rs6000: use li;x?oris to build constant

[official-gcc.git] / gcc / doc / gm2.texi

\input texinfo
@c -*-texinfo-*-
@c Copyright (C) 2001-2022 Free Software Foundation, Inc.
@c This is part of the GM2 manual.

@c User level documentation for GNU Modula-2
@c
@c header

@setfilename gm2.info
@settitle The GNU Modula-2 Compiler

@set version-python  3.5

@include gcc-common.texi

@c Copyright years for this manual.
@set copyrights-gm2 1999-2022

@copying
@c man begin COPYRIGHT
Copyright @copyright{} @value{copyrights-gm2} 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 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
A copy of the license is included in the
@c man end
section entitled ``GNU Free Documentation License''.
@ignore
@c man begin COPYRIGHT
man page gfdl(7).
@c man end
@end ignore
@end copying

@ifinfo
@format
@dircategory Software development
@direntry
* gm2: (gm2).               A GCC-based compiler for the Modula-2 language
@end direntry
@end format

@insertcopying
@end ifinfo

@titlepage
@title The GNU Modula-2 Compiler
@versionsubtitle
@author Gaius Mulley

@page
@vskip 0pt plus 1filll
Published by the Free Software Foundation @*
51 Franklin Street, Fifth Floor@*
Boston, MA 02110-1301, USA@*
@sp 1
@insertcopying
@end titlepage
@contents
@page

@c `Top' Node and Master Menu

@node Top, Overview, (dir), (dir)
@top Introduction

@menu
* Overview::         What is GNU Modula-2.
* Using::            Using GNU Modula-2.
* License::          License of GNU Modula-2
* Copying::          GNU Public License V3.
* Contributing::     Contributing to GNU Modula-2
* Internals::        GNU Modula-2 internals.
* EBNF::             EBNF of GNU Modula-2
* Libraries::        PIM and ISO library definitions.
* Indices::          Document and function indices.
@end menu

@node Overview, Using, Top, Top
@chapter Overview of GNU Modula-2

@menu
* What is GNU Modula-2::  Brief description of GNU Modula-2.
* Why use GNU Modula-2::  Advantages of GNU Modula-2.
* Development::           How to get source code using git.
* Features::              GNU Modula-2 Features
@end menu

@node What is GNU Modula-2, Why use GNU Modula-2, , Using
@section What is GNU Modula-2

GNU Modula-2 is a @uref{http://gcc.gnu.org/frontends.html, front end}
for the GNU Compiler Collection (@uref{http://gcc.gnu.org/, GCC}).
The GNU Modula-2 compiler is compliant with the PIM2, PIM3, PIM4 and
ISO dialects.  Also implemented are a complete set of free ISO
libraries and PIM libraries.

@footnote{The four Modula-2 dialects supported are defined in the following
references:

PIM2: 'Programming in Modula-2', 2nd Edition, Springer Verlag, 1982,
1983 by Niklaus Wirth (PIM2).

PIM3: 'Programming in Modula-2', 3rd Corrected Edition, Springer Verlag,
1985 (PIM3).

PIM4: 'Programming in Modula-2', 4th Edition, Springer Verlag, 1988
(@uref{http://freepages.modula2.org/report4/modula-2.html, PIM4}).

ISO: the ISO Modula-2 language as defined in 'ISO/IEC Information
technology - programming languages - part 1: Modula-2 Language,
ISO/IEC 10514-1 (1996)'
}

@node Why use GNU Modula-2, Release map, What is GNU Modula-2, Using
@section Why use GNU Modula-2

There are a number of advantages of using GNU Modula-2 rather than
translate an existing project into another language.

The first advantage is of maintainability of the original sources
and the ability to debug the original project source code using a
combination of gm2 and gdb.

The second advantage is that gcc runs on many processors and
platforms.  gm2 builds and runs on powerpc64le, amd64, i386, aarch64
to name but a few processors.

gm2 can produce swig interface headers to allow access from Python and
other scripting languages.  It can also be used with C/C++ and
generate shared libraries.

The compiler provides semantic analysis and run time checking (full ISO
Modula-2 checking is implemented) and there is a plugin which can,
under certain conditions, detect run time errors at compile time.

The compiler supports PIM2, PIM3, PIM4 and ISO dialects of Modula-2,
work is underway to implement M2R10.  Many of the GCC builtins are
available and access to assembly programming is achieved using the
same syntax as that used by GCC.

The gm2 driver allows third party libraries to be installed alongside
gm2 libraries.  For example if the user specifies library @code{foo}
using @code{-flibs=foo} the driver will check the standard GCC install
directory for a sub directory @code{foo} containing the library
contents.  The library module search path is altered accordingly
for compile and link.

@node Release map, Development, Why use GNU Modula-2, Using
@section Release map

GNU Modula-2 is now part of GCC and therefore will adopt the GCC
release schedule.  It is intended that GNU Modula-2 implement more of
the GCC builtins (vararg access) and GCC features.

There is an intention to implement the ISO generics and the M2R10
dialect of Modula-2.  It will also implement all language changes.  If
you wish to see something different please email
@email{gm2@@nongnu.org} with your ideas.

@node Development, Features, Release map, Using
@section How to get source code using git

GNU Modula-2 is now in the @url{https://gcc.gnu.org/git.html, GCC git
tree}.

@node Features, Documentation, Development, Using
@section GNU Modula-2 Features

@itemize @bullet

@item
the compiler currently complies with Programming in Modula-2 Edition
2, 3, 4 and ISO Modula-2.  Users can switch on specific language
features by using: @samp{-fpim}, @samp{-fpim2}, @samp{-fpim3},
@samp{-fpim4} or @samp{-fiso}.

@item
the option @samp{-fswig} will automatically create a swig interface
file which corresponds to the definition module of the file being
compiled.

@item
exception handling is compatible with C++ and swig.  Modula-2 code can
be used with C or C++ code.

@item
Python can call GNU Modula-2 modules via swig.

@item
shared libraries can be built.

@item
fixed sized types are now available from @samp{SYSTEM}.

@c @item
@c support for dynamic @code{ARRAY}s has been added into @samp{gdb}.

@item
variables can be declared at addresses.

@item
much better dwarf-2 debugging support and when used with
@samp{gdb} the programmer can display @code{RECORD}s,
@code{ARRAY}s, @code{SET}s, subranges and constant char literals
in Modula-2 syntax.

@item
supports sets of any ordinal size (memory permitting).

@item
easy interface to C, and varargs can be passed to C routines.

@item
many Logitech libraries have been implemented and can be accessed via:
@samp{-flibs=m2log,m2pim,m2iso}.

@item
coroutines have been implemented in the PIM style and these are
accessible from SYSTEM.  A number of supporting libraries (executive
and file descriptor mapping to interrupt vector libraries are
available through the @samp{-flibs=m2iso,m2pim} switch).

@item
can be built as a cross compiler (for embedded microprocessors
such as the AVR and the ARM).

@end itemize

@node Documentation, Regression tests, Features, Using
@section Documentation

The GNU Modula-2 documentation is available on line
@url{https://www.nongnu.org/gm2/homepage.html,at the gm2 homepage}
or in the pdf, info, html file format.

@node Regression tests, Limitations, Documentation, Using
@section Regression tests for gm2 in the repository

The regression testsuite can be run from the gcc build directory:

@example
$ cd build-gcc
$ make check -j 24
@end example

which runs the complete testsuite for all compilers using 24 parallel
invocations of the compiler.  Individual language testsuites can be
run by specifying the language, for example the Modula-2 testsuite can
be run using:

@example
$ cd build-gcc
$ make check-m2 -j 24
@end example

Finally the results of the testsuite can be emailed to the
@url{https://gcc.gnu.org/lists.html, gcc-testresults} list using the
@file{test_summary} script found in the gcc source tree:

@example
$ @samp{directory to the sources}/contrib/test_summary
@end example

@node Limitations, Objectives, Regression tests, Using
@section Limitations

Logitech compatibility library is incomplete.  The principle modules
for this platform exist however for a comprehensive list of completed
modules please check the documentation
@url{gm2.html}.

@node Objectives, FAQ, , Using
@section Objectives

@itemize @bullet

@item
The intention of GNU Modula-2 is to provide a production Modula-2
front end to GCC.

@item
It should support all Niklaus Wirth PIM Dialects [234] and also ISO
Modula-2 including a re-implementation of all the ISO modules.

@item
There should be an easy interface to C.

@item
Exploit the features of GCC.

@item
Listen to the requests of the users.
@end itemize

@node FAQ, Community, Objectives, Using
@section FAQ

@subsection Why use the C++ exception mechanism in GCC, rather than a bespoke Modula-2 mechanism?

The C++ mechanism is tried and tested, it also provides GNU Modula-2
with the ability to link with C++ modules and via swig it can raise
Python exceptions.

@node Community, Other languages, FAQ, Using
@section Community

You can subscribe to the GNU Modula-2 mailing by sending an
email to:
@email{gm2-subscribe@@nongnu.org}
or by
@url{http://lists.nongnu.org/mailman/listinfo/gm2}.
The mailing list contents can be viewed
@url{http://lists.gnu.org/archive/html/gm2}.

@node Other languages, , Community, Using
@section Other languages for GCC

These exist and can be found on the frontends web page on the
@uref{http://gcc.gnu.org/frontends.html, gcc web site}.

@node Using, , Community, Top
@chapter Using GNU Modula-2

@menu
* Example usage::         Example compile and link.
* Compiler options::      GNU Modula-2 compiler options.
* Linking::               Linking options in more detail.
* Elementary data types:: Data types supported by GNU Modula-2.
* Standard procedures::   Permanently accessible base procedures.
* Dialect::               GNU Modula-2 supported dialects.
* Exceptions::            Exception implementation
* Semantic checking::     How to detect run time problems at compile time.
* Extensions::            GNU Modula-2 language extensions.
* Type compatibility::    Data type compatibility.
* Unbounded by reference::Explanation of a language optimization.
* Building a shared library:: How to build a shared library.
* Interface for Python::  How to produce swig interface files.
* Producing a Python module::  How to produce a Python module.
* Interface to C::        Interfacing GNU Modula-2 to C.
* Assembly language::     Interface to assembly language.
* Alignment::             Data type alignment.
* Packed::                Packing data types.
* Built-ins::             Accessing GNU Modula-2 Built-ins.
* The PIM system module:: SYSTEM data types and procedures.
* The ISO system module:: SYSTEM data types, procedures and run time.
* Other languages::       Other languages for GCC.
* What is GNU Modula-2::  Brief description of GNU Modula-2.
* Why use GNU Modula-2::  Advantages of GNU Modula-2.
@ifnothtml
@c omit these nodes if generating gm2 webpage as these are hand written.
* Release map:: Release map.
* Development:: Development.
* Features:: Features of the implementation.
* Documentation:: Placeholder for how to access the documentation online.
* Regression tests:: How to run the testsuite.
* Limitations:: Current limitations.
* Objectives:: Objectives of the implementation.
* FAQ:: Frequently asked questions.
* Community:: How to join the community.
@end ifnothtml
@end menu

This document contains the user and design issues relevant to the
Modula-2 front end to gcc.

