Google Git
Sign in
fuchsia / third_party / swig / refs/heads/upstream/gsoc2008-jezabek / . / Doc / Manual / Chicken.html
blob: 98372a0f7e6541787877e8b66a600afbd4cdd326 [file] [log] [blame]
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<!-- Hand-written HTML -->
<html>
<head>
<title>SWIG and Chicken</title>
<link rel="stylesheet" type="text/css" href="style.css">
</head>
<body bgcolor="#ffffff">
<H1><a name="Chicken"></a>19 SWIG and Chicken</H1>
<!-- INDEX -->
<div class="sectiontoc">
<ul>
<li><a href="#Chicken_nn2">Preliminaries</a>
<ul>
<li><a href="#Chicken_nn3">Running SWIG in C mode</a>
<li><a href="#Chicken_nn4">Running SWIG in C++ mode</a>
</ul>
<li><a href="#Chicken_nn5">Code Generation</a>
<ul>
<li><a href="#Chicken_nn6">Naming Conventions</a>
<li><a href="#Chicken_nn7">Modules</a>
<li><a href="#Chicken_nn8">Constants and Variables</a>
<li><a href="#Chicken_nn9">Functions</a>
<li><a href="#Chicken_nn10">Exceptions</a>
</ul>
<li><a href="#Chicken_nn11">TinyCLOS</a>
<li><a href="#Chicken_nn12">Linkage</a>
<ul>
<li><a href="#Chicken_nn13">Static binary or shared library linked at compile time</a>
<li><a href="#Chicken_nn14">Building chicken extension libraries</a>
<li><a href="#Chicken_nn15">Linking multiple SWIG modules with TinyCLOS</a>
</ul>
<li><a href="#Chicken_nn16">Typemaps</a>
<li><a href="#Chicken_nn17">Pointers</a>
<ul>
<li><a href="#collection">Garbage collection</a>
</ul>
<li><a href="#Chicken_nn18">Unsupported features and known problems</a>
<ul>
<li><a href="#Chicken_nn19">TinyCLOS problems with Chicken version &lt;= 1.92</a>
</ul>
</ul>
</div>
<!-- INDEX -->
<p>
This chapter describes SWIG's support of CHICKEN. CHICKEN is a
Scheme-to-C compiler supporting most of the language features as
defined in the <i>Revised^5 Report on Scheme</i>. Its main
attributes are that it
</p>
<ol>
<li>generates portable C code</li>
<li>includes a customizable interpreter</li>
<li>links to C libraries with a simple Foreign Function Interface</li>
<li>supports full tail-recursion and first-class continuations</li>
</ol>
<p>
When confronted with a large C library, CHICKEN users can use
SWIG to generate CHICKEN wrappers for the C library. However,
the real advantages of using SWIG with CHICKEN are its
<strong>support for C++</strong> -- object-oriented code is
difficult to wrap by hand in CHICKEN -- and its <strong>typed
pointer representation</strong>, essential for C and C++
libraries involving structures or classes.
</p>
<H2><a name="Chicken_nn2"></a>19.1 Preliminaries</H2>
<p>
CHICKEN support was introduced to SWIG in version 1.3.18. SWIG
relies on some recent additions to CHICKEN, which are only
present in releases of CHICKEN with version number
<strong>greater than or equal to 1.89</strong>.
To use a chicken version between 1.40 and 1.89, see the <a href="#collection">Garbage collection</a>
section below.
</p>
<p>
You may want to look at any of the examples in Examples/chicken/
or Examples/GIFPlot/Chicken for the basic steps to run SWIG
CHICKEN.
</p>
<H3><a name="Chicken_nn3"></a>19.1.1 Running SWIG in C mode</H3>
<p>
To run SWIG CHICKEN in C mode, use
the -chicken option.
</p>
<div class="shell">
<pre>% swig -chicken example.i</pre>
</div>
<p>
To allow the wrapper to take advantage of future CHICKEN code
generation improvements, part of the wrapper is direct CHICKEN
function calls (<tt>example_wrap.c</tt>) and part is CHICKEN
Scheme (<tt>example.scm</tt>). The basic Scheme code must
be compiled to C using your system's CHICKEN compiler or
both files can be compiled directly using the much simpler <tt>csc</tt>.
