2001-08-08 Alexandre Petit-Bianco <apbianco@redhat.com>

[official-gcc.git] / gcc / java / gcj.texi

@\input texinfo @c -*-texinfo-*-
@setfilename gcj.info
@settitle Guide to GNU gcj

@c Note: When reading this manual you'll find lots of strange
@c circumlocutions like ``compiler for the Java language''.
@c This is necessary due to Sun's restrictions on the use of
@c the word ``Java'.

@c When this manual is copyrighted.
@set copyrights-gcj 2001

@c Versions
@set version-gcc 3.1
@set which-gcj GCC-@value{version-gcc}

@ifinfo
@format
@dircategory Programming
@direntry
* Gcj: (gcj).               Ahead-of-time compiler for the Java language
@end direntry

@dircategory Individual utilities
@direntry
* gcjh: (gcj)Invoking gcjh.
                            Generate header files from Java class files
* jv-scan: (gcj)Invoking jv-scan.
                            Print information about Java source files
* jcf-dump: (gcj)Invoking jcf-dump.
                            Print information about Java class files
* gij: (gcj)Invoking gij.   GNU interpreter for Java bytecode
@end direntry
@end format

Copyright (C) @value{copyrights-gcj} Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.1 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``GNU General Public License'', the Front-Cover
texts being (a) (see below), and with the Back-Cover Texts being (b)
(see below).  A copy of the license is included in the section entitled
``GNU Free Documentation License''.

(a) The FSF's Front-Cover Text is:

     A GNU Manual

(b) The FSF's Back-Cover Text is:

     You have freedom to copy and modify this GNU Manual, like GNU
     software.  Copies published by the Free Software Foundation raise
     funds for GNU development.
@end ifinfo

@titlepage
@title GNU gcj
@author Tom Tromey

@page
@vskip 0pt plus 1filll
Copyright @copyright{} @value{copyrights-gcj} Free Software Foundation, Inc.
@sp 2
For the @value{which-gcj} Version*
@sp 1
Published by the Free Software Foundation @*
59 Temple Place - Suite 330@*
Boston, MA 02111-1307, USA@*
@sp 1
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.1 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``GNU General Public License'', the Front-Cover
texts being (a) (see below), and with the Back-Cover Texts being (b)
(see below).  A copy of the license is included in the section entitled
``GNU Free Documentation License''.

(a) The FSF's Front-Cover Text is:

     A GNU Manual

(b) The FSF's Back-Cover Text is:

     You have freedom to copy and modify this GNU Manual, like GNU
     software.  Copies published by the Free Software Foundation raise
     funds for GNU development.
@end titlepage
@contents
@page


@node Top
@top Introduction

This manual describes how to use @code{gcj}, the GNU compiler for the
Java programming language.  @code{gcj} can generate both @file{.class}
files and object files, and it can read both Java source code and
@file{.class} files.

@menu
* Copying::             The GNU General Public License
* GNU Free Documentation License::
                        How you can share and copy this manual
* Invoking gcj::        Compiler options supported by @code{gcj}
* Compatibility::       Compatibility between gcj and other tools for Java
* Invoking gcjh::       Generate header files from class files
* Invoking jv-scan::    Print information about source files
* Invoking jcf-dump::   Print information about class files
* Invoking gij::        Interpreting Java bytecodes
* Resources::           Where to look for more information
@end menu


@include gpl.texi

@include fdl.texi


@node Invoking gcj
@chapter Invoking gcj

As @code{gcj} is just another front end to @code{gcc}, it supports many
of the same options as gcc.  @xref{Option Summary, , Option Summary,
gcc, Using the GNU Compiler Collection}.  This manual only documents the
options specific to @code{gcj}.