@node Example usage, Compiler options, Using, Using
@section Example compile and link

@ignore
@c man begin SYNOPSIS gm2
gm2 [@option{-c}|@option{-S}] [@option{-g}] [@option{-pg}]
    [@option{-O}@var{level}] [@option{-W}@var{warn}@dots{}]
    [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
    [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
    [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}

Only the most useful options are listed here; see below for the
remainder.
@c man end
@c man begin SEEALSO
gpl(7), gfdl(7), fsf-funding(7), gcc(1)
and the Info entries for @file{gm2} and @file{gcc}.
@c man end
@end ignore

@c man begin DESCRIPTION gm2

The @command{gm2} command is the GNU compiler for the Modula-2 language and
supports many of the same options as @command{gcc}.  @xref{Option Summary, ,
Option Summary, gcc, Using the GNU Compiler Collection (GCC)}.
This manual only documents the options specific to @command{gm2}.

@c man end

This section describes how to compile and link a simple hello world
program.  It provides a few examples of using the different options
mentioned in @pxref{Compiler options, , ,gm2}.  Assuming that you have
a file called @file{hello.mod} in your current directory which
contains:

@example
MODULE hello ;

FROM StrIO IMPORT WriteString, WriteLn ;

BEGIN
   WriteString ('hello world') ; WriteLn
END hello.
@end example

You can compile and link it by: @samp{gm2 -g hello.mod}.
The result will be an @samp{a.out} file created in your directory.

You can split this command into two steps if you prefer.  The compile
step can be achieved by: @samp{gm2 -g -c -fscaffold-main hello.mod}
and the link via: @samp{gm2 -g hello.o}.

@footnote{To see all the compile actions taken by @samp{gm2} users can also
add the @samp{-v} flag at the command line, for example:

@samp{gm2 -v -g -I. hello.mod}

This displays the sub processes initiated by @samp{gm2} which can be useful
when trouble shooting.}

@node Compiler options, Elementary data types, Example usage, Using
@section Compiler options

This section describes the compiler options specific to GNU Modula-2
for generic flags details @xref{Invoking GCC, , ,gcc}.

@c man begin OPTIONS

For any given input file, the file name suffix determines what kind of
compilation is done.  The following kinds of input file names are supported:

@table @gcctabopt
@item @var{file}.mod
Modula-2 implementation or program source files.  See the
@samp{-fmod=} option if you wish to compile a project which uses a
different source file extension.
@item @var{file}.def
Modula-2 definition module source files.  Definition modules are not
compiled separately, in GNU Modula-2 definition modules are parsed as
required when program or implementation modules are compiled.  See the
@samp{-fdef=} option if you wish to compile a project which uses a
different source file extension.
@end table

You can specify more than one input file on the @command{gm2} command line,

@table @code

@item -g
create debugging information so that debuggers such as @file{gdb}
can inspect and control executable.

@item -I
used to specify the search path for definition and implementation
modules.  An example is:  @code{gm2 -g -c -I.:../../libs foo.mod}.
If this option is not specified then the default path is added
which consists of the current directory followed by the appropriate
language dialect library directories.

@c ordered list of options from here.

@item -fauto-init
turns on auto initialization of pointers to NIL.  Whenever a block is
created all pointers declared within this scope will have their
addresses assigned to NIL.

@item -fbounds
turns on run time subrange, array index and indirection via @code{NIL}
pointer checking.

@item -fcase
turns on compile time checking to check whether a @code{CASE}
statement requires an @code{ELSE} clause when on was not specified.

@item -fcpp
preprocess the source with @samp{cpp -lang-asm -traditional-cpp}
For further details about these options @xref{Invocation, , ,cpp}.
If @samp{-fcpp} is supplied then all definition modules and
implementation modules which are parsed will be prepossessed by
@samp{cpp}.

@c fcpp-end
@c Modula-2
@c passed to the preprocessor if -fcpp is used (internal switch)

@c fcpp-begin
@c Modula-2
@c passed to the preprocessor if -fcpp is used (internal switch)

@item -fdebug-builtins
call a real function, rather than the builtin equivalent.  This can
be useful for debugging parameter values to a builtin function as
it allows users to single step code into a real function.

@c fd
@c Modula-2
@c turn on internal debugging of the compiler (internal switch)

@c fdebug-trace-quad
@c Modula-2
@c turn on quadruple tracing (internal switch)

@c fdebug-trace-api
@c Modula-2
@c turn on the Modula-2 api tracing (internal switch)

@c fdebug-function-line-numbers
@c Modula-2
@c turn on the Modula-2 function line number generation (internal switch)

@item -fdef=
recognize the specified suffix as a definition module filename.
The default implementation and module filename suffix is @file{.def}.
If this option is used GNU Modula-2 will still fall back to this
default if a requested definition module is not found.

@item -fdump-system-exports
display all inbuilt system items.
This is an internal command line option.

@item -fexceptions
turn on exception handling code.  By default this option is on.
Exception handling can be disabled by @samp{-fno-exceptions}
and no references are made to the run time exception libraries.

@item -fextended-opaque
allows opaque types to be implemented as any type.  This is a GNU
Modula-2 extension and it requires that the implementation module
defining the opaque type is available so that it can be resolved when
compiling the module which imports the opaque type.

@item -ffloatvalue
turns on run time checking to check whether a floating point number is
about to exceed range.

@item -fgen-module-list=@file{filename}
attempt to find all modules when linking and generate a module list.
If the @file{filename} is @samp{-} then the contents are not written
and only used to force the linking of all module ctors.
This option cannot be used if @samp{-fuse-list=} is enabled.

@item -findex
generate code to check whether array index values are out of bounds.
Array index checking can be disabled via @samp{-fno-index}.

@item -fiso
turn on ISO standard features.  Currently this enables the ISO
@code{SYSTEM} module and alters the default library search path so
that the ISO libraries are searched before the PIM libraries.  It also
effects the behavior of @code{DIV} and @code{MOD} operators.
@xref{Dialect, , ,gm2}.

@item -flibs=
modifies the default library search path.  The libraries supplied are:
m2pim, m2iso, m2min, m2log and m2cor.  These map onto the
Programming in Modula-2 base libraries, ISO standard libraries, minimal
library support, Logitech compatible library and Programming in
Modula-2 with coroutines.
Multiple libraries can be specified and are comma separated with precedence
going to the first in the list.  It is not necessary to use -flibs=m2pim or
-flibs=m2iso if you also specify -fpim, -fpim2, -fpim3, -fpim4 or
-fiso.  Unless you are using -flibs=m2min you should include m2pim as
the they provide the base modules which all other dialects utilize.
The option @samp{-fno-libs=-} disables the @samp{gm2} driver from
modifying the search and library paths.

@c flocation=
@c Modula-2 Joined
@c set all location values to a specific value (internal switch)

@item -fm2-g
improve the debugging experience for new programmers at the expense
of generating @code{nop} instructions if necessary to ensure single
stepping precision over all code related keywords.  An example
of this is in termination of a list of nested @code{IF} statements
where multiple @code{END} keywords are mapped onto a sequence of
@code{nop} instructions.

@item -fm2-lower-case
render keywords in error messages using lower case.

@item -fm2-plugin
insert plugin to identify run time errors at compile time (default on).

@item -fm2-statistics
generates quadruple information: number of quadruples generated,
number of quadruples remaining after optimization and number of source
lines compiled.

@item -fm2-strict-type
experimental flag to turn on the new strict type checker.

@item -fm2-whole-program
compile all implementation modules and program module at once.  Notice
that you need to take care if you are compiling different dialect
modules (particularly with the negative operands to modulus).  But
this option, when coupled together with @code{-O3}, can deliver huge
performance improvements.

@item -fmod=
recognize the specified suffix as implementation and module filenames.
The default implementation and module filename suffix is @file{.mod}.
If this option is used GNU Modula-2 will still fall back to this
default if it needs to read an implementation module and the specified
suffixed filename does not exist.

@item -fnil
generate code to detect accessing data through a @code{NIL} value
pointer.  Dereferencing checking through a @code{NIL} pointer can be
disabled by @samp{-fno-nil}.

@item -fpim
turn on PIM standard features.  Currently this enables the PIM
@code{SYSTEM} module and determines which identifiers are pervasive
(declared in the base module).  If no other @samp{-fpim[234]} switch is
used then division and modulus operators behave as defined in PIM4.
@xref{Dialect, , ,gm2}.

@item -fpim2
turn on PIM-2 standard features.  Currently this removes @code{SIZE}
from being a pervasive identifier (declared in the base module).  It
places @code{SIZE} in the @code{SYSTEM} module.  It also effects the
behavior of @code{DIV} and @code{MOD} operators.
@xref{Dialect, , ,gm2}.

@item -fpim3
turn on PIM-3 standard features.  Currently this only effects the
behavior of @code{DIV} and @code{MOD} operators.
@xref{Dialect, , ,gm2}.

@item -fpim4
turn on PIM-4 standard features.  Currently this only effects the
behavior of @code{DIV} and @code{MOD} operators.
@xref{Dialect, , ,gm2}.

@item -fpositive-mod-floor-div
forces the @code{DIV} and @code{MOD} operators to behave as defined by PIM4.
All modulus results are positive and the results from the division are
rounded to the floor.
@xref{Dialect, , ,gm2}.

@item -fpthread
link against the pthread library.  By default this option is on.  It
can be disabled by @samp{-fno-pthread}.  GNU Modula-2 uses the GCC
pthread libraries to implement coroutines (see the SYSTEM
implementation module).

@c -fq
@c -Modula-2
@c -internal compiler debugging information, dump the list of quadruples

@item -frange
generate code to check the assignment range, return value range
set range and constructor range.  Range checking can be disabled
via @samp{-fno-range}.

@item -freturn
generate code to check that functions always exit with a @code{RETURN}
and do not fall out at the end.  Return checking can be disabled
via @samp{-fno-return}.

@item -fruntime-modules=
specify, using a comma separated list, the run time modules and their
order.  These modules will initialized first before any other modules
in the application dependency.  By default the run time modules list is
set to @code{Storage,SYSTEM,M2RTS,RTExceptions,IOLink}.  Note that
these modules will only be linked into your executable if they are
required.  So adding a long list of dependent modules will not effect
the size of the executable it merely states the initialization order
should they be required.

@item -fscaffold-dynamic
the option ensures that @samp{gm2} will generate a dynamic scaffold
infrastructure when compiling implementation and program modules.
By default this option is on.  Use @samp{-fno-scaffold-dynamic}
to turn it off or select @samp{-fno-scaffold-static}.

@item -fscaffold-c
generate a C source scaffold for the current module being compiled.

@item -fscaffold-c++
generate a C++ source scaffold for the current module being compiled.

@item -fscaffold-main
force the generation of the @samp{main} function.  This is not
necessary if the @samp{-c} is omitted.

@item -fscaffold-static
the option ensures that @samp{gm2} will generate a static scaffold
within the program module.  The static scaffold consists of sequences
of calls to all dependent module initialization and finalization
procedures.  The static scaffold is useful for debugging and single
stepping the initialization blocks of implementation modules.

@item -fshared
generate a shared library from the module.

@item -fsoft-check-all
turns on all run time checks.  This is the same as invoking
GNU Modula-2 using the command options
@code{-fnil} @code{-frange} @code{-findex}
@code{-fwholevalue}
@code{-fwholediv} @code{-fcase} @code{-freturn}.

