phone book information used in

Example 12.6

is transformed into an

HTML table.

Example 12.7. Transforming XML to HTML (phone_html.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/xsltaglib" prefix="xsl" %>

<html>

<head>

<title>Phone List</title>

</head>

<body bgcolor="white">

<xsl:apply xsl="/ch12/htmltable.xsl">

<?xml version="1.0" encoding="ISO-8859-1"?>

<employees>

<employee id="123">

<first-name>Hans</first-name>

<last-name>Bergsten</last-name>

<telephone>310-555-1212</telephone>

</employee>

<employee id="456">

<first-name>Robert</first-name>

<last-name>Eckstein</last-name>

<telephone>512-555-5678</telephone>

</employee>

<employee id="789">

<first-name>Paula</first-name>

<last-name>Ferguson</last-name>

<telephone>213-555-1234</telephone>

</employee>

</employees>

</xsl:apply>

</body>

</html>

The transformation is performed by a custom action, named <xsl:apply>, from the Jakarta Taglibs project. The binary version of the Jakarta XSL tag library and the necessary JAR files with XML and

XSL processing classes are bundled with the examples for this book, and can be downloaded from the book's catalog page ( http://www.oreilly.com/catalog%20/jserverpages/ ). The body of the

<xsl:apply>

action contains an XML document with elements representing information about employees. The xsl attribute specifies an XSL stylesheet with XSLT elements that transform the

XML document into an HTML table. The resulting table is inserted into the JSP page. Descriptions of all the XSLT elements would fill an entire book, but

Example 12.8

shows the stylesheet used here to give you a glimpse of what XSLT looks like.

Example 12.8. XSL Stylesheet that Generates an HTML Table (htmltable.xsl)

<?xml version="1.0"?>

<xsl:stylesheet version="1.0"

xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

<xsl:template match="employees">

<html>

<head>

<title>Phone List</title>

</head>

<body bgcolor="white">

<table border="1" width="100%">

<tr>

<th>ID</th>

<th>Employee Name</th>

<th>Phone Number</th>

</tr>

<xsl:for-each select="employee">

<tr>

<td>

<xsl:value-of select="@id"/>

</td>

<td>

<xsl:value-of select="last-name"/>,

<xsl:value-of select="first-name"/>

</td>

<td>

<xsl:value-of select="telephone"/>

</td>

</tr>

</xsl:for-each>

</table>

</body>

</html>

</xsl:template>

</xsl:stylesheet>

The <xsl:template> uses the non-XSLT elements in its body as a template to generate a new document from the <employees> element in the source XML document. The <xsl:for-each> element loops over all <employee> elements in the source, and the <xsl:value-of> elements extracts the values of attributes and nested elements. You get the idea.

The <xsl:apply> action, together with other actions in the Jakarta Taglibs XSL library, can apply a stylesheet to XML documents from other sources than its body, such as an external file or a database column value saved as a String in one of the JSP scopes. You can read more about the Jakarta XSL tag library and download the source code from http://jakarta.apache.org/taglibs/index.html

.

12.3.3 Transforming XML into a Request-Dependent Format

As a final example of using XML with JSP, let's look at a page that uses the <xsl:apply> action to apply different stylesheets depending on if the page is requested by a WML browser or an HTML browser.

Example 12.9

shows such a page.

Example 12.9. XSL Stylesheet that Generates HTML or WML (phone.jsp)

<%@ taglib uri="/xsltaglib" prefix="xsl" %><%

String xslURI = null;

if (request.getHeader("User-Agent").indexOf("WAP") != -1) {

xslURI = "/ch12/wml.xsl";

response.setContentType("text/vnd.wap.wml");

}

else {

xslURI = "/ch12/html.xsl";

response.setContentType("text/html");

}

%><xsl:apply

xsl="<%= xslURI %>">

<?xml version="1.0" encoding="ISO-8859-1"?>

<employees>

<employee id="123">

<first-name>Hans</first-name>

<last-name>Bergsten</last-name>

<telephone>310-555-1212</telephone>

</employee>

<employee id="456">

<first-name>Robert</first-name>

<last-name>Eckstein</last-name>

<telephone>512-555-5678</telephone>

</employee>

<employee id="789">

<first-name>Paula</first-name>

<last-name>Ferguson</last-name>

<telephone>213-555-1234</telephone>

</employee>

</employees>

</xsl:apply>

There are a number of things to note here. First, this page uses the HTTP User-Agent header to figure out which type of browser is requesting the page, and selects an appropriate XSL stylesheet to transform the XML data for the current type of browser. Be aware that this test may not work for all

WML browsers. The WML browser in Nokia's WAP Toolkit happens to include the WAP acronym in the User-Agent header, but that's not necessarily the case for other WML browsers. The two stylesheets used here, wml.xsl and html.xsl, generate complete WML and HTML pages, respectively.

Since the page can serve both HTML and WML content, the page directive's contentType attribute cannot be used to set the content type as we have done in all other examples. Instead, the content type is set to the appropriate type using the setContentType( ) method of the implicit response

object, depending on the type of browser asking for the page.

Finally, note how the start tags for all JSP directives, scriptlets, and custom actions on this page are written on the same line as the end tag for the preceding element. This is to ensure that no extra linefeeds are added to the response. As described earlier, leading whitespace in a WML page can cause a WML browser to reject the page.

For a simple example like this, letting an XSLT stylesheet transform the XML source into a complete web page works fine. However, on most real web sites, the HTML version of the site differs significantly from the WML version. You want to provide a rich interface for HTML browsers with a nice layout, navigation bars, images, colors, nice fonts, and typically as much content as you can fit on each page. A WML browser, on the other hand, has a very small screen with limited layout, font, and graphics capabilities. Developing an efficient interface for this type of device is very different. A more practical approach for combining XML, XSL, and JSP to serve different types of browsers is to keep the actual content (articles, product information, phone lists, etc.) in a device-independent XML format, but use separate JSP pages for each device type. The JSP pages can then use a custom action like the

<xsl:apply>

action to transform the key content and merge it with the device dependent template

text to form a complete page suitable for each specific device type, like in Example 12.9

.

© 2001, O'Reilly & Associates, Inc.

12.4 Mixing Client-Side and Server-Side Code

I touched on the differences between server-side code and client-side code in

Chapter 3

. JSP is a serverside technology, so all JSP elements such as actions and scriptlets execute on the server before the resulting page is sent to the browser. A page can also contain client-side code, such as JavaScript code or Java applets. This code is executed by the browser itself. There is no way that a JavaScript event handler such as onClick or onSelect can directly invoke a JSP element such as an action, a scriptlet, or a Java method declared with a JSP declaration.

However, a JSP page can generate JavaScript code dynamically the same way it generates HTML,

WML, or any type of text content. Therefore, you can add client-side scripting code to your JSP pages to provide a more interactive user interface. You can also use applets on your pages to provide a more interesting and easier to use interface than what's possible with pure HTML.

12.4.1 Generating JavaScript Code

Example 12.10

shows a modified version of the User Info page used in the examples in Chapter 5 .

Example 12.10. Input Form with Client-Side Validation Code (clientscript.jsp)

<%@ page language="java" contentType="text/html" %>

<%@ page import="com.ora.jsp.util.*" %>

<script language="JavaScript">

<!-- Hide from browsers without JavaScript support

function isValidForm(theForm) {

if (isEmpty(theForm.userName.value)) {

theForm.userName.focus( );

return false;

}

if (!isValidDate(theForm.birthDate.value)) {

theForm.birthDate.focus( );

return false;

}

if (!isValidEmailAddr(theForm.emailAddr.value)) {

theForm.emailAddr.focus( );

return false;

}

var choices = new Array("male", "female");

if (!isValidChoice(theForm.sex.value, choices)) {

theForm.sex.focus( );

return false;

}

if (!isValidNumber(theForm.luckyNumber.value, 1, 100)) {

theForm.luckyNumber.focus( );

return false;

}

return true;

}

function isEmpty(aStr) {

if (aStr.length == 0) {

alert("Mandatory field is empty");

return true;

}

return false;

}

function isValidDate(dateStr) {

var matchArray = dateStr.match(/^[0-9]+-[0-1][0-9]-[0-3][0-9]$/)

if (matchArray == null) {

alert("Invalid date: " + dateStr);

return false;

}

return true;

}

function isValidEmailAddr(emailStr) {

var matchArray = emailStr.match(/^(.+)@(.+)\.(.+)$/)

if (matchArray == null) {

alert("Invalid email address: " + emailStr);

return false;

}

return true;

}

function isValidNumber(numbStr, start, stop) {

var matchArray = numbStr.match(/^[0-9]+$/)

if (matchArray == null) {

alert("Invalid number: " + numbStr);

return false;

}

if (numbStr < start || numbStr > stop) {

alert("Number not within range (" + start + "-" +

stop + "): " + numbStr);

return false;

}

return true;

}

function isValidChoice(choiceStr, choices) {

var isValid = false;

for (var i = 0; i < choices.length; i++) {

if (choices[i].toLowerCase( ) == choiceStr.toLowerCase( )) {

isValid = true;

break;

}

}

if (isValid == false) {

alert("Invalid choice: " + choiceStr);

}

return isValid;

}

-->

</script>

<html>

<head>

<title>User Info Entry Form</title>

</head>

<body bgcolor="white">

<jsp:useBean id="userInfo"

class="com.ora.jsp.beans.userinfo.UserInfoBean"

scope="request" />

</jsp:useBean>

<%-- Output list of values with invalid format, if any --%>

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

</font>

<%-- Output form with submitted valid values --%>

<form action="userinfovalidate.jsp" method="post"

onSubmit="return isValidForm(this)">

<table>

<tr>

<td>Name:</td>

<td><input type="text" name="userName"

value="<%= StringFormat.toHTMLString(

userInfo.getUserName( )) %>" >

</td>

</tr>

...

<tr>

<td colspan=2><input type="submit"></td>

</tr>

</table>

</form>

</body>

</html>

When the user submits the form, the JavaScript isValidForm( ) method is first executed by the browser to validate all input field values. Only if all values pass the test is the form actually submitted to the userinfovalidate.jsp page specified as the form's action URI. In this way, the user is alerted to mistakes much faster, and the server is relieved from processing invalid requests.

However, the validation is also performed by the server when the form is finally submitted, in exactly

the same way as described in Chapter 5 . This is important, because you don't know if the user's browser

supports JavaScript or if scripting has been disabled in the browser.

Note that the JavaScript validation code shown in Example 12.10

is far from perfect. It's really intended only as an example. You can find much better validation code on sites such as the JavaScript Source

( http://javascript.internet.com/ ).

In

Example 12.10

, all JavaScript code is written as static template text. However, nothing prevents you

from generating parts of the JavaScript code, for instance a JavaScript array, with values retrieved from a database by the JSP page. Just remember which code executes where and when. To the code in the JSP page executing on the server, the JavaScript code it generates is just plain text; it doesn't even try to understand it. It's only when the page that contains the dynamically generated JavaScript code reaches the browser that it becomes meaningful and can be executed by the browser. The browser, on the other hand, couldn't care less that the JavaScript code was created by a JSP page; it has no idea how the code was created. It should be clear, then, that JavaScript code cannot call Java code in the JSP page, and vice versa.

12.4.2 Using Java Applets

A Java applet is a Java class that is embedded in an HTML page and executed by the browser. It can be used to provide a nice user interface on a web page. The problem here is that the native Java support in

the web browsers doesn't keep up with the Java release cycles. Many users still have browsers that support only JDK 1.0, and more current browsers have so many limitations and bugs in their implementations that you're still limited to JDK 1.0 features to make the applet work.

To address this issue, Sun provides a Java runtime environment that can be integrated in a browser using the browser's native plug-in API. The product is appropriately named the Java Plug-in, and as of this writing the JDK 1.3 version is available for Netscape Navigator and Internet Explorer on Windows 95,

98, and NT, Linux, and Solaris. For an up-to-date list of supported platforms, visit Sun's Java Plug-in page at http://java.sun.com/products/plugin/index.html

.

With the Java Plug-in, you can use the latest Java features in your applets, such as the Swing GUI classes, collection classes, enhanced security, and more. But there's one more hurdle you have to jump.

The HTML element you need in a page to get the Java Plug-in (or any plug-in component) installed and loaded by the browser differs between Internet Explorer and Netscape Navigator. For Netscape, you need to use the <embed> element, while Internet Explorer requires the <object> element.

Fortunately, JSP provides an easy solution to this problem, namely the <jsp:plugin> action.

The <jsp:plugin> action looks at the User-Agent request header to figure out which type of browser is requesting the page, and inserts the appropriate HTML element for using the Java Plug-in to

run the applet. Example 12.11

shows an example borrowed from the Tomcat JSP examples.

Example 12.11. Embedding an Applet in a JSP Page (applet.jsp)

<%@ page language="java" contentType="text/html" %>

<html>

<head>

<title>Embedding an applet</title>

</head>

<body bgcolor="white">

<h1>Embedding an applet</h1>

<jsp:plugin type="applet" code="Clock2.class"

codebase="applet"

jreversion="1.2" width="160" height="150" >

<jsp:params>

<jsp:param name="bgcolor" value="ccddff" />

</jsp:params>

<jsp:fallback>

Plugin tag OBJECT or EMBED not supported by browser.

</jsp:fallback>

</jsp:plugin>

</body>

</body>

The <jsp:plugin> action has three mandatory attributes: type, code, and codebase. The type attribute must be set to either applet or bean (to include a JavaBeans object), code is used to specify the class name, and codebase is the absolute or relative URL for the directory or archive file that contains the class. Note that the applet class must be stored in a directory that can be accessed by the web browser; that is, it must be part of the public web page structure for the application. As you may recall, class files for beans and custom actions are typically stored in the WEB-INF lib and classes subdirectories, accessible only to the container. The different locations make sense when you think about where the code is executed: the applet is loaded and executed by the browser, and beans and custom action classes are loaded and executed by the container.

The <jsp:plugin> action also has a number of optional attributes, such as the width, height, and jreversion

attributes used here. Appendix A , contains a description of all attributes.

The body of the action element can contain nested elements. The <jsp:params> element, which in turn contains one or more <jsp:param> elements, is used to provide parameter values to the applet.

In

Example 12.11

, the applet's bgcolor parameter is set to the hexadecimal RGB value for light blue.

The <jsp:fallback> element can optionally be used to specify text that should be displayed instead of the applet in a browser that doesn't support the HTML <object> or <embed> element.

Figure 12.3

shows what the page in

Example 12.11

looks like in a browser.

Figure 12.3. A page with an applet using the Java Plug-in

An applet can communicate with the server in many different ways, but how it's done is off-topic for this book. If you would like to learn how to develop applets that communicate with a servlet, I suggest you read Jason Hunter and William Crawford's Java Servlet Programming (O'Reilly).

© 2001, O'Reilly & Associates, Inc.

12.5 Precompiling JSP Pages

To avoid hitting your site visitors with the delay caused by converting a JSP page into a servlet on the first access, you can precompile all pages in the application. Another use of precompilation is if you do not want anyone to change the pages in a JSP-based application after the application is deployed. In this case you can precompile all pages, define URI mappings for all JSP pages in the WEB-INF/web.xml file, and install the Java class files only for the compiled pages. We look at both these scenarios in this section.

One way of precompiling all pages in an application is to simply run through the application in a development environment and make sure you hit all pages. You can then copy the class files together with all the other application pages to the production server when you deploy the application. Where the class files are stored varies between containers. However, Tomcat stores all JSP page implementation classes in its work directory by default, in a subdirectory for the particular web application. As long as the modification dates of the class files are more recent than for the corresponding JSP pages, the production server uses the copied class files.

The JSP specification also defines a special request parameter that can be used to give the JSP container a hint that the page should be compiled without letting the page process the request. An advantage of using this method is that you can automatically invoke each page, perhaps using a simple load testing tool, without having to provide all the regular request parameters the pages use. Since the pages are not executed, application logic that requires pages to be invoked in a certain order or enforces similar rules cannot interfere with the compilation. The request parameter name is jsp_precompile, and valid values are true and false, or no value at all. In other words, the following URIs are all valid:

/ora/ch12/applet.jsp?jsp_precompile

/ora/ch12/applet.jsp?jsp_precompile=true

/ora/ch12/applet.jsp?jsp_precompile=false

The third example is not very useful, since if the parameter value is false, the request is treated exactly as any other request, and is therefore processed by the JSP page. A JSP container that receives a request like the ones in the first and second examples should compile the JSP page (go through the translation phase) but not allow the page to process the request. Most JSP containers support this feature, even though the specification doesn't require it. A compliant JSP container is allowed to ignore the compilation request, as long as it doesn't let a JSP page process a request that includes a jsp_precompile parameter with the value true or with no value at all.

When you have compiled the JSP pages, you can package your application without the JSP pages themselves by using only the generated servlet class files. You do this by adding URI mapping definitions in the WEB-

INF/web.xml file for the applications, so that a request for a certain JSP page is served directly by the corresponding servlet instead. There are two reasons why you might want to do this. One is that using the servlet files directly is slightly faster, since the container doesn't have to go through the JSP container code to figure out which servlet class to use. The other is that if you do not include the JSP pages in the application packet, no one can change the application. This can be an advantage if you resell prepackaged JSP-based applications.

Unfortunately, it's much harder to do this than it should be if you use Tomcat as your web container. This is because Tomcat's JSP container uses a very creative naming convention for the class files it generates. Because

Tomcat is such a widely used container, I describe this problem in detail here, even though other containers may handle this in a different way.

Tomcat stores all class files for an application's JSP pages in a subdirectory to its work directory, using filenames composed of the URI path for each JSP page plus a lot of extra characters to make sure the name doesn't contain any special characters that can cause problems. Here is an example:

_0002fch_00031_00032_0002fhello_0002ejsphello.class

This is the name Tomcat picks for a JSP file with the URI /ch12/hello.jsp. The problem is that the filename does not match the Java class name, something the standard Java class loader expects. For instance, the class file here contains a class named: ch_00031_00032._0002fch_00031_00032_0002fhello_0002ejsphello_jsp_0

When you let the JSP container handle the class, this name mismatch doesn't cause a problem because the container has its own class loader that's able to deal with this kind of class file. If, however, you want to use the generated class files as regular servlets, handled by a class loader that understands only the standard naming scheme, you have to rename the files. Here are the steps you need to go through to make the class files usable as regular servlets.

First, use the javap command (part of the Java runtime environment) to get the real class name for each class file.

javap _0002fch_00031_00032_0002fhello_0002ejsphello.class

This gives an error message that includes the real class name:

Error: Binary file '_0002fch_00031_00032_0002fhello_0002ejsphello' contains ch_00031_00032._0002fch_00031_00032_0002fhello_0002ejsphello_jsp_0

Then move the class file to the WEB-INF/classes directory, using the real class name as the filename, in a subdirectory matching the package name, if any. In this example, the class file should be moved to a subdirectory named ch_00031_00032, like this:

WEB-INF/

classes/

ch_00031_00032/

_0002fch_00031_00032_0002fhello_0002ejsphello_jsp_0.class

Finally, add a URI mapping rule for the JSP page in the WEB-INF/web.xml file. For this example, it should look like this:

<servlet-mapping>

<servlet-name>

ch_00031_00032._0002fch_00031_00032_0002fhello_0002ejsphello_jsp_0

</servlet-name>

<url-pattern>/ch12/hello.jsp</url-pattern>

</servlet-mapping>

You can then remove the JSP page file and the application will use the servlet class file directly instead.

Some containers, such as Allaire's JRun, provide proprietary programs you can use to convert JSP pages into servlets. Tomcat 3.2 includes an early version of a command-line tool for converting JSP pages into servlet Java files. The tool is named jspc , and it's invoked with the jspc.bat (Windows) or jspc.sh (Unix) script files in

Tomcat's bin directory. It's not yet fully tested and currently doesn't compile the servlet source files it generates.

These kinds of tools may eventually make the packaging and mapping of precompiled JSP pages easier.

There is one more thing to be aware of. The technique described in this section works fine as long as you compile and deploy the generated servlet classes using the same web container product, for instance generating the files in one Tomcat installation and deploying in another Tomcat installation. But a web container is allowed to use its own internal classes in the generated servlets, which means that you may not be able to generate the servlets with one web container (such as Tomcat) and deploy them in another (such as Unify's ServletExec).

© 2001, O'Reilly & Associates, Inc.

12.6 Preventing Caching of JSP Pages

A browser can cache web pages so that it doesn't have to get them from the server every time the user asks for them. Proxy servers can also cache pages that are frequently requested by all users going through the proxy. Caching helps cut down the network traffic and server load, and provides the user with faster responses. But caching can also cause problems in a web application where you really want the user to see the latest version of a dynamically generated page.

Both browsers and proxy servers can be told not to cache a page by setting response headers. You can use a scriptlet like this in your JSP pages to set these headers:

<%

response.addHeader("Pragma", "No-cache");

response.addHeader("Cache-Control", "no-cache");

response.addDateHeader("Expires", 1);

%>

An alternative is to use a custom action that's included with the book examples:

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:noCache/>

The <ora:noCache> action sets the exact same headers as the scriptlet example, but it's cleaner.

© 2001, O'Reilly & Associates, Inc.

12.7 How URLs Are Interpreted

One thing that can be confusing in a JSP-based application is the different types of URIs used in the

HTML and JSP elements. The confusion stems from a combination of conflicting terms used to describe URIs in the HTTP, servlet, and JSP specifications, as well as the fact that some types of URIs are interpreted differently in the HTML and the servlet world.

In HTML, URIs are used as attribute values in elements like <a>, <img>, and <form>. JSP elements that use URI attribute values are the page, include, and taglib directives and the

<jsp:forward>

and <jsp:include> actions. Custom actions can also define attributes that take

URI values.

The HTTP/1.1 specification (RFC 2616, with more details in RFC 2396) defines a Uniform Resource

Identifier (URI) as a string, following certain rules, that uniquely identifies a resource of some kind. A

Uniform Resource Locator (URL) is just a special kind of URI that includes a location (such as the server name in an HTTP URL). An absolute URI is a URI that starts with the name of a so called scheme, such as http or https, followed by a colon (:) and the rest of the resource identifier. An example of an absolute URI for a resource accessed through the HTTP protocol is: http://localhost:8080/ora/ch12/login.jsp

Here, http is the scheme, localhost:8080 is the location (a server name and a port number), and

/ora/ch12/login.jsp

is the path.

The URIs used in the HTML elements generated by a JSP page are interpreted by the browser. A browser needs the absolute URI to figure out how to send the requests for the resources referenced by the HTML elements. It uses the scheme to select the correct protocol, and the location to know where to send the request. The path is sent as part of the request to the server, so the server can figure out which resource is requested. But when you write a URI in an HTML document, such as the action attribute of a form element or the src attribute of an image element, you don't have to specify an absolute URI if the resource is located on the same server. Instead you can use just the URI path, like this:

<img src="/images/hello.gif">

This type of URI is called an absolute path, meaning it contains the complete path for the resource within a server; the only difference compared to an absolute URI is that the scheme and location are not specified. The browser interprets an absolute path URI as a reference to a resource on the same server,

so it adds the scheme and location it used to make the request that returned the page to the absolute path URI it finds in the page. It then has the absolute URI it needs to make a request for the referenced resource.

Another type of URI is a relative path, interpreted relative to the path of the current page. A relative path is a path that does not start with a slash (/):

<form action="process.jsp">

<img src="../images/hello.gif"

Here the action attribute references a JSP file at the same level in the path structure as the page that contains the reference. The src attribute value uses the ../ notation to refer to a resource one level up in the structure. The browser interprets a relative path URI as relative to the URI path for the request that produced the page. If the two relative paths in this example are used in a page generated by a request for http://localhost:8080/ora/ch12/login.jsp, the browser interprets them as the following absolute URIs:

http://localhost:8080/ora/ch12/process.jsp

http://localhost:8080/ora/images/hello.gif

Relative URI paths offer a lot of flexibility. If all references between the web resources in an application are relative, you can move the application to a different part of the path structure without changing any URIs. For instance, if you move the pages from /ora/ch12 to /foo/bar, the relative paths still reference the same resources.

So far, so good. Now let's see what happens in a Java web container when it receives a request. The first part of a URI for a servlet or JSP page has a special meaning. It's called the context path— one example is the /ora path used for all examples in this book. As described in

Chapter 2

, a servlet container can contain multiple web applications, handled by a corresponding servlet context. Each web application is associated with a unique context path assigned when the web application is installed.

When a request is received by the web container, it uses the context path to select the servlet context that's responsible for handling the request. The container hands over the request to the selected context, which then uses the URI path minus the context path to locate the requested resource (a servlet or a JSP page) within the context. For instance, an absolute URI like http://localhost:8080/ora/ch12/login.jsp is interpreted by the container as a request for a JSP page named /ch12/login.jsp within the context path

/ora.

Because a web application can be assigned any context path when the application is installed, the context path must not be part of the URIs used in JSP elements (and servlet methods) to refer to other parts of the same application. You can always use a relative-path URI, just as you do in HTML, for example to refer to another page in a <jsp:include> action:

<jsp:include page="navigation.jsp" flush="true" />

This type of URI is called a page-relative path in the JSP specification. It's interpreted by the container as relative to the page where it's used.

Sometimes it's nice to be able to refer to an internal application resource with a URI that is not interpreted relative to the containing page. An example is a reference to a customized error page that is used by all pages in the application independent of where in the path structure they are located:

<%@ page errorPage="/errorMsg.jsp" %>

A URI used as a JSP element attribute that starts with a slash is interpreted by the container as relative to the application's context path. The JSP specification calls this type of URI a context-relative path.

This type of URI is useful for all sorts of common resources used in an application, such as error pages and images, that have fixed URIs within the application path structure.

In summary, a URI used in an HTML element can be:

An absolute URI (including the scheme and server name)

An absolute-path URI (a path starting with a slash), interpreted as the path to a resource on the same server

A relative-path URI (a path not starting with a slash), interpreted as relative to the path for the page where it's used

A URI used in a JSP element (or a servlet method) can be:

A context-relative path (a path starting with a slash), interpreted as relative to the application's context path

A page-relative path (a path not starting with a slash), interpreted as relative to the path for the page where it's used

As long as you remember that URIs used in HTML elements are interpreted by the browser, and URIs used in JSP elements are interpreted by the web container, it's not so hard to figure out which type of

URI to use in any given situation.

© 2001, O'Reilly & Associates, Inc.

13.1 The Java 2 Enterprise Edition Model

At the JavaOne conference in San Francisco in June 1999, Sun Microsystems announced a new architecture for Java, with separate editions for different types of applications: the Java 2 Platform,

Standard Edition ( J2SE) for desktop and workstation devices; the Java 2 Platform, Micro Edition (

J2ME) for small devices like cell phones, pagers, and Personal Digital Assistants (PDAs); and the Java

2 Platform, Enterprise Edition ( J2EE) for server-based applications.

J2EE is a compilation of various Java APIs that have previously been offered as separate packages, an

Application Programming Model (APM), also known as the J2EE Blueprints, that describes how they can all be combined, and a test suite that J2EE vendors can use to test their products for compatibility.

J2EE includes the following enterprise-specific APIs:

JavaServer Pages ( JSP)

Java Servlets

Enterprise JavaBeans (EJB)

Java Database Connection ( JDBC)

Java Transaction API ( JTA) and Java Transaction Service ( JTS)

Java Naming and Directory Interface ( JNDI)

Java Message Service ( JMS)

Java IDL and Remote Method Invocation (RMI)

Java XML

In addition, all the J2SE APIs can be used when developing a J2EE application. These groups of APIs can be used in numerous combinations. The first three APIs—EJB, JSP, and servlets—represent different component technologies, managed by what the J2EE documents call containers . As you may remember from

Chapter 2

and Chapter 3

, servlets and JSP pages are managed by a container that provides the runtime environment for these components, translating requests and responses into standard Java objects. This container is called a web container . EJB components are similarly handled by an EJB container. Components in the two types of containers use the other APIs to access databases

( JDBC and JTA/JTS), locate various resources (JNDI), and communicate with other server resources

(JMS, Java IDL, RMI, and XML). Figure 13.1

shows a high-level view of all the pieces and their relationships.

Figure 13.1. J2EE overview

Enterprise applications are often divided into a set of tiers, and J2EE identifies three: the client tier, the middle tier, and the Enterprise Information System (EIS) tier. The middle tier can be further divided into the web tier and the EJB tier. This logical separation, with well-defined interfaces, makes it possible to build scalable applications. Initially one or more tiers can be running on the same physical server. With increased demands, the tiers can be separated and distributed over multiple servers without modifying the code, just by changing the configuration.

The client tier contains browsers as well as regular GUI applications. A browser communicates with the web container using HTTP. A standalone application can also use HTTP or communicate directly with the EJB container using RMI or IIOP (a CORBA protocol). Another type of client that's becoming more and more popular is the extremely thin client, such as a cellular phone or PDA. This type of client typically uses the Wireless Access Protocol (WAP), often converted into HTTP via a gateway, to communicate with the web container.

The middle tier provides client services through the web container and the EJB container. A client that communicates through HTTP with the server uses components in the web container, such as servlets and JSP pages, as entry points to the application. Many applications can be implemented solely as web container components. In other applications, the web components just act as an interface to the application logic implemented by EJB components. A standalone application, written in Java or any other programming language, can also communicate directly with the EJB components. General guidelines for when to use the different approaches are discussed later in this chapter. Components in this tier can access databases and communicate with other server applications using all of the other

J2EE APIs.

The Enterprise Information System (EIS) tier holds the application's business data. Typically, it consists of one or more relational database management servers, but other types of databases such as

IMS databases, applications such as Enterprise Resource Planning (ERP), and mainframe transaction processing systems such as CICS, are also included in this tier. The middle tier uses J2EE APIs such as

JDBC and JTA/JTS to interact with the EIS tier.

© 2001, O'Reilly & Associates, Inc.

13.2 The MVC Model

In addition to the separation of responsibilities into different tiers, J2EE also encourages the use of the

Model-View-Controller (MVC) design model, briefly introduced in Chapter 3 , when designing

applications.

The MVC model was first described by Xerox in a number of papers published in the late 1980s in conjunction with the Smalltalk language. But this model has since been used for GUI applications developed in all popular programming languages. Let's review: the basic idea is to separate the application data and business logic, the presentation of the data, and the interaction with the data into distinct entities labeled the Model, the View, and the Controller, respectively. This makes for a flexible design, where multiple presentations (Views) can be provided and easily modified, and changes in the business rules or physical representation of the data (the Model) can be made without touching any of the user interface code.

Even though the model was originally developed for standalone GUI applications, it translates fairly well into the multitier application domain of J2EE. The user interacts with the Controller to ask for things to be done, and the Controller relays these requests to the Model in a client-type independent way. Say, for instance, that you have two types of clients: an HTTP client such as a browser, and a GUI client application using IIOP to talk to the server. In this scenario you can have one Controller for each protocol that receives the requests and extracts the request information in a protocol-dependent manner.

Both Controllers then call the Model the same way; the Model doesn't need to know what kind of client it was called by. The result of the request is then presented to the two types of clients using different

Views. The HTTP client typically gets an HTTP response message, possibly created by a JSP page, while the GUI application may include a View component that communicates directly with the Model to get its new state and render it on the screen.

The assignment of roles to the different types of J2EE components varies depending on the scenario, the types of clients supported, and whether or not EJB is used. The following sections describe possible role assignments for the three most common scenarios where JSP pages play an important role.

13.2.1 Using Only JSP

The J2EE platform includes many APIs and component types, as I have just shown. However, there's no reason to use them all for a specific application. You can pick and choose the technology that makes most sense for your application scope and functionality, the longevity of the application, the skills in your development team, and so on.

As you saw in Part II

of this book, there are all sorts of applications that can be developed using just

JSP pages, a few JavaBeans components, and custom actions. If you're primarily a page author working alone, with limited or no Java knowledge, you can still develop fairly sophisticated applications using the custom actions in this book to access databases, serve localized content, perform authentication and access control, and so forth. And as JSP's popularity grows, I'm sure you'll see more and more generic tag libraries offered by both commercial companies and open source projects, making it possible to do even more with just the JSP part of the J2EE platform.

A pure JSP approach can be a good approach even for a team, if most of the team members are skilled in page design and layout and only a few are Java programmers. The programmers can then develop application-specific beans and custom actions to complement the generic components and minimize the amount of SQL and Java code in the JSP pages.

Using pure JSP is also a suitable model for testing out new ideas and prototyping. Using generic components, a bit of scripting code, and a few application-specific beans and actions is often the fastest way to reach a visible result. Once the ideas have been proven and the team has a better understanding of the problems, a decision can be made about the ultimate application architecture for the real thing.

The danger here is that the last step—evaluating the prototype and deciding how it should be redesigned—never happens; I have seen prototypes being relabeled as production systems overnight too many times, and also experienced the inevitable maintenance nightmares that follow.

Figure 13.2. MVC roles in a pure JSP scenario

The MVC model makes sense even for a pure JSP play. I recommend that you use separate JSP pages for presentation (the View) and request processing (the Controller), and place all business logic in

beans, (the Model), as shown in Figure 13.2

. Let the Controller pages initialize the beans, and the View

pages generate a response by reading its properties. That's the model used in most examples in Part II

.

If you follow this model, it's easy to move to a combination of servlets and JSP the day you find that the pure JSP application is becoming hard to maintain.

13.2.2 Using Servlets and JSP

The combination of servlets and JSP is a powerful tool for developing well-structured applications that are easy to maintain and extend as new requirements surface. Since a servlet is a regular Java class, you can use the full power of the Java language to implement the request processing, using standard Java development and debugging environments instead of debugging JSP pages filled with scripting code.

JSP pages can then be used for what they are best at: rendering the response by including information collected or generated by the servlets.

A common combination of servlets and JSP is to use a servlet as the Controller (or front component, as it's called in the J2EE documents) for an application, with a number of JSP pages acting as Views. The advantage of this model, compared to the pure JSP approach, becomes apparent as the application gets more complex. For instance, if you need to roll your own authentication and access control code, centralizing the security controls in a servlet instead of counting on everyone remembering to put custom actions in all protected pages is less error-prone. A servlet as the single entry point to the application also makes it easy to do application-specific logging (for instance, collecting statistics in a database), maintain a list of currently active users, and other things.

Figure 13.3. MVC roles in a servlet/JSP scenario

In this scenario, all requests are sent to the servlet acting as the Controller with an indication about what needs to be done. The indication can be in the form of a request parameter or as a part of the URI

path. Figure 13.3

shows how the MVC roles are allocated in this scenario. The same way as in the pure

JSP scenario, beans are used to represent the Model. The servlet locates the appropriate bean and asks it to perform the requested action. Depending on the result of the request, the Controller servlet picks an appropriate JSP page to generate a response to the user (a View). For instance, if a request to delete a document in a document archive is executed successfully, the servlet can pick a JSP page that shows the updated archive contents. If the request fails, it can pick a JSP page that describes exactly why it

failed. We look at this approach in more detail later in Chapter 14 .

13.2.3 Using Servlets, JSP, and EJB

An application based on Enterprise JavaBeans (EJB) is commonly viewed as the Holy Grail today.

However, this is the most complex model of the ones described in this chapter, and it therefore comes with overhead in the development, deployment, operation, and administration areas. Still, EJB may be the way to go for many types of applications.

What EJB brings to the table is primarily transaction management and a client type-independent component model. Even though it's impossible to say that a specific type of application should use EJB, if you develop an application with numerous database write-access operations accessed through different types of clients (such as browsers, standalone applications, PDAs, or another server in a B2B application), EJB is probably the way to go. An EJB-based application also enforces the separation between the Model, View, and Controller aspects, leading to an application that's easy to extend and maintain.

There are two primary types of EJB components: entity beans and session beans. An entity bean represents a specific piece of business data such as an employee or a customer. Each entity bean has a unique identity, and all clients that need access to the entity represented by the bean use the same bean instance. Session beans, on the other hand, are intended to handle business logic and are used only by the client that created them. Typically, a session bean operates on entity beans on behalf of its client.