</p>
<div class="shell">
<pre>
% chicken example.scm -output-file oexample.c
</pre>
</div>
<p>
So for the C mode of SWIG CHICKEN, <tt>example_wrap.c</tt> and
<tt>oexample.c</tt> are the files that must be compiled to
object files and linked into your project.
</p>
<H3><a name="Chicken_nn4"></a>19.1.2 Running SWIG in C++ mode</H3>
<p>
To run SWIG CHICKEN in C++ mode, use
the -chicken -c++ option.
</p>
<div class="shell">
<pre>% swig -chicken -c++ example.i</pre>
</div>
<p>
This will generate <tt>example_wrap.cxx</tt> and
<tt>example.scm</tt>. The basic Scheme code must be
compiled to C using your system's CHICKEN compiler or
both files can be compiled directly using the much simpler <tt>csc</tt>.
</p>
<div class="shell">
<pre>% chicken example.scm -output-file oexample.c</pre>
</div>
<p>
So for the C++ mode of SWIG CHICKEN, <tt>example_wrap.cxx</tt>
and <tt>oexample.c</tt> are the files that must be compiled to
object files and linked into your project.
</p>
<H2><a name="Chicken_nn5"></a>19.2 Code Generation</H2>
<H3><a name="Chicken_nn6"></a>19.2.1 Naming Conventions</H3>
<p>
Given a C variable, function or constant declaration named
<tt>Foo_Bar</tt>, the declaration will be available
in CHICKEN as an identifier ending with
<tt>Foo-Bar</tt>. That is, an underscore is converted
to a dash.
</p>
<p>
You may control what the CHICKEN identifier will be by using the
<tt>%rename</tt> SWIG directive in the SWIG interface file.
</p>
<H3><a name="Chicken_nn7"></a>19.2.2 Modules</H3>
<p>
The name of the module must be declared one of two ways:
<ul>
<li>Placing <tt>%module example</tt> in the SWIG interface
file.</li>
<li>Using <tt>-module example</tt> on the SWIG command
line.</li>
</ul>
<p>
The generated example.scm file then exports <code>(declare (unit modulename))</code>.
If you do not want SWIG to export the <code>(declare (unit modulename))</code>, pass
the -nounit option to SWIG.
<p>
CHICKEN will be able to access the module using the <code>(declare
(uses <i>modulename</i>))</code> CHICKEN Scheme form.
</p>
<H3><a name="Chicken_nn8"></a>19.2.3 Constants and Variables</H3>
<p>
Constants may be created using any of the four constructs in
the interface file:
</p>
<ol>
<li><code>#define MYCONSTANT1 ...</code></li>
<li><code>%constant int MYCONSTANT2 = ...</code></li>
<li><code>const int MYCONSTANT3 = ...</code></li>
<li><code>enum { MYCONSTANT4 = ... };</code></li>
</ol>
<p>
In all cases, the constants may be accessed from within CHICKEN
using the form <tt>(MYCONSTANT1)</tt>; that is, the constants
may be accessed using the read-only parameter form.
</p>
<p>
Variables are accessed using the full parameter form.
For example, to set the C variable "int my_variable;", use the
Scheme form <tt>(my-variable 2345)</tt>. To get the C variable,
use <tt>(my-variable)</tt>.
</p>
<p>
The <tt>%feature("constasvar")</tt> can be applied to any constant
or immutable variable. Instead of exporting the constant as
a function that must be called, the constant will appear as a
scheme variable. This causes the generated .scm file to just contain the code
<tt>(set! MYCONSTANT1 (MYCONSTANT1))</tt>. See
<a href="Customization.html#features">Features and the %feature directive</a>
for info on how to apply the %feature.