@item -fsources
displays the path to the source of each module.  This option
can be used at compile time to check the correct definition module
is being used.

@item -fswig
generate a swig interface file.

@item -funbounded-by-reference
enable optimization of unbounded parameters by attempting to pass non
@code{VAR} unbounded parameters by reference.  This optimization
avoids the implicit copy inside the callee procedure.  GNU Modula-2
will only allow unbounded parameters to be passed by reference if,
inside the callee procedure, they are not written to, no address is
calculated on the array and it is not passed as a @code{VAR}
parameter.  Note that it is possible to write code to break this
optimization, therefore this option should be used carefully.
For example it would be possible to take the address of an array, pass
the address and the array to a procedure, read from the array in
the procedure and write to the location using the address parameter.

Due to the dangerous nature of this option it is not enabled
when the @samp{-O} option is specified.

@item -fuse-list=@file{filename}
if @samp{-fscaffold-static} is enabled then use the file
@file{filename} for the initialization order of modules.  Whereas if
@samp{-fscaffold-dynamic} is enabled then use this file to force
linking of all module ctors.
This option cannot be used if @samp{-fgen-module-list=} is enabled.

@item -fwholediv
generate code to detect whole number division by zero or modulus by
zero.

@item -fwholevalue
generate code to detect whole number overflow and underflow.

@c the following warning options are complete but need to be
@c regression tested against all other front ends
@c to ensure the options do not conflict.

@c @item -Wall
@c turn on all Modula-2 warnings.

@c @item -Wpedantic
@c forces the compiler to reject nested @code{WITH} statements
@c referencing the same record type.  Does not allow multiple imports of
@c the same item from a module.  It also checks that: procedure variables
@c are written to before being read; variables are not only written to
@c but read from; variables are declared and used.  If the compiler
@c encounters a variable being read before written it will terminate with
@c a message.  It will check that @code{FOR} loop indices are not used
@c outside the end of this loop without being reset.

@c @item -Wpedantic-cast
@c warns if the ISO system function is used and if the size of
@c the variable is different from that of the type.  This is legal
@c in ISO Modula-2, however it can be dangerous.  Some users may prefer
@c to use @code{VAL} instead in these situations and use @code{CAST}
@c exclusively for changes in type on objects which have the same size.

@c @item -Wpedantic-param-names
@c procedure parameter names are checked in the definition module
@c against their implementation module counterpart.  This is not
@c necessary in ISO or PIM versions of Modula-2.

@c @item -Wstyle
@c checks for poor programming style.  This option is aimed at new users of
@c Modula-2 in that it checks for situations which might cause confusion
@c and thus mistakes.  It checks whether variables of the same name are
@c declared in different scopes and whether variables look like keywords.
@c Experienced users might find this option too aggressive.

@c @item -Wunused-variable
@c warns if a variable has been declared and it not used.

@c @item -Wunused-parameter
@c warns if a parameter has been declared and it not used.

@c @item -Wverbose-unbounded
@c inform the user which non @code{VAR} unbounded parameters will be
@c passed by reference.  This only produces output if the option
@c @samp{-funbounded-by-reference} is also supplied on the command line.

@end table

@c man end

@node Elementary data types, Standard procedures, Compiler options, Using
@section Elementary data types

This section describes the elementary data types supported by GNU
Modula-2.  It also describes the relationship between these data types
and the equivalent C data types.

The following data types are supported: @code{INTEGER},
@code{LONGINT}, @code{SHORTINT}, @code{CARDINAL}, @code{LONGCARD},
@code{SHORTCARD}, @code{BOOLEAN}, @code{REAL}, @code{LONGREAL},
@code{SHORTREAL}, @code{COMPLEX}, @code{LONGCOMPLEX},
@code{SHORTCOMPLEX} and @code{CHAR}.

An equivalence table is given below:

@example
GNU Modula-2              GNU C
======================================
INTEGER                   int
LONGINT                   long long int
SHORTINT                  short int
CARDINAL                  unsigned int
LONGCARD                  long long unsigned int
SHORTCARD                 short unsigned int
BOOLEAN                   int
REAL                      double
LONGREAL                  long double
SHORTREAL                 float
CHAR                      char
SHORTCOMPLEX              complex float
COMPLEX                   complex double
LONGCOMPLEX               complex long double
@end example

Note that GNU Modula-2 also supports fixed sized data types which are
exported from the @code{SYSTEM} module.
@xref{The PIM system module, , ,gm2}.
@xref{The ISO system module, , ,gm2}.

@node Standard procedures, Dialect, Elementary data types, Using
@section Permanently accessible base procedures.

This section describes the procedures and functions which are
always visible.

@subsection Standard procedures and functions common to PIM and ISO

The following procedures are implemented and conform with Programming
in Modula-2 and ISO Modula-2: @code{NEW}, @code{DISPOSE}, @code{INC},
@code{DEC}, @code{INCL}, @code{EXCL} and @code{HALT}.  The standard
functions are: @code{ABS}, @code{CAP}, @code{CHR}, @code{FLOAT},
@code{HIGH}, @code{LFLOAT}, @code{LTRUNC}, @code{MIN}, @code{MAX},
@code{ODD}, @code{SFLOAT}, @code{STRUNC} @code{TRUNC} and
@code{VAL}.  All these functions and procedures (except @code{HALT},
@code{NEW}, @code{DISPOSE} and, under non constant conditions,
@code{LENGTH}) generate in-line code for efficiency.

@example

(*
   ABS - returns the positive value of i.
*)

@findex ABS
PROCEDURE ABS (i: <any signed type>) : <any signed type> ;

@end example

@example

(*
   CAP - returns the capital of character ch providing
         ch lies within the range 'a'..'z'.  Otherwise ch
         is returned unaltered.
*)

@findex CAP
PROCEDURE CAP (ch: CHAR) : CHAR ;

@end example

@example

(*
   CHR - converts a value of a <whole number type> into a CHAR.
         CHR(x) is shorthand for VAL(CHAR, x).
*)

@findex CHR
PROCEDURE CHR (x: <whole number type>) : CHAR ;

@end example

@example

(*
   DISPOSE - the procedure DISPOSE is replaced by:
             DEALLOCATE(p, TSIZE(p^)) ;
             The user is expected to import the procedure DEALLOCATE
             (normally found in the module, Storage.)

             In:  a variable p: of any pointer type which has been
                  initialized by a call to NEW.
             Out: the area of memory
                  holding p^ is returned to the system.
                  Note that the underlying procedure DEALLOCATE
                  procedure in module Storage will assign p to NIL.
*)

@findex DISPOSE
PROCEDURE DISPOSE (VAR p:<any pointer type>) ;
@end example

@example

(*
   DEC - can either take one or two parameters.  If supplied
         with one parameter then on the completion of the call to
         DEC, v will have its predecessor value.  If two
         parameters are supplied then the value v will have its
         n'th predecessor.  For these reasons the value of n
         must be >=0.
*)

@findex DEC
PROCEDURE DEC (VAR v: <any base type>; [n: <any base type> = 1]) ;
@end example

@example

(*
   EXCL - excludes bit element e from a set type s.
*)

@findex EXCL
PROCEDURE EXCL (VAR s: <any set type>; e: <element of set type s>) ;
@end example

@example

(*
   FLOAT - will return a REAL number whose value is the same as o.
*)

@findex FLOAT
PROCEDURE FLOAT (o: <any whole number type>) : REAL ;
@end example

@example

(*
   FLOATS - will return a SHORTREAL number whose value is the same as o.
*)

@findex FLOATS
PROCEDURE FLOATS (o: <any whole number type>) : REAL ;
@end example

@example

(*
   FLOATL - will return a LONGREAL number whose value is the same as o.
*)

@findex FLOATL
PROCEDURE FLOATL (o: <any whole number type>) : REAL ;
@end example

@example

(*
   HALT - will call the HALT procedure inside the module M2RTS.
          Users can replace M2RTS.
*)

@findex HALT
PROCEDURE HALT ;
@end example

@example

(*
   HIGH - returns the last accessible index of an parameter declared as
          ARRAY OF CHAR.  Thus

          PROCEDURE foo (a: ARRAY OF CHAR) ;
          VAR
             c: CARDINAL ;
          BEGIN
             c := HIGH(a)
          END foo ;

          BEGIN
             foo('hello')
          END

          will cause the local variable c to contain the value 4
*)

@findex HIGH
PROCEDURE HIGH (a: ARRAY OF CHAR) : CARDINAL ;
@end example

@example

(*
   INC - can either take one or two parameters.  If supplied
         with one parameter then on the completion of the call to
         INC, v will have its successor value.  If two
         parameters are supplied then the value v will have its
         n'th successor.  For these reasons the value of n
         must be >=0.
*)

@findex INC
PROCEDURE INC (VAR v: <any base type>; [n: <any base type> = 1]) ;
@end example

@example

(*
   INCL - includes bit element e to a set type s.
*)

@findex INCL
PROCEDURE INCL (VAR s: <any set type>; e: <element of set type s>) ;
@end example

@example

(*
   LFLOAT - will return a LONGREAL number whose value is the same as o.
*)

@findex LFLOAT
PROCEDURE LFLOAT (o: <any whole number type>) : LONGREAL ;
@end example

@example

(*
   LTRUNC - will return a LONG<type> number whose value is the
            same as o.  PIM2, PIM3 and ISO Modula-2 will return
            a LONGCARD whereas PIM4 returns LONGINT.
*)

@findex LTRUNC
PROCEDURE LTRUNC (o: <any floating point type>) : LONG<type> ;
@end example

@example

(*
   MIN - returns the lowest legal value of an ordinal type.
*)

@findex MIN
PROCEDURE MIN (t: <ordinal type>) : <ordinal type> ;

@end example

@example

(*
   MAX - returns the largest legal value of an ordinal type.
*)

@findex MAX
PROCEDURE MAX (t: <ordinal type>) : <ordinal type> ;

@end example

@example

(*
   NEW - the procedure NEW is replaced by:
         ALLOCATE(p, TSIZE(p^)) ;
         The user is expected to import the procedure ALLOCATE
         (normally found in the module, Storage.)