With EJB in the picture, the MVC roles typically span over multiple components in the web container and EJB container. In a web-based interface to an EJB application, requests are sent to a servlet just as in the servlet/JSP scenario. But instead of the servlet locating a JavaBeans component to process the request, it asks an EJB session bean (or a JavaBeans component that acts as an interface to an EJB session bean) to do its thing. The Controller role therefore spans the servlet and the EJB session bean,

as illustrated in Figure 13.4

. The Model can also span multiple components. Typically, JavaBeans

components in the web tier are used to mirror the data maintained by EJB entity beans to avoid expensive communication between the web tier and the EJB tier. The session bean may update a number of the EJB entity beans as a result of processing the request. The JavaBeans components in the web tier get notified so they can refresh their state, and are then used in a JSP page to generate a response. With this approach, the Model role is shared by the EJB entity beans and the web tier

JavaBeans components.

Figure 13.4. MVC roles in a servlet/JSP/EJB scenario

We have barely scratched the surface of how to use EJB in an application here. If you believe that this model fits your application, I recommend that you read the J2EE Blueprints

( http://java.sun.com/j2ee/blueprints/ ) and a book dedicated to this subject, such as Richard Monson-

Haefel's Enterprise JavaBeans (O'Reilly).

© 2001, O'Reilly & Associates, Inc.

13.3 Scalability

For a large, complex application, there are many reasons to move to a model that includes Enterprise

JavaBeans components. But, contrary to popular belief, scalability and great performance should not be the number one deciding factor. There are many ways to develop scalable applications using just JSP or the servlet/JSP combination, often with better performance than an EJB-based application, since the communication overhead between the web tier and EJB tier is avoided.

Scalability means that an application can deal with more and more users by changing the hardware configuration rather than the application itself. Typically this means, among other things, that it's partitioned into pieces that can run on separate servers. Most servlet- and JSP-based applications use a database to handle persistent data, so the database is one independent piece. They also use a mixture of static and dynamically generated content. Static content, such as images and regular HTML pages, is handled by a web server, while dynamic content is generated by the servlets and JSP pages running within a web container. So without even trying, we have three different pieces that can be deployed separately.

Initially, you can run all three pieces on the same server. However, both the web container and the database use a lot of memory. The web container needs memory to load all servlet and JSP classes, session data, and shared application information. The database server needs memory to work efficiently with prepared statements, cached indexes, statistics used for query optimization, etc. The server requirements for these two types of servers are also different; for instance, the web server must be able to cope with a large number of network connections, and the database server needs fast disk access.

Therefore, the first step in scaling a web application is typically to use one server for the web server and servlet container, and another for the database.

If this is not enough, you can distribute the client requests over a set of servers processing HTTP requests. There are two common models: distributing requests only for dynamic content (servlet and

JSP requests), or distributing requests for all kinds of content.

If the web server can keep up with the requests for static content, such as images, regular HTML, and audio and video files, but not with the servlet and JSP requests, you can spread the dynamic content processing over multiple web containers on separate servers, as shown in

Figure 13.5

. A number of

load balancing web container modules are available for the major web servers (Apache, iPlanet Web

Server, and Microsoft's Internet Information Server): for instance, Apache's Tomcat™

( http://jakarta.apache.org/ ), BEA's WebLogic™ ( http://www.bea.com/ ), Caucho Technology's Resin™

( http://www.caucho.com/ ), and Unify's ServletExec™ ( http://www.unify.com/ ).

The tricky part when distributing servlet loads over multiple servers is ensuring that session data is

handled appropriately. Most containers keep session data in memory. The load balance module therefore picks the server with the lowest load to serve the first request from a client. If a session is created by this request, all subsequent requests within the same session are sent to the same server. But a container can also save session data on disk or in a database instead of in memory. It can then freely distribute each request over all servers in the cluster, and can also offer failure recovery in case a server crashes. A container is allowed to move a session from one server to another only for applications marked as distributable, as described in the next section.

Figure 13.5. Web server distributing load over multiple web containers

For a high-traffic site, you may need to distribute requests for both static and dynamic content over

multiple servers, as illustrated in Figure 13.6

. You can then place a load balancing server in front of a set of servers, each running a web server and a servlet container. As with the previous configuration, session data must be considered when selecting a server for the request. The easiest way to deal with this is to use a load balancing product that sends all requests from the same client to the same server.

This is not ideal, though, since all clients behind the same proxy or firewall appear as the same host.

Some load balancing products try to solve this problem by using cookies or SSL sessions to identify individual clients behind proxies and firewalls. In this configuration, you get the best performance from a web server than runs a servlet container in the same process, eliminating the process-to-process communication between the web server and the servlet container. Most of the servlet containers mentioned above can also be used in-process with all the major web servers. Another alternate configuration is a pure Java server that acts both as a web server and a servlet container. Examples are

Apache's Tomcat, Sun's Java Web Server ( http://www.sun.com/software/jwebserver / ), and Gefion software's LiteWebServer ( http://www.gefionsoftware.com/LiteWebServer/ ). Compared to adding a servlet container to a standard web server, this all-in-one alternative is easier to configure and maintain.

The traditional servers written in C or C++ may still be faster for serving static content, but with faster and faster Java runtimes, pure Java servers will soon be just as fast.

Figure 13.6. Load balancing server distributing requests over multiple servers

You should not rely on configuration strategies alone to handle the scalability needs of your application. The application must also be designed for scalability, using all the traditional tricks of the trade. Finally, you must load-test your application with the configuration you will deploy it on to make sure it can handle the expected load. There are many pure Java performance testing tools to choose from, from the simple but powerful Apache JMeter ( http://java.apache.org/jmeter/index.html

) to sophisticated tools like Innovative IT Development's PureLoad

( http://www.ideit.com/products/pureload/ ), that support data-driven, session-aware tests to be executed on a cluster of test machines.

13.3.1 Preparing for Distributed Deployment

As described in the previous section, some web containers can distribute requests for a web application's resources over multiple servers, each server running its own Java Virtual Machines (

JVM). Of course, this has implications for how you develop your application. So, by default, a web container must use only one JVM for an application.

If you want to take advantage of web container-controlled load balancing, you must do two things: mark the application as distributable, and follow the rules for a distributed application defined by the

Servlet 2.2 API.

To mark an application as distributable means adding the following element to the WEB-INF/web.xml file for the application:

<web-app>

...

<distributable/>

...

</web-app>

By doing so, you're telling the web container that your application adheres to the rules for distributed applications. According to the Servlet 2.2 API, a distributed application must be able to work within the following constraints:

Each JVM has a unique servlet instance for each servlet definition. In case the servlet implements the javax.servlet.SingleThreadModel interface, each JVM may contain multiple instances of the same servlet.

Each JVM has a unique instance of the javax.servlet.ServletContext class.

Each object stored in the session must be serializable (must implement the java.io.Serializable

interface, and all variables must be serializable).

This means that your servlets cannot rely on instance variables to keep data shared by all requests for a certain servlet. It also means that application scope objects (ServletContext attributes) are not shared between JVMs. In most cases, this is not a problem. For instance, if you use the application scope to provide shared access to cached read-only data, it just means that each JVM has its own copy.

If you really need to share some resource between JVMs, you must share it through an external mechanism, such as a directory server accessed through JNDI, a database, or a file in a filesystem that is available to all servers.

The most interesting part about distributed applications is how sessions are handled. The web container allows only one server at a time to handle a request that's part of a session. But since all objects put into the session must be serializable, the container can save them on disk or in a database as well as in memory. If the server that handles a session gets overloaded or crashes, the container can therefore move the responsibility for the session to another server. The new server simply loads all serialized session data and picks up where the previous server left off.

© 2001, O'Reilly & Associates, Inc.

14.1 Using a Servlet as the Controller

In an application in which all requests must be preprocessed or postprocessed in some way, using a servlet as the common entry point—the Controller—for all requests makes a lot of sense. Examples of this type of processing are application-controlled authentication and access control, application-specific logging, accessing a database, and so forth. Another common reason for using servlets instead of JSP pages for parts of an application is that the request processing requires so much code that putting it in a

JSP page makes the resulting application hard to develop, debug, and maintain. If you find yourself spending too much time trying to locate the source of code-based syntax errors or trying to figure out why the code in a page doesn't behave as it should, you may want to move the code to a servlet instead.

Since a servlet is just a regular Java class, you can make use of a Java compiler and debugger to fix the problems.

The Project Billboard application introduced in

Chapter 10 is a good candidate for using a servlet as the

Controller; it uses application-controlled authentication and contains code for accessing a database. In this chapter, it's used as a concrete example of the servlet-based approach. The components of the new

design are shown in Figure 14.1

.

Figure 14.1. Project Billboard application combining servlet and JSP components

Compare this design with the one described in

Chapter 10 . Note that the pure presentation JSP pages

remain the same, but the request processing pages are replaced by the servlet. So, as before, the user first requests the login.jsp page. This page still contains a form with fields for username and password, but with the new design it invokes the controller servlet instead of a request processing JSP page. The servlet is invoked with a request parameter named action with the value authenticate. The servlet performs the authentication, and if successful, it creates an EmployeeBean object and saves it in the session scope as proof of authentication. It then redirects the browser back to itself, this time with action

set to showPage and a page parameter set to the URL of an application JSP page. When the

servlet receives a request with a showPage action, it forwards the request to the specified page. (I will explain why it's done this way instead of just redirecting straight to the JSP page shortly.) As before, the selected page depends on whether the user loaded the login.jsp page or tried to access an application page directly, without first logging in.

The main.jsp page contains a form for updating the project subscription list, and links for posting a new

message or logging out. The difference from the page in Chapter 10

is that the form and the links now invoke the servlet instead of request processing JSP pages. Different action parameter values are used to distinguish each type of request. For all types of requests, the servlet first verifies that the user is authenticated. If not, it forwards the request to the login.jsp page, where the URL for the requested page is saved as a hidden field in the same way as in

Chapter 10

. If the user is authenticated, the servlet performs the requested action and redirects to itself with the showPage action and the appropriate page

parameter value, just as after a successful authentication. Redirecting to the servlet with a showPage

action instead of directly invoking the JSP page lets the servlet perform access control for the JSP pages as well.

This example illustrates a number of interesting things: servlet-initialized application scope objects, centralized request processing, how to map a URL pattern to a servlet, and additional access-control requirements for protected JSP pages when using a servlet as the Controller.

14.1.1 Initializing Application Scope Objects

The Project Billboard application uses two business logic beans that must be available to all users; in other words, they must be available as application scope objects. You may remember the NewsBean from

Chapter 10 . This bean is the repository for all news items relating to projects, and it's used as the

source for the personalized message list. The other business logic bean is called

EmployeeRegistryBean

. It acts as an abstraction of the database with employee information, and contains methods for authenticating a user and retrieving and saving employee information. The

EmployeeRegistryBean

class is described in more detail in

Chapter 15 .

The controller servlet, named PBControllerServlet, needs access to these beans to do its work, so it's a good idea to let it create and initialize the beans. By placing references to the beans in

ServletContext

attributes, they are made available to all JSP pages in the application as application scope beans. That's because the application scope is just an abstraction of the ServletContext attributes. If a servlet places a reference to a bean in a context attribute, the bean can be accessed by any

JSP page using the <jsp:useBean> action with the scope attribute set to application.

Example

14.1

shows the PBControllerServlet's init( ) method, where the two beans are initialized and

saved as context attributes.

Example 14.1. Initialization of Application Scope Beans

package com.ora.jsp.servlets;

...

import javax.servlet.*; import javax.sql.*; import com.ora.jsp.beans.emp.*; import com.ora.jsp.beans.news.*; import com.ora.jsp.sql.*; public class PBControllerServlet extends HttpServlet {

public void init( ) throws ServletException {

DataSource ds = null;

try {

ds = new DataSourceWrapper("sun.jdbc.odbc.JdbcOdbcDriver",

"jdbc:odbc:example", null, null);

}

catch (Exception e) {} // Ignore all in this example

EmployeeRegistryBean empReg = new EmployeeRegistryBean( );

empReg.setDataSource(ds);

getServletContext( ).setAttribute("empReg", empReg);

NewsBean news = new NewsBean( );

getServletContext( ).setAttribute("news", news);

}

...

First, a javax.sql.DataSource instance is created. The DataSource, an interface that's part of the JDBC 2.0 Standard Extension (SE) package, allows you to open JDBC database connections to retrieve and modify database data. It can represent a connection pool, letting connections be reused instead of opened and closed over and over again. Many JDBC driver vendors offer connection pool

DataSource

implementations, but here we use a simple wrapper class that implements its own connection pool based on standard JDBC 1.0 classes. The wrapper class is discussed in more detail in

Chapter 17

, where I also describe how to use a vendor-provided DataSource implementation. Also

note that in Example 14.1

, the DataSource instance is created with hardcoded values for the JDBC

driver class and URL. Of course, that is not the best way. A better approach is to specify this information externally, for instance as servlet initialization parameters. Again, more about this in

Chapter 17

.

The DataSource is used by the EmployeeRegistryBean, which is the bean the Project Billboard application uses instead of accessing the database directly. An instance of the bean is created, initialized with the DataSource, and saved as a context attribute named empReg. Next, the NewsBean instance is created and saved as a context attribute named news. The implementation used in this book keeps all messages in memory as opposed to in a database. If a database were used, the NewsBean

would also need to be initialized with the DataSource.

Servlets as well as JSP pages can now access these two beans by asking the context for the corresponding attribute. A servlet uses the getAttribute( ) method:

NewsBean newsBean =

(NewsBean) getServletContext( ).getAttribute("news");

Note that the getAttribute( ) method has the return type Object, so you have to cast the attribute value to the correct type.

A JSP page uses the <jsp:useBean> action to make the bean available to other action elements and scriptlets in the page:

<jsp:useBean id="news" scope="application"

class="com.ora.jsp.beans.news.NewsBean" />

Since the id attribute value matches the context attribute value, the <jsp:useBean> action retrieves the bean created by the servlet instead of creating a new one.

In the application described in this chapter, only one servlet is used. But that's not always the case. In a large application, it's sometimes better to use different controller servlets for different parts of the application, for instance, one controller for all functions used by an administrator, and another for end users. If that's the case, you can use a separate servlet with the sole purpose of taking care of all initialization of beans for the application. Whether you use the one-servlet approach described here or a separate application initialization servlet, the servlet that creates the beans used by the rest of the application must be loaded by the servlet container before any other parts of the application. You can tell a container to load a servlet when the web application is initialized by using elements in the application deployment descriptor, the WEB-INF/web.xml file:

<web-app>

...

<servlet>

<servlet-name>pbController</servlet-name>

<servlet-class>

com.ora.jsp.servlets.PBControllerServlet

</servlet-class>

<load-on-startup>1</load-on-startup>

</servlet>

...

Within the <servlet> element, which is described in further detail later, the <load-on-startup> element specifies that this servlet must be loaded when the web application is started by the container.

The value is a positive integer, indicating when this servlet should be loaded relative to other servlets.

Servlets with low values are always loaded before servlets with higher values.

A servlet that creates and initializes shared beans should also make sure that the beans are being removed and shut down gracefully, if needed. This is done in the servlet's destroy( ) method, as

shown in Example 14.2

.

Example 14.2. Removing Application Scope Beans

public void destroy( ) {

getServletContext( ).removeAttribute("empReg");

getServletContext( ).removeAttribute("news");

}

14.1.2 Centralized Request Processing

When all requests for an application are first processed by one servlet, it's easy to handle common functions in one place. In this example, the controller servlet is used for authentication and access control. But there are other possibilities. Say you have a site where you charge the users based on what they do, or need to keep an audit log over all database modifications. A servlet serving as a common entry point for all application requests is a good place to put this logic.

The controller servlet needs to handle both GET and POST requests, so its doGet( ) method simply calls the doPost( ) method: public void doGet(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

doPost(request, response);

}

It's in the doPost( ) method, shown in Example 14.3

, that all processing takes place.

Example 14.3. Centralized Request Processing

public void doPost(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

String action = request.getParameter("action");

// Check if the user is authenticated

if (!isAuthenticated(request) &&

(!"authenticate".equals(action) ||

"logout".equals(action))) {

doForwardToLogin(request, response);

}

else {

if ("authenticate".equals(action)) {

doAuthenticate(request, response);

}

else if ("logout".equals(action)) {

doLogout(request, response);

}

else if ("storeMsg".equals(action)) {

doStoreMsg(request, response);

}

else if ("updateProfile".equals(action)) {

doUpdateProfile(request, response);

}

else if ("showPage".equals(action)) {

doShowPage(request, response);

}

else {

response.sendError(

HttpServletResponse.SC_NOT_IMPLEMENTED);

}

}

}

In this example, a request parameter named action is used to decide what to do. An alternative to this is using part of the URI itself as the indication about which action to perform, as you will see in the next section when we look at the URI used to invoke the servlet. If the user isn't already authenticated, and the request is not to authenticate or log off, the request is forwarded to the login page.

The isAuthenticated( ) method simply checks if an authentication token is available in the session: private boolean isAuthenticated(HttpServletRequest request) {

boolean isAuthenticated = false;

HttpSession session = request.getSession( );

if (session.getAttribute("validUser") != null) {

isAuthenticated = true;

}

return isAuthenticated;

}

In this example, a bean named validUser is used as an authentication token. This bean is placed in the session scope by the servlet's doAuthenticate( ) method if the user is successfully authenticated. The isAuthenticated( ) method looks for the bean. If it finds it, it knows that the user has already been authenticated.

If the user is not authenticated, the doForwardToLogin( ) method is called: private void doForwardToLogin(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

String origURL = HttpUtils.getRequestURL(request).toString( );

String queryString = request.getQueryString( );

if (queryString != null) {

origURL += "?" + queryString;

}

String loginURL = "login.jsp" + "?origURL=" +

URLEncoder.encode(origURL) +

"&errorMsg=" + URLEncoder.encode("Please log in first");

forward(loginURL, request, response);

}

In order for the login page to save the URL of the requested page so it can be shown automatically after successful authentication, the URL must be passed to the login.jsp page as a parameter. The URL is constructed from the getRequestURL( ) method, which returns the URL of the current request minus the query string; a call is made to getQueryString( ) to append the query string if it's present. The complete URL is then used as the value of the origURL parameter. Another parameter, errorMsg

, is added with a message to display on the login page. Both parameter values are encoded using the java.net.URLEncoder class. This is necessary to convert special characters, such as the space character, into encoded values that are accepted in a URL.

The request is finally forwarded to the URL for the login.jsp page, including both parameters, using the forward( )

method: private void forward(String url, HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

RequestDispatcher rd = request.getRequestDispatcher(url);

rd.forward(request, response);

}

The RequestDispatcher is a servlet API class used to programmatically invoke another servlet or a

JSP page to continue the processing of a request. The servlet's forward( ) method gets an instance of the RequestDispatcher class that represents the login.jsp page and calls its forward( ) method.

This is almost exactly what the <jsp:forward> action element in a JSP page does. The difference is that the action element also aborts processing of the rest of the page. When a request is forwarded, the originating servlet delegates all processing of the request to the target servlet (or JSP page). The originating servlet is not allowed to modify the response in any way, neither before calling forward(

)

nor when the method returns. In most cases, it should simply return after calling forward( ), possibly after doing some cleanup that does not involve modifying the response. As you can see in

Example 14.3

, the PBControllerServlet doesn't do anything after calling

doForwardToLogin( )

, which in turn calls the forward( ) method.

If the user is authenticated, the requested action is performed by calling the corresponding method.

Some of the methods are described in the next section. You can look at the source code for the

PBControllerServlet

class to see how the other actions are performed.

14.1.3 Mapping a URI Pattern to a Servlet

Using a URI starting with /servlet to invoke a servlet is a convention introduced by Sun's Java Web

Server (JWS), the first product to support servlets before the API was standardized. This convention is supported by most servlet containers today. But using this type of URI has a couple of problems. First, it makes it perfectly clear to a user (at least a user who knows about servlets) what technology is used to implement the application. Not that you shouldn't be proud of using servlets, but a hint like this can help a hacker explore possible security holes. Even though no servlet-related security issues are known at this time, it never hurts to be a bit paranoid when it comes to security. The other problem is of a more practical nature.

As I described in

Chapter 12 , using relative URIs to refer to resources within an application makes life a

lot easier. As you know, a relative URI is interpreted as relative to the URI of the current request. But if a servlet, like the controller servlet in this chapter, is invoked using a URI like

/ora/servlet/PBController, it can't easily use a relative URI to refer to a View JSP page, such as

/ora/ch14/main.jsp. This is because the two resources are identified by URIs at different levels in the application's URI path structure. If you invoke the servlet with this type of URI, it has to refer to the JSP page with a relative path like ../ch14/main.jsp, or a context-relative path like /ch14/main.jsp. This works, but if you later decide to change the path structure, for instance to use the path /billboard/main.jsp instead, you have to modify all path references in the servlet.

A simple solution to this problem is to define a URI pattern mapping for the servlet, so it can be

accessed using a URI with the same path structure as the JSP pages. This type of mapping has been supported by servlet containers in a proprietary way for a long time, but the Servlet 2.2 API standardizes the rules. The mapping is done with XML elements in the web application's deployment descriptor (the

WEB-INF/web.xml file) like this:

<web-app>

...

<servlet>

<servlet-name>pbController</servlet-name>

<servlet-class>

com.ora.jsp.servlets.PBControllerServlet

</servlet-class>

</servlet>

...

<servlet-mapping>

<servlet-name>pbController</servlet-name>

<url-pattern>/ch14/process</url-pattern>

</servlet-mapping>

...

Two main elements are used here. The <servlet> element, with subelements <servlet-name> and <servlet-class>, are used to associate the servlet class with a symbolic name. The name must be unique within the application. The <servlet-mapping> element is then used to define the path pattern used to invoke the servlet. The <servlet-name> subelement value must match the symbolic name of the servlet, and the <url-pattern> subelement defines the pattern. The pattern can be a context-relative path, as in this example. This type of pattern tells the container to invoke the named servlet only if the request URI path matches the pattern exactly (a query string is of course allowed, but no other path parts).

This is the pattern used for the servlet in the Project Billboard application described in this chapter. To see how it works, let's first look at the action attribute in the login.jsp page:

<form action="process?action=authenticate" method="post">

The action attribute contains a relative path (i.e., it doesn't start with a slash). Since the login.jsp page is requested with the URI /ora/ch14/login.jsp, the browser converts the relative path into the absolute path

/ora/ch14/process?action=authenticate. The /ora part is the context path, and the rest of the path,

/ch14/process, matches the context-relative path pattern for the servlet in the web.xml file. If you decide to install this web application with a different context path, the action URI still invokes the correct servlet. If you then decide to change the internal application path structure, you only have to change the pattern in the web.xml file; no changes to paths in the JSP pages or the servlet are needed.

Other types of patterns let you use a part of the URI path to indicate what the servlet is asked to do instead of using an action parameter. A path pattern matches a URI starting with the specified path

(as opposed to the exact path match required in the previous case):

...

<servlet-mapping>

<servlet-name>pbController</servlet-name>

<url-pattern>/ch14/process/*</url-pattern>

</servlet-mapping>

...

Note that the pattern ends with /*, which is a wildcard sequence matching any subpath. With a path pattern, the servlet can be invoked with a URI such as /ch14/process/autheticate, and it can get the action name specified by the last part of the URI with this code: public void doPost(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

String action = request.getPathInfo( );

...

Another popular way to invoke the controller server is with a URI that ends with a special extension, like /ch14/authenticate.do, instead of a special URI prefix. To have all requests with the selected extension processed by the controller servlet, use an extension pattern:

...

<servlet-mapping>

<servlet-name>pbController</servlet-name>

<url-pattern>*.do</url-pattern>

</servlet-mapping>

...

The servlet then gets the action name by stripping off the extension: public void doPost(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

String action = request.getServletPath( );

int extension = action.lastIndexOf(".");

if (extension != -1) {

action = action.substring(0, extension);

}

...

All pattern-mapping types described here allow you to use a relative path to invoke the controller servlet, making it easier to refer to all the pieces of the application.

Relative paths are also supported by the servlet API methods used to forward and redirect to another resource within the same application. To forward a request, you first need to get a

RequestDispatcher

. There are two ways to obtain one. To use a relative path URI, call the

ServletRequest

's getRequestDispatcher( ) method, like in the forward( ) method in the servlet described earlier: private void forward(String url, HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

RequestDispatcher rd = request.getRequestDispatcher(url);

rd.forward(request, response);

}

If you look at how this method is called in the previous section, you see that a relative path (login.jsp) is passed as the value of the url parameter. The request object knows the absolute path that was used to make the current request, so it can use a relative path to locate the target resource. The javax.servlet.ServletContext

class also has a getRequestDispatcher( ) method, but it works only with context-relative paths, since the context has no knowledge about the absolute path for the current request.

In versions of the servlet API prior to 2.2, the sendRedirect( ) method, used to send a redirect response to the browser, required that an absolute URL was used to specify the target resource.

However, in the 2.2 version, this requirement was relaxed, so both absolute and relative paths are now also supported (note that context-relative paths are not supported, for backward-compatibility reasons: there's no way to distinguish an absolute path from a context-relative path, since both start with a slash).

The method that handles the logout action in the controller servlet uses a relative path to redirect to the login page: private void doLogout(HttpServletRequest request,

HttpServletResponse response) throws IOException {

HttpSession session = request.getSession( );

session.invalidate( );

response.sendRedirect("login.jsp");

}

14.1.4 Access Control for JSP Pages

You may have noticed that the access control performed by the controller servlet in Example 14.3

doesn't prevent a user from invoking one of the protected JSP pages directly, for instance with a URL like /ch14/main.jsp instead of /ch14/process?action=showPage&page=main.jsp. In most applications, this is not as bad as it looks. When you use a servlet as the controller, it is responsible for retrieving all protected data, and the JSP page displays only the data it receives from the servlet. Invoking the JSP page directly, without going through the controller servlet, therefore yields a page with only static template text.

However, if some of your application contains JSP pages or HTML pages with static data that needs to be protected, you can configure the application with a security constraint in the WEB-INF/web.xml file that makes it impossible for anyone to load the pages directly:

<security-constraint>

<web-resource-collection>

<web-resource-name>no-access</web-resource-name>

<url-pattern>/ch14/main.jsp</url-pattern>

<url-pattern>/ch14/entermsg.jsp</url-pattern>

</web-resource-collection>

<auth-constraint>

<role-name>nobody</role-name>

</auth-constraint>

</security-constraint>

This protects the two JSP pages in the example application, identified by the <url-pattern> elements. The <auth-constraint> element defines a single role named nobody. If no user is assigned this role, nobody can access these pages directly. When the controller servlet is invoked with the showPage action, however, it uses a RequestDispatcher to forward control to the pages.

Since forwarding is an internal affair, the security constraint doesn't have any effect.

© 2001, O'Reilly & Associates, Inc.

14.2 A More Modular Design Using Action Objects

The approach in Example 14.3

is to implement each action as a separate method in the servlet, so every

time you add a new action, you also have to update the servlet. In a large application, you may find yourself adding more and more code to the controller servlet, eventually ending up with a class that's hard to maintain.

A more modular approach is to treat each action as a separate class that implements a common interface, for instance called Action.

[1]

The Action interface may look like this:

[1]

There is an Action interface already defined in the Java Swing ( JFC) classes. However, since you will probably not be using the Swing GUI classes in this context, the two should not conflict. import java.io.*; import javax.servlet.*; import javax.servlet.http.*; public interface Action {

public void perform(HttpServlet servlet,

HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException;

}

The single method, perform( ) , has arguments that give the action class access to all the same objects as a regular servlet: the request and response objects. In addition, the servlet argument lets the action class access the servlet context and servlet configuration objects if needed.

With this approach, each action ends up as a simple class. For instance, the logout action is handled by a nice little class like this: import java.io.*; import javax.servlet.*; import javax.servlet.http.*; public class LogoutAction implements Action {

public void perform(HttpServlet servlet,

HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

HttpSession session = request.getSession( );

session.invalidate( );

response.sendRedirect("login.jsp");

}

}

In most cases, you can develop the action classes so that they do not keep any state information in instance variables. The controller servlet therefore only needs one instance of each class. They can be created when the first request is received for a certain action ( "lazy instantiation") or in the controller servlet's init( ) method: public void init( ) throws ServletException {

...

initActions( );

} private void initActions( ) {

actions = new Hashtable( );

actions.put("authenticate", new AuthenticateAction( ));

actions.put("logout", new LogoutAction( ));

actions.put("storeMsg", new StoreMsgAction( ));

actions.put("updateProfile", new UpdateProfileAction( ));

actions.put("showPage", new ShowPageAction( ));

actions.put("login", new LoginAction( ));

}

In this example, all action class names are hardcoded. An alternative is to use servlet initialization parameters or a separate configuration file to specify the mapping between action names and action class names. If you do that, you can add new actions to the application without touching the controller servlet code at all.

With the action class approach, the controller servlet's doPost( ) method gets the action name from an action parameter (or a part of the URL, as discussed earlier), finds the corresponding action class, and calls its perform( ) method: public void doPost(HttpServletRequest request,

HttpServletResponse response)

throws IOException, ServletException {

String actionName = request.getParameter("action");

if (actionName == null) {

response.sendError(HttpServletResponse.SC_NOT_ACCEPTABLE);

return;

}

Action action = (Action) actions.get(actionName);

if (action == null) {

response.sendError(HttpServletResponse.SC_NOT_IMPLEMENTED);

return;

}

// Use the login action if the user is not authenticated

if (!isAuthenticated(request) &&

(!"authenticate".equals(actionName) ||

"logout".equals(actionName))) {

action = (Action) actions.get("login");

}

action.perform(this, request, response);

}

The source code for a controller servlet and action classes using this approach is available in the code package for this book, so you can look at the details of all classes yourself. A framework for applications using a servlet with action classes and JSP pages is also available as part of the Apache

Jakarta project, at http://jakarta.apache.org/struts/index.html

.

© 2001, O'Reilly & Associates, Inc.

14.3 Sharing Data Between Servlets and JSP Pages

When you use servlets for request processing and JSP pages to render the user interface, you often need a way to let the different components access the same data. The model I recommend is having the servlet create beans and pass them to a JSP page for display.

As I described earlier, the application scope is just a JSP abstraction of javax.servlet.ServletContext

attributes. Similarly, the request and session scopes are JSP abstractions for attributes associated with javax.servlet.ServletRequest and javax.servlet.http.HttpSession

, respectively. All three classes provide setAttribute( )

, getAttribute( ), and removeAttribute( ) methods. A servlet uses the setAttribute( ) method to make a bean available to a JSP page. For instance, a servlet can create a bean, save it as a request attribute, and then forward control to a JSP page like this: public void doGet(HttpServletRequest request,

HttpServletResponse response) throws ServletException,

IOException {

String userName = request.getParameter("userName");

UserInfoBean userInfo = userReg.getUserInfo(userName);

request.setAttribute("userInfo", userInfo);

RequestDispatcher rd =

request.getRequestDispatcher("welcome.jsp");

rd.forward(request, response);

}

To the JSP page, the bean appears as a request scope variable. It can therefore obtain the bean using the

<jsp:useBean>

action and then access the properties of the bean as usual, in this case using

<jsp:getProperty>

:

<h1>Welcome

<jsp:useBean id="userInfo" scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" />

<jsp:getProperty name="userInfo" property="userName" />

</h1>

The <jsp:useBean> action, with an id attribute value matching the request attribute name set by the servlet, is needed to make the bean known to the JSP container before other actions or scripting

code can access it.

If the bean needs to be available throughout the session, the servlet uses an HttpSession attribute instead:

HttpSession session = request.getSession( ); session.setAttribute("userInfo", userInfo);

Using a ServletContext attribute, the bean becomes available in the application scope:

ServletContext context = getServletContext( ); context.setAttribute("userInfo", userInfo);

The only difference in the JSP page is that the scope attribute for the <jsp:useBean> action must match the scope used by the servlet.

Passing beans in the other direction, from a JSP page to a servlet, is not so common, but it can be done.

Here's how. The JSP page creates the bean using <jsp:useBean> and sets the properties using

<jsp:setProperty>

:

<jsp:useBean id="userInfo" scope="request"

class="com.ora.jsp.beans.userinfo.UserInfoBean" >

<jsp:setProperty name="userInfo" property="*" />

</jsp:useBean>

<jsp:forward page="/myServlet" />

It then forwards the request to the servlet (mapped to the URI /myServlet) using

<jsp:forward>

, and the servlet retrieves the bean using getAttribute( ):

UserInfoBean userInfo =

(UserInfo) request.getAttribute("userInfo");

© 2001, O'Reilly & Associates, Inc.

14.4 Using a JSP Error Page for All Runtime Errors

Even if an application is developed with different types of components, it should deal with runtime errors in a consistent way. Recall from

Chapter 7 , how the page directive can specify a JSP page to be

used when code in the JSP page throws an exception. The error page gets access to the exception through an implicit variable named exception, and can display a user-friendly message as well as log details about the problem. The way the error page is invoked and how the implicit variable gets its value can easily be mimicked by a servlet.

Example 14.4

shows how to do this in the controller servlet used in previous examples in this chapter.

Example 14.4. Invoking a JSP Error Page from a Servlet

public void doPost(HttpServletRequest request,

HttpServletResponse response) {

...

try {

action.perform(this, request, response);

}

catch (Throwable t) {

request.setAttribute("javax.servlet.jsp.jspException", t);

RequestDispatcher rd =

getServletContext( ).getRequestDispatcher("/error.jsp");

rd.forward(request, response);

}

}

The servlet calls the action's perform( ) method within a try block. If any type of exception occurs while executing an action, the servlet catches it, sets the javax.servlet.jsp.jspException

request attribute to the exception object, and forwards the request to the error JSP page. The javax.servlet.jsp.jspException attribute name is reserved for the exception object in the JSP specification, so the JSP container picks up the exception object from the request attribute and makes it available to the JSP page through the exception variable.

© 2001, O'Reilly & Associates, Inc.

15.1 JavaBeans as JSP Components

JavaBeans are simply regular Java classes designed according to a set of guidelines. By following these guidelines, development tools can figure out how the bean is intended to be used and how it can be linked to other beans. The JavaBeans specification characterizes beans as classes that:

Support introspection so that a builder tool can analyze how a bean works

Support customization, so that when using an application builder, a user can customize the appearance and behavior of a bean

Support events as a simple communication metaphor that can be used to notify beans of interesting things

Support properties, both for customization through a tool and for programmatic use

Support persistence, so that a bean can be customized in an application builder and then have its state saved away and reloaded later

Introspection means that information about a class, such as details about its methods and their arguments and return type, can be discovered by another class. By following certain naming conventions for the methods, the external class can figure out how the bean class is intended to be used. Specifically, the bean's properties and the events it generates or observes can be found using introspection. For GUI beans, introspection is typically used by a builder tool to allow properties to be set by the user in a property window. In a JSP scenario, the JSP standard actions and custom actions use introspection to find the methods used for reading or writing property values, and to declare variables of appropriate types.

A property is an attribute of a bean that can be read or written by the bean's client through regular methods named according to the JavaBeans guidelines. Typically, the property value is represented by an instance variable in the bean, but a read-only property can also represent a value that's calculated at runtime. The property methods are intended to be used to customize the bean, for instance, setting the label text for a bean used as a button in a GUI application, or setting the name of the data source to be used for a faceless serverside bean. Besides property access methods, a bean class can have regular methods that perform actions such as saving the bean's properties in a database or sending a mail composed from its properties.

A bean can generate or observe events. In a GUI bean, typical events are "button clicked" and "item selected." A server-side bean can deal with events indicating that the source of the data it represents has been updated so it can refresh its copy.

Support for persistence means that a bean should implement the java.io.Serializable interface.

