<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<link href="../../../style.css" rel="stylesheet" type="text/css"/>
<title>Gluegen Manual</title>
</head>
<body>
<div id="container">
<div id="header">
<div id="slogan">Gluegen Manual</div>
<div id="logo"><a href="http://jogamp.org/">Gluegen</a></div>
</div>
<div id="menu">
<ul>
<li><a href="http://jogamp.org/">Home</a></li>
<li><a href="../../../gluegen/www/">Gluegen</a></li>
<li><a href="../../../joal/www/">JOAL</a></li>
<li><a href="../../../jocl/www/">JOCL</a></li>
<li><a href="../../../jogl/www/">JOGL</a></li>
<li><a href="../../../demos/">Demos</a></li>
<li><a href="../../../wiki/">Wiki</a></li>
<li><a href="../../../deployment/gluegen-next/javadoc/">JavaDoc</a></li>
</ul>
</div>
<div id="main">
<div id="text" class="fill">
<h1>GlueGen Manual</h1>
<h2> Table of Contents </h2>
<a href="#Chapter1">Chapter 1</a> - Introduction
<ul>
<li> <a href = "#SecIntroduction">Introduction</a></li>
<li> <a href = "#SecStructure">Structure of the Generated Glue Code</a></li>
<li> <a href = "#SecUnique">Unique Features</a></li>
<li> <a href = "#SecBackground">Background and Design Principles</a></li>
</ul>
<a href="#Chapter2">Chapter 2</a> - Using GlueGen
<ul>
<li> <a href = "#SecAcquiring">Acquiring and Building GlueGen</a>
<ul>
<li> <a href = "#SecCommon">Common Build Problems</a></li>
</ul>
</li>
<li> <a href = "#SecBasic">Basic Operation</a></li>
<li> <a href = "#SecAnt">Running GlueGen as an Ant Task</a></li>
<li> <a href = "#SecPCPP">PCPP</a></li>
<li> <a href = "#SecError">Error Reporting</a></li>
<li> <a href = "#SecStub">Stub Headers</a></li>
<li> <a href = "#Sec32">32- and 64-bit Considerations</a></li>
<li> <a href = "#SecOpaque">Opaque Directives</a></li>
<li> <a href = "#SecSubstitution">Argument Name Substitution</a></li>
<li> <a href = "#SecConfiguration">Configuration File Directives</a>
<ul>
<li> <a href = "#SecJavaEmitter">JavaEmitter Configuration</a></li>
<li> <a href = "#SecProcAddressEmitter">ProcAddressEmitter Configuration</a></li>
</ul>
</li>
</ul>
<a href="#Chapter3">Chapter 3</a> - Configuration File Examples<br/>
<ul>
<li> <a href = "#SecSimplest">Simplest possible example</a></li>
<li> <a href = "#SecArrays">Arrays and buffers</a></li>
<li> <a href = "#SecString">String handling</a></li>
<li> <a href = "#SecMemory">Memory allocation</a></li>
<li> <a href = "#SecStructs">Ingoing and outgoing structs</a></li>
<li> <a href = "#SecStructArrays">Returned arrays of structs</a></li>
<li> <a href = "#SecPointerArrays">Returned arrays of pointers</a></li>
</ul>
<h2> <a name="Chapter1">Chapter 1 - Introduction</a> </h2>
<h3> <a name="SecIntroduction">Introduction</a> </h3>
<p>
GlueGen is a tool which automatically generates the Java and JNI code
necessary to call C libraries. It reads as input ANSI C header files
and separate configuration files which provide control over many
aspects of the glue code generation. GlueGen uses a complete ANSI C
parser and an internal representation (IR) capable of representing all
C types to represent the APIs for which it generates interfaces. It
has the ability to perform significant transformations on the IR
before glue code emission. GlueGen is currently powerful enough to
bind even low-level APIs such as the Java Native Interface (JNI) and
the AWT Native Interface (JAWT) back up to the Java programming
language.
</p>
<p>
GlueGen is currently used to generate the JOGL interface to the OpenGL
3D graphics API and the JOAL interface to the OpenAL audio library. In
the case of JOGL, GlueGen is used not only to bind OpenGL to Java, but
also the low-level windowing system APIs on the Windows, X11 and Mac
OS X platforms. The implementation of the JOGL library is thereby
written in the Java programming language rather than in C, which has
offered considerable advantages during the development of the library.
</p>
<p>
GlueGen is designed in modular form and can be extended to alter the
glue code emission style or to generate interface code for other
languages than Java.
</p>
<p>
This manual describes how to use GlueGen to bind new C libraries to
the Java programming language.
</p>
<h3> <a name="SecStructure">Structure of the Generated Glue Code</a> </h3>
<p>
GlueGen supports two basic styles of glue code generation: everything
in one class, or a separate interface and implementing class. The
first mode, "AllStatic", exposes the underlying C functions as a set
of static Java methods in a concrete class. This is a straightforward
binding mechanism, but has the disadvantage of tying users to a
concrete class (which may or may not be a problem) and makes it more
difficult to support certain kinds of call-through-function-pointer
semantics required by certain C APIs. The second mode,
"InterfaceAndImpl", exposes the C functions as methods in an interface
and emits the implementation of that interface into a separate class
and package. The implementing class is not intended to be in the
public API; this more strongly separates the user from the
implementation of the API. Additionally, because it is necessary to
hold an instance of the implementing class in order to access the
underlying C routines, it is easier to support situations where
call-through-function-pointer semantics must be followed, in
particular where those function pointers might change from instance to
instance.
</p>
The generated glue code follows some basic rules in binding C APIs to
Java:
<ul>
<li> C primitive types are exposed as the corresponding Java primitive
type.
</li>
<li> Pointers to typed C primitives (<code>int*</code>,
<code>float*</code>) are bound to java.nio Buffer subclasses
(<code>IntBuffer</code>, <code>FloatBuffer</code>) and optionally
to Java arrays (<code>int[]</code>, <code>float[]</code>).
<ul>
<li> If a C function takes such a pointer as an outgoing argument,
two method overloadings will generally be produced; one which
accepts a Buffer, and one which accepts a primitive array plus
an integer offset argument. The variant taking a Buffer may
accept either a "direct" NIO Buffer or a non-direct one
(wrapping a Java array). The exception is when such a routine
is specified by the <a href="#NioDirectOnly">NioDirectOnly</a>
directive to keep a persistent pointer to the passed storage,
in which case only the Buffer variant will be generated, and
will only accept a direct Buffer as argument.
</li>
<li> If a C function returns such a pointer as its result, it will
be exposed as the corresponding Buffer type. In this case it is
also typically necessary to specify to GlueGen via the <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> directive
the number of addressable elements in the resulting array.
</li>
</ul>
</li>
<li> Pointers to <code>void*</code> are bound to java.nio.Buffer.
<ul>
<li> By default any C function accepting a <code>void*</code>
argument will allow either a direct or non-direct java.nio
Buffer to be passed as argument. If the <a
href="#NioDirectOnly">NioDirectOnly</a> directive is specified,
however, only a direct Buffer will be accepted.
</li>
<li> Similar rules for <code>void*</code> return values apply to
those for pointers to typed primitives.
</li>
</ul>
</li>
<li> To avoid an explosion in the number of generated methods, if a
particular API accepts more than one typed primitive pointer
argument, only two overloadings continue to be produced: one
accepting all arrays as arguments and one accepting all Buffers
as arguments. When calling the variant accepting Buffers, all of
the Buffers passed in a particular call must be either direct or
non-direct. Mixing of direct and non-direct Buffers in a given
function call is not supported.
</li>
<li> When a java.nio Buffer is passed from Java to C, the position of
the Buffer is taken into account. The resulting pointer passed to
C is equal to the base address of the Buffer plus the position
scaled appropriately for the size of the primitive elements in
the Buffer. This feature is called "auto-slicing", as it mimics
the behavior of calling Buffer.slice() without the overhead of
explicit object creation.
</li>
<li> Pointers to constant <code>char*</code> may be bound to
java.lang.String using the <a
href="#ArgumentIsString">ArgumentIsString</a> or <a
href="#ReturnsString">ReturnsString</a> directives.
</li>
<li> <code>#define</code> statements in header files mapping names to
constant values are exposed as public static final constant
values in either the generated interface or AllStatic class.
</li>
<li> C structs encountered during the glue code generation process and
referenced by the C functions are exposed as Java classes of the
same name (typically the name to which the struct is typedefed).
Each primitive field in the struct is exposed as two methods; a
getter, which accepts no arguments, and a setter, which accepts
as argument a primitive value of the type of the field. Static
factory methods are exposed allowing allocation of these structs
from Java code. The backing storage for these Java classes is a
direct java.nio Buffer. GlueGen fully supports returning of
pointers to C structs up to Java.
</li>
</ul>
<h3><a name="SecUnique">Unique Features</a></h3>
GlueGen contains several unique features making it both a powerful and
easy-to-use tool.
<ul>
<li> C structs are exposed as Java classes. The generated code for
these classes supports both 32-bit and 64-bit platforms.
</li>
<li> C structs containing function pointers are exposed as Java
classes with methods. This makes it easy to interact with
low-level C APIs such as the AWT Native Interface (JAWT) from the
Java programming language level.
<ul>
<li> In this context, GlueGen automatically detects which argument
to the various function pointers indicates the "this" pointer,
hiding it at the Java level and passing it automatically.
</li>
<li> GlueGen offers automatic handling of JNI-specific data types
such as <code>JNIEnv*</code> and <code>jobject</code>. The tool
understands that the <code>JNIEnv*</code> argument is implicit
and that <code>jobject</code> maps to java.lang.Object at the
Java programming language level. While this is most useful when
binding JDK-internal APIs such as the JAWT to Java, there may
be other JNI libraries which expose C functions taking these
data types, and GlueGen can very easily bind to them.
</li>
</ul>
</li>
</ul>
<h3><a name="SecBackground">Background and Design Principles</a></h3>
<p>
This section provides motivation for the design of the GlueGen tool
and is not necessary to understand how to use the tool.
</p>
<p>
There are many tools available for assisting in the autogeneration of
foreign function interfaces for various high-level languages. Only a
few examples include <a
href="http://alumni.media.mit.edu/~kbrussel/Header2Scheme/">Header2Scheme</a>,
an early tool allowing binding of a limited subset of C++ to the
Scheme programming language; <a href="http://www.swig.org/">SWIG</a>,
a tool released at roughly the same time as Header2Scheme which by now
supports binding C and C++ libraries to a variety of scripting
languages; <a href="http://www.jniwrapper.com/">JNIWrapper</a>, a
commercial tool automating the binding of C APIs to Java; and <a
href="http://www.noodleglue.org/noodleglue/noodleglue.html">NoodleGlue</a>,
a recently-released tool automating the binding of C++ APIs to
Java. Other language-specific tools such as Perl's XS, Boost.Python
and many others exist.