         In:  a variable p: of any pointer type.
         Out: variable p is set to some allocated memory
              which is large enough to hold all the contents of p^.
*)

@findex NEW
PROCEDURE NEW (VAR p:<any pointer type>) ;
@end example

@example

(*
   ODD - returns TRUE if the value is not divisible by 2.
*)

@findex ODD
PROCEDURE ODD (x: <whole number type>) : BOOLEAN ;

@end example

@example

(*
   SFLOAT - will return a SHORTREAL number whose value is the same
            as o.
*)

@findex SFLOAT
PROCEDURE SFLOAT (o: <any whole number type>) : SHORTREAL ;
@end example

@example

(*
   STRUNC - will return a SHORT<type> number whose value is the same
            as o.  PIM2, PIM3 and ISO Modula-2 will return a
            SHORTCARD whereas PIM4 returns SHORTINT.
*)

@findex STRUNC
PROCEDURE STRUNC (o: <any floating point type>) : SHORT<type> ;
@end example

@example

(*
   TRUNC - will return a <type> number whose value is the same as o.
           PIM2, PIM3 and ISO Modula-2 will return a CARDINAL
           whereas PIM4 returns INTEGER.
*)

@findex TRUNC
PROCEDURE TRUNC (o: <any floating point type>) : <type> ;
@end example

@example

(*
   TRUNCS - will return a <type> number whose value is the same
            as o.  PIM2, PIM3 and ISO Modula-2 will return a
            SHORTCARD whereas PIM4 returns SHORTINT.
*)

@findex TRUNCS
PROCEDURE TRUNCS (o: <any floating point type>) : <type> ;
@end example

@example

(*
   TRUNCL - will return a <type> number whose value is the same
            as o.  PIM2, PIM3 and ISO Modula-2 will return a
            LONGCARD whereas PIM4 returns LONGINT.
*)

@findex TRUNCL
PROCEDURE TRUNCL (o: <any floating point type>) : <type> ;
@end example

@example

(*
   VAL - converts data i of <any simple data type 2> to
         <any simple data type 1> and returns this value.
         No range checking is performed during this conversion.
*)

@findex VAL
PROCEDURE VAL (<any simple data type 1>,
               i: <any simple data type 2>) : <any simple data type 1> ;

@end example

@subsection ISO specific standard procedures and functions

The standard function @code{LENGTH} is specific to ISO Modula-2 and
is defined as:

@example

(*
   IM - returns the imaginary component of a complex type.
        The return value will the same type as the imaginary field
        within the complex type.
*)

@findex IM
PROCEDURE IM (c: <any complex type>) : <floating point type> ;
@end example

@example

(*
   INT - returns an INTEGER value which has the same value as v.
         This function is equivalent to: VAL(INTEGER, v).
*)

@findex INT
PROCEDURE INT (v: <any ordinal type>) : INTEGER ;
@end example

@example

(*
   LENGTH - returns the length of string a.
*)

@findex LENGTH
PROCEDURE LENGTH (a: ARRAY OF CHAR) : CARDINAL ;
@end example

This function is evaluated at compile time, providing that string
@code{a} is a constant.  If @code{a} cannot be evaluated then a call is
made to @code{M2RTS.Length}.

@example

(*
   ODD - returns a BOOLEAN indicating whether the whole number
         value, v, is odd.
*)

@findex ODD
PROCEDURE ODD (v: <any whole number type>) : BOOLEAN ;
@end example

@example

(*
   RE - returns the real component of a complex type.
        The return value will the same type as the real field
        within the complex type.
*)

@findex RE
PROCEDURE RE (c: <any complex type>) : <floating point type> ;
@end example

@node Dialect, Exceptions, Standard procedures, Using
@section GNU Modula-2 supported dialects

This section describes the dialects understood by GNU Modula-2.
It also describes the differences between the dialects and
any command line switches which determine dialect behaviour.

The GNU Modula-2 compiler is compliant with four dialects of Modula-2.
The language as defined in 'Programming in Modula-2' 2nd Edition,
Springer Verlag, 1982, 1983 by Niklaus Wirth (PIM2), 'Programming in
Modula-2', 3rd Corrected Edition, Springer Verlag, 1985 (PIM3) and
'Programming in Modula-2', 4th Edition, Springer Verlag, 1988 (PIM4)
@uref{http://freepages.modula2.org/report4/modula-2.html} and the ISO
Modula-2 language as defined in ISO/IEC Information technology -
programming languages - part 1: Modula-2 Language, ISO/IEC 10514-1
(1996) (ISO).

The command line switches @samp{-fpim2}, @samp{-fpim3}, @samp{-fpim4}
and @samp{-fiso} can be used to force mutually exclusive
features.  However by default the compiler will not aggressively fail
if a non mutually exclusive feature is used from another dialect.  For
example it is possible to specify @samp{-fpim2} and still utilize
@samp{DEFINITION} @samp{MODULES} which have no export list.

Some dialect differences will force a compile time error, for example
in PIM2 the user must @code{IMPORT} @code{SIZE} from the module
@code{SYSTEM}, whereas in PIM3 and PIM4 @code{SIZE} is a pervasive
function.  Thus compiling PIM4 source code with the @samp{-fpim2}
switch will cause a compile time error.  This can be fixed quickly
with an additional @code{IMPORT} or alternatively by compiling with
the @samp{-fpim4} switch.

However there are some very important differences between the dialects
which are mutually exclusive and therefore it is vital that users
choose the dialects with care when these language features are used.

@subsection Integer division, remainder and modulus

The most dangerous set of mutually exclusive features found in the
four dialects supported by GNU Modula-2 are the @code{INTEGER}
division, remainder and modulus arithmetic operators.  It is important
to note that the same source code can be compiled to give different
run time results depending upon these switches!  The reference manual
for the various dialects of Modula-2 are quite clear about this
behavior and sadly there are three distinct definitions.

The table below illustrates the problem when a negative operand is
used.

@example
                  Pim2/3          Pim4                ISO
               -----------    -----------    ----------------------
lval    rval   DIV     MOD    DIV     MOD    DIV    MOD    /    REM
 31      10      3       1      3       1      3      1     3     1
-31      10     -3      -1     -4       9     -4      9    -3    -1
 31     -10     -3       1     -3       1     Exception    -3     1
-31     -10      3      -1      4       9     Exception     3    -1
@end example

See also P24 of PIM2, P27 of PIM3, P29 of PIM4 and P201 of the ISO
Standard.  At present all dialect division, remainder and modulus are
implemented as above, apart from the exception calling in the ISO
dialect.  Instead of exception handling the results are the same as the
PIM4 dialect.  This is a temporary implementation situation.

@node Exceptions, Semantic checking, Dialect, Using
@section Exception implementation

This section describes how exceptions are implemented in GNU Modula-2
and how command line switches affect their behavior.  The option
@samp{-fsoft-check-all} enables all software checking of nil
dereferences, division by zero etc.  Additional code is produced to
check these conditions and exception handlers are invoked if the
conditions prevail.

Without @samp{-fsoft-check-all} these exceptions will be caught by
hardware (assuming the hardware support exists) and a signal handler
is invoked.  The signal handler will in turn @code{THROW} an exception
which will be caught by the appropriate Modula-2 handler.  However the
action of throwing an exception from within a signal handler is
implementation defined (according to the C++ documentation).  For
example on the x86_64 architecture this works whereas on the i686
architecture it does not.  Therefore to ensure portability it is
recommended to use @samp{-fsoft-check-all}.

@footnote{@samp{-fsoft-check-all} can be effectively combined with
@samp{-O2} to semantically analyze source code for possible run time
errors at compile time.}

@node Semantic checking, Extensions, Exceptions, Using
@section How to detect run time problems at compile time

Consider the following program:

@example
MODULE assignvalue ;  (*!m2iso+gm2*)

PROCEDURE bad () : INTEGER ;
VAR
   i: INTEGER ;
BEGIN
   i := -1 ;
   RETURN i
END bad ;

VAR
   foo: CARDINAL ;
BEGIN
   (* The m2rte plugin will detect this as an error, post
      optimization.  *)
   foo := bad ()
END assignvalue.
@end example

here we see that the programmer has overlooked that the return value
from @samp{bad} will cause an overflow to @samp{foo}.  If we compile
the code with the following options:

@example
$ gm2 -g -fsoft-check-all -O2 -c assignvalue.mod
assignvalue.mod:16:0:inevitable that this error will occur at run time,
assignment will result in an overflow
@end example

The gm2 semantic plugin is automatically run and will generate a
warning message for every exception call which is known as reachable.
It is highly advised to run the optimizer (@samp{-O2} or @samp{-O3})
with @samp{-fsoft-check-all} so that the compiler is able to run the
optimizer and perform variable and flow analysis before the semantic
plugin is invoked.

@node Extensions, Type compatibility, Semantic checking, Using
@section GNU Modula-2 language extensions

This section introduces the GNU Modula-2 language extensions.
The GNU Modula-2 compiler allows abstract data types to be any type,
not just restricted to a pointer type providing the
@samp{-fextended-opaque} option is supplied
@xref{Compiler options, , ,gm2}.

Declarations can be made in any order, whether they are
types, constants, procedures, nested modules or variables.
@c (@xref{Passes, , ,}.)

GNU Modula-2 also allows programmers to interface to @code{C} and
assembly language.

GNU Modula-2 provides support for the special tokens @code{__LINE__},
@code{__FILE__}, @code{__FUNCTION__} and @code{__DATE__}.  Support for
these tokens will occur even if the @samp{-fcpp} option is not
supplied.  A table of these identifiers and their data type and values
is given below:

@example
Scope       GNU Modula-2 token      Data type and example value

anywhere    __LINE__                Constant Literal compatible
                                    with CARDINAL, INTEGER and WORD.
                                    Example 1234

anywhere    __FILE__                Constant string compatible
                                    with parameter ARRAY OF CHAR or
                                    an ARRAY whose SIZE is >= string
                                    length.  Example
                                    "hello.mod"

procedure   __FUNCTION__            Constant string compatible
                                    with parameter ARRAY OF CHAR or
                                    an ARRAY whose SIZE is >= string
                                    length.  Example
                                    "calc"

module      __FUNCTION__            Example
                                    "module hello initialization"

anywhere    __DATE__                Constant string compatible
                                    with parameter ARRAY OF CHAR or
                                    an ARRAY whose SIZE is >= string
                                    length.  Example
                                    "Thu Apr 29 10:07:16 BST 2004"

anywhere   __COLUMN__               Gives a constant literal number
                                    determining the left hand column
                                    where the first _ appears in
                                    __COLUMN__.  The left most column
                                    is 1.

@end example

The preprocessor @samp{cpp} can be invoked via the @samp{-fcpp}
command line option.  This in turn invokes @samp{cpp} with the
following arguments @samp{-traditional -lang-asm}.  These options
preserve comments and all quotations.  @samp{gm2} treats a @samp{#}
character in the first column as a preprocessor directive.

For example here is a module which calls @code{FatalError}
via the macro @code{ERROR}.

@example
MODULE cpp ;

FROM SYSTEM IMPORT ADR, SIZE ;
FROM libc IMPORT exit, printf, malloc ;

PROCEDURE FatalError (a, file: ARRAY OF CHAR;
                         line: CARDINAL;
                         func: ARRAY OF CHAR) ;
BEGIN
   printf ("%s:%d:fatal error, %s, in %s\n",
            ADR (file), line, ADR (a), ADR (func)) ;
   exit (1)
END FatalError ;

#define ERROR(X)  FatalError(X, __FILE__, __LINE__, __FUNCTION__)

VAR
   pc: POINTER TO CARDINAL;
BEGIN
   pc := malloc (SIZE (CARDINAL)) ;
   IF pc = NIL
   THEN
      ERROR ('out of memory')
   END
END cpp.
@end example

Another use for the C preprocessor in Modula-2 might be to turn on
debugging code.  For example the library module
@file{FormatStrings.mod} uses procedures from @file{DynamicStrings.mod}
and to track down memory leaks it was useful to track the source file
and line where each string was created.  Here is a section of
@file{FormatStrings.mod} which shows how the debugging code was
enabled and disabled by adding @code{-fcpp} to the command line.