</p>
<H3><a name="Chicken_nn9"></a>19.2.4 Functions</H3>
<p>
C functions declared in the SWIG interface file will have
corresponding CHICKEN Scheme procedures. For example, the C
function "int sqrt(double x);" will be available using the
Scheme form <tt>(sqrt 2345.0)</tt>. A <code>void</code> return
value will give C_SCHEME_UNDEFINED as a result.
</p>
<p>
A function may return more than one value by using the
<code>OUTPUT</code> specifier (see Lib/chicken/typemaps.i).
They will be returned as multiple values using <code>(values)</code> if there is more than one
result (that is, a non-void return value and at least one argout
parameter, or a void return value and at least two argout
parameters). The return values can then be accessed with <code>(call-with-values)</code>.
</p>
<H3><a name="Chicken_nn10"></a>19.2.5 Exceptions</H3>
<p>The SWIG chicken module has support for exceptions thrown from
C or C++ code to be caught in scheme.
See <a href="Customization.html#exception">Exception handling with %exception</a>
for more information about declaring exceptions in the interface file.
</p>
<p>Chicken supports both the <code>SWIG_exception(int code, const char *msg)</code> interface
as well as a <code>SWIG_ThrowException(C_word val)</code> function for throwing exceptions from
inside the %exception blocks. <code>SWIG_exception</code> will throw a list consisting of the code
(as an integer) and the message. Both of these will throw an exception using <code>(abort)</code>,
which can be handled by <code>(handle-exceptions)</code>. See
<a href="http://www.call-with-current-continuation.org/manual/Exceptions.html#Exceptions">Chicken manual on Exceptions</a>
and <a href="http://srfi.schemers.org/srfi-12/srfi-12.html">SFRI-12</a>. Since the exception values are thrown
directly, if <code>(condition-case)</code> is used to catch an exception the exception will come through in the <code>val ()</code> case.
</p>
<p>The following simple module</p>
<div class="code"><pre>
%module exception_test
%inline %{
void test_throw(int i) throws (int) {
if (i == 1) throw 15;
}
%}
</pre></div>
<p>could be run with</p>
<div class="targetlang"><pre>
(handle-exceptions exvar
(if (= exvar 15)
(print "Correct!")
(print "Threw something else " exvar))
(test-throw 1))
</pre></div>
<H2><a name="Chicken_nn11"></a>19.3 TinyCLOS</H2>
<p>
The author of TinyCLOS, Gregor Kiczales, describes TinyCLOS as:
"Tiny CLOS is a Scheme implementation of a `kernelized' CLOS, with a
metaobject protocol. The implementation is even simpler than
the simple CLOS found in `The Art of the Metaobject Protocol,'
weighing in at around 850 lines of code, including (some)
comments and documentation."
</p>
<p>
Almost all good Scheme books describe how to use metaobjects and
generic procedures to implement an object-oriented Scheme
system. Please consult a Scheme book if you are unfamiliar
with the concept.
</p>
<p>
CHICKEN has a modified version of TinyCLOS, which SWIG CHICKEN
uses if the -proxy argument is given. If -proxy is passed, then
the generated example.scm file will contain TinyCLOS class definitions.
A class named Foo is declared as &lt;Foo&gt;, and each member variable
is allocated a slot. Member functions are exported as generic functions.
<p>
Primitive symbols and functions (the interface that would be presented if
-proxy was not passed) are hidden and no longer accessible. If the -unhideprimitive
command line argument is passed to SWIG, then the primitive symbols will be
available, but each will be prefixed by the string "primitive:"
<p>
The exported symbol names can be controlled with the -closprefix and -useclassprefix arguments.
If -useclassprefix is passed to SWIG, every member function will be generated with the class name
as a prefix. If the -closprefix mymod: argument is passed to SWIG, then the exported functions will
be prefixed by the string "mymod:". If -useclassprefix is passed, -closprefix is ignored.