</p>
<p>
GlueGen was designed with a few key principles in mind. The most
fundamental was to support binding of the lowest-level APIs on a given
platform up to the Java programming language. The intended goal, in
the context of the JOGL project, was to allow subsets of the Win32 and
X11 APIs to be exposed to Java, and to use those APIs to write the
behind-the-scenes OpenGL context creation and management code in Java
instead of C. This informed several other design goals:
</p>
<ul>
<li> Avoid touching the C headers as much as possible. This makes it
easier to upgrade to a more recent version of the C API just by
copying in a new set of headers.
</li>
<li> Avoid touching the generated glue code completely.</li>
<li> Avoid having to hand-write a lot of generated glue code. Instead,
handle many complex constructs automatically and provide
sufficient control over the glue code generation to avoid having
to handwrite certain native methods where one or two lines of
tweaking would suffice.
</li>
<li> Support all C constructs in the parser and intermediate
representation. The rationale is that it is acceptable to cut
corners in the number of constructs supported in the Java
binding, but not whether the tool can internally represent it in
its C type system. This design goal implies starting with
complete a ANSI C parser coupled with a complete C type system.
</li>
<li> As the tool is targetting the Java programming language, build
the tool in the Java programming language.
</li>
</ul>
In order to make the problem more tractable, support for binding C++
to the Java programming language was not considered. C++ adds many
constructs over ANSI C which make it much more difficult to reason
about and to find a useful subset to support binding to Java.
Additionally, it seems that there are relatively few C++-specific
libraries in general use which could be usefully bound to Java,
although this may be a matter of opinion.
<p>
GlueGen was designed with the Java programming language in mind, but
is not necessarily restricted to generating glue code for the Java
language. The tool is divided into separate parse and code generation
phases, and the internal representation is fairly easy to iterate
over. The core driver of GlueGen may therefore be useful in producing
other tools which autogenerate foreign function interfaces to C
libraries for other languages.
</p>
<h2> <a name="Chapter2">Chapter 2 - Using GlueGen</a> </h2>
<h3><a name="SecAcquiring">Acquiring and Building GlueGen</a></h3>
The source code for GlueGen may be obtained by cloning the Git repository:
<pre>
$git clone git://github.com/mbien/gluegen.git gluegen
</pre>
To build GlueGen, cd into the gluegen/make folder and invoke ant.
<pre>
$ant clean all test
</pre>
Ant 1.8 or later and a Java 6 compatible JDK is required.
<h4><a name="SecCommon">Common Build Problems</a></h4>
<dl>
<dt> <strong>CharScanner; panic: ClassNotFoundException: com.sun.gluegen.cgram.CToken</strong></dt>
<dd> This occurs because ANTLR was dropped into the Extensions
directory of the JRE/JDK. On Windows and Linux, delete any ANTLR
jars from jre/lib/ext, and on Mac OS X, delete them from
/Library/Java/Extensions. Use the antlr.jar property in the
build.xml to point to a JRE-external location of this jar file.
</dd>
</dl>
<h3><a name="SecBasic">Basic Operation</a></h3>
<p>
GlueGen can be run either as an executable jar file (<code>java -jar
gluegen.jar</code>; note that antlr.jar must be in the same directory
as gluegen.jar in order for this invocation to work) or from within
Ant as described in the following section. When run from the command
line, GlueGen accepts four kinds of command-line arguments:
</p>
<ul>
<li> -I<em>dir</em> (optional) adds <em>dir</em> to the include path.
Similarly to a C compiler or preprocessor, GlueGen scans a set of
directories to locate header files it encounters in
<code>#include</code> directives. Unlike most C preprocessors,
however, GlueGen has no default include path, so it is typically
necessary to supply at least one <code>-I</code> option on the
command line in order to handle any <code>#include</code>
directives in the file being parsed.
</li>
<li> -E<em>emitterClassName</em> (optional) uses
<em>emitterClassName</em> as the fully-qualified name of the
emitter class which will be used by GlueGen to generate the glue
code. The emitter class must implement the
<code>com.sun.gluegen.GlueEmitter</code> interface. If this
option is not specified, a
<code>com.sun.gluegen.JavaEmitter</code> will be used by default.
</li>
<li> -C<em>cfgFile</em> adds <em>cfgFile</em> to the list of
configuration files used to set up the chosen emitter. This is
the means by which a large number of options are passed in to the
GlueGen tool and to the emitter in particular. Configuration
files are discussed more in the following section.
</li>
<li> [ filename | - ] selects the file or standard input from which
GlueGen should read the C header file for which glue code should
be generated. This must be the last command-line argument, and
only one filename argument is supported. To cause multiple header
files to be parsed, write a small .c file #including the multiple
headers and point GlueGen at the .c file.</li>
</ul>
<h3><a name="SecAnt">Running GlueGen as an Ant Task</a></h3>
<p>
GlueGen can also be invoked as a subtask within Ant. In order to do
so, a path element should be defined as follows:
</p>
<pre>
<path id="gluegen.classpath">
<pathelement location="${gluegen.jar}" />
<pathelement location="${antlr.jar}" />
</path>
</pre>
where the <code>gluegen.jar</code> and <code>antlr.jar</code>
properties point to the respective jar files. A taskdef defining the
GlueGen task should then be specified as follows:
<pre>
<taskdef name="gluegen"
classname="com.sun.gluegen.ant.GlueGenTask"
classpathref="gluegen.classpath" />
</pre>
At this point GlueGen may be invoked as follows:
<pre>
<gluegen src="[header to parse]"
config="[configuration file]"
includeRefid="[dirset for include path]"
emitter="com.sun.gluegen.JavaEmitter">
<classpath refid="gluegen.classpath" />
</gluegen>
</pre>
Please see the <a href="http://jogl.dev.java.net/">JOGL</a> and <a
href="http://joal.dev.java.net/">JOAL</a> build.xml files for
concrete, though non-trivial, examples of how to invoke GlueGen via
Ant.
<h3><a name="SecPCPP">PCPP</a></h3>
<p>
GlueGen contains and uses a minimal C preprocessor called the "Pseudo
C Pre-Processor", or PCPP. A slightly specialized C preprocessor is
required for correct glue code generation with most libraries.
Constant values intended for use by end users are defined in many C
libraries' headers using <code>#define</code>s rather than constant
int declarations, and if the header is processed by a full C
preprocessor then the #define statements will be stripped become
unavailable for processing by the glue code generator.
</p>
<p>
PCPP is largely an invisible part of the glue code generation process;
however, it has certain limitations which make it difficult to parse
certain header files. First, it does not support macro evaluation in
any form, so if a header relies on macro evaluation in order to
generate code, PCPP will fail. It is possible that PCPP may fail
silently in this situation, causing GlueGen to simply not produce code
for the associated constructs. If GlueGen's output is not as expected
and there is heavy use of the C preprocessor in the header, run PCPP
against the header directly (PCPP takes simply the -I and filename
arguments accepted by GlueGen) and examine the output.
</p>
<p>
Second, PCPP contains only limited support for <code>#if</code>
clauses. Generally speaking, its handling of <code>#if defined(foo) ||
defined(bar)</code> constructs is limited to approximately what is
required to handle the OpenGL header files. If the header being parsed
relies on moderately complicated expressions being evaluated by the C
preprocessor, check the output from PCPP and ensure it is as expected.
</p>
<p>
Contributions to PCPP would be especially welcome. It would be very
desirable to turn it into a full-blown C preprocessor with simply the
option of passing through #define statements unchanged.
</p>
<h3><a name="SecError">Error Reporting</a></h3>
<p>
Error reporting by GlueGen's parser is currently less than ideal.
Because PCPP makes <code>#include</code> directives disappear
completely with respect to the C parser (it appears that the
<code>#line</code> directives it emits are not being consumed properly
-- an area which needs more investigation), the line numbers reported
in parse failures are incorrect in all but the simplest cases. This
makes it difficult to determine in exactly what header file and on
exactly what construct the C parser failed.
</p>
<p>
Fortunately, there is a relatively simple workaround. PCPP can be run
with all of the same -I arguments passed to GlueGen and the result
piped to a new .c file. GlueGen can then be invoked on that .c file
(now containing no <code>#include</code> directives) and the line
numbers on any parse failures will be correct.
</p>
<h3><a name="SecStub">Stub Headers</a></h3>
<p>
As much as is possible, GlueGen is intended to operate on unmodified C
header files, so that it is easy to upgrade the given C API being
bound to Java simply by dropping in a new set of header files.
However, most C headers contain references to standard headers like
<code>stdio.h</code>, and if this header is parsed by GlueGen, the
tool will automatically attempt to generate Java entry points for such
routines as <code>fread</code> and <code>fwrite</code>, among others.
It is impractical to exclude these APIs on a case by case basis.
Therefore, the suggested technique to avoid polluting the binding with
these APIs is to "stub out" the headers.
</p>
<p>
GlueGen searches the include path for headers in the order the include
directories were specified to the tool. Placing another directory in
front of the one in which the bulk of the headers are found allows,
for example, an alternative <code>stdio.h</code> to be inserted which
contains few or no declarations but which satisfies the need of the
dependent header to find such a file.
</p>
<p>
GlueGen uses a complete ANSI and GNU C parser written by John Mitchell
and Monty Zukowski from the set of grammars available for the ANTLR
tool by Terrence Parr. As a complete C parser, this grammar requires
all data types encountered during the parse to be fully defined. Often
a particular header will be included by another one in order to pick
up data type declarations rather than API declarations. Stubbing out
the header with a smaller one providing a "fake" type declaration is a
useful technique for avoiding the binding of unnecessary APIs during
the glue code process.
</p>
<p>
Here's an example from the JOGL glue code generation process. The
<code>glext.h</code> header defining OpenGL extensions references
<code>stddef.h</code> in order to pick up the <code>ptrdiff_t</code>
data type. We choose to not include the real stddef.h but instead to
swap in a stub header. The contents of this header are therefore as
follows:
</p>
<pre>
#if defined(_WIN64)
typedef __int64 ptrdiff_t;
#elif defined(__ia64__) || defined(__x86_64__)
typedef long int ptrdiff_t;
#else
typedef int ptrdiff_t;
#endif
</pre>
<p>
This causes the ptrdiff_t data type to be defined appropriately for
the current architecture. It will be referenced during the glue code
generation and cause a Java value of the appropriate type (int or
long) to be used to represent it.