@menu
* Input and output files::
* Input Options::               How gcj finds files
* Encodings::                   Options controlling source file encoding
* Warnings::                    Options controlling warnings specific to gcj
* Code Generation::             Options controlling the output of gcj
* Configure-time Options::      Options you won't use
@end menu

@node Input and output files
@section Input and output files

A @code{gcj} command is like a @code{gcc} command, in that it
consists of a number of options and file names.  The following kinds
of input file names are supported:

@table @code
@item @var{file}.java
Java source files.
@item @var{file}.class
Java bytecode files.
@item @var{file}.zip
@itemx @var{file}.jar
An archive containing one or more @code{.class} files, all of
which are compiled.  The archive may be compressed.
@item @@@var{file}
A file containing a whitespace-separated list of input file names.
(Currently, these must all be @code{.java} source files, but that
may change.)
Each named file is compiled, just as if it had been on the command line.
@item @var{library}.a
@itemx @var{library}.so
@itemx -l@var{libname}
Libraries to use when linking.  See the @code{gcc} manual.
@end table

You can specify more than one input file on the @code{gcj} command line,
in which case they will all be compiled.  If you specify a
@code{-o @var{FILENAME}}
option, all the input files will be compiled together, producing a
single output file, named @var{FILENAME}.
This is allowed even when using @code{-S} or @code{-c},
but not when using @code{-C}.
(This is an extension beyond the what plain @code{gcc} allows.)
(If more than one input file is specified, all must currently
be @code{.java} files, though we hope to fix this.)

@node Input Options
@section Input Options

@cindex class path

@code{gcj} has options to control where it looks to find files it needs.
For instance, @code{gcj} might need to load a class that is referenced
by the file it has been asked to compile.  Like other compilers for the
Java language, @code{gcj} has a notion of a @dfn{class path}.  There are
several options and environment variables which can be used to
manipulate the class path.  When @code{gcj} looks for a given class, it
searches the class path looking for matching @file{.class} or
@file{.java} file.  @code{gcj} comes with a built-in class path which
points at the installed @file{libgcj.jar}, a file which contains all the
standard classes.

In the below, a directory or path component can refer either to an
actual directory on the filesystem, or to a @file{.zip} or @file{.jar}
file, which @code{gcj} will search as if it is a directory.

@table @code
@item -I@var{dir}
All directories specified by @code{-I} are kept in order and prepended
to the class path constructed from all the other options.  Unless
compatibility with tools like @code{javac} is imported, we recommend
always using @code{-I} instead of the other options for manipulating the
class path.

@item --classpath=@var{path}
This sets the class path to @var{path}, a colon-separated list of paths
(on Windows-based systems, a semicolon-separate list of paths).

@item --CLASSPATH=@var{path}
This sets the class path to @var{path}, a colon-separated list of paths
(on Windows-based systems, a semicolon-separate list of paths).  This
differs from the @code{--classpath} option in that it also suppresses
the built-in system path.

@item CLASSPATH
This is an environment variable which holds a list of paths.
@end table

The final class path is constructed like so:

@itemize @bullet
@item
First come all directories specified via @code{-I}.

@item
If @code{--classpath} is specified, its value is appended and processing
stops.  That is, @code{--classpath} suppresses all the options mentioned
later in this list.

@item
If @code{--CLASSPATH} is specified, its value is appended and the
@code{CLASSPATH} environment variable is suppressed.

@item
If the @code{CLASSPATH} environment variable is specified (and was not
suppressed by @code{--CLASSPATH}), then its value is appended.

@item
Finally, the built-in system directory, @file{libgcj.jar}, is appended.
@end itemize

The classfile built by @code{gcj} for the class @code{java.lang.Object}
(and placed in @code{libgcj.jar}) contains a special zero length
attribute @code{gnu.gcj.gcj-compiled}. The compiler looks for this
attribute when loading @code{java.lang.Object} and will report an error
if it isn't found, unless it compiles to bytecode (the option
@code{-fforce-classes-archive-check} can be used to overide this
behavior in this particular case.)