This interface flags a class that can be saved in an external format, such as a file. When tools are used to customize a bean, it is possible to save the customized state during application development and then let the

customized bean be instantiated in runtime. The <jsp:useBean> action allows you to take advantage of this feature, but it's not commonly used today since no JSP authoring tools provide a customization interface yet. There's another reason for supporting persistence in JSP beans, however. A servlet container can support session persistence, saving all session data when a servlet context is shut down and reloading it again when the context is restarted. This works only if the beans you save in the session scope implement

Serializable

. In addition, beans (or any other object) placed in the session scope of an application marked as being distributable must be serializable, so that the container can migrate the session from one server to another.

15.1.1 JavaBeans Naming Conventions

As we mentioned early in the book, a Java bean is a class that has a no-argument constructor and conforms to the JavaBeans naming conventions. The bean properties are accessed through getter and setter methods, collectively known as a bean's accessor methods. Getter and setter method names are composed of the word

get or set, respectively, plus the property name, with the first character of each word capitalized. A regular getter method has no arguments, but returns a value of the property's type, while a setter method takes a single argument of the property's type and has a void return type. Here's an example: public class CustomerBean implements java.io.Serializable {

private String firstName;

private String lastName;

private int accountNumber;

private int[] categories;

private boolean preferred;

public String getFirstName( ) {

return firstName;

}

public void setFirstName(String firstName) {

this.firstName = firstName;

}

A readable property has getter methods, a writable property has setter methods. Depending on the combination of getter and setter methods, a property is read-only, write-only, or read/write.

A read-only property doesn't necessarily have to match an instance variable exactly. Instead, it can combine instance variable values, or any values, and return a computed value: public String getFullName( ) {

return (new StringBuffer(firstName).append(" ")

.append(lastName).toString( ));

}

The type of a property can be a Java class, interface, or a primitive type such as int: public int getAccountNumber( ) {

return accountNumber;

}

Besides simple single-value properties, beans can also have multivalue properties represented by an array of any type. This is called an indexed property in the specification. Two types of access methods can be used for an indexed property: methods reading or writing the whole array, or methods working with just one element, specified by an index: public int[] getCategories( ) {

return categories;

}

public void setCategories(int[] categories) {

this.categories = categories;

}

public int getCategories(int i) {

return categories[i];

}

public void setCategories(int i, int category) {

this.categories[i] = category;

}

The naming convention for a Boolean property getter method is different than for all other types. You could use the regular getter name patterns, but the recommendation is to use the word is combined with the property name, to form a question: public boolean isPreferred( ) {

return preferred;

}

This helps to make the source code more readable. The setter method for a Boolean property follows the regular pattern: public void setPreferred(boolean preferred) {

this.preferred = preferred;

}

}

Event handling is based on observers implementing a listener interface, and generators providing methods for observers to register their interest in the events. A listener interface defines the methods a listener needs

to implement to be notified when the corresponding event is triggered. A bean identifies itself as a listener by declaring that it's implementing a listener interface, and an event source is identified by its listener registration methods.

Let's look at an example. A listener interface for observing events related to the customer data handled by the example bean looks like this: import java.util.EventListener; public interface CustomerUpdatedListener extends EventListener {

void customerUpdated(CustomerUpdatedEvent e);

}

The interface shown here defines only one event, but an interface may also group a number of related events.

The CustomerBean identifies itself as an observer of the event by implementing the interface: public class CustomerBean implements CustomerUpdatedListener {

...

public void customerUpdated(CustomerUpdatedEvent e) {

if (e.getAccountNumber( ) == accountNumber) {

// Refresh local copy

}

}

Another bean, perhaps one acting as the gatekeeper to the customer database, identifies itself as a source for the event by defining methods for registration of listeners: import java.util.Vector; public class CustomerRegister {

private Vector listeners = new Vector( );

public

void addCustomerUpdatedListener(CustomerUpdatedListener cul) {

listeners.addElement(cul);

}

public

void removeCustomerUpdatedListener(CustomerUpdatedListener cul) {

listeners.removeElement(cul);

}

public void updateCustomer(CustomerBean customer) {

// Update persistent customer storage

notifyUpdated(customer);

}

It notifies all listeners when the customer data is modified, like this: protected void notifyUpdated(CustomerBean customer) {

Vector l;

CustomerUpdatedEvent e =

new CustomerUpdatedEvent(this, customer.getAccountNumber( ));

synchronized(listeners) {

l = (Vector)listeners.clone( );

}

for (int i = 0; i < l.size( ); i++) {

((CustomerUpdatedListener)l.elementAt(i)).customerUpdated(e);

}

}

}

By following these simple naming conventions, the JSP standard actions <jsp:getProperty> and

<jsp:setProperty>

, as well as custom actions like <ora:loop>, can discover how to use your beans correctly. At this time, no JSP features rely on the event naming conventions, but future development tools may do so. So if your beans need to handle events, it's a good idea to follow the conventions. Besides, it's a well-known design pattern, so using it makes your code more readable to other developers familiar with this design.

15.1.1.1 Handling session events

A bean used in a JSP application can actually register itself to receive session-related events. The servlet API includes an interface called javax.servlet.http.HttpSessionBindingListener . An object that implements this interface is notified when the object is placed in or removed from a session, through these two methods: public void valueBound(HttpSessionBindingEvent event); public void valueUnbound(HttpSessionBindingEvent event);

The valueBound( ) method is called when the object is added to a session, and the valueUnbound(

)

method when it's removed. The HttpSessionBindingEvent class contains these two methods: public String getName( ); public HttpSession getSession( );

The getName( ) method returns the name used for the object in the session, and the getSession( ) method returns a reference to the session object itself.

One way to use a bean that implements the HttpSessionBindingListener interface is to keep track of currently logged-in users. A bean representing a user is added to the session after a successful authentication, and is given access to the servlet context for the application through a write-access property.

In the valueBound( ) method, the bean adds a reference to itself to the list of users maintained by a context attribute object, such as a java.util.Vector. The valueUnbound( ) method removes the reference: import java.io.*; import java.util.*; import javax.servlet.*; import javax.servlet.http.*; public class UserBean implements HttpSessionBindingListener,

Serializable {

private ServletContext context;

private String name;

...

public void setContext(ServletContext context) {

this.context = context;

}

public void setName(String name) {

this.name = name;

}

...

public void valueBound(HttpSessionBindingEvent e) {

Vector currentUsers =

(Vector) context.getAttribute("currentUsers");

if (currentUsers == null) {

currentUsers = new Vector( );

}

currentUsers.addElement(this);

}

public void valueUnbound(HttpSessionBindingEvent e) {

Vector currentUsers =

(Vector) context.getAttribute("currentUsers");

currentUsers.removeElement(this);

}

...

}

A JSP page can then generate a list of all current users by looping through the Vector, available as an

application scope object.

© 2001, O'Reilly & Associates, Inc.

15.2 JSP Bean Examples

In a JSP-based application, two types of beans are primarily used: value beans and utility beans. A value bean encapsulates all information about some entity, such as a user or a product. A utility bean performs some action, such as saving information in a database or sending email. Utility beans can use value beans as input, or produce value beans as a result of an action.

If you develop JavaBeans for your application, you're also preparing for migration to a full-blown J2EE application. The utility beans can be changed into proxies for one or more EJB session beans, acting as controllers for the application.

Value beans may be used as-is, acting as what are called Value Objects in Sun's paper, "Developing

Enterprise Applications With the Java 2 Platform, Enterprise Edition," also known as the J2EE blueprint. In

EJB-based applications, the application's data is represented by EJB entity beans. Getting a property value from an EJB entity bean requires a remote call, consuming both system resources and bandwidth. Instead of making a remote call for each property value that is needed, the web component can make one remote call to an EJB session bean (possibly via a JSP utility bean) that returns all properties of interest packaged as a value bean. The web component can then get all the properties from the value bean with inexpensive local calls. The value bean can also act as cache in the web container to minimize remote calls even more, and can combine information from multiple EJB entity beans that is meaningful to the web interface. If you plan to eventually move to the EJB model, I recommend that you read the J2EE blueprint paper

( http://java.sun.com/j2ee/blueprints/ ) before you design your application, so that you can make the migration as smooth as possible.

15.2.1 Value Beans

Value beans are useful even without EJB. They are handy for capturing form input, since the

<jsp:setProperty>

JSP action automatically sets all properties with names corresponding to request

parameter names, as described in Chapter 5

. In addition, the <jsp:getProperty> action lets you include the property values in the response without using scriptlets.

Another benefit of value beans is that they can be used to minimize expensive database accesses for entities that rarely change their value. By placing a value bean in the application scope, all users of your application can use the cached value instead.

Example 15.1

shows the source code for the ProductBean used in

Chapter 8 , to represent products in an online shopping application. This is a pure value bean, with only

property accessor methods, that could represent data stored in a database.

Example 15.1. ProductBean

package com.ora.jsp.beans.shopping; import java.io.*;

public class ProductBean implements Serializable {

private int id;

private String name;

private String descr;

private float price;

public int getId( ) {

return id;

}

public String getName( ) {

return name;

}

public String getDescr( ) {

return descr;

}

public float getPrice( ) {

return price;

}

void setId(int id) {

this.id = id;

}

void setName(String name) {

this.name = name;

}

void setDescr(String descr) {

this.descr = descr;

}

void setPrice(float price) {

this.price = price;

}

}

This bean is created and initialized by the single instance of the CatalogBean. All setter methods have package accessibility, while the getter methods are public. Using package accessibility for the setter methods ensures that only the CatalogBean can set the property values. For instance, a JSP page can read the product information, but not change the price.

Another example of a value bean is the UserInfoBean introduced in

Chapter 5 . Part of this bean is shown

in

Example 15.2

. Besides encapsulating the property values of the entity it represents, it also provides

methods for validating the data.

Example 15.2. Part of the UserInfoBean

package com.ora.jsp.beans.userinfo; import java.io.*; import java.util.*; import com.ora.jsp.util.*; public class UserInfoBean implements Serializable {

// Validation constants

private static String DATE_FORMAT_PATTERN = "yyyy-MM-dd";

private static String[] SEX_LIST = {"male", "female"};

private static int MIN_LUCKY_NUMBER = 0;

private static int MAX_LUCKY_NUMBER = 100;

// Properties

private String birthDate;

private String birthDateInput;

private String emailAddr;

private String emailAddrInput;

private String[] interests;

private String luckyNumber;

private String luckyNumberInput;

private String sex;

private String sexInput;

private String userName;

private boolean isInitialized;

public String getBirthDate( ) {

return birthDate;

}

public void setBirthDate(String birthDate) {

isInitialized = true;

birthDateInput = birthDate;

if (StringFormat.isValidDate(birthDate, DATE_FORMAT_PATTERN)) {

this.birthDate = birthDate;

}

}

...

All setter methods save the value passed as the argument in an instance variable representing the input value.

They then validate the value and, if it's valid, save it in another instance variable representing the property.

The getter methods return the value of the property value variable. The effect is that only validated values can be accessed through the getter methods. The saved input values are, however, used for another purpose, as shown in

Example 15.3

.

Example 15.3. UserInfoBean Validation Methods

public String getPropertyStatusMsg( ) {

StringBuffer msg = new StringBuffer( );

if (!isInitialized( )) {

msg.append("Please enter values in all fields");

}

else if (!isValid( )) {

msg.append("The following values are missing or invalid: ");

msg.append("<ul>");

if (birthDate == null) {

if (birthDateInput == null) {

msg.append("<li>Birth date is missing");

}

else {

msg.append("<li>Birth date value is invalid: " +

birthDateInput);

}

}

if (emailAddr == null) {

if (emailAddrInput == null) {

msg.append("<li>Email address is missing");

}

else {

msg.append("<li>Email address value is invalid: " +

emailAddrInput);

}

}

...

}

else {

msg.append("Thanks for telling us about yourself!");

}

return msg.toString( );

}

public boolean isInitialized( ) {

return isInitialized;

}

public boolean isValid( ) {

return isInitialized( ) &&

getBirthDate( ) != null &&

getEmailAddr( ) != null &&

getLuckyNumber( ) != null &&

getSex( ) != null &&

getUserName( ) != null;

}

The getPropertyStatusMsg( ) method returns a string with a message about the validation status of the properties. If one or more properties have invalid values, the message contains an HTML list with a list item element for each invalid value, showing the input value plus an explanation about why it was not accepted.

This behavior serves a purpose in the examples where the bean is used, as it allows me to introduce one JSP concept at a time. But it may not be ideal for a real application. First, it may be better to always save the input value as the property value, and let the isValid( ) method, instead of the setter method, handle all validation. This way, the getter methods return the input value whether it's valid or not. When the bean is used to fill out an input form, like in the last two examples in

Chapter 5

, the user will then be able to correct

the invalid values instead of typing new values from scratch. Example 15.4

shows an outline of these changes.

Example 15.4. Centralized UserInfoBean Validation

package com.ora.jsp.beans.userinfo; import java.io.*; import java.util.*; import com.ora.jsp.util.*; public class UserInfoBean implements Serializable {

// Validation constants

private static String DATE_FORMAT_PATTERN = "yyyy-MM-dd";

private static String[] SEX_LIST = {"male", "female"};

private static int MIN_LUCKY_NUMBER = 0;

private static int MAX_LUCKY_NUMBER = 100;

// Properties

private String birthDate;

private String emailAddr;

private String[] interests;

private String luckyNumber;

private String sex;

private String userName;

private boolean isInitialized;

public String getBirthDate( ) {

return birthDate;

}

public void setBirthDate(String birthDate) {

isInitialized = true;

this.birthDate = birthDate;

}

...

public boolean isValid( ) {

return

StringFormat.isValidDate(birthDate, DATE_FORMAT_PATTERN) &&

StringFormat.isValidEmailAddr(emailAddr) &&

StringFormat.isValidInteger(luckyNumber, MIN_LUCKY_NUMBER,

MAX_LUCKY_NUMBER) &&

StringFormat.isValidString(sex, SEX_LIST, true) &&

userName != null;

}

Secondly, a bean is easier to reuse if it's not tied to one form of presentation. It's therefore a good idea to remove the HTML in the getPropertyStatusMsg( ) method. One way of doing this is by splitting

the method into two methods, as outlined in Example 15.5

.

Example 15.5. HTML Free Status Messages

public String getPropertyStatusMsg( ) {

String msg = "Thanks for telling us about yourself!";

if (!isInitialized( )) {

msg ="Please enter values in all fields";

}

else if (!isValid( )) {

msg = "The following values are missing or invalid: ";

}

return msg;

}

public String[] getPropertyStatusDetails( ) {

Vector details = new Vector( );

if (isInitialized( ) && !isValid( )) {

if (birthDate == null) {

details.addElement("Birth date is missing");

}

else if (!StringFormat.isValidDate(birthDate,

DATE_FORMAT_PATTERN)) {

details.addElement("Birth date value is invalid: " +

birthDate);

}

...

}

String[] arr = new String[details.size( )];

details.copyInto(arr);

return arr;

}

With this approach, a JSP page can render the messages in any format that fits the overall layout of the page.

First it gets the main message, and then it loops through the indexed propertyStatusDetails values:

<font color="red">

<jsp:getProperty name="userInfo" property="propertyStatusMsg" />

<ul>

<ora:loop name=userInfo" property="propertyStatusDetails"

loopId="details" className="String" >

<li><%= details %>

</ora:loop>

</ul>

</font>

15.2.2 Utility Beans

A utility bean performs some action, such as processing information, as opposed to simply acting as a container for information.

The UserInfoBean contains processing code, namely code for encoding special characters such as HTML character entities and for validation of all property values. We can move this code to a separate bean so that the UserInfoBean becomes a pure property value container. Whether this change makes sense or not depends on the application that uses it.

One set of property getter methods used in

Chapter 5 returns the UserInfoBean property values with all

special characters that can cause problems in an HTML form converted to the corresponding HTML character entity codes. Here's one example: public String getUserNameFormatted( ) {

return StringFormat.toHTMLString(getUserName( ));

}

The actual conversion is performed by a method in the com.ora.jsp.util.StringFormat utility class: public static String toHTMLString(String in) {

StringBuffer out = new StringBuffer( );

for (int i = 0; in != null && i < in.length( ); i++) {

char c = in.charAt(i);

if (c == '\'') {

out.append("&#39;");

}

else if (c == '\"') {

out.append("&#34;");

}

else if (c == '<') {

out.append("&lt;");

}

else if (c == '>') {

out.append("&gt;");

}

else if (c == '&') {

out.append("&amp;");

}

else {

out.append(c);

}

}

return out.toString( );

}

Moving this HTML-specific code out of the bean and instead using the utility class directly in the JSP page makes the bean easier to reuse in parts of the application that do not render the values as HTML.

The way the UserInfoBean bean is used in this book, it's perfectly okay to keep the validation code in the bean itself. However, let's say you would like to add a property referencing another bean, a friends property for instance, that holds an array of other UserInfoBean objects. It may then be better to let a utility bean that knows about all users in the application perform the validation, including verifying that the friends exist.

A bean used for validation is one example of a utility bean that makes the application easy to maintain. The

CatalogBean

used in Chapter 8 is another example. The version developed for this book simply creates a

set of ProductBean objects with hardcoded values, and provides a method that returns all products in the catalog. In a real application, it would likely get the information from a database and have methods for updating catalog information, such as adding and removing products or changing the information about a product, as well as methods that return only products matching a search criterion. If all catalog update requests go through the CatalogBean, it can create, delete, and update the ProductBean objects so that they always match the information stored in the database. The number of database accesses can be greatly reduced this way.

Chapter 9 , and

Chapter 10 , offer another example of how you can use a utility bean. The purpose of these

chapters is to show you how to use the generic database custom actions to access a database, but a Java

programmer may want to encapsulate all database-access code in a bean instead. Example 15.6

shows part of a utility bean that handles all database interactions needed for these two chapters.

Example 15.6. Employee RegistryBean

package com.ora.jsp.beans.emp; import java.io.*; import java.sql.*; import java.text.*; import java.util.*; import javax.sql.*; import com.ora.jsp.sql.*; import com.ora.jsp.sql.value.*; import com.ora.jsp.util.*; public class EmployeeRegistryBean implements Serializable {

private DataSource dataSource;

/**

* Sets the dataSource property value.

*/

public void setDataSource(DataSource dataSource) {

this.dataSource = dataSource;

}

/**

* Returns true if the specified user name and password

* match an employee in the database.

*/

public boolean authenticate(String userName, String password)

throws SQLException {

EmployeeBean empInfo = getEmployee(userName);

if (empInfo != null &&

empInfo.getPassword( ).equals(password)) {

return true;

}

return false;

}

/**

* Returns an EmployeeBean initialized with the information

* found in the database for the specified employee, or null if

* not found.

*/

public EmployeeBean getEmployee(String userName)

throws SQLException {

// Get the user info from the database

Connection conn = dataSource.getConnection( );

Row empRow = null;

Vector projectRows = null;

try {

empRow = getSingleValueProps(userName, conn);

projectRows = getProjects(userName, conn);

}

finally {

try {

conn.close( );

}

catch (SQLException e) {} // Ignore

}

// Create a EmployeeBean if the user was found

if (empRow == null) {

// Not found

return null;

}

EmployeeBean empInfo = new EmployeeBean( );

try {

empInfo.setDept(empRow.getString("Dept"));

empInfo.setEmpDate(empRow.getString("EmpDate"));

empInfo.setEmailAddr(empRow.getString("EmailAddr"));

empInfo.setFirstName(empRow.getString("FirstName"));

empInfo.setLastName(empRow.getString("LastName"));

empInfo.setPassword(empRow.getString("Password"));

empInfo.setUserName(empRow.getString("UserName"));

empInfo.setProjects(toProjectsArray(projectRows));

}

catch (NoSuchColumnException nsce) {} // Cannot happen here

catch (UnsupportedConversionException nsce) {}

return empInfo;

}

/**

* Inserts the information about the specified employee, or

* updates the information if it's already defined.

*/

public void saveEmployee(EmployeeBean empInfo) throws SQLException {

// Save the user info from the database

Connection conn = dataSource.getConnection( );

conn.setAutoCommit(false);

Row userRow = null;

Vector interestRows = null;

try {

saveSingleValueProps(empInfo, conn);

saveProjects(empInfo, conn);

conn.commit( );

}

finally {

try {

conn.setAutoCommit(true);

conn.close( );

}

catch (SQLException e) {} // Ignore

}

}

/**

* Returns a Row with all information about the specified

* employee except the project list, or null if not found.

*/

private Row getSingleValueProps(String userName, Connection conn)

throws SQLException {

if (userName == null) {

return null;

}

SQLCommandBean sqlCommandBean = new SQLCommandBean( );

sqlCommandBean.setConnection(conn);

StringBuffer sql = new StringBuffer( );

sql.append("SELECT * FROM Employee ")

.append("WHERE UserName = ?");

sqlCommandBean.setSqlValue(sql.toString( ));

Vector values = new Vector( );

values.addElement(new StringValue(userName));

sqlCommandBean.setValues(values);

Vector rows = null;

try {

rows = sqlCommandBean.executeQuery( );

}

catch (UnsupportedTypeException e) {} // Cannot happen here

if (rows == null || rows.size( ) == 0) {

// User not found

return null;

}

return (Row) rows.firstElement( );

}

...

The EmployeeRegistryBean has one property, dataSource, that needs to be set when the bean is

created. Chapter 14

, describes how a servlet can create the bean and initialize it with a DataSource when the application starts, and then save it in the application scope where all JSP pages can reach it. The other

public methods in this bean perform the same functions as the generic database actions in Chapter 9

and

Chapter 10 . The getSingleValueProps( ) method, as well other private methods not shown in

Example 15.6

, use an SQLCommandBean to execute the SQL statement. This bean is included in the source code package for this book, so you can use it in your own beans as well. We will take a look at the

implementation in Chapter 17 .

Using the EmployeeRegistryBean instead of the generic database actions means you can greatly

simplify the JSP pages. For instance, the authentication code in Example 10.3

can be reduced to this:

<jsp:useBean id="empReg" scope="application"

class="com.ora.jsp.beans.emp.EmployeeRegistryBean" />

<%

String userName = request.getParameter("userName");

String password = request.getParameter("password");

if (!userReg.authenticate(userName, password)) {

%>

<ora:redirect page="index.jsp" >

<ora:param name="errorMsg"

value="The User Name and Password you entered are not valid." />

</ora:redirect>

<%

}

else {

com.ora.jsp.beans.emp.EmployeeBean validUser =

empReg.getEmployee(userName);

session.setAttribute("validUser", validUser);

%>

The <jsp:useBean> action finds the single instance in the application scope and associates it with a scripting variable named empReg, which is then used in the scriptlets to call the bean's methods.

15.2.3 Multithreading Considerations

As you have seen, putting business logic in beans leads to a more structured and maintainable application.

But there's one thing you need to be aware of: beans shared between multiple pages must be thread-safe.

Multithreading is an issue only for beans in the session and application scopes. Beans in the page and request scopes are executed by only one thread at a time. A bean in the session scope can be executed by more than one thread, initiated by requests from the same client. This may happen if the user brings up multiple browsers, repeatedly clicks a Submit button in a form, or if the application uses frames to request multiple

JSP pages at the same time. Application scope beans are shared by all application users, hence it's very likely that more than one thread is using an application scope bean.

Java provides mechanisms for dealing with concurrent access to resources, such as synchronized blocks and thread notification methods. But there are other ways to avoid multithreading issues in the type of beans used in JSP pages.

Value beans are typically placed in the request or session scope, as containers of related information used in multiple pages. In most cases, they are created and initialized in one place only, such as by a Controller servlet or by a <jsp:useBean> and <jsp:setProperty> combination in the request processing page invoked by a form, or by a custom action or utility bean. In all other places, the bean is used only with

<jsp:getProperty>

or scripting elements to read its property values. Since only one thread writes to the bean and all others just read it, you don't have to worry about different threads overwriting each other.

But if you have a value bean that can be updated, such as the NewsBean used in

Chapter 10 , you have to be

careful. The NewsBean contains an instance variable that holds a list of NewsItemBean objects and has methods for retrieving new items matching a search criterion, as well as for adding and removing new items.

If one thread calls removeNewsItem( ) while another is executing getNewsItems( ), a runtime exception may occur.

Example 15.7

shows how you can use synchronization to guard against this problem.

Example 15.7. Synchronized Access to Instance Variable

package com.ora.jsp.beans.news; import java.io.*; import java.util.*; import com.ora.jsp.util.*; public class NewsBean implements Serializable {

private Vector newsItems = new Vector( );

private int[] idSequence = new int[1];

...

public NewsItemBean[] getNewsItems(String[] categories) {

Vector matches = new Vector( );

synchronized (newsItems) {

for (int i = 0; i < newsItems.size( ); i++) {

NewsItemBean item =

(NewsItemBean) newsItems.elementAt(i);

if (ArraySupport.contains(categories,

item.getCategory( ))) {

matches.addElement(item);

}

}

}

NewsItemBean[] matchingItems =

new NewsItemBean[matches.size( )];

matches.copyInto(matchingItems);

return matchingItems;

}

public void setNewsItem(NewsItemBean newsItem) {

synchronized (idSequence) {

newsItem.setId(idSequence[0]++);

}

newsItems.addElement(newsItem);

}

public void removeNewsItem(int id) {

synchronized (newsItems) {

for (int i = 0; i < newsItems.size( ); i++) {

NewsItemBean item =

(NewsItemBean) newsItems.elementAt(i);

if (id == item.getId( )) {

newsItems.removeElementAt(i);

break;

}

}

}

}

...

Both the getNewsItems( ) and the removeNewsItem( ) methods synchronize on the newsItems object, and the addElement( ) method used in setNewsItem( ) is a synchronized method. The effect is that while one thread is manipulating the list of news items through one of these methods, all other threads wait until the current thread leaves the synchronized block.

The setNewsItem( ) method also synchronizes on idSequence, a variable used to generate a unique

ID for each item. idSequence is an int array with one component. It's a neat trick to be able to use synchronization for an integer value: Java doesn't allow synchronization on primitive types, only on objects,

but an array is an object. You could use an Integer object instead, but you can't change the value of an

Integer

. To increment the value, a new Integer must be created. Using an array avoids these repeated object creations (and creating an object is a fairly expensive operation in Java).

Another approach that avoids multithreading problems is used in the utility beans in this book, such as the

CounterBean

used in Chapter 8 and the EmployeeRegistryBean described in the previous section.

These beans only define setter methods for customization that takes place when the bean is created, and define all data needed to perform a function as method arguments instead of properties. Each thread has its own copies of argument values and local variables, so with this approach there's no risk of one thread stepping on another.

© 2001, O'Reilly & Associates, Inc.

15.3 Unexpected <jsp:setProperty> Behavior

The <jsp:setProperty> action can be used to automatically set all properties in a bean with names matching the names of the parameters received with the request. This is a great feature that's used in many of the examples in this book. But unless you know how this works behind the scenes, you could be in for a surprise.

When the <jsp:setProperty> code is invoked, it gets a list of all request parameter names and uses bean introspection to find the corresponding property setter methods. It then calls all setter methods to set the properties to the values of the parameters. This means that if you have a property in your bean that doesn't match a parameter, the setter method for this property is not called. In most cases, this is not surprising. However, if the parameter is present in some requests but not in others, things may get a bit confusing. This is the case with parameters corresponding to checkbox, radio button, and selection list elements in an HTML form. If this type of element is selected, the browser sends a parameter with the element's name and the value of the selected item. But if the element is not selected, it doesn't send a parameter at all.

For example, let's say you have a bean with an indexed property, such as the projects property in

the com.ora.jsp.beans.emp.EmployeeBean used in Chapter 10

. This bean is kept in the session scope. The user can change the value of the property through a group of checkboxes in a form.

To unregister all projects, a user deselects all checkboxes and submits the form. You may think the following code would then clear the property (setting it to null):

<jsp:setProperty name="validUser" property="projects"

param="projects" />

Yet it doesn't. Without any checkbox selections, the projects parameter is not sent and the

corresponding property setter method is not called. The workaround used in Chapter 10 is to use a

request-time attribute expression to explicitly set the property to either the array of selected checkboxes or null if none is selected:

<jsp:setProperty name="validUser" property="projects"

value='<%= request.getParameterValues("projects") %>' />

If you have been developing web applications for a while, you may not think this is so surprising. The

<jsp:setProperty>

action behaves the same way, however, even when a parameter matching a property is received but its value is an empty string. This happens for text fields that the user leaves empty.

If you have properties matching text fields, make sure the code that uses the values of the corresponding properties can deal with null values, or initialize them to empty strings. If you keep a bean like this in a scope other than the page and request scopes (where a new instance is created for each request), also be aware that the user cannot clear the property by erasing the field in a form. One possible workaround is to define a reset property, with a setter method that clears all properties.

Then call it explicitly in the JSP page before setting the other properties, like this:

<jsp:setProperty name="validUser" property="reset"

value="any value" />

<jsp:setProperty name="validUser" property="*" />

This way, all properties are first reset by the first <jsp:setProperty> action, and then all properties matching request parameters are set by the second action.

© 2001, O'Reilly & Associates, Inc.

16.1 Tag Extension Basics

A custom action—actually a tag handler class for a custom action—is basically a bean with property setter methods corresponding to the custom action element's attributes. In addition, the tag handler class must implement one of two Java interfaces defined by the JSP specification.

All the interfaces and classes you need to implement a tag handler are defined in the javax.servlet.jsp.tagext

package. The two primary interfaces are named Tag and

BodyTag

. The Tag interface defines the methods you need to implement for any action. The

BodyTag

interface extends the Tag interface and adds methods used to access the body of an action element. To make it easier to develop a tag handler, two support classes are defined by the API:

TagSupport

and BodyTagSupport, as shown in

Figure 16.1

. These classes provide default implementations for the methods in the corresponding interface.

Figure 16.1. The primary tag extension interfaces and support classes

The reason the specification defines both interfaces and the support classes that implement those interfaces is simply to cover all the bases. If you already have a class with functionality that you want to access as a custom action, you can specify that it implements the appropriate interface and add the few methods defined by that interface. In practice, though, I recommend that you implement your tag handlers as extensions to the support classes. This way, you get most of the methods implemented for free, and you can still reuse the existing classes by calling them from the tag handler.

A tag library is a collection of custom actions. For instance, all custom actions used in this book are packaged as one tag library. Besides the tag handler class files, a tag library must contain a Tag Library

Descriptor (TLD) file. This is an XML file that maps all custom action names to the corresponding tag

handler classes, and describes all attributes supported by each custom action. The class files and the

TLD can be packaged in a JAR file to make it easy to install. We look at the TLD syntax and packaging details at the end of this chapter.

Before we get into all the intricate details, let's take a brief look at what it takes to develop, deploy, and use a custom action. First, you implement a tag handler class, like the following: package com.mycompany; import java.io.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; public class HelloTag extends TagSupport {

private String name = "World";

public void setName(String name) {

this.name = name;

}

public int doEndTag( ) {

try {

pageContext.getOut( ).println("Hello " + name);

}

catch (IOException e) {} // Ignore it

return EVAL_PAGE;

}

}

The tag handler class contains a setter method for an attribute named name. The doEndTag( ) method (defined by the Tag interface) simply writes "Hello" plus the name attribute value to the response. You compile the class and place the resulting class file in the WEB-INF/classes directory for the application.

Next, you create the TLD file. The following is a minimal TLD file for a library with just one custom action element:

<?xml version="1.0" encoding="ISO-8859-1" ?>

<!DOCTYPE taglib

PUBLIC "-//Sun Microsystems, Inc.//DTD JSP Tag Library 1.1//EN"

"http://java.sun.com/j2ee/dtds/web-jsptaglibrary_1_1.dtd">

<taglib>

<tlibversion>1.0</tlibversion>

<jspversion>1.1</jspversion>

<shortname>test</shortname>

<tag>

<name>hello</name>

<tagclass>com.mycompany.HelloTag</tagclass>

<bodycontent>empty</bodycontent>

<attribute>

<name>name</name>

</attribute>

</tag>

</tablib>

The TLD maps the custom action name hello to the tag handler class com.mycompany.HelloTag

, and defines the name attribute. Place the TLD file in the application's WEB-INF/tlds directory, for instance with the filename mylib.tld.

Now you're ready to use the custom action in a JSP page, like this:

<%@ taglib uri="/WEB-INF/mylib.tld" prefix="test" %>

<html>

<body bgcolor="white">

<test:hello name="Hans" />

</body>

</html>

When the page is requested, the JSP container uses the TLD to figure out which class to execute for the custom action. It then calls all the appropriate methods, resulting in the text "Hello Hans" being added to the response. That's all there's to it for the most simple case. In the remainder of this chapter, we go through all of this in greater detail.

© 2001, O'Reilly & Associates, Inc.

16.2 Developing a Simple Action

As you have seen in the previous chapters, a custom action element in a JSP page consists of a start tag

(possibly with attributes), a body, and an end tag:

<prefix:actionName attr1="value1" attr2="value2">

The body

</prefix:actionName>

If the action element doesn't have a body, the following shorthand notation can be used instead of the start tag and the end tag:

<prefix:actionName attr1="value1" attr2="value2" />

A tag handler is the object invoked by the JSP container when a custom action is found in a JSP page.

In order for the tag handler to do anything interesting, it needs access to all information about the request and the page, as well as the action element's attribute values (if any). At a minimum, the tag handler must implement the Tag interface, which contains methods for giving it access to the request and page information, as well as methods called when the start tag and end tag are encountered. Note that an action element supported by a tag handler that implements the Tag interface may have a body, but the tag handler has more limited control over the body content than a tag handler that implements the BodyTag interface. For the attribute values, the JSP container treats the tag handler as a bean and calls a property setter method corresponding to each attribute, as shown in

Figure 16.2

.

Figure 16.2. Tag interface methods and property setter methods

Here are the most important methods of the Tag interface: public void setPageContext(PageContext pageContext); public int doStartTag( ) throws JspException;

public int doEndTag( ) throws JspException;

To be complete, let's first look at the implementation of these methods provided by the TagSupport class. This is the class that most simple tag handlers extend, so it's important to know how

TagSupport

implements the methods a tag handler inherits.

The first method of interest is the setPageContext( ) method: public class TagSupport implements Tag, Serializable {

...

protected PageContext pageContext;

...

public void setPageContext(PageContext pageContext) {

this.pageContext = pageContext;

}

This method is called by the JSP container before the tag handler is used. The TagSupport implementation simply sets an instance variable to the current PageContext object. The

PageContext

provides access to the request and response objects and all the JSP scope variables, and it implements a number of utility methods that the tag handler may use.

Appendix B , includes a

complete list of all PageContext methods.

When the start tag is encountered, the JSP container calls the doStartTag( ) method, implemented like this in the TagSupport class: public int doStartTag( ) throws JspException {

return SKIP_BODY;

}

This method gives the tag handler a chance to initialize itself, perhaps verifying that all attributes have valid values. Another use for this method is to decide what to do with the element's body content, if a body exists. The method returns an int, which must be one of two values defined by the Tag interface: SKIP_BODY or EVAL_BODY_INCLUDE. The default implementation returns SKIP_BODY.

As the name implies, this tells the JSP container to ignore the body completely. If

EVAL_BODY_INCLUDE

is returned instead, the JSP container processes the body (for instance, executes scripting elements and other actions in the body) and includes the result in the response. A simple conditional tag—a replacement for a scriptlet with an if statement—can be created by testing some condition (set by action attributes) in the doStartTag( ) and returning either SKIP_BODY or

EVAL_BODY_INCLUDE

, depending on if the condition is true or false.

No matter which value the doStartTag( ) method returns, the JSP container calls doEndTag( ) when it encounters the end tag: public int doEndTag( ) throws JspException {

return EVAL_PAGE;

}

This is the method that most tag handlers override to do the real work. It can also return one of two int

values defined by the Tag interface. The TagSupport class returns EVAL_PAGE, telling the

JSP container to continue to process the rest of the page. But a tag handler can also return

SKIP_PAGE

, which aborts the processing of the rest of the page. This is appropriate for a tag handler that forwards processing to another page or that sends a redirect response to the browser, like the

<ora:redirect>

custom action introduced in Chapter 8 .

