1 \input texinfo @c -*-texinfo-*-
3 @setfilename gfortran.info
4 @set copyrights-gfortran 1999-2007
6 @include gcc-common.texi
8 @settitle The GNU Fortran Compiler
10 @c Create a separate index for command line options
12 @c Merge the standard indexes into a single one.
19 @c TODO: The following "Part" definitions are included here temporarily
20 @c until they are incorporated into the official Texinfo distribution.
21 @c They borrow heavily from Texinfo's \unnchapentry definitions.
28 \vglue\titlepagetopglue
30 \leftline{Part #1:@* #2}
31 \vskip4pt \hrule height 4pt width \hsize \vskip4pt
33 \writetocentry{part}{#2}{#1}
36 \writetocentry{blankpart}{}{}
38 % Part TOC-entry definition for summary contents.
39 \gdef\dosmallpartentry#1#2#3#4{%
40 \vskip .5\baselineskip plus.2\baselineskip
43 \tocentry{Part #2: #1}{\doshortpageno\bgroup#4\egroup}
46 \gdef\dosmallblankpartentry#1#2#3#4{%
47 \vskip .5\baselineskip plus.2\baselineskip
49 % Part TOC-entry definition for regular contents. This has to be
50 % equated to an existing entry to not cause problems when the PDF
52 \gdef\dopartentry#1#2#3#4{%
53 \unnchapentry{Part #2: #1}{}{#3}{#4}
55 \gdef\doblankpartentry#1#2#3#4{}
60 @c Use with @@smallbook.
62 @c %** start of document
64 @c Cause even numbered pages to be printed on the left hand side of
65 @c the page and odd numbered pages to be printed on the right hand
66 @c side of the page. Using this, you can print on both sides of a
67 @c sheet of paper and have the text on the same part of the sheet.
69 @c The text on right hand pages is pushed towards the right hand
70 @c margin and the text on left hand pages is pushed toward the left
72 @c (To provide the reverse effect, set bindingoffset to -0.75in.)
75 @c \global\bindingoffset=0.75in
76 @c \global\normaloffset =0.75in
80 Copyright @copyright{} @value{copyrights-gfortran} Free Software Foundation, Inc.
82 Permission is granted to copy, distribute and/or modify this document
83 under the terms of the GNU Free Documentation License, Version 1.1 or
84 any later version published by the Free Software Foundation; with the
85 Invariant Sections being ``GNU General Public License'' and ``Funding
86 Free Software'', the Front-Cover
87 texts being (a) (see below), and with the Back-Cover Texts being (b)
88 (see below). A copy of the license is included in the section entitled
89 ``GNU Free Documentation License''.
91 (a) The FSF's Front-Cover Text is:
95 (b) The FSF's Back-Cover Text is:
97 You have freedom to copy and modify this GNU Manual, like GNU
98 software. Copies published by the Free Software Foundation raise
99 funds for GNU development.
103 @dircategory Software development
105 * gfortran: (gfortran). The GNU Fortran Compiler.
107 This file documents the use and the internals of
108 the GNU Fortran compiler, (@command{gfortran}).
110 Published by the Free Software Foundation
111 51 Franklin Street, Fifth Floor
112 Boston, MA 02110-1301 USA
118 @setchapternewpage odd
120 @title Using GNU Fortran
122 @author The @t{gfortran} team
124 @vskip 0pt plus 1filll
125 Published by the Free Software Foundation@*
126 51 Franklin Street, Fifth Floor@*
127 Boston, MA 02110-1301, USA@*
128 @c Last printed ??ber, 19??.@*
129 @c Printed copies are available for $? each.@*
135 @c TODO: The following "Part" definitions are included here temporarily
136 @c until they are incorporated into the official Texinfo distribution.
139 \global\let\partentry=\dosmallpartentry
140 \global\let\blankpartentry=\dosmallblankpartentry
145 \global\let\partentry=\dopartentry
146 \global\let\blankpartentry=\doblankpartentry
152 @c ---------------------------------------------------------------------
153 @c TexInfo table of contents.
154 @c ---------------------------------------------------------------------
161 This manual documents the use of @command{gfortran},
162 the GNU Fortran compiler. You can find in this manual how to invoke
163 @command{gfortran}, as well as its features and incompatibilities.
166 @emph{Warning:} This document, and the compiler it describes, are still
167 under development. While efforts are made to keep it up-to-date, it might
168 not accurately reflect the status of the most recent GNU Fortran compiler.
172 @comment When you add a new menu item, please keep the right hand
173 @comment aligned to the same column. Do not use tabs. This provides
174 @comment better formatting.
179 Part I: Invoking GNU Fortran
180 * Invoking GNU Fortran:: Command options supported by @command{gfortran}.
181 * Runtime:: Influencing runtime behavior with environment variables.
183 Part II: Language Reference
184 * Fortran 2003 status:: Fortran 2003 features supported by GNU Fortran.
185 * Extensions:: Language extensions implemented by GNU Fortran.
186 * Intrinsic Procedures:: Intrinsic procedures supported by GNU Fortran.
188 * Contributing:: How you can help.
189 * Copying:: GNU General Public License says
190 how you can copy and share GNU Fortran.
191 * GNU Free Documentation License::
192 How you can copy and share this manual.
193 * Funding:: How to help assure continued work for free software.
194 * Index:: Index of this documentation.
198 @c ---------------------------------------------------------------------
200 @c ---------------------------------------------------------------------
203 @chapter Introduction
205 @c The following duplicates the text on the TexInfo table of contents.
