Pawn The Language

embedded scripting language
The Language
June 2011
ITB CompuPhase
“Java” is a trademark of Sun Microsystems, Inc.
“Microsoft” and “Microsoft Windows” are registered trademarks of Microsoft Corporation.
“Linux” is a registered trademark of Linus Torvalds.
“CompuPhase” is a registered trademark of ITB CompuPhase.
“Unicode” is a registered trademark of Unicode, Inc.
c 1997–2011, ITB CompuPhase
Copyright Eerste Industriestraat 19–21, 1401VL Bussum The Netherlands (Pays Bas)
telephone: (+31)-(0)35 6939 261
e-mail:, www:
The documentation is licensed under the Creative Commons Attribution-ShareAlike 2.5 License.
A summary of this license is in appendix D. For more information on this licence, visit
or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305,
The information in this manual and the associated software are provided “as is”. There are no
guarantees, explicit or implied, that the software and the manual are accurate. Requests for
corrections and additions to the manual and the software can be directed to ITB CompuPhase
at the above address.
Typeset with TEX in the “Computer Modern” and “Palatino” typefaces at a base size of 11
Table of contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
A tutorial introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Data and declarations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
The preprocessor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
General syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Operators and expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Directives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Proposed function library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Pitfalls: differences from C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Assorted tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
A: Error and warning messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
B: The compiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
C: Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
D: License . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Table of contents
pawn is a simple, typeless, 32-bit “scripting” language with a C-like syntax.
Execution speed, stability, simplicity and a small footprint were essential
design criteria for both the language and the interpreter/abstract machine
that a pawn program runs on.
An application or tool cannot do or be everything for all users. This not
only justifies the diversity of editors, compilers, operating systems and many
other software systems, it also explains the presence of extensive configuration options and macro or scripting languages in applications. My own applications have contained a variety of little languages; most were very simple,
some were extensive. . . and most needs could have been solved by a general
purpose language with a special purpose library. Hence, pawn.
The pawn language was designed as a flexible language for manipulating
objects in a host application. The tool set (compiler, abstract machine)
were written so that they were easily extensible and would run on different
software/hardware architectures.
pawn is a descendent of the original Small C by Ron Cain and James Hendrix, which at its turn was a subset of C. Some of the modifications that
I did to Small C, e.g. the removal of the type system and the substitution
of pointers by references, were so fundamental that I could hardly call my
language a “subset of C” or a “C dialect” any more. Therefore, I stripped
off the “C” from the title and used the name “Small” for the name of the
language in my publication in Dr. Dobb’s Journal and the years since. During development and maintenance of the product, I received many requests
for changes. One of the frequently requested changes was to use a different
name for the language —searching for information on the Small scripting
language on the Internet was hindered by “small” being such a common
word. The name change occurred together with a significant change in the
language: the support of “states” (and state machines).
I am indebted to Ron Cain and James Hendrix (and more recently, Andy
Yuen), and to Dr. Dobb’s Journal to get this ball rolling. Although I must
have touched nearly every line of the original code multiple times, the Small
C origins are still clearly visible.
A detailed treatise of the design goals and compromises is in appendix C;
here I would like to summarize a few key points. As written in the previous
paragraphs, pawn is for customizing applications (by writing scripts), not
for writing applications. pawn is weak on data structuring because pawn
programs are intended to manipulate objects (text, sprites, streams, queries,
. . . ) in the host application, but the pawn program is, by intent, denied
direct access to any data outside its abstract machine. The only means that
a pawn program has to manipulate objects in the host application is by
calling subroutines, so called “native functions”, that the host application
pawn is flexible in that key area: calling functions. pawn supports default values for any of the arguments of a function (not just the last), callby-reference as well as call-by-value, and “named” as well as “positional”
function arguments. pawn does not have a “type checking” mechanism,
by virtue of being a typeless language, but it does offer in replacement a
“classification checking” mechanism, called “tags”. The tag system is especially convenient for function arguments because each argument may specify
multiple acceptable tags.
For any language, the power (or weakness) lies not in the individual features,
but in their combination. For pawn, I feel that the combination of named
arguments —which lets you specify function arguments in any order, and
default values —which allows you to skip specifying arguments that you are
not interested in, blend together to a convenient and “descriptive” way to
call (native) functions to manipulate objects in the host application.
A tutorial introduction
pawn is a simple programming language with a syntax reminiscent to the
“C” programming language. A pawn program consists of a set of functions
and a set of variables. The variables are data objects and the functions
contain instructions (called “statements”) that operate on the data objects
or that perform tasks.
The first program in almost any computer language is one that prints a
simple string; printing “Hello world” is a classic example. In pawn, the
program would look like:
Compiling and
running scripts:
see 169
printf "Hello world\n"
This manual assumes that you know how to run a pawn program; if not,
please consult the application manual (more hints are at page 169).
A pawn program starts execution in an “entry” function∗ —in nearly all
examples of this manual, this entry function is called “main”. Here, the
function main contains only a single instruction, which is at the line below
the function head itself. Line breaks and indenting are insignificant; the
invocation of the function print could equally well be on the same line as
the head of function main.
The definition of a function requires that a pair of parentheses follow the
function name. If a function takes parameters, their declarations appear
between the parentheses. The function main does not take any parentheses. The rules are different for a function invocation (or a function call);
parentheses are optional in the call to the print function.
The single argument of the print function is a string, which must be enclosed
in double quotes. The characters \n near the end of the string form an
escape sequence, in this case they indicate a “newline” symbol. When print
encounters the newline escape sequence, it advances the cursor to the first
column of the next line. One has to use the \n escape sequence to insert a
“newline” into the string, because a string may not wrap over multiple lines.
This should not be confused with the “state” entry functions, which are called entry, but
serve a different purpose —see page 40.
String literals: 99
Escape sequence:
A tutorial introduction
pawn is a “case sensitive” language: upper and lower case letters are considered to be different letters. It would be an error to spell the function printf
in the above example as “PrintF”. Keywords and predefined symbols, like
the name of function “main”, must be typed in lower case.
If you know the C language, you may feel that the above example does not
look much like the equivalent “Hello world” program in C/C++. pawn can
also look very similar to C, though. The next example program is also valid
pawn syntax (and it has the same semantics as the earlier example):
Listing: hello.p — C style
#include <console>
printf("Hello world\n");
These first examples also reveal a few differences between pawn and the C
⋄ there is usually no need to include any system-defined “header file”;
⋄ semicolons are optional (except when writing multiple statements on one
⋄ when the body of a function is a single instruction, the braces (for a
compound instruction) are optional;
⋄ when you do not use the result of a function in an expression or assignment,
parentheses around the function argument are optional.
As an aside, the few preceding points refer to optional syntaxes. It is your
choice what syntax you wish to use: neither style is “deprecated” or “considered harmful”. The examples in this manual position the braces and use
an indentation that is known as the “Whitesmith’s style”, but pawn is a
free format language and other indenting styles are just as good.
More function
descriptions at
page 123
Because pawn is designed to be an extension language for applications, the
function set/library that a pawn program has at its disposal depends on the
host application. As a result, the pawn language has no intrinsic knowledge of any function. The print function, used in this first example, must
be made available by the host application and be “declared” to the pawn
parser.† It is assumed, however, that all host applications provide a minimal
In the language specification, the term “parser” refers to any implementation that processes and runs on conforming Pawn programs —either interpreters or compilers.
A tutorial introduction
set of common functions, like print and printf.
In some environments, the display or terminal must be enabled before any
text can be output onto it. If this is the case, you must add a call to the
function “console” before the first call to function print or printf. The
console function also allows you to specify device characteristics, such as
the number of lines and columns of the display. The example programs in
this manual do not use the console functions, because many platforms do
not require or provide it.
• Arithmetic
Fundamental elements of most programs are calculations, decisions (conditional execution), iterations (loops) and variables to store input data, output
data and intermediate results. The next program example illustrates many
of these concepts. The program calculates the greatest common divisor of
two values using an algorithm invented by Euclides.
Listing: gcd.p
The greatest common divisor of two values,
using Euclides’ algorithm .
print "Input two values\n"
new a = getvalue()
new b = getvalue()
while (a != b)
if (a > b)
a = a - b
b = b - a
printf "The greatest common divisor is %d\n", a
Function main now contains more than just a single “print” statement. When
the body of a function contains more than one statement, these statements
must be embodied in braces —the “{” and “}” characters. This groups
the instructions to a single compound statement. The notion of grouping
statements in a compound statement applies as well to the bodies of if–
else and loop instructions.
The new keyword creates a variable. The name of the variable follows new.
It is common, but not imperative, to assign a value to the variable already at
Compound statement: 111
Data declarations
are covered in
detail starting at
page 60
A tutorial introduction
the moment of its creation. Variables must be declared before they are used
in an expression. The getvalue function (also common predefined function)
reads in a value from the keyboard and returns the result. Note that pawn
is a typeless language, all variables are numeric cells that can hold a signed
integral value.
The getvalue function name is followed by a pair of parentheses. These
are required because the value that getvalue returns is stored in a variable.
Normally, the function’s arguments (or parameters) would appear between
the parentheses, but getvalue (as used in this program) does not take any
explicit arguments. If you do not assign the result of a function to a variable
or use it in a expression in another way, the parentheses are optional. For
example, the result of the print and printf statements are not used. You
may still use parentheses around the arguments, but it is not required.
“while” loop: 115
“if–else”: 113
Loop instructions, like “while”, repeat a single instruction as long as the
loop condition (the expression between parentheses) is “true”. One can
execute multiple instructions in a loop by grouping them in a compound
statement. The if–else instruction has one instruction for the “true” clause
and one for the “false”.
Observe that some statements, like while and if–else, contain (or “fold
around”) another instruction —in the case of if–else even two other instructions. The complete bundle is, again, a single instruction. That is:
⋄ the assignment statements “a = a - b” below the if and “b = b - a”
below the else are statements;
⋄ the if–else statement folds around these two assignment statements and
forms a single statement of itself;
⋄ the while statement folds around the if–else statement and forms, again,
a single statement.
It is common to make the nesting of the statements explicit by indenting
any sub-statements below a statement in the source text. In the “Greatest
Common Divisor” example, the left margin indent increases by four space
characters after the while statement, and again after the if and else keywords. Statements that belong to the same level, such as both printf
invocations and the while loop, have the same indentation.
Relational operators: 106
The loop condition for the while loop is “(a != b)”; the symbol != is the
“not equal to” operator. That is, the if–else instruction is repeated until
“a” equals “b”. It is good practice to indent the instructions that run under
control of another statement, as is done in the preceding example.
A tutorial introduction
The call to printf, near the bottom of the example, differs from the print
call right below the opening brace (“{”). The “f” in printf stands for
“formatted”, which means that the function can format and print numeric
values and other data (in a user-specified format), as well as literal text.
The %d symbol in the string is a token that indicates the position and the
format that the subsequent argument to function printf should be printed.
At run time, the token %d is replaced by the value of variable “a” (the second
argument of printf).
Function print can only print text; it is quicker than printf. If you want
to print a literal “%” at the display, you have to use print, or you have to
double it in the string that you give to printf. That is:
print "20% of the personnel accounts for 80% of the costs\n"
printf "20%% of the personnel accounts for 80%% of the costs\n"
print the same string.
• Arrays & constants
Next to simple variables with a size of a single cell, pawn supports “array variables” that hold many cells/values. The following example program
displays a series of prime numbers using the well known “sieve of Eratosthenes”. The program also introduces another new concept: symbolic constants. Symbolic constants look like variables, but they cannot be changed.
/* Print all primes below 100, using the "Sieve of Eratosthenes" */
const max_primes = 100
new series[max_primes] = { true, ... }
for (new i = 2; i < max_primes; ++i)
if (series[i])
printf "%d ", i
/* filter all multiples of this "prime" from the list */
for (new j = 2 * i; j < max_primes; j += i)
series[j] = false
Constant declaration: 101
Progressive initiallers: 63
“for” loop: 112
An overview of all
operators: 103
A tutorial introduction
When a program or sub-program has some fixed limit built-in, it is good
practice create a symbolic constant for it. In the preceding example, the
symbol max_primes is a constant with the value 100. The program uses the
symbol max_primes three times after its definition: in the declaration of the
variable series and in both for loops. If we were to adapt the program to
print all primes below 500, there is now only one line to change.
Like simple variables, arrays may be initialized upon creation. pawn offers a
convenient shorthand to initialize all elements to a fixed value: all hundred
elements of the “series” array are set to true —without requiring that the
programmer types in the word “true” a hundred times. The symbols true
and false are predefined constants.
When a simple variable, like the variables i and j in the primes sieve example, is declared in the first expression of a for loop, the variable is valid only
inside the loop. Variable declaration has its own rules; it is not a statement
—although it looks like one. One of those rules is that the first expression
of a for loop may contain a variable declaration.
Both for loops also introduce new operators in their third expression. The
++ operator increments its operand by one; meaning that, ++i is equal to
i = i + 1. The += operator adds the expression on its right to the variable
on its left; that is, j += i is equal to j = j + i.
There is an “off-by-one” issue that you need to be aware if when working with arrays. The first element in the series array is series[0], so
if the array holds max_primes elements, the last element in the array is
series[max_primes-1]. If max_primes is 100, the last element, then, is
series[99]. Accessing series[100] is invalid.
• Functions
Larger programs separate tasks and operations into functions. Using functions increases the modularity of programs and functions, when well written,
are portable to other programs. The following example implements a function to calculate numbers from the Fibonacci series.
The Fibonacci sequence was discovered by Leonardo “Fibonacci” of Pisa, an
Italian mathematician of the 13th century—whose greatest achievement was
popularizing for the Western world the Hindu-Arabic numerals. The goal
of the sequence was to describe the growth of a population of (idealized)
A tutorial introduction
rabbits; and the sequence is 1, 1, 2, 3, 5, 8, 13, 21,. . . (every next value is
the sum of its two predecessors).
/* Calculation of Fibonacci numbers by iteration */
print "Enter a value: "
new v = getvalue()
if (v > 0)
printf "The value of Fibonacci number %d is %d\n",
v, fibonacci(v)
printf "The Fibonacci number %d does not exist\n", v
assert n > 0
new a = 0, b = 1
for (new i = 2; i < n; i++)
new c = a + b
a = b
b = c
return a + b
The assert instruction at the top of the fibonacci function deserves explicit mention; it guards against “impossible” or invalid conditions. A negative Fibonacci number is invalid, and the assert statement flags it as a
programmer’s error if this case ever occurs. Assertions should only flag programmer’s errors, never user input errors.
The implementation of a user-defined function is not much different than that
of function main. Function fibonacci shows two new concepts, though: it
receives an input value through a parameter and it returns a value (it has a
Function parameters are declared in the function header; the single parameter in this example is “n”. Inside the function, a parameter behaves as a
local variable, but one whose value is passed from the outside at the call to
the function.
The return statement ends a function and sets the result of the function. It
need not appear at the very end of the function; early exits are permitted.
“assert” statement: 111
Functions: properties & features:
Native function
interface: 83
A tutorial introduction
The main function of the Fibonacci example calls predefined “native” functions, like getvalue and printf, as well as the user-defined function fibonacci. From the perspective of calling a function (as in function main),
there is no difference between user-defined and native functions.
The Fibonacci numbers sequence describes a surprising variety of natural
phenomena. For example, the two or three sets of spirals in pineapples, pine
cones and sunflowers usually have consecutive Fibonacci numbers between
5 and 89 as their number of spirals. The numbers that occur naturally
in branching patterns (e.g. that of plants) are indeed Fibonacci numbers.
Finally, although the Fibonacci sequence is not a geometric sequence, the
further the sequence is extended, the more closely the ratio between successive terms approaches the Golden Ratio, of 1.618. . . ∗ that appears so often
in art and architecture.
• Call-by-reference & call-by-value
Dates are a particularly rich source of algorithms and conversion routines,
because the calenders that a date refers to have known such a diversity,
through time and around the world.
The “Julian Day Number” is attributed to Josephus Scaliger† and it counts
the number of days since November 24, 4714 BC (proleptic Gregorian calendar‡ ). Scaliger chose that date because it marked the coincidence of three
well-established cycles: the 28-year Solar Cycle (of the old Julian calendar),
the 19-year Metonic Cycle and the 15-year Indiction Cycle (periodic taxes
or governmental requisitions in ancient Rome), and because no literature or
recorded history was known to pre-date that particular date in the remote
past. Scaliger used this concept to reconcile dates in historic documents,
The exact value for the Golden Ratio is 1/2( 5 + 1). The relation between Fibonacci
numbers and the Golden Ratio also allows for a “direct” calculation of any sequence
number, instead of the iterative method described here.
There is some debate on exactly what Josephus Scaliger invented and who or what he
called it after.
The Gregorian calendar was decreed to start on 15 October 1582 by pope Gregory XIII,
which means that earlier dates do not really exist in the Gregorian calendar. When
extending the Gregorian calendar to days before 15 October 1582, we refer to it as the
proleptic Gregorian calendar.
A tutorial introduction
later astronomers embraced it to calculate intervals between two events more
Julian Day numbers (sometimes denoted with unit “jd”) should not be confused with Julian Dates (the number of days since the start of the same
year), or with the Julian calendar that was introduced by Julius Caesar.
Below is a program that calculates the Julian Day number from a date in the
(proleptic) Gregorian calendar, and vice versa. Note that in the proleptic
Gregorian calendar, the first year is 1 AD (Anno Domini) and the year
before that is 1 BC (Before Christ): year zero does not exist! The program
uses negative year values for BC years and positive (non-zero) values for AD
/* calculate Julian Day number from a date, and vice versa */
new d, m, y, jdn
print "Give a date (dd-mm-yyyy): "
d = getvalue(_, ’-’, ’/’)
m = getvalue(_, ’-’, ’/’)
y = getvalue()
jdn = DateToJulian(d, m, y)
printf("Date %d/%d/%d = %d JD\n", d, m, y, jdn)
print "Give a Julian Day Number: "
jdn = getvalue()
JulianToDate jdn, d, m, y
printf "%d JD = %d/%d/%d\n", jdn, d, m, y
DateToJulian(day, month, year)
/* The first year is 1. Year 0 does not exist: it is 1 BC (or -1) */
assert year != 0
if (year < 0)
/* move January and February to the end of the previous year */
if (month <= 2)
year--, month += 12
new jdn = 365*year + year/4 - year/100 + year/400
+ (153*month - 457) / 5
+ day + 1721119
return jdn
A tutorial introduction
JulianToDate(jdn, &day, &month, &year)
jdn -= 1721119
/* approximate year, then adjust in a loop */
year = (400 * jdn) / 146097
while (365*year + year/4 - year/100 + year/400 < jdn)
year-/* determine month */
jdn -= 365*year + year/4 - year/100 + year/400
month = (5*jdn + 457) / 153
/* determine day */
day = jdn - (153*month - 457) / 5
/* move January and February to start of the year */
if (month > 12)
month -= 12, year++
/* adjust negative years (year 0 must become 1 BC, or -1) */
if (year <= 0)
Function main starts with creating variables to hold the day, month and
year, and the calculated Julian Day number. Then it reads in a date —three
calls to getvalue— and calls function DateToJulian to calculate the day
number. After calculating the result, main prints the date that you entered
and the Julian Day number for that date. Now, let us focus on function
DateToJulian. . .
“Call by value”
versus “call by
reference”: 70
Near the top of function DateToJulian, it increments the year value if
it is negative; it does this to cope with the absence of a “zero” year in
the proleptic Gregorian calendar. In other words, function DateToJulian
modifies its function arguments (later, it also modifies month). Inside a
function, an argument behaves like a local variable: you may modify it.
These modifications remain local to the function DateToJulian, however.
Function main passes the values of d, m and y into DateToJulian, who maps
them to its function arguments day, month and year respectively. Although
DateToJulian modifies year and month, it does not change y and m in
function main; it only changes local copies of y and m. This concept is called
“call by value”.
The example intentionally uses different names for the local variables in the
functions main and DateToJulian, for the purpose of making the above
A tutorial introduction
explanation easier. Renaming main’s variables d, m and y to day, month and
year respectively, does not change the matter: then you just happen to have
two local variables called day, two called month and two called year, which
is perfectly valid in pawn.
The remainder of function DateToJulian is, regarding the pawn language,
uninteresting arithmetic.
Returning to the second part of the function main we see that it now asks
for a day number and calls another function, JulianToDate, to find the
date that matches the day number. Function JulianToDate is interesting
because it takes one input argument (the Julian Day number) and needs to
calculate three output values, the day, month and year. Alas, a function can
only have a single return value —that is, a return statement in a function
may only contain one expression. To solve this, JulianToDate specifically
requests that changes that it makes to some of its function arguments are
copied back to the variables of the caller of the function. Then, in main, the
variables that must hold the result of JulianToDate are passed as arguments
to JulianToDate.
Function JulianToDate marks the appropriate arguments for being “copied
back to caller” by prefixing them with an & symbol. Arguments with an & are
copied back, arguments without is are not. “Copying back” is actually not
the correct term. An argument tagged with an & is passed to the function
in a special way that allows the function to directly modify the original
variable. This is called “call by reference” and an argument that uses it is a
“reference argument”.
In other words, if main passes y to JulianToDate —who maps it to its function argument year— and JulianToDate changes year, then JulianToDate
really changes y. Only through reference arguments can a function directly
modify a variable that is declared in a different function.
To summarize the use of call-by-value versus call-by-reference: if a function
has one output value, you typically use a return statement; if a function has
more output values, you use reference arguments. You may combine the two
inside a single function, for example in a function that returns its “normal”
output via a reference argument and an error code in its return value.
As an aside, many desktop application use conversions to and from Julian
Day numbers (or varieties of it) to conveniently calculate the number of
days between to dates or to calculate the date that is 90 days from now
A tutorial introduction
—for example.
• Rational numbers
All calculations done up to this point involved only whole numbers —integer
values. pawn also has support for numbers that can hold fractional values:
these are called “rational numbers”. However, whether this support is enabled depends on the host application.
Rational numbers can be implemented as either floating-point or fixed-point
numbers. Floating-point arithmetic is commonly used for general-purpose
and scientific calculations, while fixed-point arithmetic is more suitable for
financial processing and applications where rounding errors should not come
into play (or at least, they should be predictable). The pawn toolkit has both
a floating-point and a fixed-point module, and the details (and trade-offs)
for these modules in their respective documentation. The issue is, however,
that a host application may implement either floating-point or fixed-point,
or both or neither.∗ The program below requires that at least either kind of
rational number support is available; it will fail to run if the host application
does not support rational numbers at all.
#include <rational>
new Rational: Celsius
new Rational: Fahrenheit
print "Celsius\t Fahrenheit\n"
for (Celsius = 5; Celsius <= 25; Celsius++)
Fahrenheit = (Celsius * 1.8) + 32
printf "%r \t %r\n", Celsius, Fahrenheit
The example program converts a table of degrees Celsius to degrees Fahrenheit. The first directive of this program is to import definitions for rational
number support from an include file. The file “rational” includes either
Actually, this is already true of all native functions, including all native functions that
the examples in this manual use.
A tutorial introduction
support for floating-point numbers or for fixed-point numbers, depending on
what is available.
The variables Celsius and Fahrenheit are declared with a tag “Rational:”
between the keyword new and the variable name. A tag name denotes the
purpose of the variable, its permitted use and, as a special case for rational
numbers, its memory lay-out. The Rational: tag tells the pawn parser that
the variables Celsius and Fahrenheit contain fractional values, rather than
whole numbers.
The equation for obtaining degrees Fahrenheit from degrees Celsius is
F =
+ 32 ◦ C
The program uses the value 1.8 for the quotient 9/5. When rational number
support is enabled, pawn supports values with a fractional part behind the
decimal point.
The only other non-trivial change from earlier programs is that the format
string for the printf function now has variable placeholders denoted with
“%r” instead of “%d”. The placeholder %r prints a rational number at the
position; %d is only for integers (“whole numbers”).
I used the include file “rational” rather than “float” or “fixed” in an
attempt to make the example program portable. If you know that the host
application supports floating point arithmetic, it may be more convenient
to “#include” the definitions from the file float and use the tag Float:
instead of Rational —when doing so, you should also replace %r by %f in
the call to printf. For details on fixed point and floating point support,
please see the application notes “Fixed Point Support Library” and “Floating
Point Support Library” that are available separately.
• Strings
pawn has no intrinsic “string” type; character strings are stored in arrays,
with the convention that the array element behind the last valid character is
zero. Working with strings is therefore equivalent with working with arrays.
Among the simplest of encryption schemes is the one called “ROT13” —
actually the algorithm is quite “weak” from a cryptographical point of view.
It is most widely used in public electronic forums (BBSes, Usenet) to hide
Tag names: 66
A tutorial introduction
texts from casual reading, such as the solution to puzzles or riddles. ROT13
simply “rotates” the alphabet by half its length, i.e. 13 characters. It is a
symmetric operation: applying it twice on the same text reveals the original.
/* Simple encryption, using ROT13 */
printf "Please type the string to mangle: "
new str[100]
getstring str, sizeof str, .pack = false
rot13 str
printf "After mangling, the string is: \"%s\"\n", str
for (new index = 0; string[index]; index++)
if (’a’ <= string[index] <= ’z’)
string[index] = (string[index] - ’a’ + 13) % 26 + ’a’
else if (’A’ <= string[index] <= ’Z’)
string[index] = (string[index] - ’A’ + 13) % 26 + ’A’
In the function header of rot13, the parameter “string” is declared as an
array, but without specifying the size of the array —there is no value between
the square brackets. When you specify a size for an array in a function
header, it must match the size of the actual parameter in the function call.
Omitting the array size specification in the function header removes this
restriction and allows the function to be called with arrays of any size. You
must then have some other means of determining the (maximum) size of the
array. In the case of a string parameter, one can simply search for the zero
The for loop that walks over the string is typical for string processing functions. Note that the loop condition is “string[index]”. The rule for true/
false conditions in pawn is that any value is “true”, except zero. That is,
when the array cell at string[index] is zero, it is “false” and the loop
The ROT13 algorithm rotates only letters; digits, punctuation and special
characters are left unaltered. Additionally, upper and lower case letters must
be handled separately. Inside the for loop, two if statements filter out the
characters of interest. The way that the second if is chained to the “else”
A tutorial introduction
clause of the first if is noteworthy, as it is a typical method of testing for
multiple non-overlapping conditions.
Earlier in this chapter, the concept of “call by value” versus “call by reference” were discussed. When you are working with strings, or arrays in
general, note that pawn always passes arrays by reference. It does this to
conserve memory and to increase performance —arrays can be large data
structures and passing them by value requires a copy of this data structure
to be made, taking both memory and time. Due to this rule, function rot13
can modify its function parameter (called “string” in the example) without
needing to declare as a reference argument.
Another point of interest are the conditions in the two if statements. The
first if, for example, holds the condition “’a’ <= string[index] <=
’z’”, which means that the expression is true if (and only if) both ’a’
<= string[index] and string[index] <= ’z’ are true. In the combined
expression, the relational operators are said to be “chained”, as they chain
multiple comparisons in one condition.
Finally, note how the last printf in function main uses the escape sequence
\" to print a double quote. Normally a double quote ends the literal string;
the escape sequence “\"” inserts a double quote into the string.
Staying on the subject of strings and arrays, below is a program that separates a string of text into individual words and counts them. It is a simple
program that shows a few new features of the pawn language.
/* word count: count words on a string that the user types */
#include <string>
print "Please type a string: "
new string[100]
getstring string, sizeof string
new count = 0
new word[20]
new index
for ( ;; )
word = strtok(string, index)
if (strlen(word) == 0)
A function that
takes an array
as an argument
and that does not
change it, may
mark the argument as “const”;
see page 71
Relational operators: 106
Escape sequence:
A tutorial introduction
printf "Word %d: ’%s’\n", count, word
printf "\nNumber of words: %d\n", count
strtok(const string[], &index)
new length = strlen(string)
/* skip leading white space */
while (index < length && string[index] <= ’ ’)
/* store the word letter for letter */
new offset = index
/* save start position of token */
new result[20]
/* string to store the word in */
while (index < length
&& string[index] > ’ ’
&& index - offset < sizeof result - 1)
result[index - offset] = string[index]
result[index - offset] = EOS
/* zero-terminate the string */
return result
“for” loop: 112
Function main first displays a message and retrieves a string that the user
must type. Then it enters a loop: writing “for (;;)” creates a loop without initialisation, without increment and without test —it is an infinite loop,
equivalent to “while (true)”. However, where the pawn parser will give
you a warning if you type “while (true)” (something along the line “redundant test expression; always true”), “for (;;)” passes the parser without
A typical use for an infinite loop is a case where you need a loop with the
test in the middle —a hybrid between a while and a do. . . while loop, so to
speak. pawn does not support loops-with-a-test-in-the middle directly, but
you can imitate one by coding an infinite loop with a conditional break. In
this example program, the loop:
⋄ gets a word from the string —code before the test;
⋄ tests whether a new word is available, and breaks out of the loop if not
—the test in the middle;
⋄ prints the word and its sequence number —code after the test.
A tutorial introduction
As is apparent from the line “word = strtok(string, index)” (and the
declaration of variable word), pawn supports array assignment and functions returning arrays. The pawn parser verifies that the array that strtok
returns has the same size and dimensions as the variable that it is assigned
Function strlen is a native function (predefined), but strtok is not: it
must be implemented by ourselves. The function strtok was inspired by
the function of the same name from C/C++ , but it does not modify the
source string. Instead it copies characters from the source string, word for
word, into a local array, which it then returns.
• Arrays and symbolic subscripts (structured data)
In a typeless language, we might assign a different purpose to some array
elements than to other elements in the same array. pawn supports symbolic
substripts that allow to assign specific tag names or ranges to individual
array elements.
The example to illustrate symbolic subscripts and arrays is longer than previous pawn programs, and it also displays a few other features, such as global
variables and named parameters.
Listing: queue.p
/* Priority queue (for simple text strings) */
#include <string>
new msg[.text{40}, .priority]
/* insert a few items (read from console input) */
printf "Please insert a few messages and their priorities; " ...
"end with an empty string\n"
for ( ;; )
printf "Message: "
getstring msg.text, .pack = true
if (strlen(msg.text) == 0)
printf "Priority: "
msg.priority = getvalue()
if (!insert(msg))
printf "Queue is full, cannot insert more items\n"
A tutorial introduction
/* now print the messages extracted from the queue */
printf "\nContents of the queue:\n"
while (extract(msg))
printf "[%d] %s\n", msg.priority, msg.text
const queuesize = 10
new queue[queuesize][.text{40}, .priority]
new queueitems = 0
insert(const item[.text{40}, .priority])
/* check if the queue can hold one more message */
if (queueitems == queuesize)
return false
/* queue is full */
/* find the position to insert it to */
new pos = queueitems
/* start at the bottom */
while (pos > 0 && item.priority > queue[pos-1].priority)
/* higher priority: move up a slot */
/* make place for the item at the insertion spot */
for (new i = queueitems; i > pos; --i)
queue[i] = queue[i-1]
/* add the message to the correct slot */
queue[pos] = item
return true
extract(item[.text{40}, .priority])
/* check whether the queue has one more message */
if (queueitems == 0)
return false
/* queue is empty */
/* copy the topmost item */
item = queue[0]
/* move the queue one position up */
for (new i = 0; i < queueitems; ++i)
queue[i] = queue[i+1]
return true
Function main starts with a declaration of array variable msg. The array has
two fields, “.text” and “.priority”; the “.text” field is declared as a subarray holding 40 characters. The period is required for symbolic subscripts
and there may be no space between the period and the name.
A tutorial introduction
When an array is declared with symbolic subscripts, it may only be indexed
with these subscripts. It would be an error to say, for example, “msg[0]”.
On the other hand, since there can only be a single symbolic subscript between the brackets, the brackets become optional. That is, you can write
“msg.priority” as a shorthand for “msg.[priority]”.
Further in main are two loops. The for loop reads strings and priority values
from the console and inserts them in a queue. The while loop below that
extracts element by element from the queue and prints the information on
the screen. The point to note, is that the for loop stores both the string and
the priority number (an integer) in the same variable msg; indeed, function
main declares only a single variable. Function getstring stores the message
text that you type starting at array msg.text while the priority value is
stored (by an assignment a few lines lower) in msg.priority. The printf
function in the while loop reads the string and the value from those positions
as well.