</p>
<H2><a name="Chicken_nn12"></a>19.4 Linkage</H2>
<p>
Please refer to <em>CHICKEN - A practical and portable Scheme
system - User's manual</em> for detailed help on how to link
object files to create a CHICKEN Scheme program. Briefly, to
link object files, be sure to add <tt>`chicken-config
-extra-libs -libs`</tt> or <tt>`chicken-config -shared
-extra-libs -libs`</tt>to your linker options. Use the
<tt>-shared</tt> option if you want to create a dynamically
loadable module. You might also want to use the much simpler
<tt>csc</tt> or <tt>csc.bat</tt>.
</p>
<p>Each scheme file that is generated
by SWIG contains <code>(declare (uses <i>modname</i>))</code>. This means that to load the
module from scheme code, the code must include <code>(declare (uses <i>modname</i>))</code>.
</p>
<H3><a name="Chicken_nn13"></a>19.4.1 Static binary or shared library linked at compile time</H3>
<p>We can easily use csc to build a static binary.</p>
<div class="shell">
<pre>
$ swig -chicken example.i
$ csc -v example.scm example_impl.c example_wrap.c test_script.scm -o example
$ ./example
</pre>
</div>
<p>Similar to the above, any number of <tt>module.scm</tt> files could be compiled
into a shared library, and then that shared library linked when compiling the
main application.</p>
<div class="shell">
<pre>
$ swig -chicken example.i
$ csc -sv example.scm example_wrap.c example_impl.c -o example.so
</pre>
</div>
<p>The <tt>example.so</tt> file can then linked with <tt>test_script.scm</tt> when it
is compiled, in which case <tt>test_script.scm</tt> must have <code>(declare (uses example))</code>.
Multiple SWIG modules could have been linked into <tt>example.so</tt> and each
one accessed with a <code>(declare (uses ... ))</code>.
</p>
<div class="shell">
<pre>
$ csc -v test_script.scm -lexample
</pre>
</div>
<p>An alternative is that the test_script.scm can have the code <code>(load-library 'example "example.so")</code>,
in which case the test script does not need to be linked with example.so. The test_script.scm file can then
be run with <tt>csi</tt>.
</p>
<H3><a name="Chicken_nn14"></a>19.4.2 Building chicken extension libraries</H3>
<p>Building a shared library like in the above section only works if the library
is linked at compile time with a script containing <code>(declare (uses ...))</code> or is
loaded explicitly with <code>(load-library 'example "example.so")</code>. It is
not the format that CHICKEN expects for extension libraries and eggs. The problem is the
<code>(declare (unit <i>modname</i>))</code> inside the <tt>modname.scm</tt> file. There are
two possible solutions to this.</p>
<p>First, SWIG accepts a <tt>-nounit</tt> argument, in which case the <code>(declare (unit <i>modname</i>))</code>
is not generated. Then, the <tt>modname.scm</tt> and <tt>modname_wrap.c</tt> files <b>must</b> be compiled into
their own shared library.</p>
<div class="shell">
<pre>
$ csc -sv modname.scm modname_wrap.c modname_impl.c -o modname.so
</pre>
</div>
<p>This library can then be loaded by scheme code with the <code>(require 'modname)</code> function.