An example of a custom action that can be implemented as a simple tag handler is the

<ora:addCookie>

action, introduced in Chapter 10 . The tag handler class is called

com.ora.jsp.tags.generic.AddCookieTag

and extends the TagSupport class to inherit most of the Tag interface method implementations: package com.ora.jsp.tags.generic; import javax.servlet.http.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import com.ora.jsp.util.*; public class AddCookieTag extends TagSupport {

The <ora:addCookie> action has two mandatory attributes, name and value, and one optional attribute, maxAge. Each attribute is represented by an instance variable and a standard property setter method: private String name;

private String value;

private String maxAgeString;

public void setName(String name) {

this.name = name;

}

public void setValue(String value) {

this.value = value;

}

public void setMaxAge(String maxAgeString) {

this.maxAgeString = maxAgeString;

}

The purpose of the custom action is to create a new javax.servlet.Cookie object, with the name, value, and max age values specified by the attributes, and to add the cookie to the response. The tag handler class overrides the doEndTag( ) method to carry out this work: public int doEndTag( ) throws JspException {

int maxAge = -1;

if (maxAgeString != null) {

try {

maxAge = Integer.valueOf(maxAgeString).intValue( );

}

catch (NumberFormatException e) {

throw new JspException("Invalid maxAge: " +

e.getMessage( ));

}

}

CookieUtils.sendCookie(name, value, maxAge,

(HttpServletResponse) pageContext.getResponse( ));

return EVAL_PAGE;

}

The maxAge attribute is optional, so before the corresponding String value is converted into an int

, a test is performed to see if it is set or not. You may wonder why similar tests are not done for the name

and value variables. The reason is that the JSP container verifies that all mandatory attributes are set in the custom action. If a mandatory attribute is not set, the JSP container refuses to process the page, so you can always be sure that a variable corresponding to a mandatory attribute has a value. I describe how to specify a mandatory attribute at the end of this chapter.

The code that actually creates the Cookie object and adds it to the response object is executed by the sendCookie( )

method in the com.ora.jsp.util.CookieUtils class. This is a pretty common practice; the tag handler is just a simple adapter for logic that's implemented in another class, providing a JSP-specific interface to the reusable class.

One last thing to note in this example is that the property setter method for the maxAge attribute, and the corresponding instance variable, is of type String, even though it's later converted to an int before it's used. In a regular bean, you would likely make it a property of type int to begin with

instead. Using a String property and converting it to an int in the tag handler is not necessarily the best implementation strategy, but it's the safest. A JSP 1.1-compliant container should automatically

convert a literal string attribute value to the appropriate type, as shown in Table 16.1

.

Property Type

Table 16.1, Conversion of String Value to Property Type

Conversion Method

boolean or Boolean Boolean.valueOf(String) byte or Byte char or Character double or Double int or Integer float or Float long or Long

Byte.valueOf(String)

String.charAt(int)

Double.valueOf(String)

Integer.valueOf(String)

Float.valueOf(String)

Long.valueOf(String)

This is a very recent clarification of the specification, documented in the specification errata document available at http://java.sun.com/products/jsp/ . Even though Tomcat 3.2 works according to the updated specification, other early implementations may not. If the conversion from a String to the appropriate type is not done by the container, a page author has to use a request-time attribute expression to set a non-String attribute value:

<ora:addCookie name="myCookie"

value="myValue"

maxAge="<%= 2592000 %>" />

That's likely to cause at least some confusion; it can be avoided by taking care of the conversion in the tag handler instead.

Whether to count on the container to do the conversion or to do it in the tag handler depends on how mature container implementations are when you read this. Letting the container take care of it is easiest, of course, but if the containers you plan to deploy with your application are still first-generation

JSP 1.1 implementations, you should test to make sure they handle the conversion correctly.

The tag handler class should also implement the release( ) method, to release all references to objects that it has acquired: public void release( ) {

name = null;

value = null;

maxAgeString = null;

super.release( );

}

The release( ) method is called when the tag handler is no longer needed. The AddCookieTag class sets all its properties to null and calls super.release( ) to let the TagSupport class do the same. This makes all property objects available for garbage collection.

© 2001, O'Reilly & Associates, Inc.

16.3 Processing the Action Body

As you can see, it's easy to develop a tag handler that doesn't need to do anything with the action element's body. For a tag handler that does need to process the body, however, just a few more methods are needed. They are defined by the BodyTag interface, which extends the Tag interface.

The action element's body has many possible uses. It can be used for input values spanning multiple

lines; the SQL custom actions introduced in Chapter 9 , use the body this way. The SQL statement is

often large, so it's better to let the page author write it in the action body instead of forcing it to fit on one line, which is a requirement for an attribute value. The body can also contain nested actions that rely on the enclosing action in some way. The <ora:sqlTransaction> action, also from

Chapter

9

, provides the nested SQL actions with the DataSource object they use to communicate with the database, and ensures that the SQL statements in all actions are treated as one transaction that either fails or succeeds.

A third example is an action that processes the body content in one way or another before it's added to

the response. Chapter 12

, contains an example of an action that processes its XML body using the XSL stylesheet specified as an attribute. Later in this section we look at an action that replaces characters that have special meanings in HTML with the corresponding HTML character entities.

As with the Tag interface, there's a BodyTagSupport class that implements all the methods of the

BodyTag

interface, plus a few utility methods: public class BodyTagSupport extends TagSupport implements BodyTag {

A tag handler that implements the BodyTag interface is at first handled the same way as a tag handler implementing the Tag interface: the container calls all property setter methods and the doStartTag( )

method. But then things divert, as illustrated in

Figure 16.3

.

Figure 16.3. BodyTag interface methods

First of all, the BodyTagSupport class overrides the doStartTag( ) method inherited from the

TagSupport

class:

public int doStartTag( ) throws JspException {

return EVAL_BODY_TAG;

}

Instead of returning SKIP_BODY, like the TagSupport class does, it returns EVAL_BODY_TAG .

The EVAL_BODY_TAG value is valid only for a tag handler that implements the BodyTag interface. It means that not only should the action's body be processed, but the container must also make the result available to the tag handler.

To satisfy this requirement, the container uses a BodyContent object. This is a subclass of the

JspWriter

, the class used to write text to the response body. In addition to the inherited methods for writing to the object, the BodyContent class has methods that the tag handler can use to read the content.

This is how it works. The JSP container assigns a reference to a JspWriter to the implicit out variable at the top of the page. Everything that's added to the response body—either explicitly by JSP elements or implicitly by the JSP container (template text)—is written to out, so it ends up in the

JspWriter

before it's sent to the browser. When the JSP container encounters a custom action with a tag handler that implements the BodyTag interface, it temporarily reassigns out to a BodyContent object until the action's end tag is encountered. The content produced when the element body is processed is therefore buffered in the BodyContent object where the tag handler can read it.

The tag handler gets a reference to the BodyContent object through the setBodyContent( ) method:

...

protected BodyContent bodyContent;

...

public void setBodyContent(BodyContent b) {

this.bodyContent = b;

}

The BodyTagSupport class simply saves the reference to the BodyContent object in an instance variable.

Next, the container gives the tag handler a chance to initialize itself before the body is processed by calling doInitBody( ):

public void doInitBody( ) throws JspException {

}

The implementation in BodyTagSupport does nothing. A tag handler can, however, use this method to prepare for the first pass through the action body, perhaps initializing scripting variables that it makes available to the body. We look at this in more detail later. A tag handler that doesn't introduce variables rarely overrides this method.

When the body has been processed, the doAfterBody( ) method is invoked:

public int doAfterBody( ) throws JspException {

return SKIP_BODY;

}

A tag handler can use this method to read the buffered body content and process it in some way. This method also gives the tag handler a chance to decide whether the body should be processed again. If so, it returns the EVAL_BODY_TAG value. We'll look at an example of an iteration action that takes advantage of this later. The BodyTagSupport implementation returns SKIP_BODY to let the processing continue to the doEndTag( ) method. As with a tag handler implementing the Tag interface, this method returns either EVAL_PAGE or SKIP_PAGE.

Let's look at a tag handler class that extends the BodyTagSupport class. The EncodeHTMLTag class is the tag handler class for a custom action called <ora:encodeHTML>. This action reads its body, replaces all characters with special meanings in HTML (single quotes, double quotes, less-than and greater-than symbols, and ampersands) with their corresponding HTML character entities (&#39;,

&#34;

, &lt;, &gt;, and &amp;) and inserts the result in the response body.

Example 16.1

shows

how the action can be used in a JSP page, and Figure 16.4

what the processed result looks like in a browser.

Example 16.1. A JSP Page Using the <ora:encodeHTML> Action

<%@ page language="java" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<head>

<title>Encoded HTML Example</title>

</head>

<body>

<h1>Encoded HTML Example</h1>

The following text is encoded by the &lt;ora:encodeHTML&gt;

custom action:

<pre>

<ora:encodeHTML>

HTML 3.2 Documents start with a <!DOCTYPE>

declaration followed by an HTML element containing

a HEAD and then a BODY element:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">

<HTML>

<HEAD>

<TITLE>A study of population dynamics</TITLE>

... other head elements

</HEAD>

<BODY>

... document body

</BODY>

</HTML>

</ora:encodeHTML>

</pre>

</body>

</html>

Figure 16.4. A JSP page with HTML source processed by the <ora:encodeHTML> action

Note how the body of the <ora:encodeHTML> action in

Example 16.1

contains HTML elements.

Unless the special characters were converted to HTML character entities, the browser would interpret the HTML and show the result instead of the elements themselves. Besides static text, the action body can contain any JSP element. A more realistic example of the use of this action is to insert text from a database into a JSP page, without having to worry about how special characters in the text are interpreted by the browser.

The tag handler class is very trivial, as shown in

Example 16.2

.

Example 16.2. The EncodeHTMLTag Class

package com.ora.jsp.tags.generic; import java.io.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import com.ora.jsp.util.*; public class EncodeHTMLTag extends BodyTagSupport {

public int doAfterBody( ) throws JspException {

BodyContent bc = getBodyContent( );

JspWriter out = getPreviousOut( );

try {

out.write(StringFormat.toHTMLString(bc.getString( )));

}

catch (IOException e) {} // Ignore

return SKIP_BODY;

}

}

The action doesn't have any attributes, so the tag handler doesn't need any instance variables and property access methods. The tag handler can reuse all BodyTag methods implemented by the

BodyTagSupport

class except for the doAfterBody( ) method.

In the doAfterBody( ) method, two utility methods provided by the BodyTagSupport class are used. The getBodyContent( ) method returns a reference to the BodyContent object that contains the result of processing the action's body. The getPreviousOut( ) method returns the

BodyContent

of the enclosing action (if any) or the main JspWriter for the page if the action is at the top level.

You may be wondering why the method is called getPreviousOut( ) as opposed to getOut(

)

. The name is intended to emphasize the fact that you want to use the object assigned as the output to the enclosing element in a hierarchy of nested action elements. Say you have the following action elements in a page:

<xmp:foo>

<xmp:bar>

Some template text

</xmp:bar>

</xmp:foo>

The JSP container first creates a JspWriter and assigns it to the out variable for the page. When it encounters the <xmp:foo> action, it creates a BodyContent object and temporarily assigns it to the out

variable. It creates another BodyContent for the <xmp:bar> action and, again, assigns it to out

. The container keeps track of this hierarchy of output objects. Template text and output produced by the standard JSP elements end up in the current output object. Each element can get access to its own BodyContent object by calling the getBodyContent( ) method and reading the content.

For the <xmp:bar> element, the content is the template text. After processing the content, it can write it to the <xmp:foo> body by getting the BodyContent for this element through the getPreviousOut( )

method. Finally, the <xmp:foo> element can process the content provided by the <xmp:bar> element and add it to the top-level output object: the JspWriter object it gets by calling the getPreviousOut( ) method.

The tag handler in Example 16.2

converts all the special characters it finds in its BodyContent

object using the toHTMLString( ) method in the com.ora.jsp.utils.StringFormat

class, introduced in Chapter 6 . It gets the content of the BodyContent by using the getString( )

method, and uses it as the argument to the toHTMLString( ) method. The result is written to the

JspWriter

obtained by calling getPreviousOut( ). The doAfterBody( ) method then returns SKIP_BODY, since no iteration is needed.

© 2001, O'Reilly & Associates, Inc.

16.4 Letting Actions Cooperate

Now that you've seen how to develop basic tag handlers, let's discuss some more advanced features. In this section, we look at tag handlers that let a page author use custom actions that cooperate with each other.

You have seen examples of this throughout this book. For instance, in Chapter 9 , various types of value

actions are nested within the body of an <ora:sqlQuery> action to set the values of place holders in the SQL statement. Another example is the <ora:encodeURL> action with nested <ora:param> actions, which are used in

Chapter 8

:

<ora:encodeURL url="product1.jsp">

<ora:param name="id" value="<%= product.getId( )%>" />

</ora:encodeURL>

How does the <ora:param> action tell the enclosing <ora:encodeURL> action about the parameter it defines? The answer to this question lies in a couple of Tag interface methods and a utility method implemented by the TagSupport class that I skipped earlier.

The Tag interface methods are setParent( ) and getParent( ), implemented like this by the

TagSupport

class:

...

private Tag parent;

...

public void setParent(Tag t) {

parent = t;

}

public Tag getParent( ) {

return parent;

}

These two methods are standard accessor methods for the parent instance variable. For a nested action element, the setParent( ) method is always called on the tag handler with the value of the enclosing Tag as its value. This way, a nested tag handler always has a reference to its parent. So a tag handler at any nesting level can ask for its parent using getParent( ), and then ask for the parent's parent, and so on until it reaches a Tag that doesn't have a parent (that is, getParent( ) returns

null

). This means it has reached the top level.

This is part of the puzzle. However, a tag handler is usually interested only in finding a parent it's been designed to work with. It would be nice to have a method that works its way up the hierarchy until it finds the parent of interest. That's exactly what the findAncestorWithClass( ) method implemented by the TagSupport class does: public static final Tag findAncestorWithClass(Tag from, Class klass) {

boolean isInterface = false;

if (from == null ||

klass == null ||

(!Tag.class.isAssignableFrom(klass) &&

!(isInterface = klass.isInterface( )))) {

return null;

}

for (;;) {

Tag tag = from.getParent( );

if (tag == null) {

return null;

}

if ((isInterface && klass.isInstance(tag)) ||

klass.isAssignableFrom(tag.getClass( )))

return tag;

else

from = tag;

}

}

First of all, note that this is a static method. Consequently, it can be used even by tag handlers that implement the Tag interface directly, instead of extending the TagSupport class. The method takes two arguments: the tag handler instance to start searching from, and the class or interface of the parent.

After making sure that all parameters are valid, it starts working its way up the nested tag handlers. It stops when it finds a tag handler of the specified class or interface and returns it. If the specified parent type is not found, the method returns null.

This is all that's needed to let a nested action communicate with its parent: the parent accessor

methods, and the method that walks the action hierarchy to find the parent of interest. Example 16.3

shows how the ParamTag class uses this mechanism to find the enclosing EncodeURLTag instance.

Example 16.3. The ParamTag Class

package com.ora.jsp.tags.generic; import java.net.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; public class ParamTag extends TagSupport {

private String name;

private String value;

public void setName(String name) {

this.name = name;

}

public void setValue(String value) {

this.value = value;

}

public int doEndTag( ) throws JspException {

Tag parent = findAncestorWithClass(this, ParamParent.class);

if (parent == null) {

throw new JspException("The param action is not " +

"enclosed by a supported action type");

}

ParamParent paramParent = (ParamParent) parent;

paramParent.setParam(name, URLEncoder.encode(value));

return EVAL_PAGE;

}

}

The class has two instance variables, name and value, and the corresponding setter methods. The most interesting method is the doEndTag( ) method. This method first uses the findAncestorWithClass( )

method to try to locate the enclosing EncodeURLTag instance.

Note that this is not the class name used as the argument value. Instead, the ParamParent interface is used. The reason is that the <ora:param> action is supported in the body of other actions besides

<ora:encodeURL>

, such as the <ora:redirect> action. The ParamParent interface is implemented by all tag handlers for actions that can contain nested <ora:param> actions: package com.ora.jsp.tags.generic; public interface ParamParent {

void setParam(String name, String value);

}

The interface defines one method: the setParam( ) method. This is the method the nested

ParamTag

tag handler uses to communicate with its parent. For each nested <ora:param> action, the setParam( ) method gets called when the parent's action body is processed. The name and value for each <ora:param> action are accumulated in the parent tag handler, ready to be used when the

parent's doEndTag( ) method is called. Example 16.4

shows the setParam( ) and doEndTag(

)

methods implemented by the EncodeURLTag class.

Example 16.4. EncodeURLTag

...

private Vector params;

...

public void setParam(String name, String value) {

if (params == null) {

params = new Vector( );

}

Param param = new Param(name, value);

params.addElement(param);

}

public int doEndTag( ) throws JspException {

StringBuffer encodedURL = new StringBuffer(url);

if (params != null && params.size( ) > 0) {

encodedURL.append('?');

boolean isFirst = true;

Enumeration e = params.elements( );

while (e.hasMoreElements( )) {

Param p = (Param) e.nextElement( );

if (!isFirst) {

encodedURL.append('&');

}

encodedURL.append(p.getName( )).append('=').

append(p.getValue( ));

isFirst = false;

}

}

try {

HttpServletResponse res =

(HttpServletResponse) pageContext.getResponse( );

JspWriter out = pageContext.getOut( );

out.print(res.encodeURL(encodedURL.toString( )));

}

catch (IOException e) {}

return SKIP_BODY;

}

In setParam( ), the parameter name and value are saved as instances of a simple value holder class named Param, held in a Vector. In the doEndTag( ) method, each parameter's name/value pair is added to the URL before the complete URL is encoded to support session tracking through URL rewriting. If you don't remember what all of this means, you can refresh your memory by looking at

Chapter 8 again.

© 2001, O'Reilly & Associates, Inc.

16.5 Creating New Variables Through Actions

Actions can also cooperate through objects available in the standard JSP scopes (page, request, session, and application). One example of this type of cooperation is illustrated by the three standard JSP actions:

<jsp:useBean>

, <jsp:setProperty>, and <jsp:getProperty>. The <jspUseBean> action creates a new object and makes it available in one of the JSP scopes. The other two actions can then access the properties of the object by searching for it in the scopes. Besides making the object available in one of the scopes, the <jsp:useBean> action also makes it available as a scripting variable, so it can be accessed by scripting elements in the page.

The JSP 1.1 specification defines that an attribute named id typically is used to name a variable created

by an action.

[1]

The value of the id attribute must be unique within the page. Because it's used as a scripting variable name, it must also follow the variable name rules for the scripting language. For Java, this means it must start with a letter followed by a combination of letters and digits, and must not contain special characters, such as a dot or a plus sign. The attribute used in another action to refer to the variable can be named anything, but the convention established by the standard actions is to call it name.

[1]

If an action creates more than one variable, the id attribute is typically used to name one of them.

When a custom action creates a variable, it must cooperate with the JSP container to make it happen. To understand how this works, recall that the JSP page is turned into a servlet by the JSP container. The JSP container needs to generate code that declares the scripting variable in the generated servlet and assigns the variable a value. Before getting into how the tag handler and the container cooperate, let's look at the kind of code that is generated for the <ora:useProperty> custom action introduced in

Chapter 8 .

Here's a JSP page fragment:

<jsp:useBean id="catalog"

class="com.ora.jsp.beans.shopping.CatalogBean" />

<ora:useProperty name="catalog" id="prod"

property="product" arg="1"

className="com.ora.jsp.beans.shopping.ProductBean" />

<jsp:getProperty name="prod" property="name" />

<% prod.getName( ); %>

The <jsp:useBean> action creates an instance of the CatalogBean (or locates an existing instance) and saves it in the page scope with the name catalog. It also declares a scripting variable named catalog

and sets it to the same CatalogBean instance. The <ora:useProperty> custom action

retrieves the product property from the CatalogBean and introduces it as a page scope object named prod

and a scripting variable with the same name, in the same manner as the <jsp:useBean> action.

Finally, the value of prod is added to the response twice: first using the <jsp:getProperty> action, and then again using a JSP expression.

This JSP page fragment results in code similar to Example 16.5

in the generated servlet.

Example 16.5. Code Generated for JSP Actions

// Code for <jsp:useBean> com.ora.jsp.beans.shopping.CatalogBean catalog = null; catalog= (com.ora.jsp.beans.shopping.CatalogBean)

pageContext.getAttribute("catalog",PageContext.PAGE_SCOPE); if ( catalog == null ) {

try {

catalog = (com.ora.jsp.beans.shopping.CatalogBean)

Beans.instantiate(getClassLoader( ),

"com.ora.jsp.beans.shopping.CatalogBean");

} catch (Exception exc) {

throw new ServletException ("Cannot create bean of class "+

"com.ora.jsp.beans.shopping.CatalogBean");

}

pageContext.setAttribute("catalog", catalog, PageContext.PAGE_SCOPE);

}

...

// Code for <ora:useProperty> com.ora.jsp.tags.generic.UsePropertyTag _jspx_th_ora_useProperty_1 =

new com.ora.jsp.tags.generic.UsePropertyTag( );

_jspx_th_ora_useProperty_1.setPageContext(pageContext);

_jspx_th_ora_useProperty_1.setParent(null);

_jspx_th_ora_useProperty_1.setId("prod");

_jspx_th_ora_useProperty_1.setName("catalog");

_jspx_th_ora_useProperty_1.setProperty("product");

_jspx_th_ora_useProperty_1.setArg("1");

_jspx_th_ora_useProperty_1.setClassName(

"com.ora.jsp.beans.shopping.ProductBean"); try {

_jspx_th_ora_useProperty_1.doStartTag( );

if (_jspx_th_ora_useProperty_1.doEndTag( ) == Tag.SKIP_PAGE)

return;

} finally {

_jspx_th_ora_useProperty_1.release( );

}

com.ora.jsp.beans.shopping.ProductBean prod = null; prod = (com.ora.jsp.beans.shopping.ProductBean)

pageContext.findAttribute("prod");

...

// Code for <jsp:getProperty> out.print(pageContext.findAttribute("prod"), "name")));

...

// Code for <%= prod.getName( ) %> out.print( prod.getName( ) );

The <jsp:useBean> action results in code for locating or creating the CatalogBean, and declaring and assigning a Java variable named catalog. But since we're talking about custom actions here, let's focus on the code generated for the <ora:useProperty> action.

First, a tag handler instance is created and initialized with the standard properties (pageContext and parent

) plus all properties corresponding to the action attributes. Next, the doStartTag( ) and doEndTag( )

methods are called. Then comes the code that makes the object created by the action available as a scripting variable. Note how a variable with the name specified by the id attribute (prod) is declared, using the type specified by the className attribute. Also note that the variable is declared at the top level of the method. This means that it's available to scripting elements anywhere on the page after the action element. The variable is then assigned the value of the object with same name located in one of the standard JSP scopes, using the findAttribute( ) method. This method searches through the scopes, in the order page, request, session, and application, until it finds the specified object.

With the object available in the JSP page scope, the code generated for the <jsp:getProperty> action can find it. Since it's also assigned to a Java variable, the JSP expression works correctly as well.

At least two things are required for a tag handler to create a new object and make it accessible for other actions and JSP scripting code:

1. The JSP container must know the name and the Java type for the object, so it can generate the code for the variable declaration.

2. The object must be placed in one of the JSP scopes, so it can be found by findAttribute( ) and assigned to the variable.

The first requirement is fulfilled by a class called TagExtraInfo . When you develop a tag handler for an action that introduces an object, you must also create a subclass of the TagExtraInfo class. The

JSP container consults an instance of this class when it generates the code.

Example 16.6

shows the class

associated with the <ora:useProperty> action.

Example 16.6. UsePropertyTagExtraInfo Class

package com.ora.jsp.tags.generic; import javax.servlet.jsp.tagext.*; public class UsePropertyTagExtraInfo extends TagExtraInfo {

public VariableInfo[] getVariableInfo(TagData data) {

return new VariableInfo[]

{

new VariableInfo(data.getAttributeString("id"),

data.getAttributeString("className"),

true,

VariableInfo.AT_END)

};

}

}

The method used by the JSP container during code generation is called getVariableInfo( ). It returns an array of VariableInfo objects, one per variable introduced by the tag handler.

The VariableInfo class is a simple bean with four properties, all of them initialized to the values passed as arguments to the constructor: varName, className, declare, and scope. The meaning of the first two is not hard to guess: the name of the variable and the name of its class. The declare property is a boolean, in which true means that a new variable is created by the action. In other words, a declaration of the variable must be added to the generated servlet. A value of false means that the variable has already been created by another action or by another occurrence of the same action, so the generated code already contains the declaration. This is all the information the JSP container needs to generate the code for the variable declaration; the first requirement is satisfied.

The scope property has nothing to do with the JSP scopes we have seen so far (page, request, session, and application). Instead, it defines where the new variable is available to JSP scripting elements. A value of AT_BEGIN means that it is available from the action's start tag and stays available after the action's end tag. AT_END means it is not available until after the action's end tag. A variable with scope NESTED is available only in the action's body, between the start and the end tags. The scope therefore controls where the variable declaration and value assignment code is generated, and the tag handler class must make sure the variable is available in one of the JSP scopes at the appropriate time.

The UsePropertyTagExtraInfo class sets the scope to AT_END . As you can see in Example 16.5

, this results in the variable declaration and assignment code being added after the doEndTag( ) call. To satisfy the second requirement, the tag handler must therefore give the variable a value and save it in one of the JSP scopes at the very latest in the doEndTag( ) method.

Example 16.7

shows the doEndTag(

)

method for the UsePropertyTag class.

Example 16.7. Saving the New Object in a JSP Scope

public int doEndTag( ) throws JspException {

Object obj = pageContext.findAttribute(name);

if (obj == null) {

throw new JspException("Variable " + name + " not found");

}

Object propObj = getProperty(obj, property, className);

pageContext.setAttribute(id, propObj);

return SKIP_BODY;

}

The value is added to the page scope by calling the setAttribute( ) method on the current

PageContext

object, using the name specified by the id attribute.

If the scope is specified as AT_BEGIN instead, the declaration is added before the doStartTag( ) call and the assignment code is added right after the call. In this case, the tag handler must save the variable in a JSP scope in the doStartTag( ) method. If the tag handler implements BodyTag, assignment code is also added so that it is executed for every evaluation of the body, and after the call to doAfterBody( )

. This allows the tag handler to modify the variable value in the doAfterBody(

)

method, so each evaluation of the body has a new value. When we look at an iteration action later, you'll see why this is important.

Finally, if the scope is set to NESTED, both the declaration and the value assignment code are inserted in the code block representing the action body. The tag handler must therefore make the variable available in either the doStartTag( ) method or the doInitBody( ) method, and can also modify the value in the doAfterBody( ) method.

The UsePropertyTagExtraInfo class sets the varName and className properties of the

VariableInfo

bean to the values of the id and className attributes specified by the page author in the JSP page. This is done using another simple class named TagData , passed as the argument to the getVariableInfo( )

method, as shown in

Example 16.6

. The TagData instance is created by the

JSP container and contains information about all action attributes that the page author specified in the JSP page. It has two methods of interest. First, the getAttributeString( ) method, used in

Example

16.6

, simply returns the specified attribute as a String. But some attribute values may be specified by a

JSP expression instead of a string literal, so-called request-time attributes. Since these values are not known during the translation phase, the TagData class also provides the getAttribute( ) method to indicate if an attribute value is a literal string, a request-time attribute, or not set at all. The getAttribute( )

method returns an Object. If the attribute is specified as a request-time value, the

special REQUEST_TIME_VALUE object is returned. Otherwise, a String is returned, or null if the attribute is not set.

© 2001, O'Reilly & Associates, Inc.

16.6 Developing an Iterating Action

As I alluded to earlier, a tag handler can iterate over the element's body until some condition is true.

The evaluation of the body may be different for each iteration, since the tag handler can introduce a variable (used in the body) that changes its value. An example of an iterating action is the

<ora:loop>

used in this book. It can be used to iterate over the element body once for each value in an array, a java.util.Vector, a java.util.Dictionary, or a java.util.Enumeration

. Here's an example of how the <ora:loop> action can be used:

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<html>

<body bgcolor="white">

<%

String[] test = new String[4];

test[0] = "first";

test[1] = "second";

test[2] = "third";

test[3] = "fourth";

pageContext.setAttribute("test", test);

%>

<pre>

<ora:loop name="test" loopId="x" className="java.lang.String">

Current value: <%= x %>

</ora:loop>

</pre>

</body>

</html>

Here, the <ora:loop> tag iterates over the elements of a String array, adding the current value to the response using a JSP expression in the action's body.

The com.ora.jsp.tags.generic.LoopTag class is the tag handler class for the

<ora:loop>

action. It extends BodyTag support and has four properties: public class LoopTag extends BodyTagSupport {

private String name;

private String loopId;

private String className;

private String property;

...

A standard property setter method is provided for each property. This is no different than in previous examples, so it's not shown here. The name, loopId, and className properties are mandatory.

The name is the name of a JSP scope variable of one of the types listed earlier. The current value of the data structure is made available in the element body through a variable with the name specified by loopId

, of the type specified by className. Optionally, property can be specified. If it is, it's used to get the data structure from the specified property of the bean named by name, instead of using the name object itself as the data structure.

To make the loopId variable available in the element's body, a TagExtraInfo subclass is needed, as described in the previous section. The LoopTagExtraInfo class looks like this: public class LoopTagExtraInfo extends TagExtraInfo {

public VariableInfo[] getVariableInfo(TagData data) {

return new VariableInfo[]

{

new VariableInfo(data.getAttributeString("loopId"),

data.getAttributeString("className"),

true,

VariableInfo.NESTED)

};

}

}

It introduces a variable named by the loopId attribute, with the type specified by the className attribute. The scope is specified as NESTED, meaning the variable is available only within the action element's body.

In addition to the property variables, the tag handler class has an Enumeration instance variable: private Enumeration enum;

This variable is initiated by the doStartTag( ) method: public int doStartTag( ) throws JspException {

Object obj = pageContext.findAttribute(name);

if (obj == null) {

throw new JspException("Variable " + name + " not found");

}

Object mvObj = obj;

try {

// Get the multi-value object using the specified property

// getter method, if any

if (property != null) {

mvObj = getProperty(obj, property);

}

enum = getEnumeration(mvObj);

}

catch (JspException e) {

throw new JspException("Error getting loop data from " +

name + ": " + e.getMessage( ));

}

// Set the first loop value, if any

if (enum != null && enum.hasMoreElements( )) {

Object currValue = enum.nextElement( );

pageContext.setAttribute(loopId, currValue);

return EVAL_BODY_TAG;

}

else {

return SKIP_BODY;

}

}

After verifying that there really is an object with the specified name, a test is done to see if a property name is specified. If it is, the getProperty( ) method is called to retrieve the property value from the specified object so it can be used for the iteration. If a property name is not specified, the object itself is used. All the supported data structure types can be turned into an Enumeration. That's done by calling the getEnumeration( ) method. The getProperty( ) method and the getEnumeration( )

method are not shown here, because this code is just plain Java code that has nothing to do with implementing iteration in a tag handler. You can look at the source code to see how they work.

When the Enumeration has been created, the doStartTag( ) method initializes the loopId variable and places it in the JSP page scope. As you learned in the previous section, the code generated for the page uses the information gained from the LoopTagExtraInfo class to declare a Java variable and assign it the value it finds in one of the JSP scopes, right after the doStartTag( ) call.

When the body has been evaluated, the doAfterBody( ) method is called: public int doAfterBody( ) throws JspException {

if (enum.hasMoreElements( )) {

Object currValue = enum.nextElement( );

pageContext.setAttribute(loopId, currValue);

return EVAL_BODY_TAG;

}

else {

return SKIP_BODY;

}

}

The Enumeration is tested to see if it contains any more values. If it does, the loopId page scope variable is reassigned to the new value, and EVAL_BODY_TAG is returned to evaluate the body again.

When the end of the Enumeration is reached, SKIP_BODY is returned to break the iteration.

When the doAfterBody( ) method returns SKIP_BODY, the doEndTag( ) method is called: public int doEndTag( ) throws JspException {

// Test if bodyContent is set, since it will be null if the

// body was never evaluated (doStartTag returned SKIP_BODY)

if (getBodyContent( ) != null) {

try {

getPreviousOut().print(getBodyContent( ).getString( ));

}

catch (IOException e) {}

}

return EVAL_PAGE;

}

For every iteration, the content of the evaluated body is buffered in the BodyContent instance assigned to the tag handler. In the doEndTag( ), the content is simply moved to the parent's

BodyContent

instance or the main JspWriter instance for the page. An alternative to accumulating the content until the doEndTag( ) method is called is to write it to the parent's output stream already in the doAfterBody( ) method, using the same code as shown here.

class Versus className

You may have noticed that all the custom actions in this book use an attribute named className

to specify a class name, while all the standard JSP actions use an attribute named class for the same purpose.

The reason for this inconsistency is the fact that tag handlers are handled as JavaBeans components with regards to the attributes, combined with an unfortunate name clash.

The attribute is used to specify a class name, in other words a String. If the attribute name class is used, the corresponding property setter method must be named setClass( )

, with a String as its argument. The Object class, however, implements a method named getClass( ) that returns a Class object. The java.beans.Introspector

class, used to figure out which properties a bean supports by looking for accessor methods, doesn't approve of what it sees as a type mismatch between the setter and getter methods for the class property. It therefore refuses to accept that class is a valid bean property.

To work around this problem, all custom actions in this book use an attribute called className

instead of class.

© 2001, O'Reilly & Associates, Inc.

16.7 Creating the Tag Library Descriptor

Now you have a good idea about what the code for a tag handler looks like. But when the JSP container converts custom action elements into code that creates and calls the correct tag handler, it needs information about which tag handler implements which custom action element. It gets this information from the Tag Library Descriptor (TLD). As you will see in the next section, the JSP container also uses the TLD information to verify that the attribute list for an action element is correct.

The TLD is an XML file with information about all custom actions in one library. A JSP page that uses custom actions must identify the corresponding TLD and the namespace prefix used for the actions in the page with the taglib directive (this is described in more detail later).

<%@ taglib uri="/WEB-INF/tlds/orataglib_1_0.tld" prefix="ora" %>

...

<ora:redirect page="main.jsp" />

The JSP page then uses the TLD to find the information it needs when it encounters an action element with a matching prefix.

Example 16.8

shows a part of the TLD for the custom actions in this book.

Example 16.8. Tag Library Descriptor (TLD)

<?xml version="1.0" encoding="ISO-8859-1" ?>

<!DOCTYPE taglib

PUBLIC "-//Sun Microsystems, Inc.//DTD JSP Tag Library 1.1//EN"

"http://java.sun.com/j2ee/dtds/web-jsptaglibrary_1_1.dtd">

<taglib>

<tlibversion>1.0</tlibversion>

<jspversion>1.1</jspversion>

<shortname>ora</shortname>

<uri>

/orataglib

</uri>

<info>

A tab library for the examples in the O'Reilly JSP book

</info>

<tag>

<name>redirect</name>

<tagclass>com.ora.jsp.tags.generic.RedirectTag</tagclass>

<bodycontent>JSP</bodycontent>

<info>

Encodes the url attribute and possible param tags in the body

and sets redirect headers.

</info>

<attribute>

<name>page</name>

<required>true</required>

<rtexprvalue>true</rtexprvalue>

</attribute>

</tag>

...

</taglib>

At the top of the TLD file, you find a standard XML declaration and a DOCTYPE declaration, specifying the Document Type Definition (DTD) for this file. A DTD defines the rules for how elements in an XML file must be used, such as the order of the elements, which elements are mandatory and which are optional, if an element can be included multiple times, etc. If you're not familiar with XML, don't worry about this. Just accept the fact that you need to copy the first two

elements of Example 16.8

faithfully into your own TLD files. Regarding the order of the elements, just

follow the same order as in

Example 16.8

. Whether an element is mandatory or optional is spelled out in the following descriptions of each element.