At the same time, the msg array is an entity on itself: it is passed in its
entirety to function insert. That function, in turn, says near the end
“queue[queueitems] = item”, where item is an array with the same declaration as the msg variable in main, and queue is a two-dimensional array
that holds queuesize elements, with the minor dimension having symbolic
subscripts. The declaration of queue and queuesize are just above function
At several spots in the example program, the same symbolic subscripts are
repeated. In practice, a program would declare the list of symbolic constants in a #define directive and declare the arrays using this text-substition
macro. This saves typing and makes modifications of the declaration easier to maintain. Concretely, when adding near the top of the program the
following line:
#define MESSAGE[.text{40}, .priority]
you can declare all arrays with these symbolic subscripts as “msg[MESSAGE]”.
The example implements a “priority queue”. You can insert a number of
messages into the queue and when these messages all have the same priority,
they are extracted from the queue in the same order. However, when the
messages have different priorities, the one with the highest priority comes
out first. The “intelligence” for this operation is inside function insert: it
first determines the position of the new message to add, then moves a few
A tutorial introduction
messages one position upward to make space for the new message. Function
extract simply always retrieves the first element of the queue and shifts all
remaining elements down by one position.
Note that both functions insert and extract work on two shared variables,
queue and queueitems. A variable that is declared inside a function, like
variable msg in function main can only be accessed from within that function.
A “global variable” is accessible by all functions, and that variable is declared
outside the scope of any function. Variables must still be declared before
they are used, so main cannot access variables queue and queueitems, but
both insert and extract can.
Function extract returns the messages with the highest priority via its function argument item. That is, it changes its function argument by copying
the first element of the queue array into item. Function insert copies in
the other direction and it does not change its function argument item. In
such a case, it is advised to mark the function argument as “const”. This
helps the pawn parser to both check for errors and to generate better (more
compact, quicker) code.
Named parameters: 73
getstring: 128
A final remark on this latest sample is the call to getstring in function
main: if you look up the function declaration, you will see that it takes three
parameters, two of which are optional. In this example, only the first and
the last parameters are passed in. Note how the example avoids ambiguity
about which parameter follows the first, by putting the argument name in
front of the value. By using “named parameters” rather than positional
parameters, the order in which the parameters are listed is not important.
Named parameters are convenient in specifying —and deciphering— long
parameter lists.
• Bit operations to manipulate “sets”
A few algorithms are most easily solved with “set operations”, like intersection, union and inversion. In the figure below, for example, we want to
design an algorithm that returns us the points that can be reached from
some other point in a specified maximum number of steps. For example, if
we ask it to return the points that can be reached in two steps starting from
B, the algorithm has to return C, D, E and F, but not G because G takes
three steps from B.
Our approach is to keep, for each point in the graph, the set of other points
A tutorial introduction
that it can reach in one step —this is the “next_step” set. We also have a
“result” set that keeps all points that we have found so far. We start by
setting the result set equal to the next_step set for the departure point.
Now we have in the result set all points that one can reach in one step.
Then, for every point in our result set, we create a union of the result set
and the next_step set for that point. This process is iterated for a specified
number of loops.
An example may clarify the procedure outlined above. When the departure
point is B, we start by setting the result set to D and E —these are the
points that one can reach from B in one step. Then, we walk through the
result set. The first point that we encounter in the set is D, and we check
what points can be reached from D in one step: these are C and F. So we
add C and F to the result set. We knew that the points that can be reached
from D in one step are C and F, because C and F are in the next_step set
for D. So what we do is to merge the next_step set for point D into the
result set. The merge is called a “union” in set theory. That handles D.
The original result set also contained point E, but the next_step set for
E is empty, so no more point is added. The new result set therefore now
contains C, D, E and F.
A set is a general purpose container for elements. The only information that
a set holds of an element is whether it is present in the set or not. The order
of elements in a set is insignificant and a set cannot contain the same element
multiple times. The pawn language does not provide a “set” data type or
A tutorial introduction
operators that work on sets. However, sets with up to 32 elements can be
simulated by bit operations. It takes just one bit to store a “present/absent”
status and a 32-bit cell can therefore maintain the status for 32 set elements
—provided that each element is assigned a unique bit position.
The relation between set operations and bitwise operations is summarized
in the following table. In the table, an upper case letter stands for a set and
a lower case letter for an element from that set.
empty set
mathematical notation pawn expression
A & B
(1 << x) & A
To test for membership —that is, to query whether a set holds a particular
element, create a set with just one element and take the intersection. If the
result is 0 (the empty set) the element is not in the set. Bit numbering starts
typically at zero; the lowest bit is bit 0 and the highest bit in a 32-bit cell
is bit 31. To make a cell with only bit 7 set, shift the value 1 left by seven
—or in a pawn expression: “1 << 7”.
Below is the program that implements the algorithm described earlier to find
all points that can be reached from a specific departure in a given number
of steps. The algorithm is completely in the findtargets function.
/* Set operations, using bit arithmetic */
{ A
new nextstep[] =
[ C | E,
/* A can reach C and E */
D | E,
/* B "
D and E */
/* C "
G */
A tutorial introduction
C | F,
E | F,
C and F */
none */
none */
E and F */
print "The departure point: "
new start = clamp( .value = toupper(getchar()) - ’A’,
.min = 0,
.max = sizeof nextstep - 1
print "\nThe number of steps: "
new steps = getvalue()
/* make the set */
new result = findtargets(start, steps, nextstep)
printf "The points in range of %c in %d steps: ", start + ’A’, steps
for (new i = 0; i < sizeof nextstep; i++)
if (result & 1 << i)
printf "%c ", i + ’A’
findtargets(start, steps, nextstep[], numpoints = sizeof nextstep)
new result = 0
new addedpoints = nextstep[start]
while (steps-- > 0 && result != addedpoints)
result = addedpoints
for (new i = 0; i < numpoints; i++)
if (result & 1 << i)
addedpoints |= nextstep[i]
return result
The const statement just below the header of the main function declares
the constants for the nodes A to G, using binary radix so that that only a
single bit is set in each value.
“const” statement: 101
When working with sets, a typical task that pops up is to determine the
number of elements in the set. A straightforward function that does this is
“cellbits” constant: 101
simple bitcount function
new count = 0
for (new i = 0; i < cellbits; i++)
if (set & (1 << i))
A tutorial introduction
return count
With a cell size of 32 bits, this function’s loop iterates 32 times to check for
a single bit at each iteration. With a bit of binary arithmetic magic, we can
reduce it to loop only for the number of bits that are “set”. That is, the
following function iterates only once if the input value has only one bit set:
Listing: improved bitcount function
new count = 0
if (set)
while ((set = set & (set - 1)))
return count
• A simple RPN calculator
Algebraic notation is also called
“infix” notation
The common mathematical notation, with expressions like “26−3×(5+2)”,
is known as the algebraic notation. It is a compact notation and we have
grown accustomed to it. pawn and by far most other programming languages
use the algebraic notation for their programming expressions. The algebraic
notation does have a few disadvantages, though. For instance, it occasionally
requires that the order of operations is made explicit by folding a part of
the expression in parentheses. The expression at the top of this paragraph
can be rewritten to eliminate the parentheses, but at the cost of nearly
doubling its length. In practice, the algebraic notation is augmented with
precedence level rules that say, for example, that multiplication goes before
addition and subtraction.∗ Precedence levels greatly reduce the need for
parentheses, but it does not fully avoid them. Worse is that when the number
of operators grows large, the hierarchy of precedence levels and the particular
precedence level for each operator becomes hard to memorize —which is
why an operator-rich language as APL does away with precedence levels
Around 1920, the Polish mathematician Jan Ĺukasiewicz demonstrated that
by putting the operators in front of their operands, instead of between them,
These rules are often summarized in a mnemonic like “Please Excuse My Dear Aunt Sally”
(Parentheses, Exponentiation, Multiplication, Division, Addition, Subtraction).
A tutorial introduction
precedence levels became redundant and parentheses were never necessary.
This notation became known as the “Polish Notation”.† Charles Hamblin
proposed later to put operators behind the operands, calling it the “Reverse
Polish Notation”. The advantage of reversing the order is that the operators
are listed in the same order as they must be executed: when reading the
operators from the left to the right, you also have the operations to perform
in that order. The algebraic expression from the beginning of this section
would read in rpn as:
26 3 5 2 + × −
When looking at the operators only, we have: first an addition, then a
multiplication and finally a subtraction. The operands of each operator are
read from right to left: the operands for the + operator are the values 5 and
2, those for the × operator are the result of the previous addition and the
value 3, and so on.
It is helpful to imagine the values to be stacked on a pile, where the operators
take one or more operands from the top of the pile and put a result back
on top of the pile. When reading through the rpn expression, the values
26, 3, 5 and 2 are “stacked” in that order. The operator + removes the top
two elements from the stack (5 and 2) and pushes the sum of these values
back —the stack now reads “26 3 7”. Then, the × operator removes 3 and 7
and pushes the product of the values onto the stack —the stack is “26 21”.
Finally, the − operator subtracts 21 from 26 and stores the single value 5,
the end result of the expression, back onto the stack.
Reverse Polish Notation became popular because it was easy to understand
and easy to implement in (early) calculators. It also opens the way to operators with more than two operands (e.g. integration) or operators with more
than one result (e.g. conversion between polar and Cartesian coordinates).
The main program for a Reverse Polish Notation calculator is below:
/* a simple RPN calculator */
#include strtok
#include stack
#include rpnparse
Polish Notation is completely unrelated to “Hungarian Notation” —which is just the habit
of adding “type” or “purpose” identification warts to names of variables or functions.
Reverse Polish
Notation is also
called “postfix”
A tutorial introduction
print "Type expressions in Reverse Polish Notation " ...
"(or an empty line to quit)\n"
new string{100}
while (getstring(string, .pack = true))
rpncalc string
The main program contains very little code itself; instead it includes the
required code from three other files, each of which implements a few functions
that, together, build the rpn calculator. When programs or scripts get
larger, it is usually advised to spread the implementation over several files,
in order to make maintenance easier.
Function main first puts up a prompt and calls the native function getstring
to read an expression that the user types. Then it calls the custom function
rpncalc to do the real work. Function rpncalc is implemented in the file, reproduced below:
/* main rpn parser and lexical analysis, part of the RPN calculator */
#include <rational>
#include <string>
#define Token [
Rational: .value,
/* operator or token type */
/* value, if t_type is "Number" */
/* raw string */
const Number
= ’0’
const EndOfExpr = ’#’
rpncalc(const string{})
new index
new field[Token]
for ( ;; )
field = gettoken(string, index)
switch (field.type)
case Number:
push field.value
case ’+’:
push pop() + pop()
case ’-’:
push - pop() + pop()
case ’*’:
A tutorial introduction
push pop() * pop()
case ’/’, ’:’:
push 1.0 / pop() * pop()
case EndOfExpr:
/* exit "for" loop */
printf "Unknown operator ’%s’\n", field.word
printf "Result = %r\n", pop()
if (clearstack())
print "Stack not empty\n", red
gettoken(const string{}, &index)
/* first get the next "word" from the string */
new word{20}
word = strtok(string, index)
/* then parse it */
new field[Token]
field.word = word
if (strlen(word) == 0)
field.type = EndOfExpr /* special "stop" symbol */
field.value = 0
else if (’0’ <= word{0} <= ’9’)
field.type = Number
field.value = rval(word)
field.type = word{0}
field.value = 0
return field
The rpn calculator uses rational number support and includes the “rational” file for that purpose. Almost all of the operations on
rational numbers is hidden in the arithmetic. The only direct references to
rational numbers are the “%r” format code in the printf statement near the
bottom of function rpncalc and the call to rationalstr halfway function
Near the top in the file is a preprocessor macro that declares
the symbolic subscripts for an array. The macro name, “Token” will be used
Rational numbers,
see also the “Celsius to Fahrenheit” example on
page 14
macro: 92
A tutorial introduction
throughout the program to declare arrays with those fields. For example,
function rpncalc declares variable field as an array using the macro to
declare the field names.
Arrays with symbolic subscripts were already introduced in the section Arrays and symbolic subscripts on page 19; this script shows another feature
of symbolic subscripts: individual substripts may have a tag name of their
own. In this example, .type is a simple cell, .value is a rational value (with
a fractional part) that is tagged as such, and .word can hold a string of 20
characters (includding the terminating zero byte). See, for example, the line:
printf "Unknown operator ’%s’\n", field.word
how the .word subscript of the field variable is used as a string.
“switch” statement: 114
If you know C/C++ or Java, you may want to look at the switch statement.
The switch statement differs in a number of ways from the other languages
that provide it. The cases are not fall-through, for example, which in turn
means that the break statement for the case EndOfExpr breaks out of the
enclosing loop, instead of out of the switch.
On the top of the for loop in function rpncalc, you will find the instruction “field = gettoken(string, index)”. As already exemplified in the
wcount.p (“word count”) program on page 17, functions may return arrays.
It gets more interesting for a similar line in function gettoken:
field.word = word
where word is an array for 20 characters and field is an array with 3 (symbolic) subscripts. However, as the .word subscript is declared as having a
size of 20 characters, the expression “field.word” is considered a sub-array
of 20 characters, precisely matching the array size of word.
/* extract words from a string (words are separated by white space) */
#include <string>
strtok(const string{}, &index)
new length = strlen(string)
/* skip leading white space */
while (index < length && string{index} <= ’ ’)
/* store the word letter for letter */
new offset = index
/* save start position of token */
const wordlength = 20
/* maximum word length */
new result{wordlength}
/* string to store the word in */
while (index < length
A tutorial introduction
&& string{index} > ’ ’
&& index - offset < wordlength)
result{index - offset} = string{index}
result{index - offset} = EOS
/* zero-terminate the string */
return result
Function strtok is the same as the one used in the wcount.p example. It
is implemented in a separate file for the rpn calculator program. Note that
the strtok function as it is implemented here can only handle words with
up to 19 characters —the 20th character is the zero terminator. A truly
general purpose re-usable implementation of an strtok function would pass
the destination array as a parameter, so that it could handle words of any
size. Supporting both packed and unpack strings would also be a useful
feature of a general purpose function.
When discussing the merits of Reverse Polish Notation, I mentioned that a
stack is both an aid in “visualizing” the algorithm as well as a convenient
method to implement an rpn parser. This example rpn calculator, uses a
stack with the ubiquitous functions push and pop. For error checking and
resetting the stack, there is a third function that clears the stack.
/* stack functions, part of the RPN calculator */
#include <rational>
static Rational: stack[50]
static stackidx = 0
push(Rational: value)
assert stackidx < sizeof stack
stack[stackidx++] = value
Rational: pop()
assert stackidx > 0
return stack[--stackidx]
assert stackidx >= 0
if (stackidx == 0)
return false
wcount.p: 17
A tutorial introduction
stackidx = 0
return true
The file includes the file rational again. This is technically
not necessary ( already included the definitions for rational
number support), but it does not do any harm either and, for the sake of
code re-use, it is better to make any file include the definitions of the libraries
that it depends on.
Notice how the two global variables stack and stackidx are declared as
“static” variables; using the keyword static instead of new. Doing this
makes the global variables “visible” in that file only. For all other files in
a larger project, the symbols stack and stackidx are invisible and they
cannot (accidentally) modify the variables. It also allows the other modules
to declare their own private variables with these names, so it avoids name
The rpn calculator is actually still a fairly small program, but it has been
set up as if it were a larger program. It was also designed to demonstrate a
set of elements of the pawn language and the example program could have
been implemented more compactly.
• Event-driven programming
All of the example programs that were developed in this chapter so far, have
used a “flow-driven” programming model: they start with main and the code
determines what to do and when to request input. This programming model
is easy to understand and it nicely fits most programming languages, but it is
also a model does not fit many “real life” situations. Quite often, a program
cannot simply process data and suggest that the user provides input only
when it is ready for him/her. Instead, it is the user who decides when to
provide input, and the program or script should be prepared to process it in
an acceptable time, regardless of what it was doing at the moment.
The above description suggests that a program should therefore be able to
interrupt its work and do other things before picking up the original task.
In early implementations, this was indeed how such functionality was implemented: a multi-tasking system where one task (or thread) managed the
background tasks and a second task/thread that sits in a loop continuously
requesting user input. This is a heavy-weight solution, however. A more
A tutorial introduction
light-weight implementation of a responsive system is what is called the
“event-driven” programming model.
In the event-driven programming model, a program or script decomposes any
lengthy (background) task into short manageable blocks and in between, it
is available for input. Instead of having the program poll for input, however, the host application (or some other sub-system) calls a function that
is attached to the event —but only if the event occurs.
A typical event is “input”. Observe that input does not only come from human operators. Input packets can arrive over serial cables, network stacks,
internal sub-systems such as timers and clocks, and all kinds of other equipment that you may have attached to your system. Many of the apparatus
that produce input, just send it. The arrival of such input is an event, just
like a key press. If you do not catch the event, a few of them may be stored
in an internal system queue, but once the queue is saturated the events are
simply dropped.
pawn directly supports the event-driven model, because it supports multiple entry points. The sole entry point of a flow-driven program is main; an
event-driven program has an entry point for every event that it captures.
When compared to the flow-driven model, event-driven programs often appear “bottom-up”: instead of your program calling into the host application
and deciding what to do next, your program is being called from the outside
and it is required to respond appropriately and promptly.
pawn does not specify a standard library, and so there is no guarantee
that in a particular implementation, functions like printf and getvalue.
Although it is suggested that every implementation provides a minimal console/terminal interface with a these functions, their availability is ultimately
implementation-dependent. The same holds for the public functions —the
entry points for a script. It is implementation-dependent which public functions a host application supports. The script in this section may therefore
not run on your platform (even if all previous scripts ran fine). The tools
in the standard distribution of the pawn system support all scripts developed in this manual, provided that your operating system or environment
supports standard terminal functions such as setting the cursor position.
An early programming language that was developed solely for teaching the
concepts of programming to children was “Logo”. This dialect of LISP
made programming visual by having a small robot, the “turtle”, drive over
the floor under control of a simple program. This concept was then copied
Public functions:
A tutorial introduction
to moving a (usually triangular) cursor of the computer display, again under
control of a program. A novelty was that the turtle now left a trail behind
it, allowing you to create drawings by properly programming the turtle —it
became known as turtle graphics. The term “turtle graphics” was also used
for drawing interactively with the arrow keys on the keyboard and a “turtle”
for the current position. This method of drawing pictures on the computer
was briefly popular before the advent of the mouse.
/* get current position */
new x, y
wherexy x, y
/* determine how the update the current position */
switch (key)
case ’u’: y-/* up */
case ’d’: y++
/* down */
case ’l’: x-/* left */
case ’r’: x++
/* right */
case ’\e’: exit /* Escape = exit */
/* adjust the cursor position and draw something */
moveturtle x, y
moveturtle(x, y)
gotoxy x, y
print "*"
gotoxy x, y
The entry point of the above program is @keypressed —it is called on a
key press. If you run the program and do not type any key, the function
@keypressed never runs; if you type ten keys, @keypressed runs ten times.
Contrast this behaviour with main: function main runs immediately after
you start the script and it runs only once.
It is still allowed to add a main function to an event-driven program: the
main function will then serve for one-time initialization. A simple addition
to this example program is to add a main function, in order to clear the
console/terminal window on entry and perhaps set the initial position of the
“turtle” to the centre.
A tutorial introduction
Support for function keys and other special keys (e.g. the arrow keys) is
highly system-dependent. On ANSI terminals, these keys produce different
codes than in a Windows “DOS box”. In the spirit of keeping the example program portable, I have used common letters (“u” for up, “l” for left,
etc.). This does not mean, however, that special keys are beyond pawn’s
In the “turtle” script, the “Escape” key terminates the host application
through the instruction exit. For a simple pawn run-time host, this will
indeed work. With host applications where the script is an add-on, or hostapplications that are embedded in a device, the script usually cannot terminate the host application.
• Multiple events
The advantages of the event-driven programming model, for building reactive
programs, become apparent in the presence of multiple events. In fact, the
event-driven model is only useful if you have more that one entry point; if
your script just handles a single event, it might as well enter a polling loop
for that single event. The more events need to be handled, the harder the
flow-driven programming model becomes. The script below implements a
bare-bones “chat” program, using only two events: one for sending and one
for receiving. The script allows users on a network (or perhaps over another
connection) to exchange single-line messages.
The script depends on the host application to provide the native and public
functions for sending and receiving “datagrams” and for responding to keys
that are typed in. How the host application sends its messages, over a serial
line or using TCP/IP, the host application may decide itself. The tools in the
standard pawn distribution push the messages over the TCP/IP network,
and allow for a “broadcast” mode so that more than two people can chat
with each other.
Listing: chat.p
#include <datagram>
const cellchars = cellbits / charbits
@receivestring(const message[], const source[])
printf "[%s] says: %s\n", source, message
static string{100}
static index
A tutorial introduction
if (key == ’\e’)
/* quit on ’Esc’ key */
echo key
if (key == ’\r’ || key == ’\n’ || index == sizeof string * cellchars)
string{index} = ’\0’
/* terminate string */
sendstring string
index = 0
string{index} = ’\0’
string{index++} = key
new string{2} = { 0 }
string{0} = key == ’\r’ ? ’\n’ : key
printf string
The bulk of the above script handles gathering received key-presses into a
string and sending that string after seeing the enter key. The “Escape” key
ends the program. The function echo serves to give visual feedback of what
the user types: it builds a zero-terminated string from the key and prints it.
Despite its simplicity, this script has the interesting property that there is
no fixed or prescribed order in which the messages are to be sent or received
—there is no query–reply scheme where each host takes its turn in talking
& listening. A new message may even be received while the user is typing
its own message.∗
• State programming
In a program following the event-driven model, events arrive individually,
and they are also responded to individually. On occasion, though, an event
is part of a sequential flow, that must be handled in order. Examples are data
transfer protocols over, for example, a serial line. Each event may carry a
command, a snippet of data that is part of a larger file, an acknowledgement,
As this script makes no attempt to separate received messages from typed messages (for
example, in two different scrollable regions), the terminal/console will look confusing when
this happens. With an improved user-interface, this simple script could indeed be a nice
message-base chat program.
A tutorial introduction
or other signals that take part in the protocol. For the stream of events (and
the data packets that they carry) to make sense, the event-driven program
must follow a precise hand-shaking protocol.
To adhere to a protocol, an event-driven program must respond to each
event in compliance with the (recent) history of events received earlier and
the responses to those events. In other words, the handling of one event
may set up a “condition” or “environment” for the handling any one or
more subsequent events.
A simple, but quite effective, abstraction for constructing reactive systems
that need to follow (partially) sequential protocols, is that of the “automaton” or state machine. As the number of states are usually finite, the theory
often refers to such automatons as Finite State Automatons or Finite State
Machines. In an automaton, the context (or condition) of an event is its
state. An event that arrives may be handled differently depending on the
state of the automaton, and in response to an event, the automaton may
switch to another state —this is called a transition. A transition, in other
words, as a response of the automaton to an event in the context of its state.
Automatons are very common in software as well as in mechanical devices
(you may see the Jacquard Loom as an early state machine). Automatons, with a finite number of states, are deterministic (i.e. predictable in
behaviour) and their relatively simple design allows a straightforward implementation from a “state diagram”.
In a state diagram, the states are usually represented as circles or rounded
rectangles and the arrows represent the transitions. As transitions are the
A tutorial introduction
response of the automaton to events, an arrow may also be seen as an event
“that does something”. An event/transition that is not defined in a particular state is assumed to have no effect —it is silently ignored. A filled
dot represents the entry state, which your program (or the host application)
must set in start-up. It is common to omit in a state diagram all event arrows that drop back into the same state, but for the preceding figure I have
chosen to make the response to all events explicit.
This state diagram is for “parsing” comments that start with “/*” and end
with “*/”. There are states for plain text and for text inside a comment, plus
two states for tentative entry into or exit from a comment. The automaton
is intended to parse the comments interactively, from characters that the
user types on the keyboard. Therefore, the only events that the automaton
reacts on are key presses. Actually, there is only one event (“key-press”)
and the state switches are determined by event’s parameter: the key.
pawn supports automatons and states directly in the language. Every function∗ may optionally have one or more states assigned to it. pawn also
supports multiple automatons, and each state is part of a particular automaton. The following script implements the preceding state diagram (in a
single, anonymous, automaton). To differentiate plain text from comments,
both are output in a different colour.
/* parse C comments interactively, using events and a state machine */
state plain
@keypressed(key) <plain>
state (key == ’/’) slash
if (key != ’/’)
echo key
@keypressed(key) <slash>
state (key != ’/’) plain
state (key == ’*’) comment
echo ’/’
/* print ’/’ held back from previous state */
if (key != ’/’)
echo key
With the exception of “native functions” and user-defined operators.
A tutorial introduction
@keypressed(key) <comment>
echo key
state (key == ’*’) star
@keypressed(key) <star>
echo key
state (key != ’*’) comment
state (key == ’/’) plain
echo(key) <plain, slash>
printchar key, yellow
echo(key) <comment, star>
printchar key, green
printchar(ch, colour)
setattr .foreground = colour
printf "%c", ch
Function main sets the starting state to main and exits; all logic is eventdriven. When a key arrives in state plain, the program checks for a slash
and conditionally prints the received key. The interaction between the states
plain and slash demonstrates a complexity that is typical for automatons:
you must decide how to respond to an event when it arrives, without being
able to “peek ahead” or undo responses to earlier events. This is usually
the case for event-driven systems —you neither know what event you will
receive next, nor when you will receive it, and whatever your response to the
current event, there is a good chance that you cannot erase it on a future
event and pretend that it never happened.
In our particular case, when a slash arrives, this might be the start of a
comment sequence (“/*”), but it is not necessarily so. By inference, we
cannot decide on reception of the slash character what colour to print it in.
Hence, we hold it back. However, there is no global variable in the script that
says that a character is held back —in fact, apart from function parameters,
no variable is declared at all in this script. The information about a character
being held back is “hidden” in the state of the automaton.
As is apparent in the script, state changes may be conditional. The condition
is optional, and you can also use the common if–else construct to change
A tutorial introduction
Being state-dependent is not reserved for the event functions. Other functions may have state declarations as well, as the echo function demonstrates.
When a function would have the same implementation for several states, you
just need to write a single implementation and mention all applicable states.
For function echo there are two implementations to handle the four states.†
That said, an automaton must be prepared to handle all events in any state.
Typically, the automaton has neither control over which events arrive nor
over when they arrive, so not handling an event in some state could lead
to wrong decisions. It frequently happens, then, that a some events are
meaningful only in a few specific states and that they should trigger an
error or “reset” procedure in all other cases. The function for handling the
event in such “error” condition might then hold a lot of state names, if you
were to mention them explicitly. There is a shorter way: by not mentioning
any name between the angle brackets, the function matches all states that
have not explicit implementation elsewhere. So, for example, you could use
the signature “echo(key) <>” for either of the two implementations (but
not for both).
A single anonymous automaton is pre-defined. If a program contains more
than one automaton, the others must be explicitly mentioned, both in the
state classifier of the function and in the state instruction. To do so, add the
name of the automaton in front of the state name and separate the names of
the automaton and the state with a colon. That is, “parser:slash” stands
for the state slash of the automaton parser. A function can only be part
of a single automaton; you can share one implementation of a function for
several states of the same automaton, but you cannot share that function
for states of different automatons.
• Entry functions and automata theory
State machines, and the foundation of “automata theory”, originate from
mechanical design and pneumatic/electric switching circuits (using relays
rather than transistors). Typical examples are coin acceptors, traffic light
control and communication switching circuits. In these applications, robustness and predictability are paramount, and it was found that these goals
A function that has the same implementation for all states, does not need a state classifierat all —see printchar.
A tutorial introduction
Figure 1: Pedestrian crossing lights
were best achieved when actions (output) were tied to the states rather than
to the events (input). In this design, entering a state causes activity —events
cause state changes, but do not carry out other operations.
In a pedestrian crossing lights system, the lights for the vehicles and the
pedestrians must be synchronized. Technically, there are six possible combinations, but obviously the combination of a green light for the traffic and
a “walk” sign for the pedestrians is recipe for disaster. We can also immediately dismiss the combination of yellow /walk as too dangerous. Thus, four
combinations remain to be handled. The figure below is a state diagram for
the pedestrian crossing lights. The entire process is activated with a button,
and operates on a timer.
A tutorial introduction
When the state red /walk times out, the state cannot immediately go back to
green/wait, because the pedestrians that are busy crossing the road at that
moment need some time to clear the road —the state red /wait allows for
this. For purpose of demonstration, this pedestrian crossing has the added
functionality that when a pedestrian pushes the button while the light for the
traffic is already red, the time that the pedestrian has for crossing is lengthened. If the state is red /wait and the button is pressed, it switches back to
red /walk. The enfolding box around the states red /walk and red /wait for
handling the button event is just a notational convenience: I could also have
drawn two arrows from either state back to red /walk. The script source code
(which follows below) reflects this same notational convenience, though.
In the implementation in the pawn language, the event functions now always
have a single statement, which is either a state change or an empty statement.
Events that do not cause a state change are absent in the diagram, but they
must be handled in the script; hence, the “fall-back” event functions that
do nothing. The output, in this example program only messages printed on
the console, is all done in the special functions entry. The function entry
may be seen as a main for a state: it is implicitly called when the state that
it is attached to is entered. Note that the entry function is also called when
“switching” to the state that the automaton is already in: when the state
is red_walk an invocation of the @keypressed sets the state to red_walk
(which it is already in) and causes the entry function of red_walk to run
—this is a re-entry of the state.
/* traffic light synchronizer, using states in an event-driven model */
#include <time>
state green_wait
@keypressed(key) <green_wait>
state yellow_wait
@keypressed(key) <red_walk, red_wait> state red_walk
@keypressed(key) <>
{} /* fallback */
entry() <green_wait>
print "Green / Don’t walk\n"
entry() <yellow_wait>
print "Yellow / Don’t walk\n"
settimer 2000
{} /*
fallback */
A tutorial introduction
entry() <red_walk>
print "Red / Walk\n"
settimer 5000
entry() <red_wait>
print "Red / Don’t walk\n"
settimer 2000
This example program has an additional dependency on the host application/environment: in addition to the “@keypressed” event function, the
host must also provide an adjustable “@timer” event. Because of the timing
functions, the script includes the system file near the top of the
The event functions with the state changes are all on the top part of the
script. The functions are laid out to take a single line each, to suggest a
table-like structure. All state changes are unconditional in this example,
but conditional state changes may be used with entry functions too. The
bottom part are the event functions.
Two transitions to the state red_walk exist —or three if you consider the
affection of multiple states to a single event function as a mere notational
convenience: from yellow_wait and from the combination of red_walk and
red_wait. These transitions all pass through the same entry function,
thereby reducing and simplifying the code.
In automata theory, an automaton that associates activity with state entries,
such as this pedestrian traffic lights example, is a “Moore automaton”; an
automaton that associates activity with (state-dependent) events or transitions is a “Mealy automaton”. The interactive comment parser on page 38
is a typical Mealy automaton. The two kinds are equivalent: a Mealy automaton can be converted to a Moore automaton and vice versa, although a
Moore automaton may need more states to implement the same behaviour.
In practice, the models are often mixed, with an overall “Moore automaton”
design, and a few “Mealy states” where that saves a state.
• State variables
A tutorial introduction
The model of a pedestrian crossing light in the previous example is not very
realistic (its only goal is to demonstrate a few properties of state programming with pawn). The first thing that is lacking is a degree of fairness:
pedestrians should not be able to block car traffic indefinitely. The car traffic should see a green light for a period of some minimum duration after
pedestrians have had their time slot for crossing the road. Secondly, many
traffic lights have a kind of remote control ability, so that emergency traffic
(ambulance, fire truck, . . . ) can force green lights on their path. A wellknown example of such remote control is the mirt system (Mobile Infra-Red
Transmitter) but other systems exist —the Netherlands use a radiographic
system called vetag for instance.
The new state diagram for the pedestrian crossing light has two more states,
but more importantly: it needs to save data across events and share it
between states. When the pedestrian presses the button while the state is
red_wait, we neither want to react on the button immediately (this was
our “fairness rule”), nor the button to be ignored or “forgotten”. In other
words, we move to the state green_wait_interim regardless of the button
press, but memorize the press for a decision made at the point of leaving
state green_wait_interim.
Automatons excel in modelling control flow in reactive/interactive systems,
but data flow has traditionally been a weak point. To see why, consider
that each event is handled individually by a function and that the local variables in that function disappear when the function returns. Local variables
can, hence, not be used to pass data from one event to the next. Global
variables, while providing a work-around, have drawbacks: global scope and
an “eternal” lifespan. If a variable is used only in the event handlers of a
A tutorial introduction
single state, it is desirable to hide it from the other states, in order to protect it from accidental modification. Likewise, shortening the lifespan to the
state(s) that the variable is active in, reduces the memory footprint. “State
variables” provide this mix of variable scope and variable lifespan that are
tied to a series of states, rather than to functions or modules.
pawn enriches the standard finite state machine (or automaton) with variables that are declared with a state classifier. These variables are only accessible from the listed states and the memory these variable hold may be
reused by other purposes while the automaton is in a different state (different than the ones listed). Apart from the state classifier, the declaration of
a state variable is similar to that of a global variable. The declaration of the
variable button_memo in the next listing illustrates the concept.