@table @code
@item -fforce-classes-archive-check
This forces the compiler to always check for the special zero length
attribute @code{gnu.gcj.gcj-compiled} in @code{java.lang.Object} and
issue an error if it isn't found.
@end table

@node Encodings
@section Encodings

The Java programming language uses Unicode throughout.  In an effort to
integrate well with other locales, @code{gcj} allows @file{.java} files
to be written using almost any encoding.  @code{gcj} knows how to
convert these encodings into its internal encoding at compile time.

You can use the @code{--encoding=@var{NAME}} option to specify an
encoding (of a particular character set) to use for source files.  If
this is not specified, the default encoding comes from your current
locale.  If your host system has insufficient locale support, then
@code{gcj} assumes the default encoding to be the @samp{UTF-8} encoding
of Unicode.

To implement @code{--encoding}, @code{gcj} simply uses the host
platform's @code{iconv} conversion routine.  This means that in practice
@code{gcj} is limited by the capabilities of the host platform.

The names allowed for the argument @code{--encoding} vary from platform
to platform (since they are not standardized anywhere).  However,
@code{gcj} implements the encoding named @samp{UTF-8} internally, so if
you choose to use this for your source files you can be assured that it
will work on every host.


@node Warnings
@section Warnings

@code{gcj} implements several warnings.  As with other generic
@code{gcc} warnings, if an option of the form @code{-Wfoo} enables a
warning, then @code{-Wno-foo} will disable it.  Here we've chosen to
document the form of the warning which will have an effect -- the
default being the opposite of what is listed.

@table @code
@item -Wredundant-modifiers
With this flag, @code{gcj} will warn about redundant modifiers.  For
instance, it will warn if an interface method is declared @code{public}.

@item -Wextraneous-semicolon
This causes @code{gcj} to warn about empty statements.  Empty statements
have been deprecated.

@item -Wno-out-of-date
This option will cause @code{gcj} not to warn when a source file is
newer than its matching class file.  By default @code{gcj} will warn
about this.

@item -Wunused
This is the same as @code{gcc}'s @code{-Wunused}.

@item -Wall
This is the same as @code{-Wredundant-modifiers -Wextraneous-semicolon
-Wunused}.
@end table


@node Code Generation
@section Code Generation

In addition to the many @code{gcc} options controlling code generation,
@code{gcj} has several options specific to itself.

@table @code
@item --main=@var{CLASSNAME}
This option is used when linking to specify the name of the class whose
@code{main} method should be invoked when the resulting executable is
run.  @footnote{The linker by default looks for a global function named
@code{main}.  Since Java does not have global functions, and a
collection of Java classes may have more than one class with a
@code{main} method, you need to let the linker know which of those
@code{main} methods it should invoke when starting the application.}

@item -D@var{name}[=@var{value}]
This option can only be used with @code{--main}.  It defines a system
property named @var{name} with value @var{value}.  If @var{value} is not
specified then it defaults to the empty string.  These system properties
are initialized at the program's startup and can be retrieved at runtime
using the @code{java.lang.System.getProperty} method.

@item -C
This option is used to tell @code{gcj} to generate bytecode
(@file{.class} files) rather than object code.

@item -d @var{directory}
When used with @code{-C}, this causes all generated @file{.class} files
to be put in the appropriate subdirectory of @var{directory}.  By
default they will be put in subdirectories of the current working
directory.

@item -fno-bounds-check
By default, @code{gcj} generates code which checks the bounds of all
array indexing operations.  With this option, these checks are omitted.
Note that this can result in unpredictable behavior if the code in
question actually does violate array bounds constraints.

@item -fjni
With @code{gcj} there are two options for writing native methods: CNI
and JNI.  By default @code{gcj} assumes you are using CNI.  If you are
compiling a class with native methods, and these methods are implemented
using JNI, then you must use @code{-fjni}.  This option causes
@code{gcj} to generate stubs which will invoke the underlying JNI
methods.