</p>
<p>
This is not the best example because it involves a data type which
changes size between 32- and 64-bit platforms, and there are otner
considerations to take into account in these situations (see the
section <a href="#Sec32">32- and 64-bit considerations</a>). Here's
another example, again from the JOGL source tree. JOGL binds the AWT
Native Interface, or JAWT, up to the Java programming language so that
the low-level code which binds OpenGL contexts to Windows device
contexts may be written in Java. The JDK's <code>jawt_md.h</code> on
the Windows platform includes <code>windows.h</code> to pick up the
definitions of data types such as <code>HWND</code> (window handle)
and <code>HDC</code> (handle to device context). However, it is
undesirable to try to parse the real <code>windows.h</code> just to
pick up these typedefs; not only does this header contain thousands of
unneeded APIs, but it also uses certain macro constructs not supported
by GlueGen's <a href="#SecPCPP">minimal C preprocessor</a>. To avoid
these problems, a "stub" <code>windows.h</code> header is placed in
GlueGen's include path containing only the necessary typedefs:
</p>
<pre>
typedef struct _handle* HANDLE;
typedef HANDLE HDC;
typedef HANDLE HWND;
</pre>
<p>
Note that it is essential that the type being specified to GlueGen is
compatible at least in semantics with the real definition of the
HANDLE typedef in the real <code>windows.h</code>, so that during
compilation of GlueGen's autogenerated C code, when the real
<code>windows.h</code> is referenced by the C compiler, the
autogenerated code will compile correctly.
</p>
<p>
This example is not really complete as it also requires <a
href="#Sec32">consideration of the size of data types on 32- and
64-bit platforms</a> as well as a discussion of how certain <a
href="#SecOpaque">opaque data types</a> are described to GlueGen and
exposed in its autogenerated APIs. Nonetheless, it illustrates at a
basic level why using a stub header is necessary and useful in certain
situations.
</p>
<h3><a name="Sec32">32- and 64-bit Considerations</a></h3>
<p>
When binding C functions to the Java programming language, it is
important that the resulting Java code support execution on a 64-bit
platform if the associated native methods are compiled appropriately.
In other words, the public Java API should not change if the
underlying C data types change to another data model such as LP64 (in
which longs and pointers become 64-bit).
</p>
<p>
GlueGen internally maintains two descriptions of the underlying C data
model: one for 32-bit architectures and one for 64-bit architectures.
These machine descriptions are used when deciding the mapping between
integral C types such as int and long and the corresponding Java
types, as well as when laying out C structs for access by the Java
language. For each autogenerated C struct accessor, both a 32-bit and
64-bit variant are generated behind the scenes, ensuring that the
resulting Java code will run correctly on both 32-bit and 64-bit
architectures.
</p>
<p>
When generating the main class containing the bulk of the method
bindings, GlueGen uses the 64-bit machine description to map C data
types to Java data types. This ensures that the resulting code will
run properly on 64-bit platforms. Note that it also generally means
that C <code>long</code>s will be mapped to Java <code>long</code>s,
since an LP64 data model is assumed.
</p>
<p>
If <a href="#SecOpaque">Opaque directives</a> are used to cause a
given C integer or pointer data type to be mapped directly to a Java
primitive type, care should be taken to make sure that the Java
primitive type is wide enough to hold all of the data even on 64-bit
platforms. Even if the data type is defined in the header file as
being only a 32-bit C integer, if there is a chance that on a 64-bit
platform the same header may define the data type as a 64-bit C
integer or long, the Opaque directive should map the C type to a Java
long.
</p>
<h3><a name="SecOpaque">Opaque Directives</a></h3>
<p>
Complex header files may contain declarations for certain data types
that are either too complex for GlueGen to handle or unnecessarily
complex from the standpoint of glue code generation. In these
situations a stub header may be used to declare a suitably compatible
typedef for the data type. An <a href="#Opaque">Opaque</a> directive
can be used to map the resulting typedef to a Java primitive type if
it is undesirable to expose it as a full-blown Java wrapper class.
</p>
<p>
GlueGen hashes all typedefs internally down to their underlying
primitive type. (This is probably not really correct according to the
C type system, but is correct enough from a glue code generation
standpoint, where if the types are compatible they are considered
equivalent.) This means that if the parser encounters
</p>
<pre>
typedef void* LPVOID;
</pre>
<p>
then an Opaque directive stating
</p>
<pre>
Opaque long LPVOID
</pre>
<p>
will cause all <code>void*</code> or <code>LPVOID</code> arguments in
the API to be mapped to Java longs, which is almost never desirable.
Unfortunately, it is not currently possible to distinguish between the
LPVOID typedef and the underlying <code>void*</code> data type in this
situation.
</p>
<p>
A similar problem occurs for other data types for which Opaque
directives may be desired. For example, a Windows HANDLE equates to a
typedef to <code>void*</code>, but performing this typedef in a stub
header and then adding the Opaque directive
</p>
<pre>
Opaque long HANDLE
</pre>
<p>
will cause all void* arguments to be exposed as Java longs instead of
Buffers, which is again undesirable. Attempting to work around the
problem by typedef'ing HANDLE to an integral type, as in:
</p>
<pre>
typedef long HANDLE;
</pre>
<p>
may itself have problems, because GlueGen will assume the two integral
types are compatible and not perform any intermediate casts between
HANDLE and jlong in the autogenerated C code. (When casting between a
pointer type and a JNI integral type such as jlong in C code, GlueGen
automatically inserts casts to convert the pointer first to an
"intptr_t" and then to the appropriate JNI type, in order to silence
compiler warnings and/or errors.)
</p>
<p>
What is desired is to produce a new type name distinct from all others
but still compatible with the pointer semantics of the original type.
Then an Opaque directive can be used to map the new type name to, for
example, a Java long.
</p>
<p>
To implement this in the context of the HANDLE example, the following
typedef may be inserted into the stub header:
</p>
<pre>
typedef struct _handle* HANDLE;
</pre>
<p>
This uses a pointer to an anonymous struct name to produce a new
pointer type. This is legal ANSI C and is supported by GlueGen's
parser without having seen a declaration for "struct _handle".
Subsequently, an Opaque directive can be used to map the HANDLE data
type to a Java long:
</p>
<pre>
Opaque long HANDLE
</pre>
<p>
Now HANDLEs are exposed to Java as longs as desired. A similar
technique is used to expose XIDs on the X11 platform as Java longs.
</p>
<h3><a name="SecSubstitution">Argument Name Substitution</a></h3>
<p>
Certain configuration file directives allow the insertion of Java or C
code at various places in the generated glue code, to both eliminate
the need to hand-edit the generated glue code as well as to minimize
the hand-writing of glue code, which sidesteps the GlueGen process. In
some situations the inserted code may reference incoming arguments to
compute some value or perform some operation. Examples of directives
supporting this substitution include <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> and <a
href="#ReturnedArrayLength">ReturnedArrayLength</a>.
</p>
<p>
The expressions in these directives may contain Java MessageFormat
expressions like <code>{0}</code> which refer to the incoming argument
names to the function. <code>{0}</code> refers to the first incoming
argument.
</p>
<p>
Strongly-typed C primitive pointers such as <code>int*</code>, which
ordinarily expand to overloaded Java methods taking
e.g. <code>int[]</code> as well as <code>IntBuffer</code>, present a
problem. The expansion to <code>int[] arr</code> also generates an
<code>int arr_offset</code> argument to be able to pass a pointer into
the middle of the array down to C. To allow the same MessageFormat
expression to be used for both cases, the subsitution that occurs when
such a primitive array is referenced is the string <code>arr,
arr_offset</code>; in other words, the subtituted string contains a
comma. This construct may be used in the following way: the code being
manually inserted may itself contain a method call taking
e.g. <code>{3}</code> (the incoming argument index of the primitive
array or buffer). The user should supply two overloaded versions of
this method, one taking a strongly-typed Buffer and one taking e.g. an
<code>int[] arr</code> and <code>int arr_offset</code> argument. The
implementation of <code>RangeCheck</code>s for primitive arrays and
strongly-typed buffers uses this construct.
</p>
<p>
It should be noted that in the autogenerated C code the offset
argument is expressed in bytes while at the Java level it is expressed
in elements. Most uses of GlueGen will probably not have to refer to
the primitive array arguments in C code so this slight confusion
should be minor.
</p>
<h3><a name="SecConfiguration">Configuration File Directives</a></h3>
<p>
In addition to the C headers, GlueGen requires a certain amount of
metadata in the form of configuration files in order to produce its
glue code. There are three basic reasons for this: first, GlueGen must
be informed into which Java classes the C methods are to be bound;
second, there are many configuration options for the generated glue
code, and passing them all on the command line is infeasible; and
third, there are ambiguities in many constructs in the C programming
language which must be resolved before a Java binding can be produced.
</p>
<p>
The contents of the configuration file are dependent on the class of
emitter specified to GlueGen. Currently there are three built-in
emitter classes: JavaEmitter, which produces a basic, static Java
binding of C functions; ProcAddressEmitter, which extends JavaEmitter
by calling the underlying C functions through function pointers,
resulting in more dynamic behavior and supporting C APIs with optional
functionality; and GLEmitter, which specializes ProcAddressEmitter to
support some OpenGL-specific constructs. The GLEmitter will be ignored
in this manual as it is specialized for JOGL and provides very little
additional functionality beyond the ProcAddressEmitter. The
JavaEmitter and ProcAddressEmitter support many options in their
configuration files. As the ProcAddressEmitter is a subclass of
JavaEmitter, all of the constructs in the JavaEmitter's configuration
files are also legal in the ProcAddressEmitter's configuration files.
</p>
<p>
The configuration files have a very simple line-by-line structure, and
are parsed by a very rudimentary, hand-written parser. Each
non-whitespace and non-comment line (note: comment lines begin with
'#') contains a directive like <code>Package</code>,
<code>Style</code> or <code>JavaClass</code> followed by arguments to
that directive. There are a certain set of directives that are
required for any code generation; others are optional and their
omission results in some default behavior. Directives are
case-insensitive.
</p>
<p>
The following is an exhaustive list of the options currently supported
by each of these emitters' configuration files. It is difficult to see
exactly how to use the tool based simply on these descriptions, so the
<a href="#Chapter3">examples</a> may be more helpful in seeing exactly
how to structure a configuration file for proper glue code generation.
</p>
<h4><a name="SecJavaEmitter">JavaEmitter Configuration</a></h4>
<p>
Note that only a very few of the following directives are specified as
being "required" rather than "optional"; these indicate the minimal
directives needed for a valid configuration file to begin to get glue
code to be produced. In general, these are <a
href="#Package">Package</a>, <a href="#ImplPackage">ImplPackage</a>,
<a href="#JavaClass">JavaClass</a>, <a
href="#ImplJavaClass">ImplJavaClass</a>, and <a
href="#Style">Style</a>. Other directives such as <a
href="#NioDirectOnly">NioDirectOnly</a> are required in some
circumstances for the glue code to be correct, and some such as <a
href="#ReturnedArrayLength">ReturnedArrayLength</a>, <a
href="#ReturnValueCapacity">ReturnValueCapacity</a>, and <a
href="#ReturnValueLength">ReturnValueLength</a> should be specified in
some situations in order for certain return values to be useful at the
Java level.