@example
FROM DynamicStrings IMPORT String, InitString, InitStringChar, Mark,
                           ConCat, Slice, Index, char,
                           Assign, Length, Mult, Dup, ConCatChar,
                           PushAllocation, PopAllocationExemption,
                           InitStringDB, InitStringCharStarDB,
                           InitStringCharDB, MultDB, DupDB, SliceDB ;

(*
#define InitString(X) InitStringDB(X, __FILE__, __LINE__)
#define InitStringCharStar(X) InitStringCharStarDB(X, __FILE__, \
                                                   __LINE__)
#define InitStringChar(X) InitStringCharDB(X, __FILE__, __LINE__)
#define Mult(X,Y) MultDB(X, Y, __FILE__, __LINE__)
#define Dup(X) DupDB(X, __FILE__, __LINE__)
#define Slice(X,Y,Z) SliceDB(X, Y, Z, __FILE__, __LINE__)
*)

PROCEDURE doDSdbEnter ;
BEGIN
   PushAllocation
END doDSdbEnter ;

PROCEDURE doDSdbExit (s: String) ;
BEGIN
   s := PopAllocationExemption (TRUE, s)
END doDSdbExit ;

PROCEDURE DSdbEnter ;
BEGIN
END DSdbEnter ;

PROCEDURE DSdbExit (s: String) ;
BEGIN
END DSdbExit ;

(*
#define DBsbEnter doDBsbEnter
#define DBsbExit  doDBsbExit
*)

PROCEDURE Sprintf1 (s: String; w: ARRAY OF BYTE) : String ;
BEGIN
   DSdbEnter ;
   s := FormatString (HandleEscape (s), w) ;
   DSdbExit (s) ;
   RETURN s
END Sprintf1 ;
@end example

It is worth noting that the overhead of this code once @code{-fcpp} is
not present and -O2 is used will be zero since the local empty
procedures @code{DSdbEnter} and @code{DSdbExit} will be thrown away by
the optimization passes of the GCC backend.

@subsection Optional procedure parameter

GNU Modula-2 allows the last parameter to a procedure or function
parameter to be optional.  For example in the ISO library
@file{COROUTINES.def} the procedure @code{NEWCOROUTINE} is defined as
having an optional fifth argument (@code{initProtection}) which, if
absent, is automatically replaced by @code{NIL}.

@example
@findex NEWCOROUTINE
PROCEDURE NEWCOROUTINE (procBody: PROC; workspace: SYSTEM.ADDRESS;
                        size: CARDINAL; VAR cr: COROUTINE;
                        [initProtection: PROTECTION = NIL]);

  (* Creates a new coroutine whose body is given by procBody,
     and returns the identity of the coroutine in cr.
     workspace is a pointer to the work space allocated to
     the coroutine; size specifies the size of this workspace
     in terms of SYSTEM.LOC.

     The optional fifth argument may contain a single parameter
     which specifies the initial protection level of the coroutine.
  *)
@end example

The implementation module @file{COROUTINES.mod} implements this
procedure using the following syntax:

@example
PROCEDURE NEWCOROUTINE (procBody: PROC; workspace: SYSTEM.ADDRESS;
                        size: CARDINAL; VAR cr: COROUTINE;
                        [initProtection: PROTECTION]);
BEGIN

END NEWCOROUTINE ;
@end example

Note that it is illegal for this declaration to contain an initializer
value for @code{initProtection}.  However it is necessary to surround
this parameter with the brackets @code{[} and @code{]}.  This serves to
remind the programmer that the last parameter was declared as optional
in the definition module.

Local procedures can be declared to have an optional final parameter
in which case the initializer is mandatory in the implementation or
program module.

GNU Modula-2 also provides additional fixed sized data types which
are all exported from the @code{SYSTEM} module.
@xref{The PIM system module, , ,gm2}.
@xref{The ISO system module, , ,gm2}.

@node Type compatibility, Unbounded by reference, Extensions, Using
@section Type compatibility

This section discuss the issues surrounding assignment, expression
and parameter compatibility, their effect of the additional
fixed sized datatypes and also their effect of run time checking.
The data types supported by the compiler are:

@example
GNU Modula-2              scope      switches
=============================================
INTEGER                   pervasive
LONGINT                   pervasive
SHORTINT                  pervasive
CARDINAL                  pervasive
LONGCARD                  pervasive
SHORTCARD                 pervasive
BOOLEAN                   pervasive
BITSET                    pervasive
REAL                      pervasive
LONGREAL                  pervasive
SHORTREAL                 pervasive
CHAR                      pervasive
SHORTCOMPLEX              pervasive
COMPLEX                   pervasive
LONGCOMPLEX               pervasive

LOC                       SYSTEM     -fiso
BYTE                      SYSTEM
WORD                      SYSTEM
ADDRESS                   SYSTEM

The following extensions are supported for
most architectures (please check SYSTEM.def).
=============================================
INTEGER8                  SYSTEM
INTEGER16                 SYSTEM
INTEGER32                 SYSTEM
INTEGER64                 SYSTEM
CARDINAL8                 SYSTEM
CARDINAL16                SYSTEM
CARDINAL32                SYSTEM
CARDINAL64                SYSTEM
BITSET8                   SYSTEM
BITSET16                  SYSTEM
BITSET32                  SYSTEM
WORD16                    SYSTEM
WORD32                    SYSTEM
WORD64                    SYSTEM
REAL32                    SYSTEM
REAL64                    SYSTEM
REAL96                    SYSTEM
REAL128                   SYSTEM
COMPLEX32                 SYSTEM
COMPLEX64                 SYSTEM
COMPLEX96                 SYSTEM
COMPLEX128                SYSTEM
@end example

The Modula-2 language categorizes compatibility between entities of
possibly differing types into three sub components: expressions,
assignments, and parameters.  Parameter compatibility is further
divided into two sections for pass by reference and pass by value
compatibility.

For more detail on the Modula-2 type compatibility see the Modula-2
ISO standard BS ISO/IEC 10514-1:1996 page 121-125.  For detail on the
PIM type compatibility see Programming in Modula-2 Edition 4 page 29,
(Elementary Data Types).

@subsection Expression compatibility

Modula-2 restricts the types of expressions to the same type.
Expression compatibility is a symmetric relation.

For example two sub expressions of @code{INTEGER} and @code{CARDINAL}
are not expression compatible
(@uref{http://freepages.modula2.org/report4/modula-2.html} and ISO
Modula-2).

In GNU Modula-2 this rule is also extended across all fixed sized data
types (imported from SYSTEM).

@subsection Assignment compatibility

This section discusses the assignment issues surrounding assignment
compatibility of elementary types (@code{INTEGER}, @code{CARDINAL},
@code{REAL} and @code{CHAR} for example).  The information here is
found in more detail in the Modula-2 ISO standard BS ISO/IEC
10514-1:1996 page 122.

Assignment compatibility exists between the same sized elementary
types.

Same type family of different sizes are
also compatible as long as the @code{MAX(}type@code{)} and
@code{MIN(}type@code{)} is known.  So for example this includes the
@code{INTEGER} family, @code{CARDINAL} family and the @code{REAL}
family.

The reason for this is that when the assignment is performed
the compiler will check to see that the expression (on the right of
the @code{:=}) lies within the range of the designator type (on the
left hand side of the @code{:=}).  Thus these ordinal types can be
assignment compatible.  However it does mean that @code{WORD32} is not
compatible with @code{WORD16} as @code{WORD32} does not have a minimum
or maximum value and therefore cannot be checked.  The compiler does
not know which of the two bytes from @code{WORD32} should be copied
into @code{WORD16} and which two should be ignored.  Currently the
types @code{BITSET8}, @code{BITSET16} and @code{BITSET32} are
assignment incompatible.  However this restriction maybe lifted when
further run time checking is achieved.

Modula-2 does allow @code{INTEGER} to be assignment compatible with
@code{WORD} as they are the same size.  Likewise GNU Modula-2 allows
@code{INTEGER16} to be compatible with @code{WORD16} and the same for
the other fixed sized types and their sized equivalent in either
@code{WORD}n, @code{BYTE} or @code{LOC} types.  However it prohibits
assignment between @code{WORD} and @code{WORD32} even though on many
systems these sizes will be the same.  The reasoning behind this rule
is that the extended fixed sized types are meant to be used by
applications requiring fixed sized data types and it is more portable
to forbid the blurring of the boundaries between fixed sized and
machine dependent sized types.

Intermediate code run time checking is always generated by the front
end.  However this intermediate code is only translated into actual
code if the appropriate command line switches are specified.  This
allows the compiler to perform limited range checking at compile time.
In the future it will allow the extensive GCC optimizations to
propagate constant values through to the range checks which if they
are found to exceed the type range will result in a compile time
error message.

@subsection Parameter compatibility

Parameter compatibility is divided into two areas, pass by value and
pass by reference (@code{VAR}).  In the case of pass by value the
rules are exactly the same as assignment.  However in the second case,
pass by reference, the actual parameter and formal parameter must be
the same size and family.  Furthermore @code{INTEGER} and
@code{CARDINAL}s are not treated as compatible in the pass by
reference case.

The types @code{BYTE}, @code{LOC}, @code{WORD} and @code{WORD}n
derivatives are assignment and parameter compatible with any data type
of the same size.

@node Unbounded by reference, Building a shared library, Type compatibility, Using
@section Unbounded by reference

This section documents a GNU Modula-2 compiler switch which implements
a language optimization surrounding the implementation of unbounded
arrays.  In GNU Modula-2 the unbounded array is implemented by
utilizing an internal structure @code{struct @{dataType *address,
unsigned int high@}}.  So given the Modula-2 procedure declaration:

@example
PROCEDURE foo (VAR a: ARRAY OF dataType) ;
BEGIN
   IF a[2]= (* etc *)
END foo ;
@end example

it is translated into GCC @code{tree}s, which can be represented
in their C form thus:

@example
void foo (struct @{dataType *address, unsigned int high@} a)
@{
   if (a.address[2] == /* etc */
@}
@end example

Whereas if the procedure @code{foo} was declared as:

@example
PROCEDURE foo (a: ARRAY OF dataType) ;
BEGIN
   IF a[2]= (* etc *)
END foo ;
@end example

then it is implemented by being translated into the following
GCC @code{tree}s, which can be represented in their C form thus:

@example
void foo (struct @{dataType *address, unsigned int high@} a)
@{
   dataType *copyContents = (dataType *)alloca (a.high+1);
   memcpy(copyContents, a.address, a.high+1);
   a.address = copyContents;

   if (a.address[2] == /* etc */
@}
@end example

This implementation works, but it makes a copy of each non VAR
unbounded array when a procedure is entered.  If the unbounded array
is not changed during procedure @code{foo} then this implementation
will be very inefficient.  In effect Modula-2 lacks the @code{REF}
keyword of Ada.  Consequently the programmer maybe tempted to
sacrifice semantic clarity for greater efficiency by declaring the
parameter using the @code{VAR} keyword in place of @code{REF}.