To reset the memorized button press, the script uses an “exit” function.
Just like an entry function is called when entering a state, the exit function
is called when leaving a state.
/* a more realistic traffic light synchronizer, including an
* "override" for emergency vehicles
#include <time>
state green_wait_interim
new bool: button_memo <red_wait, green_wait_interim>
switch (key)
case ’ ’: button_press
case ’*’: mirt_detect
button_press() <green_wait>
state yellow_wait
button_press() <red_wait, green_wait_interim>
button_memo = true
button_press() <>
state mirt_override
@timer() <yellow_wait>
state red_walk
/* fallback */
A tutorial introduction
@timer() <red_walk>
state red_wait
@timer() <red_wait>
state green_wait_interim
@timer() <green_wait_interim>
state (!button_memo) green_wait
state (button_memo) yellow_wait
@timer() <mirt_override>
state green_wait
@timer() <>
/* fallback */
entry() <green_wait_interim>
print "Green / Don’t walk\n"
settimer 5000
exit() <green_wait_interim>
button_memo = false
entry() <yellow_wait>
print "Yellow / Don’t walk\n"
settimer 2000
entry() <red_walk>
print "Red / Walk\n"
settimer 5000
entry() <red_wait>
print "Red / Don’t walk\n"
settimer 2000
entry() <mirt_override>
print "Green / Don’t walk\n"
settimer 5000
A tutorial introduction
• State programming wrap-up
The common notation used in state diagrams is to indicate transitions with
arrows and states with circles or rounded rectangles. The circle/rounded
rectangle optionally also mentions the actions of an entry or exit function
and of events that are handled internally —without causing a transition.
The arrow for a transition contains the name of the event (or pseudo-event),
an optional condition between square brackets and an optional action behind
a slash (“/”).
States are ubiquitous, even if we do not always recognize them as such.
The concept of finite state machines has traditionally been applied mostly
to programs mimicking mechanical apparatus and software that implements
communication protocols. With the appearance of event-driven windowing systems, state machines now also appear in the GUI design of desktop
programs. States abound in web programs, because the browser and the
web-site scripting host have only a weak link. That said, the state machine
in web applications is typically implemented in an ad-hoc manner.
States can also be recognized in common problems and riddles. In the well
known riddle of the man that must move a cabbage, a sheep and a wolf
across a river,∗ the states are obvious —the trick of the riddle is to avoid
the forbidden states.
But now that we are discovering “states” everywhere, we must be careful
not to overdo it. For example, in the second implementation of a pedestrian
crossing light, see page 45, I used a variable (button_memo) to hold a criterion
for a decision made at a later time. An alternative implementation would
be to throw in a couple of more states to hold the situations “red-wait-&button-pressed” and “green-wait-interim-&-button-pressed”. No more variable would then be needed, but at the cost of a more complex state diagram
and implementation. In general, the number of states should be kept small.
Although automata provide a good abstraction to model reactive and interactive systems, coming to a correct diagram is not straightforward —and
sometimes just outright hard. Too often, the “sunny day scenario” of states
and events is plotted out first, and everything straying from this path is then
A man has to ferry a wolf, a sheep and a cabbage across a river in a boat, but the boat
can only carry the man and a single additional item. If left unguarded, the wolf will eat
the sheep and the sheep will eat the cabbage. How can the man ferry them across the
A tutorial introduction
added on an impromptu basis. This approach carries the risk that some combinations of events & states are forgotten, and indeed I have encountered
two comment parser diagrams (like the one at page 38) by different book/
magazine authors that were flawed in such way. Instead, I advise to focus on
the events and on the responses for individual events. For every state, every
event should be considered; do not route events through a general purpose
fall-back too eagerly.
It has become common practice, unfortunately, to introduce automata theory with applications for which better solutions exist. One, oft repeated,
example is that of an automaton that accumulates the value of a series of
coins, or that “calculates” the remainder after division by 3 of a binary
number. These applications may have made sense in mechanical/pneumatic
design where “the state” is the only memory that the automaton has, but
in software, using variables and arithmetic operations is the better choice.
Another typical example is that of matching words or patterns using a state
machine: every next letter that is input switches to a new state. Lexical
scanners, such as the ones that compilers and interpreters use to interpret
source code, might use such state machines to filter out “reserved words”.
However, for any practical set of reserved words, such automatons become
unwieldy, and no one will design them by hand. In addition, there is no
reason why a lexical scanner cannot peek ahead in the text or jump back to
a mark that it set earlier —which is one of the criteria for choosing a state
implementation in the first place, and finally, solutions like “trie lookups”
are likely simpler to design and implement while being at least as quick.
• Program verification
Should the compiler/interpreter not catch all bugs? This rhetorical question
has both technical and philosophical sides. I will forego all non-technical
aspects and only mention that, in practice, there is a trade-off between the
“expressiveness” of a computer language and the “enforced correctness” (or
“provable correctness’) of programs in that language. Making a language
very “strict” is not a solution if work needs to be done that exceeds the size
of a toy program. A too strict language leaves the programmer struggling
with the language, whereas the “problem to solve” should be the real struggle
and the language should be a simple means to express the solution in.
The goal of the pawn language is to provide the developer with an informal,
and convenient to use, mechanism to test whether the program behaves
A tutorial introduction
as was intended. This mechanism is called “assertions” and, although the
concept of assertions pre-dates the idea of “design by contract”, it is most
easily explained through the design-by-contract methodology.
The “design by contract” paradigm provides an alternative approach for
dealing with erroneous conditions. The premise is that the programmer
knows the task at hand, the conditions under which the software must operate and the environment. In such an environment, each function specifies
the specific conditions, in the form of assertions, that must hold true before
a client may execute the function. In addition, the function may also specify
any conditions that hold true after it completes its operation. This is the
“contract” of the function.
The name “design by contract” was coined by Bertrand Meyer and its principles trace back to predicate logic and algorithmic analysis.
⋄ Preconditions specify the valid values of the input parameters and environmental attributes;
⋄ Postconditions specify the output and the (possibly modified) environment;
⋄ Invariants indicate the conditions that must hold true at key points in a
function, regardless of the path taken through the function.
For example, a function that computes a square root of a number may specify
that its input parameter be non-negative. This is a precondition. It may also
specify that its output, when squared, is the input value ±0.01%. This is a
postcondition; it verifies that the routine operated correctly. A convenient
way to calculate a square root is via “bisection”. At each iteration, this
algorithm gives at least one extra bit (binary digit) of accuracy. This is an
invariant (it might be an invariant that is hard to check, though).
Preconditions, postconditions and invariants are similar in the sense that
they all consist of a test and that a failed test indicates an error in the implementation. As a result, you can implement preconditions, postconditions
and invariants with a single construct: the “assertion”. For preconditions,
write assertions at the very start of the routine; for invariants, write an assertion where the invariant should hold; for post conditions, write an assertion
before each “return” statement or at the end of the function.
In pawn, the instruction is called assert; it is a simple statement that
contains a test. If the test outcome is “true”, nothing happens. If the
Example square
root function (using bisection): 77
A tutorial introduction
outcome is “false”, the assert instruction terminates the program with a
message containing the details of the assertion that failed.
Assertions are checks that should never fail. Genuine errors, such as user
input errors, should be handled with explicit tests in the program, and not
with assertions. As a rule, the expressions contained in assertions should
be free of side effects: an assertion should never contain code that your
application requires for correct operation.
This does have the effect, however, that assertions never fire in a bug-free
program: they just make the code fatter and slower , without any user-visible
benefit. It is not this bad, though. An additional feature of assertions is that
you can build the source code without assertions simply using a flag or option
to the pawn parser. The idea is that you enable assertions during development and build the “retail version” of the code without assertions. This is
a better approach than removing the assertions, because all assertions are
automatically “back” when recompiling the program —e.g. for maintenance.
During maintenance, or even during the initial development, if you catch a
bug that was not trapped by an assertion, before fixing the bug, you should
think of how an assertion could have trapped this error. Then, add this
assertion and test whether it indeed catches the bug before fixing the bug.
By doing this, the code will gradually become sturdier and more reliable.
• Documentation comments
When programs become larger, documenting the program and the functions
becomes vital for its maintenance, especially when working in a team. The
pawn language tools have some features to assist you in documenting the
code in comments. Documenting a program or library in its comments has a
few advantages —for example: documentation is more easily kept up to date
with the program, it is efficient in the sense that programming comments
now double as documentation, and the parser helps your documentation
efforts in generating syntax descriptions and cross references.
Comment syntax:
Every comment that starts with three slashes (“/// ”) followed by whitespace, or that starts with a slash and two stars (“/** ”) followed by whitespace is a special documentation comment. The pawn compiler extracts
documentation comments and optionally writes these to a “report” file. See
the application documentation, or appendix B, how to enable the report
A tutorial introduction
As an aside, comments that start with “/**” must still be closed with “*/”.
Single line documentation comments (“///”) close at the end of the line.
The report file is an XML file that can subsequently be transformed to
HTML documentation via an XSL/XSLT stylesheet, or be run through other
tools to create printed documentation. The syntax of the report file is compatible with that of the “.Net” developer products —except that the pawn
compiler stores more information in the report than just the extracted documentation strings. The report file contains a reference to the “smalldoc.xsl” stylesheet.
The example below illustrates documentation comments in a simple script
that has a few functions. You may write documentation comments for a
function above its declaration or in its body. All documentation comments
that appear before the end of the function are attributed to the function.
You can also add documentation comments to global variables and global
constants —these comments must appear above the declaration of the variable or constant. Figure 2 shows part of the output for this (rather long)
example. The style of the output is adjustable in the cascading style sheet
(CSS-file) associated with the XSLT transformation file.
* This program illustrates Zeller’s congruence algorithm to calculate
* the day of the week given a date.
* <summary>
The main program: asks the user to input a date and prints on
what day of the week that date falls.
* </summary>
new day, month, year
if (readdate(day, month, year))
new wkday = weekday(day, month, year)
printf "The date %d-%d-%d falls on a ", day, month, year
switch (wkday)
case 0:
print "Saturday"
case 1:
print "Sunday"
case 2:
A tutorial introduction
case 3:
case 4:
case 5:
case 6:
print "Invalid date"
print "\n"
* <summary>
The core function of Zeller’s congruence algorithm. The function
works for the Gregorian calender.
* </summary>
* <param name="day">
The day in the month, a value between 1 and 31.
* </param>
* <param name="month">
The month: a value between 1 and 12.
* </param>
* <param name="year">
The year in four digits.
* </param>
* <returns>
The day of the week, where 0 is Saturday and 6 is Friday.
* </returns>
* <remarks>
This function does not check the validity of the date; when the
date in the parameters is invalid, the returned "day of the week"
will hold an incorrect value.
This equation fails in many programming languages, notably most
implementations of C, C++ and Pascal, because these languages have
a loosely defined "remainder" operator. Pawn, on the other hand,
provides the true modulus operator, as defined in mathematical
theory and as was intended by Zeller.
* </remarks>
weekday(day, month, year)
A tutorial introduction
* <remarks>
For Zeller’s congruence algorithm, the months January and
February are the 13th and 14th month of the <em>preceding</em>
year. The idea is that the "difficult month" February (which
has either 28 or 29 days) is moved to the end of the year.
* </remarks>
if (month <= 2)
month += 12, --year
new j = year % 100
new e = year / 100
return (day + (month+1)*26/10 + j + j/4 + e/4 - 2*e) % 7
* <summary>
Reads a date and stores it in three separate fields.
* </summary>
* <param name="day">
Will hold the day number upon return.
* </param>
* <param name="month">
Will hold the month number upon return.
* </param>
* <param name="year">
Will hold the year number upon return.
* </param>
* <returns>
<em>true</em> if the date is valid, <em>false</em> otherwise;
if the function returns <em>false</em>, the values of
<paramref name="day"/>, <paramref name="month"/> and
<paramref name="year"/> cannot be relied upon.
* </returns>
bool: readdate(&day, &month, &year)
print "Give a date (dd-mm-yyyy): "
day = getvalue(_,’-’,’/’)
month = getvalue(_,’-’,’/’)
year = getvalue()
return 1 <= month <= 12 && 1 <= day <= daysinmonth(month,year)
* <summary>
Returns whether a year is a leap year.
* </summary>
* <param name="year">
The year in 4 digits.
A tutorial introduction
* </param>
* <remarks>
A year is a leap year:
<li> if it is divisable by 4, </li>
<li> but <strong>not</strong> if it is divisable by 100, </li>
<li> but it <strong>is</strong> it is divisable by 400. </li>
* </remarks>
bool: isleapyear(year)
return year % 400 == 0 || year % 100 != 0 && year % 4 == 0
* <summary>
Returns the number of days in a month (the month is an integer
in the range 1 .. 12). One needs to pass in the year as well,
because the function takes leap years into account.
* </summary>
* <param name="month">
The month number, a value between 1 and 12.
* </param>
* <param name="year">
The year in 4 digits.
* </param>
daysinmonth(month, year)
static daylist[] = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ]
assert 1 <= month <= 12
return daylist[month-1] + _:(month == 2 && isleapyear(year))
The format of the XML file created by “.Net” developer products is documented in the Microsoft documentation. The pawn parser creates a minimal
description of each function or global variable or constant that is used in a
project, regardless of whether you used documentation comments on that
function/variable/constant. The parser also generates few tags of its own:
Attributes for a function, such as “native” or “stock”.
The automaton that the function belongs to (if any).
The names of the symbols (other functions, global variables
and/global constants) that the function requires. If desired, a
call tree can be constructed from the dependencies.
A tutorial introduction
Figure 2: Documentation generated from the source code
Function parameters. When you add a parameter description in a documentation comment, this description is combined
with the auto-generated content for the parameter.
Tags and array or reference information on a parameter.
All functions that refer to this symbol; i.e., all functions that
use or call this variable/function. This information is sufficient to serve as a “cross-reference” —the “referrer” tree is the
inverse of the “dependency” tree.
The estimated number of cells that the function will allocate
on the stack and heap. This stack usage estimate excludes
the stack requirements of any functions that are “called” from
the function to which the documentation applies. For example,
A tutorial introduction
function readdate is documented as taking 6 cells on the stack,
but it also calls daysinmonth which takes 4 additional cells —
and in turn calls isleapyear. To calculate the total stack
requirements for function readdate, the call tree should be
In addition to the local variables and function parameters, the
compiler also uses the stack for storing intermediate results in
complex expressions. The stack space needed for these intermediate results are also excluded from this report. In general,
the required overhead for the intermediate results is not cumulative (over all functions), which is why it would be inaccurate
to add a “safety margin” to every function. For the program
as a whole, a safety margin would be highly advised. See appendix B (page 169) for the -v option which can tell you the
maximum estimate stack usage, based on the call tree.
The tag of the constant, variable, function result or function
The transitions that the function provokes and their conditions
—see the section State programming on page 36.
All text in the documentation comment(s) is also copied to each function,
variable or constant to which it is attached. The text in the documentation
comment is copied without further processing —with one exception, see
below. As the rest of the report file is in XML format, and the most suitable
way to process XML to on-line documentation is through an XSLT processor
(such as a modern browser), you may choose to do any formatting in the
documentation comments using HTML tags. Note that you will often need
to explicitly close any HTML tags; the HTML standard does not require
this, but XML/XSLT processors usually do. The pawn toolkit comes with
an example XSLT file (with a matching style sheet) which supports the
following XML/HTML tags:
<code> </code>
Formatted source code in a monospaced
font; although the “&”, “<” and “>” must
be typed as “&amp;”, “&lt;” and “&rt;”
<example> </example>
Text set under the topic “Example”.
<param name="..."> </param> A parameter description, with the parameter name appearing inside the opening
A tutorial introduction
<paramref name="..." />
<section> </section>
<subsection> </subsection>
tag (the “name=” option) and the parameter description following it.
A reference to a parameter, with the parameter name appearing inside the opening tag (the “name=” option).
Text set under the topic “Remarks”.
Text set under the topic “Returns”.
Text set under the topic “See also”.
Text set immediately below the header of
the symbol.
Sets the text in a header. This should
only be used in documentation that is not
attached to a function or a variable.
Sets the text in a sub-header. This should
only be used in documentation that is not
attached to a function or a variable.
The following additional HTML tags are supported for general purpose formatting text inside any of the above sections:
<c> </c>
Text set in a monospaced font.
<em> </em>
Text set emphasized, usually in italics.
<p> </p>
Text set in a new paragraph. Instead of wrapping <p>
and </p> around every paragraph, inserting <p/> as a
separator between two paragraphs produces the same
<para> </para>
An alternative for <p> </p>
<ul> </ul>
An unordered (bulleted) list.
<ol> </ol>
An ordered (numbered) list.
<li> </li>
An item in an ordered or unordered list.
As stated, there is one exception in the processing of documentation comments: if your documentation comment contains a <param ...> tag (and a
matching </param>), the pawn parser looks up the parameter and combines
your description of the parameter with the contents that it has automatically
• Warnings and errors
The big hurdle that I have stepped over is how to actually compile the
code snippets presented in this chapter. The reason is that the procedure
A tutorial introduction
depends on the system that you are using: in some applications there is a
“Make” or “Compile script” command button or menu option, while in other
environments you have to type a command like “pawncc myscript” on a
command prompt. If you are using the standard pawn toolset, you will
find instructions of how to use the compiler and run-time in the companion
booklet “The pawn booklet — Implementer’s Guide”. If you are using
Microsoft Windows, it may prove the most convenient to use the Quincy
IDE that comes with pawn for writing, running and debugging scripts.
Regardless of the differences in launching the compile, the phenomenon that
results from launching the compile are likely to be very similar between all
⋄ either the compile succeeds and produces an executable program —that
may or may not run automatically after the compile;
⋄ or the compile gives a list of warning and error messages.
Mistakes happen and the pawn parser tries to catch as many of them as it
can. When you inspect the code that the pawn parser complains about, it
may on occasion be rather difficult for you to see why the code is erroneous
(or suspicious). The following hints may help:
⋄ Each error or warning number is numbered. You can look up the error
message with this number in appendix A, along with a brief description
on what the message really means.
⋄ If the pawn parser produces a list of errors, the first error in this list is a
true error, but the diagnostic messages below it may not be errors at all.
After the pawn parser sees an error, it tries to step over it and complete the
compilation. However, the stumbling on the error may have confused the
pawn parser so that subsequent legitimate statements are misinterpreted
and reported as errors too.
When in doubt, fix the first error and recompile.
⋄ The pawn parser checks only the syntax (spelling/grammar), not the semantics (i.e. the “meaning”) of the code. When it detects code that does
not comply to the syntactical rules, there may actually be different ways
in which the code can be changed to be “correct”, in the syntactical sense
of the word —even though many of these “corrections” would lead to
nonsensical code. The result is, though, that the pawn parser may have
difficulty to precisely locate the error: it does not know what you meant
A tutorial introduction
to write. Hence, the parser often outputs two line numbers and the error
is somewhere in the range (between the line numbers).
⋄ Remember that a program that has no syntactical errors (the pawn parser
accepts it without error & warning messages) may still have semantical and
logical errors which the pawn parser cannot catch. The assert instruction
(page 111) is meant to help you catch these “run-time” errors.
• In closing
If you know the C programming language, you will have seen many concepts
that you are familiar with, and a few new ones. If you don’t know C, the
pace of this introduction has probably been quite high. Whether you are
new to C or experienced in C, I encourage you to read the following pages
carefully. If you know C or a C-like language, by the way, you may want to
consult the chapter Pitfalls (page 133) first.
This booklet attempts to be both an informal introduction and a (more formal) language specification at the same time, perhaps succeeding at neither.
Since it is also the standard book on pawn,∗ the focus of this booklet is on
being accurate and complete, rather than being easy to grasp.
The double nature of this booklet shows through in the order in which it
presents the subjects. The larger conceptual parts of the language, variables
and functions, are covered first. The operators, the statements and general
syntax rules follow later —not that they are less important, but they are
easier to learn, to look up, or to take for granted.
It is no longer the only book on Pawn.
Data and declarations
pawn is a typeless language. All data elements are of type “cell”, and a
cell can hold an integral number. The size of a cell (in bytes) is system
dependent —usually, a cell is 32-bits.
The keyword new declares a new variable. For special declarations, the keyword new is replaced by static, public or stock (see below). A simple
variable declaration creates a variable that occupies one “cell” of data memory. Unless it is explicitly initialized, the value of the new variable is zero.
“for” loop: 112
Compound statement: 111
A variable declaration may occur:
⋄ at any position where a statement would be valid —local variables;
⋄ at any position where a function declaration (native function declarations)
or a function implementation would be valid —global variables;
⋄ in the first expression of a for loop instruction —also local variables.
Local declarations
A local declaration appears inside a compound statement. A local
variable can only be accessed from within the compound statement,
and from nested compound statements. A declaration in the first
expression of a for loop instruction is also a local declaration.
Global declarations
A global declaration appears outside a function and a global variable is accessible to any function. Global data objects can only be
initialized with constant expressions.
• State variable declarations
A state variable is a global variable with a state classifier appended at the
end. The scope and the lifespan of the variable are restricted to the states
that are listed in the classifier. Fall-back state specifiers are not permitted
for state variables.
State variables may not be initialized. In contrast to normal variables (which
are zero after declaration —unless explicitly initialized), state variables hold
an indeterminate value after declaration and after first entering a state in
its classifier. Typically, one uses the state entry function(s) to properly
initialize the state variable, and the exit function(s) to reset these variables.
Data and declarations
• Static local declarations
A local variable is destroyed when the execution leaves the compound block
in which the variable was created. Local variables in a function only exist
during the run time of that function. Each new run of the function creates
and initializes new local variables. When a local variable is declared with
the keyword static rather than new, the variable remains in existence after
the end of a function. This means that static local variables provide private, permanent storage that is accessible only from a single function (or
compound block). Like global variables, static local variables can only be
initialized with constant expressions.
• Static global declarations
A static global variable behaves the same as a normal global variable, except
that its scope is restricted to the file that the declaration resides in. To
declare a global variable as static, replace the keyword new by static.
• Stock declarations
A global variable may be declared as “stock”. A stock declaration is one
that the parser may remove or ignore if the variable turns out not to be used
in the program.
Stock variables are useful in combination with stock functions. A public
variable may be declared as “stock” as well —declaring public variables as
“public stock” enables you to declare al public variables that a host application provides in an include file, with only those variables that the script
actually uses winding up in the P-code file.
• Public declarations
Global “simple” variables (no arrays) may be declared “public” in two ways:
⋄ declare the variable using the keyword public instead of new;
⋄ start the variable name with the “@” symbol.
Public variables behave like global variables, with the addition that the host
program can also read and write public variables. A (normal) global variable
can only be accessed by the functions in your script —the host program is
unaware of them. As such, a host program may require that you declare a
Stock functions:
Data and declarations
variable with a specific name as “public” for special purposes —such as the
most recent error number, or the general program state.
• Constant variables
Symbolic constants: 101
It is sometimes convenient to be able to create a variable that is initialized
once and that may not be modified. Such a variable behaves much like a
symbolic constant, but it still is a variable.
To declare a constant variable, insert the keyword const between the keyword that starts the variable declaration —new, static, public or stock—
and the variable name.
new const address[4] = { 192, 0, 168, 66 }
public const status
/* initialized to zero */
Three typical situations where one may use a constant variable are:
⋄ To create an “array” constant; symbolic constants cannot be indexed.
⋄ For a public variable that should be set by the host application, and only
by the host application. See the preceding section for public variables.
⋄ A special case is to mark array arguments to functions as const. Array
arguments are always passed by reference, declaring them as const guards
against unintentional modification. Refer to page 71 for an example of
const function arguments.
• Arrays (single dimension)
See also “multidimensional arrays”, page 65,
and “symbolic
subscripts”, page
The syntax name[constant] declares name to be an array of “constant”
elements, where each element is a single cell. The name is a placeholder
of an identifier name of your choosing and constant is a positive non-zero
value; constant may be absent. If there is no value between the brackets,
the number of elements is set equal to the number of initiallers —see the
example below.
The array index range is “zero based” which means that the first element is
at name[0] and the last element is name[constant-1].
The syntax name{constant} also declares name as an array of constant
elements, but now the elements are characters rather than cells. The number
of characters that fit in a cell depends on the configuration of the pawn
Data and declarations
• Initialization
Data objects can be initialized at their declaration. The initialler of a global
data object must be a constant. Arrays, global or local, must also be initialized with constants.
Uninitialized data defaults to zero.
good declaration
new i = 1
new j
new k = ’a’
/* j is zero */
/* k has character code for letter ’a’ */
new a[] = [1,4,9,16,25]
new s1[20] = [’a’,’b’]
/* a has 5 elements */
/* the other 18 elements are 0 */
new s2[] = ’’Hello world...’’
/* an unpacked string */
Examples of invalid declarations:
bad declarations
c[3] = 4
i = "Good-bye"
p[2] = { i + j, k - 3 }
an array cannot be set to a value */
only an array can hold a string */
unknown size of array */
array initiallers must be constants */
• Progressive initiallers for arrays
The ellipsis operator continues the progression of the initialisation constants
for an array, based on the last two initialized elements. The ellipsis operator
(three dots, or “...”) initializes the array up to its declared size.
array initializers
a[10] = { 1, ... }
b[10] = { 1, 2, ... }
c[8] = { 1, 2, 40, 50, ... }
d[10] = { 10, 9, ... }
sets all ten elements to 1
b = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
c = 1, 2, 40, 50, 60, 70, 80, 90
d = 10, 9, 8, 7, 6, 5, 4, 3, 2, 1
Constants: 97
Data and declarations
• Symbolic subscripts for arrays
An array may be declared with a list of symbols instead of a value for its
size: an example of this is the “priority queue” sample program on page 19.
An individual subscript may also be interpreted as a sub-arrays, for example,
see the rpn calculator program at page 28.
The sub-array syntax applies as well to the initialization of an array with
symbolic subscripts. Referring again to the “priority queue” sample program, to initialize a “message” array with fixed values, the syntax is:
array initializers
new msg[.text{40}, .priority] = { "new message", 1 }
The initialler consists of a string (a literal array) and a value; these go into
the fields “.text” and “.priority” respectively.
An array dimension that is declared as a list of symbolic subscripts, may only
be indexed with these subscripts. From the above declaration of variable
“msg”, we may use:
array initializers
msg[.text] = "another message"
msg[.priority] = 10 - msg[.priority]
It is an error, however, to use a (numeric) expression to index “msg”. For
example, “msg[1]” is an invalid expression.
Since an array with symbolic subscripts may not be indexed with an expression, the square brackets that enclose the expression become optional. These
brackets may be omitted. The snippet below is equivalent to the previous
array initializers
msg.text = "another message"
msg.priority = 10 - msg.priority
Tag names: 66
A subscript may have an explicit tag name as well. This tag will then
override the default tag for array elements. The rpn calculator program
makes use of this feature to mark one of the subscripts as a rational value.
In the declaration in the snippet below, the expression “field.type” is
a plain integer (without tag), but the expression “field.value” has tag
Data and declarations
array initializers
new field[
Rational: .value,
/* operator or token type */
/* value, if t_type is "Number" */
/* raw string */
• Multi-dimensional arrays
Multi-dimensional arrays are arrays that contain references to the sub-arrays.
That is, a two-dimensional array is an “array of single-dimensional arrays”.∗
Below are a few examples of declarations of two-dimensional arrays.
two-dimensional arrays
[ 1, 2 ], [ 3, 4 ], [ 5, 6 ] ]
[ 1 ], [ 2, ...], [ 3, 4, ... ] ]
"agreement", "dispute" ]
’’OK’’, ’’Cancel’’ ]
’’OK’’, ’’Cancel’’ ]
As the last two declarations (variable “e” en “f”) show, the final dimension
of an array may have an unspecified length, in which case the length of
each sub-array is determined from the related initializer. Every sub-array
may have a different size; in this particular example, “e[1][5]” contains
the letter “l” from the word “Cancel”, but “e[0][5]” is invalid because the
length of the sub-array “e[0]” is only three cells (containing the letters “O”,
“K” and a zero terminator).
The difference between the declarations for arrays “e” and “f” is that we
let the compiler count the number of initializers for the major dimension
—“sizeof f” is 2, like “sizeof e” (see the next section on the sizeof
• Arrays and the sizeof operator
The sizeof operator returns the size of a variable in “elements”. For a
simple (non-compound) variable, the result of sizeof is always 1, because an
element is a cell for a simple variable.
The current implementation of the Pawn compiler supports only arrays with up to three
Data and declarations
An array with one dimension holds a number of cells and the sizeof operator
returns that number. The snippet below would therefore print “5” at the
display, because the array “msg” holds four characters (each in one cell) plus
a zero-terminator:
Listing: sizeof operator
new msg[] = ’’Help’’
printf(’’%d’’, sizeof msg);
The sizeof operator always returns the number of cells, even for a packed
array. That is, in the next snippet, the value printed would be less than “5”
—although there are five characters in the array, those are packed in fewer
Listing: sizeof operator
new msg{} = "Help"
printf(’’%d’’, sizeof msg);
With multi-dimensional arrays, the sizeof operator can return the number
of elements in each dimension. For the last (minor) dimension, an element
will again be a cell, but for the major dimension(s), an element is a sub-array.
In the following code snippet, observe that the syntax sizeof matrix refers
to the major dimension of the two-dimensional array and the syntax sizeof
matrix[] refers to the minor dimension of the array. The values that this
snippet prints are 3 and 2 (for the major and minor dimensions respectively):
Listing: sizeof operator and multidimensional arrays
new matrix[3][2] = { { 1, 2 }, { 3, 4 }, { 5, 6 } }
printf(’’%d %d’’, sizeof matrix, sizeof matrix[]);
Default function
arguments and
sizeof: 75
The application of the sizeof operator on multi-dimensional arrays is especially convenient when used as a default value for function arguments.
• Tag names
A tag is a label that denotes the objective of —or the meaning of— a variable,
a constant or a function result. Tags are optional, their only purpose is to
allow a stronger compile-time error checking of operands in expressions, of
function arguments and of array indices.
Label syntax: 111
A tag consists of a symbol name followed by a colon; it has the same syntax as
a label. A tag precedes the symbol name of a variable, constant or function.
In an assignment, only the right hand of the “=” sign may be tagged.
Examples of valid tagged variable and constant definitions are:
Data and declarations
tag names
new bool:flag = true
/* "flag" can only hold "true" or "false" */
const error:success = 0
const error:fatal= 1
const error:nonfatal = 2
error:errno = fatal
The sequence of the constants success, fatal and nonfatal could more
conveniently be declared by grouping the constants in a compount block, as
illustrated below. The declaration below creates the same three constants,
all with the tag error:. It is required to specify a value for the first constant
of the list, the subsequent constants are automatically assigned a value that
is the value of the previous constant +1 —unless an explicit value is present.
enumerated constants
const error: {
notice = 0,
new error: code
After declaring variable “code” with tag name “error:”, you can assign
any of the constants with that same tag name to it; however, writing “code
= 2” will give a parser diagnostic (a warning or error message). A tag
override (or a tag cast) adjusts an expression to the desired tag name. As a
somewhat contrived example, the next snippet elevates “code” to a higher
level (a “more serious error”) —note how the literal value 1 is forced to the
tag name “error:”
tag override
if (code < fatal)
code = code + error:1
Tag names introduced so far started with a lower case letter; these are “weak”
tags. Tag names that start with an upper case letter are “strong” tags.
The difference between weak and strong tags is that weak tags may, in
a few circumstances, be dropped implicitly by the pawn parser —so that a
weakly tagged expression becomes an untagged expression. The tag checking
mechanism verifies the following situations:
“const” statement: 101
“lvalue”: the variable on the left
side in an assignment, see page
Data and declarations
⋄ When the expressions on both sides of a binary operator have a different tag, or when one of the expressions is tagged and the other is not,
the compiler issues a “tag mismatch” diagnostic. There is no difference
between weak and strong tags in this situation.
⋄ There is a special case for the assignment operator: the compiler issues a
diagnostic if the variable on the left side of an assignment operator has a
tag, and the expression on the right side either has a different tag or is
untagged. However, if the variable on the left of the assignment operator
is untagged, it accepts an expression (on the right side) with a weak tag.