</p>
<p>
The following directives are specified in alphabetical order, although
this is not necessarily the best semantic order.
</p>
<dl>
<dt><strong><a name="AccessControl">AccessControl</a></strong></dt>
<dd> Syntax: <code>AccessControl [method name] [ PUBLIC | PROTECTED | PRIVATE | PACKAGE_PRIVATE ]</code> <br/>
(optional) Controls the access control of a certain Java method
corresponding to a C function. The access control of all APIs defaults
to public. This is useful when using the C binding of a particular
function only as one implementation strategy of the real public API
and using <a href="#CustomJavaCode">CustomJavaCode</a> to write the
exposed API. In this case is most useful in conjunction with <a
href="#RenameJavaMethod">RenameJavaMethod</a>.
</dd>
<dt><strong><a name="ArgumentIsString">ArgumentIsString</a></strong></dt>
<dd> Syntax: <code>ArgumentIsString [function name]
[indices...]</code> where the first argument index is 0 <br/>
(optional) For a C function with one or more outgoing
<code>char*</code> (or compatible data type) arguments, indicates that
those arguments are semantically null-terminated C strings rather than
arbitrary arrays of bytes. The generated glue code will be modified to
emit those arguments as java.lang.String objects rather than
<code>byte[]</code> or <code>ByteBuffer</code>.
</dd>
<dt><strong><a name="ClassJavadoc">ClassJavadoc</a></strong></dt>
<dd> Syntax: <code>ClassJavadoc [class name] [code...]</code> <br/>
(optional) Causes the specified line of code to be emitted in the
appropriate place in the generated code to become the per-class
Javadoc for the specified class. By default GlueGen produces no
Javadoc for its generated classes, so this is the mechanism by which a
user can emit Javadoc for these classes. The specified Javadoc
undergoes no transformation by GlueGen, so the initial
<code>/**</code> and trailing <code>*/</code> must be included in the
correct place. Each line of Javadoc is emitted in the order
encountered during parsing of the configuration files.
</dd>
<dt><strong><a name="CustomCCode">CustomCCode</a></strong></dt>
<dd>Syntax: <code>CustomCCode [code...]</code> <br/>
(optional) Causes the specified line of C code to be emitted into the
generated native code for the implementing class. Currently there is
no way (and no real need) to be able to emit custom C code into any
other generated .c file, so the class name in the <a
href="#CustomJavaCode">CustomJavaCode</a> directive is omitted.
</dd>
<dt><strong><a name="CustomJavaCode">CustomJavaCode</a></strong></dt>
<dd>Syntax: <code>CustomJavaCode [class name] [code...]</code> <br/>
(optional) Causes the specified line of Java code to be emitted into
the specified generated Java class. Can be used to emit code into any
generated class: the public interface, the implementing class, the
sole concrete class (in the case of the AllStatic <a
href="#Style">Style</a>), or any of the Java classes corresponding to
referenced C structs in the parsed headers. This usage is somewhat
verbose, and the <a href="#IncludeAs">IncludeAs</a> directive provides
a more concise way of including large bodies of Java code into the
generated code.
</dd>
<dt><strong><a name="EmitStruct">EmitStruct</a></strong></dt>
<dd>Syntax: <code>EmitStruct [C struct type name]</code> <br/>
(optional) Forces a Java class to be emitted for the specified C
struct. Normally only those structs referenced directly by the parsed
C APIs have corresponding Java classes emitted.
</dd>
<dt><strong><a name="GlueGenRuntimePackage">GlueGenRuntimePackage</a></strong></dt>
<dd>Syntax: <code>GlueGenRuntimePackage [package name, like com.jogamp.gluegen.runtime]</code> <br/>
(optional) Changes the package in which the generated glue code
expects to find its run-time helper classes (like Buffers, CPU,
StructAccessor). Defaults to <code>com.jogamp.gluegen.runtime</code> (no
quotes). This is useful if you want to bundle the runtime classes in
your application without the possibility of interfering with other
versions elsewhere in the system.
</dd>
<dt><strong><a name="Extends">Extends</a></strong></dt>
<dd>Syntax: <code>Extends [Java interface name] [interface name to extend] </code> <br/>
(optional) Causes the specified autogenerated Java interface to
declare that it extends another one. This directive may only be
applied to autogenerated interfaces, not concrete classes. For
concrete classes, use the <a href="#Implements">Implements</a>
directive.
</dd>
<dt><strong><a name="HierarchicalNativeOutput">HierarchicalNativeOutput</a></strong></dt>
<dd>Syntax: <code>HierarchicalNativeOutput true</code> <br/>
(optional) If "true", makes subdirectories for the generated native
code matching the package names of the associated classes. This is
typically not needed (or desired, as it complicates the compilation
process for this native code) and defaults to false.
</dd>
<dt><strong><a name="Ignore">Ignore</a></strong></dt>
<dd> Syntax: <code>Ignore [regexp]</code> <br/>
(optional) Ignores one or more functions or data types matching the
regexp argument which are encountered during parsing of the C
headers. By default GlueGen will emit all encountered C functions as
well as Java classes corresponding to all C structs referenced by
those functions. Related directives are <a
href="#IgnoreNot">IgnoreNot</a>, <a href="#Unignore">Unignore</a> and
<a href="#EmitStruct">EmitStruct</a>.
</dd>
<dt><strong><a name="IgnoreField">IgnoreField</a></strong></dt>
<dd> Syntax: <code>IgnoreField [struct type name] [field name]</code>
<br/>
(optional) Causes the specified field of the specified struct type
to be ignored during code generation, typically because it is too
complex for GlueGen to handle.
</dd>
<dt><strong><a name="IgnoreNot">IgnoreNot</a></strong></dt>
<dd> Syntax: see <a href="#Ignore">Ignore</a>.
(optional) Similar to the <a href="#Ignore">Ignore</a> directive, but
evaluates the negation of the passed regexp when deciding whether to
ignore the given function or data type. The <a
href="#Unignore">Unignore</a> mechanism may be used with IgnoreNot as
well. NOTE: the IgnoreNot mechanism may ultimately turn out to be
superfluous; the authors do not have sufficient experience with
regular expressions to know whether general negation of a regexp is
possible. Feedback in this area would be appreciated.
</dd>
<dt><strong><a name="Implements">Implements</a></strong></dt>
<dd> Syntax: <code>Implements [Java class name] [interface name to
implement]</code> <br/>
(optional) Causes the specified autogenerated Java concrete class to
declare that it implements the specified interface. This directive may
only be applied to autogenerated concrete classes, not interfaces. For
interfaces, use the <a href="#Extends">Extends</a> directive.
</dd>
<dt><strong><a name="ImplJavaClass">ImplJavaClass</a></strong></dt>
<dd> Syntax: <code>ImplJavaClass [class name]</code> <br/>
(optional) Specifies the name of the typically non-public,
implementation Java class which contains the concrete Java and native
methods for the glue code. If the emission style is AllStatic, there
is no distinction between the public and implementation class and
ImplJavaClass should not be specified. Otherwise, if the ImplJavaClass
is unspecified, it defaults to the JavaClass name plus "Impl". (If
both are unspecified in this configuration, an error is reported.) See
also <a href="#JavaClass">JavaClass</a>.
</dd>
<dt><strong><a name="ImplPackage">ImplPackage</a></strong></dt>
<dd> Syntax: <code>ImplPackage [package name]</code> <br/>
(optional) Specifies the package name into which the implementing
class containing the concrete Java and native methods will be emitted,
assuming an emission style of InterfaceAndImpl or ImplOnly. If
AllStatic, there is no separate implementing class from the public
interface. If the emission style is not AllStatic and the ImplPackage
is not specified, it defaults to the Package plus ".impl". See also <a
href="#Package">Package</a>.
</dd>
<dt><strong><a name="Import">Import</a></strong></dt>
<dd>Syntax: <code>Import [package name]</code> (no trailing semicolon)
<br/>
(optional) Adds an import statement at the top of each generated Java
source file.
</dd>
<dt><strong><a name="Include">Include</a></strong></dt>
<dd> Syntax: <code>Include [filename]</code> <br/>
(optional) Causes another configuration file to be read at the current
point in parsing the current configuration file. The filename argument
may be either absolute or relative; in the latter case it is specified
relative to the location of the current configuration file.
</dd>
<dt><strong><a name="IncludeAs">IncludeAs</a></strong></dt>
<dd>Syntax: <code>IncludeAs [prefix tokens] [filename]</code> <br/>
(optional) Similar to the <a href="#Include">Include</a> directive,
but prepends the specified prefix tokens on to every line of the file
to be read. The last token parsed is the name of the file to be
read. This allows, for example, <a
href="#CustomJavaCode">CustomJavaCode</a> to be stored as Java source
rather than in the configuration file; in this example the
configuration file might contain <code>IncludeAs CustomJavaCode
MyClass MyClass-CustomJavaCode.java</code>.
</dd>
<dt><strong><a name="JavaClass">JavaClass</a></strong></dt>
<dd> Syntax: <code>JavaClass [class name]</code> <br/>
(optional / required) Specifies the name of the public,
non-implementation Java class or interface into which the glue code
will be generated. If the emission style is not ImplOnly, the
JavaClass directive is required. See also <a
href="#ImplJavaClass">ImplJavaClass</a>.
</dd>
<dt><strong><a name="JavaEpilogue">JavaEpilogue</a></strong></dt>
<dd>Syntax: <code>JavaEpilogue [C function name] [code...]</code> <br/>
(optional) Adds the specified code as an epilogue in the Java method
for the specified C function; this code is run after the underlying C
function has been called via the native method but before any result
is returned. As in the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> and other
directives, <a href="#SecSubstitution">argument name
substitution</a> is performed on MessageFormat expressions in the
specified code. See also <a href="#JavaPrologue">JavaPrologue</a>.
</dd>
<dt><strong><a name="JavaOutputDir">JavaOutputDir</a></strong></dt>
<dd> Syntax: <code>JavaOutputDir [directory name]</code> <br/>
(optional) Specifies the root directory into which the emitted
Java code will be produced. Subdirectories for the packages of the
associated Java classes will be automatically created. If unspecified,
defaults to the current working directory.
</dd>
<dt><strong><a name="JavaPrologue">JavaPrologue</a></strong></dt>
<dd> Syntax: <code>JavaPrologue [C function name] [code...]</code>
<br/>
(optional) Adds the specified code as a prologue in the Java method
for the specified C function; this code is run before the underlying C
function is called via the native method. As in the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> and other
directives, <a href="#SecSubstitution">argument name
substitution</a> is performed on MessageFormat expressions in the
specified code. See also <a href="#JavaEpilogue">JavaEpilogue</a>.