After the two declarations, the first element in the TLD file must be the <taglib> element. This is the main element for the TLD, enclosing all more specific elements that describe the library. Within the body of the <taglib> element, you can specify elements that describe the library as such, as well as each individual tag handler. Let's start with the five elements that describe the library itself.

The <tlibversion> element is mandatory and is used to specify the tag library version. The version should be specified as a series of numbers separated by dots. In other words, the normal conventions for software version numbers, such as 1.1 or 2.0.3, should be used.

The <jspversion> element, specifying the version of the JSP specification that the library depends on, is optional. The default value is 1.1.

The <shortname> element is intended to be used by page authoring tools. It's a mandatory element that should contain the default prefix for the action elements. In

Example 16.8

the value is ora, meaning that an authoring tool by default generates custom action elements using the ora prefix, for instance <ora:redirect page="main.jsp">. This element value can also be used by a tool as

the value of the prefix attribute if it generates the taglib directive in the JSP page. The element value must not include whitespace or other special characters, or start with a digit or underscore.

The <uri> element is also intended to benefit authoring tools. The value can be used as the default value for the uri attribute in a taglib directive. It's an optional element, following the same character rules as the <shortname> element.

The last element that describes the library as such is the optional <info> element. It can be used to provide a short description of the library, perhaps something a tool might display to help users decide if the library is what they are looking for.

Besides the general elements, the TLD must include at least one <tag> element. The <tag> element contains other elements that describe different aspects of the custom action: <name>, <tagclass>,

<teiclass>

, <bodycontent>, <info>, and <attribute>.

The <name> element is mandatory and contains the unique name for the corresponding custom action element.

The <tagclass> element, also mandatory, contains the fully qualified class name for the tag handler class.

If the action introduces variables or needs to do additional syntax validation as described in the next section, the optional <teiclass> element is used to specify the fully qualified class name for the

TagExtraInfo

subclass.

Another optional element is <bodycontent>. It can contain one of three values. A value of empty means that the action body must be empty. If the body can contain JSP elements, such as standard or custom actions or scripting elements, the JSP value should be used. All JSP elements in the body are processed, and the result is handled as specified by the tag handler (i.e., processed by the tag handler or sent through to the response body). This is also the default value, in case you omit the

<bodycontent>

element. The third alternative is tagdependent. This value means that possible

JSP elements in the body are not processed. Typically, this value is used when the body is processed by the tag handler and the content may contain characters that could be confused with JSP elements, for example, SELECT * FROM MyTable WHERE Name LIKE '<%>'. If a tag that expects this kind of body content is declared as JSP, the <%> is likely to confuse the JSP container. The tagdependent

value can be used to avoid this risk for confusion.

The <info> element can optionally be used to describe the purpose of the action.

The <tag> element must also contain an <attribute> element for each action attribute. Each element in turn contains other elements that describe the attribute: <name>, <required>, and

<rtexprvalue>

.

The mandatory <name> element contains the attribute name. The optional <required> element tells if the attribute is required or not. The values true, false, yes, and no are valid, with false being the default. Finally, the <rtexprvalue> element is an optional element that can have the same values as the <required> element. If the value is true or yes, a request-time attribute expression can be used to specify the attribute value, for instance 'attr="<%= request.getParameter("par") %>

'. The default value is false.

© 2001, O'Reilly & Associates, Inc.

16.8 Validating Syntax

The TLD for a tag library contains information about the attributes each action element supports.

Therefore, the JSP container can help by verifying that the custom action is used correctly by the page author, at least with respect to the attributes.

When the JSP container converts a JSP page to a servlet, it compares each custom action element to the specification of the action element in the TLD. First, it makes sure that the action name matches the name of an action specified in the TLD corresponding to the action element's prefix. It then looks at the attribute list in the page and compares it to the attribute specification in the TLD. If a required attribute is missing, or an attribute is used in the page but not specified in the TLD, it reports it as an error so the page author can correct the mistake.

But for some actions, it's not that simple. Some attributes may depend on the presence of other attributes. Attributes may be mutually exclusive, so that if one is used, the other must not be used. Or an optional attribute may require that another optional attribute is used as well. To be able to verify these kinds of dependencies, the JSP container asks the tag handler's TagExtraInfo subclass for assistance.

After the JSP container has checked everything it can on its own, it looks for a TagExtraInfo subclass, defined by the <teiclass> element, for the action. If one is defined, it puts all attribute information in an instance of the TagData class and calls the TagExtraInfo isValid( ) method: public boolean isValid(TagData data) {

// Mutually exclusive attributes

if (data.getAttribute("attr1") != null &&

data.getAttribute("attr2" != null) {

return false;

}

// Dependent optional attributes

if (data.getAttribute("attr3") != null &&

data.getAttribute("attr4" == null) {

return false;

}

return true;

}

A TagExtraInfo subclass can use the TagData instance to verify that all attribute dependencies are okay, as in this example. In JSP 1.1, unfortunately, there's no way to generate an appropriate error message; the method can only return false to indicate that something is not quite right. This will hopefully be rectified in a future version of JSP.

© 2001, O'Reilly & Associates, Inc.

16.9 How Tag Handlers May Be Reused

Creating new objects is a relatively expensive operation in Java. For high-performance applications, it's common to try to minimize the number of objects created and reuse the same objects instead. The JSP

1.1 specification describes how a tag handler instance can be reused within the code generated for a

JSP page if the same type of custom action appears more than once. The reuse is subject to a number of restrictions and relies on tag handler classes dealing with their internal state as specified. It's important to understand the reuse rules, so your tag handler classes behave as expected in a JSP implementation that takes advantage of this mechanism.

As discussed in the previous sections of this chapter, a tag handler's state is initiated through property setter methods corresponding to the action element's attributes. The tag handler is then offered a chance to do its thing in various stages, represented by the doStartTag( ), doInitBody( ), doAfterBody( )

, and doEndTag( ) methods. It's clear that the property values must be kept at least until the tag handler has done what it intends to do. But when can it safely reset its state? If a tag handler implements all logic in the doStartTag( ) method, can it reset all instance variables before it returns from this method? Or should it wait until the doEndTag( ) method is called? The answer is that it must not reset the state until the release( ) method is called. Let's use a JSP page fragment to discuss why:

<test:myAction attr1="one" attr2="two" />

<test:myAction attr1="one" attr2="new" />

In this case, a JSP container is allowed to use one instance of the tag handler for both

<test:myAction>

action elements, with generated code similar to this:

// Code for first occurrence

MyActionTag _jspx_th_test_myAction_1 = new MyActionTag( );

_jspx_th_test_myAction_1.setPageContext(pageContext);

_jspx_th_test_myAction_1.setParent(null);

_jspx_th_test_myAction_1.setAttr1("one");

_jspx_th_test_myAction_1.setAttr2("two");

_jspx_th_test_myAction_1.doStartTag( ); if (_jspx_th_test_myAction_1.doEndTag( ) == Tag.SKIP_PAGE)

return;

// Code for second occurrence

_jspx_th_test_myAction_1.setAttr2("new");

_jspx_th_test_myAction_1.doStartTag( );

if (_jspx_th_test_myAction_1.doEndTag( ) == Tag.SKIP_PAGE)

return;

_jspx_th_test_myAction_1.release( );

As you can see, all the property setter methods are called to initialize the instance for the first occurrence of the element. But for the second occurrence, only the setter method for the property with a different value is called. The release( ) method is called when the tag handler has been used for both occurrences. If the tag handler class resets all property variables in any method other than release( )

, the processing of the second action element fails.

The only scenario in which a tag handler can be reused in JSP 1.1 is the one described above. If the same action element is used multiple times on the same page but with different sets of attributes, the state of the tag handler is not guaranteed to be correct if the same instance is reused.

Reuse between pages, using a tag handler object pool, is not explicitly supported in JSP 1.1. For this reason, most JSP containers do not implement tag handler pooling today. To get your tag handler classes to work with the few that do, you must reset all properties before the tag handler is used to handle a new request. I recommend that you do this in the release( ) method, as shown in the examples in this chapter. Note that if some properties must have a default value set instead of null, you must set it in the release( ) method as well. A typical example is a primitive type property, such as an int property: public void release( ) {

aStringProperty = null;

anIntProperty = -1;

}

To make it easier for a container to reuse tag handlers, both within a page and between pages, a future version of JSP will likely introduce a method that resets all properties in a controlled manner.

© 2001, O'Reilly & Associates, Inc.

16.10 Packaging and Installing a Tag Library

During development, you may want to let the tag library classes and the TLD file reside as-is in the filesystem, since it makes it easy to change the TLD and modify and recompile the classes. Just make sure the class files are stored in a directory that's part of the classpath for the JSP container, such as the

WEB-INF/classes directory for the web application. The TLD must also be in a directory where the JSP container can find it. The recommended location is the WEB-INF/tlds directory. To identify the library with the TLD stored in this location, use a taglib directive in the JSP pages like this:

<%@ taglib uri="/WEB-INF/tlds/orataglib_1_0.tld" prefix="ora" %>

Here the uri attribute refers directly to the TLD file's location.

When you're done with the development, you may want to package all tag handler classes,

TagExtraInfo

classes, beans used by the tag handler classes, and the TLD in a JAR file. This makes it easier to install the library in an application. The TLD must be saved as /META-INF/taglib.tld within the JAR file.

To create the JAR file, first arrange the files in a directory with a structure like this:

META-INF/

taglib.tld

com/

ora/

jsp/

tags/

generic/

EncodeHTMLTag.class

...

util/

StringFormat.class

...

The structure for the class files must match the package names for your classes. Here a few of the classes in the tag library for this book are shown as an example.

With the file structure in place, use the jar command to create the JAR file: jar cvf orataglib_1_0.jar META-INF com

This command creates a JAR file named orataglib_1_0.jar containing the files in the META-INF and

com directories. Use any JAR filename that makes sense for your own tag library. Including the version number for the library is also a good idea, since it lets the users know which version of the library they are using.

You can now use the packaged tag library in any application. Just copy the JAR file to the application's

WEB-INF/lib directory and use a taglib directive like this in the JSP pages:

<%@ taglib uri="/WEB-INF/lib/orataglib_1_0.jar" prefix="ora" %>

Note that the uri attribute now refers to the JAR file instead of the TLD file. A JSP 1.1 container is supposed to be able to find the TLD file in the JAR file, but this is a fairly recent clarification of the specification. If the JSP container you use doesn't support this notation yet, you have to extract the

TLD file from the JAR file, save it somewhere else, for instance in WEB-INF/tlds, and let the uri attribute refer to the TLD file instead.

Instead of letting the taglib directive point directly to the TLD or JAR file, you can specify a symbolic name as the uri attribute value, and provide a mapping between this name and the real location in the WEB-INF/web.xml file for the application:

<%@ taglib uri="/orataglib" prefix="ora" %>

The WEB-INF/web.xml file must then contain the following elements:

<web-app>

...

<taglib>

<taglib-uri>

/orataglib

</taglib-uri>

<taglib-location>

/WEB-INF/lib/orataglib_1_0.jar

</taglib-location>

</taglib>

...

</web-app>

The <taglib-uri> element contains the symbolic name, and the <taglib-location> element contains the path to either the JAR file or the extracted TLD file.

© 2001, O'Reilly & Associates, Inc.

17.1 Using Connections and Connection Pools

In a JDBC-based application, a lot revolves around the java.sql.Connection interface. Before any database operations can take place, the application must create a Connection to the database. It then acts as the communication channel between the application and the database, carrying the SQL statements sent by the application and the results returned by the database. A Connection is associated with a database user account, to allow the database to enforce access control rules for the

SQL statements submitted through the Connection. Finally, the Connection is also the boundary for database transactions. Only SQL statements executed through the same Connection can make up a transaction. A transaction consists of a number of SQL statements that must either all succeed or all fail as one atomic operation. A transaction can be committed (the changes resulting from the statements are permanently saved) or rolled back (all changes are ignored) by calling Connection methods.

In a standalone application, a Connection is typically created once and kept open until the application is shut down. This is not surprising, since a standalone application serves only one user at a time, and all database operations initiated by a single user are typically related to each other. In a server application that deals with unrelated requests from many different users, it's not so obvious how to deal with connections. There are three things to consider: a Connection is time-consuming to create, it must be used for only one user at a time to avoid transaction clashes, and it is expensive to keep open.

Creating a Connection is an operation that can actually take a second or two to perform. Besides establishing a network connection to the database, the database engine must authenticate the user and create a context with various data structures to keep track of transactions, cached statements, results, and so forth. Creating a new Connection for each request received by the server, while simple to implement, is far too time-consuming in a high-traffic server application.

One way to minimize the number of times a connection needs to be created is to keep one

Connection

per servlet or JSP page that need access to the database. A Connection can be created when the web resource is initialized, and kept in an instance variable until the application is shut down.

As you will discover when you deploy an application based on this approach, this route will lead to numerous multithreading issues. Each request executes as a separate thread through the same servlet or

JSP page. Many JDBC drivers do not support multiple threads accessing the same Connection, causing all kinds of runtime errors. Others support it by serializing all calls, leading to poor scalability.

An even more serious problem with this approach is that requests from multiple users, all using the same

Connection

, operate within the same transaction. If one request leads to a rollback, all other database operations using the same Connection are also rolled back.

A connection is expensive to keep open in terms of server resources such as memory. Many commercial

database products also use licenses that are priced based on the number of simultaneously open connections, so a connection can also be expensive in terms of real money. Therefore, it's wise to try to minimize the number of connections the application needs. An alternative to the "one Connection per resource" approach is to create a Connection for each user when the first request is received and keep it as a session scope object. However, a drawback with this approach is that the Connection will be inactive most of the time, since the user needs time to look at the result of one request before making the next.

The best alternative is to use a connection pool. A connection pool contains a number of Connection objects shared by all servlets and JSP pages. For each request, one Connection is checked out, used, and checked back in. Using a pool solves the problems described for the other alternatives:

It's time-consuming to create a Connection.

A pooled Connection is created only once and then reused. Most pool implementations let you specify an initial number of Connection objects to create at startup, as well as a max number. New Connection objects are created as needed up to the max number. Once the max number has been reached, the pool clients wait until an existing Connection object becomes available instead of creating a new one.

There are multithreading problems with a shared Connection.

Each request gets its own Connection, so it's used by only one thread at a time, eliminating any potential multithreading issues.

A Connection is a limited resource.

With a pool, each Connection is used efficiently. It never sits idle if there are requests pending. If the pool allows you to specify a max number of Connection objects, you can also balance a license limit for the number of simultaneous connections against acceptable response times.

A connection pool, however, doesn't solve all problems. Since all users are using the same

Connection

objects, you cannot rely on the database engine to limit access to protected data on a peruser basis. Instead, you have to define data access rules in terms of roles (groups of users with the same access rights). You can then use separate pools for different roles, each pool creating Connection objects with a user account that represents the role.

17.1.1 Using a JDBC 2.0 Optional Package Connection Pool

Connection pools exist in many forms. You can find them in books, articles, and on the Web. Yet prior to JDBC 2.0, there was no standard defined for how a Java application would interact with a connection pool. The JDBC 2.0 Optional Package (formerly known as a Standard Extension) changes this by introducing a set of interfaces that connection pools should implement: javax.sql.DataSource

A DataSource represents a database. This is the interface the application always uses to get a

Connection

. The class that implements the interface can provide connection pooling capabilities or hand out regular, unpooled Connection objects; the application code is identical for both cases, as described later.

javax.sql.ConnectionPoolDataSource

A DataSource implementation that provides pooling capabilities uses a class that implements the ConnectionPoolDataSource interface. A ConnectionPoolDataSource is a factory for PooledConnection objects.

javax.sql.PooledConnection

The objects that a DataSource keeps in its pool implement the PooledConnection interface. When the application asks the DataSource for a Connection, it locates an available PooledConnection object, or gets a new one from its

ConnectionPoolDataSource

if the pool is empty.

The PooledConnection provides a getConnection( ) method that returns a

Connection

object. The DataSource calls this method and returns the Connection to the application. This Connection object behaves like a regular Connection with one exception: when the application calls the close( ) method, instead of closing the connection to the database, it informs the PooledConnection it belongs to that it's no longer being used. The

PooledConnection

relays this information to the DataSource, which returns the

PooledConnection

to the pool.

Figure 17.1

outlines how an application uses implementations of these interfaces to obtain a pooled connection and how to return that connection to the pool.

Figure 17.1. Application using a JDBC 2.0 connection pool

The application calls the DataSource getConnection( ) method. The DataSource looks for an available PooledConnection object in its pool. If it doesn't find one, it uses its

ConnectionPoolDataSource

object to create a new one. It then calls the getConnection( ) method on the PooledConnection object and returns the Connection object associated with the

PooledConnection

. The application uses the Connection, and calls its close( ) method when it's done. This results in a notification event being sent to the DataSource, which puts the corresponding PooledConnection object back in the pool. If you would like to learn more about the

JDBC 2.0 connection pool model, you can download the JDBC 2.0 Optional Package specification from http://java.sun.com/products/jdbc/ .

The real beauty of these interfaces is that the application doesn't have to be aware that it's using a connection pool. All configuration data, such as which JDBC driver and JDBC URL to use, the initial and maximum numbers of pooled connections, and the database account name and password, can be set by a server administrator. The completely configured DataSource object is registered as a JNDI resource, and the application can obtain a reference to it with the following code:

Context ctx = new InitialContext( );

DataSource ds = (DataSource) ctx.lookup("jdbc/EmployeeDB");

It then gets a Connection, uses it, and returns it with the following code:

Connection conn = ds.getConnection( );

// Uses the Connection conn.close( ); // Returns the Connection to the pool

By implementing these JDBC 2.0 interfaces, JDBC driver and middleware vendors can offer portable connection pooling implementations. Sun's JDBC driver list contains roughly ten different companies that claim to either offer implementations of connection pools today or have announced products to be delivered during 2000.

17.1.2 Making a JDBC 1.0 Connection Pool Behave as a JDBC 2.0 Connection Pool

If you can't find a JDBC 2.0 connection pool implementation for your database, there are plenty of implementations based on JDBC 1.0 available. I describe one in an article I wrote for the Web

Developer's Journal, titled "Improved Performance With a Connection Pool," available at http://www.webdevelopersjournal.com/columns/connection_pool.html

. Another is the

DBConnectionBroker

, available at http://www.javaexchange.com/ . It's easy to develop a couple of wrapper classes for one of these implementations so that it can be used in place of a JDBC 2.0 connection pool implementation. This way, you can switch out the JDBC 1.0 pool with a JDBC 2.0 pool when one becomes available from your database vendor or a third party.

The interaction between the wrapper classes and a connection pool implementation is illustrated in

Figure 17.2

.

Figure 17.2. A JDBC 1.0 connection pool wrapped with JDBC 2.0 interface classes

Figure 17.2

can be explained like this: the application calls the DataSourceWrapper getConnection( )

method. The DataSourceWrapper obtains a Connection object from its

ConnectionPool

object. The ConnectionPool either finds an available Connection in its pool or creates a new one. The DataSourceWrapper creates a new ConnectionWrapper object for the Connection it obtained, and returns the ConnectionWrapper to the application. The application uses the ConnectionWrapper object as a regular Connection. The

ConnectionWrapper

relays all calls to the corresponding method in the Connection it wraps, except for the close( ) method. When the application calls the close( ) method, the

ConnectionWrapper

returns its Connection to the DataSourceWrapper, which in turn returns it to its ConnectionPool.

In this example, I show you how to wrap the connection pool described in Jason Hunter and William

Crawford's Java Servlet Programming (O'Reilly). It's a simple implementation, intended only to illustrate the principles of connection pooling. The source code for the connection pool is included with

the code for this book, but I will not discuss the implementation of the pool itself, only how to make it look like a JDBC 2.0 connection pool. For production use, I recommend that instead of this code, you use a pool intended for real use, such as one of the implementations mentioned earlier. The first wrapper

class is called com.ora.jsp.sql.ConnectionWrapper , shown in Example 17.1

.

Example 17.1. The ConnectionWrapper Class

package com.ora.jsp.sql; import java.sql.*; import java.util.*; class ConnectionWrapper implements Connection {

private Connection realConn;

private DataSourceWrapper dsw;

private boolean isClosed = false;

public ConnectionWrapper(Connection realConn,

DataSourceWrapper dsw) {

this.realConn = realConn;

this.dsw = dsw;

}

/**

* Inform the DataSourceWrapper that the ConnectionWrapper

* is closed.

*/

public void close( ) throws SQLException {

isClosed = true;

dsw.returnConnection(realConn);

}

/**

* Returns true if the ConnectionWrapper is closed, false

* otherwise.

*/

public boolean isClosed( ) throws SQLException {

return isClosed;

}

/*

* Wrapped methods.

*/

public void clearWarnings( ) throws SQLException {

if (isClosed) {

throw new SQLException("Pooled connection is closed");

}

realConn.clearWarnings( );

}

...

}

An instance of this class is associated with a real Connection object, retrieved from a connection pool, through the constructor. The constructor also provides a reference to the DataSourceWrapper instance that creates it, described next.

The ConnectionWrapper class implements the Connection interface. The implementations of all the methods except two simply relay the call to the real Connection object so it can perform the requested database operation. The implementation of the close( ) method, however, doesn't call the real Connect object's method. Instead, it calls the DataSourceWrapper object's return-

Connection( )

method, to return the Connection to the pool. The isClosed( ) method, finally, returns the state of the ConnectionWrapper object as opposed to the real Connection object.

Example 17.2

shows how the com.ora.jsp.sql.DataSourceWrapper gets a connection from a

pool, and returns it when the pool client is done with it.

Example 17.2. The DataSourceWrapper Class

package com.ora.jsp.sql; import java.io.*; import java.sql.*; import javax.sql.*; public class DataSourceWrapper implements DataSource {

private ConnectionPool pool;

public DataSourceWrapper(String driverClass, String url,

String user, String pw)

throws ClassNotFoundException, InstantiationException,

SQLException, IllegalAccessException {

pool = new ConnectionPool(url, user, pw, driverClass, 1, 1);

}

/**

* Gets a connection from the pool and returns it wrapped in

* a ConnectionWrapper.

*/

public Connection getConnection( ) throws SQLException {

return new ConnectionWrapper(pool.getConnection( ), this);

}

/**

* Returns a Connection to the pool. This method is called by

* the ConnectionWrapper's close( ) method.

*/

public void returnConnection(Connection conn) {

pool.returnConnection(conn);

}

/**

* Always throws a SQLException. Username and password are set

* in the constructor and can not be changed.

*/

public Connection getConnection(String username, String password)

throws SQLException {

throw new SQLException("Not supported");

}

public int getLoginTimeout( ) throws SQLException {

throw new SQLException("Not supported");

}

public PrintWriter getLogWriter( ) throws SQLException {

throw new SQLException("Not supported");

}

public void setLoginTimeout(int seconds) throws SQLException {

throw new SQLException("Not supported");

}

public synchronized void setLogWriter(PrintWriter out)

throws SQLException {

throw new SQLException("Not supported");

}

}

The DataSourceWrapper class implements the DataSource interface, so that it can be used in place of a pure JDBC 2.0 connection pool implementation. The constructor creates an instance of the

real connection pool class, using the provided JDBC driver, URL, user and password information.

Besides the constructor, the two most interesting methods are getConnection( ) and returnConnection( )

.

The pool client application calls the getConnection( ) method, and the DataSourceWrapper relays the call to the connection pool class. It then wraps the Connection object it receives in a

ConnectionWrapper

object and returns it to the client application.

As described earlier, the ConnectionWrapper object calls the return-Connection( ) method when the pool client calls close( ) on the ConnectionWrapper object. The returnConnection( )

method hands over the Connection to the real connection pool so it can be returned to the pool.

All other DataSource interface methods are implemented to throw an SQLException. If you use the wrapper classes presented here to wrap a more sophisticated connection pool, you may be able to relay some of these method calls to the real connection pool instead.

17.1.3 Making a Connection Pool Available to Application Components

Through a DataSource object, the servlets and JSP pages in an application can get the Connection they need to access a database. What's missing is how they get access to the DataSource . I touched on this in

Chapter 14 , but let's recap and add a few details.

The place for resources that all components in an application need access to is the application scope, corresponding to ServletContext attributes in the servlet world. The current versions of the servlet and JSP specifications, 2.2 and 1.1 respectively, do not provide a specific mechanism for automatic creation and release of application scope objects when the application starts and stops (but this is being discussed as a feature for future versions of the specifications). A regular servlet can, however, fill this need nicely.

[1]

[1]

Theoretically, a web container is allowed to unload a servlet at any time, for instance to preserve memory. This could cause the shared resources to be removed while other parts of the application are still active and need access to them. In practice, though, none of the major web containers unloads a servlet before the application as such is shut down.

As described in

Chapter 14 , the container can be configured to load and initialize a servlet when the

application starts. Such a servlet can create objects and make them available to other application components in its init( ) method before any user requests are received. The servlet is also informed when the application is shut down by a call to its destroy( ) method, allowing it to release all shared objects. Finally, a servlet can read configuration data, defined as servlet initialization parameters, so that it can work in different settings. In this section, we look at how all of this can be used to make a

DataSource

object available to all components of an application.

The servlet used to manage the shared DataSource can be defined like this in the application's WEB-

INF/web.xml file:

<web-app>

<servlet>

<servlet-name>appInit</servlet-name>

<servlet-class>com.mycompany.AppInitServlet</servlet-class>

<init-param>

<param-name>jdbcDriverClassName</param-name>

<param-value>sun.jdbc.odbc.JdbcOdbcDriver</param-value>

</init-param>

<init-param>

<param-name>jdbcURL</param-name>

<param-value>jdbc:odbc:example</param-value>

</init-param>

<init-param>

<param-name>dbUserName</param-name>

<param-value>foo</param-value>

</init-param>

<init-param>

<param-name>dbUserPassword</param-name>

<param-value>bar</param-value>

</init-param>

<load-on-startup>1</load-on-startup>

</servlet>

...

The servlet class, defined by the <servlet-class> element, is given a name through the

<servlet-name>

element. A number of <init-param> elements, with nested <param-name> and <param-value> elements, are used to define the following initialization parameters: jdbcDriverClassName

, jdbcURL, dbUserName, and dbUserPassword. If you use a JDBC

2.0 connection pool, you need to define the URL used to get a reference from JNDI to it instead of all these parameters. The last servlet element, <load-on-startup> , tells the container that this servlet should be initialized when the web application is started. The container initializes servlets in the relative order specified by this element, from the lowest number to the highest. If two servlets have the same value, their relative start order is undefined.

The servlet reads all the initialization parameters in its init( ) method, creates a

DataSourceWrapper

instance, and sets it as a ServletContext attribute named exampleDS: public void init( ) throws ServletException {

ServletConfig config = getServletConfig( );

String jdbcDriverClassName =

config.getInitParameter("jdbcDriverClassName");

String jdbcURL = config.getInitParameter("jdbcURL");

String dbUserName = config.getInitParameter("dbUserName");

String dbUserPassword = config.getInitParameter("dbUserPassword");

// Make sure a driver class and JDBC URL is specified

if (jdbcDriverClassName == null || jdbcURL == null) {

throw new UnavailableException("Init params missing");

}

DataSource ds = null;

try {

ds = new DataSourceWrapper(jdbcDriverClassName, jdbcURL,

dbUserName, dbUserPassword);

}

catch (Exception e) {

throw new UnavailableException("Cannot create connection pool"

+ ": " + e.getMessage( ));

}

getServletContext( ).setAttribute("exampleDS", ds);

}

All servlets and JSP pages in the application can now obtain a reference to the DataSource. Servlets use the ServletContext getAttribute( ) method to accomplish this. For JSP pages, the

DataSource

appears as an application scope object. All the database custom actions introduced in

Chapter 9 look for a DataSource in the application scope, so all you have to do to use the one created

by the initialization servlet is to provide the name:

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<ora:sqlQuery id="empList" dataSource="exampleDS" scope="request">

SELECT * FROM Employee

WHERE FirstName LIKE ?

AND LastName LIKE ?

AND Dept LIKE ?

ORDER BY LastName

<ora:sqlStringValue param="firstName" prefix="%" suffix="%" />

<ora:sqlStringValue param="lastName" prefix="%" suffix="%" />

<ora:sqlStringValue param="dept" prefix="%" suffix="%" />

</ora:sqlQuery>

<jsp:forward page="list.jsp" />

Note how the dataSource attribute value matches the name of the ServletContext attribute holding the reference to the DataSource, set by the initialization servlet.

It's much better to let an initialization servlet create the DataSource, as described here, than to use the

<ora:useDataSource>

custom action described in Chapter 9 . With a servlet, all information about

the JDBC driver class, URL, user and password is in one place (the WEB-INF/web.xml file), as opposed to being repeated in every JSP page that uses the database custom actions. This makes it easier to change the information when needed. Also, if you decide at some point to use another connection pool implementation, such as a true JDBC 2.0 connection pool available from your JDBC driver or database vendor, you can easily change the servlet's init( ) method. So even for a pure JSP application, I recommend that you use an application initialization servlet like the one described here.

The initialization servlet should also clean up when the application is shut down. The web container calls the destroy( ) method: public void destroy( ) {

getServletContext( ).removeAttribute("exampleDS");

}

Most connection pools used in production provide a method that should be called at shutdown to let it close all connections. If you use such a pool, you need to call this method in the servlet's destroy( ) method as well. The example pool used here doesn't provide a shutdown method.

© 2001, O'Reilly & Associates, Inc.

17.2 Using a Generic Database Bean

All the database custom action tag handler classes described later in this chapter are based on a generic database bean named com.ora.jsp.sql.SQLCommandBean . This bean uses a number of other classes.

Figure

17.3

shows the relationship between all these classes.

Figure 17.3. The SQLCommandBean and related classes

The SQLCommandBean takes care of setting all values in a JDBC java.sql.PreparedStatement and executing the statement. For SELECT statements, it also processes the result by creating com.ora.jsp.sql.Row

objects containing a com.ora.jsp.sql.Column object for each column in the result. The rows returned by the SELECT statement are returned to the caller as a java.util.Vector

with Row objects. The EmployeeRegistryBean described in Chapter 15

is one example of how to use this bean, and other examples follow in this chapter. Let's look at each class in detail, starting with the

SQLCommandBean

itself.

17.2.1 The SQLCommandBean and Value Classes

The SQLCommandBean has three write-only properties.

Example 17.3

shows the beginning of the class file

with the setter methods.

Example 17.3. SQLCommandBean Property Setter Methods

package com.ora.jsp.sql; import java.util.*; import java.sql.*; import com.ora.jsp.sql.value.*; public class SQLCommandBean {

private Connection conn;

private String sqlValue;

private Vector values;

private boolean isExceptionThrown = false;

public void setConnection(Connection conn) {

this.conn = conn;

}

public void setSqlValue(String sqlValue) {

this.sqlValue = sqlValue;

}

public void setValues(Vector values) {

this.values = values;

}

...

The connection property holds the Connection to use, and the sqlValue property is set to the SQL statement to execute, with question marks as placeholders for variable values, if any. The placeholders are then replaced with the values defined by the values property, a Vector with one com.ora.jsp.sql.Value object per placeholder. Before we look at the other SQLCommandBean methods, let's look at the Value class.

The Value class is an abstract class used as a superclass for classes representing specific Java types, as shown in

Figure 17.3

. It contains default implementations of methods for getting the specific type of value a subclass

represents.

Example 17.4

shows two of the methods.

Example 17.4. Two Value Class Methods

public abstract class Value {

public BigDecimal getBigDecimal( )

throws UnsupportedConversionException {

throw new UnsupportedConversionException(

"No conversion to BigDecimal");

}

public boolean getBoolean( )

throws UnsupportedConversionException {

throw new UnsupportedConversionException(

"No conversion to boolean");

}

...

The default implementation for each method simply throws a com.ora.jsp.sql.UnsupportedConversionException

. Each subclass implements the method that returns the value of the type it represents, as well as the getString( ) method. The getString( ) method returns the value converted to a String.

Example 17.5

shows the

com.ora.jsp.sql.value.IntValue

subclass.

Example 17.5. The IntValue Class

package com.ora.jsp.sql.value; import com.ora.jsp.sql.Value; public class IntValue extends Value {

private int value;

public IntValue(int value) {

this.value = value;

}

public int getInt( ) {

return value;

}

public String getString( ) {

return String.valueOf(value);

}

}

An application that uses the SQLCommandBean can create Value objects and set the bean's properties like this:

SQLCommandBean sqlBean = new SQLCommandBean( ); sqlBean.setConnection(ds.getConnection( ));

String sqlValue =

"SELECT * FROM MyTable WHERE IntCol = ? AND TextCol = ?"; sqlBean.setSqlValue(sqlValue);

Vector values = new Vector( ); values.addElement(new IntValue(10)); values.addElement(new StringValue("Hello!")); sqlBean.setValues(values);

One of two methods in the SQLCommandBean is used to execute the SQL statement: the executeQuery(

)

method for a SELECT statement, and the executeUpdate( ) method for all other types of statements.

Example 17.6

shows the executeQuery( ) method.

Example 17.6. The SQLCommandBean's executeQuery( ) Method

public Vector executeQuery( ) throws SQLException,

UnsupportedTypeException {

Vector rows = null;

ResultSet rs = null;

PreparedStatement pstmt = null;

Statement stmt = null;

try {

if (values != null && values.size( ) > 0) {

// Use a PreparedStatement and set all values

pstmt = conn.prepareStatement(sqlValue);

setValues(pstmt, values);

rs = pstmt.executeQuery( );

}

else {

// Use a regular Statement

stmt = conn.createStatement( );

rs = stmt.executeQuery(sqlValue);

}

// Save the result in a Vector of Row object

rows = toVector(rs);

}

finally {

try {

if (rs != null) {

rs.close( );

}

if (stmt != null) {

stmt.close( );

}

if (pstmt != null) {

pstmt.close( );

}

}

catch (SQLException e) {

// Ignore. Probably caused by a previous

// SQLException thrown by the outer try block

}

}

return rows;

}

If the values property is set, a JDBC PreparedStatement is needed to associate the values with the question mark placeholders in the SQL statement. A method named setValues( ) takes care of setting all values, using the appropriate JDBC method for the datatype represented by each Value object. If the values property is not set, a regular JDBC Statement is created instead. In both cases, the JDBC driver is asked to execute the statement, and the resulting ResultSet is turned into a Vector with Row objects by the toVector( )

method. The Vector is then returned to the caller.

You may wonder why the ResultSet is not returned directly instead of creating a Vector with Row objects.

The reason is that a ResultSet is tied to the Connection that was used to generate it. When the

Connection

is closed or used to execute a new SQL statement, all open ResultSet objects for the

Connection

are released. You must therefore make sure to save the information from the ResultSet in a new data structure before reusing the Connection or returning it to the pool.

The code for creating the PreparedStatement or Statement object and executing the statement is enclosed in a try/finally block. This is important, because if something fails (due to an invalid SQL statement, for instance), the JDBC methods throw an SQLException . You want the exception to be handled by the application using the SQLCommandBean, but first you must make sure that all JDBC resources are released and the Connection object is returned to the pool. Using a try block with a finally clause but no catch clause gives this behavior. If an exception is thrown, the finally clause is executed, and then the exception is automatically thrown to the object that called the executeQuery( ) method. In the finally clause, the ResultSet object and either the PreparedStatement or Statement object are closed. It should be enough to close the statement object according to the JDBC specification (closing the statement should also close the ResultSet associated with the statement), but doing it explicitly doesn't hurt and makes the code work even with a buggy JDBC driver.

Example 17.7

shows a part of the setValues( ) method.

Example 17.7. The SQLCommandBean's setValues( ) Method

private void setValues(PreparedStatement pstmt, Vector values)

throws SQLException {

for (int i = 0; i < values.size( ); i++) {

try {

Value v = (Value) values.elementAt(i);

// Set the value using the method corresponding to

// the type.