In other words, a weak tag is dropped in an assignment when the lvalue
is untagged.
⋄ Passing arguments to functions follows the rule for assignments. The
compiler issues a diagnostic when the formal parameter (in a function
definition) has a tag and the actual parameter (in the function call) either
is untagged or has a different tag. However, if the formal parameter is
untagged, it also accepts a parameter with any weak tag.
A function declaration specifies the name of the function and, between parentheses, its formal parameters. A function may also return a value. A function
declaration must appear on a global level (i.e. outside any other functions)
and is globally accessible.
If a semicolon follows the function declaration (rather than a statement),
the declaration denotes a forward declaration of the function.
The return statement sets the function result. For example, function sum
(see below) has as its result the value of both its arguments added together.
The return expression is optional for a function, but one cannot use the
value of a function that does not return a value.
The preferred way
to declare forward
functions is at
page 80
sum function
sum(a, b)
return a + b
Arguments of a function are (implicitly declared) local variables for that
function. The function call determines the values of the arguments.
Another example of a complete definition of the function leapyear (which
returns true for a leap year and false for a non-leap year):
leapyear function
return y % 4 == 0 && y % 100 != 0
y % 400 == 0
The logical and arithmetic operators used in the leapyear example are
covered on pages 107 and 103 respectively.
Usually a function contains local variable declarations and consists of a compound statement. In the following example, note the assert statement to
guard against negative values for the exponent.
power function (raise to a power)
power(x, y)
/* returns x raised to the power of y */
assert y >= 0
new r = 1
for (new i = 0; i < y; i++)
r *= x
return r
“assert” statement: 111
A function may contain multiple return statements —one usually does this
to quickly exit a function on a parameter error or when it turns out that the
function has nothing to do. If a function returns an array, all return statements must specify an array with the same size and the same dimensions.
• Function arguments (call-by-value versus call-by-reference)
Another example is function
JulianToDate at
page 11
The “faculty” function in the next program has one parameter which it uses
in a loop to calculate the faculty of that number. What deserves attention
is that the function modifies its argument.
/* Calculation of the faculty of a value */
print "Enter a value: "
new v = getvalue()
new f = faculty(v)
printf "The faculty of %d is %d\n", v, f
assert n >= 0
new result = 1
while (n > 0)
result *= n-return result
Whatever (positive) value that “n” had at the entry of the while loop in
function faculty, “n” will be zero at the end of the loop. In the case of the
faculty function, the parameter is passed “by value”, so the change of “n”
is local to the faculty function. In other words, function main passes “v”
as input to function faculty, but upon return of faculty, “v” still has the
same value as before the function call.
Arguments that occupy a single cell can be passed by value or by reference.
The default is “pass by value”. To create a function argument that is passed
by reference, prefix the argument name with the character &.
swap function
swap(&a, &b)
new temp = b
b = a
a = temp
To pass an array to a function, append a pair of brackets to the argument
name. You may optionally indicate the size of the array; doing so improves
error checking of the parser.
addvector function
addvector(a[], const b[], size)
for (new i = 0; i < size; i++)
a[i] += b[i]
Arrays are always passed by reference. As a side note, array b in the above
example does not change in the body of the function. The function argument
has been declared as const to make this explicit. In addition to improving
error checking, it also allows the pawn parser to generate more efficient code.
Constant variables: 62
To pass an array of literals to a function, use the same syntax as for array
initiallers: a literal string or the series of array indices enclosed in braces
(see page 99; the ellipsis for progressive initiallers cannot be used). Literal
arrays can only have a single dimension.
The following snippet calls addvector to add five to every element of the
array “vect”:
addvector usage
new vect[3] = [ 1, 2, 3 ]
addvector(vect, [5, 5, 5], 3)
/* vect[] now holds the values 6, 7 and 8 */
The invocation of function printf with the string "Hello world\n" in the
first ubiquitous program is another example of passing a literal array to a
“Hello world” program: 3
• Calling functions
When inserting a function name with its parameters in a statement or expression, the function will get executed in that statement/expression. The
statement that refers to the function is the “caller” and the function itself,
at that point, is the “callee”: the one being called.
Function power:
The standard syntax for calling a function is to write the function’s name,
followed by a list with all explicitly passed parameters between parentheses.
If no parameters are passed, or if the function does not have any, the pair
of parentheses behind the function name are still present. For example, to
try out the power function, the following program calls it thus:
example program for the power function
print "Please give the base value and the power to raise it to:"
new base = getvalue()
new power = getvalue()
new result = power(base, power)
printf "%d raised to the power %d is %d", base, power, result
Functions sum &
leapyear: 69
Function swap: 70
A function may optionally return a value. The sum, leapyear and power
functions all return a value, but the swap function does not. Even if a
function returns a value, the caller may ignore it.
For the situation that the caller ignores the function’s return value, there
is an alternative syntax to call the function, which is also illustrated by
the preceding example program calls the power function. The parentheses
around all function arguments are optional if the caller does not use the
return value. In the last statement, the example program reads
printf "%d raised to the power %d is %d", base, power, result
rather than
printf("%d raised to the power %d is %d", base, power, result)
which does the same thing.
The syntax without parentheses around the parameter list is called the “procedure call” syntax. You can use it only if:
⋄ the caller does not assign the function’s result to a variable and does not
use it in an expression, or as the “test expression” of an if statement for
⋄ the first parameter does not start with an opening parenthesis;
⋄ the first parameter is on the same line as the function name, unless you
use named parameters (see the next section).
As you may observe, the procedure call syntax applies to cases where a
function call behaves rather as a statement, like in the calls to print and
printf in the preceding example. The syntax is aimed at making such
statements appear less cryptic and friendlier to read, but not that the use
of the syntax is optional.
As a side note, all parentheses in the example program presented in this
section are required: the return values of the calls to getvalue are stored
in two variables, and therefore an empty pair of parentheses must follow the
function name. Function getvalue has optional parameters, but none are
passed in this example program.
• Named parameters versus positional parameters
In the previous examples, the order of parameters of a function call was important, because each parameter is copied to the function argument with the
same sequential position. For example, with the function weekday (which
uses Zeller’s congruence algorithm) defined as below, you would call weekday(12,31,1999) to get the week day of the last day of the preceding century.
weekday function
weekday(month, day, year)
/* returns the day of the week: 0=Saturday, 1=Sunday, etc. */
if (month <= 2)
month += 12, --year
new j = year % 100
new e = year / 100
return (day + (month+1)*26/10 + j + j/4 + e/4 - 2*e) % 7
Date formats vary according to culture and nation. While the format month/
day/year is common in the United States of America, European countries
often use the day/month/year format, and technical publications sometimes
standardize on the year/month/day format (ISO/IEC 8824). In other words,
no order of arguments in the weekday function is “logical” or “conventional”.
That being the case, the alternative way to pass parameters to a function
is to use “named parameters”, as in the next examples (the three function
calls are equivalent):
weekday usage —positional parameters
new wkday1 = weekday( .month = 12, .day = 31, .year = 1999)
new wkday2 = weekday( .day = 31, .month = 12, .year = 1999)
new wkday3 = weekday( .year = 1999, .month = 12, .day = 31)
With named parameters, a period (“.”) precedes the name of the function
argument. The function argument can be set to any expression that is
valid for the argument. The equal sign (“=”) does in the case of a named
parameter not indicate an assignment; rather it links the expression that
follows the equal sign to one of the function arguments.
One may mix positional parameters and named parameters in a function call
with the restriction that all positional parameters must precede any named
• Default values of function arguments
Public functions
do not support
default argument
values; see page
A function argument may have a default value. The default value for a
function argument must be a constant. To specify a default value, append
the equal sign (“=”) and the value to the argument name.
When the function call specifies an argument placeholder instead of a valid
argument, the default value applies. The argument placeholder is the underscore character (“_”). The argument placeholder is only valid for function
arguments that have a default value.
The rightmost argument placeholders may simply be stripped from the function argument list. For example, if function increment is defined as:
increment function —default values
increment(&value, incr=1) value += incr
the following function calls are all equivalent:
increment usage
increment(a, _)
increment(a, 1)
Default argument values for passed-by-reference arguments are useful to
make the input argument optional. For example, if the function divmod
is designed to return both the quotient and the remainder of a division operation through its arguments, default values make these arguments optional:
divmod function —default values for reference parameters
divmod(a, b, &quotient=0, &remainder=0)
quotient = a / b
remainder = a % b
With the preceding definition of function divmod, the following function calls
are now all valid:
Listing: divmod usage
new p, q
divmod 10,
Default arguments for array arguments are often convenient to set a default
string or prompt to a function that receives a string argument. For example:
print error function
print_error(const message[], const title[] = "Error: ")
print title
print message
print "\n"
The next example adds the fields of one array to another array, and by
default increments the first three elements of the destination array by one:
addvector function, revised
addvector(a[], const b[] = {1, 1, 1}, size = 3)
for (new i = 0; i < size; i++)
a[i] += b[i]
• sizeof operator & default function arguments
A default value of a function argument must be a constant, and its value is
determined at the point of the function’s declaration. Using the “sizeof”
operator to set the default value of a function argument is a special case:
the calculation of the value of the sizeof expression is delayed to the point
of the function call and it takes the size of the actual argument rather than
that of the formal argument. When the function is used several times in a
“sizeof” operator
program, with different arguments, the outcome of the “sizeof” expression
is potentially different at every call —which means that the “default value”
of the function argument may change.
Below is an example program that draws ten random numbers in the range of
0–51 without duplicates. An example for an application for drawing random
numbers without duplicates is in card games —those ten numbers could
represent the cards for two “hands” in a poker game. The virtues of the
algorithm used in this program, invented by Robert W. Floyd, are that it is
efficient and unbiased —provided that the pseudo-random number generator
is unbiased as well.
Listing: randlist.p
new HandOfCards[10]
FillRandom(HandOfCards, 52)
print "A draw of 10 numbers from a range of 0 to 51 \
(inclusive) without duplicates:\n"
for (new i = 0; i < sizeof HandOfCards; i++)
printf "%d ", HandOfCards[i]
“random” is a
proposed core
function, see page
FillRandom(Series[], Range, Number = sizeof Series)
assert Range >= Number
/* cannot select 50 values
* without duplicates in the
* range 0..40, for example */
new Index = 0
for (new Seq = Range - Number; Seq < Range; Seq++)
new Val = random(Seq + 1)
new Pos = InSeries(Series, Val, Index)
if (Pos >= 0)
Series[Index] = Series[Pos]
Series[Pos] = Seq
Series[Index] = Val
InSeries(Series[], Value, Top = sizeof Series)
for (new i = 0; i < Top; i++)
if (Series[i] == Value)
return i
return -1
Function main declares the array HandOfCards with a size of ten cells and
then calls function FillRandom with the purpose that it draws ten positive
random numbers below 52. Observe, however, that the only two parameters that main passes into the call to FillRandom are the array HandOfCards, where the random numbers should be stored, and the upper bound
“52”. The number of random numbers to draw (“10”) is passed implicitly
to FillRandom.
Array declarations: 62
The definition of function FillRandom below main specifies for its third parameter “Number = sizeof Series”, where “Series” refers to the first
parameter of the function. Due to the special case of a “sizeof default
value”, the default value of the Number argument is not the size of the formal argument Series, but that of the actual argument at the point of the
function call: HandOfCards.
Note that inside function FillRandom, asking the “sizeof” the function
argument Series would (still) evaluate in zero, because the Series array is
declared with unspecified length (see page 108 for the behaviour of sizeof).
Using sizeof as a default value for a function argument is a specific case.
If the formal parameter Series were declared with an explicit size, as in
Series[10], it would be redundant to add a Number argument with the
array size of the actual argument, because the parser would then enforce
that both formal and actual arguments have the size and dimensions.
• Arguments with tag names
A tag optionally precedes a function argument. Using tags improves the
compile-time error checking of the script and it serves as “implicit documentation” of the function. For example, a function that computes the square
root of an input value in fixed point precision may require that the input
parameter is a fixed point value and that the result is fixed point as well.
The function below uses the fixed point extension module, and an approximation algorithm known as “bisection” to calculate the square root. Note
the use of tag overrides on numeric literals and expression results.
sqroot function —strong tags
Fixed: sqroot(Fixed: value)
new Fixed: low = 0.0
new Fixed: high = value
Tag names: 66
Fixed point arithmetic: 89;
see also the application note “Fixed
Point Support Library”
while (high - low > Fixed: 1)
new Fixed: mid = (low + high) >> 1
if (fmul(mid, mid) < value)
low = mid
high = mid
return low
With the above definition, the pawn parser issues a diagnostic if one calls
the sqroot function with a parameter with a tag different from “Fixed:”, or
when it tries to store the function result in a variable with a “non-Fixed:”
The bisection algorithm is related to binary search, in the sense that it
continuously halves the interval in which the result must lie. A “successive
substitution” algorithm like Newton-Raphson, that takes the slope of the
function’s curve into account, achieves precise results more quickly, but at
the cost that a stopping criterion is more difficult to state. State of the
art algorithms for computing square roots combine bisection and NewtonRaphson algorithms.
• Variable arguments
A function that takes a variable number of arguments, uses the “ellipsis”
operator (“...”) in the function header to denote the position of the first
variable argument. The function can access the arguments with the predefined functions numargs, getarg and setarg (see page 123).
Function sum returns the summation of all of its parameters. It uses a
variable length parameter list.
sum function, revised
new result = 0
for (new i = 0; i < numargs(); ++i)
result += getarg(i)
return result
This function could be used in:
sum function usage
new v = sum(1, 2, 3, 4, 5)
A tag may precede the ellipsis to enforce that all subsequent parameters
have the same tag, but otherwise there is no error checking with a variable
argument list and this feature should therefore be used with caution.
Tag names: 66
The functions getarg and setarg assume that the argument is passed “by
reference”. When using getarg on normal function parameters (instead of
variable arguments) one should be cautious of this, as neither the compiler
nor the abstract machine can check this. Actual parameters that are passed
as part of a “variable argument list” are always passed by reference.
• Coercion rules
the function argument, as per the function definition (or its declaration),
a “value parameter”, the caller can pass as a parameter to the function:
a value, which is passed by value;
a reference, whose dereferenced value is passed;
an (indexed) array element, which is a value.
If the function argument is a reference, the caller can pass to the function:
⋄ a value, whose address is passed;
⋄ a reference, which is passed by value because it has the type that the
function expects;
⋄ an (indexed) array element, which is a value.
If the function argument is an array, the caller can pass to the function:
⋄ an array with the same dimensions, whose starting address is passed;
⋄ an (indexed) array element, in which case the address of the element is
• Recursion
A faculty example function earlier in this chapter used a simple loop. An
example function that calculated a number from the Fibonacci series also
used a loop and an extra variable to do the trick. These two functions are the
most popular routines to illustrate recursive functions, so by implementing
these as iterative procedures, you might be inclined to think that pawn does
not support recursion.
“faculty”: 70
“fibonacci”: 9
Well, pawn does support recursion, but the calculation of faculties and of
Fibonacci numbers happen to be good examples of when not to use recursion.
Faculty is easier to understand with a loop than it is with recursion. Solving
Fibonacci numbers by recursion indeed simplifies the problem, but at the
cost of being extremely inefficient: the recursive Fibonacci calculates the
same values over and over again.
There exists an
intriguing iterative
solution to the
Towers of Hanoi.
The program below is an implementation of the famous “Towers of Hanoi”
game in a recursive function:
/* The Towers of Hanoi, a game solved through recursion */
print "How many disks: "
new disks = getvalue()
move 1, 3, 2, disks
move(from, to, spare, numdisks)
if (numdisks > 1)
move from, spare, to, numdisks-1
printf "Move disk from pillar %d to pillar %d\n", from, to
if (numdisks > 1)
move spare, to, from, numdisks-1
• Forward declarations
Forbidden userdefined operators:
For standard functions, the current “reference implementation” of the pawn
compiler does not require functions to be declared before their first use.∗
User-defined operators are special functions, and unlike standard functions
they must be declared before use. In many cases it is convenient to put
the implementation of a user-defined operator in an include file, so that the
implementation and declaration precedes any call/invocation. Sometimes, it
may however be required (or convenient) to declare a user- defined operator
first and implement it elsewhere. A particular use of this technique is to
implement “forbidden” user-defined operators.
Other implementations of the Pawn language (if they exist) may use “single pass” parsers,
requiring functions to be defined before use.
To create a forward declaration, precede the function name and its parameter list with the keyword forward. For compatibility with early versions
of pawn, and for similarity with C/C++ , an alternative way to forwardly
declare a function is by typing the function header and terminating it with
a semicolon (which follows the closing parenthesis of the parameter list).
The full definition of the function, with a non-empty body, is implemented
elsewhere in the source file (except for forbidden user-defined operators).
State classifiers are ignored on forward declarations.
• State classifiers
All functions except native functions may optionally have a state attribute.
This consists of a list of state (and automata) names between angle brackets
behind the function header. The names are separated by commas. When
the state is part of a non-default automaton, the name of the automaton
and a colon separator must precede the state; for example, “parser:slash”
stands for the state slash of the automaton parser.
If a function has states, there must be several “implementations” of the
function in the source code. All functions must have the same function
header (excluding the state classifier list).
As a special syntax, when there are no names between the angle brackets,
the function is linked to all states that are not attributed to other implementations of the function. The function that handles “all states not handled
elsewhere” is the so-called fall-back function.
• Public functions, function main
A stand-alone program must have the function main. This function is the
starting point of the program. The function main may not have arguments.
A function library need not to have a main function, but it must have it
either a main function, or at least one public function. Function main is
the primary entry point into the compiled program; the public functions
are alternative entry points to the program. The virtual machine can start
execution with one of the public functions. A function library may have a
main function to perform one-time initialization at start-up.
To make a function public, prefix the function name with the keyword public. For example, a text editor may call the public function “onkey” for
Example: 38
every key that the user typed in, so that the user can change (or reject)
keystrokes. The onkey function below would replace every “~” character
(code 126 in the ISO Latin-1 character set) by the “hard space” code in the
ANSI character table:
onkey function
public onkey(keycode)
if (key==’~’)
return 160
return key
// replace ~ by hard space (code 160 in Latin-1)
// leave other keys unaltered
Functions whose name starts with the “@” symbol are also public. So an
alternative way to write the public function onkey function is:
@onkey function
return key==’~’ ? 160 : key
The “@” character, when used, becomes part of the function name; that is,
in the last example, the function is called “@onkey”. The host application
decides on the names of the public functions that a script may implement.
Default values
of function arguments: 74
Arguments of a public function may not have default values. A public function interfaces the host application to the pawn script. Hence, the arguments passed to the public function originate from the host application, and
the host application cannot know what “default values” the script writer
plugged for function arguments —which is why the pawn parser flags the
use of default values for arguments of public functions as an error. The issue
of default values in public function arguments only pops up in the case that
you wish to call public functions from the script itself.
• Static functions
When the function name is prefixed with the keyword static, the scope of
the function is restricted to the file that the function resides in.
The static attribute can be combined with the “stock” attribute.
• Stock functions
A “stock” function is a function that the pawn parser must “plug into”
the program when it is used, and that it may simply “remove” from the
program (without warning) when it is not used. Stock functions allow a
compiler or interpreter to optimize the memory footprint and the file size of
a (compiled) pawn program: any stock function that is not referred to, is
completely skipped —as if it were lacking from the source file.
A typical use of stock functions, hence, is in the creation of a set of “library”
functions. A collection of general purpose functions, all marked as “stock”
may be put in a separate include file, which is then included in any pawn
script. Only the library functions that are actually used get “linked” in.
To declare a stock function, prefix the function name with the keyword
stock. Public functions and native functions cannot be declared “stock”.
When a stock function calls other functions, it is usually a good practice to
declare those other functions as “stock” too —with the exception of native
functions. Similarly, any global variables that are used by a stock function
should in most cases also be defined “stock”. The removal of unused (stock)
functions can cause a chain reaction in which other functions and global
variables are not longer accessed either. Those functions are then removed
as well, thereby continuing the chain reaction until only the functions that
are used, directly or indirectly, remain.
• Native functions
A pawn program can call application-specific functions through a “native
function”. The native function must be declared in the pawn program by
means of a function prototype. The function name must be preceded by the
keyword native.
native getparam(a[], b[], size)
native multiply_matrix(a[], b[], size)
native openfile(const name[])
Public variables
can be declared
Stock variables:
The names “getparam”, “multiply_matrix” and “openfile” are the internal names of the native functions; these are the names by which the
functions are known in the pawn program. Optionally, you may also set an
external name for the native function, which is the name of the function as
the “host application” knows it. To do so, affix an equal sign to the function
prototype followed by the external name. For example:
native getparam(a[], b[], size) = host_getparam
native multiply_matrix(a[], b[], size) = mtx_mul
When a native function returns an array, the dimensions and size of the
array must be explicitly declared. The array specification occurs between
the function name and the parameter list. For example:
#define Rect [ .left, .top, .right, .bottom ]
native intersect[Rect](src1[Rect], src2[Rect])
An example of
a native userdefined operator is
on page 88
Unless specified explicitly, the external name is equal to the internal name
of a native function. One typical use for explicit external names is to set a
symbolic name for a user-defined operator that is implemented as a native
See the “Implementer’s Guide” for implementing native functions in C/C++
(on the “host application” side).
Native functions may not have state specifiers.
• User-defined operators
Tags: 66
The only data type of pawn is a “cell”, typically a 32-bit number or bit
pattern. The meaning of a value in a cell depends on the particular application —it need not always be a signed integer value. pawn allows to attach
a “meaning” to a cell with its “tag” mechanism.
Based on tags, pawn also allows you to redefine operators for cells with a
specific purpose. The example below defines a tag “ones” and an operator to
add two “ones” values together (the example also implements operators for
subtraction and negation). The example was inspired by the checksum algorithm of several protocols in the TCP/IP protocol suite: it simulates one’s
complement arithmetic by adding the carry bit of an arithmetic overflow
back to the least significant bit of the value.
forward ones: operator+(ones: a, ones: b)
forward ones: operator-(ones: a, ones: b)
forward ones: operator-(ones: a)
new ones: chksum = ones: 0xffffffff
print "Input values in hexadecimal, zero to exit\n"
new ones: value
print ">> "
value = ones: getvalue(.base=16)
chksum = chksum + value
printf "Checksum = %x\n", chksum
while (value)
stock ones: operator+(ones: a, ones: b)
const ones: mask = ones: 0xffff
/* word mask */
const ones: shift = ones: 16
/* word shift */
/* add low words and high words separately */
new ones: r1 = (a & mask) + (b & mask)
new ones: r2 = (a >>> shift) + (b >>> shift)
new ones: carry
/* code label (goto target) */
/* add carry of the new low word to the high word, then
* strip it from the low word
carry = (r1 >>> shift)
r2 += carry
r1 &= mask
/* add the carry from the new high word back to the low
* word, then strip it from the high word
carry = (r2 >>> shift)
r1 += carry
r2 &= mask
/* a carry from the high word injected back into the low
* word may cause the new low to overflow, so restart in
* that case
if (carry)
goto restart
return (r2 << shift) | r1
stock ones: operator-(ones: a)
return (a == ones: 0xffffffff) ? a : ~a
stock ones: operator-(ones: a, ones: b)
return a + -b
The notable line in the example is the line “chksum = chksum + value” in
the loop in function main. Since both the variables chksum and value have
the tag ones, the “+” operator refers to the user-defined operator (instead
of the default “+” operator). User-defined operators are merely a notational
convenience. The same effect is achieved by calling functions explicitly.
The definition of an operator is similar to the definition of a function, with
the difference that the name of the operator is composed by the keyword
“operator” and the character of the operator itself. In the above example,
both the unary “-” and the binary “-” operators are redefined. An operator function for a binary operator must have two arguments, one for an
unary operator must have one argument. Note that the binary “-” operator
adds the two values together after inverting the sign of the second operand.
The subtraction operator thereby refers to both the user-defined “negation”
(unary “-”) and addition operators.
Forward declaration: 80
A redefined operator must adhere to the following restrictions:
⋄ A user-defined operator must be declared before use (this is in contrast
to “normal” functions): either put the implementation of the user-defined
operator above the functions that use it, or add a forward declaration near
the top of the file.
⋄ Only the following operators may be redefined: +, -, *, /, %, ++, --, ==,
!=, <, >, <=, >=, ! and =. That is, the sets of arithmetic and relational
operators can be overloaded, but the bitwise operators and the logical
operators cannot. The = and ! operators are a special case.
⋄ You cannot invent new operators; you cannot define operator “#” for example.
⋄ The precedence level and associativity of the operators, as well as their
“arity” remain as defined. You cannot make an unary “+” operator, for
⋄ The return tag of the relational operators and of the “!” operator must
be “bool:”.
⋄ The return tag of the arithmetic operators is at your choosing, but you
cannot redefine an operator that is identical to another operator except
for its return tag. For example, you cannot make both
alpha: operator+(alpha: a, alpha: b)
beta: operator+(alpha: a, alpha: b)
(The assignment operator is an exception to this rule.)
⋄ pawn already defines operators to work on untagged cells, you cannot
redefine the operators with only arguments without tags.
⋄ The arguments of the operator function must be non-arrays passed by
value. You cannot make an operator work on arrays.
In the example given above, both arguments of the binary operators have
the same tag. This is not required; you may, for example, define a binary
“+” operator that adds an integer value to a “ones:” number.
Essentially, the operation of the pawn parser is to look up the tag(s) of the
operand(s) that the operator works on and to look up whether a user-defined
operator exists for the combination of the operator and the tag(s). However,
the parser recognizes special situations and provides the following features:
⋄ The parser recognizes operators like “+=” as a sequence of “+” and “=” and
it will call a user-defined operator “+” if available and/or a user-defined
operator “=”. In the example program, the line “chksum = chksum +
value” might have been abbreviated to “chksum += value”.
⋄ The parser recognizes commutative operators (“+”, “*”, “==”, and “!=”)
and it will swap the operands of a commutative operator if that produces
a fit with a user-defined operator. For example, there is usually no need
to implement both
ones:operator+(ones:a, b)
ones:operator+(a, ones:b)
(implementing both functions is valid, and it is useful in case the userdefined operator should not be commutative).
⋄ Prefix and postfix operators are handled automatically. You only need to
define one user operator for the “++” and “--” operators for a tag.
⋄ The parser calls the “!” operator implicitly in case of a test without
explicit comparison. For example, in the statement “if (var) ...”
when “var” has tag “ones:”, the user-defined operator “!” will be called
for var. The “!” operator thus doubles as a “test for zero” operator. (In
“Call by value”
versus “call by
reference”: 70
Native functions:
one’s complement arithmetic, both the “all-ones” and the “all-zeros” bit
patterns represent zero.)
⋄ The user-defined assignment operator is implicitly called for a function
argument that is passed “by value” when the tag names of the formal and
the actual arguments match the tag names of the left and right hand sides
of the operator. In other words, the pawn parser simulates that “pass
by value” happens through assignment. The user-defined operator is not
called for function arguments that are passed “by reference”.
⋄ If you wish to forbid an operation, you can “forward declare” the operator
without ever defining it (see page 80). This will flag an error when the
user-defined operator is invoked. For example, to forbid the “%” operator
(remainder after division) on floating point values, you can add the line:
forward Float: operator%(Float: a, Float: b)
User-defined operators can optionally be declared “stock” or “native”.
In the case of a native operator function, the definition should include an
external name. For example (when, on the host’s side, the native function
is called float_add):
native operator+ function
native Float: operator+(Float: val, Float: val) = float_add
The user-defined assignment operator is a special case, because it is an operator that has a side effect. Although the operator has the appearance
of a binary operator, its “expression result” is the value at the right hand
—the assignment operator would be a “null”-operator if it weren’t for its
side-effect. In pawn a user-defined assignment operator is declared as:
operator= function
ones: operator=(a)
return ones: ( (a >= 0) ? a : ~(-a) )
The user-defined “=” operator looks like a unary operator in this definition,
but it is a special case nevertheless. In contrast to the other operators, the
tag of the return value for the user-defined operator is important: the pawn
parser uses the tags of the argument and the return value to find a matching
user-defined operator.
The example function above is a typical application for a user-defined assignment operator: to automatically coerce/convert an untagged value to
a tagged value, and to optionally change the memory representation of the
value in the process. Specifically, the statement “new ones:A = -5” causes
the user-defined operator to run, and for the constant -5 the operator will
return “~(- -5)”, or ~5, or −6.∗
• Floating point and fixed point arithmetic
pawn only has intrinsic support for integer arithmetic (the Z-domain, or
“whole numbers”, both positive and negative). Support for floating point
arithmetic or fixed point arithmetic must be implemented through (native)
functions. User operators, then, allow a more natural notation of expressions
with fixed or floating point numbers.
The pawn parser has support for literal values with a fractional part, which
it calls “rational numbers”. Support for rational literals must be enabled
explicitly with a #pragma. The #pragma indicates how the rational numbers
must be stored —floating point or fixed point. For fixed point rational values,
the #pragma also specifies the precision in decimals. Two examples for the
#pragma are:
#pragma rational Float
/* floating point format */
#pragma rational Fixed(3)
/* fixed point, with 3 decimals */
Since a fixed point value must still fit in a cell, the number of decimals has
a direct influence of the range of a fixed point value. For a fixed point value
with 3 decimals, the range would be −2, 147, 482 . . . + 2, 147, 482.
The format for a rational number may only be specified once for the entire
pawn program. In an implementation one typically chooses either floating
point support or fixed point support. As stated above, for the actual implementation of the floating point or fixed point arithmetic, pawn requires the
help of (native) functions and user-defined operators. A good place to put
the #pragma for rational number support would be in the include file that
also defines the functions and operators.
The include file
for fixed point arithmetic contains definitions like:
Modern CPUs use two’s complement integer arithmetic. For positive values, the bitwise
representation of a value is the same in one’s complement and two’s complement, but the
representations differ for negative values. For instance, the same bit pattern that means
-5 in one’s complement stands for -6 in two’s complement.
See the application note “Fixed Point Support Library” for where to obtain the include
Rational literals:
#pragma rational:
native Fixed: operator*(Fixed: val1, Fixed: val2) = fmul
native Fixed: operator/(Fixed: val1, Fixed: val2) = fdiv
The user-defined operators for multiplication and division of two fixed point
numbers are aliased directly to the native functions fmul and fdiv. The
host application must, then, provide these native functions.
Another native user-defined operator is convenient to transform an integer to
fixed point automatically, if it is assigned to a variable tagged as “Fixed:”:
native Fixed: operator=(oper) = fixed
User-defined operators: 84
With this definition, you can say “new Fixed:
fract = 3” and the
value will be transformed to 3.000 when it is stored in variable fract. As
explained in the section on user-defined operators, the assignment operator
also runs for function arguments that are passed by value. In the expression
“new Fixed: root = sqroot(16)” (see the implementation of function
sqroot on page 77), the user-defined assignment operator is called on the
argument 16.
For adding two fixed point values together, the default “+” operator is sufficient, and the same goes for subtraction. Adding a normal (integer) number
to a fixed point number is different: the normal value must be scaled before
adding it. Hence, the include file implements operators for that purpose too:
additive operators, commutative and non-commutative
stock Fixed: operator+(Fixed: val1, val2)
return val1 + fixed(val2)
stock Fixed: operator-(Fixed: val1, val2)
return val1 - fixed(val2)
stock Fixed: operator-(val1, Fixed: val2)
return fixed(val1) - val2
The “+” operator is commutative, so one implementation handles both cases.
For the “-” operator, both cases must be implemented separately.
Finally, the include file forbids the use of the remainder operator (“%”) on
fixed point values: the remainder is only applicable to integer divisions:
forbidden operators on fixed point values
forward Fixed: operator%(Fixed: val1, Fixed: val2)
forward Fixed: operator%(Fixed: val1, val2)
forward Fixed: operator%(val1, Fixed: val2)
Because of the presence of the (forward) declaration of the operator, the
pawn parser will attempt to use the user-defined operator rather than the
default “%” operator. By not implementing the operator, the parser will
subsequently issue an error message.
The preprocessor
The first phase of compiling a pawn source file to the executable P-code is
“preprocessing”: a general purpose text filter that modifies/cleans up the
text before it is fed into the parser. The preprocessing phase removes comments, strips out “conditionally compiled” blocks, processes the compiler
directives and performs find-&-replace operations on the text of the source
file. The compiler directives are summarized on page 116 and the text substitution (“find-&-replace”) is the topic of this chapter.
The preprocessor is a process that is invoked on all source lines immediately
after they are read. No syntax checking is performed during the text substitutions. While the preprocessor allows powerful tricks in the pawn language,
it is also easy to shoot yourself in the foot with it.