</dd>
<dt><strong><a name="ManuallyImplement">ManuallyImplement</a></strong></dt>
<dd> Syntax: <code>ManuallyImplement [function name]</code> <br/>
(optional) Indicates to GlueGen to not produce a method into the
implementing class for the specified C function; the user must provide
one via the <a href="#CustomJavaCode">CustomJavaCode</a> directive. If
the emission style is InterfaceAndImpl or InterfaceOnly, a public
method will still be generated for the specified function.
</dd>
<dt><strong><a name="NativeOutputDir">NativeOutputDir</a></strong></dt>
<dd> Syntax: <code>NativeOutputDir [directory name]</code> <br/>
(optional) Specifies the root directory into which the emitted JNI
code will be produced. If unspecified, defaults to the current working
directory. See also <a
href="#HierarchicalNativeOutput">HierarchicalNativeOutput</a>.
</dd>
<dt><strong><a name="NioDirectOnly">NioDirectOnly</a></strong></dt>
<dd> Syntax: <code>NioDirectOnly [function name]</code> <br/>
(required when necessary) When passing a pointer down to a C API, it
is semantically undefined whether the underlying C code expects to
treat that pointer as a persistent pointer, living past the point of
return of the function call, or whether the pointer is used only
during the duration of the function call. For APIs taking C primitive
pointers such as <code>void*</code>, <code>float*</code>, etc.,
GlueGen will typically generate up to two overloaded Java methods, one
taking a <code>Buffer</code> or <code>Buffer</code> subclass such as
<code>FloatBuffer</code>, and one taking a primitive array such as
<code>float[]</code>. (In the case of <code>void*</code> outgoing
arguments, GlueGen produces only one variant taking a Buffer.)
Normally the generated glue code accepts either a "direct" or
non-"direct" buffer (according to the New I/O APIs) as argument.
However, if the semantics of the C function are that it either expects
to hold on to this pointer past the point of the function call, or if
it can block while holding on to the pointer, the
<code>NioDirectOnly</code> directive <strong>must</strong> be
specified for this C function in order for the generated glue code to
be correct. Failing to observe this requirement may cause JVM hangs or
crashes.
</dd>
<dt><strong><a name="Opaque">Opaque</a></strong></dt>
<dd> Syntax: <code>Opaque [Java primitive data type] [C data
type]</code> <br/>
(optional) Causes a particular C data type to be exposed in opaque
form as a Java primitive type. This is most useful for certain pointer
types for which it is not desired to generate full Java classes but
instead expose them to Java as e.g. <code>long</code>s. It is also
useful for forcing certain integral C data types to be exposed as e.g.
<code>long</code> to Java to ensure 64-bit cleanliness of the
generated glue code. See the <a href="#Chapter3">examples</a>. The C
data type may be a multiple-level pointer type; for example
<code>Opaque long void**</code>. Note that it is not currently
supported to make a given data type opaque for just a few functions;
the Opaque directive currently applies to all C functions in the
headers being parsed. This means that sweeping Opaque declarations
like <code>Opaque long void*</code> will likely have unforseen and
undesirable consequences.
</dd>
<dt><strong><a name="Package">Package</a></strong></dt>
<dd> Syntax: <code>Package [package name]</code> (no trailing
semicolon) <br/>
(optional / required) Specifies the package into which the public
interface or class for the autogenerated glue code will be
generated. Required whenever the emission style is not ImplOnly. See
also <a href="#ImplPackage">ImplPackage</a>.
</dd>
<dt><strong><a name="RangeCheck">RangeCheck</a></strong></dt>
<dd> Syntax: <code>RangeCheck [C function name] [argument number] [expression]</code> <br/>
(optional) Causes a range check to be performed on the specified array
or Buffer argument of the specified autogenerated Java method. This
range check ensures, for example, that a certain number of elements
are remaining in the passed Buffer, knowing that the underlying C API
will access no more than that number of elements. For range checks
that should be expressed in terms of a number of bytes rather than a
number of elements, see the <a
href="#RangeCheckBytes">RangeCheckBytes</a> directive. As in the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> and other
directives, <a href="#SecSubstitution">argument name substitution</a>
is performed on MessageFormat expressions.
</dd>
<dt><strong><a name="RangeCheckBytes">RangeCheckBytes</a></strong></dt>
<dd> Syntax: <code>RangeCheckBytes [C function name] [argument number]
[expression]</code> <br/>
(optional) Same as the <a href="#RangeCheck">RangeCheck</a> directive,
but the specified expression is treated as a minimum number of bytes
remaining rather than a minimum number of elements remaining. This
directive may not be used with primitive arrays.
</dd>
<dt><strong><a name="RenameJavaMethod">RenameJavaMethod</a></strong></dt>
<dd> Syntax: <code>RenameJavaMethod [from name] [to name]</code> <br/>
(optional) Causes the specified C function to be emitted under a
different name in the Java binding. This is most useful in conjunction
with the <a href="#AccessControl">AccessControl</a> directive when the
C function being bound to Java is only one potential implementation of
the public API, or when a considerable amount of Java-side custom code
is desired to wrap the underlying C native method entry point.
</dd>
<dt><strong><a name="RenameJavaType">RenameJavaType</a></strong></dt>
<dd> Syntax: <code>RenameJavaType [from name] [to name]</code> <br/>
(optional) Causes the specified C struct to be exposed as a Java class
under a different name. This only applies to autogenerated classes
corresponding to C structs encountered during glue code generation;
full control is provided over the name of the top-level classes
associated with the set of C functions via the <a
href="#JavaClass">JavaClass</a> and <a
href="#ImplJavaClass">ImplJavaClass</a> directives.
</dd>
<dt><strong><a name="ReturnedArrayLength">ReturnedArrayLength</a></strong></dt>
<dd> Syntax: <code>ReturnedArrayLength [C function name]
[expression]</code> where <code>expression</code> is a legal Java
expression with MessageFormat specifiers such as "{0}". These
specifiers will be replaced in the generated glue code with the
incoming argument names where the first argument to the method is
numbered 0. See the section on <a href="#SecSubstitution"> argument
name substitution</a>.<br/>
(optional) For a function returning a compound C pointer type such as
an <code>XVisualInfo*</code>, indicates that the returned pointer is
to be treated as an array and specifies the length of the returned
array as a function of the arguments passed to the function. Note that
this directive differs subtly from <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> and
ReturnValueLength. It is also sometimes most useful in conjunction
with the <a
href="#TemporaryCVariableDeclaration">TemporaryCVariableDeclaration</a>
and TemporaryCVariableAssignment directives.
</dd>
<dt><strong><a name="ReturnsString">ReturnsString</a></strong></dt>
<dd> Syntax: <code>ReturnsString [function name]</code> <br/>
(optional) Indicates that the specified C function which returns a
<code>char*</code> or compatible type actually returns a
null-terminated C string which should be exposed as a
java.lang.String. NOTE: currently does not properly handle the case
where this storage needs to be freed by the end user. In these
situations the data should be returned as a direct ByteBuffer, the
ByteBuffer converted to a String using custom Java code, and the
ByteBuffer freed manually using another function bound to Java.
</dd>
<dt><strong><a name="ReturnValueCapacity">ReturnValueCapacity</a></strong></dt>
<dd> Syntax: <code>ReturnValueCapacity [C function name]
[expression]</code> <br/>
(optional) Specifies the capacity of a java.nio <code>Buffer</code> or
subclass wrapping a C primitive pointer such as <code>char*</code> or
<code>float*</code> being returned from a C function. Typically
necessary in order to properly use such pointer return results from
Java. As in the <a href="#ReturnedArrayLength">ReturnedArrayLength</a>
directive, <a href="#SecSubstitution">argument name substitution</a>
is performed on MessageFormat expressions.
</dd>
<dt><strong><a name="ReturnValueLength">ReturnValueLength</a></strong></dt>
<dd> Syntax: <code>ReturnValueLength [C function name] [expression]</code> <br/>
(optional) Specifies the length of a returned array of pointers,
typically to C structs, from a C function. This differs from the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> directive in the
pointer indirection to the array elements. The <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> directive handles
slicing up of a linear array of structs, while the ReturnValueLength
directive handles boxing of individual elements of the array (which
are pointers) in to the Java class which wraps that C struct type. See
the <a href="#Chapter3">examples</a> for a concrete example of usage.
As in the <a href="#ReturnedArrayLength">ReturnedArrayLength</a>
directive, <a href="#SecSubstitution">argument name substitution</a>
is performed on MessageFormat expressions.
</dd>
<dt><strong><a name="RuntimeExceptionType">RuntimeExceptionType</a></strong></dt>
<dd> Syntax: <code>RuntimeExceptionType [class name]</code> <br/>
(optional) Specifies the class name of the exception type which should
be thrown when run-time related exceptions occur in the generated glue
code, for example if a non-direct Buffer is passed to a method for
which <a href="#NioDirectOnly">NioDirectOnly</a> was
specified. Defaults to <code>RuntimeException</code>.
</dd>
<dt><strong><a name="StructPackage">StructPackage</a></strong></dt>
<dd> Syntax: <code>StructPackage [C struct type name] [package
name]</code>. Package name contains no trailing semicolon. <br/>
(optional) Indicates that the specified Java class corresponding to
the specified C struct should be placed in the specified package. By
default, these autogenerated Java classes corresponding to C structs
are placed in the main package (that defined by <a
href="#PackageName">PackageName</a>).
</dd>
<dt><strong><a name="Style">Style</a></strong></dt>
<dd> Syntax: <code> Style [ AllStatic | InterfaceAndImpl |InterfaceOnly | ImplOnly ] </code> <br/>
(optional) Defines how the Java API for the parsed C headers is
structured. If AllStatic, one concrete Java class will be generated
containing static methods corresponding to the C entry points. If
InterfaceAndImpl, a public Java interface will be generated into the
<a href="#Package">Package</a> with non-static methods corresponding
to the C functions, and an "implementation" concrete Java class
implementing this interface will be generated into the <a
href="#ImplPackage">ImplPackage</a>. If InterfaceOnly, the
InterfaceAndImpl code generation style will be followed, but only the
interface will be generated. If ImplOnly, the InterfaceAndImpl code
generation style will be followed, but only the concrete implementing
class will be generated. The latter two options are useful when
generating a public API in which certain operations are unimplemented
on certain platforms; platform-specific implementation classes can be
generated which implement or leave unimplemented various parts of the
API.