The @code{-funbounded-by-reference} switch instructs the compiler to
check and see if the programmer is modifying the content of any
unbounded array.  If it is modified then a copy will be made upon
entry into the procedure.  Conversely if the content is only read and
never modified then this non @code{VAR} unbounded array is a candidate
for being passed by reference.  It is only a candidate as it is still
possible that passing this parameter by reference could alter the
meaning of the source code.  For example consider the following case:

@example
PROCEDURE StrConCat (VAR a: ARRAY OF CHAR; b, c: ARRAY OF CHAR) ;
BEGIN
   (* code which performs string a := b + c *)
END StrConCat ;

PROCEDURE foo ;
VAR
   a: ARRAY [0..3] OF CHAR ;
BEGIN
   a := 'q' ;
   StrConCat(a, a, a)
END foo ;
@end example

In the code above we see that the same parameter, @code{a}, is being
passed three times to @code{StrConCat}.  Clearly even though parameters
@code{b} and @code{c} are never modified it would be incorrect to
implement them as pass by reference.  Therefore the compiler checks to
see if any non @code{VAR} parameter is type compatible with any
@code{VAR} parameter and if so it generates run time procedure entry
checks to determine whether the contents of parameters @code{b} or
@code{c} matches the contents of @code{a}.  If a match is detected
then a copy is made and the @code{address} in the unbounded
@code{struct}ure is modified.

The compiler will check the address range of each candidate against
the address range of any @code{VAR} parameter, providing they are type
compatible.  For example consider:

@example
PROCEDURE foo (a: ARRAY OF BYTE; VAR f: REAL) ;
BEGIN
   f := 3.14 ;
   IF a[0]=BYTE(0)
   THEN
      (* etc *)
   END
END foo ;

PROCEDURE bar ;
BEGIN
   r := 2.0 ;
   foo(r, r)
END bar ;
@end example

Here we see that although parameter, @code{a}, is a candidate for the
passing by reference, it would be incorrect to use this
transformation.  Thus the compiler detects that parameters, @code{a}
and @code{f} are type compatible and will produce run time checking
code to test whether the address range of their respective contents
intersect.

@node Linking, Building a shared library, Unbounded by reference, Using

This section describes the linking related options.  There are three
linking strategies available which are dynamic scaffold, static
scaffold and user defined.  The dynamic scaffold is enabled by default
and each module will register itself to the run time @samp{M2RTS} via
a constructor.  The static scaffold mechanism will invoke each modules
@samp{_init} and @samp{_finish} function in turn via a sequence of
calls from within @samp{main}.  Lastly the user defined strategy
can be implemented by turning off the dynamic and static options via
@samp{-fno-scaffold-dynamic} and @samp{-fno-scaffold-static}.

In the simple test below:

@example
$ gm2 hello.mod
@end example

the driver will add the options @samp{-fscaffold-dynamic} and
@samp{-fgen-module-list=-} which generate a list of application
modules and also creates the @samp{main} function with calls to
@samp{M2RTS}.  It can be useful to add the option @samp{-fsources}
which displays the source files as they are parsed and summarizes
whether the source file is required for compilation or linking.

If you wish to split the above command line into a compile and link
then you could use these steps:

@example
$ gm2 -c -fscaffold-main hello.mod
$ gm2 hello.o
@end example

The @samp{-fscaffold-main} informs the compiler to generate the
@samp{main} function and scaffold.  You can enable the environment
variable @samp{GCC_M2LINK_RTFLAG} to trace the construction and
destruction of the application.  The values for
@samp{GCC_M2LINK_RTFLAG} are shown in the table below:

@example
value   | meaning
=================
all     | turn on all flags below
module  | trace modules as they register themselves
pre     | generate module list prior to dependency resolution
dep     | trace module dependency resolution
post    | generate module list after dependency resolution
force   | generate a module list after dependency and forced
        | ordering is complete
@end example

The values can be combined using a comma separated list.

One of the advantages of the dynamic scaffold is that the driver
behaves in a similar way to the other front end drivers.
For example consider a small project consisting of 4 definition
implementation modules (@samp{a.def}, @samp{a.mod}, @samp{b.def},
@samp{b.mod}, @samp{c.def}, @samp{c.mod}, @samp{d.def}, @samp{d.mod})
and a program module @samp{program.mod}.

To link this project we could:

@example
$ gm2 -g -c a.mod
$ gm2 -g -c b.mod
$ gm2 -g -c c.mod
$ gm2 -g -c d.mod
$ gm2 -g program.mod a.o b.o c.o d.o
@end example

The module initialization sequence is defined by the ISO standard to
follow the import graph traversal.  The initialization order is the
order in which the corresponding separate modules finish the
processing of their import lists.

However, if required, you can override this using
@samp{-fruntime-modules=a,b,c,d} for example which forces the
initialization sequence to @samp{a}, @samp{b}, @samp{c} and @samp{d}.

@node Building a shared library, Interface for Python, Unbounded by reference, Using
@section Building a shared library

This section describes building a tiny shared library implemented in
Modula-2 and built with @file{libtool}.  Suppose a project consists of
two definition modules and two implementation modules and a program
module @file{a.def}, @file{a.mod}, @file{b.def}, @file{b.mod} and
@file{c.mod}.  The first step is to compile the modules using position
independent code.  This can be achieved by the following three
commands:

@example
libtool --tag=CC --mode=compile gm2 -g -c a.mod -o a.lo
libtool --tag=CC --mode=compile gm2 -g -c b.mod -o b.lo
libtool --tag=CC --mode=compile gm2 -g -c c.mod -o c.lo
@end example

The second step is to generate the shared library initialization and
finalization routines.  We can do this by asking gm2 to generate a
list of dependent modules and then use this to generate the scaffold.
We also must compile the scaffold.

@example
gm2 -c -g -fmakelist c.mod
gm2 -c -g -fmakeinit -fshared c.mod
libtool --tag=CC --mode=compile g++ -g -c c_m2.cpp -o c_m2.lo
@end example

The third step is to link all these @file{.lo} files.

@example
libtool --mode=link gcc -g c_m2.lo a.lo b.lo c.lo \
        -L$(prefix)/lib64 \
        -rpath `pwd` -lgm2 -lstdc++ -lm -o libabc.la
@end example

At this point the shared library @file{libabc.so} will have been
created inside the directory @file{.libs}.

@node Interface for Python, Producing a Python module, Building a shared library, Using
@section How to produce swig interface files

This section describes how Modula-2 implementation modules can be
called from Python (and other scripting languages such as TCL and
Perl).  GNU Modula-2 can be instructed to create a swig interface when
it is compiling an implementation module.  Swig then uses the
interface file to generate all the necessary wrapping to that the
desired scripting language may access the implementation module.

Here is an example of how you might call upon the services of the
Modula-2 library module @code{NumberIO} from Python3.

The following commands can be used to generate the Python3 module:

@example
export src=@samp{directory to the sources}
export prefix=@samp{directory to where the compiler is installed}
gm2 -I$@{src@} -c -g -fswig $@{src@}/../../../gm2-libs/NumberIO.mod
gm2 -I$@{src@} -c -g -fmakelist $@{src@}/../../../gm2-libs/NumberIO.mod

gm2 -I$@{src@} -c -g -fmakeinit -fshared \
   $@{src@}/../../../gm2-libs/NumberIO.mod

swig -c++ -python3 NumberIO.i

libtool --mode=compile g++ -g -c -I$@{src@} NumberIO_m2.cpp \
  -o NumberIO_m2.lo

libtool --tag=CC --mode=compile gm2 -g -c \
  -I$@{src@}../../../gm2-libs \
  $@{src@}/../../../gm2-libs/NumberIO.mod -o NumberIO.lo

libtool --tag=CC --mode=compile g++ -g -c NumberIO_wrap.cxx \
  -I/usr/include/python3 -o NumberIO_wrap.lo

libtool --mode=link gcc -g NumberIO_m2.lo NumberIO_wrap.lo \
   -L$@{prefix@}/lib64 \
   -rpath `pwd` -lgm2 -lstdc++ -lm -o libNumberIO.la

cp .libs/libNumberIO.so _NumberIO.so
@end example

The first four commands, generate the swig interface file
@file{NumberIO.i} and python wrap files @file{NumberIO_wrap.cxx} and
@file{NumberIO.py}.  The next three @file{libtool} commnads compile
the C++ and Modula-2 source code into @file{.lo} objects.  The last
@file{libtool} command links all the @file{.lo} files into a
@file{.la} file and includes all shared library dependencies.

Now it is possible to run the following Python script
(called @file{testnum.py}):

@example
import NumberIO

print ("1234 x 2 =", NumberIO.NumberIO_StrToInt("1234")*2)
@end example

like this:

@example
$ python3 testnum.py
1234 x 2 = 2468
@end example

@xref{Producing a Python module, , ,gm2} for another example which
uses the @code{UNQUALIFIED} keyword to reduce the module name clutter
from the viewport of Python3.

@subsection Limitations of automatic generated of Swig files

This section discusses the limitations of automatically generating
swig files.  From the previous example we see that the module
@code{NumberIO} had a swig interface file @file{NumberIO.i}
automatically generated by the compiler.  If we consider three of the
procedure definitions in @file{NumberIO.def} we can see the
success and limitations of the automatic interface generation.

@example
PROCEDURE StrToHex (a: ARRAY OF CHAR; VAR x: CARDINAL) ;
PROCEDURE StrToInt (a: ARRAY OF CHAR; VAR x: INTEGER) ;
PROCEDURE ReadInt (VAR x: CARDINAL) ;
@end example

Below are the swig interface prototypes:

@example
extern void NumberIO_StrToHex (char *_m2_address_a,
                               int _m2_high_a, unsigned int *OUTPUT);
/*  parameters: x is known to be an OUTPUT */
extern void NumberIO_StrToInt (char *_m2_address_a,
                               int _m2_high_a, int *OUTPUT);
/*  parameters: x is guessed to be an OUTPUT */
extern void NumberIO_ReadInt (int *x);
/*  parameters: x is unknown */
@end example

In the case of @code{StrToHex} it can be seen that the compiler
detects that the last parameter is an output.  It explicitly tells
swig this by using the parameter name @code{OUTPUT} and in the
following comment it informs the user that it knows this to be an
output parameter.  In the second procedure @code{StrToInt} it marks
the final parameter as an output, but it tells the user that this is
only a guess.  Finally in @code{ReadInt} it informs the user that
it does not know whether the parameter, @code{x}, is an output, input
or an inout parameter.

The compiler decides whether to mark a parameter as either:
@code{INPUT}, @code{OUTPUT} or @code{INOUT} if it is read before
written or visa versa in the first basic block.  At this point
it will write output that the parameter is known.  If it is not
read or written in the first basic block then subsequent basic blocks
are searched and the result is commented as a guess.  Finally if
no read or write occurs then the parameter is commented as unknown.
However, clearly it is possible to fool this mechanism.  Nevertheless
automatic generation of implementation module into swig interface files
was thought sufficiently useful despite these limitations.

In conclusion it would be wise to check all parameters in any
automatically generated swig interface file.  Furthermore you can
force the automatic mechanism to generate correct interface files by
reading or writing to the @code{VAR} parameter in the first basic
block of a procedure.

@node Producing a Python module, Interface to C, Interface for Python, Using
@section How to produce a Python module

This section describes how it is possible to produce a Python module
from Modula-2 code.  There are a number of advantages to this
approach, it ensures your code reaches a wider audience, maybe it is
easier to initialize your application in Python.