In this chapter, I will refer to the C/C++ language on several occasions
because pawn’s preprocessor is similar to the one in C/C++. That said, the
pawn preprocessor is incompatible with the C/C++ preprocessor.
The #define directive defines the preprocessor macros. Simple macros are:
#define maxsprites
#define CopyRightString
"(c) Copyright 2004 by me"
In the pawn script, you can then use them as you would use constants. For
#define maxsprites 25
#define CopyRightString "(c) Copyright 2004 by me"
print( Copyright )
new sprites[maxsprites]
By the way, for these simple macros there are equivalent pawn constructs:
const maxsprites = 25
stock const CopyRightString[] = "(c) Copyright 2004 by me"
These constant declarations have the advantage of better error checking and
the ability to create tagged constants. The syntax for a string constant is
an array variable that is declared both “const” and “stock”. The const
attribute prohibits any change to the string and the stock attribute makes
the declaration “disappear” if it is never referred to.
Substitution macros can take up to 10 parameters. A typical use for macros
with parameters is to simulate tiny functions:
The preprocessor
Listing: the “min” macro
#define min(%1,%2)
((%1) < (%2) ? (%1) : (%2))
If you know C/C++ , you will recognize the habit of enclosing each argument
and the whole substitution expression in parentheses.
If you use the above macro in a script in the following way:
Listing: bad usage of the “min” macro
new a = 1, b = 4
new min = min(++a,b)
the preprocessor translates it to:
new a = 1, b = 4
new min = ((++a) < (b) ? (++a) : (b))
which causes “a” to possibly be incremented twice. This is one of the traps
that you can trip into when using substitution macros (this particular problem is well known to C/C++ programmers). Therefore, it may be a good
idea to use a naming convention to distinguish macros from functions. In
C/C++ it is common practice to write preprocessor macros in all upper case.
To show why enclosing macro arguments in parentheses is a good idea, consider the macro:
#define ceil_div(%1,%2) (%1 + %2 - 1) / %2
This macro divides the first argument by the second argument, but rounding
upwards to the nearest integer (the divide operator, “/”, rounds downwards).
If you use it as follows:
new a = 5
new b = ceil_div(8, a - 2)
the second line expands to “new b = (8 + a - 2 - 1) / a - 2”, which,
considering the precedence levels of the pawn operators, leads to “b” being
set to zero (if “a” is 5). What you would have expected from looking at the
macro invocation is eight divided by three (“a - 2”), rounded upwards —
hence, that “b” would be set to the value 3. Changing the macro to enclose
each parameter in parentheses solves the problem. For similar reasons, it is
also advised to enclose the complete replacement text in parentheses. Below
is the ceil_div macro modified accordingly:
#define ceil_div(%1,%2) ( ((%1) + (%2) - 1) / (%2) )
The pattern matching is subtler than matching strings that look like function
calls. The pattern matches text literally, but accepts arbitrary text where
the pattern specifies one or more parameter(s). You can create patterns like:
Operator precedence: 109
The preprocessor
macro that translates a syntax for variable assignment to a function call
#define Field.%1=%2;
When the expansion of a macro contains text that matches other macros,
the expansion is performed at invocation time, not at definition time. Thus
the code:
#define a(%1)
#define b(%1)
new c = a(8)
will evaluate to “new c = (1+(2*(8)))”, even though the macro “b” was
not defined at the time of the definition of “a”.
If an argument in the replacement text has a “#” immediately in front of
it, the argument is converted to a packed string constant —meaning that
double quotes are tagged at the beginning and the end. For example, if you
have the definition:
#define log(%1) "ERR: " ... #%1 ... "\n"
See page 100 for
literal strings with
the “...” operator
Directives: 116
then the expression log(test) will result in "ERR: " . . . "test" . . . "\n". The
“#” operator is also called the “stringize” operator, as it converts arguments
to (packed) strings.
In the preceding examples, the pattern and the substitution text fit on a
single line, as is the case for all directives. For macros where this is inconvenient, an alternative syntax is to enclose the substitution text between
square brackets. The substitution text must still start at the same line as
the pattern, but it may be split over multiple lines.
The pattern matching is constrained to the following rules:
⋄ There may be no square brackets and no space characters in the pattern.
If you must match a space, you need to use the “\32;” escape sequence.
The substitution text, on the other hand, may contain space characters
and brackets. Due to the matching rules of the macro pattern (explained
below), matching a space character is rarely needed.
⋄ Escape sequences may appear in the pattern, and can be used, for example,
to match a literal “%” character.
⋄ The pattern may not end with a parameter; a pattern like “set:%1=%2”
is illegal. If you wish to match with the end of a statement, you can add a
semicolon at the end of the pattern. If semicolons are optional at the end
of each statement, the semicolon will also match a newline in the source.
The preprocessor
⋄ The pattern must start with a letter, an underscore, or an “@” character
The first part of the pattern that consists of alphanumeric characters (plus
the “_” and/“@”) is the “name” or the “prefix” of the macro. On the
defined operator and the #undef directive, you specify the macro prefix.
⋄ When matching a pattern, the preprocessor ignores white space between
non-alphanumeric symbols and white space between an alphanumeric symbol and a non-alphanumeric one, with one exception: between two identical symbols, white space is not ignored. Therefore:
the pattern abc(+-) matches “abc ( + - )”
the pattern abc(--) matches “abc ( -- )” but does not match
“abc(- -)”
⋄ There are up to 10 parameters, denoted with a “%” and a single digit (1
to 9 and 0). The order of the parameters in a pattern is not important.
⋄ The #define symbol is a parser directive. As with all parser directives,
the pattern definition must fit on a single line. You can circumvent this
with a “\” on the end of the line. The text to match must also fit on a
single line.
Note that in the presence of (parametrized) macros, lines of source code
may not be what they appear: what looks like an array access may be
“preprocessed” to a function call, and vice versa.
A host application that embeds the pawn parser may provide an option
to let you check the result of text substitution through macros. If you are
using the standard pawn toolset, you will find instructions of how to use
the compiler and run-time in the companion booklet “The pawn booklet —
Implementer’s Guide”.
Directives: 116
General syntax
Identifiers, numbers and tokens are separated by spaces, tabs, carriage returns and “form feeds”. Series of one or more of these separators are called white space.
Optional semicolons: 120
Optional semicolons
Semicolons (to end a statement) are optional if they occur at the end
of a line. Semicolons are required to separate multiple statements
on a single line. An expression may still wrap over multiple lines,
but postfix operators (++ and --) must appear on the same line as
their operand.
Text between the tokens /* and */ (both tokens may be at the
same line or at different lines) and text behind // (up to the end of
the line) is a programming comment. The parser treats a comment
as white space. Comments may not be nested.
A comment that starts with “/** ” (two stars and white-space
behind the second star) and ends with “*/” is a documentation
comment. A comment that starts with “/// ” (three slashes and
white-space behind the third slash) is also a documentation comment. The parser may treat documentation comments in a special
way; for example, it may construct on-line help from it.
Names of variables, functions and constants. Identifiers consist of
the characters a. . . z, A. . . Z, 0. . . 9, _ or @; the first character may
not be a digit. The characters @ and _ by themselves are not valid
identifiers, i.e. “_Up” is a valid identifier, but “_” is not.
pawn is case sensitive.
A parser may truncate an identifier after a maximum length. The
number of significant characters is implementation defined, but it is
at least 16 characters.
Reserved words (keywords)
Statements Operators
General syntax
Next to reserved words, pawn also has several predefined constants,
you cannot use the symbol names of the predefined constants for
variable or function names.
Constants (literals)
Integer numeric constants
0b followed by a series of the digits 0 and 1
a series of digits between 0 and 9
0x followed by a series of digits between 0 and 9 and
the letters a to f
In all number radices, the single quote may be used as a
“digit group separator”. For decimal numbers, the quote
separates 3 digits (thousands separator), for binary numbers,
the quote separates 8 digits and for hexadecimal numbers,
a quote separates 4 (hexadecimal) digits. Using the single
quote as a digit group separator is optional, but if used, the
number of digits after the quote must conform to the above
Predefined constants: 101
Rational numbers
are also called
“real numbers”
or “floating point
#pragma rational:
General syntax
Rational number constants
A rational number is a number with a fractional part. A
rational number starts with one or more digits, contains a
decimal point and has at least one digit following the decimal
point. For example, “12.0” and “0.75” are valid rational
numbers. Optionally, an exponent may be appended to the
rational number; the exponent notation is the letter “e”
(lower case) followed by a signed integer numeric constant.
For example, “3.12e4” is a valid rational number with an
Support for rational numbers must be enabled with #pragma
rational directive. Depending on the options set with this
directive, the rational number represents a floating point or
a fixed point number.
The single quote may be used as a “thousands separator”
(separating 3 digits) in the “whole part” of the number.
Character constants
A single ASCII character surrounded by single quotes is
a character constant (for example: ’a’, ’7’, ’$’). The
single quote that starts a character constant may either be a
normal (forward) single quote or a reverse single quote. The
terminating quote must always be a forward single quote.
Character constants are assumed to be numeric constants.
Escape sequences
Audible alarm (beep)
Form feed
Carriage Return
Horizontal tab
Vertical tab
the escape character
single quote
double quote
percent sign
General syntax
character code with decimal code “ddd ”
character code with hexadecimal code “hhh”
The semicolon after the \ddd; and \xhhh; codes is optional.
Its purpose is to give the escape sequence sequence an explicit termination symbol when it is used in a string constant.
The backslash (“\”) is the default “escape” character. If
desired, you can set a different escape character with the
#pragma ctrlchar directive (page 119).
String constants
String constants are assumed to be arrays with a size that
is sufficient to hold all characters plus a terminating ’\0’.
Each string is stored at a unique position in memory; there
is no elimination of duplicate strings.
An unpacked string is a series of zero or more ASCII characters surrounded by doubled single quotes. Each array element contains a single character. Both the forward and the
reversed single quotes are acceptable to start a string. An
unpacked string can hold characters in a multi-byte character set, such as Unicode or UCS-4.
unpacked string constant:
‘‘the quick brown fox...’’
A packed string literal series of zero or more ASCII characters surrounded by double quotes.
packed string constant:
"...packed the lazy dog in a bag"
In the case of a packed string, the parser packs as many
characters in a cell as will fit. A character is not addressable
as a single unit, instead each element of the array contains
multiple characters. The first character in a “pack” occupies
the highest bits of the array element. In environments that
store memory words with the high byte at the lower address
(Big Endian, or Motorola format), the individual characters
are stored in the memory cells in the same order as they are
in the string. A packed string ends with a zero character
General syntax
and the string is padded (with zero bytes) to a multiple of
A packed string can only hold characters from a single-byte
character set, such as ascii or one of the extended ascii sets
from the ISO 8859 norm.
Escape sequences may be used within strings. See the section on character constants (page 98) for a list of escape
There is an alternative syntax for “plain strings”. In a plain
string, every character is taken as-is and escape sequences
are not recognized. Plain strings are convenient to store
file/resource names, especially in the case where the escape
character is also used as a special character by the operating
system or host application.
The syntax for a plain string is the escape character followed
by the string in double quotes. The backslash (“\”) is the default “escape” character. You cannot enter escape sequences
in a plain string: all characters will be taken literally.
plain (unpacked) string constant:
\’’C:\all my work\novel.rtf’’
In the above example, the occurrences of “\a” and “\n” do
not indicate escape sequences, but rather the literal character pairs “\” and “a”, and “\” and “n”.
Plain packed strings exist as well:
\"C:\all my work\novel.rtf"
Two string literals may be concatenated by inserting with an
ellipsis operator (three dots, or “...”) between the strings.
For example:
"The quick" ... "brown fox"
This syntax works with packed, unpacked and raw strings.
Different kinds of string literals should not be combined in
literal string concatenation, though. String concatenation
is valid for string literals only —string variables must be
concatenated with a library function or run-time code.
General syntax
Array constants
A series of numeric constants between braces is an array
constant. Array constants can be used to initialize array
variables with (see page 63) and they can be passed as function arguments (see page 70).
Symbolic constants
A source file declares symbolic constants with the const instruction. A single const keyword may declare a list of constants with
sequentially incremented values and sharing the same tag name.
const identifier = constant expression
Creates a single symbolic constant with the value of the
constant expression on the right hand of the assignment operator. The constant can be used at any place where a literal
number is valid (for example: in expressions, in array declarations and in directives like “#if” and “#assert”).
const tagname { constant list }
A list of symbolic names, grouped by braces, may follow
the const keyword. The constant list is a series of identifiers separated by commas. The first indentifier must be
explicitly assigned a (numeric) value. Unless overruled, every subsequent constant has the value of its predecessor plus
Examples: 7, 19
Identifiers: 96
Example: 24
The optional tagname token that follows the const keyword
is used as the default tag name for every symbol in the constant list. The symbols in the constant list may have an
explicit tag, which overrules the default tag name.
A symbolic constant that is defined locally, is valid throughout the
block. A local symbolic constant may not have the same name as a
variable (local or global), a function, or another constant (local or
Predefined constants
cellbits The size of a cell in bits; usually 32.
cellmax The largest valid positive value that a cell can hold; usually 2147483647.
See page 67 for
examples of enumerated “const”
General syntax
The largest valid negative value that a cell can hold;
usually -2147483648.
charbits The size of a packed character in bits; usually 8.
charmax The largest valid packed character value; a packed character is usually 8-bit and the maximum valid value is
thus 255.
charmin The smallest valid character value, for both packed and
unpacked values; currently set to zero (0).
The debug level: 2 if the parser creates full symbolic information plus run-time bounds checking, 1 if the parser
generates run-time checking only (assertions and array
bounds checks), and 0 (zero) if all debug support and
run-time checking was turned off.
0 (this constant is tagged as bool:).
The current line number in the source file.
The version number of the pawn compiler in Binary
Coded Decimals (BCD) —that is, for version 2.8.1 the
constant is “0x281”.
1 (this constant is tagged as bool:).
ucharmax The largest unpacked character value, its value depends
on the size of a cell. A typical use for this constant is
in checking whether a string is packed or unpacked, see
page 136.
Identifiers: 96
Tag names
A tag consists of an identifier followed by a colon. There may be
no white space between the identifier and the colon.
Predefined tag names
For “true/false” flags. The predefined constants true
and false have this tag.
Rational numbers have this tag when fixed point support
is enabled (page 120).
Rational numbers have this tag when floating point support is enabled (page 120).
Operators and expressions
• Notational conventions
The operation of some operators depends on the specific kinds of operands.
Therefore, operands are notated thus:
any expression;
any expression to which a value can be assigned (“lvalue” expressions);
an array;
a function;
a symbol —which is a variable, a constant or a function.
• Expressions
An expression consists of one or more operands with an operator. The
operand can be a variable, a constant or another expression. An expression
followed by a semicolon is a statement.
Listing: examples of expressions
f(a1, a2)
v = (ia1 * ia2) / ia3
• Arithmetic
e1 + e2
Results in the addition of e1 and e2.
e1 - e2
Results in the subtraction of e1 and e2.
Results in the arithmetic negation of a (two’s complement).
e1 * e2
Results in the multiplication of e1 and e2.
Operators and expressions
e1 / e2
Results in the division of e1 by e2. The result is truncated
to the nearest integral value that is less than or equal to the
quotient. Both negative and positive values are rounded down,
i.e. towards −∞.
e1 % e2
Results in the remainder of the division of e1 by e2. The sign
of the remainder follows the sign of e2. Integer division and
remainder have the Euclidean property: D = q*d + r, where
q = D/d and r = D%d.
increments v by 1; the result if the expression is the value of
v before it is incremented.
increments v by 1; the result if the expression is the value of
v after it is incremented.
v-decrements v by 1; the result if the expression is the value of
v before it is decremented.
decrements v by 1; the result if the expression is the value of
v after it is decremented.
The unary + is not defined in pawn.
The operators ++ and -- modify the operand. The operand
must be an lvalue.
• Bit manipulation
results in the one’s complement of e.
e1 >> e2
results in the arithmetic shift to the right of e1 by e2 bits.
The shift operation is signed: the leftmost bit of e1 is copied
to vacant bits in the result.
Operators and expressions
e1 >>> e2
results in the logical shift to the right of e1 by e2 bits. The
shift operation is unsigned: the vacant bits of the result are
filled with zeros.
e1 << e2
results in the value of e1 shifted to the left by e2 bits; the
rightmost bits are set to zero. There is no distinction between
an arithmetic and a logical left shift
e1 & e2
results in the bitwise logical “and” of e1 and e2.
e1 | e2
results in the bitwise logical “or” of e1 and e2.
e1 ^ e2
results in the bitwise “exclusive or” of e1 and e2.
• Assignment
The result of an assignment expression is the value of the left operand after
the assignment. The left operand may not be tagged.
v = e
assigns the value of e to variable v.
v = a
assigns array a to variable v; v must be an array with the
same size and dimensions as a; a may be a string or a literal
the following operators combine an assignment with an arithmetic or a bitwise operation; the result of the expression is
the value of the left operand after the arithmetic or bitwise
v += e
increments v with e.
v -= e
decrements v with e
Tag names: 66
Operators and expressions
v *= e
multiplies v with e
v /= e
divides v by e.
v %= e
assigns the remainder of the division of v by e to v.
v >>= e
shifts v arithmetically to the right by e bits.
v >>>= e
shifts v logically to the right by e bits.
v <<= e
shifts v to the left by e bits.
v &= e
applies a bitwise “and” to v and e and assigns the result to v.
v |= e
applies a bitwise “or” to v and e and assigns the result to v.
v ^= e
applies a bitwise “exclusive or” to v and e and assigns the
result to v.
• Relational
A logical “false” is represented by an integer value of 0; a logical “true” is
represented by any value other than 0. Value results of relational expressions
are either 0 or 1, and their tag is set to “bool:”.
e1 == e2
results in a logical “true” if e1 is equal to e2.
e1 != e2
results in a logical “true” if e1 differs from e2.
the following operators may be “chained”, as in the expression
“e1 <= e2 <= e3”, with the semantics that the result is “1”
if all individual comparisons hold and “0” otherwise.
e1 < e2
results in a logical “true” if e1 is smaller than e2.
Operators and expressions
e1 <= e2
results in a logical “true” if e1 is smaller than or equal to e2.
e1 > e2
results in a logical “true” if e1 is greater than e2.
e1 >= e2
results in a logical “true” if e1 is greater than or equal to e2.
• Boolean
A logical “false” is represented by an integer value of 0; a logical “true” is
represented by any value other than 0. Value results of Boolean expressions
are either 0 or 1, and their tag is set to “bool”.
results to a logical “true” if e was logically “false”.
e1 || e2
results to a logical “true” if either e1 or e2 (or both) are
logically “true”. The expression e2 is only evaluated if e1 is
logically “false”.
e1 && e2
results to a logical “true” if both e1 and e2 are logically
“true”. The expression e2 is only evaluated if e1 is logically
• Miscellaneous
array index: results to cell e from array a.
array index: results to character e from “packed” array a.
Operators and expressions
results to the value returned by the function f. The function
is called with the arguments e1, e2, . . . eN. The order of evaluation of the arguments is undefined (an implementation may
choose to evaluate function arguments in reversed order).
e1 ? e2 : e3
results in either e2 or e3, depending on the value of e1. The
conditional expression is a compound expression with a two
part operator, “?” and “:”. Expression e2 is evaluated if e1
is logically “true”, e3 is evaluated if e1 is logically “false”.
tagname: e
tag override; the value of the expression e does not change,
but its tag changes. See page 66 for more information.
e1, e2
results in e2, e1 is evaluated before e2. If used in function
argument lists or a conditional expression, the comma expression must be surrounded by parentheses.
defined defined s
results in the value 1 if the symbol is defined. The symbol
may be a constant (page 97), or a global or local variable.
The tag of this expression is bool:.
Example: 76
sizeof sizeof s
results in the size in “elements” of the specified variable. For
simple variables and for arrays with a single dimension, an
element is a cell. For multi-dimensional arrays, the result is
the number of array elements in that dimension —append []
to the array name to indicate a lower/more minor dimension.
If the size of a variable is unknown, the result is zero.
When used in a default value for a function argument, the expression is evaluation at the point of the function call, instead
of in the function definition.
Operators and expressions
state s
See also page 114
where s is the name of a state that is optionally prefixed with for state specifiers
the automaton name, this operator results in the value 1 if
the automatons is in the indicated state and in 0 otherwise.
The tag of this expression is bool:.
tagof s
results in the a unique number that represents the tag of the
variable, the constant, the function result or the tag label.
When used in a default value for a function argument, the expression is evaluation at the point of the function call, instead
of in the function definition.
• Operator precedence
The table beneath groups operators with equal precedence, starting with the
operator group with the highest precedence at the top of the table.
If the expression evaluation order is not explicitly established by parentheses,
it is determined by the association rules. For example: a*b/c is equivalent
with (a*b)/c because of the left-to-right association, and a=b=c is equivalent
with a=(b=c).
Operators and expressions
function call
array index (cell)
array index (character)
logical not
one’s complement
two’s complement (unary minus)
tag override
symbol definition status
symbol size in “elements”
automaton/state condition
unique number for the tag
arithmetic shift right
logical shift right
shift left
bitwise and
bitwise exclusive or
bitwise or
smaller than
smaller than or equal to
greater than
greater than or equal to
logical and
logical or
? :
*= /= %= += -= >>= >>>= <<= &= ^= |=
A statement may take one or more lines, whereas one line may contain two
or more statements.
Control flow statements (if, if–else, for, while, do–while and switch)
may be nested.
Statement label
A label consists of an identifier followed by a colon (“:”). A label
is a “jump target” of the goto statement.
Identifiers: 96
Each statement may be preceded by a label. There must be a statement after the label; an empty statement is allowed.
The scope of a label is the function in which it is declared (a goto
statement cannot therefore jump out off the current function to another function).
Compound statement
A compound statement is a series of zero or more statements surrounded by braces ({ and }). The final brace (}) should not be
followed by a semicolon. Any statement may be replaced by a compound statement. A compound statement is also called a block. A
compound statement with zero statements is a special case, and it
is called an “empty statement”.
Expression statement
Any expression becomes a statement when a semicolon (“;”) is appended to it. An expression also becomes a statement when only
white space follows it on the line and the expression cannot be extended over the next line.
Empty statement
An empty statement performs no operation and consists of a compound block with zero statements; that is, it consists of the tokens
“{ }”. Empty statements are used in control flow statements if there
is no action (e.g. while (!iskey()) {}) or when defining a label
just before the closing brace of a compound statement. An empty
statement does not end with a semicolon.
assert expression
Aborts the program with a run-time error if the expression evaluates
to logically “false”.
Example: 9
Example: 19
Terminates and exits the smallest enclosing do, for or while statement from any point within the loop other than the logical end.
The break statement moves program control to the next statement
outside the loop.
Terminates the current iteration of the smallest enclosing do, for or
while statement and moves program control to the condition part of
the loop. If the looping statement is a for statement, control moves
to the third expression in the for statement (and thereafter to the
second expression).
Example: 26
do statement while ( expression )
Executes a statement before the condition part (the while clause) is
evaluated. The statement is repeated while the condition is logically
“true”. The statement is at least executed once.
exit expression
Abort the program. The expression is optional, but it must start
on the same line as the exit statement if it is present. The exit
instruction returns the expression value (plus the expression tag) to
the host application, or zero if no exit expression is present. The
significance and purpose of exit codes is implementation defined.
Examples: 7, 9, 19
Variable declarations: 60
for ( expression 1 ; expression 2 ; expression 3 ) statement
All three expressions are optional.
expression 1 Evaluated only once, and before entering the loop.
This expression may be used to initialize a variable.
This expression may also hold a variable declaration,
using the new syntax. A variable declared in this expression exists only in the for loop.
You cannot combine and expression (using already existing variables) and a declaration of new variables in
this field —either all variables in this field must already
exist, or they must all be created in this field.
expression 2 Evaluated before each iteration of the loop and ends
the loop if the expression results to logically “false”.
If omitted, the result of expression 2 is assumed to be
logically “true”.
expression 3 Evaluated after each execution of the statement. Program control moves from expression 3 to expression 2
for the next (conditional) iteration of the loop.
The statement for( ; ; ) is equivalent with while (true).
goto label
Moves program control (unconditionally) to the statement that follows the specified label. The label must be within the same function
as the goto statement (a goto statement cannot jump out of a function).
if ( expression ) statement 1 else statement 2
Executes statement 1 if the expression results to logically “true”.
The else clause of the if statement is optional. If the expression
results to logically “false” and an else clause exists, the statement
associated with the else clause (statement 2) executes.
Example: 5
When if statements are nested and else clauses are present, a given
else is associated with the closest preceding if statement in the
same block.
return expression
Terminates the current function and moves program control to the
statement following the calling statement. The value of the expression is returned as the function result. The expression may be an
array variable or a literal array.
The expression is optional, but it must start on the same line as the
return statement if it is present. If absent, the value of the function
is zero.
sleep expression
Abort the program, but leave it in a re-startable state. The expression is optional. If included, the sleep instruction returns the
expression value (plus the expression tag) to the host application.
The significance and purpose of exit codes/tags is implementation
defined; typically, an application uses the sleep instruction to allow
for light-weight multi-tasking of several concurrent pawn programs,
or to implement “latent” functions.
Examples: 9, 19
state ( expression ) automaton :name
Changes the current state in the specified automaton. The expression between parentheses is optional; if it is absent, the parentheses
must be omitted as well. The name of the automaton is optional as
well, when changing the state of the default, anonymous, automaton;
if the automaton name is absent, the colon (“:”) must be omitted
as well.
Below are two examples of unconditional state changes. The first is
for the default automaton:
state handshake
and the second for a specific automaton:
state gps:handshake
Often, whether or not a state changes depends on parameters of the
event or the condition of the automaton as a whole. Since conditional state changes are so frequent, the condition may be in the
state instruction itself.∗ The condition follows the keyword state,
between parentheses. The state will only change if the condition is
logically “true”.
See page 38 for
examples of conditional state
The state instruction causes an implied call to the exit function
of the current state and to the entry function for the new state —if
such exit and entry functions exist.
“entry” functions:
42, “exit” functions: 45
switch ( expression ) { case list }
Transfers control to different statements within the switch body
depending on the value of the switch expression. The body of the
switch statement is a compound statement, which contains a series
of “case clauses”.
Each “case clause” starts with the keyword case followed by a constant list and one statement. The constant list is a series of expressions, separated by comma’s, that each evaluates to a constant
value. The constant list ends with a colon. To specify a “range”
in the constant list, separate the lower and upper bounds of the
range with a double period (“..”). An example of a range is: “case
The alternative is to fold unconditional state changes in the common if–else construct.
The switch statement moves control to a “case clause” if the value
of one of the expressions in the constant list is equal to the switch
expression result.
The “default clause” consists of the keyword default and a colon.
The default clause is optional, but if it is included, it must be the last
clause in the switch body. The switch statement moves control to
the “default clause” if none of the case clauses match the expression
switch (weekday(12,31,1999))
case 0, 1:
/* 0 == Saturday, 1 == Sunday */
case 2:
case 3:
case 4:
case 5:
case 6:
print("invalid week day")
while ( expression ) statement
Evaluates the expression and executes the statement if the expression result yields logically “true”. After the statement has executed,
program control returns to the expression again. The statement is
thus executed while the expression is true.
Examples: 5, 19,
Line continuation:
See also “Predefined constants”
on page 101
All directives must appear first on a line (they may be preceded by white
space, but not by any other characters). All directives start with the character # and the complete instruction may not span more than one line —with
an exception for #define.
#assert constant expression
Issues a compile time error if the supplied constant expression evaluates to zero. The #assert directive is most useful to guard against
implementation defined constructs on which a program may depend,
such as the cell size in bits, or the number of packed characters per
#define pattern replacement
Defines a text substitution macro. The pattern is matched to all
lines read from the source files; the sections that match are replaced
by the replacement texts. The pattern and the replacement texts
may contain parameters, denoted by “%0” to “%9”. See page 92 for
details and examples on text substitution.
Closes the current file and thereby ignores all the text below the
#endinput directive.
message: Signals a “user error” with the specified message. User errors are fatal errors and they serve a similar purpose as the #assert
#file name
Adjusts the name for the current file. This directive is used implicitly by the text preprocessor; there is usually no need to set a
filename explicitly.
#if constant expression, #elseif, #else, #endif
Portions of a program may be parsed or be ignored depending on
certain conditions. The pawn parser (compiler or interpreter) generates code only for those portions for which the condition is true.
The directive #if must be followed by a constant expression. To
check whether a variable or constant is defined, use the defined
Zero or more #elseif directives may follow the initial #if directive.
These blocks are skipped if any of the preceding #if or #elseif
blocks were parsed (i.e. not skipped). As with the #if directive, a
constant expression must follow the #elseif expression.
The #else causes the parser to skip all lines up to #endif if the preceding #if or any of the preceding #elseif directives were “true”,
and the parses these lines if all preceding blocks were skipped. The
#else directive may be omitted; if present, there may be only be
one #else associated with each #if.
The #endif directive terminates a program portion that is parsed
conditionally. Conditional directives can be nested and each #if
directive must be ended by an #endif directive.
#include filename or <filename>
Inserts the contents of the specified file at the current position within
the current file. A filename between angle brackets (“<” and “>”)
refers to a system file; the pawn parser (compiler or interpreter) will
search for such files only in a pre-set list of directories and not in the
“current” directory. Filenames that are unquoted or that appear in
double quotes are normal include files, for which a pawn parser will
look in the current directory first.
The pawn parser first attempts to open the file with the specified
name. If that fails, it tries appending the extensions “.inc”, “.p”
and “.pawn” to the filename (in that order). The proposed default
extension of include files is “.inc”.
When the file can be opened successfully, the #include directive
defines a constant with the name “_inc_” plus the base name of
the file (the filename without path and extension) and the value 1.
If the constant already exists, the #include directive skips opening
and including the file, thus preventing a double inclusion. To force
a double include, remove the constant definition with the #undef
directive before the second inclusion of the file.
#line number
The current line number (in the current file). This directive is used
implicitly by the text preprocessor; there is usually no need to set
the line number explicitly.
#pragma extra information
A “pragma” is a hook for a parser to specify additional settings,
such as warning levels or extra capabilities. Common #pragmas are:
#pragma align
Aligns the next declaration to the offset set with the alignment compiler option. Some (native) functions may perform
better with parameters that are passed by reference when
these are on boundaries of 8, 16, or even 32 bytes. Alignment requirements are dependent of the host applications.
Putting the #pragma align line in front of a declaration
of a global or a static variable aligns this variable to the
boundary set with the compiler option. This #pragma aligns
only the variable that immediately follows the #pragma. The
alignment of subsequent variables depends on the size and
alignment of the variables that precede it. For example,
if a global array variable of 2 cells is aligned on a 16-byte
boundary and a cell is 4 bytes, the next global variable is
located 8 bytes further.
Putting the #pragma align line in front of a declaration of
a function will align the stack frame of that function to the
boundary specified earlier, with the result that the first local, non-“static”, variable is aligned to that boundary. The
alignment of subsequent variables depends on the size and
alignment of the variables that precede it. In practice, to
align a local non-static variable, you must align the function’s stack frame and declare that variable before any other
#pragma amxlimit value
Sets the maximum size, in bytes, that the compiled script
may grow to. This pragma is useful for (embedded) environments where the maximum size of a script is bound to a
hard upper limit.
If there is no setting for the amount of RAM for the data
and stack (see the pragma amxram), this refers to the total
memory requirements; if the amount of RAM is explicitly
set, this value only gives the amount of memory needed for
the code and the static data.
#pragma amxram value
Sets the maximum memory requirements, in bytes, for data
and stack that a compiled script may have. This value is
is useful for (embedded) environments where the maximum
data size of a script is bound to a hard upper limit. Especially in the case where the pawn script runs from ROM,
the sizes for the code and data sections need both to be set.
#pragma codepage name/value
The pawn parser can translate characters in character constants and in unpacked strings to Unicode/UCS-4 “wide”
characters. This #pragma indicates the codepage that must
be used for the translation. See the section Internationalization on page 138 for details and required extra resources for
the codepage translation.
#pragma ctrlchar character
Defines the character to use to indicate the start of a “escape
sequence”. By default, the control character is “\”.
For example
#pragma ctrlchar ’$’
You may give the new value for the control character as a
character constant (between single quotes) or as a decimal
or hexadecimal value. When you omit the value of the new
control character, the parser reverts to the default control
#pragma deprecated value
The subsequent symbol is flagged as “deprecated”. If a
script uses it, the parser issues a warning.