@item -fno-optimize-static-class-initialization
When the optimization level is greather or equal to @code{-O2},
@code{gcj} will try to optimize the way calls into the runtime are made
to initialize static classes upon their first use (this optimization
isn't carried out if @code{-C} was specified.) When compiling to native
code, @code{-fno-optimize-static-class-initialization} will turn this
optimization off, regardless of the optimization level in use.
@end table


@node Configure-time Options
@section Configure-time Options

Some @code{gcj} code generations options affect the resulting ABI, and
so can only be meaningfully given when @code{libgcj}, the runtime
package, is configured.  @code{libgcj} puts the appropriate options from
this group into a @samp{spec} file which is read by @code{gcj}.  These
options are listed here for completeness; if you are using @code{libgcj}
then you won't want to touch these options.

@table @code
@item -fuse-boehm-gc
This enables the use of the Boehm GC bitmap marking code.  In particular
this causes @code{gcj} to put an object marking descriptor into each
vtable.

@item -fhash-synchronization
By default, synchronization data (the data used for @code{synchronize},
@code{wait}, and @code{notify}) is pointed to by a word in each object.
With this option @code{gcj} assumes that this information is stored in a
hash table and not in the object itself.

@item -fuse-divide-subroutine
On some systems, a library routine is called to perform integer
division.  This is required to get exception handling correct when
dividing by zero.

@item -fcheck-references
On some systems it's necessary to insert inline checks whenever
accessing an object via a reference.  On other systems you won't need
this because null pointer accesses are caught automatically by the
processor.
@end table


@node Compatibility
@chapter Compatibility with the Java Platform

As we believe it is important that the Java platform not be fragmented,
@code{gcj} and @code{libgcj} try to conform to the relevant Java
specifications.  However, limited manpower and incomplete and unclear
documentation work against us.  So, there are caveats to using
@code{gcj}.

This list of compatibility issues is by no means complete.

@itemize @bullet
@item
@code{gcj} implements the JDK 1.1 language.  It supports inner classes,
though these are known to still be buggy.  It does not yet support the
Java 2 @code{strictfp} keyword (it recognizes the keyword but ignores
it).

@item
@code{libgcj} is largely compatible with the JDK 1.2 libraries.
However, @code{libgcj} is missing many packages, most notably
@code{java.awt}.  There are also individual missing classes and methods.
We currently do not have a list showing differences between
@code{libgcj} and the Java 2 platform.

@item
Sometimes the @code{libgcj} implementation of a method or class differs
from the JDK implementation.  This is not always a bug.  Still, if it
affects you, it probably makes sense to report it so that we can discuss
the appropriate response.
@end itemize


@node Invoking gcjh
@chapter Invoking gcjh

The @code{gcjh} program is used to generate header files from class
files.  It can generate both CNI and JNI header files, as well as stub
implementation files which can be used as a basis for implementing the
required native methods.

@table @code
@item -stubs
This causes @code{gcjh} to generate stub files instead of header files.
By default the stub file will be named after the class, with a suffix of
@samp{.cc}.  In JNI mode, the default output file will have the suffix
@samp{.c}.

@item -jni
This tells @code{gcjh} to generate a JNI header or stub.  By default,
CNI headers are generated.

@item -add @var{text}
Inserts @var{text} into the class body.  This is ignored in JNI mode.

@item -append @var{text}
Inserts @var{text} into the header file after the class declaration.
This is ignored in JNI mode.

@item -friend @var{text}
Inserts @var{text} into the class as a @code{friend} declaration.
This is ignored in JNI mode.

@item -prepend @var{text}
Inserts @var{text} into the header file before the class declaration.
This is ignored in JNI mode.

@item --classpath=@var{path}
@itemx --CLASSPATH=@var{path}
@itemx -I@var{directory}
@itemx -d @var{directory}
@itemx -o @var{file}
These options are all identical to the corresponding @code{gcj} options.