// Note! Set methods are indexed from 1, so we add

// 1 to i

if (v instanceof BigDecimalValue) {

pstmt.setBigDecimal(i + 1, v.getBigDecimal( ));

}

else if (v instanceof BooleanValue) {

pstmt.setBoolean(i + 1, v.getBoolean( ));

}

...

}

catch (UnsupportedConversionException e) {

// Can not happen here since we test the type first

}

}

}

The setValue( ) method loops through all elements in the Vector with values. For each element, it tests which Value subclass it is and uses the corresponding JDBC method to set the value for the

PreparedStatement

object. You may wonder why a PreparedStatement is used here, since it's used only once. It's true that a PreparedStatement is intended to be reused over and over again to execute the same SQL statement with new values. But it offers a convenient solution to the problem of different syntax for values of type date/time and numbers when represented by a string literal. When a PreparedStatement is used, the variable values in the SQL statement can be represented by Java variables of the appropriate types without worrying about what literal representation a certain JDBC driver supports. So even though it's used only once, a PreparedStatement still has an advantage over a regular Statement.

The toVector( ) method is shown in Example 17.8

.

Example 17.8. The SQLCommandBean's toVector( ) Method

private Vector toVector(ResultSet rs) throws SQLException,

UnsupportedTypeException {

Vector rows = new Vector( );

while (rs.next( )) {

Row row = new Row(rs);

rows.addElement(row);

}

return rows;

}

This method simply walks through the ResultSet and adds a new Row object for each row to a Vector that it then returns. As you will see later, the Row constructor reads all column values and creates a Column object for each.

The executeUpdate( ) method, shown in

Example 17.9

, is very similar to the executeQuery( )

method.

Example 17.9. The SQLCommandBean's executeUpdate( ) Method

public int executeUpdate( ) throws SQLException, UnsupportedTypeException {

int noOfRows = 0;

ResultSet rs = null;

PreparedStatement pstmt = null;

Statement stmt = null;

try {

if (values != null && values.size( ) > 0) {

// Use a PreparedStatement and set all values

pstmt = conn.prepareStatement(sqlValue);

setValues(pstmt, values);

noOfRows = pstmt.executeUpdate( );

}

else {

// Use a regular Statement

stmt = conn.createStatement( );

noOfRows = stmt.executeUpdate(sqlValue);

}

}

finally {

try {

if (rs != null) {

rs.close( );

}

if (stmt != null) {

stmt.close( );

}

if (pstmt != null) {

pstmt.close( );

}

}

catch (SQLException e) {

// Ignore. Probably caused by a previous

// SQLException thrown by the outer try block.

}

}

return noOfRows;

}

The main difference is that the executeUpdate( ) method is used to execute SQL statements that do not return rows, only the number of rows affected by the statement. Examples of such statements are UPDATE,

INSERT

, and DELETE. In the same way as the executeQuery( ) method, a PreparedStatement is created and initialized with the values defined by the values property, if set. Otherwise a regular

Statement

is used. The statement is executed and the number of affected rows is returned to the caller.

17.2.2 The Row and Column Classes

Let's now look at the Row and Column classes.

Example 17.10

shows a part of the Row class constructor.

Example 17.10. The Row Class Constructor

package com.ora.jsp.sql; import java.util.*; import java.sql.*; import java.sql.Date; import java.math.*; import com.ora.jsp.sql.column.*; public class Row {

private Column[] columns;

public Row(ResultSet rs) throws SQLException,

UnsupportedTypeException {

ResultSetMetaData rsmd = rs.getMetaData( );

int cols = rsmd.getColumnCount( );

columns = new Column[cols];

// Note! Columns are numbered from 1 in the ResultSet

for (int i = 1; i <= cols; i++) {

int type = rsmd.getColumnType(i);

switch (type) {

case Types.DATE:

columns[i - 1] =

new DateColumn(rsmd.getColumnName(i),

rs.getDate(i));

break;

case Types.TIME:

columns[i - 1] =

new TimeColumn(rsmd.getColumnName(i),

rs.getTime(i));

break;

...

default:

throw new

UnsupportedTypeException("Unsupported SQL " +

"data type: " + type);

}

}

}

The Row class keeps all column values as an array of Column objects. The constructor is called with a

ResultSet

that has been positioned at a new row by the caller using the next( ) method. It loops through all columns in the row and creates a Column subclass instance for each. The column's datatype, retrieved from the ResultSetMetaData object, is used to decide which Column subclass to create. Similarly to the

Value

class structure, the Column class structure contains subclasses corresponding to JDBC column datatypes, as shown in

Figure 17.3

.

Two methods provide access to the number of columns and the array of Column objects, shown in Example

17.11

.

Example 17.11. The Row's getColumnCount( ) and getColumns( ) Methods

public int getColumnCount( ) {

return columns.length;

}

public Column[] getColumns( ) {

return columns;

}

Another set of methods can be used to retrieve the value of an individual column, given its name or index. This

set of methods contains one pair per supported datatype. Example 17.12

shows the methods for the

BigDecimal

, boolean, and String types.

Example 17.12. The Row's Column Value Access Methods

public BigDecimal getBigDecimal(int columnIndex)

throws NoSuchColumnException, UnsupportedConversionException {

Column col = null;

try {

col = columns[columnIndex - 1];

}

catch (ArrayIndexOutOfBoundsException e) {

throw

new NoSuchColumnException(String.valueOf(columnIndex));

}

return col.getBigDecimal( );

}

public BigDecimal getBigDecimal(String columnName)

throws NoSuchColumnException, UnsupportedConversionException {

return getBigDecimal(getIndex(columnName));

}

public boolean getBoolean(int columnIndex)

throws NoSuchColumnException, UnsupportedConversionException {

Column col = null;

try {

col = columns[columnIndex - 1];

}

catch (ArrayIndexOutOfBoundsException e) {

throw

new NoSuchColumnException(String.valueOf(columnIndex));

}

return col.getBoolean( );

}

public boolean getBoolean(String columnName)

throws NoSuchColumnException, UnsupportedConversionException {

return getBoolean(getIndex(columnName));

}

...

public String getString(int columnIndex)

throws NoSuchColumnException {

Column col = null;

try {

col = columns[columnIndex - 1];

}

catch (ArrayIndexOutOfBoundsException e) {

throw

new NoSuchColumnException(String.valueOf(columnIndex));

}

return col.getString( );

}

public String getString(String columnName)

throws NoSuchColumnException {

return getString(getIndex(columnName));

}

All these methods locate the Column subclass instance specified by the argument and call the corresponding method on the instance. Except for the getString( ) method, this call results in an

UnsupportedConversionException

if the column is not of the requested type. All types can be converted to a String, however, so a getString( ) call is successful provided that the requested column exists.

The Column class is an abstract class, very similar to the Value class shown in

Example 17.4

. It contains

access methods for all datatypes, with default implementations that throw an

UnsupportedConversionException

. Each subclass provides a real implementation of the access

method corresponding to its type, plus the getString( ) method. Example 17.13

shows the IntColumn class.

Example 17.13. The IntColumn Class

package com.ora.jsp.sql.column; import com.ora.jsp.sql.Column; public class IntColumn extends Column {

private int value;

public IntColumn(String name, int value) {

super(name);

this.value = value;

}

public int getInt( ) {

return value;

}

public String getString( ) {

return String.valueOf(value);

}

}

The constructor takes the column name and value as arguments. The name is used to initialize the Column superclass and is returned by its getName( ) method.

© 2001, O'Reilly & Associates, Inc.

17.3 Developing Generic Database Custom Actions

The database custom actions introduced in Chapter 9 can be used like this in a JSP page:

<ora:sqlTransaction dataSource="example">

<ora:sqlUpdate>

UPDATE Account SET Balance = Balance - ?

WHERE AccountNumber = ?

<ora:floatValue param="amount" />

<ora:intValue param="account" />

</ora:sqlUpdate>

<ora:sqlUpdate>

UPDATE Account SET Balance = Balance + ?

WHERE AccountNumber = ?

<ora:floatValue param="amount" />

<ora:intValue param="account" />

</ora:sqlUpdate>

</ora:sqlTransaction>

The database custom actions use all of the classes described previously in this chapter. A DataSource available in the application scope is used to get a Connection, and an SQLCommandBean is used to execute the SQL statement specified in the database action element body. The nested value actions create

Value

subclass instances and add them to a list held by the parent action tag handler. The

<ora:sqlQuery>

action saves the result as a Vector of Row objects in the scope specified by the page author. In this section, we first look at how the tag handlers for the <ora:sqlQuery>,

<ora:sqlUpdate>

, and <ora:sqlIntValue> actions are implemented. All value actions follow the same pattern as <ora:sqlIntValue>, so they are not described here. At the end of this section, we also look at the tag handler for the <ora:sqlTransatction> action to see how it provides a transaction scope for the database actions nested in its body.

17.3.1 The <ora:sqlQuery> and <ora:sqlUpdate> Actions

The <ora:sqlQuery> and <ora:sqlUpdate> actions both need access to the action body to read the

SQL statement. Hence, the corresponding tag handlers extend the BodyTagSupport class described in

Chapter 16

. They also implement an interface called com.ora.jsp.sql.ValueParent, which is used by the nested value actions to find the correct parent, following the pattern described for cooperating actions in

Chapter 16

.

These two actions share the same set of attributes and have almost the same behavior, so a common superclass called com.ora.jsp.tags.sql.DBTag implements most of the tag handler functionality for both actions.

Example 17.14

shows the top part of the DBTag class, with the class declaration and all property setter methods.

Example 17.14. The DBTag Declaration and Properties

package com.ora.jsp.tags.sql; import java.util.*; import java.sql.*; import javax.sql.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import com.ora.jsp.sql.*; import com.ora.jsp.sql.value.*; public abstract class DBTag extends BodyTagSupport

implements ValueTagParent {

private SQLCommandBean sqlCommandBean = new SQLCommandBean( );

private String dataSourceName;

private String id;

private int scope = PageContext.PAGE_SCOPE;

private String sqlValue;

private Vector values;

private boolean isExceptionThrown = false;

private boolean isPartOfTransaction = false;

public void setDataSource(String dataSourceName) {

this.dataSourceName = dataSourceName;

}

public void setId(String id) {

this.id = id;

}

public void setScope(String scopeName) {

if ("page".equals(scopeName)) {

scope = PageContext.PAGE_SCOPE;

}

else if ("request".equals(scopeName)) {

scope = PageContext.REQUEST_SCOPE;

}

else if ("session".equals(scopeName)) {

scope = PageContext.SESSION_SCOPE;

}

else if ("application".equals(scopeName)) {

scope = PageContext.APPLICATION_SCOPE;

}

}

An instance of the SQLCommandBean is kept as a private instance variable. The bean is used to perform all database operations; the tag handler just provides an easy-to-use interface to the bean for page authors.

The setter methods for the dataSource and id properties set the corresponding instance variables. The setter method for the scope property converts the String value to the corresponding scope int value defined by the PageContext class before saving the value, since the int value is later needed in the doEndTag( )

method.

The page author specifies the SQL statement to execute in the action element's body. Because the tag handler implements the BodyTag interface (by extending the BodyTagSupport class), it can read the

SQL statement in the doAfterBody( ) method as shown in

Example 17.15

.

Example 17.15. The DBTag's doAfterBody( ) Method

public int doAfterBody( ) throws JspException {

sqlValue = bodyContent.getString( );

return SKIP_BODY;

}

The SQL statement may contain question marks as placeholders for values set by nested value actions. As you will see later, the value actions create the appropriate Value subclass and call the DBTag 's addValue( )

method, shown in

Example 17.16

.

Example 17.16. The DBTag's addValue( ) Method

public void addValue(Value value) {

if (values == null) {

values = new Vector( );

}

values.addElement(value);

}

This method creates a Vector to hold all values the first time it's called, and then adds the Value object to the Vector. When called by subsequent value action tag handlers, the Value objects are simply added to the list.

The real processing happens in the doEndTag( ) method, shown in

Example 17.17

. This method is

called by the container when the action element's body has been processed and the end tag is encountered.

Example 17.17. The DBTag's doEndTag( ) Method

public int doEndTag( ) throws JspException {

Connection conn = getConnection( );

sqlCommandBean.setConnection(conn);

sqlCommandBean.setSqlValue(sqlValue);

sqlCommandBean.setValues(values);

Object result = null;

try {

result = execute(sqlCommandBean);

}

catch (SQLException e) {

...

}

catch (UnsupportedTypeException e) {

...

}

finally {

...

}

// Save the result with the specified id in the specified scope

if (id != null) {

pageContext.setAttribute(id, result, scope);

}

return EVAL_PAGE;

}

The private getConnection( ) method is used to get a Connection. The Connection is retrieved either from the DataSource specified by the dataSource attribute value for the <ora:sqlQuery> or the <ora:sqlUpdate> action itself, or from an enclosing <ora:sqlTransaction> element.

We'll return to getConnection( ) and the exception handling code later when we look at transaction support. The Connection, the SQL statement, and all values (if any) are passed to the bean. Then, the abstract execute( ) method is called to ask the bean to execute the SQL statement. The Object returned by execute( ) is saved in the scope specified by the scope attribute, using the name specified by the id attribute.

The implementation of the execute( ) method is the only thing that differs between the tag handlers for the <ora:sqlQuery> action and the <ora:sqlUpdate> action. The corresponding tag handler classes, QueryTag and UpdateTag, extend the DBTag class and implement the execute( ) method.

Example 17.18

shows the implementation in the QueryTag tag handler.

Example 17.18. The QueryTag's execute( ) Method

public Object execute(SQLCommandBean sqlCommandBean)

throws SQLException, UnsupportedTypeException {

return sqlCommandBean.executeQuery( );

}

This method simply calls the bean's executeQuery( ) method.

The UpdateTag tag handler uses the bean's executeUpdate( ) method instead, and wraps the

returned int in an Integer object, as shown in Example 17.19

.

Example 17.19. The UpdateTag's execute( ) Method

public Object execute(SQLCommandBean sqlCommandBean)

throws SQLException, UnsupportedTypeException {

return new Integer(sqlCommandBean.executeUpdate( ));

}

The reason for wrapping the int in an Integer is that only real objects can be saved as JSP scope objects; primitive types are not supported.

The <ora:sqlQuery> and the <sql:Update> actions define separate TagExtraInfo classes, named QueryTagExtraInfo and UpdateTagExtraInfo, respectively. They both extend a class called com.ora.jsp.tags.sql.DBTagExtraInfo. The isValid( ) method for the

DBTagExtraInfo

class makes sure a valid value is specified for the scope attribute. It's shown in

Example 17.20

.

Example 17.20. The DBTagExtraInfo Class

package com.ora.jsp.tags.sql; import javax.servlet.jsp.tagext.*; public class DBTagExtraInfo extends TagExtraInfo {

/**

* Returns true only if a valid scope value is specified:

* page, request, session or application.

*/

public boolean isValid(TagData data) {

boolean isValid = false;

String scope = data.getAttributeString("scope");

if (scope == null || scope.equals("page") ||

scope.equals("request") ||

scope.equals("session") || scope.equals("application")) {

isValid = true;

}

return isValid;

}

}

The QueryTagExtraInfo and UpdateTagExtraInfo classes implement the getVariableInfo( )

method to tell the container about the result variables they create. Here's the

UpdateTagExtraInfo

class: package com.ora.jsp.tags.sql; import javax.servlet.jsp.tagext.*; public class UpdateTagExtraInfo extends DBTagExtraInfo {

public VariableInfo[] getVariableInfo(TagData data) {

if (data.getAttributeString("id") == null) {

return new VariableInfo[0];

}

else {

return new VariableInfo[]

{

new VariableInfo(data.getAttributeString("id"),

"java.lang.Integer",

true,

VariableInfo.AT_END)

};

}

}

}

The QueryTagExtraInfo class is almost identical. The only difference is that it sets the class name to com.ora.jsp.sql.Row

instead of java.lang.Integer.

17.3.2 The <ora:intValue> Action

A set of value actions can be used inside the body of an <ora:sqlQuery> or <ora:sqlUpdate> action to set the values for placeholders in the SQL statement. The example tag library contains value actions for date/time values as well as for numeric and string values. The tag handlers for all these actions have a great deal in common, so they all extend a common superclass called com.ora.jsp.tags.sql.value.ValueTag

. This superclass contains instance variables and

property setter methods for the attributes shared by all value actions: stringValue, pattern, param, name

, and property. In addition, it contains methods used by the subclasses to get the value when it's specified as a parameter name or a bean name plus a property name.

Let's look at the tag handler class for the <ora:intValue> action as an example. All other value actions follow the same pattern. The com.ora.jsp.tags.sql.IntValueTag class extends the ValueTag class as shown in

Example 17.21

.

Example 17.21. The IntValueTag Class

package com.ora.jsp.tags.sql.value; import java.lang.reflect.*; import java.text.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import com.ora.jsp.sql.value.*; import com.ora.jsp.tags.sql.ValueTagParent; import com.ora.jsp.util.*; public class IntValueTag extends ValueTag {

private int value;

public void setValue(int value) {

this.value = value;

}

public int doEndTag( ) throws JspException {

if (stringValue != null) {

value = toInt(stringValue, pattern);

}

else if (param != null) {

String paramValue = getParameter(param);

value = toInt(paramValue, pattern);

}

else if (name != null) {

value = getInt(name, property, pattern);

}

ValueTagParent parent =

(ValueTagParent) findAncestorWithClass(this,

ValueTagParent.class);

if (parent == null) {

throw new JspException("The sqlIntValue action is not " +

"enclosed by a supported action type");

}

parent.addValue(new IntValue(value));

return EVAL_PAGE;

}

...

}

The Java datatype is different for each value action, so all subclasses implement their own value property setter methods to store the value in an instance variable. The value property for the IntValueTag is of course an int.

Besides the value attribute, all value actions support three other ways to set the value: as a String value specified by the stringValue attribute, as a request parameter specified by the param attribute, or as a bean property specified by the name and property attributes. The ValueTag superclass contains the setter methods and instance variables for these attributes, and the ValueTagExtraInfo class makes sure that a page author uses only one of these alternatives to specify the value. The doEndTag( ) method finds out which alternative is used by looking at all property instance variables in turn. If it's a

String

value, defined by the stringValue or param attributes, the method converts the value to an int

using a private toInt( ) method. If it's specified as a bean property, another private method, getInt( )

, is used to invoke the bean to get the value. These private methods are not shown here, but you can look at the source code if you want to see how they work.

The enclosing <ora:sqlQuery> or <ora:sqlUpdate> action's tag handler is then located, as

described in Chapter 16 , using the findAncestorWithClass( ) method. The tag handlers for the

enclosing actions implement the ValueTagParent interface used as the parent class argument. If an enclosing action is found, its addValue( ) method is called to add the int value wrapped in an

IntValue

object to the parent's value list.

17.3.3 The <ora:sqlTransaction> Action

The final database action in the example tag library is the <ora:sqlTransaction> action. A database transaction consists of the execution of a number of SQL statements on the same Connection; they either all succeed or all fail. As you may recall, the data changes resulting from executing all statements are then either permanently saved by committing the transaction, or ignored by rolling back the transaction.

The task for the <ora:sqlTransaction> action is to provide a Connection to all database actions enclosed in its body, and commit the transaction when they have all been executed. The nested database actions must cooperate with the <ora:sqlTransaction> action tag handler by retrieving the shared

Connection

and rolling back the transaction if they fail.

Let's look at how the Connection is handled first.

Example 17.22

shows the TransactionTag class declaration, the dataSource property setter method, and the doStartTag( ) method.

Example 17.22. Part of the TransactionTag Class

package com.ora.jsp.tags.sql; import java.io.*; import java.util.*; import java.sql.*; import javax.sql.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; public class TransactionTag extends TagSupport {

private String dataSourceName;

private Connection conn;

public void setDataSource(String dataSourceName) {

this.dataSourceName = dataSourceName;

}

public int doStartTag( ) throws JspException {

conn = getTransactionConnection( );

return EVAL_BODY_INCLUDE;

}

The TransactionTag class extends the TagSupport class. It doesn't need to access its body content; it only needs to tell the JSP container to execute all actions and possible scripting elements in the body, so the TagSupport class is the proper choice. The dataSource property corresponds to the attribute with the same name. The page author sets it to the name of the DataSource object in the application scope to use for all nested database actions.

The doStartTag( ) method, invoked by the container before the actions in the body are processed, sets the conn instance variable by calling the get-TransactionConnection( ) method, shown in

Example 17.23

.

Example 17.23. The TransactionTag's getTransactionConnection( ) and getConnection( ) Methods

private Connection getTransactionConnection( ) throws JspException {

DataSource dataSource = (DataSource)

pageContext.getAttribute(dataSourceName,

PageContext.APPLICATION_SCOPE);

if (dataSource == null) {

throw new JspException("dataSource " + dataSourceName +

" not found");

}

try {

conn = dataSource.getConnection( );

conn.setAutoCommit(false);

}

catch (SQLException e) {

throw new JspException("SQL error: " + e.getMessage( ));

}

return conn;

}

Connection getConnection( ) {

return conn;

}

This method retrieves the DataSource object from the application scope and gets a Connection. A

Connection

automatically commits each SQL statement by default. In order to use the Connection to execute more than one statement within the same transaction, the setAutoCommit( ) method is called with the value false.

The package scope getConnection( ) method, also shown in Example 17.23

, is used by the tag handlers for the <ora:sqlQuery> and <ora:sqlUpdate> actions. To see how it's used, let's look at the method with the same name in the DBTag class that we skipped earlier. The DBTag 's getConnection( )

method is shown in Example 17.24

.

Example 17.24. The DBTag's getConnection( ) Method

private Connection getConnection( ) throws JspException {

Connection conn = null;

TransactionTag transactionTag = (TransactionTag)

findAncestorWithClass(this, TransactionTag.class);

if (transactionTag != null) {

conn = transactionTag.getConnection( );

isPartOfTransaction = true;

if (dataSourceName != null) {

throw new JspException("A dataSource must not be " +

"specified when the action is part of a " +

"transaction");

}

}

else {

DataSource dataSource = (DataSource)

pageContext.getAttribute(dataSourceName,

PageContext.APPLICATION_SCOPE);

if (dataSource == null) {

throw new JspException("dataSource " +

dataSourceName + " not found");

}

try {

conn = dataSource.getConnection( );

}

catch (SQLException e) {

throw new JspException("SQL error: " + e.getMessage( ));

}

}

return conn;

}

The findAncestorWithClass( ) is used to find out whether or not the action is nested in the body of an <ora:sqlTransaction> tag. If a TransactionTag is found, the action is part of a transaction, so the TransactionTag's get-Connection( ) is used to retrieve the shared

Connection

, and a boolean flag is set to remember that the action is part of a transaction. If a parent tag handler is not found, the action is not part of a transaction. In this case, the DBTag's dataSource property value is used instead to locate the DataSource in the application scope and get a

Connection

.

The doEndTag( ) method in the DBTag class contains some details related to database transactions that

we also skipped earlier. Let's revisit this method, shown in Example 17.25

.

Example 17.25. The DBTag's doEndTag( ) Method

public int doEndTag( ) throws JspException {

Connection conn = getConnection( );

sqlCommandBean.setConnection(conn);

sqlCommandBean.setSqlValue(sqlValue);

sqlCommandBean.setValues(values);

Object result = null;

try {

result = execute(sqlCommandBean);

}

catch (SQLException e) {

try {

isExceptionThrown = true;

conn.rollback( );

}

catch (SQLException se) {

// Ignore, probably a result of the main exception

}

throw new JspException("SQL error: " + e.getMessage( ));

}

catch (UnsupportedTypeException e) {

try {

isExceptionThrown = true;

conn.rollback( );

}

catch (SQLException se) {

// Ignore, probably caused by the main exception

}

throw new JspException("Query result error: " +

e.getMessage( ));

}

finally {

try {

if (isPartOfTransaction && isExceptionThrown) {

// Reset auto commit in case the connection is

// pooled before it's returned to the pool by close

conn.setAutoCommit(true);

conn.close( );

}

else if (!isPartOfTransaction) {

// If we're not part of a transaction, the

// connection is in auto commit mode so we only

// close it

conn.close( );

}

}

catch (SQLException e) {

e.printStackTrace(System.err);

}

}

// Save the result with the specified id in the specified scope

if (id != null) {

pageContext.setAttribute(id, result, scope);

}

return EVAL_PAGE;

}

The interesting code is in the catch and finally clauses of the try block. If the execution of the SQL statement causes an exception to be thrown, the transaction is rolled back and a JspException is thrown. This aborts the processing of the rest of the page and informs the user about the error. A boolean flag is also set to be able to handle this case in the finally clause. The finally clause is executed whether or not an exception is thrown. If this action is part of a transaction and an exception is thrown by

the execute( ) method, the Connection is returned to the pool by calling the close( ) method after auto commit is turned on again to reset it to its default state. If the action is not part of a transaction, there's no need to reset the auto commit since it has never been changed; the Connection is just returned to the pool by calling the close( ) method. If the action is part of a transaction and no exception is thrown, the result is saved in the specified scope, and processing continues with the next nested database action. Note that the Connection is not closed in this case, as the same Connection must be used for all SQL statements in the transaction.

If all actions execute successfully, the TransactionTag 's doEndTag( ), shown in

Example 17.26

, is

invoked.

Example 17.26. The TransactionTag's doEndTag( ) Method

public int doEndTag( ) throws JspException {

try {

conn.commit( );

conn.setAutoCommit(true);

conn.close( );

}

catch (SQLException e) {

throw new JspException("SQL error: " + e.getMessage( ));

}

return EVAL_PAGE;

}

The doEndTag( ) method commits the transaction, resets the auto commit for the Connection, and returns the Connection to the pool by calling the close( ) method.

© 2001, O'Reilly & Associates, Inc.

17.4 Developing Application-Specific Database Components

The classes described in this chapter can also be used for application-specific components that access a

database. Chapter 15

includes one example of an application-specific bean, the

EmployeeRegisterBean

, that uses the SQLCommandBean to execute its SQL statements.

You can also use these classes in your application-specific custom actions. One example is the custom action that's mentioned in

Chapter 9

as an alternative to the generic database actions for inserting or updating employee information:

<%@ page language="java" contentType="text/html" %>

<%@ taglib uri="/orataglib" prefix="ora" %>

<%@ taglib uri="/mytaglib" prefix="myLib" %>

<myLib:saveEmployeeInfo dataSource="example" />

<%-- Get the new or updated data from the database --%>

<ora:sqlQuery id="newEmpDbInfo" dataSource="example" scope="session">

SELECT * FROM Employee

WHERE UserName = ?

<ora:sqlStringValue param="userName" />

</ora:sqlQuery>

<%-- Redirect to the confirmation page --%>

<ora:redirect page="confirmation.jsp" />

Example 17.27

shows one way to implement this custom action.

Example 17.27. SaveEmployeeInfoTag Class

package com.mycompany.tags; import java.sql.*; import java.text.*; import java.util.Vector; import javax.sql.*; import javax.servlet.*; import javax.servlet.jsp.*; import javax.servlet.jsp.tagext.*; import com.ora.jsp.sql.*; import com.ora.jsp.sql.value.*;

import com.ora.jsp.util.*; public class SaveEmployeeInfoTag extends TagSupport {

private String dataSourceName;

public void setDataSource(String dataSourceName) {

this.dataSourceName = dataSourceName;

}

public int doEndTag( ) throws JspException {

// Get all request parameters

ServletRequest request = pageContext.getRequest( );

String userName = request.getParameter("userName");

String password = request.getParameter("password");

String firstName = request.getParameter("firstName");

String lastName = request.getParameter("lastName");

String dept = request.getParameter("dept");

String empDate = request.getParameter("empDate");

String emailAddr = request.getParameter("emailAddr");

if (userName == null || password == null ||

firstName == null || lastName == null ||

dept == null || empDate == null ||

emailAddr == null) {

throw new JspException("Missing a mandatory parameter");

}

SQLCommandBean sqlCommandBean = new SQLCommandBean( );

DataSource dataSource = (DataSource)

pageContext.getAttribute(dataSourceName,

PageContext.APPLICATION_SCOPE);

if (dataSource == null) {

throw new JspException("The data source " + dataSource +

" is not found in the application scope");

}

Connection conn = null;

try {

conn = dataSource.getConnection( );

sqlCommandBean.setConnection(conn);

// See if the user exists

String sqlValue =

"SELECT * FROM Employee WHERE UserName = ?";

Vector values = new Vector( );

values.addElement(new StringValue(userName));

sqlCommandBean.setSqlValue(sqlValue);

sqlCommandBean.setValues(values);

Vector rows = sqlCommandBean.executeQuery( );

// Create values for insert/update

values.removeAllElements( );

values.addElement(new StringValue(password));

values.addElement(new StringValue(firstName));

values.addElement(new StringValue(lastName));

values.addElement(new StringValue(dept));

values.addElement(new DateValue(

new Date(StringFormat.toDate(empDate,

"yyyy-MM-dd").getTime( ))));

values.addElement(new StringValue(emailAddr));

values.addElement(new TimestampValue(

new Timestamp(System.currentTimeMillis( ))));

values.addElement(new StringValue(userName));

if (rows.size( ) == 0) {

// New user. Insert

StringBuffer sb = new StringBuffer( );

sb.append("INSERT INTO Employee ").

append("(Password, FirstName, LastName, Dept, ").

append("EmpDate, EmailAddr, ModDate, UserName) ").

append("VALUES(?, ?, ?, ?, ?, ?, ?, ?)");

sqlCommandBean.setSqlValue(sb.toString( ));

}

else {

// Existing user. Update

StringBuffer sb = new StringBuffer( );

sb.append("UPDATE Employee ").

append("SET Password = ?, FirstName = ?, ").

append("LastName = ?, Dept = ?, EmpDate = ?, ").

append("EmailAddr = ?, ModDate = ? ").

append("WHERE UserName = ?");

sqlCommandBean.setSqlValue(sb.toString( ));

}

sqlCommandBean.executeUpdate( );

}

catch (SQLException e) {

throw new JspException("SQL error: " + e.getMessage( ));

}

catch (UnsupportedTypeException e) {

throw new JspException("Query result error: " +

e.getMessage( ));

}

catch (ParseException e) {

throw new JspException("Invalid empDate format: " +

e.getMessage( ));

}

finally {

try {

if (conn != null) {

conn.close( );

}

}

catch (SQLException e) {

// Ignore

}

}

return EVAL_PAGE;

}

public void release( ) {

dataSourceName = null;

super.release( );

}

}

This tag handler has one property, named dataSource. It's marked as required in the TLD for the tag, so it will always be set:

...

<tag>

<name>saveEmployeeInfo</name>

<tagclass>com.mycompany.tags.SaveEmployeeInfoTag</tagclass>

<bodycontent>empty</bodycontent>

<info>

</info>

<attribute>

<name>dataSource</name>

<required>true</required>

</attribute>

</tag>

...

In the doEndTag( ) method, all request parameters with information about the employee are first retrieved. If a parameter is missing, an exception is thrown. Then an SQLCommandBean instance is created, the DataSource object fetched from the application scope, and a Connection created and provided to the bean.

The bean is used to execute a SELECT statement to find out if the specified employee is already defined in the database. If not, the bean is used to execute an INSERT statement with all the information provided through the request parameters. Otherwise, the bean is used to execute an UPDATE statement.

The tag handler class described here is intended only to show how you can use the database access classes to implement your own custom actions. The tag handler class could be improved in several ways.

For instance, it could provide default values for missing parameters, such as the current date for a missing employment date, or an email address based on the employee's first and last names if the email address is missing. You could also use a bean as input to the action instead of reading request parameters directly. This would allow the bean to be used as described in

Chapter 6

, and

Chapter 8

, to capture and validate the user input until all information is valid, and then pass it on to the custom action for permanent storage of the information in a database.

© 2001, O'Reilly & Associates, Inc.

Part IV: Appendixes

In this part of the book, you find reference material, such as descriptions of JSP elements and classes, all book example components, the web application deployment descriptor, and more.

Appendix A

Appendix B

Appendix C

Appendix D

Appendix E

© 2001, O'Reilly & Associates, Inc.

A.1 Directive Elements

Directive elements are used to specify information about the page itself, especially information that doesn't differ between requests for the page. The general directive syntax is:

<%@ directiveName attr1="value1" attr2="value2" %>

The attribute values can be enclosed with single quotes instead of double quotes. The directive name and all attribute names are case-sensitive.

A.1.1 include Directive

The include directive includes a static file, merging its content with the including page before the combined result is converted to a JSP page implementation class. The include directive supports the

attribute described in Table A.1

.

Attribute Name Default

Table A.1, include Directive Attribute

Description

file

No default A page-relative or context-relative URI path for the file to include

A page can contain multiple include directives. The including page and all included pages together form what is called a JSP translation unit.

Example:

<%@ include file="header.html" %>

A.1.2 page Directive

The page directive defines page-dependent attributes, such as scripting language, error page, and

buffering requirements. It supports the attributes described in Table A.2

.

Table A.2, page Directive Attributes

Attribute Name Default

autoFlush true

Description

Set to true if the page buffer should be flushed automatically when it's full, or to false

if an exception should be thrown when it's full.

buffer 8kb

Specifies the buffer size for the page. The value must be expressed as the size in kilobytes followed by kb, or be the keyword none to disable buffering.

contentType text/html

The MIME type for the response generated by the page, and optionally the charset for the source page as well as the response; e.g., text/html;charset=Shift_JIS.

errorPage

No default

A page-relative or context-relative URI path for the JSP page to forward to if an exception is thrown by code in the page.

extends import

No default

The fully qualified name of a Java class that the generated JSP page implementation class extends. The class must implement the JspPage or HttpJspPage interface in the javax.servlet.jsp package.

Note that the recommendation is to not use this attribute. Specifying your own superclass restricts the JSP container's ability to provide a specialized, highperformance superclass.

No default

A Java import declaration, i.e., a comma-separated list of fully qualified class names or package names followed by .* (for all public classes in the package).

info isErrorPage isThreadSafe true language

No default

Text that a web container may use as a description of the page in its administration user interface.

false java

Set to true for a page that is used as an error page, to make the implicit exception

variable available to scripting elements. Use false for regular JSP pages.

Set to true if the container is allowed to run multiple threads through the page (i.e., lets the page serve parallel requests). If set to false, the container serializes all requests for the page. It may also use a pool of page implementation class instances to serve more than one request at a time. The recommendation is to always use true, and handle multithread issues by avoiding JSP declarations and ensuring that all objects used by the page are thread-safe.

Defines the scripting language used in the page.

session true

Set to true if the page should participate in a user session. If set to false, the implicit session variable is not available to scripting elements in the page.

A translation unit (the JSP source file and any files included via the include directive) can contain more than one page directive, as long as there is only one occurrence of an attribute, with the exception of the import attribute. If multiple import attribute values are used, they are combined into one list of import definitions.

Example:

<%@ page language="java" contentType="text/html;charset=Shift_JIS"%>

<%@ page import="java.util.*, java.text.*" %>

<%@ page import="java.sql.Date" %>

A.1.3 taglib Directive

Declares a tag library, containing custom actions, that is used in the page. The taglib directive supports the attributes described in

Table A.3

.

Table A.3, taglib Directive Attributes

Attribute Name Default Description

prefix

No default Mandatory. The prefix to use in the action element names for all actions in the library.

uri

No default

Mandatory. Either a symbolic name for the tag library defined in the web.xml file for the application, or a page-relative or context-relative URI path for the library's TLD file or

JAR file.

Example:

<%@ taglib uri="/orataglib" prefix="ora" %>

© 2001, O'Reilly & Associates, Inc.

A.2 Scripting Elements

Scripting elements let you add small pieces of code to a JSP page, such as an if statement that generates different HTML depending on some condition. The scripting code must be written in the language defined by the page directive. It is executed when the JSP page is requested.