The example application here is a pedagogical two dimensional gravity
next event simulation.  The Python module needs to have a clear API
which should be placed in a single definition module.  Furthermore the
API should only use fundamental pervasive data types and strings.
Below the API is contained in the file @file{twoDsim.def}:

@example
DEFINITION MODULE twoDsim ;

EXPORT UNQUALIFIED gravity, box, poly3, poly5, poly6, mass,
                   fix, circle, pivot, velocity, accel, fps,
                   replayRate, simulateFor ;
(*
   gravity - turn on gravity at: g m^2
*)

PROCEDURE gravity (g: REAL) ;


(*
   box - place a box in the world at (x0,y0),(x0+i,y0+j)
*)

PROCEDURE box (x0, y0, i, j: REAL) : CARDINAL ;


(*
   poly3 - place a triangle in the world at:
           (x0,y0),(x1,y1),(x2,y2)
*)

PROCEDURE poly3 (x0, y0, x1, y1, x2, y2: REAL) : CARDINAL ;


(*
   poly5 - place a pentagon in the world at:
           (x0,y0),(x1,y1),(x2,y2),(x3,y3),(x4,y4)
*)

PROCEDURE poly5 (x0, y0, x1, y1,
                 x2, y2, x3, y3, x4, y4: REAL) : CARDINAL ;


(*
   poly6 - place a hexagon in the world at:
           (x0,y0),(x1,y1),(x2,y2),(x3,y3),(x4,y4),(x5,y5)
*)

PROCEDURE poly6 (x0, y0, x1, y1,
                 x2, y2, x3, y3,
                 x4, y4, x5, y5: REAL) : CARDINAL ;


(*
   mass - specify the mass of an object and return the, id.
*)

PROCEDURE mass (id: CARDINAL; m: REAL) : CARDINAL ;


(*
   fix - fix the object to the world.
*)

PROCEDURE fix (id: CARDINAL) : CARDINAL ;


(*
   circle - adds a circle to the world.  Center
            defined by: x0, y0 radius, r.
*)

PROCEDURE circle (x0, y0, r: REAL) : CARDINAL ;


(*
   velocity - give an object, id, a velocity, vx, vy.
*)

PROCEDURE velocity (id: CARDINAL; vx, vy: REAL) : CARDINAL ;


(*
   accel - give an object, id, an acceleration, ax, ay.
*)

PROCEDURE accel (id: CARDINAL; ax, ay: REAL) : CARDINAL ;


(*
   fps - set frames per second.
*)

PROCEDURE fps (f: REAL) ;


(*
   replayRate - set frames per second during replay.
*)

PROCEDURE replayRate (f: REAL) ;


(*
   simulateFor - render for, t, seconds.
*)

PROCEDURE simulateFor (t: REAL) ;


END twoDsim.
@end example

The keyword @code{UNQUALIFIED} can be used to ensure that the
compiler will provide externally accessible functions
@code{gravity}, @code{box}, @code{poly3}, @code{poly5}, @code{poly6},
@code{mass}, @code{fix}, @code{circle}, @code{pivot}, @code{velocity},
@code{accel}, @code{fps}, @code{replayRate}, @code{simulateFor}
rather than name mangled alternatives.
Hence in our Python3 application we could write:

@example
#!/usr/bin/env python3

from twoDsim import *

b = box (0.0, 0.0, 1.0, 1.0)
b = fix (b)
c1 = circle (0.7, 0.7, 0.05)
c1 = mass (c1, 0.01)
c2 = circle (0.7, 0.1, 0.05)
c2 = mass (c2, 0.01)
c2 = fix (c2)
gravity (-9.81)
fps (24.0*4.0)
replayRate (24.0)
print ("creating frames")
try:
    simulateFor (1.0)
    print ("all done")
except:
    print ("exception raised")
@end example

which accesses the various functions defined and implemented by the
module @code{twoDsim}.  The Modula-2 source code is compiled via:

@example
$ gm2 -g -fiso -c -fswig twoDsim.mod
$ gm2 -g -fiso -c -fmakelist twoDsim.mod
$ gm2 -g -fiso -c -fmakeinit twoDsim.mod
@end example

The first command both compiles the source file creating
@file{twoDsim.o} and produces a swig interface file @file{swig.i}.  We
now use @code{swig} and @code{g++} to produce and compile the
interface wrappers:

@example
$ libtool --mode=compile g++ -g -c twoDsim_m2.cpp -o twoDsim_m2.lo
$ swig -c++ -python3 twoDsim.i
$ libtool --mode=compile g++ -c -fPIC twoDsim_wrap.cxx \
   -I/usr/include/python3 -o twoDsim_wrap.lo
$ libtool --mode=compile gm2 -g -fPIC -fiso -c deviceGnuPic.mod
$ libtool --mode=compile gm2 -g -fPIC -fiso -c roots.mod
$ libtool --mode=compile gm2 -g -fPIC -fiso -c -fswig \
   twoDsim.mod -o twoDsim.lo
@end example

Finally the application is linked into a shared library:

@example
$ libtool --mode=link gcc -g twoDsim_m2.lo twoDsim_wrap.lo \
  roots.lo deviceGnuPic.lo \
   -L$@{prefix@}/lib64 \
   -rpath `pwd` -lgm2 -lstdc++ -lm -o libtwoDsim.la
cp .libs/libtwoDsim.so _twoDsim.so
@end example

The library name must start with @code{_} to comply with the Python3
module naming scheme.

@node Interface to C, Assembly language, Producing a Python module, Using
@section Interfacing GNU Modula-2 to C

The GNU Modula-2 compiler tries to use the C calling convention
wherever possible however some parameters have no C equivalent and
thus a language specific method is used.  For example unbounded arrays
are passed as a @code{struct @{void *address, unsigned int high@}} and
the contents of these arrays are copied by callee functions when they
are declared as non @code{VAR} parameters.  The @code{VAR} equivalent
unbounded array parameters need no copy, but still use the
@code{struct} representation.

The recommended method of interfacing GNU Modula-2 to C is by telling
the definition module that the implementation is in the C language.
This is achieved by using the tokens @code{DEFINITION MODULE FOR "C"}.
Here is an example @file{libprintf.def}.

@example
DEFINITION MODULE FOR "C" libprintf ;

EXPORT UNQUALIFIED printf ;

PROCEDURE printf (a: ARRAY OF CHAR; ...) : [ INTEGER ] ;

END libprintf.
@end example

the @code{UNQUALIFIED} keyword in the definition module informs
GNU Modula-2 not to prefix the module name to exported references
in the object file.

The @code{printf} declaration states that the first parameter
semantically matches @code{ARRAY OF CHAR} but since the module is for
the C language it will be mapped onto @code{char *}.  The token
@code{...} indicates a variable number of arguments (varargs) and all
parameters passed here are mapped onto their C equivalents.  Arrays and
constant strings are passed as pointers.  Lastly @code{[ INTEGER ]}
states that the caller can ignore the function return result if desired.

The hello world program can be rewritten as:

@example
MODULE hello ;

FROM libprintf IMPORT printf ;

BEGIN
   printf ("hello world\n")
END hello.
@end example

and it can be compiled by:

@samp{gm2 -g hello.mod -lc}

In reality the @samp{-lc} is redundant as libc is always included in the
linking process.  It is shown here to emphasize that the C library or
object file containing @code{printf} must be present.  The search path
for modules can be changed by using @samp{-I}.

If a procedure function is declared using varargs then some parameter
values are converted.  The table below summarizes the default conversions
and default types used.

@example
Actual Parameter       |  Default conversion  |   Type of actual
                       |                      |   value passed
===============================================================
123                    |  none                |   long long int
"hello world"          |  none                |   const char *
a: ARRAY OF CHAR       |  ADR (a)             |   char *
a: ARRAY [0..5] OF CHAR|  ADR (a)             |   char *
3.14                   |  none                |   long double
@end example

If you wish to pass @code{int} values then you should explicitly
convert the constants using one of the conversion mechanisms.
For example:  @code{INTEGER(10)} or @code{VAL(INTEGER, 10)} or
@code{CAST(INTEGER, 10)}.

@node Assembly language, Alignment, Interface to C, Using
@section Interface to assembly language

The interface for GNU Modula-2 to assembly language is almost
identical to GNU C.  The only alterations are that the keywords
@code{asm} and @code{volatile} are in capitals, following the Modula-2
convention.

A simple, but highly non optimal, example is given below.  Here we want
to add the two @code{CARDINAL}s @code{foo} and @code{bar} together and
return the result.  The target processor is assumed to be executing
the x86_64 instruction set.

@example
PROCEDURE Example (foo, bar: CARDINAL) : CARDINAL ;
VAR
   myout: CARDINAL ;
BEGIN
   ASM VOLATILE ("movq %1,%%rax; addq %2,%%rax; movq %%rax,%0"
      : "=rm" (myout)            (* outputs *)
      : "rm" (foo), "rm" (bar)   (* inputs  *)
      : "rax") ;                 (* we trash *)
   RETURN( myout )
END Example ;
@end example

For a full description of this interface we refer the reader to the GNU C manual.

@xref{Extended Asm, ,Extensions to the C Language Family,gcc}.

The same example can be written using the newer extensions of naming
the operands rather than using numbered arguments.

@example
PROCEDURE Example (foo, bar: CARDINAL) : CARDINAL ;
VAR
   myout: CARDINAL ;
BEGIN
   ASM VOLATILE (
    "movq %[left],%%rax; addq %[right],%%rax; movq %%rax,%[output]"
      : [output] "=rm" (myout)                  (* outputs *)
      : [left] "rm" (foo), [right] "rm" (bar)   (* inputs  *)
      : "rax") ;                                (* we trash *)
   RETURN( myout )
END Example ;
@end example

Both examples generate exactly the same code.  It is worth noting that
the specifier ``rm'' indicates that the operand can be either a
register or memory.  Of course you must choose an instruction which
can take either, but this allows the compiler to take make more
efficient choices depending upon the optimization level given to the
compiler.

@node Alignment, Packed, Assembly language, Using
@section Data type alignment

GNU Modula-2 allows you to specify alignment for types and variables.
The syntax for alignment is to use the ISO pragma directives @code{<*}
@code{bytealignment (} expression @code{)} and @code{*>}.  These directives
can be used after type and variable declarations.

The ebnf of the alignment production is:

@example
Alignment := [ ByteAlignment ] =:
ByteAlignment := '<*' AttributeExpression '*>' =:
AlignmentExpression := "(" ConstExpression ")" =:
@end example

The @code{Alignment} ebnf statement may be used during construction of
types, records, record fields, arrays, pointers and variables.  Below
is an example of aligning a type so that the variable @code{bar} is
aligned on a 1024 address.

@example
MODULE align ;

TYPE
   foo = INTEGER <* bytealignment(1024) *> ;

VAR
   z  : INTEGER ;
   bar: foo ;
BEGIN
END align.
@end example

The next example aligns a variable on a 1024 byte boundary.

@example
MODULE align2 ;

VAR
   x  : CHAR ;
   z  : ARRAY [0..255] OF INTEGER <* bytealignment(1024) *> ;
BEGIN
END align2.
@end example

Here the example aligns a pointer on a 1024 byte boundary.