</dd>
<dt><strong><a name="TemporaryCVariableAssignment">TemporaryCVariableAssignment</a></strong></dt>
<dd> Syntax: <code>TemporaryCVariableAssignment [C function name][code...]</code> <br/>
(optional) Inserts a C variable assignment declared using the <a
href="#TemporaryCVariableDeclaration">TemporaryCVariableDeclaration</a>
directive in to the body of a particular autogenerated native
method. The assignment is performed immediately after the call to the
underlying C function completes. This is typically used in
conjunction with the <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> or <a
href="#ReturnValueLength">ReturnValueLength</a> directives to capture
the size of a returned C buffer or array of pointers. See the <a
href="#Chapter3">examples</a> for a concrete example of usage of this
directive. Note that unlike, for example, the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> directive, no
substitution is performed on the supplied code, so the user must
typically have previously looked at the generated code and seen what
work needed to be done and variables needed to be examined at exactly
that line.
</dd>
<dt><strong><a name="TemporaryCVariableDeclaration">TemporaryCVariableDeclaration</a></strong></dt>
<dd> Syntax: <code>TemporaryCVariableDeclaration [C function name]
[code...]</code> <br/>
(optional) Inserts a C variable declaration in to the body of a
particular autogenerated native method. This is typically used in
conjunction with the <a
href="#TemporaryCVariableAssignment">TemporaryCVariableAssignment</a>
and <a href="#ReturnValueCapacity">ReturnValueCapacity</a> or <a
href="#ReturnValueLength">ReturnValueLength</a> directives to capture
the size of a returned C buffer or array of pointers. See the <a
href="#Chapter3">examples</a> for a concrete example of usage of this
directive.
</dd>
<dt><strong><a name="Unignore">Unignore</a></strong></dt>
<dd> Syntax: <code>Unignore [regexp]</code> <br/>
(optional) Removes a previously-defined <a href="#Ignore">Ignore</a>
directive. This is useful when one configuration file includes
another and wishes to disable some of the Ignores previously
specified.
</dd>
<dt><strong><a name="Unimplemented">Unimplemented</a></strong></dt>
<dd> Syntax: <code>Unimplemented [regexp]</code> <br/>
(optional) Causes the binding for the functions matching the passed
regexp to have bodies generated which throw the stated <a
href="#RuntimeExceptionType">RuntimeExceptionType</a> indicating that
this function is unimplemented. This is most useful when an API
contains certain functions that are not supported on all platforms and
there are multiple implementing classes being generated, one per
platform.
</dd>
</dl>
<h4><a name="SecProcAddressEmitter">ProcAddressEmitter Configuration</a></h4>
<p>
The ProcAddressEmitter is a subclass of the core JavaEmitter which
knows how to call C functions through function pointers. In
particular, the ProcAddressEmitter detects certain constructs in C
header files which imply that the APIs are intended to be called
through function pointers, and generates the glue code appropriately
to support that.
</p>
<p>
The ProcAddressEmitter detects pairs of functions and function pointer
typedefs in a set of header files. If it finds a matching pair, it
converts the glue code emission style for that API to look for the
function to call in an autogenerated table called a ProcAddressTable
rather than linking the autogenerated JNI code directly to the
function. It then changes the calling convention of the underlying
native method to pass the function pointer from Java down to C, where
the call-through-function-pointer is performed.
</p>
<p>
The ProcAddressEmitter discovers the function and function pointer
pairs by being informed of the mapping between their names by the
user. In the OpenGL and OpenAL libraries, there are fairly simple
mappings between the functions and function pointers. For example, in
the OpenGL <code>glext.h</code> header file, one may find the
following pair:
</p>
<pre>
GLAPI void APIENTRY glFogCoordf (GLfloat);
...
typedef void (APIENTRYP PFNGLFOGCOORDFPROC) (GLfloat coord);
</pre>
<p>
Therefore the mapping rule between the function name and the function
pointer typedef for the OpenGL extension header file is "PFN +
Uppercase(funcname) + PROC". Similarly, in the OpenAL 1.1 header
files, one may find the following pair:
</p>
<pre>
AL_API void AL_APIENTRY alEnable( ALenum capability );
...
typedef void (AL_APIENTRY *LPALENABLE)( ALenum capability );
</pre>
<p>
Therefore the mapping rule between the function name and the function
pointer typedef for the OpenAL header files is "LP +
Uppercase(funcname)".
</p>
<p>
These are the two principal function pointer-based APIs toward which
the GlueGen tool has currently been applied. It may turn out to be
that this simple mapping heuristic is insufficient, in which case it
will need to be extended in a future version of the GlueGen tool.
</p>
<p>
Note that it is currently the case that in order for the
ProcAddressEmitter to notice that a given function should be called
through a function pointer, it must see both the function prototype as
well as the function pointer typedef. Some headers, in particular the
OpenAL headers, have their <code>#ifdefs</code> structured in such a
way that either the declaration or the typedef is visible, but not
both simultaneously. Because the <a href="#SecPCPP">PCPP</a> C
preprocessor GlueGen uses obeys <code>#ifdefs</code>, it is in a
situation like this that the headers would have to be modified to
allow GlueGen to see both declarations.
</p>
<p>
The following directives are specified in alphabetical order, although
this is not necessarily the best semantic order. The
ProcAddressEmitter also accepts all of the directives supported by the
JavaEmitter. The required directives are <a
href="#GetProcAddressTableExpr">GetProcAddressTableExpr</a> and <a
href="#ProcAddressNameExpr">ProcAddressNameExpr</a>.
</p>
<dl>
<dt><strong><a name="EmitProcAddressTable">EmitProcAddressTable</a></strong></dt>
<dd> Syntax: <code>EmitProcAddressTable [true | false]</code> <br/>
(optional) Indicates whether to emit the ProcAddressTable during glue
code generation. Defaults to false.
</dd>
<dt><strong><a name="ForceProcAddressGen">ForceProcAddressGen</a></strong></dt>
<dd> Syntax: <code>ForceProcAddressGen [function name]</code> <br/>
(optional) Indicates that a ProcAddressTable entry should be produced
for the specified function even though it does not have an associated
function pointer typedef in the header. This directive does not
currently cause the autogenerated Java and C code to change to
call-through-function-pointer style, which should probably be
considered a bug. (FIXME)
</dd>
<dt><strong><a name="GetProcAddressTableExpr">GetProcAddressTableExpr</a></strong></dt>
<dd> Syntax: <code>GetProcAddressTableExpr [expression]</code> <br/>
(required) Defines the Java code snippet used by the generated glue
code to fetch the ProcAddressTable containing the function pointers
for the current API. It is up to the user to decide where to store the
ProcAddressTable. Common places for it include in an instance field of
the implementing class, in an associated object with which there is a
one-to-one mapping, or in a static field of another class accessed by
a static method. In the JOGL project, for example, each GLImpl
instance has an associated GLContext in an instance field called
"_context", so the associated directive is
<code>GetProcAddressTableExpr _context.getGLProcAddressTable()</code>.
In the JOAL project, the ProcAddressTables are currently held in a
separate class accessed via static methods, so one of the associated
directives is <code>GetProcAddressTableExpr
ALProcAddressLookup.getALCProcAddressTable()</code>.
</dd>
<dt><strong><a name="ProcAddressNameExpr">ProcAddressNameExpr</a></strong></dt>
<dd> Syntax: <code>ProcAddressNameExpr [expression]</code> <br/>
(required) Defines the mapping from function name to function pointer
typedef to be able to properly identify this function as needing
call-through-function-pointer semantics. The supplied expression uses
a set of simple commands to describe certain operations on the
function name:
<ul>
<li> <code>$UpperCase(arg)</code> converts the argument to
uppercase. "UpperCase" is case-insensitive.</li>
<li> <code>$LowerCase(arg)</code> converts the argument to
lowercase. "LowerCase" is case-insensitive.</li>
<li> <code>{0}</code> represents the name of the function.</li>
<li> Any other string represents a constant string.</li>
<li> Concatenation is implicit.</li>
</ul>
The corresponding ProcAddressNameExpr for the OpenGL extension
functions as described at the start of this section is <code>PFN
$UPPERCASE({0}) PROC</code>. The ProcAddressNameExpr for the OpenAL
functions as described at the start of this section is <code>LP
$UPPERCASE({0})</code>.
</dd>
<dt><strong><a name="ProcAddressTableClassName">ProcAddressTableClassName</a></strong></dt>
<dd> Syntax: <code>ProcAddressTableClassName [class name]</code> <br/>
(optional) Specifies the class name into which the table containing
the function pointers will be emitted. Defaults to "ProcAddressTable".
</dd>
<dt><strong><a name="ProcAddressTablePackage">ProcAddressTablePackage</a></strong></dt>
<dd> Syntax: <code>ProcAddressTablePackage [package name] (no
trailing semicolon)</code> <br/>
(optional) Specifies the package into which to produce the
ProcAddressTable for the current set of APIs. Defaults to the
implementation package specified by the <a
href="#ImplPackage">ImplPackage</a> directive.
</dd>
<dt><strong><a name="SkipProcAddressGen">SkipProcAddressGen</a></strong></dt>
<dd> Syntax: <code>SkipProcAddressGen [function name]</code> <br/>
(optional) Indicates that the default behavior of
call-through-function-pointer should be skipped for this function
despite the fact that it has an associated function pointer typedef in
the header.
</dd>
</dl>
<h2> <a name="Chapter3">Chapter 3 - Configuration File Examples</a> </h2>
<h3><a name="SecSimplest">Simplest possible example</a></h3>
Files:
<ul>
<li> <a href="example1/function.c">function.c</a></li>
<li> <a href="example1/function.h">function.h</a></li>
<li> <a href="example1/function.cfg">function.cfg</a></li>
<li> <a href="example1/gen.sh">gen.sh</a></li>
</ul>
<p> This example shows the simplest possible usage of GlueGen; a
single routine taking as arguments and returning only primitive
types. The signature of the C function we are interested in binding is
</p>
<pre>
int one_plus(int a);
</pre>
<p> To bind this function to Java, we only need a configuration file
with very basic settings, indicating the style of glue code emission,
the package and class into which the glue code will be generated, and
the output directories for the Java and native code. The contents of
the configuration file are as follows: </p>
<pre>
Package testfunction
Style AllStatic
JavaClass TestFunction
JavaOutputDir gensrc/java
NativeOutputDir gensrc/native
</pre>
<p> GlueGen can then be invoked with approximately the following
command line:
</p>
<pre>
java -cp gluegen.jar:antlr.jar com.sun.gluegen.GlueGen \
-I. -Ecom.sun.gluegen.JavaEmitter -Cfunction.cfg function.h
</pre>
<p> The resulting Java and native code needs to be compiled, and the
application needs to load the native library for the Java binding
before attempting to invoke the native method by calling
<code>System.load()</code> or <code>System.loadLibrary()</code>. </p>
<h3><a name="SecArrays">Arrays and buffers</a></h3>
Files:
<ul>
<li> <a href="example2/function.c">function.c</a></li>
<li> <a href="example2/function.h">function.h</a></li>
<li> <a href="example2/function.cfg">function.cfg</a></li>
<li> <a href="example2/gen.sh">gen.sh</a></li>
</ul>
<p> This example shows how C primitive arrays are bound to Java. The
header file contains three functions to bind: </p>
<pre>
float process_data(float* data, int n);
void set_global_data(float* data);
float process_global_data(int n);
</pre>
<p> The semantics of <code>process_data</code> are that it takes in a
pointer to a set of primitive <code>float</code> values and the number
of elements in the array and performs some operation on them,
returning a floating-point value as the result. Afterward the passed
data is no longer referenced.