A.2.1 Declaration

A declaration starts with <%! and ends with %>. The content between the start and the end characters must be a complete, valid declaration in the scripting language defined by the page directive. The JSP implicit variables are not visible in a declaration element.

A declaration can be used to declare a scripting language variable or method. When the scripting language is Java, a variable declared by a declaration element ends up as an instance variable in the JSP page implementation class. It is therefore visible to parallel threads (requests) processing the page, and needs to be handled in a thread-safe manner. A thread-safe alternative is to declare variables within a scriptlet element instead. It then becomes a local variable of the method in the page implementation class used to process each request, and is not shared by parallel threads.

Example:

<%! int globalCounter = 0; %>

A.2.2 Expression

An expression starts with <%= and ends with %>. The content between the start and the end characters must be a complete, valid expression in the scripting language defined by the page directive that results in or can be converted to a string. All JSP implicit variables are visible in an expression element.

Example:

<%= globalCounter++ %>

A.2.3 Scriptlet

A scriptlet starts with <% and ends with %>. The content between the start and the end characters must be a code fragment in the scripting language defined by the page directive. Scriptlet code fragments are combined with code for sending the template data between them to the browser. The combination

of all scriptlets in a page must form valid scripting language statements. All JSP implicit variables are visible in a scripting element.

Example:

<% java.util Date clock = new java.util.Date( ) %>

<% if (clock.getHours( ) < 12) { %>

Good morning!

<% } else if (clock.getHours( ) < 17) { %>

Good day!

<% } else { %>

Good evening!

<% } %>

© 2001, O'Reilly & Associates, Inc.

A.3 Action Elements

Actions are executed when the JSP page is requested by a client. They are inserted in a page using

XML element syntax, and encapsulate functionality such as input validation using beans, database access, or passing control to another page. The JSP specification defines a few standard action elements, described in this section, and also includes a framework for developing custom action elements.

An action element consists of a start tag (optionally with attributes), a body, and an end tag. Other elements can be nested in the body. Here's an example:

<jsp:forward page="nextPage.jsp">

<jsp:param name="aParam" value="aValue" />

</jsp:forward>

If the action element doesn't have a body, a shorthand notation can be used in which the start tag ends with /> instead of >, as shown by the <jsp:param> action in this example. The action element name and attribute names are case-sensitive.

Some action attributes accept a request-time attribute value, using the JSP expression syntax:

<% String headerPage = currentTemplateDir + "/header.jsp"; %>

<jsp:include page="<%= headerPage %>" />

The attribute descriptions for each action in this section define whether a request-time attribute value is accepted or not.

A.3.1 <jsp:fallback>

The <jsp:fallback> action can only be used in the body of a <jsp:plugin> action. Its body is used to specify the template text to use for browsers that do not support the HTML <embed> or

<object>

elements. This action supports no attributes.

Example:

<jsp:plugin type="applet" code="Clock2.class"

codebase="applet"

jreversion="1.2" width="160" height="150" >

<jsp:fallback>

Plugin tag OBJECT or EMBED not supported by browser.

</jsp:fallback>

</jsp:plugin>

A.3.2 <jsp:forward>

The <jsp:forward> action passes the request processing control to another JSP page or servlet in the same web application. The execution of the current page is terminated, giving the target resource full control over the request.

If any response content has been buffered when the <jsp:forward> action is executed, the buffer is cleared first. If the response has already been committed (i.e., partly sent to the browser), the forwarding fails with an IllegalStateException.

The URI path information available through the implicit request object is adjusted to reflect the URI path information for the target resource. All other request information is left untouched, so the target resource has access to all the original parameters and headers passed with the request. Additional parameters can be passed to the target resource through <jsp:param> elements in the

<jsp:forward>

element's body.

The <jsp:forward> action supports the attributes described in Table A.4

.

Table A.4, <jsp:forward>Attributes

Attribute Name Java Type Request-Time Value Accepted Description

page String

Yes

Mandatory. A page-relative or context-relative URI path for the resource to forward to.

Example:

<jsp:forward page="list.jsp" />

A.3.3 <jsp:getProperty>

The <jsp:getProperty> action adds the value of a bean property, converted to a string, to the

response generated by the page. The attributes described in Table A.5

are supported.

Table A.5, <jsp:getProperty>Attributes

Attribute Name Java Type Request-Time Value Accepted Description

name String

No

Mandatory. The name assigned to a bean in one of the

JSP scopes.

property String

No

Mandatory. The name of the bean's property to include in the page.

Example:

<jsp:getProperty name="clock" property="hours" />

A.3.4 <jsp:include>

The <jsp:include> action includes the response from another JSP page, servlet, or static file in the same web application. The execution of the current page continues after including the response generated by the target resource.

If any response content has been buffered when the <jsp:include> action is executed, the buffer is flushed first. Even though this behavior can be controlled by the flush attribute, the only valid value in

JSP 1.1 is true. This limitation will likely be lifted in a future version of JSP.

The URI path information available through the implicit request object reflects the URI path information for the source JSP page even in the target resource. All other request information is also left untouched, so the target resource has access to all the original parameters and headers passed with the request. Additional parameters can be passed to the target resource through <jsp:param> elements in the <jsp:include> element's body.

The <jsp:include> action supports the attributes described in Table A.6

.

Table A.6, <jsp:include>Attributes

Attribute Name Java Type Request-Time Value Accepted Description

page String

Yes

Mandatory. A page-relative or context-relative URI path for the resource to include.

flush boolean No

Mandatory in JSP 1.1 with true as the only accepted value.

Example:

<jsp:include page="navigation.jsp" />

A.3.5 <jsp:param>

The <jsp:param> action can be used in the body of a <jsp:forward> or <jsp:include> action to specify additional request parameters for the target resource, as well as in the body of a

<jsp:params>

action to specify applet parameters. The attributes described in

Table A.7

are

supported.

Attribute Name

Table A.7, <jsp:param>Attributes

Java Type Request-Time Value Accepted Description

name String

No Mandatory. The parameter name.

value String

Yes Mandatory. The parameter value.

Example:

<jsp:include page="navigation.jsp">

<jsp:param name="bgColor" value="<%= currentBGColor %>" />

</jsp:include>

A.3.6 <jsp:params>

The <jsp:params> action can only be used in the body of a <jsp:plugin> action to enclose a set of <jsp:param> actions that are used to specify applet parameters. This action supports no attributes.

Example:

<jsp:plugin type="applet" code="Clock2.class"

codebase="applet"

jreversion="1.2" width="160" height="150" >

<jsp:params>

<jsp:param name="bgcolor" value="ccddff" />

</jsp:params>

</jsp:plugin>

A.3.7 <jsp:plugin>

The <jsp:plugin> action generates HTML <embed> or <object> elements (depending on the browser type) that result in the download of the Java Plugin software (if required) and subsequent execution of the specified Java Applet or JavaBeans component. The body of the action can contain a

<jsp:params>

element to specify applet parameters, and a <jsp:fallback> element to specify the text shown in browsers that do not support the <embed> or <object> HTML elements. For more information about the Java Plugin, see http://java.sun.com/products/plugin/ .

The attributes described in Table A.8

are supported.

Table A.8, <jsp:plugin>Attributes

Attribute Name Java Type

Request-Time Value

Accepted

Description

align String

No

Optional. Alignment of the applet area. One of bottom, middle , or top.

archive code codebase String

No height hspace

String

No

String

No

String

No

String

No

Optional. A comma-separated list of URIs for archives containing classes and other resources that will be preloaded.

The classes are loaded using an instance of an

AppletClassLoader

with the given codebase. Relative

URIs for archives are interpreted with respect to the applet's codebase

.

Mandatory. The fully qualified class name for the object.

Mandatory. The relative URL for the directory that contains the class file. The directory must be a subdirectory to the directory holding the page, according to the HTML 4.0 specification.

Optional. The height of the applet area, in pixels or percentage.

Optional. The amount of whitespace to be inserted to the left and right of the applet area, in pixels.

iepluginurl String

No jreversion String

No name String

No nspluginurl String

No title String

No type vspace width

String

No

String

No

String

No

Optional. The URL for the location of the Internet Explorer

Java Plugin. The default is implementation-dependent.

Optional. Identifies the spec version number of the JRE the component requires in order to operate. The default is 1.1.

Optional. The applet name, used by other applets on the same page that need to communicate with it.

Optional. The URL for the location of the Netscape Java

Plugin. The default is implementation-dependent.

Optional. Text to be rendered by the browser for the applet in a some way, for instance as a "tool tip."

Mandatory. The type of object to embed, one of applet or bean

.

Optional. The amount of whitespace to be inserted above and below the applet area, in pixels.

Optional. The width of the applet area, in pixels or percentage.

Example:

<jsp:plugin type="applet" code="Clock2.class"

codebase="applet"

jreversion="1.2" width="160" height="150" >

<jsp:params>

<jsp:param name="bgcolor" value="ccddff" />

</jsp:params>

<jsp:fallback>

Plugin tag OBJECT or EMBED not supported by browser.

</jsp:fallback>

</jsp:plugin>

A.3.8 <jsp:setProperty>

The <jsp:setProperty> action sets the value of one or more bean properties. The attributes described in

Table A.9

are supported.

Table A.9, <jsp:setProperty>Attributes

Attribute Name Java Type Request-Time Value Accepted Description

name property param value

String

No

String

No

String

No

See below Yes

Mandatory. The name assigned to a bean in one of the

JSP scopes.

Mandatory. The name of the bean's property to set, or an asterisk (*) to set all properties with names matching request parameters.

Optional. The name of a request parameter that holds the value to use for the specified property. If omitted, the parameter name and the property name must be the same.

Optional. An explicit value to assign to the property.

This attribute cannot be combined with the param attribute.

The property type can be any valid Java type, including primitive types and arrays (i.e., an indexed property). If a runtime attribute value is specified by the value attribute, the type of the expression must match the property's type.

If the value is a string, either in the form of a request parameter value or explicitly specified by the value

attribute, it is converted to the property's type as described in

Table A.10

.

Property Type

Table A.10, Conversion of String Value to Property Type

Conversion Method

boolean

or Boolean

Boolean.valueOf(String) byte

or Byte

Byte.valueOf(String) char

or Character

String.charAt(int) double

or Double Double.valueOf(String)

int

or Integer float

or Float long

or Long

Integer.valueOf(String)

Float.valueOf(String)

Long.valueOf(String)

Example:

<jsp:setProperty name="user" property="*" />

<jsp:setProperty name="user" property="modDate"

value="<%= new java.util.Date( ) %>" />

A.3.9 <jsp:useBean>

The <jsp:useBean> action associates a Java bean with a name in one of the JSP scopes and also makes it available as a scripting variable. An attempt is first made to find a bean with the specified name in the specified scope. If it's not found, a new instance of the specified class is created. The

attributes described in Table A.11

are supported.

Table A.11, <jsp:useBean>Attributes

Attribute Name Java Type

Request-Time Value

Accepted

Description

beanName String

Yes

Optional. The name of the bean, as expected by the instantiate( )

method of the Beans class in the java.beans

package.

class String

No id scope

String

No

String

No

Optional. The fully qualified class name for the bean.

Mandatory. The name to assign to the bean in the specified scope, as well as the name of the scripting variable.

Optional. The scope for the bean, one of page, request , session, or application. The default is page.

type String

No

Optional. The fully qualified type name for the bean

(i.e., a superclass or an interface implemented by the bean's class).

Of the optional attributes, at least one of class or type must be specified. If both are specified, class

must be assignable to type. The beanName attribute must be combined with the type attribute, and is not valid with the class attribute.

The action is processed in these steps:

1. Attempt to locate an object based on the id and scope attribute values.

2. Define a scripting language variable with the given id of the specified type or class.

3. If the object is found, the variable's value is initialized with a reference to the located object, cast to the specified type or class. This completes the processing of the action. If the action

element has a nonempty body, it is ignored.

4. If the object is not found in the specified scope and neither class nor beanName is specified, a InstantiationException is thrown. This completes the processing of the action.

5. If the object is not found in the specified scope, and the class attribute specifies a nonabstract class with a public no-arg constructor, a new instance of the class is created and associated with the scripting variable and with the specified name in the specified scope. After this, Step 7 is performed.

If the object is not found and the specified class doesn't fulfill the requirements, a

InstantiationException

is thrown. This completes the processing of the action.

6. If the object is not found in the specified scope and the beanName attribute is specified, the instantiate( )

method of the java.beans.Beans class is invoked, with the

ClassLoader

of the JSP implementation class instance and the beanName as arguments. If the method succeeds, the new object reference is associated with the scripting variable and with the specified name in the specified scope. After this, Step 7 is performed.

7. If the action element has a nonempty body, the body is processed. The scripting variable is initialized and available within the scope of the body. The text of the body is treated as elsewhere: any template text is passed through to the response, and scriptlets and action tags are evaluated.

A common use of a nonempty body is to complete initializing the created instance; in that case, the body typically contains <jsp:setProperty> actions and scriptlets. This completes the processing of the action.

Example:

<jsp:useBean id="clock" class="java.util.Date" />

A.3.10 Custom Actions

A custom action element can be developed by a programmer to extend the JSP language. The examples in this book use custom actions for database access, internationalization, access control, and more.

They are described in Appendix C

. The general syntax for custom actions is the same as for the JSP standard actions: a start tag (optionally with attributes), a body, and an end tag. Other elements and template text can be nested in the body. Here's an example:

<ora:loop name="anArray" loopId="current" className="String">

<li><%= current %>

</ ora:loop >

The tag library containing the custom actions must be declared by the taglib directive, assigning a prefix for the custom action elements (ora in this example) before a custom action can be used in a

JSP page.

© 2001, O'Reilly & Associates, Inc.

A.4 Comments

You can use JSP comments in JSP pages to describe what a scripting element or action is doing:

<%-- This is a comment --%>

All text between the start and stop tags is ignored by the JSP container and is not included in the response. The comment text can be anything except the character sequence representing the closing tag: --%>.

Besides describing what's going on in the JSP page, comments can also be used to "comment out" portions of the JSP page, for instance during testing:

<jsp:useBean id="user" class="com.mycompany.UserBean" />

<%--

<jsp:setProperty name="user" property="*" />

<jsp:setProperty name="user" property="modDate"

value="<%= new java.util.Date( ) %>" />

<% boolean isValid = user.isValid( ); %>

--%>

The action and scripting elements within the comment are not executed.

© 2001, O'Reilly & Associates, Inc.

A.5 Escape Characters

Since certain character sequences are used to represent start and stop tags, you sometimes need to escape a character so the container doesn't interpret it as part of a special character sequence.

In a scripting element, if you need to use the characters %> literally, you must escape the greater-than character with a backslash:

<% String msg = "Literal %\> must be escaped"; %>

To avoid the character sequence <% in template text being interpreted as the start of a scripting element, you must escape the percent sign:

This is template text, and <\% is not a start of a scriptlet.

In an attribute value, you must use the following escapes: attr='a value with an escaped \' single quote' attr="a value with an escaped \" double quote" attr="a value with an escaped \\ backslash" attr="a value with an escaped %\> scripting end tag" attr="a value with an escaped <\% scripting start tag"

© 2001, O'Reilly & Associates, Inc.

B.1 Implicit Variables

The JSP specification defines a number of implicit variables. Most of the implicit variables have types defined by classes and interfaces in the servlet specification's javax.servlet.http package, but two are part of the JSP javax.servlet.jsp package and one is part of the Java core API. Scripting elements in a JSP page can use these objects to access request and response information as well as objects saved in one of the JSP scopes: page, request, session, and application.

application

Synopsis

Variable Name:

Interface Name:

Extends:

Implemented by:

JSP Page Type: application javax.servlet.ServletContext

None

Internal container-dependent class

Available in both regular JSP pages and error pages

Description

The ServletContext provides resources shared within a web application. It holds attribute values representing the JSP application scope. An attribute value can be an instance of any valid Java class. It also defines a set of methods that a JSP page or a servlet uses to communicate with its container, for example, to get the MIME type of a file, dispatch requests, or write to a log file. The web container is responsible for providing an implementation of the ServletContext interface.

A ServletContext is assigned a specific URI path prefix within a web server. For example, a context could be responsible for all resources under http://www.mycorp.com/catalog . All requests that start with the /catalog request path, which is known as the context path, are routed to this servlet context.

Only one instance of a ServletContext may be available to the servlets and JSP pages in a web application. If the web application indicates that it is distributable, there must be only one instance of the

ServletContext

object in use per application per Java Virtual Machine.

Interface Declaration

public interface ServletContext {

public Object getAttribute(String name);

public Enumeration getAttributeNames( );

public ServletContext getContext(String uripath);

public String getInitParameter(String name);

public Enumeration getInitParameterNames( );

public int getMajorVersion( );

public String getMimeType(String filename);

public int getMinorVersion( );

public RequestDispatcher getNamedDispatcher(String name);

public String getRealPath(String path);

public RequestDispatcher getRequestDispatcher(String path);

public URL getResource(String path) throws MalformedURLException;

public InputStream getResourceAsStream(String path);

public String getServerInfo( );

public void log(String message);

public void log(String message, Throwable cause);

public void removeAttribute(String name);

public void setAttribute(String name, Object attribute);

// Deprecated methods

public Servlet getServlet(String name) throws ServletException;

public Enumeration getServlets( );

public Enumeration getServletNames( );

public void log(Exception exception, String message);

}

Methods

public Object getAttribute(String name)

Returns the servlet context attribute with the specified name, or null if there is no attribute by that name. Context attributes can be set by a servlet or a JSP page, representing the JSP application scope. A container can also use attributes to provide information that is not already available through methods in this interface.

public java.util.Enumeration getAttributeNames( )

Returns an Enumeration of String objects containing the attribute names available within this servlet context.

public ServletContext getContext(String uripath)

Returns a ServletContext object that corresponds to a specified URI in the web container.

This method allows servlets and JSP pages to gain access to contexts other than its own. The URI path must be absolute (beginning with /) and is interpreted based on the container's document root. In a security-conscious environment, the container may return null for a given URI.

public String getInitParameter(String name)

Returns a String containing the value of the named context-wide initialization parameter, or null

if the parameter does not exist. Context initialization parameters can be defined in the web application deployment descriptor.

public java.util.Enumeration getInitParameterNames( )

Returns the names of the context's initialization parameters as an Enumeration of String objects, or an empty Enumeration if the context has no initialization parameters.

public int getMajorVersion( )

Returns the major version of the Java servlet API that this web container supports. For example, a container that complies with the Servlet 2.3 API returns 2.

public String getMimeType(String filename)

Returns the MIME type of the specified file, or null if the MIME type is not known. The

MIME type is determined by the configuration of the web container, and may be specified in a web application deployment descriptor.

public int getMinorVersion( )

Returns the minor version of the Java servlet API that this web container supports. For example, a container that complies with the Servlet 2.3 API returns 3.

public RequestDispatcher getNamedDispatcher(String name)

Returns a RequestDispatcher object that acts as a wrapper for the named servlet or JSP page. Names can be defined for servlets and JSP pages in the web application deployment descriptor.

public String getRealPath(String path)

Returns a String containing the filesystem path for a specified context-relative path. This method returns null if the web container cannot translate the path to a filesystem path for any reason (such as when the content is being made available from a WAR archive).

public RequestDispatcher getRequestDispatcher(String path)

Returns a RequestDispatcher object that acts as a wrapper for the resource located at the specified context-relative path. The resource can be dynamic (servlet or JSP) or static (for instance, a regular HTML file).

public java.net.URL getResource(String path) throws

MalformedURLException

Returns a URL to the resource that is mapped to the specified context-relative path. This method allows the web container to make a resource available to servlets and JSP pages from other sources than a local filesystem, such as a database or a WAR file.

The URL provides access to the resource content directly, so be aware that requesting a JSP page returns a URL for the JSP source code as opposed to the processed result. Use a

RequestDispatcher

instead to include results of an execution.

This method returns null if no resource is mapped to the pathname.

public java.io.InputStream getResourceAsStream(String path)

Returns the resource mapped to the specified context-relative path as an InputStream object.

See getResource( ) for details.

public String getServerInfo( )

Returns the name and version of the servlet container on which the servlet or JSP page is running as a String with the format "servername/versionnumber" (for example, "Tomcat/3.2").

Optionally, a container may include other information, such as the Java version and operating system information, within parentheses.

public void log(String message)

Writes the specified message to a container log file. The name and type of the log file is containerdependent.

public void log(String message, Throwable cause)

Writes the specified message and a stack trace for the specified Throwable to the servlet log file. The name and type of the log file is container-dependent.

public void removeAttribute(String name)

Removes the attribute with the specified name from the servlet context.

public void setAttribute(String name, Object attribute)

Binds an object to the specified attribute name in this servlet context. If the specified name is already used for an attribute, this method removes the old attribute and binds the name to the new attribute.

The following methods are deprecated: public Servlet getServlet(String name) throws ServletException

This method was originally defined to retrieve a servlet from a ServletContext. As of the

Servlet 2.1 API, this method always returns null, and remains only to preserve binary compatibility. This method will be permanently removed in a future version of the Java servlet

API.

public Enumeration getServlets( )

This method was originally defined to return an Enumeration of all the servlets known to this servlet context. As of the Servlet 2.1 API, this method always returns an empty Enumeration, and remains only to preserve binary compatibility. This method will be permanently removed in a future version of the Java servlet API.

public Enumeration getServletNames( )

This method was originally defined to return an Enumeration of all the servlet names known to this context. As of the Servlet 2.1 API, this method always returns an empty Enumeration, and remains only to preserve binary compatibility. This method will be permanently removed in a future version of the Java servlet API.

public void log(Exception exception, String message)

This method was originally defined to write an exception's stack trace and an explanatory error message to the web container log file. As of the Servlet 2.1 API, the recommendation is to use log(String, Throwable)

instead.

config

Synopsis

Variable Name:

Interface Name:

Extends:

Implemented by:

JSP Page Type: config javax.servlet.

ServletConfig

None

Internal container-dependent class

Available in both regular JSP pages and error pages

Description

A ServletConfig instance is used by a web container to pass information to a servlet or JSP page during initialization. The configuration information contains initialization parameters (defined in the web application deployment descriptor) and the ServletContext object representing the web application the servlet or JSP page belongs to.

Interface Declaration

public interface ServletConfig {

public String getInitParameter(String name);

public Enumeration getInitParameterNames( );

public ServletContext getServletContext( );

public String getServletName( );

}

Methods

public String getInitParameter(String name)

Returns a String containing the value of the specified servlet or JSP page initialization parameter, or null if the parameter does not exist.

public java.util.Enumeration getInitParameterNames( )

Returns the names of the servlet's or JSP page's initialization parameters as an Enumeration of

String

objects, or an empty Enumeration if the servlet has no initialization parameters.

public ServletContext getServletContext( )

Returns a reference to the ServletContext that the servlet or JSP page belongs to.

public String getServletName( )

Returns the name of the servlet instance or JSP page. The name may be assigned in the web application deployment descriptor. For an unregistered (and thus unnamed) servlet instance or

JSP page, the servlet's class name is returned.

exception

Synopsis

Variable Name: exception

Class Name: java.lang.

Throwable

Extends:

Implements:

None java.io.Serializable

Implemented by: Part of the standard Java library

JSP Page Type: Available only in a page marked as an error page using the page directive isErrorPage attribute

Description

The exception variable is assigned to the subclass of Throwable that caused the error page to be invoked. The Throwable class is the superclass of all errors and exceptions in the Java language. Only objects that are instances of this class (or of one of its subclasses) are thrown by the Java Virtual

Machine or by the Java throw statement. See the Java documentation at http://java.sun.com/docs/index.html

for a description of the Throwable class.

out

Synopsis

Variable Name: out

Class Name: javax.servlet.jsp.JspWriter

Extends: java.io.Writer

Implements: None

Implemented by: A concrete subclass of this abstract class is provided as an internal container-dependent class

JSP Page Type: Available in both regular JSP pages and error pages

Description

The out variable is assigned to a concrete subclass of the JspWriter abstract class by the web container. JspWriter emulates some of the functionality found in the java.io.BufferedWriter

and java.io.PrintWriter classes. It differs, however, in that it throws a java.io.IOException from the print methods, while the PrintWriter does not.

If the page directive attribute autoflush is set to true, all the I/O operations on this class automatically flush the contents of the buffer when it's full. If autoflush is set to false, all the I/O operations on this class throw an IOException when the buffer is full.

Class Summary

public abstract class JspWriter extends java.io.Writer {

// Constructor

protected JspWriter(int bufferSize, boolean autoFlush);

// Methods

public abstract void clear( ) throws java.io.IOException;

public abstract void clearBuffer( ) throws java.io.IOException;

public abstract void close( )throws java.io.IOException;

public abstract void flush( ) throws java.io.IOException;

public int getBufferSize( );

public abstract int getRemaining( );

public boolean isAutoFlush( );

}

Constructor

protected JspWriter(int bufferSize, boolean autoFlush)

Creates an instance with at least the specified buffer size and auto-flush behavior.

Methods

public abstract void clear( ) throws java.io.IOException

Clears the contents of the buffer. If the buffer has already been flushed, throws an

IOException

to signal that some data has already been irrevocably written to the client response stream.

public abstract void clearBuffer( ) throws java.io.IOException

Clears the current contents of the buffer. Unlike clear( ), this method does not throw an

IOException

if the buffer has already been flushed. It just clears the current content of the buffer and returns.

public abstract void close( ) throws java.io.IOException

Closes the JspWriter after flushing it. A call to flush( ) or write( ) after a call to close( )

results in an IOException being thrown. If close( ) is called on a previously closed JspWriter, it is ignored.

public abstract void flush( ) throws java.io.IOException

Flushes the current contents of the buffer to the underlying writer, and flushes the underlying writer as well. This means the buffered content is delivered to the client immediately.

public int getBufferSize( )

Returns the size of the buffer in bytes, or 0 if it is not buffered.

public abstract int getRemaining( )

Returns the number of bytes unused in the buffer.

public boolean isAutoFlush( )

Returns true if this JspWriter is set to auto-flush the buffer, false otherwise.

page

Synopsis

Variable Name:

Class Name:

Extends:

Implements:

Implemented by:

JSP Page Type: page

Object

None

None

Part of the standard Java library

Available in both regular JSP pages and error pages

Description

The page variable is assigned to the instance of the JSP implementation class declared as an Object.

This variable is rarely, if ever, used. See the Java documentation at http://java.sun.com/docs/index.html

for a description of the Object class.

pageContext

Synopsis

Variable Name: pageContext

Class Name: javax.servlet.jsp.PageContext

Extends:

Implements:

None

None

Implemented by: A concrete subclass of this abstract class is provided as an internal container-dependent class

JSP Page Type: Available in both regular JSP pages and error pages

Description

A PageContext instance provides access to all the JSP scopes and several page attributes, and offers a layer above the container implementation details to enable a container to generate portable JSP implementation classes. The JSP page scope is represented by PageContext attributes. A unique instance of this object is created by the web container and assigned to the pageContext variable for each request.

Class Summary

public abstract class PageContext {

// Constants

public static final int APPLICATION_SCOPE;

public static final int PAGE_SCOPE;

public static final int REQUEST_SCOPE;

public static final int SESSION_SCOPE;

// Constructor

public PageContext( );

// Methods

public abstract java.lang.Object findAttribute(String name);

public abstract void forward(String relativeUrlPath)

throws ServletException, java.io.IOException;

public abstract Object getAttribute(String name);

public abstract Object getAttribute(String name,

int scope);

public abstract

java.util.Enumeration getAttributeNamesInScope(int scope);

public abstract int getAttributesScope(String name);

public abstract Exception getException( );

public abstract JspWriter getOut( );

public abstract Object getPage( );

public abstract ServletRequest getRequest( );

public abstract ServletResponse getResponse( );

public abstract ServletConfig getServletConfig( );

public abstract ServletContext getServletContext( );

public abstract HttpSession getSession( );

public abstract void handlePageException(Exception e)

throws ServletException, java.io.IOException;

public abstract void include(String relativeUrlPath)

throws ServletException, java.io.IOException;

public abstract void initialize(Servlet servlet,

ServletRequest request, ServletResponse response,

String errorPageURL, boolean needsSession, int bufferSize,

boolean autoFlush) throws java.io.IOException,

IllegalStateException, IllegalArgumentException;

public JspWriter popBody( );

public BodyContent pushBody( );

public abstract void release( );

public abstract void removeAttribute(String name);

public abstract void removeAttribute(String name, int scope);

public abstract void setAttribute(String name, Object attribute);

public abstract void setAttribute(String name, Object o, int scope);

}

Constructor

public PageContext( )

Creates an instance of the PageContext class. Typically, an instance is created and initialized by the JspFactory class.

Methods

public abstract Object findAttribute(String name)

Searches for the named attribute in page, request, session (if valid), and application scopes, in order, and returns the associated value, or null if the attribute is not found.

public abstract void forward(String relativeUrlPath) throws

ServletException, java.io.IOException

This method is used to forward the current request to another active component, such as a servlet or JSP page, in the application. If the specified URI starts with a slash, it's interpreted as a context-relative path; otherwise, it's interpreted as a page-relative path.

The response must not be modified after calling this method, since the response is committed before this method returns.

public abstract Object getAttribute(String name)

Returns the object associated with the specified attribute name in the page scope, or null if the attribute is not found.

public abstract Object getAttribute(String name, int scope)

Returns the object associated with the specified attribute name in the specified scope, or null if the attribute is not found. The scope argument must be one of the int values specified by the

PageContext

static scope variables.

public abstract java.util.Enumeration getAttributeNamesInScope(int scope)

Returns an Enumeration of String objects containing all attribute names for the specified scope. The scope argument must be one of the int values specified by the PageContext static scope variables.

public abstract int getAttributesScope(String name)

Returns one of the int values specified by the PageContext static scope variables for the scope of the object associated with the specified attribute name, or if the attribute is not found.

public abstract Exception getException( )

Returns the Exception that caused the current page to be invoked if its page directive isErrorPage

is set to true.

public abstract JspWriter getOut( )

Returns the current JspWriter for the page. When this method is called by a tag handler that implements BodyTag or is nested in the body of another action element, the returned object may be an instance of the BodyContent subclass.

public abstract Object getPage( )

Returns the object that represents the JSP page implementation class instance that this

PageContext

is associated with.

public abstract ServletRequest getRequest( )

Returns the current ServletRequest.

public abstract ServletResponse getResponse( )

Returns the current ServletResponse.

public abstract ServletConfig getServletConfig( )

Returns the ServletConfig for this JSP page implementation class instance.

public abstract ServletContext getServletContext( )

Returns the ServletContext for this JSP page implementation class instance.

public abstract HttpSession getSession( )

Returns the current HttpSession, or null if the page directive session attribute is set to false

.

public abstract void handlePageException(Exception e) throws

ServletException, java.io.IOException

This method is intended to be called by the JSP page implementation class only to process unhandled exceptions by forwarding the request exception to either the error page specified by the page directive's errorPage attribute, or to perform an implementation-dependent action if no error page is specified.

public abstract void include(String relativeUrlPath) throws

ServletException, java.io.IOException

Causes the specified resource to be processed as part of the current request. The current

JspWriter

is flushed before invoking the target resource, and the output of the target resource's processing of the request is written directly to the current ServletResponse object's writer. If the specified URI starts with a slash, it's interpreted as a context-relative path, otherwise as a page-relative path.

public abstract void initialize(Servlet servlet, ServletRequest request, ServletResponse response, String errorPageURL, boolean needsSession, int bufferSize, boolean autoFlush) throws java.io.IOException, IllegalStateException, IllegalArgumentException

This method is called to initialize a PageContext object so that it can be used by a JSP implementation class to service an incoming request. This method is typically called from the

JspFactory.getPageContext( )

method.

public JspWriter popBody( )

This method is intended to be called by the JSP page implementation class only to reassign the previous JspWriter, saved by the matching pushBody( ) method, as the current

JspWriter

.

public BodyContent pushBody( )

This method is intended to be called by the JSP page implementation class only to get a new

BodyContent

object, and to save the current JspWriter on the PageContext object's internal stack.

public abstract void release( )

Resets the internal state of a PageContext, releasing all internal references and preparing the

PageContext

for potential reuse by a later invocation of initialize( ). This method is typically called from the JspFactory.releasePageContext( ) method.

public abstract void removeAttribute(String name)

Removes the object reference associated with the specified attribute name in the page scope.

public abstract void removeAttribute(String name, int scope)

Removes the object reference associated with the specified attribute name in the specified scope.

The scope argument must be one of the int values specified by the PageContext static scope variables.

public abstract void setAttribute(String name, Object attribute)

Saves the specified attribute name and object in the page scope.

public abstract void setAttribute(String name, Object o, int scope)

Saves the specified attribute name and object in the specified scope. The scope argument must be one of the int values specified by the PageContext static scope variables.

request

Synopsis

Variable Name:

Interface Name:

Extends:

Implemented by:

JSP Page Type: request javax.servlet.http.HttpServletRequest

javax.servlet.ServletRequest

Internal container-dependent class

Available in both regular JSP pages and error pages

Description

The request variable is assigned a reference to an internal container-dependent class that implements a protocol-dependent interface extending the javax.servlet.ServletRequest. Since HTTP is the only protocol supported by JSP 1.1, the class always implements the javax.servlet.http.HttpServletRequest

interface. The method descriptions in this section include all methods from both interfaces.

Interface Declarations

public interface ServletRequest {

public Object getAttribute(String name);

public java.util.Enumeration getAttributeNames( );

public String getCharacterEncoding( );

public int getContentLength( );

public String getContentType( );

public ServletInputStream getInputStream( )

throws java.io.IOException;

public java.util.Locale getLocale( );

public java.util.Enumeration getLocales( );

public String getParameter(String name);

public java.util.Enumeration getParameterNames( );

public String[] getParameterValues( );

public String getProtocol( );

public java.io.BufferedReader getReader( )

throws java.io.IOException;

public String getRemoteAddr( );

public String getRemoteHost( );

public RequestDispatcher getRequestDispatcher(String path);

public String getScheme( );

public String getServerName( );

public int getServerPort( );

public boolean isSecure( );

public void removeAttribute(String name);

public Object setAttribute(String name, Object attribute);

// Deprecated methods

public String getRealPath( );

} public interface HttpServletRequest extends ServletRequest {

public String getAuthType( );

public String getContextPath( );

public Cookie[] getCookies( );

public long getDateHeader(String name);

public String getHeader(String name);

public java.util.Enumeration getHeaderNames( );

public java.util.Enumeration getHeaders(String name);

public int getIntHeader(String name);

public String getMethod( );

public String getPathInfo( );

public String getPathTranslated( );

public String getQueryString( );

public String getRequestedSessionId( );

public String getRequestURI( );

public String getRemoteUser( );

public String getServletPath( );

public HttpSession getSession( );

public HttpSession getSession(boolean create);

public java.security.Principal getUserPrincipal( );

public boolean isRequestedSessionIdFromCookie( );

public boolean isRequestedSessionIdFromURL( );

public boolean isRequestedSessionIdValid( );

public boolean isUserInRole(String role);

// Deprecated methods

public boolean isRequestSessionIdFromUrl( );

}

Methods

public Object getAttribute(String name)

Returns the value of the named attribute as an Object, or null if no attribute of the given name exists.

public java.util.Enumeration getAttributeNames( )

Returns an Enumeration containing the names of the attributes available to this request. The

Enumeration

is empty if the request doesn't have any attributes.

public String getAuthType( )

Returns the name of the authentication scheme used to protect the servlet, for example BASIC or

SSL, or null if the servlet is not protected.

public String getCharacterEncoding( )

Returns the name of the character encoding used in the body of this request, or null if the request does not specify a character encoding.

public int getContentLength( )

Returns the length, in bytes, of the request body made available by the input stream, or -1 if the length is not known.

public String getContentType( )

Returns the MIME type of the body of the request, or null if the type is not known.

public String getContextPath( )

Returns the portion of the request URI that indicates the context of the request.

public Cookie[] getCookies( )

Returns an array containing all of the Cookie objects the client sent with this request, or null if the request contains no cookies.

public long getDateHeader(String name)

Returns the value of the specified request header as a long value that represents a date value, or -

1 if the header is not included in the request.

public String getHeader(String name)

Returns the value of the specified request header as a String, or null if the header is not included with the request.

public java.util.Enumeration getHeaderNames( )

Returns all the header names contained by this request as an Enumeration of String objects. The Enumeration is empty if the request doesn't have any headers.

public java.util.Enumeration getHeaders(String name)

Returns all the values of the specified request header as an Enumeration of String objects.