See <a href="http://www.call-with-current-continuation.org/manual/Loading-extension-libraries.html">
Loading-extension-libraries</a> in the eval unit inside the CHICKEN manual for more information.</p>
<p>Another alternative is to run SWIG normally and create a scheme file that contains <code>(declare (uses <i>modname</i>))</code>
and then compile that file into the shared library as well. For example, inside the <tt>mod_load.scm</tt> file,</p>
<div class="targetlang">
<pre>
(declare (uses mod1))
(declare (uses mod2))
</pre>
</div>
<p>Which would then be compiled with</p>
<div class="shell">
<pre>
$ swig -chicken mod1.i
$ swig -chicken mod2.i
$ csc -sv mod_load.scm mod1.scm mod2.scm mod1_wrap.c mod2_wrap.c mod1_impl.c mod2_impl.c -o mod.so
</pre>
</div>
<p>Then the extension library can be loaded with <code>(require 'mod)</code>. As we can see here,
<tt>mod_load.scm</tt> contains the code that gets executed when the module is loaded. All this code
does is load both mod1 and mod2. As we can see, this technique is more useful when you want to
combine a few SWIG modules into one chicken extension library, especially if modules are related by
<code>%import</code></p>
<p>In either method, the files that are compiled into the shared library could also be
packaged into an egg. The <tt>mod1_wrap.c</tt> and <tt>mod2_wrap.c</tt> files that are created by SWIG
are stand alone and do not need SWIG to be installed to be compiled. Thus the egg could be
distributed and used by anyone, even if SWIG is not installed.</p>
<p>See the <tt>Examples/chicken/egg</tt> directory in the SWIG source for an example that builds
two eggs, one using the first method and one using the second method.</p>
<H3><a name="Chicken_nn15"></a>19.4.3 Linking multiple SWIG modules with TinyCLOS</H3>
<p>Linking together multiple modules that share type information using the <code>%import</code>
directive while also using <tt>-proxy</tt> is more complicated. For example, if <tt>mod2.i</tt> imports <tt>mod1.i</tt>, then the
<tt>mod2.scm</tt> file contains references to symbols declared in <tt>mod1.scm</tt>,
and thus a <code>(declare (uses <i>mod1</i>))</code> or <code>(require '<i>mod1</i>)</code> must be exported
to the top of <tt>mod2.scm</tt>. By default, when SWIG encounters an <code>%import "modname.i"</code> directive,
it exports <code>(declare (uses <i>modname</i>))</code> into the scm file. This works fine unless mod1 was compiled with
the <tt>-nounit</tt> argument or was compiled into an extension library with other modules under a different name.</p>
<p>One option is to override the automatic generation of <code>(declare (uses mod1))</code>
by passing the <tt>-noclosuses</tt> option to SWIG when compiling <tt>mod2.i</tt>.
SWIG then provides the <code>%insert(closprefix) %{ %}</code> directive. Any scheme code inside that directive is inserted into the
generated .scm file, and if <tt>mod1</tt> was compiled with <tt>-nounit</tt>, the directive should contain <code>(require 'mod1)</code>.
This option allows for mixed loading as well, where some modules are imported with <code>(declare (uses <i>modname</i>))</code>
(which means they were compiled without -nounit) and some are imported with <code>(require 'modname)</code>.</p>
<p>The other option is to use the second idea in the above section. Compile all the modules normally, without any
<code>%insert(closprefix)</code>, <tt>-nounit</tt>, or <tt>-noclosuses</tt>. Then the modules will import each other correctly
with <code>(declare (uses ...))</code>.
To create an extension library or an egg, just create a <tt>module_load.scm</tt> file that <code>(declare (uses ...))</code>
all the modules.</p>
<H2><a name="Chicken_nn16"></a>19.5 Typemaps</H2>
<p>
The Chicken module handles all types via typemaps. This information is
read from <code>Lib/chicken/typemaps.i</code> and
<code>Lib/chicken/chicken.swg</code>.
</p>
<H2><a name="Chicken_nn17"></a>19.6 Pointers</H2>
<p>
For pointer types, SWIG uses CHICKEN tagged pointers.
A tagged pointer is an ordinary CHICKEN pointer with an
extra slot for a void *. With SWIG
CHICKEN, this void * is a pointer to a type-info
structure. So each pointer used as input or output from
the SWIG-generated CHICKEN wrappers will have type
information attached to it. This will let the wrappers
correctly determine which method should be called
according to the object type hierarchy exposed in the SWIG
interface files.
</p>
<p>
To construct a Scheme object from a C pointer, the wrapper code
calls the function
<code>SWIG_NewPointerObj(void *ptr, swig_type_info *type, int owner)</code>,
The function that calls <code>SWIG_NewPointerObj</code> must have a variable declared
<code>C_word *known_space = C_alloc(C_SIZEOF_SWIG_POINTER);</code>
It is ok to call <code>SWIG_NewPointerObj</code> more than once,
just make sure known_space has enough space for all the created pointers.
</p>
<p>
To get the pointer represented by a CHICKEN tagged pointer, the
wrapper code calls the function
<code>SWIG_ConvertPtr(C_word s, void **result, swig_type_info *type, int flags)</code>,
passing a pointer to a struct representing the expected pointer
type. flags is either zero or SWIG_POINTER_DISOWN (see below).