</p>
<p> <code>set_global_data</code>, on the other hand, takes a pointer
to the data and stores it persistently in the C code.
<code>process_global_data</code> then accepts as argument the number
of elements to process from the previously-set global data, performs
this processing and returns a result. The global data may be accessed
again afterward. As an example, these kinds of semantics are used in
certain places in the OpenGL API.
</p>
<p> From a Java binding standpoint, <code>process_data</code> may
accept data stored either inside the Java heap (in the form of a
<code>float[]</code> or non-direct <code>FloatBuffer</code>) or
outside the Java heap (in the form of a direct
<code>FloatBuffer</code>), because it does not access the data after
the function call has completed and therefore would not be affected if
garbage collection moved the data after the function call was
complete. However, <code>set_global_data</code> can cause the passed
data to be accessed after the function call is complete, if
<code>process_global_data</code> is called. Therefore the data passed
to <code>set_global_data</code> may not reside in the Java
garbage-collected heap, but must reside outside the heap in the form
of a direct <code>FloatBuffer</code>.
</p>
<p> It is straightforward to take into account these differences in
semantics in the configuration file using the <a
href="#NioDirectOnly">NioDirectOnly</a> directive:
</p>
<pre>
# The semantics of set_global_data imply that
# only direct Buffers are legal
NioDirectOnly set_global_data
</pre>
<p> Note the differences in the generated Java-side overloadings for
the two functions:
</p>
<pre>
public static void process_data(java.nio.FloatBuffer data, int n) {...}
public static void process_data(float[] data, int data_offset, int n) {...}
public static void set_global_data(java.nio.FloatBuffer data) {...}
</pre>
<p>
No overloading is produced for <code>set_global_data</code> taking
a <code>float[]</code>, as it can not handle data residing in the Java
heap. Further, the generated glue code will verify that any
<code>FloatBuffer</code> passed to this routine is direct, throwing a
<code>RuntimeException</code> if not. The type of the exception thrown
in this and other cases may be changed with the <a
href="#RuntimeExceptionType">RuntimeExceptionType</a> directive.
</p>
<h3><a name="SecString">String handling</a></h3>
Files:
<ul>
<li> <a href="example3/function.h">function.h</a></li>
<li> <a href="example3/function.cfg">function.cfg</a></li>
<li> <a href="example3/gen.sh">gen.sh</a></li>
</ul>
<p> This example shows how to pass and return C strings. The functions
involved are a bit contrived, as nobody would ever need to bind the C
library's string handling routines to Java, but they do illustrate
situations in which Java strings might need to be passed to C and C
strings returned to Java. As an example, both styles of function are
present in the OpenGL and OpenAL APIs.
</p>
<p> The included source code exposes two functions to Java: </p>
<pre>
size_t strlen(const char* str);
char* strstr(const char* str1, const char* str2);
</pre>
<p> Note that we might just as easily parse the C standard library's
<code>string.h</code> header file to pick up these function
declarations. However for the purposes of this example it is easier to
extract just the functions we need.
</p>
<p> Note that the <a href="example3/function.h">function.h</a> header
file contains a typedef for <code>size_t</code>. This is needed
because GlueGen does not inherently know about this data type. An
equivalent data type for the purposes of this example is
<code>int</code>, so we choose to tell GlueGen to use that data type
in place of <code>size_t</code> while generating glue code.
</p>
<p> The following directive in the configuration file tells GlueGen
that <code>strlen</code> takes a string as argument 0 (the first
argument):
</p>
<pre>
ArgumentIsString strlen 0
</pre>
<p> The following directive tells GlueGen that <code>strstr</code>
takes two strings as its arguments: </p>
<pre>
ArgumentIsString strstr 0 1
</pre>
<p> Finally, the following directive tells GlueGen that
<code>strstr</code> returns a string instead of an array of bytes:
</p>
<pre>
ReturnsString strstr
</pre>
<p> We also use the <a href="#CustomCCode">CustomCCode</a> directive
to cause the <code>string.h</code> header file to be #included in the
generated glue code:
</p>
<pre>
CustomCCode /* Include string.h header */
CustomCCode #include <string.h>
</pre>
<p> Now the bindings of these two functions to Java look as expected:</p>
<pre>
public static native int strlen(java.lang.String str);
public static native java.lang.String strstr(java.lang.String str1, java.lang.String str2);
</pre>
Note that the <a href="#ReturnsString">ReturnsString</a> directive
does not currently correctly handle the case where the
<code>char*</code> returned from C needs to be explicitly freed. As an
example, a binding of the C function <code>strdup</code> using a
ReturnsString directive would cause a C heap memory leak.
<h3><a name="SecMemory">Memory allocation</a></h3>
Files:
<ul>
<li> <a href="example4/function.c">function.c</a></li>
<li> <a href="example4/function.h">function.h</a></li>
<li> <a href="example4/function.cfg">function.cfg</a></li>
<li> <a href="example4/gen.sh">gen.sh</a></li>
</ul>
<p> This example shows how memory allocation is handled when binding C
to Java. It gives the example of a custom memory allocator being bound
to Java; this is a construct that at least at one point was present in
OpenGL in the NV_vertex_array_range extension. </p>
<p> The two functions we are exposing to Java are as follows: </p>
<pre>
void* custom_allocate(int num_bytes);
void custom_free(void* data);
</pre>
<p> The Java-side return type of <code>custom_allocate</code> will
necessarily be a <code>ByteBuffer</code>, as that is the only useful
way of interacting with arbitrary memory produced by C. The question
is how to inform the glue code generator of the size of the returned
sequence of memory. The semantics of <code>custom_allocate</code> are
obvious to the programmer; the incoming <code>num_bytes</code>
argument specifies the amount of returned memory. We tell GlueGen this
fact using the <a href="#ReturnValueCapacity">ReturnValueCapacity</a>
directive: </p>
<pre>
# The length of the returned ByteBuffer from custom_allocate is
# specified as the argument
ReturnValueCapacity custom_allocate {0}
</pre>
<p> Note that we name incoming argument 0 with the MessageFormat
specifier "{0}" rather than the explicit name of the parameter
("num_bytes") for generality, in case the header file is changed
later.
</p>
<p> Because <code>custom_free</code> will only ever receive Buffers
produced by custom_allocate, we use the <a
href="#NioDirectOnly">NioDirectOnly</a> directive to prevent
accidental usage with the wrong kind of Buffer: </p>
<pre>
# custom_free will only ever receive a direct Buffer
NioDirectOnly custom_free
</pre>
<p> The generated Java APIs for these functions are as follows: </p>
<pre>
public static java.nio.ByteBuffer custom_allocate(int num_bytes) {...}
public static void custom_free(java.nio.Buffer data) {...}
</pre>
<h3><a name="SecStructs">Ingoing and outgoing structs</a></h3>
Files:
<ul>
<li> <a href="example5/function.c">function.c</a></li>
<li> <a href="example5/function.h">function.h</a></li>
<li> <a href="example5/function.cfg">function.cfg</a></li>
<li> <a href="example5/gen.sh">gen.sh</a></li>
</ul>
<p> This example shows how GlueGen provides access to C structs and
supports both passing them to and returning them from C functions. The
header file defines a sample data structure that might describe the
bit depth of a given screen: </p>
<pre>
typedef struct {
int redBits;
int greenBits;
int blueBits;
} ScreenInfo;
</pre>
<p> Two functions are defined which take and return this data type:
</p>
<pre>
ScreenInfo* default_screen_depth();
void set_screen_depth(ScreenInfo* info);
</pre>
<p> The semantics of <code>default_screen_depth()</code> are that it
returns a pointer to some static storage which does not need to be
freed, which describes the default screen depth.
<code>set_screen_depth()</code> is a hypothetical function which would
take a newly-allocated <code>ScreenInfo</code> and cause the primary
display to switch to the specified bit depth. </p>
<p> The only additional information we need to tell GlueGen, beyond
that in the header file, is how much storage is returned from
<code>default_screen_depth()</code>. Note the semantic ambiguity,
where it might return a pointer to a single <code>ScreenInfo</code> or
a pointer to an array of <code>ScreenInfo</code>s. We tell GlueGen
that the return value is a single value with the <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> directive,
similarly to the <a href="#SecMemory">memory allocation</a> example
above: </p>
<pre>
# Tell GlueGen that default_screen_depth() returns a pointer to a
# single ScreenInfo
ReturnValueCapacity default_screen_depth sizeof(ScreenInfo)
</pre>
<p> Note that if <code>default_screen_depth</code> had returned
newly-allocated storage, it would be up to the user to expose a
<code>free()</code> function to Java and call it when necessary. </p>
<p> GlueGen automatically generates a Java-side
<code>ScreenInfo</code> class which supports not only access to any
such objects returned from C, but also allocation of new
<code>ScreenInfo</code> structs which can be passed (persistently)
down to C. The Java API for the ScreenInfo class looks like this: </p>
<pre>
public abstract class ScreenInfo {
public static ScreenInfo create();
public abstract ScreenInfo redBits(int val);
public abstract int redBits();
...