#pragma dynamic value
Sets the size, in cells, of the memory block for dynamic data
(the stack and the heap) to the value specified by the expression. The default size of the dynamic data block is implementation defined. An implementation may also choose to
grow the block on an as-needed basis (see the host program’s
documentation, or the “Implementer’s Guide” for details).
#pragma library name
Sets the name of the (dynamically linked) extension mod-
Escape character:
ule that contains required native functions. This #pragma
should appear above native function declarations that are
part of the extension module.
The library name parameter may be absent, in which case
any subsequent native function declarations are not associated with any extension module.
The scope of this #pragma runs from the line at which it
appears until the end of the file in which it appears. In
typical usage, a #pragma library line will appear at the
top of an include file that declares native functions for an
extension module, and the scope of the library “link” ends
at the end of that include file.
#pragma overlay value
The pawn parser can generate P-code that runs as dynamically loaded overlays. Currently, an overlay is the size of a
function. Whether the parser generates P-code for overlaid
execution or for standard execution depends on the parser
configuration (and, perhaps, user settings). This #pragma
allows the script writer to override the default. The parameter of this #pragma is the size of the overlay pool in bytes,
or zero to turn overlay support off.
Rational number
support: 98
#pragma rational tagname(value)
Enables support for rational numbers. The tagname is the
name of the tag that rational numbers will have; typically
one chooses the names “Float:” or “Fixed:”. The presence
of value in parentheses behind tagname is optional: if it is
omitted, a rational number is stored as a “floating point”
value according to the IEEE 754 norm; if it is present, a rational number is a fixed precision number (“scaled integer”)
with the specified number of decimals.
#pragma semicolon value
If value is zero, no semicolon is required to end a statement
if that statement is last on a line. Semicolons are still needed
to separate multiple statements on the same line.
When semicolons are optional (the default), a postfix operator (one of “++” and “--”) may not be the first token on a
line, as they will be interpreted as prefix operators.
#pragma tabsize value
The number of characters between two consecutive tab
positions. The default value is 8. You may need to set
the tab size to avoid warning 217 (loose indentation) if the
source code is indented alternately with spaces and with tab
characters. Alternatively, by setting the “tabsize” #pragma
to zero, the parser will no longer issue warning 217.
#pragma unused symbol,. . .
Marks the named symbol as “used”. Normally, the pawn
parser warns about unused variables and unused local constants. In most situations, these variables and constants are
redundant, and it is better to remove them for the sake of
code clarity. Especially in the case of local constants, it
may, however, be better (or required) to keep the constant
definitions. This #pragma then permits to mark the symbol
(variable or constant) as “used”, and avoid a parser warning.
The #pragma must appear after the symbol declaration —
but it need not appear immediately after the declaration.
Multiple symbol names may appear in a single #pragma; the
symbols must be separated by commas.
#section name
Starts a new section for the generated code. Any variables and
functions that are declared “static” are only visible to the section
to which they belong. By default, each source file is a separate
section and there is only one section per file.
With the #section directive, you can create multiple sections in a
source file. The name of a section is optional, if it is not set, a unique
identifier for the source file is used for the name of the section.
Any declared section ends automatically at the end of the file.
#tryinclude filename or <filename>
This directive behaves similarly as the #include directive, but it
does not give an error when the file to include does not exist —i.e.,
try to include but fail silently on error.
Warning messages: 162
#undef name
Removes a text substitution macro or a numeric constant declared
with const. The “name” parameter must be the macro “prefix”
—the alphanumeric part of the macro. See page 92 for details and
examples on text substitution.
Proposed function library
Since pawn is targeted as an application extension language, most of the
functions that are accessible to pawn programs will be specific to the host
application. Nevertheless, a small set of functions may prove useful to many
• Core functions
The “core” module consists of a set of functions that support the language
itself. Several of the functions are needed to pull arguments out of a variable
argument list (see page 78).
Force a value inside a range
clamp(value, min=cellmin, max=cellmax)
The value to force in a range.
The low bound of the range.
The high bound of the range.
value if it is in the range min – max; min if value is lower than
min; and max if value is higher than max.
See also:
max, min
Return a public function index
funcidx(const name[])
The index of the named public function. If no public function
with the given name exists, funcidx returns −1.
A host application runs a public function from the script by
passing the public function’s index to amx_Exec. With this
function, the script can query the index of a public function,
and thereby return the “next function to call” to the application.
amx Exec: see the
Get an argument
getarg(arg, index=0)
The argument sequence number, use 0 for first
The index, in case arg refers to an array.
The value of the argument.
This function retrieves an argument from a variable argument
list. When the argument is an array, the index parameter
specifies the index into the array. The return value is the
retrieved argument.
See also:
numargs, setarg
Return free heap space
The free space on the heap. The stack and the heap occupy
a shared memory area, so this value indicates the number of
bytes that is left for either the stack or the heap.
In absence of recursion, the pawn parser can also give an estimate of the required stack/heap space.
Return the highest of two numbers
max(value1, value2)
The two values for which to find the highest number.
The higher value of value1 and value2.
See also:
clamp, min
Return the lowest of two numbers
min(value1, value2)
The two values for which to find the lowest number.
The lower value of value1 and value2.
See also:
clamp, max
Return the number of arguments
The number of arguments passed to a function; numargs is
useful inside functions with a variable argument list.
See also:
getarg, setarg
Return a pseudo-random number
The limit for the random number.
A pseudo-random number in the range 0 – max-1.
The standard random number generator of pawn is likely a
linear congruential pseudo-random number generator with a
range and a period of 231 . Linear congruential pseudo-random
number generators suffer from serial correlation (especially in
the low bits) and it is unsuitable for applications that require
high-quality random numbers.
Set an argument
setarg(arg, index=0, value)
The argument sequence number, use 0 for first
The index, in case arg refers to an array.
The value to set the argument to.
true on success and false if the argument or the index are
This function sets the value of an argument from a variable
argument list. When the argument is an array, the index
parameter specifies the index into the array.
See also:
getarg, numargs
Swap bytes in a cell
The value for which to swap the bytes.
A value where the bytes in parameter “c” are swapped (the
lowest byte becomes the highest byte).
Convert a character to lower case
The character to convert to lower case.
The upper case variant of the input character, if one exists, or
the unchanged character code of “c” if the letter “c” has no
lower case equivalent.
Support for accented characters is platform-dependent.
See also:
Convert a character to upper case
The character to convert to upper case.
The lower case variant of the input character, if one exists, or
the unchanged character code of “c” if the letter “c” has no
upper case equivalent.
Support for accented characters is platform-dependent.
See also:
• Console functions
For testing purposes, the console functions that read user input and that output strings in a scrollable window or on a standard terminal display are often
convenient. Not all terminal types and implementations may implement all
functions —especially the functions that clear the screen, set foreground
and background colours and control the cursor position, require an extended
terminal control.
Clear rest of the line
This function always returns 0.
Clears the line at which the cursor is, from the position of the
cursor to the right margin of the console. This function does
not move the cursor.
See also:
Clear screen
This function always returns 0.
Clears the display and moves the cursor to the upper left corner.
See also:
Read one character
If true, the character read from the keyboard is
echoed on the display.
The numeric code for the character that is read (this is usually
the ASCII code).
See also:
Read a line
getstring(string[], size=sizeof string, bool:pack=false)
The line read from the keyboard is stored in this
The size of the string parameter in cells.
If true the function stores the line as a packed
The number of characters read, excluding the terminating null
Function getstring stops reading when either the enter key
is typed, or the maximum length is reached. The maximum
length is in cells (not characters) and it includes a terminating null character. The function can read both packed and
unpacked strings; when reading a packed string, the function
may read more characters than the size parameter specifies,
because each cell holds multiple characters.
See also:
Read a number
getvalue(base=10, end=‘\r’, ...)
Must be between 2 and 36, use 10 for decimal or
16 for hexadecimal.
The character code that terminates the input.
More than one character may be listed.
If true the function stores the line as a packed
The value that is read.
Read a value (a signed number) from the keyboard. The getvalue function allows you to read in a numeric radix from 2
to 36 (the base parameter) with decimal radix by default.
By default the input ends when the user types the enter key,
but one or more different keys may be selected (the end parameter and subsequent). In the list of terminating keys, a
positive number (like ’\r’) displays the key and terminates
input, and a negative number terminates input without displaying the terminating key.
See also:
Set cursor location
gotoxy(x=1, y=1)
The new cursor position.
true if the cursor moved and false if the requested position
is invalid.
Sets the cursor position on the console. The upper left corner
is at (1,1).
See also:
Display text on the display
print(const string[], foreground=-1, background=-1)
The string to display.
background Colour codes for the foreground and background
of the text string; see function setattr for a lost
of colours. When left at -1, the default colours
are used. Note that a terminal or a host application may not support colours.
This function always returns 0.
See also:
printf, setattr
Display formatted text on the display
printf(const format[], ...)
The string to display, including any (optional)
formatting codes.
This function always returns 0.
Prints a string with embedded codes:
print a number at this position in binary radix
print a character at this position
print a number at this position in decimal radix
print a floating point number at this position (assuming
floating point support is present)
print a fixed point number at this position (assuming fixed
point support is present)
print either a floating point number or a fixed point number at this position, depending on what is available; if
both floating point and fixed point support is present, %r
is equivalent to %f (i.e. printing a floating point number)
print a character string at this position
print a number at this position in hexadecimal radix
The printf function works similarly to the printf function
of the C language.
See also:
Set text colours
setattr(foreground=-1, background=-1)
background The colour codes for the new foreground and
background colours for text. When either of the
two parameters is negative (or absent), the respective colour setting will not be changed.
This function always returns 0.
On most systems, the colour value must be a value between
zero and seven, as per the ANSI Escape sequences, ISO 6429.
Predefined constants for the colours are black (0), red (1),
green (2), yellow (3), blue (4), magenta (5), cyan (6) and
white (7).
See also:
• Date/time functions
Functions to get and set the current date and time, as well as a millisecond
resolution “event” timer are described in an application note entitled “Time
Functions Library” that is available separately.
• File input/output
Functions for handling text and binary files, with direct support for UTF8 text files, is described in an application note entitled “File I/O Support
Library” that is available separately.
• Fixed point arithmetic
The fixed-point decimal arithmetic module for pawn is described in an application note entitled “Fixed Point Support Library” that is available separately.
• Floating point arithmetic
The floating-point arithmetic module for pawn is described in an application
note entitled “Floating Point Support Library” that is available separately.
• Process and library call interface
Functions to launch and control external applications and functions to use
general purpose DLLs or shared libraries is described in an application note
entitled “Process control and Foreign Function Interface” that is available
• String manipulation
A general set of string manipulation functions, operating on both packed
and unpacked strings, is described in an application note entitled “String
Manipulation Library” that is available separately.
Pitfalls: differences from C
⋄ pawn lacks the typing mechanism of C. pawn is an “integer-only” variety
of C; there are no structures or unions, and floating point support must be
implemented with user-defined operators and the help of native functions.
⋄ The accepted syntax for rational numbers is stricter than that of floating
point values in C. Values like “.5” and “6.” are acceptable in C, but
in pawn one must write “0.5” and “6.0” respectively. In C, the decimal
period is optional if an exponent is included, so one can write “2E8”; pawn
does not accept the upper case “E” (use a lower case “e”) and it requires
the decimal point: e.g. “2.0e8”. See page 98 for more information.
⋄ pawn does not provide “pointers”. For the purpose of passing function
arguments by reference, pawn provides a “reference” argument, (page 70).
The “placeholder” argument replaces some uses of the NULL pointer (page
⋄ Numbers can have hexadecimal, decimal or binary radix. Octal radix is
not supported. See “Constants” on page 97. Hexadecimal numbers must
start with “0x” (a lower case “x”), the prefix “0X” is invalid.
⋄ Escape sequences (“\n”, “\t”, etc.) are the same, except for “\ddd”
where “ddd” represent three decimal digits, instead of the octal digits
that C/C++ uses. The backslash (“\”) may be replaced with another
symbol; see #pragma ctrlchar on page 119 —notably, previous versions
of pawn used the caret (“^”) as the escape character.
⋄ Cases in a switch statement are not “fall through”. Only a single instruction may (and must) follow each case label. To execute multiple
instructions, you must use a compound statement. The default clause
of a switch statement must be the last clause of the switch statement.
More on page 114. In C/C++, switch is a “conditional goto”, akin to
Fortran’s calculated labels. In pawn, switch is a structured “if”.
⋄ A break statement breaks out of loops only. In C/C++ , the break statement also ends a case in a switch statement. Switch statements are
implemented differently in pawn (see page 114).
⋄ pawn supports “array assignment”, with the restriction that both arrays
must have the same size. For example, if “a” and “b” are both arrays
with 6 cells, the expression “a = b” is valid. Next to literal strings, pawn
Pitfalls: differences from C
also supports literal arrays, allowing the expression “a = {0,1,2,3,4,5}”
(where “a” is an array variable with 6 elements).
⋄ defined is an operator, not a preprocessor directive. The defined operator in pawn operates on constants (declared with const), global variables,
local variables and functions.
⋄ The sizeof operator returns the size of a variable in “elements”, not in
“bytes”. An element may be a cell or a sub-array. See page 108 for details.
⋄ The empty instruction is an empty compound block, not a semicolon (page
111). This modification avoids a frequent error.
⋄ The compiler directives differ from C’s preprocessor commands. Notably,
the #define directive is incompatible with that of C/C++, and #ifdef and
#ifndef are replaced by the more general #if directive (see “Directives”
on page 116). To create numeric constants, see also page 101; to create
string constants, see also page 92.
⋄ Text substitutions (preprocessor macros; see the #define directive) are
not matched across lines. That is, the text that you want to match and
replace with a #define macro must appear on a single line.
⋄ A division is carried out in such a way that the remainder after division
has (or would have) the same sign as the denominator. For positive denominators, this means that the direction for truncation for the operator
“/” is always towards the smaller value, where -2 is smaller than -1, and
that the “%” operator always gives a positive result —regardless of the
sign of the numerator. See page 103.
⋄ There is no unary “+” operator, which is a “no-operation” operator anyway.
⋄ Three of the bitwise operators have different precedence than in C. The
precedence levels of the “&”, “^” and | operators is higher than the relational operators (Dennis Ritchie explained that these operators got their
low precedence levels in C because early C compilers did not yet have the
logical “&&” and || operators, so the bitwise “&” and | were used instead).
⋄ The “extern” keyword does not exist in pawn; the current implementation
of the compiler has no “linking phase”. To create a program from several
source files, add all source files the compilers command line, or create one
main project script file that “#include’s” all other source files. The pawn
Pitfalls: differences from C
compiler can optimize out functions and global variables that you do not
use. See pages 61 and 83 for details.
⋄ The keyword const in pawn implements the enum functionality from C,
see page 101.
⋄ In most situations, forward declarations of functions (i.e., prototypes) are
not necessary. pawn is a two-pass compiler, it will see all functions on the
first pass and use them in the second pass. User-defined operators must
be declared before use, however.
If provided, forward declarations must match exactly with the function
definition, parameter names may not be omitted from the prototype or
differ from the function definition. pawn cares about parameter names
in prototypes because of the “named parameters” feature. One uses prototypes to call forwardly declared functions. When doing so with named
parameters, the compiler must already know the names of the parameters (and their position in the parameter list). As a result, the parameter
names in a prototype must be equal to the ones in the definition.
Assorted tips
• Working with characters and strings
Strings can be in packed or in unpacked format. In the packed format, each
cell will typically hold four characters (in common implementations, a cell
is 32-bit and a character is 8 bit). In this configuration, the first character
in a “pack” of four is the highest byte of a cell and the fourth character is
in the lowest byte of each cell.
A string must be stored in an array. For an unpacked string, the array
must be large enough to hold all characters in the string plus a terminating
zero cell. That is, in the example below, the variable ustring is defined as
having five cells, which is just enough to contain the string with which it is
unpacked string
new ustring[5] = ’’test’’
In a packed string, each cell contains several characters and the string ends
with a zero character. The example below will allocate enough cells to hold
five packed characters. In a typical implementation, there will be two cells
in the array.
packed string
new pstring{5} = "test"
In other words, the array is declared to be able to hold at least the specified
number of packed characters.
See the separate
application note
for proposed native functions that
operate on both
packed and unpacked strings
You can design routines that work on strings in both packed and unpacked
formats. To find out whether a string is packed or unpacked, look at the
first cell of a string. If its value is either negative or higher than the maximum possible value of an unpacked character, the string is a packed string.
Otherwise it is an unpacked string.
The code snippet below returns true if the input string is packed and false
ispacked function
bool: ispacked(string[])
return !(0 <= string[0] <= ucharmax)
Assorted tips
An unpacked string ends with a full zero cell. The end of a packed string is
marked with only a zero character. Since there may be up to four characters
in a 32-bit cell, this zero character may occur at any of the four positions in
the “pack”. The { } operator extracts a character from a cell in an array.
Basically, one uses the cell index operator (“[ ]”) for unpacked strings and
the character index operator (“{ }”) to work on packed strings.
For example, a routine that returns the length in characters of any string
(packed or unpacked) is:
my strlen function
new len = 0
if (ispacked(string))
while (string{len} != EOS)
while (string[len] != EOS)
return len
EOS: predefined
constant standing
for End Of String;
it has the value
/* get character from pack */
/* get cell */
If you make functions to work exclusively on either packed or unpacked
strings, it is a good idea to add an assertion to enforce this condition:
strupper function
assert ispacked(string)
for (new i=0; string{i} != EOS; ++i)
string{i} = toupper(string{i})
Although, in preceding paragraphs we have assumed that a cell is 32 bits
wide and a character is 8 bits, this should not be relied upon. The size of
a cell is implementation defined; the maximum and minimum values are in
the predefined constants cellmax and cellmin. There are similar predefined
constants for characters. One may safely assume, however, that both the size
of a character in bytes and the size of a cell in bytes are powers of two.
The predefined charbits and cellbits constants allow you to determine
how many packed characters fit in a cell. For example:
const CharsPerCell = cellbits / charbits
Predefined constants: 101
Assorted tips
• Internationalization
Programming examples in this manual have used the English language for
all output (prompts, messages, . . . ), and a Latin character set. This is not
necessarily so; one can, for example, modify the first “hello world” program
on page 3 to:
“hello world” in Greek
printf "Geiȧsou
pawn has basic support for non-Latin alphabets, but it only accepts nonLatin characters in strings and character constants. The pawn language
requires that all keywords and symbols (names of functions, variables, tags
and other elements) be encoded in the ascii character set.
For languages whose required character set is relatively small, a common
solution is to use an 8-bit extended ascii character set (the ascii character
set is 7-bit, holding 128 characters). The upper 128 codes of the extended set
contain glyphs specific for the language. For Western European languages,
a well known character set is “Latin-1”, which is standardized as ISO 8859-1
—the same set also goes by the name “codepage 1252”, at least for Microsoft
Windows.∗ Codepages have been defined for many languages; for example,
ISO 8859-2 (“Latin-2”) has glyphs used in Central and Eastern Europe, and
ISO 8859-7 contains the Greek alphabet in the upper half of the extended
ascii set.
Unfortunately, codepage selection can by confusing, as vendors of operating
systems typically created their own codepages irrespective of what already
existed. As a result, for most character sets there exist multiple incompatible
codepages. For example, codepage 1253 for Microsoft Windows also encodes
the Greek alphabet, but it is incompatible with ISO 8859-7. When writing
texts in Greek, it now becomes important to check what encoding is used,
because many Microsoft Windows applications support both.
When the character set for a language exceeds 256 glyphs, a codepage does
not suffice. Traditionally, the codepage technique was extended by reserving
special “shift” codes in the base character set that switch to a new set of
glyphs. The next character then indicates the specific glyph. In effect, the
Codepage 1252 is not exactly the same as Latin-1; Microsoft extended the standardized
set to include glyphs at code positions that Latin-1 marks as “reserved”.
Assorted tips
glyph is now identified by a 2-byte index. On the other hand, some characters
(especially the 7-bit ascii set) can still be indicated by a single byte. The
“Shift-JIS” standard, for the Japanese character set, is an example for the
variable length encoding.
Codepages become problematic when interchanging documents or data with
people in regions that use a different codepage, or when using different languages in the same document. Codepages that use “shift” characters complicate the matter further, because text processing must now take into account
that a character may take either one or two bytes. Scanning through a string
from right to left may even become impossible, as a byte may either indicate
a glyph from the base set (“unshifted”) or it may be a glyph from a shifted
set —in the latter case the preceding byte indicates the shift set, but the
meaning of the preceding character depends on the character before that.
The ISO/IEC 10646 “Universal Character Set” (UCS) standard has the ambitious goal to eventually include all characters used in all the written languages in the world, using a 31-bit character set. This solves both of the
problems related to codepages and “shifted” character sets. However, the
ISO/IEC body could not produce a standard in time, and therefore a consortium of mainly American software manufacturers started working in parallel
on a simplified 16-bit character set called “Unicode”. The rationale behind
Unicode was that it would encode abstract characters, not glyphs, and that
therefore 65,536 would be sufficient.† In practice, though, Unicode does encode glyphs and not long after it appeared, it became apparent that 65,536
code points would not be enough. To counter this, later Unicode versions
were extended with multiple “planes” and special codes that select a plane.
The combination of a plane selector and the code pointer inside that plane
is called a “surrogate pair”. The first 65,536 code points are in the “Basic
Multilingual Plane” (BMP) and characters in this set do not need a plane
Essentially, the introduction of surrogate pairs in the Unicode standard is
equivalent to the shift codes of earlier character sets —and it carries some
of the problems that Unicode was intended to solve. The UCS-4 encoding
by ISO/IEC 10646 does not have/need surrogate pairs.
Support for Unicode/UCS-4 in (host) applications and operating systems has
If Unicode encodes characters, an “Unicode font” is a contradictio in terminis —because
a font encodes glyphs.
Assorted tips
emerged in two different ways: either the internal representation of characters is multi-byte (typically 16-bit, or 2-byte), or the application stores
strings internally in UTF-8 format, and these strings are converted to the
proper glyphs only when displaying or printing them. Recent versions of
Microsoft Windows use Unicode internally; The Plan-9 operating system pioneered the UTF-8 encoding approach, which is now widely used in unix/
Linux. The advantage of UTF-8 encoding as an internal representation is
that it is physically an 8-bit encoding, and therefore compatible with nearly
all existing databases, file formats and libraries. This circumvents the need
for double entry-points for functions that take string parameters —as is the
case in Microsoft Windows, where many functions exist in an “A”nsi and
a “W”ide version. A disadvantage of UTF-8 is that it is a variable length
encoding, and many in-memory string operations are therefore clumsy (and
inefficient). That said, with the appearance of surrogate pairs, Unicode has
now also become a variable length encoding.
The pawn language requires that its keywords and symbols names are in
ascii, and it allows non-ascii characters in strings. There are five ways
that a host application could support non-ascii characters in strings and
character literals:
1 Support codepages: in this strategy the entire complexity of choosing
the correct glyphs and fonts is delegated to the host application. The
codepage support is based on codepage mapping files with a file format
of the “cross mapping tables” distributed by the Unicode consortium.
2 Support Unicode or UCS-4 and let the pawn compiler convert scripts that
were written using a codepage to “wide” characters: for this strategy, you
need to set #pragma codepage or use the equivalent compiler option. The
compiler will only correctly translate characters in unpacked strings.
3 Support Unicode or UCS-4 and let the pawn compiler convert scripts
encoded in UTF-8 to “wide” characters: when the source file for the
pawn compiler is in UTF-8 encoding, the compiler expands characters
to Unicode/UCS-4 in unpacked strings.
4 Support UTF-8 encoding internally (in the host application) and write
the source file in UTF-8 too: all strings should now be packed strings to
avoid the compiler to convert them.
For most internationalization strategies, as you can see, the host application
needs to support Unicode or UCS-4. As a side note, the pawn compiler
Assorted tips
does not generate Unicode surrogate pairs. If characters outside the BMP
are needed and the host application (or operating system) does not support
the full UCS-4 encoding, the host application must split the 32-bit character
cell provided by the pawn compiler into a surrogate pair.
The pawn compiler accepts a source file as an UTF-8 encoded text file —see
page 169. When the source file is in UTF-8 encoding, “wide” characters in
an unpacked string are stored as multi-byte Unicode/UCS-4 characters; wide
characters in a packed string remain in UTF-8 encoding. To write source
files in UTF-8 encoding, you need, of course, a (programmer’s) editor that
supports UTF-8. Codepage translation does not apply for files that are in
UTF-8 encoding.
For an occasional Unicode character in a literal string, an alternative is that
you use an escape sequence. As Unicode character tables are usually documented with hexadecimal glyph indices, the \xhhh; sequence is probably the
more convenient specification of a random Unicode character. For example,
the escape sequence “\x2209” stands for the “6∈” character.
Packed & unpacked strings:
Escape sequence:
There is a lot more to internationalization than just basic support for extended character sets, such as formatting date & time fields, reading order
(left-to-right or right-to-left) and locale-driven translation of system messages. The pawn toolkit delegates these issues to the host application.
• Working with tags
The tag name system was invented to add a “usage checking” mechanism
to pawn. A tag denotes a “purpose” of a value or variable, and the pawn
compiler issues a diagnostic message when the tag of an expression does not
match the required tag for the context of the expression.
Many modern computer languages offer variable types, where a type specifies
the memory layout and the purpose of the variable. The programming language then checks the type equivalence; the pascal language is very strict
at checking type equality, whereas the C programming language is more forgiving. The pawn language does not have types: all variables have the size
and the layout of a cell, although bit representations in the cell may depend
on the purpose of the variable. In summary:
⋄ a type specifies the memory layout and the range of variables and function
⋄ a tagname labels the purpose of variables, constants and function results
Tag names: 66
User-defined operators: 84
Assorted tips
Tags in pawn are mostly optional. A program that was “fortified” with
tag names on the variable and constant declarations will function identically
when all tag names are removed. One exception is formed by user-defined
operators: the pawn compiler uses the tags of the operands to choose between any user-defined operators and the standard operator.
The snippet below declares three variables and does three assignments, two
of which give a “tag mismatch” diagnostic message:
More tag name
rules: 66
comparing apples to oranges
new apple:elstar
new orange:valencia
new x
/* variable "elstar" with tag "apple" */
/* variable "valencia" with tag "orange" */
/* untagged variable "x" */
elstar = valencia
elstar = x
x = valencia
/* tag mismatch */
/* tag mismatch */
/* ok */
The first assignment causes a “tag mismatch” diagnostic as it assigns an
“orange” tagged variable to a variable with an “apple” tag. The second
assignment puts the untagged value of x into a tagged variable, which causes
again a diagnostic. When the untagged variable is on the left hand of the
assignment operator, as in the third assignment, there is no warning or error
message. As variable x is untagged, it can accept a value of any weak tag.
The same mechanism applies to passing variables or expressions to functions
as function operands —see page 77 for an example. In short, when a function
expects a particular tag name on an argument, you must pass an expression/
variable with a matching tag to that function; but if the function expects an
untagged argument, you may pass in arguments with any weak tag.
On occasion, it is necessary to temporarily change the tag of an expression.
For example, with the declarations of the previous code snippet, if you would
wish to compare apples with oranges (recent research indicates that comparing apples to oranges is not as absurd than popular belief holds), you could
if (apple:valencia < elstar)
valencia = orange:elstar
lvalue (definition
of ~): 103
The test expression of the if statement (between parentheses) compares
the variable valencia to the variable elstar. To avoid a “tag mismatch”
diagnostic, it puts a tag override apple: on valencia —after that, the expressions on the left and the right hands of the > operator have the same tag
name: “apple:”. The second line, the assignment of elstar to valencia,
Assorted tips
overrides the tag name of elstar or orange: before the assignment. In an
assignment, you cannot override the tag name of the destination; i.e., the left
hand of the = operator. It is an error to write “apple:valencia = elstar”.
In the assignment, valencia is an “lvalue” and you cannot override the tag
name of an lvalue.
As shown earlier, when the left hand of an assignment holds an untagged
variable, the expression on the right hand may have any weak tag name.
When used as an lvalue, an untagged variable is compatible with all weak
tag names. Or rather, a weak tag is silently dropped when it is assigned
to an untagged variable or when it is passed to a function that expects an
untagged argument. When a tag name indicates the bit pattern of a cell,
silently dropping a weak tag can hide errors. For example, the snippet below
has an error that is not immediately obvious:
Listing: bad way of using tags
#pragma rational float
new limit = -5.0
new value = -1.0
if (value < limit)
printf("Value %f below limit %f\n", value, limit)
printf("Value above limit\n")
Through the “#pragma rational”, all rational numbers receive the “float”
tag name and these numbers are encoded in the 4-byte IEEE 754 format.
The snippet declares two variables, limit and value, both of which are
untagged (this is the error). Although the literal values -5.0 and -1.0
are implicitly tagged with float:, this weak tag is silently dropped when
the values get assigned to the untagged symbols limit and value. Now,
the if statement compares value to limit as integers, using the built-in
standard < operator (a user-defined operator would be more appropriate to
compare two IEEE 754 encoded values). When run, this code snippet tells
us that “Value -1.000000 below limit -5.000000” —which is incorrect,
of course.
To avoid such subtle errors to go undetected, one should use strong tags. A
strong tag is merely a tag name that starts with an upper case letter, such as
Float: instead of float:. A strong tag is never automatically “dropped”,
but it may still be explicitly overridden. Below is a modified code snippet
with the proposed adaptations:
Assorted tips
strong tags are safer
#pragma rational Float
new Float:limit = -5.0
new Float:value = -1.0
if (value < limit)
printf("Value %f below limit %f\n", _:value, _:limit)
printf("Value above limit\n")
Forgetting the Float: tag name in the declaration of the variables limit
or value immediately gives a “tag mismatch” diagnostic, because the literal
values -5.0 and -1.0 now have a strong tag name.
printf is a general purpose function that can print strings and values in
various formats. To be general purpose, printf accepts arguments with any
weak tag name, be it apple:’s, orange:’s, or something else. The printf
function does this by accepting untagged arguments —weak tags are dropped
when an untagged argument is expected. Strong tags, however, are never
dropped, and in the above snippet (which uses the original definition of
printf), I needed to put an empty tag override, “_:”, before the variables
value and limit in the first printf call.
There is an alternative to untagging expressions with strong tag names in
general purpose functions: adjust the definition of the function to accept
both all weak tags and a selective set of strong tag names. The pawn
language supports multiple tag names for every function arguments. The
original definition of printf (from the file is:
native printf(const format[], ...);
By adding both a Float: tag and an empty tag in front of the ellipsis
(“...”), printf will accept arguments with the Float: tag name, arguments without a tag name and arguments that have a weak tag name. To
specify plural tag names, enclose all tag names without their final colon
between braces with a comma separating the tag names (see the example
below). It is necessary to add the empty tag specification to the list of
tag names, because printf would otherwise only accept arguments with a
Float: tag name. Below is the new definition of the function printf:
native printf(const format[], {Float, _}: ...);
Assorted tips
Plural tags allow you to write a single function that accepts cells with a
precisely specified subset of tags (strong and/or weak). While a function
argument may accept being passed actual arguments with diverse tags, a
variable can only have a single tag —and a formal function argument is a
local variable in the body of the function. In the presence of plural tags, the
formal function argument takes on the tag that is listed first.
On occasion, you may want to check which tag an actual function argument
had, when the argument accepts plural tags. Checking the tag of the formal
argument (in the body of the function) is of no avail, because it will always
have the first tag in the tag list in the declaration of the function argument.
You can check the tag of the actual argument by adding an extra argument
to the function, and set its default value to be the “tagof” of the argument
in question. Similar to the sizeof operator, the tagof operator has a special
meaning when it is applied in a default value of a function argument: the
expression is evaluated at the point of the function call, instead of at the
function definition. This means that the “default value” of the function
argument is the actual tag of the parameter passed to the function.
Directives: 75
Inside the body of the function, you can compare the tag to known tags by,
again, using the tagof operator.
• Concatenating lines
pawn is a free format language, but the parser directives must be on a single
line. Strings may not run over several lines either. When this is inconvenient,
you can use a backslash character (“\”) at the end of a line to “glue” that
line with the next line.
For example:
#define max_path
max_drivename + max_directorystring + \
max_filename + max_extension
You also use the concatenation character to cut long literal strings over
multiple lines. Note that the “\” eats up all trailing white space that comes
after it and leading white space on the next line. The example below prints
“Hello world” with one space between the two words (because there is a
space between ”Hello” and the backslash):
print("Hello \
Directives: 116
Assorted tips
An alternative way to concatenate literal strings is to separate strings, that
are each enclosed in pairs of double quotes, with an ellipsis. The next example is equivalent to the previous one:
print("Hello " ...