The Enumeration is empty if the request doesn't contain the specified header.

public ServletInputStream getInputStream( ) throws java.io.IOException

Retrieves the body of the request as binary data using a ServletInputStream.

public int getIntHeader(String name)

Returns the value of the specified request header as an int, or -1 if the header is not included in the request.

public java.util.Locale getLocale( )

Returns the preferred Locale in which the client accepts content, based on the Accept-

Language

header.

public java.util.Enumeration getLocales( )

Returns an Enumeration of Locale objects indicating, in decreasing order and starting with the preferred locale, the locales that are acceptable to the client based on the Accept-

Language

header.

public String getMethod( )

Returns the name of the HTTP method with which this request was made, for example GET,

POST

, or PUT.

public String getParameter(String name)

Returns the value of a request parameter as a String, or null if the parameter does not exist.

public java.util.Enumeration getParameterNames( )

Returns an Enumeration of String objects containing the names of the parameters contained in this request.

public String[] getParameterValues( )

Returns an array of String objects containing all of the values the given request parameter has, or null if the parameter does not exist.

public String getPathInfo( )

Returns any extra path information associated with the URI the client sent when it made this request, or null if there is no extra path information. For a JSP page, this method always returns null

.

public String getPathTranslated( )

Returns the result of getPathInfo( ) translated into the corresponding filesystem path.

Returns null if getPathInfo( ) returns null.

public String getProtocol( )

Returns the name and version of the protocol the request uses in the form

protocol/majorVersion.minorVersion

, for example, HTTP/1.1.

public String getQueryString( )

Returns the query string that is contained in the request URI after the path.

public java.io.BufferedReader getReader( ) throws java.io.IOException

Retrieves the body of the request as character data using a BufferedReader.

public String getRemoteAddr( )

Returns the Internet Protocol (IP) address of the client that sent the request.

public String getRemoteHost( )

Returns the fully qualified name of the client host that sent the request, or the IP address of the client if the hostname cannot be determined.

public String getRemoteUser( )

Returns the login name of the user making this request if the user has been authenticated, or null

if the user has not been authenticated.

public RequestDispatcher getRequestDispatcher(String path)

Returns a RequestDispatcher object that acts as a wrapper for the resource located at the given path. The path must be context-relative or relative to the current URL.

public String getRequestedSessionId( )

Returns the session ID specified by the client.

public String getRequestURI( )

Returns the part of this request's URL from the protocol name up to the query string in the first line of the HTTP request.

public String getScheme( )

Returns the name of the scheme (protocol) used to make this request, for example, http,

https

, or ftp.

public String getServerName( )

Returns the hostname of the server that received the request.

public int getServerPort( )

Returns the port number on which this request was received.

public String getServletPath( )

Returns the part of this request's URI that calls the servlet. For a JSP page, this is the complete context-relative path for the JSP page.

public HttpSession getSession( )

Returns the current HttpSession associated with this request. If the request does not have a session, a new HttpSession object is created, associated with the request, and returned.

public HttpSession getSession(boolean create)

Returns the current HttpSession associated with this request. If there is no current session and create is true, a new HttpSession object is created, associated with the request, and returned. If create is false and the request is not associated with a session, this method returns null.

public java.security.Principal getUserPrincipal( )

Returns a Principal object containing the name of the current authenticated user, or null if the user is not authenticated.

public boolean isRequestedSessionIdFromCookie( )

Checks whether the requested session ID came in as a cookie.

public boolean isRequestedSessionIdFromURL( )

Checks whether the requested session ID came in as part of the request URL.

public boolean isRequestedSessionIdValid( )

Checks whether the requested session ID is still valid.

public boolean isSecure( )

Returns a boolean indicating whether this request was made using a secure channel, such as

HTTPS.

public boolean isUserInRole(String role)

Returns a boolean indicating whether the authenticated user is included in the specified logical

"role." public void removeAttribute(String name)

Removes the specified attribute from this request.

public Object setAttribute(String name, Object attribute)

Stores the specified attribute in this request.

The following methods are deprecated: public String getRealPath( )

As of the Servlet 2.1 API, use ServletContext.getRealPath(String) instead.

public boolean isRequestSessionIdFromUrl( )

As of the Servlet 2.1 API, use isRequestedSessionIdFromURL( ) instead.

response

Synopsis

Variable Name:

Interface Name:

Extends:

Implemented by:

JSP Page Type: response javax.servlet.http.HttpServletResponse

javax.servlet.ServletResponse

Internal container-dependent class

Available in both regular JSP pages and error pages

Description

The response variable is assigned a reference to an internal container-dependent class that implements a protocol-dependent interface extending the javax.servlet.ServletResponse.

Since HTTP is the only protocol supported by JSP 1.1, the class always implements the javax.servlet.http.HttpServletResponse

interface. The method descriptions in this section include all methods from both interfaces.

Interface Declarations

public interface ServletResponse {

public void flushBuffer( ) throws IOException;

public int getBufferSize( );

public String getCharacterEncoding( );

public Locale getLocale( );

public ServletOutputStream getOutputStream( ) throws IOException

public PrintWriter getWriter throws IOException

public boolean isCommitted( );

public void reset( );

public void setBufferSize(int size);

public void setContentLength(int length);

public void setContentType(String type);

public void setLocale(Locale locale);

} public interface HttpServletResponse extends ServletResponse {

public void addCookie(Cookie cookie);

public void addDateHeader(String headername, long date);

public void addHeader(String headername, String value);

public void addIntHeader(String headername, int value);

public boolean containsHeader(String name);

public String encodeRedirectURL(String url);

public String encodeURL(String url);

public void sendError(int status) throws IOException;

public void sendError(int status, String message)

throws IOException;

public void sendRedirect(String location) throws IOException;

public void setDateHeader(String headername, long date);

public void setHeader(String headername, String value);

public void setIntHeader(String headername, int value);

public void setStatus(int statuscode);

// Deprecated methods

public String encodeRedirectUrl(String url);}

public String encodeUrl(String url);

public void setStatus(int statuscode, String message);

}

Methods

public void addCookie(Cookie cookie)

Adds the specified cookie to the response.

public void addDateHeader(String headername, long date)

Adds a response header with the given name and date value. The date is specified in terms of milliseconds since the epoch (January 1, 1970, 00:00:00 GMT).

public void addHeader(String headername, String value)

Adds a response header with the specified name and value.

public void addIntHeader(String headername, int value)

Adds a response header with the given name and integer value.

public boolean containsHeader(String name)

Returns a boolean indicating whether the named response header has already been set.

public String encodeRedirectURL(String url)

Encodes the specified URL for use in the sendRedirect( ) method by including the session

ID in it. If encoding (URL rewriting) is not needed, it returns the URL unchanged.

public String encodeURL(String url)

Encodes the specified URL for use in a reference element (e.g., <a>) by including the session ID in it. If encoding (URL rewriting) is not needed, it returns the URL unchanged.

public void flushBuffer( ) throws IOException

Forces any content in the response body buffer to be written to the client.

public int getBufferSize( )

Returns the actual buffer size (in bytes) used for the response. If no buffering is used, this method returns 0.

public String getCharacterEncoding( )

Returns the name of the charset used for the MIME body sent in this response.

public Locale getLocale( )

Returns the locale assigned to the response. This is either a Locale object for the server's default locale, or the Locale set with setLocale( ).

public ServletOutputStream getOutputStream( ) throws IOException

Returns a ServletOutputStream suitable for writing binary data in the response. It's recommended that this method is not used in a JSP page, since JSP pages are intended for text data.

public PrintWriter getWriter throws IOException

Returns a PrintWriter object that can send character text to the client. It's recommended that this method is not used in a JSP page, since it may interfere with the container's writer mechanism. Use the PageContext methods instead to get the current JspWriter.

public boolean isCommitted( )

Returns a boolean indicating if the response has been committed.

public void reset( )

Clears any data that exists in the buffer as well as the status code and headers. If the response has been committed, this method throws an IllegalStateException.

public void sendError(int status) throws IOException

Sends an error response to the client using the specified status. If the response has already been committed, this method throws an IllegalStateException. After using this method, the response should be considered to be committed and should not be written to.

public void sendError(int status, String message) throws IOException

Sends an error response to the client using the specified status code and a descriptive message. If the response has already been committed, this method throws an IllegalStateException.

After using this method, the response should be considered to be committed and should not be written to.

public void sendRedirect(String location) throws IOException

Sends a temporary redirect response to the client using the specified redirect location URL. This method can accept relative URLs; the servlet container converts the relative URL to an absolute

URL before sending the response to the client. If the response is already committed, this method throws an IllegalStateException. After using this method, the response should be considered to be committed and should not be written to.

public void setBufferSize(int size)

Sets the preferred buffer size (in bytes) for the body of the response. The servlet container uses a buffer at least as large as the size requested. The actual buffer size used can be found using getBufferSize( )

.

public void setContentLength(int length)

Sets the length (in bytes) of the content body in the response. In HTTP servlets, this method sets the HTTP Content-Length header. It's recommended that this method is not used in a JSP

page, since it may interfere with the container's writer mechanism.

public void setContentType(String type)

Sets the content type of the response being sent to the client.

public void setDateHeader(String headername, long date)

Sets a response header with the given name and date value. The date is specified in terms of milliseconds since the epoch (January 1, 1970, 00:00:00 GMT). If the header is already set, the new value overwrites the previous one.

public void setHeader(String headername, String value)

Sets a response header with the given name and value. If the header is already set, the new value overwrites the previous one.

public void setIntHeader(String headername, int value)

Sets a response header with the given name and integer value. If the header is already set, the new value overwrites the previous one.

public void setLocale(Locale locale)

Sets the locale of the response, setting the headers (including the Content-Type header's charset) as appropriate.

public void setStatus(int statuscode)

Sets the status code for this response. As opposed to the sendError( ) method, this method only sets the status code; it doesn't add a body or commit the response.

The following methods are deprecated: public String encodeRedirectUrl(String url)

As of the Servlet 2.1 API, use encodeRedirectURL(String url) instead.

public String encodeUrl(String url)

As of the Servlet 2.1 API, use encodeURL(String url) instead.

public void setStatus(int statuscode, String message)

Due to ambiguous meaning of the message parameter, different methods should be used as of the

Servlet 2.1 API. To set a status code, use setStatus(int); to send an error with a description, use sendError(int, String).

session

Synopsis

Variable Name: session

Interface Name: javax.servlet.http.HttpSession

Extends:

None

Implemented by: Internal container-dependent class

JSP Page Type: Available in both regular JSP pages and error pages, unless the page directive session attribute is set to false

Description

The session variable is assigned a reference to the HttpSession object that represents the current client session. Information stored as HttpSession attributes corresponds to objects in the JSP session scope.

By default, the session persists for a time period specified in the web application deployment descriptor, across more than one page request from the user. The container can maintain a session in many ways, such as using cookies or rewriting URLs.

Interface Declarations

public interface HttpSession {

public Object getAttribute(String name);

public java.util.Enumeration getAttributeNames( );

public long getCreationTime( );

public String getId( );

public long getLastAccessedTime( );

public int getMaxInactiveInterval( );

public void invalidate( );

public boolean isNew( );

public void removeAttribute(String name);

public void setAttribute(String name, Object attribute);

public void setMaxInactiveInterval(int interval);

// Deprecated methods

public HttpSessionContext getSessionContext( );

public Object getValue(String name);

public String[] getValueNames( );

public void putValue(String name, Object value);

public void removeValue(String name);

}

Methods

public Object getAttribute(String name)

Returns the object associated with the specified name in this session, or null if the object is not found.

public java.util.Enumeration getAttributeNames( )

Returns an Enumeration of String objects containing the names of all the objects in this session.

public long getCreationTime( )

Returns the time when this session was created, measured in milliseconds since the epoch

(January 1, 1970, 00:00:00 GMT).

public String getId( )

Returns a string containing the unique identifier assigned to this session.

public long getLastAccessedTime( )

Returns the last time the client sent a request associated with this session, measured in milliseconds since the epoch (January 1, 1970, 00:00:00 GMT).

public int getMaxInactiveInterval( )

Returns the maximum time interval, in seconds, that the servlet container will keep this session active between client accesses.

public void invalidate( )

Invalidates this session and unbinds any objects bound to it, calling the valueUnbound( ) methods of all objects in the session implementing the HttpSessionBindingListener interface.

public boolean isNew( )

Returns true if a request for this session has not yet been received from the client.

public void removeAttribute(String name)

Removes the object bound with the specified name from this session.

public void setAttribute(String name, Object attribute)

Associates the specified object with this session using the name specified.

public void setMaxInactiveInterval(int interval)

Specifies the time, in seconds, between client requests before the servlet container invalidates this session.

The following methods are deprecated: public HttpSessionContext getSessionContext( )

As of the Servlet 2.1 API, this method is deprecated and has no replacement.

public Object getValue(String name)

As of the Servlet 2.2 API, this method is replaced by getAttribute(String).

public String[] getValueNames( )

As of the Servlet 2.2 API, this method is replaced by getAttributeNames( ).

public void putValue(String name, Object value)

As of the Servlet 2.2 API, this method is replaced by setAttribute(String, Object).

public void removeValue(String name)

As of the Servlet 2.2 API, this method is replaced by setAttribute(String, Object).

© 2001, O'Reilly & Associates, Inc.

B.2 Servlet Classes Accessible Through Implicit Variables

This section contains descriptions of the servlet API classes that methods on the objects assigned to the implicit variables can return.

Cookie

Synopsis

Class Name: javax.servlet.http.Cookie

Extends:

Implements:

None

Cloneable

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

A Cookie object represents an HTTP cookie: a small amount of information sent by a servlet to a web browser, saved by the browser, and later sent back to the server with new requests. A cookie's value can uniquely identify a client, so cookies are commonly used for session management. A cookie has a name, a single value, and optional attributes such as comments, path and domain qualifiers, a maximum age, and a version number.

This class supports both the Version (the informal specification first introduced by Netscape) and the

Version 1 (formally defined by RFC 2109) cookie specifications. By default, cookies are created using

Version to ensure the best interoperability.

Class Summary

public class Cookie implements Cloneable {

// Constructor

public Cookie(String name, String value);

// Methods

public Object clone( );

public String getComment( );

public String getDomain( );

public int getMaxAge( );

public String getName( );

public String getPath( );

public boolean getSecure( );

public String getValue( );

public int getVersion( );

public void setComment(String comment);

public void setDomain(String domain);

public void setMaxAge(int expiry);

public void setPath(String uriPath);

public void setSecure( );

public void setValue(String value);

public void setVersion(int version);

}

Constructor

public Cookie(String name, String value)

Creates a new instance with the specified name and value.

Methods

public Object clone( )

Overrides the standard Object.clone( ) method to return a copy of this cookie.

public String getComment( )

Returns the comment describing the purpose of this cookie, or null if the cookie has no comment.

public String getDomain( )

Returns the domain name set for this cookie.

public int getMaxAge( )

Returns the maximum age of the cookie, specified in seconds. A value of -1 indicates that the cookie will persist until browser shutdown.

public String getName( )

Returns the name of the cookie.

public String getPath( )

Returns the server path to which the browser returns this cookie.

public boolean getSecure( )

Returns true if the browser is sending cookies only over a secure protocol, or false if the browser can send cookies using any protocol.

public String getValue( )

Returns the value of the cookie.

public int getVersion( )

Returns the version of the protocol this cookie complies with. A value of means that the cookie should comply with the original Netscape specification; 1 means that it should comply with

RFC 2109.

public void setComment(String comment)

Specifies a comment that describes a cookie's purpose.

public void setDomain(String domain)

Specifies the domain within which this cookie should be presented.

public void setMaxAge(int expiry)

Sets the maximum age of the cookie in seconds.

public void setPath(String uriPath)

Specifies a server path to which the client should return the cookie.

public void setSecure( )

Indicates to the browser whether the cookie should be sent only using a secure protocol, such as

HTTPS.

public void setValue(String value)

Assigns a new value to a cookie after the cookie is created.

public void setVersion(int version)

Sets the version of the cookie protocol this cookie complies with.

RequestDispatcher

Synopsis

Interface Name:

Extends:

Implemented by: javax.servlet.

RequestDispatcher

None

Internal container-dependent class

Description

The RequestDispatcher class defines an object that receives requests from the client and sends them to any resource (such as a servlet, HTML file, or JSP file) in the same web container. The container creates the RequestDispatcher object, which is used as a wrapper around a resource located at a particular URI path or identified by a particular name.

Interface Declarations

public interface RequestDispatcher {

public void forward(ServletRequest req, ServletResponse res);

public void include(ServletRequest req, ServletResponse res);

}

Methods

public void forward(ServletRequest req, ServletResponse res)

Forwards a request from a servlet to another resource (servlet, JSP file, or static file) on the server. For a RequestDispatcher obtained via the getRequestDispatcher( ) method, the ServletRequest object has its path elements and parameters adjusted to match the path of the target resource.

This method must be called before the response has been committed to the client (before response body output has been flushed). If the response has already been committed, this method throws an IllegalStateException. Uncommitted output in the response buffer is automatically cleared before the forward.

The request and response parameters must be the same objects that were passed to the calling servlet's service method.

public void include(ServletRequest req, ServletResponse res)

Includes the response generated by a resource (servlet, JSP page, static file) in the response.

The ServletRequest object's path elements and parameters remain unchanged from the caller's. The included servlet cannot change the response status code or set headers; any attempt to make a change is ignored.

The request and response parameters must be the same objects that were passed to the calling servlet's service method.

© 2001, O'Reilly & Associates, Inc.

B.3 Tag Extension Classes

The JSP specification defines a number of classes and interfaces in the javax.servlet.jsp.tagext

package. These classes are used to develop tag handler classes for

JSP custom actions. This section contains descriptions of each class and interface. Chapter 16 and

Chapter

17

show examples of how you can use these classes and interfaces to develop custom actions.

BodyContent

Synopsis

Class Name:

Extends:

Implements:

Implemented by: javax.servlet.jsp.tagext.BodyContent

javax.servlet.jsp.JspWriter

None

Internal container-dependent class

Description

The container creates an instance of the BodyContent class to encapsulate the element body of a custom action element if the corresponding tag handler implements the BodyTag interface. The container makes the BodyContent instance available to the tag handler by calling the setBodyContent( )

method, so the tag handler can process the body content.

Class Summary

public abstract class BodyContent extends JspWriter {

// Constructor

protected BodyContent(JspWriter e);

// Methods

public void clearBody( );

public void flush( ) throws java.io.IOException;

public JspWriter getEnclosingWriter( );

public abstract java.io.Reader getReader( );

public abstract String getString( );

public abstract void writeOut(java.io.Writer out)

throws java.io.IOException;

}

Constructor

protected BodyContent(JspWriter e)

Creates a new instance with the specified JspWriter as the enclosing writer.

Methods

public void clearBody( )

Removes all buffered content for this instance.

public void flush( ) throws java.io.IOException

Overwrites the behavior inherited from JspWriter to always throw an IOException, since it's invalid to flush a BodyContent instance.

public JspWriter getEnclosingWriter( )

Returns the enclosing JspWriter, which is either the top level JspWriter or the

JspWriter

(BodyContent subclass) of the parent tag handler.

public abstract java.io.Reader getReader( )

Returns the value of this BodyContent as a Reader with the content produced by evaluating the element's body.

public abstract String getString( )

Returns the value of this BodyContent as a String with the content produced by evaluating the element's body.

public abstract void writeOut(java.io.Writer out) throws

java.io.IOException

Writes the content of this BodyContent into a Writer.

BodyTag

Synopsis

Interface Name: javax.servlet.jsp.tagext.

BodyTag

Extends: javax.servlet.jsp.tagext.Tag

Implemented by: Custom action tag handler classes and javax.servlet.jsp.tagext.BodyTagSupport

Description

The BodyTag interface must be implemented by tag handler classes that need access to the body contents of the corresponding custom action element, perhaps to perform a transformation of the contents before it's included in the response. This interface must also be implemented by tag handlers that need to iterate over the body of a custom action element.

Interface Declarations

public interface BodyTag extends Tag {

// Constants

public static final int EVAL_BODY_TAG;

// Methods

public int doAfterBody( ) throws JspException;

public void doInitBody( ) throws JspException;

public void setBodyContent(BodyContent b);

}

Methods

public int doAfterBody( ) throws JspException

Performs actions after the body has been evaluated. It is invoked after every body evaluation. If this method returns EVAL_BODY_TAG, the body is evaluated again, typically after changing the value of variables used in the body. If it returns SKIP_BODY, the processing continues with a call to doEndTag( ).

If the element body is empty, or if doStartTag( ) returns SKIP_BODY, this method is not invoked.

public void doInitBody( ) throws JspException

Prepares for evaluation of the body. This method is invoked once per action invocation by the page implementation, after a new BodyContent has been obtained and set on the tag handler via the setBodyContent( )

method, just before the evaluation of the element's body.

If the element body is empty, or if doStartTag( ) returns SKIP_BODY, this method is not invoked.

public void setBodyContent(BodyContent b)

Sets the BodyContent created for this tag handler. If the element body is empty, or if doStartTag( )

returns SKIP_BODY, this method is not invoked.

BodyTagSupport

Synopsis

Class Name: javax.servlet.jsp.tagext.BodyTagSupport

Extends: javax.servlet.jsp.tagext.TagSupport

Implements:

BodyTag

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

BodyTagSupport

is a support class that provides default implementations of all BodyTag interface methods. It's intended to be used as a superclass for tag handlers that need access to the body contents of the corresponding custom action element.

Class Summary

public class BodyTagSupport extends TagSupport implements BodyTag {

// Constructor

public BodyTagSupport( );

// Methods

public int doAfterBody( ) throws JspException;

public int doEndTag( ) throws JspException;

public void doInitBody( );

public BodyContent getBodyContent( );

public JspWriter getPreviousOut( );

public void release( );

public void setBodyContent(BodyContent b);

}

Constructor

public BodyTagSupport( )

Creates a new BodyTagSupport instance.

Methods

public int doAfterBody( ) throws JspException

Returns SKIP_BODY.

public int doEndTag( ) throws JspException

Returns EVAL_PAGE.

public void doInitBody( )

This method currently does nothing. You should override this method in a class that extends

BodyTagSupport

to perform initialization.

public BodyContent getBodyContent( )

Returns the BodyContent object assigned to this instance.

public JspWriter getPreviousOut( )

Returns the enclosing writer of the BodyContent object assigned to this instance.

public void release( )

Removes the references to all objects held by this instance.

public void setBodyContent(BodyContent b)

Saves a reference to the assigned BodyContent as an instance variable.

Tag

Synopsis

Interface Name: javax.servlet.jsp.tagext.Tag

Extends:

None

Implemented by: Custom action tag handler classes and javax.servlet.jsp.tagext.TagSupport

Description

The Tag interface should be implemented by tag handler classes that do not need access to the body contents of the corresponding custom action element, and that do not need to iterate over the body of a custom action element.

Interface Declarations

public interface Tag {

// Constants

public static final int EVAL_BODY_INCLUDE;

public static final int EVAL_PAGE;

public static final int SKIP_BODY;

public static final int SKIP_PAGE;

// Methods

public int doEndTag( ) throws JspException;

public int doStartTag( ) throws JspException;

public Tag getParent( );

public void release( );

public void setPageContext(PageContext pc);

public void setParent(Tag t)

}

Methods

public int doEndTag( ) throws JspException

Performs actions when the end tag is encountered. If this method returns SKIP_PAGE, execution of the rest of the page is aborted and the _jspService( ) method of JSP page implementation class returns. If EVAL_PAGE is returned, the code following the custom action in the

_jspService( )

method is executed.

public int doStartTag( ) throws JspException

Performs actions when the start tag is encountered. This method is called by the JSP container after all property setter methods have been called. The return value from this method controls how the action's body is handled, if there is one. If it returns EVAL_BODY_INCLUDE, the JSP container evaluates the body and processes possible JSP elements. The result of the evaluation is then added to the response. If SKIP_BODY is returned, the body is ignored.

A tag handler class that implements the BodyTag interface (extending the Tag interface) can return EVAL_BODY_TAG instead of EVAL_BODY_INCLUDE. The JSP container then creates a

BodyContent

instance and makes it available to the tag handler for special processing.

public Tag getParent( )

Returns the tag handler's parent (the Tag instance for the enclosing action element, if any) or null

if the tag handler doesn't have a parent.

public void release( )

Removes the references to all objects held by this instance.

public void setPageContext(PageContext pc)

Saves a reference to the current PageContext.

public void setParent(Tag t)

Saves a reference to the tag handler's parent (the Tag instance for the enclosing action element).

TagAttributeInfo

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagAttributeInfo

Extends: None

Implements: None

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

TagAttributeInfo

instances are created by the JSP container to provide information found in the

Tag Library Descriptor (TLD) about each attribute supported by a custom action. It's primarily intended to be used by the JSP container itself during the translation phase.

Class Summary

public class TagAttributeInfo {

// Constructor

public TagAttributeInfo(String name, boolean required,

boolean rtexprvalue, String type, boolean reqTime);

// Methods

public boolean canBeRequestTime( );

public static TagAttributeInfo getIdAttribute(TagAttributeInfo[] a);

public String getName( );

public String getTypeName( );

public boolean isRequired( );

public String toString( );

}

Constructor

public TagAttributeInfo(String name, boolean required, boolean rtexprvalue, String type, boolean reqTime)

Creates a new instance with the specified information from the TLD. Instances of this class should be created only by the JSP container.

Methods

public boolean canBeRequestTime( )

Returns true if a request time attribute value can be used for this attribute.

public

static TagAttributeInfo getIdAttribute(TagAttributeInfo[] a)

Convenience method that returns the TagAttributeInfo instance in the specified array that represents an attribute named id, or null if not found.

public String getName( )

Returns the attribute name.

public String getTypeName( )

Returns the attribute's Java type (a fully qualified class or interface name).

public boolean isRequired( )

Returns true if this attribute is required, false otherwise.

public String toString( )

Returns a String representation of the attribute info.

TagData

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagData

Extends: None

Cloneable

Implements:

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

TagData

instances are created by the JSP container during the translation phase, and provide information about the attribute values specified for a custom action to the TagExtraInfo subclass for the corresponding tag handler, if any.

Class Summary

public class TagData implements Cloneable {

// Constants

public static final Object REQUEST_TIME_VALUE;

// Constructor

public TagData(Object[][] atts);

public TagData(java.util.Hashtable attrs);

// Methods

public Object getAttribute(String attName);

public String getAttributeString(String attName);

public String getId( );

public void setAttribute(String attName, Object value);

}

Constructors

public TagData(Object[][] atts)

Creates a new instance with the attribute name/value pairs specified by the Object[][].

Element of each Object[] contains the name, and Element 1 contains the value or

REQUEST_TIME_VALUE

, if the attribute value is defined as a request time value (a JSP expression).

public TagData(java.util.Hashtable attrs)

Creates a new instance with the attribute name/value pairs specified by the Hashtable.

Methods

public Object getAttribute(String attName)

Returns the specified attribute value as a String or as the REQUEST_TIME_VALUE Object, if the attribute value is defined as a request time value (a JSP expression).

public String getAttributeString(String attName)

Returns the specified attribute value as a String. A ClassCastException is thrown if the attribute value is defined as a request time value (a JSP expression).

public String getId( )

Returns the attribute named id as a String, or null if not found.

public void setAttribute(String attName, Object value)

Sets the specified attribute to the specified value.

TagExtraInfo

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagExtraInfo

Extends:

Implements:

None

None

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

For custom actions that create scripting variables or require additional translation time validation of the tag attributes, a subclass of the TagExtraInfo class must be developed for the custom action and declared in the Tag Library Descriptor. The JSP container creates an instance of the TagExtraInfo subclass during the translation phase.

Class Summary

public abstract class TagExtraInfo {

// Constructor

public TagExtraInfo( );

// Methods

public TagInfo getTagInfo( );

public VariableInfo[] getVariableInfo(TagData data);

public boolean isValid(TagData data);

public void setTagInfo(TagInfo tagInfo);

}

Constructor

public TagExtraInfo( )

Creates a new TagExtraInfo instance.

Methods

public TagInfo getTagInfo( )

Returns the TagInfo instance for the custom action associated with this TagExtraInfo

instance. The TagInfo instance is set by the setTagInfo( ) method (called by the container).

public VariableInfo[] getVariableInfo(TagData data)

Returns a VariableInfo[] with information about scripting variables created by the tag handler class associated with this TagExtraInfo instance. The default implementation returns an empty array. A subclass must override this method if the corresponding tag handler creates scripting variables.

public boolean isValid(TagData data)

Returns true if the set of attribute values specified for the custom action associated with this

TagExtraInfo

instance is valid, false otherwise. The default implementation returns true.

A subclass can override this method if the validation performed by the JSP container based on the

Tag Library Descriptor information is not enough.

public void setTagInfo(TagInfo tagInfo)

Sets the TagInfo for this instance. This method is called by the JSP container before any of the other methods are called.

TagInfo

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagInfo

Extends: None

Implements: None

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

TagInfo

instances are created by the JSP container to provide information found in the Tag Library

Descriptor (TLD) about a custom action, as well as information about the attribute values used in a JSP page for an instance of the custom action. It's primarily intended to be used by the JSP container itself during the translation phase.

Class Summary

public class TagInfo {

// Constants

public static final String BODY_CONTENT_EMPTY;

public static final String BODY_CONTENT_JSP;

public static final String BODY_CONTENT_TAG_DEPENDENT;

// Constructor

public TagInfo(String tagName, String tagClassName,

String bodycontent, String infoString, TagLibraryInfo taglib,

TagExtraInfo tagExtraInfo, TagAttributeInfo[] attributeInfo);

// Methods

public TagAttributeInfo[] getAttributes( );

public String getBodyContent( );

public String getInfoString( );

public String getTagClassName( );

public TagExtraInfo getTagExtraInfo( );

public TagLibraryInfo getTagLibrary( );

public String getTagName( );

public VariableInfo[] getVariableInfo(TagData data);

public boolean isValid(TagData data);

public String toString( );

}

Constructor

public TagInfo(String tagName, String tagClassName, String bodycontent,

String infoString, TagLibraryInfo taglib, TagExtraInfo tagExtraInfo,

TagAttributeInfo[] attributeInfo)

Creates a new instance with the specified values.

Methods

public TagAttributeInfo[] getAttributes( )

Returns information from the TLD about all attribute values, or null if no attributes are declared.

public String getBodyContent( )

Returns one of BODY_CONTENT_EMPTY, BODY_CONTENT_JSP, or

BODY_CONTENT_TAG_DEPENDENT

, based on the value in the TLD.

public String getInfoString( )

Returns the tag information string from the TLD, or null if there is no info.

public String getTagClassName( )

Returns the tag handler class name declared in the TLD.

public TagExtraInfo getTagExtraInfo( )

Returns an instance of the TagExtraInfo subclass for the tag, or null if no class is declared in the TLD.

public TagLibraryInfo getTagLibrary( )

Returns a TagLibraryInfo instance for the library the tag is part of.

public String getTagName( )

Returns the name for the tag declared in the TLD.

public VariableInfo[] getVariableInfo(TagData data)

Returns information about scripting variables created by the tag handler, or null if no variables are created. This information is obtained from the TagExtraInfo for the tag, if any.

public boolean isValid(TagData data)

Returns true if the set of attributes specified for the custom action associated with this

TagExtraInfo

instance is valid, false otherwise. This information is obtained from the

TagExtraInfo

for the tag, if any.

public String toString( )

Returns a String representation of all information held by the instance.

TagLibraryInfo

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagLibraryInfo

Extends: None

Implements: None

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

TagLibraryInfo

instances are created by the JSP container to provide information found in the Tag

Library Descriptor (TLD) about a tag library, as well as information from the taglib directive used in a

JSP page. It's primarily intended to be used by the JSP container itself during the translation phase.

Class Summary

public abstract class TagLibraryInfo {

// Constructor

protected TagLibraryInfo(String prefix, String uri);

// Methods

public String getInfoString( );

public String getPrefixString( );

public String getReliableURN( );

public String getRequiredVersion( );

public String getShortName( );

public TagInfo getTag(String shortname);

public TagInfo[] getTags( );

public String getURI( );

}

Constructor

protected TagLibraryInfo(String prefix, String uri)

Creates a new instance with the specified prefix and URI (from the taglib directive in the JSP page).

Methods

public java.lang.String getInfoString( )

Returns the information string from the TLD for the library.

public String getPrefixString( )

Returns the prefix assigned by the taglib directive for the library.

public String getReliableURN( )

Returns the URI value from the TLD for the library.

public String getRequiredVersion( )

Returns the required JSP version from the TLD for the library.

public String getShortName( )

Returns the short name from the TLD for the library.

public TagInfo getTag(String shortname)

Returns a TagInfo instance for the specified tag in the library.

public TagInfo[] getTags( )

Returns a TagInfo[] for all tags in the library.

public String getURI( )

Returns the URI assigned by the taglib directive for the library.

TagSupport

Synopsis

Class Name: javax.servlet.jsp.tagext.

TagSupport

Extends:

Implements:

None

Tag, java.io.Serializable

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).

Description

TagSupport

is a support class that provides default implementations of all Tag interface methods. It's intended to be used as a superclass for tag handlers that do not need access to the body contents of the corresponding custom action element.

Class Summary

public class TagSupport implements Tag, java.io.Serializable {

// Constructor

public TagSupport( );

// Methods

public int doEndTag( ) throws JspException;

public int doStartTag( ) throws JspException;

public static final Tag findAncestorWithClass(Tag from, Class klass);

public String getId( );

public Tag getParent( );

public Object getValue(String k);

public java.util.Enumeration getValues( );

public void release( );

public void removeValue(String k);

public void setPageContext(PageContext pageContext);

public void setId(String id);

public void setParent(Tag t);

public void setValue(String k, Object o);

}

Constructor

public TagSupport( )

Creates a new instance with the specified name and value.

Methods

public int doEndTag( ) throws JspException

Returns EVAL_PAGE. public int doStartTag( ) throws JspException

Returns SKIP_BODY.

public static final Tag findAncestorWithClass(Tag from, Class klass)

Returns the instance of the specified class, found by testing for a match of each parent in a tag handler nesting structure (corresponding to nested action elements) starting with the specified Tag instance, or null if not found.

public String getId( )

Returns the id attribute value, or null if not set.

public Tag getParent( )

Returns the parent of this Tag instance (representing the action element that contains the action element corresponding to this Tag instance), or null if the instance has no parent (at the top level in the JSP page).

public Object getValue(String k)

Returns the value for the specified attribute that has been set with the setValue( ) method, or null

if not found.

public java.util.Enumeration getValues( )

Returns an Enumeration of all attribute names for values set with the setValue( ) method.

public void release( )

Removes the references to all objects held by this instance.

public void removeValue(String k)

Removes a value set with the setValue( ) method.

public void setPageContext(PageContext pageContext)

Saves a reference to the current PageContext.

public void setId(String id)

Sets the id attribute value.

public void setParent(Tag t)

Saves a reference to the parent for this instance.

public void setValue(String k, Object o)

Saves the specified attribute with the specified value. Subclasses can use this method to save attribute values as an alternative to instance variables.

VariableInfo

Synopsis

Class Name: javax.servlet.jsp.tagext.

VariableInfo

Extends: None

Implements: None

Implemented by: Internal container-dependent class. Most containers use the reference implementation of the class

(developed in the Apache Jakarta project).