@example
MODULE align4 ;

FROM SYSTEM IMPORT ADR ;
FROM libc IMPORT exit ;

VAR
   x  : CHAR ;
   z  : POINTER TO INTEGER <* bytealignment(1024) *> ;
BEGIN
   IF ADR(z) MOD 1024=0
   THEN
      exit(0)
   ELSE
      exit(1)
   END
END align4.
@end example

In example @code{align5} record field @code{y} is aligned on a 1024
byte boundary.

@example
MODULE align5 ;

FROM SYSTEM IMPORT ADR ;
FROM libc IMPORT exit ;

TYPE
   rec = RECORD
            x: CHAR ;
            y: CHAR <* bytealignment(1024) *> ;
         END ;
VAR
   r: rec ;
BEGIN
   IF ADR(r.y) MOD 1024=0
   THEN
      exit(0)
   ELSE
      exit(1)
   END
END align5.
@end example

In the example below module @code{align6} declares @code{foo} as an
array of 256 @code{INTEGER}s.  The array @code{foo} is aligned on a
1024 byte boundary.

@example
MODULE align6 ;

FROM SYSTEM IMPORT ADR ;
FROM libc IMPORT exit ;

TYPE
   foo = ARRAY [0..255] OF INTEGER <* bytealignment(1024) *> ;

VAR
   x  : CHAR ;
   z  : foo ;
BEGIN
   IF ADR(z) MOD 1024=0
   THEN
      exit(0)
   ELSE
      exit(1)
   END
END align6.
@end example

@node Packed, Built-ins, Alignment, Using
@section Packing data types

The pragma @code{<* bytealignment(0) *>} can be used to specify that
the fields within a @code{RECORD} are to be packed.  Currently this
only applies to fields which are declared as subranges, ordinal types
and enumerated types.  Here is an example of how two subranges might
be packed into a byte.

@example
TYPE
   bits3c =  [0..7] ;
   bits3i = [-4..3] ;

   byte = RECORD
              <* bytealignment(0) *>
              x: bits3c ;
              <* bitsunused(2) *>
              y: bits3i ;
          END ;
@end example

Notice that the user has specified that in between fields @code{x} and
@code{y} there are two bits unused.

Now the user wishes to create a record with byte numbers zero and one
occupied and then an @code{INTEGER32} field which is four byte
aligned.  In this case byte numbers two and three will be unused.  The
pragma @code{bytealignment} can be issued at the start of the record
indicating the default alignment for the whole record and this can be
overridden by individual fields if necessary.

@example
   rec = RECORD
            <* bytealignment (1) *> ;
            a, b: byte ;
            x: INTEGER32 <* bytealignment(4) *> ;
         END ;
@end example

In the following example the user has specified that a record has two
fields @code{p} and @code{q} but that there are three bytes unused between
these fields.

@example
   header = RECORD
               <* bytealignment(1) *>
               p: byte ;
               <* bytesunused(3) *>
               q: byte ;
            END ;
@end example

The pragma @code{<* bytesunused(x) *>} can only be used if the current
field is on a byte boundary.  There is also a @code{SYSTEM} pseudo
procedure function @code{TBITSIZE(T)} which returns the minimum number of
bits necessary to represent type @code{T}.

Another example of packing record bit fields is given below:

@example
MODULE align21 ;

FROM libc IMPORT exit ;

TYPE
   colour = (red, blue, green, purple, white, black) ;

   soc = PACKEDSET OF colour ;

   rec = RECORD
            <* bytealignment(0) *>
            x: soc ;
            y: [-1..1] ;
         END ;

VAR
   r: rec ;
   v: CARDINAL ;
BEGIN
   v := SIZE(r) ;
   IF SIZE(r)#1
   THEN
      exit(1)
   END ;
   r.x := soc@{blue@} ;
   IF r.x#soc@{blue@}
   THEN
      exit(2)
   END
END align21.
@end example

Here we see that the total size of this record is one byte and consists
of a six bit set type followed by a 2 bit integer subrange.

@node Built-ins, The PIM system module, Packed, Using
@section Accessing GNU Modula-2 Built-ins

This section describes the built-in constants and functions defined in
GNU Modula-2.  The following compiler constants can be accessed using
the @code{__ATTRIBUTE__} @code{__BUILTIN__} keywords.  These are not
part of the Modula-2 language and they may differ depending upon the
target architecture but they provide a method whereby common
libraries can interface to a different underlying architecture.

The built-in constants are: @code{BITS_PER_UNIT}, @code{BITS_PER_WORD},
@code{BITS_PER_CHAR} and @code{UNITS_PER_WORD}.  They are integrated into
GNU Modula-2 by an extension to the @code{ConstFactor} rule:

@example
ConstFactor := ConstQualidentOrSet | Number | ConstString |
               "(" ConstExpression ")" | "NOT" ConstFactor |
               ConstAttribute =:

ConstAttribute := "__ATTRIBUTE__" "__BUILTIN__" "(" "(" Ident ")" ")" =:
@end example

Here is an example taken from the ISO library @code{SYSTEM.def}:

@example
CONST
   BITSPERLOC    = __ATTRIBUTE__ __BUILTIN__ ((BITS_PER_UNIT)) ;
   LOCSPERWORD   = __ATTRIBUTE__ __BUILTIN__ ((UNITS_PER_WORD)) ;
@end example

Built-in functions are transparent to the end user.  All built-in
functions are declared in @code{DEFINITION MODULE}s and are imported
as and when required.  Built-in functions are declared in definition
modules by using the @code{__BUILTIN__} keyword.  Here is a section of
the ISO library @code{LongMath.def} which demonstrates this feature.

@example
PROCEDURE __BUILTIN__ sqrt (x: LONGREAL): LONGREAL;
  (* Returns the square root of x *)
@end example

This indicates that the function @code{sqrt} will be implemented using
the gcc built-in maths library.  If gcc cannot utilize the built-in
function (for example if the programmer requested the address of
@code{sqrt}) then code is generated to call the alternative function
implemented in the @code{IMPLEMENTATION} @code{MODULE}.

Sometimes a function exported from the @code{DEFINITION} @code{MODULE}
will have a different name from the built-in function within gcc.  In
such cases the mapping between the GNU Modula-2 function name and the
gcc name is expressed using the keywords @code{__ATTRIBUTE__}
@code{__BUILTIN__} @code{((Ident))}.  For example the function
@code{sqrt} in @code{LongMath.def} maps onto the gcc built-in function
@code{sqrtl} and this is expressed as:

@example
PROCEDURE __ATTRIBUTE__ __BUILTIN__ ((sqrtl)) sqrt
                                    (x: LONGREAL) : LONGREAL;
  (* Returns the positive square root of x *)
@end example

The following module @code{Builtins.def} enumerates the list of
built-in functions which can be accessed in GNU Modula-2.  It also
serves to define the parameter and return value for each function:

@include m2/Builtins.texi

Although this module exists and will result in the generation of
in-line code if optimization flags are passed to GNU Modula-2, users
are advised to utilize the same functions from more generic libraries.
The built-in mechanism will be applied to these generic
libraries where appropriate.  Note for the mathematical routines to
be in-lined you need to specify the @samp{-ffast-math -O} options.

@node The PIM system module, The ISO system module, Built-ins, Using
@section The PIM system module

@include m2/SYSTEM-pim.texi

The different dialects of Modula-2 PIM-[234] and ISO Modula-2 declare
the function @code{SIZE} in different places.  PIM-[34] and ISO
Modula-2 declare @code{SIZE} as a pervasive function (declared in the
base module).  PIM-2 defined @code{SIZE} in the @code{SYSTEM} module
(as shown above).

GNU Modula-2 allows users to specify the dialect of Modula-2 by using
the @code{-fiso} and @code{-fpim2} command line switches.

The data types @code{CSIZE_T} and @code{CSSIZE_T} are also exported from
the @code{SYSTEM} module.  The type @code{CSIZE_T} is unsigned and is
mapped onto the target C data type @code{size_t} whereas the type
@code{CSSIZE_T} is mapped onto the signed C data type @code{ssize_t}.

It is anticipated that these should only be used to provide cross
platform definition modules for C libraries.

There are also a variety of fixed sized @code{INTEGER} and
@code{CARDINAL} types.  The variety of the fixed sized types will
depend upon the target architecture.

@node The ISO system module, , The PIM system module, Using
@section The ISO system module

@include m2/SYSTEM-iso.texi

The data types @code{CSIZE_T} and @code{CSSIZE_T} are also exported from
the @code{SYSTEM} module.  The type @code{CSIZE_T} is unsigned and is
mapped onto the target C data type @code{size_t} whereas the type
@code{CSSIZE_T} is mapped onto the signed C data type @code{ssize_t}.

It is anticipated that these should only be used to provide cross
platform definition modules for C libraries.

There are also a variety of fixed sized @code{INTEGER} and
@code{CARDINAL} types.  The variety of the fixed sized types will
depend upon the target architecture.

@node License, Copying, The ISO system module, Top
@section License of GNU Modula-2

GNU Modula-2 is free software, the compiler is held under the GPL v3
@uref{http://www.gnu.org/licenses/gpl.txt},
its libraries (pim, iso and Logitech compatible) are under the
GPL v3 with the GCC run time library exception clause.

Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.

You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
<http://www.gnu.org/licenses/>.

More information on how these licenses work is available
@uref{http://www.gnu.org/licenses/licenses.html} on the GNU web site.

@c Copying node is inside the gpl_v3.texi
@include gpl_v3.texi

@node Contributing, Internals, Copying, Top
@section Contributing to GNU Modula-2

Please do and please read the GNU Emacs info under

@example
* Standards: (standards).       GNU coding standards.
* Intellectual Property::       Keeping Free Software Free
* Reading Non-Free Code::       Referring to Proprietary Programs
* Contributions::               Accepting Contributions
@end example

You might consider joining the GM2 Mailing list before you start
coding.  The mailing list may be subscribed via a web interface
@uref{http://lists.nongnu.org/mailman/listinfo/gm2} or via email
@email{gm2-subscribe@@nongnu.org}.

Many thanks and enjoy your coding!

@node Internals, EBNF, Contributing, Top

This section is still being written.
@c @include gm2-internals.texi

@node EBNF, Libraries, Internals, Top
@chapter EBNF of GNU Modula-2

This chapter contains the EBNF of GNU Modula-2.  This grammar currently
supports both PIM and ISO dialects.  The rules here are automatically
extracted from the crammer files in GNU Modula-2 and serve to document
the syntax of the extensions described earlier and how they fit in
with the base language.

Note that the first six productions are built into the lexical analysis
phase.

@include m2/gm2-ebnf.texi

@node Libraries, Indices, EBNF, Top
@chapter PIM and ISO library definitions

This chapter contains M2F, PIM and ISO libraries.

@include m2/gm2-libs.texi

@node Indices, , Libraries, Top
@section Indices

@ifhtml
@menu
* Contents::    Section and subsections.
* Functions::   Function, constants, types, ebnf indices.
@end menu

@node Contents, Functions, ,
@section Section and subsections
@printindex cp

@node Functions, , Contents,
@section Function, constants, types, ebnf indices.
@end ifhtml

@printindex fn

@summarycontents
@contents
@bye