}
</pre>
<p> The <code>create()</code> method allocates a new ScreenInfo struct
which may be passed, even persistently, out to C. Its C-heap storage
will be automatically reclaimed when the Java-side ScreenInfo object
is no longer reachable, as it is backed by a direct New I/O
<code>ByteBuffer</code>. The fields of the struct are exposed as
methods which supply both getters and setters. </p>
<h3><a name="SecStructArrays">Returned arrays of structs</a></h3>
Files:
<ul>
<li> <a href="example6/function.h">function.h</a></li>
<li> <a href="example6/function.cfg">function.cfg</a></li>
<li> <a href="example6/gen.sh">gen.sh</a></li>
</ul>
<p> This example, taken from JOGL's X11 binding, illustrates how to
return an array of structs from C to Java. The
<code>XGetVisualInfo</code> function from the X library has the
following signature: </p>
<pre>
XVisualInfo *XGetVisualInfo(
Display* display,
long vinfo_mask,
XVisualInfo* vinfo_template,
int* nitems_return
);
</pre>
<p> Note that the <code>XVisualInfo</code> data structure itself
contains many elements, including a pointer to the current visual. We
use the following trick in the header file to cause GlueGen to treat
the <code>Display*</code> in the above signature as well as the
<code>Visual*</code> in the <code>XVisualInfo</code> as opaque
pointers: </p>
<pre>
typedef struct {} Display;
typedef struct {} Visual;
typedef unsigned long VisualID;
typedef struct {
Visual *visual;
VisualID visualid;
int screen;
int depth;
int c_class; /* C++ */
unsigned long red_mask;
unsigned long green_mask;
unsigned long blue_mask;
int colormap_size;
int bits_per_rgb;
} XVisualInfo;
</pre>
<p> <code>XGetVisualInfo</code> returns all of the available pixel
formats in the form of <code>XVisualInfo</code>s which match a given
template. <code>display</code> is the current connection to the X
server. <code>vinfo_mask</code> indicates which fields from the
template to match against. <code>vinfo_template</code> is a partially
filled-in <code>XVisualInfo</code> specifying the characteristics to
match. <code>nitems_return</code> is a pointer to an integer
indicating how many <code>XVisualInfo</code>s were returned. The
return value, rather than being a pointer to a single
<code>XVisualInfo</code>, is a pointer to the start of an array of
<code>XVisualInfo</code> data structures. </p>
<p> There are two basic steps to being able to return this array
properly to Java using GlueGen. The first is creating a direct
ByteBuffer of the appropriate size in the autogenerated JNI code. The
second is slicing up this ByteBuffer appropriately in order to return
an <code>XVisualInfo[]</code> at the Java level. </p>
<p> In the autogenerated JNI code, after the call to
<code>XGetVisualInfo</code> is made, the outgoing
<code>nitems_return</code> value points to the number of elements in
the returned array, which indicates the size of the direct ByteBuffer
which would need to wrap these elements. However, if we look at the
implementation of one of the generated glue code variants for this
method (specifically, the one taking an <code>int[]</code> as the
third argument), we can see a problem in trying to access this value
in the C code: </p>
<pre>
JNIEXPORT jobject JNICALL
Java_testfunction_TestFunction_XGetVisualInfo1__Ljava_nio_ByteBuffer_2JLjava_nio_ByteBuffer_2Ljava_lang_Object_2I(
JNIEnv *env, jclass _unused, jobject arg0, jlong arg1, jobject arg2, jobject arg3, jint arg3_byte_offset) {
...
int * _ptr3 = NULL;
...
if (arg3 != NULL) {
_ptr3 = (int *) (((char*) (*env)->GetPrimitiveArrayCritical(env, arg3, NULL)) + arg3_byte_offset);
}
_res = XGetVisualInfo((Display *) _ptr0, (long) arg1, (XVisualInfo *) _ptr2, (int *) _ptr3);
if (arg3 != NULL) {
(*env)->ReleasePrimitiveArrayCritical(env, arg3, _ptr3, 0);
}
if (_res == NULL) return NULL;
return (*env)->NewDirectByteBuffer(env, _res, ??? What to put here ???);
}
</pre>
<p> Note that at the point of the statement "What to put here?" the
pointer to the storage of the <code>int[]</code>, <code>_ptr3</code>,
has already been released via
<code>ReleasePrimitiveArrayCritical</code>. This means that it may not
be referenced at the point needed in the code. </p>
<p> To solve this problem we use the <a
href="#TemporaryCVariableDeclaration">TemporaryCVariableDeclaration</a>
and <a
href="#TemporaryCVariableAssignment">TemporaryCVariableAssignment</a>
directives. We want to declare a persistent integer variable down in
the C code and assign the returned array length to that variable
before the primitive array is released. While in order to do this we
unfortunately need to know something about the structure of the
autogenerated JNI code, at least we don't have to hand-edit it
afterward. We add the following directives to the configuration file: </p>
<pre>
# Get returned array's capacity from XGetVisualInfo to be correct
TemporaryCVariableDeclaration XGetVisualInfo int count;
TemporaryCVariableAssignment XGetVisualInfo count = _ptr3[0];
</pre>
<p> Now in the autogenerated JNI code the variable "count" will
contain the number of elements in the returned array. We can then
reference this variable in a <a
href="#ReturnValueCapacity">ReturnValueCapacity</a> directive: </p>
<pre>
ReturnValueCapacity XGetVisualInfo count * sizeof(XVisualInfo)
</pre>
<p> At this point the <code>XGetVisualInfo</code> binding will return
a Java-side <code>XVisualInfo</code> object whose backing ByteBuffer
is the correct size. We now have to inform GlueGen that the underlying
ByteBuffer represents not a single <code>XGetVisualInfo</code> struct,
but an array of them, using the <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> directive. This
conversion is performed on the Java side of the autogenerated code.
Here, the first element of either the passed <code>IntBuffer</code> or
<code>int[]</code> contains the number of elements in the returned
array. (Alternatively, we could examine the length of the ByteBuffer
returned from C to Java and divide by
<code>XVisualInfo.size()</code>.) Because there are two overloadings
produced by GlueGen for this method, if we reference the
<code>nitems_return</code> argument in a <a
href="#ReturnedArrayLength">ReturnedArrayLength</a> directive, we need
to handle not only the differing data types properly
(<code>IntBuffer</code> vs. <code>int[]</code>), but also the fact
that both the integer array and its offset value are substituted for
any reference to the fourth argument. </p>
<p> To solve this problem, we define a pair of private helper
functions whose purpose is to handle this overloading. </p>
<pre>
CustomJavaCode TestFunction private static int getFirstElement(IntBuffer buf) {
CustomJavaCode TestFunction return buf.get(buf.position());
CustomJavaCode TestFunction }
CustomJavaCode TestFunction private static int getFirstElement(int[] arr,
CustomJavaCode TestFunction int offset) {
CustomJavaCode TestFunction return arr[offset];
CustomJavaCode TestFunction }
</pre>
<p> Now we can simply write for the returned array length: </p>
<pre>
ReturnedArrayLength XGetVisualInfo getFirstElement({3})
</pre>
<p> That's all that is necessary. GlueGen will then produce the
following Java-side overloadings for this function: </p>
<pre>
public static XVisualInfo[] XGetVisualInfo(Display arg0,
long arg1,
XVisualInfo arg2,
java.nio.IntBuffer arg3);
public static XVisualInfo[] XGetVisualInfo(Display arg0,
long arg1,
XVisualInfo arg2,
int[] arg3, int arg3_offset);
</pre>
<p> As it happens, we don't really need the Display and Visual data
structures to be produced; they can be treated as <code>long</code>s
on the Java side. Therefore we can add the following directives to the
configuration file: </p>
<pre>
# We don't need the Display and Visual data structures to be
# explicitly exposed
Opaque long Display *
Opaque long Visual *
# Ignore the empty Display and Visual data structures (though made
# opaque, the references from XVisualInfo and elsewhere are still
# traversed)
Ignore Display
Ignore Visual
</pre>
<p> The final generated Java API is the following: </p>
<pre>
public static XVisualInfo[] XGetVisualInfo(long arg0,
long arg1,
XVisualInfo arg2,
java.nio.IntBuffer arg3);
public static XVisualInfo[] XGetVisualInfo(long arg0,
long arg1,
XVisualInfo arg2,
int[] arg3, int arg3_offset);
</pre>
<h3><a name="SecPointerArrays">Returned arrays of pointers</a></h3>
Files:
<ul>
<li> <a href="example7/function.h">function.h</a></li>
<li> <a href="example7/function.cfg">function.cfg</a></li>
<li> <a href="example7/gen.sh">gen.sh</a></li>
</ul>
<p> As with the <a href="#SecStructArrays">example above</a>, this
example is taken from JOGL's X11 binding. Here we show how to expose
to Java a C routine returning an array of pointers to a data
structure. </p>
<p> The declaration of the function we are binding is as follows: </p>
<pre>
typedef struct __GLXFBConfigRec *GLXFBConfig;
GLXFBConfig *glXChooseFBConfig( Display *dpy, int screen,
const int *attribList, int *nitems );
</pre>
<p> This function is used during allocation of a hardware-accelerated
off-screen surface ("pbuffer") on X11 platforms; its exact meaning is
not important. The semantics of the arguments and return value are as
follows. As in the <a href="#SecStructArrays">previous example</a>, it
accepts a connection to the current X display as one argument. The
screen of this display is the second argument. The
<code>attribList</code> is a zero-terminated list of integer
attributes; because it is zero-terminated, the length of this list is
not passed to the function. As in the previous example, the
<code>nitems</code> argument points to an integer into which the
number of returned <code>GLXFBConfig</code> objects is placed. The
return value is an array of <code>GLXFBConfig</code> objects. </p>
<p> Because the <code>GLXFBConfig</code> data type is typedefed as a
pointer to an opaque (undefined) struct, the construct
<code>GLXFBConfig*</code> is implicitly a "pointer-to-pointer" type.
GlueGen automatically assumes this is convertible to a Java-side array
of accessors to structs. The only configuration necessary is to tell
GlueGen the length of this array. </p>
<p> As in the previous example, we use the <a
href="#TemporaryCVariableDeclaration">TemporaryCVariableDeclaration</a>
and <a
href="#TemporaryCVariableAssignment">TemporaryCVariableAssignment</a>
directives to capture the length of the returned array: </p>
TemporaryCVariableDeclaration glXChooseFBConfig int count;
TemporaryCVariableAssignment glXChooseFBConfig count = _ptr3[0];
<p> The structure of the generated glue code for the return value is
subtly different than in the previous example. The question in this
case is not whether the return value is a pointer to a single object
vs. a pointer to an array of objects; it is what the length of the
returned array is, since we already know that the return type is
pointer-to-pointer and is therefore an array. We use the <a
href="#ReturnValueLength">ReturnValueLength</a> directive for this
case: </p>
<pre>
ReturnValueLength glXChooseFBConfig count
</pre>
We add similar Opaque directives to the previous example to yield the
resulting Java bindings for this function:
<pre>
public static GLXFBConfig[] glXChooseFBConfig(long dpy,
int screen,
java.nio.IntBuffer attribList,
java.nio.IntBuffer nitems);
public static GLXFBConfig[] glXChooseFBConfig(long dpy,
int screen,
int[] attribList, int attribList_offset,
int[] nitems, int nitems_offset);
</pre>
Note that because the GLXFBConfig data type is returned as an element
of an array, we can not use the Opaque directive to erase this data
type to <code>long</code> as we did with the <code>Display</code> data
type.
</div>
</div>
<div id="footer">
<div id="footer_left">
<span>JogAmp.org</span>
by <a href="http://jogamp.org">http://jogamp.org</a>
is licensed under a <br/>
<a href="http://creativecommons.org/licenses/by/3.0/us/">Creative Commons Attribution 3.0 License</a>.
</div>
</div>
</div>
</body>
</html>