@item -o @var{file}
Sets the output file name.  This cannot be used if there is more than
one class on the command line.

@item -td @var{directory}
Sets the name of the directory to use for temporary files.

@item --help
Print help about @code{gcjh} and exit.  No further processing is done.

@item --version
Print version information for @code{gcjh} and exit.  No further
processing is done.
@end table

All remaining options are considered to be names of classes.


@node Invoking jv-scan
@chapter Invoking jv-scan

The @code{jv-scan} program can be used to print information about a Java
source file (@file{.java} file).

@table @code
@item --complexity
This prints a complexity measure, related to cyclomatic complexity, for
each input file.

@item --encoding=@var{name}
This works like the corresponding @code{gcj} option.

@item --print-main
This prints the name of the class in this file containing a @code{main}
method.

@item --list-class
This lists the names of all classes defined in the input files.

@item --list-filename
If @code{--list-class} is given, this option causes @code{jv-scan} to
also print the name of the file in which each class was found.

@item -o @var{file}
Print output to the named file.
@end table


@node Invoking jcf-dump
@chapter Invoking jcf-dump

This is a class file examiner, similar to @code{javap}.  It will print
information about a number of classes, which are specifed by class name
or file name.

@table @code
@item -c
Disassemble method bodies.  By default method bodies are not printed.

@item --javap
Generate output in @code{javap} format.  The implementation of this
feature is very incomplete.

@item --classpath=@var{path}
@itemx --CLASSPATH=@var{path}
@itemx -I@var{directory}
@itemx -o @var{file}
These options as the same as the corresponding @code{gcj} options.
@end table


@node Invoking gij
@chapter Invoking gij

@code{gij} is a Java bytecode interpreter included with @code{libgcj}.
@code{gij} is not available on every platform; porting it requires a
small amount of assembly programming which has not been done for all the
targets supported by @code{gcj}.

The primary argument to @code{gij} is the name of a class or, with
@code{-jar}, a jar file.  Options before this argument are interpreted
by @code{gij}; remaining options are passed to the interpreted program.

If a class name is specified and this class does not have a @code{main}
method with the appropriate signature (a @code{static void} method with
a @code{String[]} as its sole argument), then @code{gij} will print an
error and exit.

If a jar file is specified then @code{gij} will use information in it to
determine which class' @code{main} method will be invoked.

@code{gij} will invoke the @code{main} method with all the remaining
command-line options.

Note that @code{gij} is not limited to interpreting code.  Because
@code{libgcj} includes a class loader which can dynamically load shared
objects, it is possible to give @code{gij} the name of a class which has
been compiled and put into a shared library on the class path.

@table @code
@item -D@var{name}[=@var{value}]
This defines a system property named @var{name} with value @var{value}.
If @var{value} is not specified then it defaults to the empty string.
These system properties are initialized at the program's startup and can
be retrieved at runtime using the @code{java.lang.System.getProperty}
method.

@item -ms=@var{number}
This sets the initial heap size 

@item -mx=@var{number}
This sets the maximum heap size.

@item -jar
This indicates that the name passed to @code{gij} should be interpreted
as the name of a jar file, not a class.
@end table


@node Resources
@chapter Resources

While writing @code{gcj} and @code{libgcj} we have, of course, relied
heavily on documentation from Sun Microsystems.  In particular we have
used The Java Language Specification (both first and second editions),
the Java Class Libraries (volumes one and two), and the Java Virtual
Machine Specification.  In addition we've used the online documentation
at @uref{http://java.sun.com/}.

The current @code{gcj} home page is
@uref{http://gcc.gnu.org/java/}.

For more information on gcc, see @uref{http://gcc.gnu.org/}.

Some @code{libgcj} testing is done using the Mauve test suite.  This is
a free software Java class library test suite which is being written
because the JCK is not free.  See
@uref{http://sources.redhat.com/mauve/} for more information.

@bye