• A program that generates its own source code
An odd, slightly academic, criterion to quantify the “expressiveness” of a
programming language is size of the smallest program that, upon execution,
regenerates its own source code. The rationale behind this criterion is that
the shorter the self-generating program, the more flexible and expressive
the language must be. Programs of this kind have been created for many
programming languages —sometimes surprisingly small, as for languages
that have a built-in reflective capabilities.
Self-generating programs are called “quines”, in honour of the philosopher
Willard Van Orman Quine who wrote self-creating phrases in natural language. The work of Van Orman Quine became well known through the books
“Gödel, Escher, Bach” and “Metamagical Themas” by Douglas Hofstadter.
The pawn quine is in the example below; it is modelled after the famous
“C” quine (of which many variations exist). At 77 characters, it is amongst
the smallest versions for the class of imperative programming languages, and
the size can be reduced to 73 characters by removing four “space” characters
that were left in for readability.
new s[]="new s[]=%c%s%c; main() printf s,34,s,34"; main() printf s,34,s,34
appendix a
Error and warning messages
When the compiler finds an error in a file, it outputs a message giving, in
this order:
⋄ the name of the file
⋄ the line number were the compiler detected the error between parentheses,
directly behind the filename
⋄ the error class (“error”, “fatal error” or “warning”)
⋄ an error number
⋄ a descriptive error message
For example:
demo.p(3) : error 001: expected token: ";", but found "{"
Note: the line number given by the compiler may specify a position behind
the actual error, since the compiler cannot always establish an error before
having analyzed the complete expression.
After termination, the return code of the compiler is:
no errors —there may be warnings, though
errors found
aborted by user
These return codes may be checked within batch processors (such as the
“make” utility).
• Error categories
Errors are separated into three classes:
Describe situations where the compiler is unable to generate
appropriate code. Errors messages are numbered from 1 to 99.
Fatal errors Fatal errors describe errors from which the compiler cannot recover. Parsing is aborted. Fatal error messages are numbered
from 100 to 199.
Warnings are displayed for unintended compiler assumptions
and common mistakes. Warning messages are numbered from
200 to 299.
Error and warning messages
• Errors
Pitfalls: 133
Compound statement: 111
Compound statement: 111
expected token: token, but found token
A required token is omitted.
only a single statement (or expression) can follow each
Every case in a switch statement can hold exactly one statement.
To put multiple statements in a case, enclose these statements
between braces (which creates a compound statement).
declaration of a local variable must appear in a compound
The declaration of a local variable must appear between braces
(“{. . . }”) at the active scope level.
When the parser flags this error, a variable declaration appears
as the only statement of a function or the only statement below
an if, else, for, while or do statement. Note that, since local
variables are accessible only from (or below) the scope that their
declaration appears in, having a variable declaration as the only
statement at any scope is useless.
Forward declaration: 80
function name is not implemented
There is no implementation for the designated function. The
function may have been “forwardly” declared —or prototyped—
but the full function definition including a statement, or statement
block, is missing.
function may not have arguments
The function main is the program entry point. It may not have
must be assigned to an array
String literals or arrays must be assigned to an array. This error
message may also indicate a missing index (or indices) at the array
on the right side of the “=” sign.
operator cannot be redefined
Only a select set of operators may be redefined, this operator is
not one of them. See page 84 for details.
Error and warning messages
must be a constant expression; assumed zero
The size of arrays and the parameters of most directives must be
constant values.
invalid array size (negative, zero or out of bounds)
The number of elements of an array must always be 1 or more. In
addition, an array that big that it does exceeds the range of a cell
is invalid too.
illegal function or declaration
The compiler expects a declaration of a global variable or of a
function at the current location, but it cannot interpret it as such.
invalid outside functions
The instruction or statement is invalid at a global level. Local
labels and (compound) statements are only valid if used within
invalid function call, not a valid address
The symbol is not a function.
no entry point (no public functions)
The file does not contain a main function or any public function.
The compiled file thereby does not have a starting point for the
invalid statement; not in switch
The statements case and default are only valid inside a switch
“default” must be the last clause in switch statement
pawn requires the default clause to be the last clause in a switch
multiple defaults in “switch”
Each switch statement may only have one default clause.
undefined symbol symbol
The symbol (variable, constant or function) is not declared.
initialization data exceeds declared size
An array with an explicit size is initialized, but the number of
initiallers exceeds the number of elements specified. For example,
in “arr[3]={1,2,3,4};” the array is specified to have three
elements, but there are four initiallers.
Initialization: 63
Symbol name syntax: 96
Escape sequence:
Error and warning messages
not a label: name
A goto statement branches to a symbol that is not a label.
invalid symbol name
A symbol may start with a letter, an underscore or an “at” sign
(“@”) and may be followed by a series of letters, digits, underscore
characters and “@” characters.
symbol already defined: identifier
The symbol was already defined at the current level.
must be lvalue (non-constant)
The symbol that is altered (incremented, decremented, assigned
a value, etc.) must be a variable that can be modified (this kind
of variable is called an lvalue). Functions, string literals, arrays
and constants are no lvalues. Variables declared with the “const”
attribute are no lvalues either.
array assignment must be simple assignment
When assigning one array to another, you cannot combine an
arithmetic operation with the assignment (e.g., you cannot use the
“+=” operator).
“break” or “continue” is out of context
The statements break and continue are only valid inside the
context of a loop (a do, for or while statement). Unlike the
languages C/C++ and Java, break does not jump out of a switch
function heading differs from prototype
The number of arguments given at a previous declaration of the
function does not match the number of arguments given at the
current declaration.
no matching “#if...”
The directive #else or #endif was encountered, but no matching
#if directive was found.
invalid character constant
One likely cause for this error is the occurrence of an unknown
escape sequence, like “\x”. Putting multiple characters between
single quotes, as in ’abc’ also issues this error message. A third
cause for this error is a situation where a character constant was
expected, but none (or a non-character expression) were provided.
Error and warning messages
invalid subscript (not an array or too many subscripts):
The subscript operators “[” and “]” are only valid with arrays.
The number of square bracket pairs may not exceed the number of
dimensions of the array.
invalid expression, assumed zero
The compiler could not interpret the expression.
compound statement not closed at the end of file (started
at line number)
An unexpected end of file occurred. One or more compound
statements are still unfinished (i.e. the closing brace “}” has not
been found). The line number where the compound statement
started is given in the message.
unknown directive
The character “#” appears first at a line, but no valid directive
was specified.
array index out of bounds
The array index is larger than the highest valid entry of the array.
array must be indexed (variable name)
An array as a whole cannot be used in a expression; you must
indicate an element of the array between square brackets.
argument does not have a default value (argument index )
You can only use the argument placeholder when the function
definition specifies a default value for the argument.
argument type mismatch (argument index )
The argument that you pass is different from the argument that the
function expects, and the compiler cannot convert the passed-in
argument to the required type. For example, you cannot pass the
literal value “1” as an argument when the function expects an
array or a reference.
empty statement
The line contains a semicolon that is not preceded by an expression.
pawn does not support a semicolon as an empty statement, use an
empty compound block instead.
Empty compound
block: 111
Error and warning messages
invalid string (possibly non-terminated string)
A string was not well-formed; for example, the final quote that
ends a string is missing, or the filename for the #include directive
was not enclosed in double quotes or angle brackets.
extra characters on line
There were trailing characters on a line that contained a directive
(a directive starts with a # symbol, see page 116).
constant symbol has no size
A variable has a size (measured in a number of cells), a constant
has no size. That is, you cannot use a (symbolic) constant with
the sizeof operator, for example.
duplicate “case” label (value value)
A preceding “case label” in the list of the switch statement
evaluates to the same value.
invalid ellipsis, array size is not known
You used a syntax like “arr[] = { 1, ... };”, which is invalid,
because the compiler cannot deduce the size of the array from the
invalid combination of class specifiers
A function or variable is denoted as both “public” and “native”,
which is unsupported. Other combinations may also be unsupported; for example, a function cannot be both “public” and
“stock” (a variable may be declared both “public” and “stock”).
character constant value exceeds range for a packed
When the error occurs on a literal string, it is usually an attempt
to store a Unicode character in a packed string where a packed
character is 8-bits. For a literal array, one of the constants does
not fit in the range for packed characters.
positional parameters must precede all named parameters
When you mix positional parameters and named parameters in a
function call, the positional parameters must come first.
too many function arguments
The maximum number of function arguments is currently limited
to 64.
Error and warning messages
unknown array size (variable name)
For array assignment, the size of both arrays must be explicitly
defined, also if they are passed as function arguments.
array sizes do not match, or destination array is too small
For array assignment, the arrays on the left and the right side of the
assignment operator must have the same number of dimensions.
In addition:
⋄ for multi-dimensional arrays, both arrays must have the same
size —note that an unpacked array does not fit in a packed array
with the same number of elements;
⋄ for single arrays with a single dimension, the array on the left
side of the assignment operator must have a size that is equal or
bigger than the one on the right side.
When passing arrays to a function argument, these rules also hold
for the array that is passed to the function (in the function call)
versus the array declared in the function definition.
When a function returns an array, all return statements must
specify an array with the same size and dimensions.
array dimensions do not match
For an array assignment, the dimensions of the arrays on both sides
of the “=” sign must match; when passing arrays to a function
argument, the arrays passed to the function (in the function call)
must match with the definition of the function arguments.
When a function returns an array, all return statements must
specify an array with the same size and dimensions.
invalid line continuation
A line continuation character (a backslash at the end of a line) is
at an invalid position, for example at the end of a file or in a single
line comment.
invalid range
A numeric range with the syntax “n1 .. n2”, where n1 and n2
are numeric constants, is invalid. Either one of the values in not a
valid number, or n1 is not smaller than n2.
Single line comment: 96
Named versus
positional parameters: 73
Error and warning messages
invalid subscript, use “[ ]” operators on major dimensions
and for named indices
You can use the “character array index” operator (braces: “{ }”
only for the last dimension, and only when indexing the array with
a number. For other dimensions, and when indexing the array
with a “symbolic index” (one that starts with a “.”), you must use
the cell index operator (square brackets: “[ ]”).
multi-dimensional arrays must be fully initialized
If an array with more than one dimension is initialized at its
declaration, then there must be equally many literal vectors/subarrays at the right of the equal sign (“=”) as specified for the major
dimension(s) of the array.
exceeding maximum number of dimensions
The current implementation of the pawn compiler only supports
arrays with one or two dimensions.
unmatched closing brace
A closing brace (“}”) was found without matching opening brace
start of function body without function header
An opening brace (“{”) was found outside the scope of a function.
This may be caused by a semicolon at the end of a preceding
function header.
arrays, local variables and function arguments cannot be
A local variable or a function argument starts with the character
“@”, which is invalid.
Unfinished expression before compiler directive
Compiler directives may only occur between statements, not inside
a statement. This error typically occurs when an expression
statement is split over multiple lines and a compiler directive
appears between the start and the end of the expression. This is
not supported.
duplicate argument; same argument is passed twice
In the function call, the same argument appears twice, possibly
through a mixture of named and positional parameters.
Error and warning messages
function argument may not have a default value (variable
All arguments of public functions must be passed explicitly. Public
functions are typically called from the host application, who has
no knowledge of the default parameter values. Arguments of user
defined operators are implied from the expression and cannot be
inferred from the default value of an argument.
multiple “#else” directives between “#if . . . #endif
Two or more #else directives appear in the body between the
matching #if and #endif.
“#elseif ” directive follows an “#else” directive
All #elseif directives must appear before the #else directive.
This error may also indicate that an #endif directive for a higher
level is missing.
number of operands does not fit the operator
When redefining an operator, the number of operands that the
operator has (1 for unary operators and 2 for binary operators)
must be equal to the number of arguments of the operator function.
function result tag of operator name must be name
Logical and relational operators are defined as having a result that
is either true (1) or false (0) and having a “bool:” tag. A user
defined operator should adhere to this definition.
cannot change predefined operators
One cannot define operators to work on untagged values, for
example, because pawn already defines this operation.
function argument may only have a single tag (argument
In a user defined operator, a function argument may not have
multiple tags.
function argument may not be a reference argument or
an array (argument number)
In a user defined operator, all arguments must be cells (non-arrays)
that are passed “by value”.
#pragma rational:
Forward declaration: 80
Error and warning messages
variable cannot be both a reference and an array (variable
A function argument may be denoted as a “reference” or as an
array, but not as both.
invalid rational number precision in #pragma
The precision was negative or too high. For floating point rational
numbers, the precision specification should be omitted.
rational number format already defined
This #pragma conflicts with an earlier #pragma that specified a
different format.
rational number support was not enabled
A rational literal number was encountered, but the format for
rational numbers was not specified.
user-defined operator must be declared before use (function name)
Like a variable, a user-defined operator must be declared before
its first use. This message indicates that prior to the declaration
of the user-defined operator, an instance where the operator was
used on operands with the same tags occurred. This may either
indicate that the program tries to make mixed use of the default
operator and a user-defined operator (which is unsupported), or
that the user-defined operator must be “forwardly declared”.
“sizeof ” operator is invalid on “function” symbols
You used something like “sizeof MyCounter” where the symbol
“MyCounter” is not a variable, but a function. You cannot request
the size of a function.
function argument must be an array (argument name)
The function argument is a constant or a simple variable, but the
function requires that you pass an array.
#define pattern must start with an alphabetic character
Any pattern for the #define directive must start with a letter,
an underscore (“_”) or an “@”-character. The pattern is the first
word that follows the #define keyword.
Error and warning messages
input line too long (after substitutions)
Either the source file contains a very long line, or text substitutions
make a line that was initially of acceptable length grow beyond its
bounds. This may be caused by a text substitution that causes
recursive substitution (the pattern matching a portion of the
replacement text, so that this part of the replacement text is also
matched and replaced, and so forth).
syntax error in the expression, or invalid function call
The expression statement was not recognized as a valid statement
(so it is a “syntax error”). From the part of the string that
was parsed, it looks as if the source line contains a function
call in a “procedure call” syntax (omitting the parentheses), but
the function result is used —assigned to a variable, passed as a
parameter, used in an expression. . .
malformed UTF-8 encoding, or corrupted file: filename
The file starts with an UTF-8 signature, but it contains encodings
that are invalid UTF-8. If the source file was created by an
editor or converter that supports UTF-8, the UTF-8 support is
function uses both “return” and “return <value>”
The function returns both with and without a return value. The
function should be consistent in always returning with a function
result, or in never returning a function result.
inconsistent return types (array & non-array)
The function returns both values and arrays, which is not allowed.
If a function returns an array, all return statements must specify
an array (of the same size and dimensions).
unknown symbol, or not a constant symbol (symbol name)
Where a constant value was expected, an unknown symbol or a
non-constant symbol (variable) was found.
user-defined operators and native functions may not have
Only standard and public functions may have states.
State specifiers:
Fall-back: 81
Error and warning messages
a function or variable may only belong to a single automaton (symbol name)
There are multiple automatons in the state declaration for the
indicated function or variable, which is not supported. In the case
of a function: all instances of the function must belong to the same
automaton. In the case of a variable: it is allowed to have several
variables with the same name belonging to different automatons,
but only in separate declarations —these are distinct variables.
state conflict: one of the states is already assigned to
another implementation (symbol name)
The specified state appears in the state specifier of two implementations of the same function.
no states are defined for symbol name
When this error occurs on a function, this function has a fall-back
implementation, but no other states. If the error refers to a
variable, this variable does not have a list of states between the <
and > characters. Use a state-less function or variable instead.
unknown automaton name
The “state” statement refers to an unknown automaton.
unknown state name for automaton name
The “state” statement refers to an unknown state (for the
specified automaton).
public variables and local variables may not have states
(symbol name)
Only standard (global) variables may have a list of states (and an
automaton) at the end of a declaration.
state variables may not be initialized (symbol name)
Variables with a state list may not have initializers. State variables
should always be initialized through an assignment (instead of at
their declaration), because their initial value is indeterminate.
public functions may not return arrays (symbol name)
A public function may not return an array. Returning arrays is
allowed only for normal functions.
Error and warning messages
first constant in an enumerated list must be initialized
(symbol name)
The first constant in a list of enumerated symbolic constants must
be set to a value. Any subsequent symbol is automatically set the
the value of the preceding symbol +1.
invalid number format
A symbol started with a digit, but is is not a valid number.
array fields with a size may only appear in the final
In the final dimension (the “minor” dimension), the fields of an
array may optionally be declared with a size that is different from
a single cell. On the major dimensions of an array, this is not
valid, however.
invalid subscript, subscript does not match array definition
regarding named indices (symbol name)
Either the array was declared with symbolic subscripts and you
are indexing it with an expression, or you are indexing the array
with a symbolic subscript which is not defined for the array.
• Fatal Errors
cannot read from file: filename
The compiler cannot find the specified file or does not have access
to it.
cannot write to file: filename
The compiler cannot write to the specified output file, probably
caused by insufficient disk space or restricted access rights (the file
could be read-only, for example).
table overflow: table name
An internal table in the pawn parser is too small to hold the
required data. Some tables are dynamically growable, which
means that there was insufficient memory to resize the table. The
“table name” is one of the following:
“staging buffer”: the staging buffer holds the code generated for
an expression before it is passed to the peephole optimizer. The
Enumerated constants: 67
Symbolic subscripts: 64
Error and warning messages
staging buffer grows dynamically, so an overflow of the staging
buffer basically is an “out of memory” error.
“loop table”: the loop table is a stack used with nested do, for, and
while statements. The table allows nesting of these statements up
to 24 levels.
“literal table”: this table keeps the literal constants (numbers,
strings) that are used in expressions and as initiallers for arrays.
The literal table grows dynamically, so an overflow of the literal
table basically is an “out of memory” error.
“compiler stack”: the compiler uses a stack to store temporary
information it needs while parsing. An overflow of this stack is
probably caused by deeply nested (or recursive) file inclusion. The
compiler stack grows dynamically, so an overflow of the compiler
stack basically is an “out of memory” error.
“option table”: in case that there are more options on the command
line or in the response file than the compiler can cope with.
See also #pragma
amxlimit on page
insufficient memory
General “out of memory” error.
incompatible options: option versus option
Two option that are passed to the pawn compiler conflict with
each other, or an option conflicts with the configuration of the
pawn compiler.
numeric overflow, exceeding capacity
A numeric constant, notably a dimension of an array, is too large
for the compiler to handle. For example, when compiled as a 16-bit
application, the compiler cannot handle arrays with more than
32767 elements.
compiled script exceeds the maximum memory size
(number bytes)
The memory size for the abstract machine that is needed to run
the script exceeds the value set with #pragma amxlimit. This
means that the script is too large to be supported by the host.
You might try reducing the script’s memory requirements by:
⋄ setting a smaller stack/heap area —see #pragma dynamic at
page 119;
Error and warning messages
⋄ using packed strings instead of unpacked strings —see pages 99
and 136;
⋄ using overlays —see pages 120 and page 170 for more information
on overlays.
⋄ putting repeated code in separate functions;
⋄ putting repeated data (strings) in global variables;
⋄ trying to find more compact algorithms to perform the same
too many error/warning messages on one line
A single line that causes several error/warning messages is often
an indication that the pawn parser is unable to “recover” from
an earlier error. In this situation, the parser is unlikely to make
any sense of the source code that follows —producing only (more)
inappropriate error messages. Therefore, compilation is halted.
codepage mapping file not found
The file for the codepage translation that was specified with the
-c compiler option or the #pragma codepage directive could not
be loaded.
invalid path: path name
A path, for example for include files or codepage files, is invalid.
Check the compiler options and, if used, the configuration file.
Configuration file:
assertion failed: expression
Compile-time assertion failed.
#assert directive:
user error: message
The parser fell on an #error directive.
#error directive:
overlay function name exceeds limit by value bytes
The size of a function is too large for the overlay system. To fix
this issue, you will have to split the function into two (or more)
#pragma codepage: 119
Error and warning messages
• Warnings
symbol is truncated to number characters
The symbol is longer than the maximum symbol length. The
maximum length of a symbol depends on whether the symbol is
native, public or neither. Truncation may cause different symbol
names to become equal, which may cause error 021 or warning
redefinition of constant/macro (symbol name)
The symbol was previously defined to a different value, or the text
substitution macro that starts with the prefix name was redefined
with a different substitution text.
number of arguments does not match definition
At a function call, the number of arguments passed to the
function (actual arguments) differs from the number of formal
arguments declared in the function heading. To declare functions
with variable argument lists, use an ellipsis (...) behind the
last known argument in the function heading; for example:
print(formatstring,...); (see page 78).
symbol is never used: identifier
A symbol is defined but never used. Public functions are excluded
from the symbol usage check (since these may be called from the
symbol is assigned a value that is never used: identifier
A value is assigned to a symbol, but the contents of the symbol
are never accessed.
redundant code: constant expression is zero
Where a conditional expression was expected, a constant expression
with the value zero was found, e.g. “while (0)” or “if (0)”.
The the conditional code below the test is never executed, and it
is therefore redundant.
redundant test: constant expression is non-zero
Where a conditional expression was expected, a constant expression
with a non-zero value was found, e.g. if (1). The test is
redundant, because the conditional code is always executed.
To create an endless loop, use for ( ;; ) instead of while (1).
Error and warning messages
unknown “#pragma”
The compiler ignores the pragma. The #pragma directives may
change between compilers of different vendors and between different
versions of a compiler of the same version.
function with tag result used before definition, forcing
When a function is “used” (invoked) before being declared, and
that function returns a value with a tag name, the parser must
make an extra pass over the source code, because the presence
of the tag name may change the interpretation of operators (in
the presence of user-defined operators). You can speed up the
parsing/compilation process by declaring the relevant functions
before using them.
function should return a value
The function does not have a return statement, or it does not have
an expression behind the return statement, but the function’s
result is used in a expression.
possible use of symbol before initialization: identifier
A local (uninitialized) variable appears to be read before a value is
assigned to it. The compiler cannot determine the actual order of
reading from and storing into variables and bases its assumption of
the execution order on the physical appearance order of statements
an expressions in the source file.
possibly unintended assignment
Where a conditional expression was expected, the assignment
operator (=) was found instead of the equality operator (==). As
this is a frequent mistake, the compiler issues a warning. To avoid
this message, put parentheses around the expression, e.g. if (
(a=2) ).
possibly unintended bitwise operation
Where a conditional expression was expected, a bitwise operator
(& or |) was found instead of a Boolean operator (&& or ||). In
situations where a bitwise operation seems unlikely, the compiler
issues this warning. To avoid this message, put parentheses around
the expression.
User-defined operators: 84
Forward declaration: 80
Tags are discussed
on page 66
Error and warning messages
tag mismatch
A tag mismatch occurs when:
⋄ assigning to a tagged variable a value that is untagged or that
has a different tag
⋄ the expressions on either side of a binary operator have different
⋄ in a function call, passing an argument that is untagged or that
has a different tag than what the function argument was defined
⋄ indexing an array which requires a tagged index with no tag or
a wrong tag name
possibly a “const” array argument was intended: identifier
Arrays are always passed by reference. If a function does not
modify the array argument, however, the compiler can sometimes
generate more compact and quicker code if the array argument is
specifically marked as “const”.
expression has no effect
The result of the expression is apparently not stored in a variable or
used in a test. The expression or expression statement is therefore
nested comment
pawn does not support nested comments.
loose indentation
Statements at the same logical level do not start in the same
column; that is, the indents of the statements are different.
Although pawn is a free format language, loose indentation
frequently hides a logical error in the control flow.
The compiler can also incorrectly assume loose indentation if the
tab size with which you indented the source code differs from the
assumed size. This may happen if the source files use a mixture
of tab and space characters to indent lines. Sometimes it is then
needed to tell the pawn parser what tab size to use, see #pragma
tabsize on page 121 or the compiler option -t on page 170.
You can also disable this warning with #pragma tabsize 0 or the
compiler option -t:0.
Error and warning messages
old style prototypes used with optional semicolon
When using “optional semicolons”, it is preferred to explicitly
declare forward functions with the forward keyword than using
terminating semicolon.
local variable identifier shadows a symbol at a preceding
A local variable has the same name as a global variable, a function,
a function argument, or a local variable at a lower precedence
level. This is called “shadowing”, as the new local variable makes
the previously defined function or variable inaccessible.
Note: if there are also error messages further on in the script
about missing variables (with these same names) or brace level
problems, it could well be that the shadowing warnings are due to
these syntactical and semantical errors. Fix the errors first before
looking at the shadowing warnings.
expression with tag override must appear between parentheses
In a case statement and in expressions in the conditional operator
(“ ? : ”), any expression that has a tag override should be enclosed between parentheses, to avoid the colon to be misinterpreted
as a separator of the case statement or as part of the conditional
label name identifier shadows tag name
A code label (for the goto instruction) has the same name as a
previously defined tag. This may indicate a faultily applied tag
override; a typical case is an attempt to apply a tag override on the
variable on the left of the = operator in an assignment statement.
number of digits exceeds rational number precision
A literal rational number has more decimals in its fractional part
than the precision of a rational number supports. The remaining
decimals are ignored.
redundant “sizeof ”: argument size is always 1 (symbol
A function argument has a as its default value the size of another
argument of the same function. The “sizeof” default value is only
useful when the size of the referred argument is unspecified in
forward declaration: 80
Error and warning messages
the declaration of the function; i.e., if the referred argument is an
#endif: 116
indeterminate array size in “sizeof ” expression (symbol
The operand of the sizeof operator is an array with an unspecified
size. That is, the size of the variable cannot be determined
at compile time. If used in an “if” instruction, consider a
conditionally compiled section, replacing if by #if.
unreachable code
The indicated code will never run, because an instruction before
(above) it causes a jump out of the function, out of a loop
or elsewhere. Look for return, break, continue and goto
instructions above the indicated line. Unreachable code can also
be caused by an endless loop above the indicated line.
a variable is assigned to itself (symbol name)
There is a statement like “x = x” in the code. The parser
checks for self assignments after performing any text and constant
substitutions, so the left and right sides of an assignment may
appear to be different at first sight. For example, if the symbol
“TWO” is a constant with the value 2, then “var[TWO] = var[2]”
is also a self-assignment.
Self-assignments are, of course, redundant, and they may hide
an error (assignment to the wrong variable, error in declaring
Note that the pawn parser is limited to performing “static checks”
only. In this case it means that it can only compare array
assignments for self-assignment with constant array indices.
more initiallers than array fields
An array that is declared with sumbolic subscripts contains more
values/fields as initiallers than there are (symbolic) subscripts.
length of initialler exceeds size of the array field
The initialler for an array element contains more values than the
size of that field allows. This occurs in an array that has symbolic
subscripts, and where a particular subscript is declared with a size.
Error and warning messages
mixing packed and unpacked array indexing or array
An array is declared as packed (with { and } braces) but indexed
as unpacked (with [ and ]), or vice versa. Or one array is assigned
to another and one is packed while the other is unpacked.
no implementation for state name in function name, no
A function is lacking an implementation for the indicated state.
The compiler cannot (statically) check whether the function will
ever be called in that state, and therefore it issues this warning.
When the function would be called for the state for which no
implementation exists, the abstract machine aborts with a run
time error.
See page 81 on how to specify a fall-back function, and page 42 for
a description and an example.
state specification on forward declaration is ignored
A state specification is redundant on forward declarations. The
function signature must be equal for all states. Only the implementations of the function are state-specific.
native function lacks a predefined index (symbol name)
The pawn compiler was configured with predefined indices for
native functions, but it encountered a declaration for which it
does not have an index declaration. This usually means that the
script uses include files that arenot appropriate for the active
state variable name shadows a global variable
The state variable has the same name as a global variable (without
state specifiers). This means that the global variable is inaccessible
for a function with one of the same states as those of the variable.
function is deprecated (symbol name)
The script uses a function which as marked as “deprecated”. The
host application can mark (native) functions as deprecated when
better alternatives for the function are available or if the function
may not be supported in future versions of the host application.
State specifiers:
Error and warning messages
public function lacks forward declaration (symbol name)
The script defines a public function, but no forward declaration of
this function is present. Possibly the function name was written
incorrectly. The requirement for forward declarations of public
functions guards against a common error.
unknown parameter in substitution (incorrect #define
A #define pattern contains a parameter in the replacement (e.g.
“%1”), that is not in the match pattern. See page 92 for the
preprocessor syntax.
recursive function name
The specified function calls itself recursively. Although this is valid
in pawn, a self-call is often an error. Note that this warning is
only generated when the pawn parser/compiler is set to “verbose”
mixing string formats in concatenation
In concatenating literals strings, strings with different formats
(such as packed versus unpacked, and “plain” versus standard
strings) were combined. This is usually an error. The parser uses
the format of the first (left-most) string in the concatenation for
the result.
appendix b
The compiler
Many applications that embed the pawn scripting language use the standalone compiler that comes with the pawn toolkit. The pawn compiler is a
command-line utility, meaning that you must run it from a “console window”, a terminal/shell, or a “DOS box” (depending on how your operating
system calls it).
• Usage
Assuming that the command-line pawn compiler is called “pawncc” (unix/
Linux) or “pawncc.exe” (DOS/Windows), the command line syntax is:
pawncc <filename> [more filenames...] [options]
The input file name is any legal filename. If no extension is given, “.pawn”
or “.p” is assumed. The compiler creates an output file with, by default,
the same name as the input file and the extension “.amx”.
After switching to the directory with the sample programs, the command:
pawncc hello
should compile the very first “hello world” example (page 3). Should, because
the command implies that:
⋄ the operating system can locate the “pawncc” program —you may need
to add it to the search path;
⋄ the pawn compiler is able to determine its own location in the file system
so that it can locate the include files —a few operating systems do not
support this and require that you use the -i option (see below).
• Input file
The input file for the pawn compiler, the “source code” file for the script/
program, must be a plain text file. All reserved words and all symbol names
(names for variables, functions, symbolic constants, tags, . . . ) must use the
ascii character set. Literal strings, i.e text between quotes, may be in
extended ascii, such as one of the sets standardized in the ISO 8859 norm
—ISO 8859-1 is the well known “Latin 1” set.
The pawn compiler also supports UTF-8 encoded text files, which are practical in an environment based on Unicode or UCS-4. The pawn compiler
strings: 99
Character constants: 98
The compiler
only recognizes UTF-8 encoded characters inside unpacked strings and character constants. The compiler interprets the syntax rules for UTF-8 files
strictly; non-conforming UTF-8 files are not recognized. The input file may
have, but does not require, a “Byte Order Mark” signature; the compiler
recognizes the UTF-8 format based on the file’s content.
• Options
Options start with a dash (“-”) or, on Microsoft Windows and DOS, with a
forward slash (“/”). In other words, all platforms accept an option written
as “-a” (see below for the purpose of this option) and the DOS/Windows
platforms accept “/a” as an alternative way to write “-a”.
All options should be separated by at least one space.
Many options accept a value —which is sometimes mandatory. A value may
be separated from the option letter by a colon or an equal sign (a “:” and
a “=” respectively), or the value may be glued to the option letter. Three
equivalent options to set the debug level to two are thus:
⋄ -d2
⋄ -d:2
⋄ -d=2
The options are:
Alignment: the memory address of global and local variables
may optionally be aligned to multiples of the given value. If
not set, all variables are aligned to cell boundaries. The optimal data alignment depends on the hardware architecture.
Assembler: generate a text file with the pseudo-assembler
code for the pawn abstract machine, instead of binary code.
The size of a cell in bits; valid values are 16, 32 and 64.
Codepage: set the codepage for translating the source file
from extended ascii to Unicode/UCS-4. The default is no
translation. The name parameter can specify a full path to a
“mapping file” or just the identifier of the codepage —in the
latter case, the compiler prefixes the identifier with the letters
“cp”, appends the extension “.txt” and loads the mapping
file from a system directory.
The compiler
Directory: the “active” directory, where the compiler should
search for its input files and store its output files.
This option is not supported on every platform. To verify
whether the pawn compiler supports this option, run the
compiler without any option or filename on the command
line. The compiler will then list its usage syntax and all
available options in alphabetical order. If the -D switch is
absent, the option is not available.
Debug level: 0 = none, 1 = bounds checking and assertions
only, 2 = full symbolic information, 3 = full symbolic information and optimizations disabled (same as the combination
-d2 and -O0).
When the debug level is 2 or 3, the pawn compiler also prints
the estimated number of stack/heap space required for the
Error file: set the name of the file into which the compiler
must write any warning and error messages; when set, there
is no output to the screen.