</p>
<H3><a name="collection"></a>19.6.1 Garbage collection</H3>
<p>If the owner flag passed to <code>SWIG_NewPointerObj</code> is 1, <code>NewPointerObj</code> will add a
finalizer to the type which will call the destructor or delete method of
that type. The destructor and delete functions are no longer exported for
use in scheme code, instead SWIG and chicken manage pointers.
In situations where SWIG knows that a function is returning a type that should
be garbage collected, SWIG will automatically set the owner flag to 1. For other functions,
the <code>%newobject</code> directive must be specified for functions whose return values
should be garbage collected. See
<a href="Customization.html#ownership">Object ownership and %newobject</a> for more information.
</p>
<p>In situations where a C or C++ function will assume ownership of a pointer, and thus
chicken should no longer garbage collect it, SWIG provides the <code>DISOWN</code> input typemap.
After applying this typemap (see the <a href="Typemaps.html">Typemaps chapter</a> for more information on how to apply typemaps),
any pointer that gets passed in will no longer be garbage collected.
An object is disowned by passing the <code>SWIG_POINTER_DISOWN</code> flag to <code>SWIG_ConvertPtr</code>.
<b>Warning:</b> Since the lifetime of the object is now controlled by the underlying code, the object might
get deleted while the scheme code still holds a pointer to it. Further use of this pointer
can lead to a crash.
</p>
<p>Adding a finalizer function from C code was added to chicken in the 1.89 release, so garbage collection
does not work for chicken versions below 1.89. If you would like the SWIG generated code to work with
chicken 1.40 to 1.89, pass the <code>-nocollection</code> argument to SWIG. This will not export code
inside the _wrap.c file to register finalizers, and will then export destructor functions which
must be called manually.
</p>
<H2><a name="Chicken_nn18"></a>19.7 Unsupported features and known problems</H2>
<ul>
<li>No director support.</li>
<li>No support for c++ standard types like std::vector.</li>
<li>The TinyCLOS wrappers for overloaded functions will not work correctly when using
<a href="SWIGPlus.html#SWIGPlus_default_args">%feature(compactdefaultargs)</a>.</li>
</ul>
<H3><a name="Chicken_nn19"></a>19.7.1 TinyCLOS problems with Chicken version &lt;= 1.92</H3>
<p>In Chicken versions equal to or below 1.92, TinyCLOS has a limitation such that generic methods do not properly work on methods
with different number of specializers: TinyCLOS assumes that every method added to a generic function
will have the same number of specializers. SWIG generates functions with different lengths of specializers
when C/C++ functions are overloaded. For example, the code</p>
<div class="code">
<pre>
class Foo {};
int foo(int a, Foo *b);
int foo(int a);
</pre></div>
<p>will produce scheme code</p>
<div class="targetlang">
<pre>
(define-method (foo (arg0 &lt;top&gt;) (arg1 &lt;Foo&gt;)) (<i>call primitive function</i>))
(define-method (foo (arg0 &lt;top&gt;)) (<i>call primitive function</i>))
</pre></div>
<p>Using unpatched TinyCLOS, the second <code>(define-method)</code> will replace the first one,
so calling <code>(foo 3 f)</code> will produce an error.</p>
<p>There are three solutions to this. The easist is to upgrade to the latest Chicken version. Otherwise, the
file <tt>Lib/chicken/tinyclos-multi-generic.patch</tt> in the SWIG source contains a patch against
tinyclos.scm inside the 1.92 chicken source to add support into TinyCLOS for multi-argument generics. (This patch was accepted into Chicken)
This requires chicken to be rebuilt and custom install of chicken. An alternative is the <tt>Lib/chicken/multi-generic.scm</tt>
file in the SWIG source. This file can be loaded after TinyCLOS is loaded, and it will override some functions
inside TinyCLOS to correctly support multi-argument generics. Please see the comments at the top of both files for more information.</p>
</body>
</html>