“HWND” (Microsoft Windows version only): the compiler can
optionally post a message to the specified window handle
upon completion of the P-code generation. Host applications
that invoke the pawn compiler can wait for the arrival of this
message or signal the user of the completion of the compile.
The message number that is sent to the window is created
with the Microsoft Windows SDK function RegisterWindowMessage using the name “PawnNotify”. The wParam of
the message holds the compiler return code: 0 = success, 1
= warnings, 2 = errors (plus possibly warnings), 3 = compilation aborted by the user.
Include path: set the path where the compiler can find the
include files. This option may appear multiple times at the
command line, to allow you to set several include paths.
Listing: perform only the file reading and preprocessing steps;
for example, to verify the effects of macro expansion and the
conditionally compiled/skipped sections.
The compiler
Optimization level: 0 = no optimizations, 1 = core instructions only, 2 = core & supplemental instructions, 3 = core,
supplemental & packed instructions. Optimization level 1 is
compatible with JIT implementations (JIT = “Just In Time”
compiler, a high-performance abstract machine).
Output file: set the name and path of the binary output file.
Prefix file: the name of the “prefix file”, this is a file that
is parsed before the input file (as a kind of implicit “include
file”). If used, this option overrides the default include file
“”. The -p option on its own (without a filename) disables the processing of any implicit include file.
Report: enable the creation of the report and optionally set
the filename to which the extracted documentation and a
cross-reference report will be written.
The report is in “XML” format. The filename parameter is
optional; if not specified, the report file has the same name
as the output file with the extension “.XML”.
#pragma dynamic: 119
Configuration file:
Stack size: the size of the stack and the heap in cells.
Skip count: the number of lines to skip in the input file before
starting to compile; for example, to skip a “header” in the
source file which is not in a valid pawn syntax.
Template file: the name of the configuration file. If no extension is present, the file extension is “.cfg”; if no path
is present, the file is loaded from the directory where the
pawn compiler resides too. The default configuration file is
tab size: the number of space characters to use for a tab
character. Without this option, the pawn parser will autodetect the tab.
Overlays: generate the necessary tables and instructions for
running the code from overlays. Using overlays reduces the
memory requirements for a compiled script, because code sections can be swapped into and out off memory dynamically.
Since overlay loading takes time, the resulting script will also
run slower. Use -V+ to enable overlays and -V- to create
The compiler
monolithic code —alternatively, use -V followed by a value
to enable overlays and set the size of the overlay pool. The
option -V without any suffix toggles the current setting.
Verbose: display informational messages during the compilation. The value can be 0 (zero) for “quiet” compile, 1 (one)
for the normal output and 2 for a code/data/stack usage report.
Warning control: the warning number following the “-w” is
enabled or disabled, depending on whether a “+” or a “-”
follows the number. When a “+” or “-” is absent, the warning
status is toggled. For example, -w225- disables the warning
for “unreachable code”, -w225+ enables it and -w225 toggles
between enabled/disabled.
Warnings: 162
Only warnings can be disabled (errors and fatal errors cannot
be disabled). By default, all warnings are enabled.
Limit for the abstract machine: the maximum memory requirements that a compiled script may have, in bytes. This
value is is useful for (embedded) environments where the maximum size of a script is bound to a hard upper limit. If there
is no setting for the amount of RAM for the data and stack,
this refers to the total memory requirements; if the amount
of RAM is explicitly set, this value only gives the amount of
memory needed for the code and the static data.
RAM limit for the abstract machine: the maximum memory
requirements for data and stack that a compiled script may
have, in bytes. This value is is useful for (embedded) environments where the maximum data size of a script is bound
to a hard upper limit. Especially in the case where the pawn
script runs from ROM, the sizes for the code and data sections
need both to be set.
Control characters start with “\” (for the sake of similarity
with C, C++ and Java).
Control characters start with “ˆ” (for compatibility with earlier versions of pawn).
See also #pragma
amxlimit on page
See also #pragma
amxram on page
The compiler
With -;+ every statement is required to end with a semicolon;
with -;-, semicolons are optional to end a statement if the
statement is the last on the line. The option -; (without +
or − suffix) toggles the current setting.
With -(+, arguments passed to a function must be enclosed
in parentheses; with -(-, parentheses are optional if the expression result of the function call is not used. The option
-( (without + or − suffix) toggles the current setting. See
the section “Calling functions” on page 72 for details on the
optional parentheses around function arguments.
define constant “sym” with the given (numeric) value, the
value is optional;
read (more) options from the specified “response file”.
• Response file
To support operating systems with a limited command line length (e.g.,
Microsoft DOS), the pawn compiler supports “response files”. A response
file is a text file that contains the options that you would otherwise put at
the command line. With the command:
pawncc @opts.txt prog.pawn
the pawn compiler compiles the file “prog.pawn” using the options that are
listed in the response file “opts.txt”.
• Configuration file
On platforms that support it (currently Microsoft DOS, Microsoft Windows
and Linux), the compiler reads the options in a “configuration file” on startup. By default, the compiler looks for the file “pawn.cfg” in the same
directory as the compiler executable program. You can specify a different
configuration file with the “-T” compiler option.
In a sense, the configuration file is an implicit response file. Options specified on the command line may overrule those in the configuration file. One
difference from a response file is that the configuration file may also contain instructions and options for an IDE, such as Quincy. These auxiliary
instructions start with “#”-character, which the pawn compiler treats as a
comment. For details of the instructions supported by Quincy, please see
the Quincy manual.
appendix c
The first issue in the presentation of a new computer language should be:
why a new language at all?
Indeed, I did look at several existing languages before I designed my own.
Many little languages were aimed at scripting the command shell (TCL,
Perl, Python). Other languages were not designed as extension languages,
and put the burden to embedding solely on the host application.
As I initially attempted to use Java as an extension language (rather than
build my own, as I have done now), the differences between pawn and Java
are illustrative for the almost reciprocal design goals of both languages. For
example, Java promotes distributed computing where “packages” reside on
diverse machines, pawn is designed so that the compiled applets can be easily stored in a compound file together with other data. Java is furthermore
designed to be architecture neutral and application independent, inversely
pawn is designed to be tightly coupled with an application; native functions
are a taboo to some extent in Java (at least, it is considered “impure”),
whereas native functions are “the reason to be” for pawn. From the viewpoint of pawn, the intended use of Java is upside down: native functions are
seen as an auxiliary library that the application —in Java— uses; in pawn,
native functions are part of “the application” and the pawn program itself
is a set of auxiliary functions that the application uses.
A language for scripting applications and devices: pawn is targeted
as an extension language, meant to write application-specific macros or subprograms with. pawn is not the appropriate language for implementing
business applications or operating systems in. pawn is designed to be easily
integrated with, and embedded in, other systems/applications. It is also designed to run in resource-constrained environments, such as on small microcontrollers.
As an extension language, pawn programs typically manipulate objects of
the host application. In an animation system, pawn scripts deal with sprites,
events and time intervals; in a communication application, pawn scripts handle packets and connections. I assume that the host application or the device
makes (a subset of) its resources and functionality available via functions,
handles, magic cookies. . . in a similar way that a contemporary operating
system provides an interface to processes written in C/C++ —e.g., the Win32
API (“handles everywhere”) or GNU/Linux’ “glibc”. To that end, pawn
has a simple and efficient interface to the “native” functions of the host application. A pawn script manipulates data objects in the host application
through function calls, but it cannot access the data of the host application
The first and foremost criteria for the pawn language were execution speed
and reliability. Reliability in the sense that a pawn program should not be
able to crash the application or tool in which it is embedded —at least, not
easily. Although this limits the capabilities of the language significantly, the
advantages are twofold:
⋄ the application vendor can rest assured that its application will not crash
due to user additions or macros,
⋄ the user is free to experiment with the language with no (or little) risk of
damaging the application files.
Speed is essential: pawn programs would probably run in an abstract machine, and abstract machines are notoriously slow. I had to make a language
that has low overhead and a language for which a fast abstract machine can
be written. Speed should also be reliable, in the sense that a pawn script
should not slow down over time or have an occasional performance hiccup.
Consequently, pawn excludes any required “background process”, such as
garbage collection, and the core of the abstract machine does not implicitly
allocate any system or application resources while it runs. That is, pawn
does not allocate memory or open files, not without the help of a native
function that the script calls explicitly.
As Dennis Ritchie said, by intent the C language confines itself to facilities
that can be mapped relatively efficiently and directly to machine instructions. The same is true for pawn, and this is also a partial explication why
pawn looks so much like C. Even though pawn runs on an abstract machine,
the goal is to keep that abstract machine small and quick. pawn is used in
tiny embedded systems with ram sizes of 32 kiB or less, as well as in highperformance games that need every processor cycle for their graphics engine
and game-play. In both environments, a heavy-weight scripting support is
difficult to swallow.
A brief analysis showed that the instruction decoding logic for an abstract
machine would quickly become the bottleneck in the performance of the
abstract machine. The quickest way to dispatch instructions would be to use
the opcode as an index in a jump table. Therefore all opcodes should have
the same size (excluding operands), and the opcode should fully specify the
instruction (including the addressing methods, size of the operands, etc.).
That meant that for each operation on a variable, the abstract machine
needed a separate opcode for every combination of variable type, storage
class and access method (direct, or dereferenced). For even three types (int,
char and unsigned int), two storage classes (global and local) and three
access methods (direct, indirect or indexed), a total of 18 opcodes (3*2*3)
are needed to simply fetch the value of a variable.
To get an abstract machine that is both small and quick, the number of
opcodes should be kept to a minimum:∗ each “virtual instruction” needs to
be handled by the abstract machine, and therefore takes code space. With
18 opcodes to load a variable in a register, 18 more to store a register into
a variable, another 18 to get the address of a variable, etc. . . the abstract
machine that I envisioned was quickly growing out of its desired proportions.
The languages bob and rexx inspired me to design a typeless language.
This saved me a lot of opcodes. At the same time, the language could no
longer be called a “subset of C”. I was changing the language. Why, then,
not go a foot further in changing the language? This is where a few more
design guidelines came into play:
⋄ give the programmer a general purpose tool, not a special purpose solution
⋄ avoid error prone language constructs; promote error checking
⋄ be pragmatic
A general purpose tool: pawn is targeted as an extension language,
without specifying exactly what it will extent. Typically, the application or
the tool that uses pawn for its extension language will provide many, optimized routines or commands to operate on its native objects, be it text,
database records or animated sprites. The extension language exists to permit the user to do what the application developer forgot, or decided not to
include. Rather than providing a comprehensive library of functions to sort
data, match regular expressions, or draw cubic Bézier splines, pawn should
supply a (general purpose) means to use, extend and combine the specific
(“native”) functions that an application provides.
pawn lacks a comprehensive standard library. By intent, pawn also lacks
features like pointers, dynamic memory allocation, direct access to the op∗
136 Opcodes are defined at this writing, plus 20 “macro” opcodes.
erating system or to the hardware, that are needed to remain competitive in
the field of general purpose application or system programming. You cannot build linked lists or dynamic tree data structures in pawn, and neither
can you access any memory beyond the boundaries of the abstract machine.
That is not to say that a pawn program can never use dynamic, sorted
symbol tables, or change a parameter in the operating system; it can do
that, but it needs to do so by calling a “native” function that an application
provides to the abstract machine.
In other words, if an application chooses to implement the well known peek
and poke functions (from basic) in the abstract machine, a pawn program
can access any byte in memory, insofar the operating system permits this.
Likewise, an application can provide native functions that insert, delete or
search symbols in a table and allows several operations on them. The proposed core functions getproperty and setproperty are an example of native functions that build a linked list in the background.
Promote error checking: As you may have noticed, one of the foremost
design criteria of the C language, “trust the programmer”, is absent from
my list of design criteria. Users of script languages may not be experienced
programmers; and even if they are, pawn will probably not be their primary
language. Most pawn programmers will keep learning the language as they
go, and will even after years not have become experts. Enough reason, hence,
to replace error prone elements from the C language (pointers) with saver,
albeit less general, constructs (references).† References are copied from C++ .
They are nothing else than pointers in disguise, but they are restricted in
various, mostly useful, ways. Turn to a C++ book to find more justification
for references.
I find it sad that many, even modern, programming languages have so little
built-in, or easy to use, support for confirming that programs do as the programmer intended. I am not referring to theoretical correctness (which is too
costly to achieve for anything bigger than toy programs), but practical, easy
to use, verification mechanisms as a help to the programmer. pawn provides
both compile time and execution time assertions to use for preconditions,
postconditions and invariants.
You should see this remark in the context of my earlier assertion that many “Pawn” programmers will be novice programmers. In my (teaching) experience, novice programmers
make many pointer errors, as opposed to experienced C/C++ programmers.
The typing mechanism that most programming languages use is also an
automatic “catcher” of a whole class of bugs. By virtue of being a typeless
language, pawn lacked these error checking abilities. This was clearly a
weakness, and I created the “tag” mechanism as an equivalent for verifying
function parameter passing, array indexing and other operations.
The quality of the tools: the compiler and the abstract machine, also have a
great impact on the robustness of code —whatever the language. Although
this is only very loosely related to the design of the language, I set out
to build the tools such that they promote error checking. The warning
system of pawn goes a step beyond simply reporting where the parser fails to
interpret the data according to the language grammar. At several occasions,
the compiler runs checks that are completely unrelated to generating code
and that are implemented specifically to catch possible errors. Likewise, the
“debugger hook” is designed right into the abstract machine, it is not an
add-on implemented as an after-thought.
Be pragmatic: The object-oriented programming paradigm has not entirely lived up to its promise, in my opinion. On the one hand, OOP solves
many tasks in an easier or cleaner way, due to the added abstraction layer.
On the other hand, contemporary object-oriented languages leave you struggling with the language. The struggle should be with implementing the
functionality for a specific task, not with the language used for the implementation. Object-oriented languages are attractive mainly because of the
comprehensive class libraries that they come with —but leaning on a standard library goes against one of the design goal for pawn. Object-oriented
programming is not a solution for a non-expert programmer with little patience for artificial complexity. The criterion “be pragmatic” is a reminder
to seek solutions, not elegance.
• Practical design criteria
The fact that pawn looks so much like C cannot be a coincidence, and it isn’t.
pawn started as a C dialect and stayed that way, because C has a proven
track record. The changes from C were mostly born out of necessity after
rubbing out the features of C that I did not want in a scripting language:
no pointers and no “typing” system.
pawn, being a typeless language, needed a different means to declare variables. In the course of modifying this, I also dropped the C requirement that
all variables should be declared at the top of a compound statement. pawn
is a little more like C++ in this respect.
C language functions can pass “output values” via pointer arguments. The
standard function scanf, for example, stores the values or strings that it
reads from the console into its arguments. You can design a function in
C so that it optionally returns a value through a pointer argument; if the
caller of the function does not care for the return value, it passes NULL as the
pointer value. The standard function strtol is an example of a function that
does this. This technique frequently saves you from declaring and passing
dummy variables. pawn replaces pointers with references, but references
cannot be NULL. Thus, pawn needed a different technique to “drop” the
values that a function returns via references. Its solution is the use of an
“argument placeholder” that is written as an underscore character (“ ”);
Prolog programmers will recognize it as a similar feature in that language.
The argument placeholder reserves a temporary anonymous data object (a
“cell” or an array of cells) that is automatically destroyed after the function
The temporary cell for the argument placeholder should still have a value,
because the function may see a reference parameters as input/output. Therefore, a function must specify for each passed-by-reference argument what
value it will have upon entry when the caller passes the placeholder instead
of an actual argument. By extension, I also added default values for arguments that are “passed-by-value”. The feature to optionally remove all
arguments with default values from the right was copied from C++ .
Support for Unicode string literals: 138
When speaking of BCPL and B, Dennis Ritchie said that C was invented
in part to provide a plausible way of dealing with character strings when
one begins with a word-oriented language. pawn provides two options for
working with strings, packed and unpacked strings. In an unpacked string,
every character fits in a cell. The overhead for a typical 32-bit implementation is large: one character would take four bytes. Packed strings store up
to four characters in one cell, at the cost of being significantly more difficult
to handle if you could only access full cells. Modern BCPL implementations
provide two array indexing methods: one to get a word from an array and
one to get a character from an array. pawn copies this concept, although
the syntax differs from that of BCPL. The packed string feature also led to
the alternative array indexing syntax.
Unicode applications often have to deal with two characters sets: 8-bit for
legacy file formats and standardized transfer formats (like many of the Internet protocols) and the 16-bit Unicode character set (or the 31-bit UCS-4
character set). Although the pawn compiler has an option that makes characters 16-bit (so only two characters fit in a 32-bit cell), it is usually more
convenient to store single-byte character strings in packed strings and multibyte strings in unpacked strings. This turns a weakness in pawn —the need
to distinguish packed strings from unpacked strings— into a strength: pawn
can make that distinction quite easily. And instead of needing two implementations for every function that deals with strings (an ascii version and a
Unicode version —look at the Win32 API, or even the standard C library),
pawn enables functions to handle both packed and unpacked strings with
Notwithstanding the above mentioned changes, plus those in the chapter
“Pitfalls: differences from C” (page 133), I have tried to keep pawn close
to C. A final point, which is unrelated to language design, but important
nonetheless, is the license: pawn is distributed under a liberal license allowing you to use and/or adapt the code with a minimum of restrictions —see
appendix D.
appendix d
The software toolkit “pawn” (the compiler, the abstract machine and
c 1997–2011 by ITB CompuPhase, and
the support routines) is copyright distributed under the “Apache License” version 2.0, which is reproduced
below, plus an exception clause regarding static linking. See the file notices
for contributions and their respective licenses.
As a special exception to the Apache License 2.0 (and referring to the definitions in Section
1 of this license), you may link, statically or dynamically, the “Work” to other modules
to produce an executable file containing portions of the “Work”, and distribute that executable file in “Object” form under the terms of your choice, without any of the additional
requirements listed in Section 4 of the Apache License 2.0. This exception applies only
to redistributions in “Object” form (not “Source” form) and only if no modifications have
been made to the “Work”.
“License” shall mean the terms and conditions for use, reproduction, and distribution as
defined by Sections 1 through 9 of this document.
“Licensor” shall mean the copyright owner or entity authorized by the copyright owner
that is granting the License.
“Legal Entity” shall mean the union of the acting entity and all other entities that control,
are controlled by, or are under common control with that entity. For the purposes of
this definition, “control” means (i) the power, direct or indirect, to cause the direction
or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty
percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.
“You” (or “Your”) shall mean an individual or Legal Entity exercising permissions granted
by this License.
“Source” form shall mean the preferred form for making modifications, including but not
limited to software source code, documentation source, and configuration files.
“Object” form shall mean any form resulting from mechanical transformation or translation
of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types.
“Work” shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached
to the work (an example is provided in the Appendix below).
“Derivative Works” shall mean any work, whether in Source or Object form, that is based on
(or derived from) the Work and for which the editorial revisions, annotations, elaborations,
or other modifications represent, as a whole, an original work of authorship. For the
purposes of this License, Derivative Works shall not include works that remain separable
from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works
“Contribution” shall mean any work of authorship, including the original version of the
Work and any modifications or additions to that Work or Derivative Works thereof, that is
intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by
an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the
purposes of this definition, “submitted” means any form of electronic, verbal, or written
communication sent to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems, and issue tracking
systems that are managed by, or on behalf of, the Licensor for the purpose of discussing
and improving the Work, but excluding communication that is conspicuously marked or
otherwise designated in writing by the copyright owner as “Not a Contribution.”
“Contributor” shall mean Licensor and any individual or Legal Entity on behalf of whom
a Contribution has been received by Licensor and subsequently incorporated within the
Grant of Copyright License. Subject to the terms and conditions of this License, each
Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royaltyfree, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in
Source or Object form.
Grant of Patent License. Subject to the terms and conditions of this License, each
Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royaltyfree, irrevocable (except as stated in this section) patent license to make, have made, use,
offer to sell, sell, import, and otherwise transfer the Work, where such license applies
only to those patent claims licensable by such Contributor that are necessarily infringed
by their Contribution(s) alone or by combination of their Contribution(s) with the Work
to which such Contribution(s) was submitted. If You institute patent litigation against
any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct or contributory patent
infringement, then any patent licenses granted to You under this License for that Work
shall terminate as of the date such litigation is filed.
Redistribution. You may reproduce and distribute copies of the Work or Derivative
Works thereof in any medium, with or without modifications, and in Source or Object
form, provided that You meet the following conditions:
You must give any other recipients of the Work or Derivative Works a copy of this
License; and
You must cause any modified files to carry prominent notices stating that You changed
the files; and
You must retain, in the Source form of any Derivative Works that You distribute,
all copyright, patent, trademark, and attribution notices from the Source form of the
Work, excluding those notices that do not pertain to any part of the Derivative Works;
If the Work includes a “NOTICE” text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices
contained within such NOTICE file, excluding those notices that do not pertain to
any part of the Derivative Works, in at least one of the following places: within a
NOTICE text file distributed as part of the Derivative Works; within the Source form
or documentation, if provided along with the Derivative Works; or, within a display
generated by the Derivative Works, if and wherever such third-party notices normally
appear. The contents of the NOTICE file are for informational purposes only and do
not modify the License. You may add Your own attribution notices within Derivative
Works that You distribute, alongside or as an addendum to the NOTICE text from
the Work, provided that such additional attribution notices cannot be construed as
modifying the License.
You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of
Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in
this License.
Submission of Contributions. Unless You explicitly state otherwise, any Contribution
intentionally submitted for inclusion in the Work by You to the Licensor shall be under the
terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate
license agreement you may have executed with Licensor regarding such Contributions.
Trademarks. This License does not grant permission to use the trade names, trademarks,
service marks, or product names of the Licensor, except as required for reasonable and
customary use in describing the origin of the Work and reproducing the content of the
NOTICE file.
Disclaimer of Warranty. Unless required by applicable law or agreed to in writing,
Licensor provides the Work (and each Contributor provides its Contributions) on an “AS
or implied, including, without limitation, any warranties or conditions of TITLE, NONINFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions
under this License.
Limitation of Liability. In no event and under no legal theory, whether in tort (including
negligence), contract, or otherwise, unless required by applicable law (such as deliberate
and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You
for damages, including any direct, indirect, special, incidental, or consequential damages
of any character arising as a result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill, work stoppage, computer
failure or malfunction, or any and all other commercial damages or losses), even if such
Contributor has been advised of the possibility of such damages.
Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support,
warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and
on Your sole responsibility, not on behalf of any other Contributor, and only if You agree
to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or
claims asserted against, such Contributor by reason of your accepting any such warranty
or additional liability.
The pawn documentation is copyright 1997–2010
by ITB CompuPhase,
and licensed under the Creative Commons Attribution/ShareAlike 3.0 License. To view a copy of this licence, visit
or send a letter to Creative Commons, 171 Second St, Suite 300, San Francisco, CA 94105 USA. Below is a “human-readable” summary of the Legal
Code (the full licence).
You are free:
to Share — to copy, distribute and transmit the work
to Remix — to adapt the work
Under the following conditions:
Attribution — You must attribute the work in the manner specified
by the author or licensor (but not in any way that suggests that they
endorse you or your use of the work).
Share Alike — If you alter, transform, or build upon this work, you may
distribute the resulting work only under the same, similar or a compatible
⋄ Nothing in this license impairs or restricts the author’s moral rights.
With the understanding that:
Waiver — Any of the above conditions can be waived if you get permission from the copyright holder.
Other Rights — In no way are any of the following rights affected by
the license:
• Your fair dealing or fair use rights;
• The author’s moral rights;
• Rights other persons may have either in the work itself or in how the
work is used, such as publicity or privacy rights.
Notice — For any reuse or distribution, you must make clear to others
the license terms of this work.
⋄ Names of persons (not products) are in italics.
⋄ Function names, constants and compiler reserved words are in typewriter
#, 94
#assert, 116
#define, 92, 95, 116, 134
#endinput, 116
#error, 116
#file, 116
#if, 116
#include, 117
#line, 117
#pragma, 118
#section, 121
#tryinclude, 121
#undef, 95, 122
@-symbol, 61, 82
. . . considered harmful, 4
Actual parameter, 16, 68
Algebraic notation, 26
Alias, See External name
Alignment (variables), 118
Anno Domini, 11
APL, 26
Argument placeholder, 74
symbolic subscripts, 19, 64
Array assignment, 105, 133
Arrays, 65
Progressive initiallers, 63
ASCII, 138, 140, 169, 170
Assertions, 9, 48, 102, 178
Automata theory, 40, 43, 45
Automaton, 37, 81, see also State
anonymous ~, 40, 114
Basic Multilingual Plane, 139
BCPL, 180
Big Endian, 99
Binary Coded Decimals, 102
Binary radix, 97, 133
Bisection, 77
bitcount, 25
Bitwise operators, 23
BOB, 177
Byte Order Mark, 170
Bytecode, See P-code
Cain, Ron, 1
Call by reference, 13, 17, 79
Call by value, 12, 17, 70, 88
Callee (functions), 72
Celsius, 14
Chained relational operators, 17,
Character constants, 98
clamp, 123
clreol, 127
clrscr, 128
Codepage, 119, 138, 140, 141, 170
Coercion rules, 79
Comments, 96
documentation ~, 50, 96
Commutative operators, 87
Compiler options, 170
Compound literals, See Literal
Conditional goto, 133
Configuration file, 172, 174
“const” variables, 62
literals, 97
predefined ~, 101
symbolic ~, 101
Counting bits, 25
Cross-reference, 50, 172
Data declarations, 60–65
arrays, 65
default initialization, 63
global ~, 60
local ~, 60
public ~, 61
stock ~, 61
~ arithmetic, 13
functions, 132
Debug level, 102
Default arguments, 74
Default initialization, 63
Design by contract, 48
Diagnostic, 67, 68, 78, 141, see
also Errors and Warnings
Digit group separator, 97, 98
Directives, 95, 116–122
DLL calls, 132
Documentation comments, 50, 96
Documentation tags, 172
Dr. Dobb’s Journal, 1
Dynamic tree, 178
Eiffel, 49
Ellipsis operator, 63, 71, 78, 100,
enum, 135
Eratosthenes, 7
Error, see also Diagnostic
Errors, 58, 147–161
Escape sequences, 98, 100
Euclides, 5
Event-driven programming model,
33, 35, 36
Extended ASCII, 138, 169
External name, 84, 88
Faculty, 70
faculty, 70
Fahrenheit, 14
Fall-back (state functions), 42, 81
Fibonacci, 8
fibonacci, 9
Fibonacci numbers, 9
File input/output, 132
Fixed point arithmetic, 77, 89, 132
Floating point arithmetic, 89, 98,
132, 133
Flow-driven programming model,
32, 35
Floyd, Robert, 76
Forbidden user-defined operators,
Foreign Function Interface, 132
Formal parameter, 68, 69
Forward declaration, 69, 80
FSM, See Automaton
funcidx, 123
Function library, 123
Functions, 69–84
call by reference, 13, 17, 70
call by value, 12, 17, 70, 88
callee, 72
caller, 72
coercion rules, 79
default arguments, 74
forward declaration, 69, 80
~ index, 123
latent ~, 113
native ~, 9, 83
public ~, 81
standard library ~, 123
state classifier, 81
state entry ~, 40, 60
state exit ~, 45
static ~, 82
stock ~, 83
variable arguments, 78
Gödel, Escher, Bach, 146
gcd, 5
getarg, 124
getchar, 128
getstring, 128
getvalue, 129
Global variables, 60
Golden ratio, 10
gotoxy, 130
Greatest Common Divisor, 5
Gregorian calendar, 10
Hamblin, Charles, 27
Hanoi, the Towers of ~, 80
heapspace, 124
Hendrix, James, 1
Hexadecimal radix, 97, 133
Hofstadter, Douglas R., 146
Host application, 61, 62, 84, 112,
113, 123, 140, 175
Identifiers, 96
Implicit conversions, See coercion
Index tag, 67
Indiction Cycle, 10
Infinite loop, 18
Infix notation, See Algebraic ~
Internationalization, 138
Internet, 180
Intersection (sets), 22
ISO 8859, 100, 138, 169
ISO/IEC 10646-1, 139
ISO/IEC 8824 (date format), 73
ispacked, 136
Jacquard Loom, 37
Java, 175
JIT, 172
Julian Day number, 10
Keywords, See reserved words
Latent function, 113
Latin-1, See ISO 8859
Leap year, 69
leapyear, 69
Leonardo of Pisa, 8
Library call, 132
Library functions, 83
License, 182
Line continuation character, 145
Linear congruential generator, 125
Linked lists, 178
Linux, 139, 175
LISP, 33
Literal array, 71
Literals, See Constants
Local variables, 60
Logo (programming language), 33
Lukasiewicz, Jan, 26
lvalue, 68, 103, 142, 143
actual ~, 16, 68
formal ~, 68, 69
Parser, 4
Placeholder, See Argument ~
Plain strings, 100
Plural tag names, 144, 145
Positional parameters, 73
power, 69
Precedence table, 109
Prefix file, 172
Preprocessor, 92–95
~ macro, 92, 116
Prime numbers, 7
print, 130
printf, 15, 130
Priority queue, 21
Procedure call syntax, 72
Process control, 132
Progressive initiallers, 63
Proleptic Gregorian calendar, 10
Pseudo-random numbers, 125
~ functions, 81, 123
~ variables, 61
Macro, 92, 116
~ prefix, 95, 122
max, 124
Mealy automata, 43
Metonic Cycle, 10
Meyer, Bertrand, 49
Micro-controllers, 175
Microsoft Windows, 139
min, 125
MIRT, 44
Moore automata, 43
Named parameters, 73
Native functions, 9, 83
external name, 84, 88
Newton-Raphson, 78
numargs, 125
Octal radix, 133
Operator precedence, 109
Operators, 103–109
commutative ~, 87
user-defined ~, 84, 142
Optional semicolons, 96
compiler ~, 170
Overlays, 120, 172
P-code, 92
Packed string, 99, 136, 180
Quincy, 58
Quincy (IDE), 174
Quine, 146
random, 125
Random sample, 76
Rational numbers, 14, 29, 98
Recursive functions, 79
Reference arguments, 13, 70, 79
Report, 172
Reserved words, 96
Response file, 174
Reverse Polish Notation, 27
REXX, 177
Ritchie, Dennis, 134, 176, 180
rot13, 16
ROT13 encryption, 16
Scaliger, Josephus, 10
Semicolons, optional, 96
Set operations, 22
setarg, 126
setattr, 131
Shadowing, 165
Shared libraries, 132
Shift-JIS, 139
sieve, 7
Single line comment, 96
sizeof operator, 108
~ in function argument, 75, 77
Small C, 1
Solar Cycle, 10
sqroot, 77
Square root, 77
Standard function library, 123
State, 37, 38
~ classifier, 40, 45, 60, 81
conditional ~, 39
~ diagram, 37, 47
~ entry function, 40, 60
~ exit function, 45
fall-back ~, 42, 81
~ notation, 47
~ operator, 109
unconditional ~, 43
~ variables, 43, 60
Statements, 111–115
~ functions, 82
~ variables, 61
~ functions, 83
~ variables, 61
~ concatenation, 100
packed ~, 99, 136, 180
plain ~, 100
unpacked ~, 99, 136, 180
String manipulation, 132
Stringize operator, 94
strtok, 17
Structure, 19, see also Enumerated
strupper, 137
Surrogate pair, 139, 140
swap, 70
swapchars, 126
Symbolic constants, 101
Symbolic information, 171
Symbolic subscripts (array), 64
Syntax rules, 96
Tag name, 15, 66, 141
~ and enumerated constant, 101
array index, 67
~ operator, 145
~ override, 67, 108, 142
plural tags, 144, 145
predefined ~, 102
strong ~, 67, 143
~ syntax, 102
untag override, 144
weak ~, 67, 142
Tag names, 178
tagof operator, 109
Text substitution, 92, 116
The Towers of Hanoi, 80
Thousands separator, 97, 98
functions, 132
tolower, 126
toupper, 127
Transition (state), 37
Turtle graphics, 34
UCS-4, 99, 139, 140, 169
Unicode, 99, 139, 140, 169, 170,
Union (sets), 22
UNIX, 139
Unpacked string, 99, 136, 180
Untag override, 144
User error, 116
User-defined operators, 84, 142
forbidden ~, 90
UTF-8, 139–141, 157, 169
Van Orman Quine, Willard, 146
Variable arguments, 78
Variables, See Data declarations
state ~, 43
Virtual Machine, See Abstract ~
Warning, see also Diagnostic
Warnings, 162–168, 173
weekday, 73, 115
White space, 96
Whitesmith’s style, 4
Wide character, 140
Word count, 17
XML, 51, 172
XSLT, 51
Year zero, 11
Zeller, 73