207 This manual documents the use of @command{gfortran}, the GNU Fortran
208 compiler. You can find in this manual how to invoke @command{gfortran},
209 as well as its features and incompatibilities.
212 @emph{Warning:} This document, and the compiler it describes, are still
213 under development. While efforts are made to keep it up-to-date, it
214 might not accurately reflect the status of the most recent GNU Fortran
219 The GNU Fortran compiler front end was
220 designed initially as a free replacement for,
221 or alternative to, the unix @command{f95} command;
222 @command{gfortran} is the command you'll use to invoke the compiler.
225 * About GNU Fortran:: What you should know about the GNU Fortran compiler.
226 * GNU Fortran and GCC:: You can compile Fortran, C, or other programs.
227 * GNU Fortran and G77:: Why we chose to start from scratch.
228 * Project Status:: Status of GNU Fortran, roadmap, proposed extensions.
229 * Standards:: Standards supported by GNU Fortran.
233 @c ---------------------------------------------------------------------
235 @c ---------------------------------------------------------------------
237 @node About GNU Fortran
238 @section About GNU Fortran
240 The GNU Fortran compiler is still in an early state of development.
241 It can generate code for most constructs and expressions,
242 but much work remains to be done.
244 When the GNU Fortran compiler is finished,
245 it will do everything you expect from any decent compiler:
249 Read a user's program,
250 stored in a file and containing instructions written
251 in Fortran 77, Fortran 90, Fortran 95 or Fortran 2003.
252 This file contains @dfn{source code}.
255 Translate the user's program into instructions a computer
256 can carry out more quickly than it takes to translate the
257 instructions in the first
258 place. The result after compilation of a program is
260 code designed to be efficiently translated and processed
261 by a machine such as your computer.
262 Humans usually aren't as good writing machine code
263 as they are at writing Fortran (or C++, Ada, or Java),
264 because is easy to make tiny mistakes writing machine code.
267 Provide the user with information about the reasons why
268 the compiler is unable to create a binary from the source code.
269 Usually this will be the case if the source code is flawed.
270 When writing Fortran, it is easy to make big mistakes.
271 The Fortran 90 requires that the compiler can point out
272 mistakes to the user.
273 An incorrect usage of the language causes an @dfn{error message}.
275 The compiler will also attempt to diagnose cases where the
276 user's program contains a correct usage of the language,
277 but instructs the computer to do something questionable.
278 This kind of diagnostics message is called a @dfn{warning message}.
281 Provide optional information about the translation passes
282 from the source code to machine code.
283 This can help a user of the compiler to find the cause of
284 certain bugs which may not be obvious in the source code,
285 but may be more easily found at a lower level compiler output.
286 It also helps developers to find bugs in the compiler itself.
289 Provide information in the generated machine code that can
290 make it easier to find bugs in the program (using a debugging tool,
291 called a @dfn{debugger}, such as the GNU Debugger @command{gdb}).
294 Locate and gather machine code already generated to
295 perform actions requested by statements in the user's program.
296 This machine code is organized into @dfn{modules} and is located
297 and @dfn{linked} to the user program.
300 The GNU Fortran compiler consists of several components:
304 A version of the @command{gcc} command
305 (which also might be installed as the system's @command{cc} command)
306 that also understands and accepts Fortran source code.
307 The @command{gcc} command is the @dfn{driver} program for
308 all the languages in the GNU Compiler Collection (GCC);
310 you can compile the source code of any language for
311 which a front end is available in GCC.
314 The @command{gfortran} command itself,
315 which also might be installed as the
316 system's @command{f95} command.
317 @command{gfortran} is just another driver program,
318 but specifically for the Fortran compiler only.
319 The difference with @command{gcc} is that @command{gfortran}
320 will automatically link the correct libraries to your program.
323 A collection of run-time libraries.
324 These libraries contain the machine code needed to support
325 capabilities of the Fortran language that are not directly
326 provided by the machine code generated by the
327 @command{gfortran} compilation phase,
328 such as intrinsic functions and subroutines,
329 and routines for interaction with files and the operating system.
330 @c and mechanisms to spawn,
331 @c unleash and pause threads in parallelized code.
334 The Fortran compiler itself, (@command{f951}).
335 This is the GNU Fortran parser and code generator,
336 linked to and interfaced with the GCC backend library.
337 @command{f951} ``translates'' the source code to
338 assembler code. You would typically not use this
340 instead, the @command{gcc} or @command{gfortran} driver
341 programs will call it for you.
345 @c ---------------------------------------------------------------------
346 @c GNU Fortran and GCC
347 @c ---------------------------------------------------------------------
349 @node GNU Fortran and GCC
350 @section GNU Fortran and GCC
351 @cindex GNU Compiler Collection
354 GNU Fortran is a part of GCC, the @dfn{GNU Compiler Collection}. GCC
355 consists of a collection of front ends for various languages, which
356 translate the source code into a language-independent form called
357 @dfn{GENERIC}. This is then processed by a common middle end which
358 provides optimization, and then passed to one of a collection of back
359 ends which generate code for different computer architectures and
362 Functionally, this is implemented with a driver program (@command{gcc})
363 which provides the command-line interface for the compiler. It calls
364 the relevant compiler front-end program (e.g., @command{f951} for
365 Fortran) for each file in the source code, and then calls the assembler
366 and linker as appropriate to produce the compiled output. In a copy of
367 GCC which has been compiled with Fortran language support enabled,
368 @command{gcc} will recognize files with @file{.f}, @file{.f90}, @file{.f95},
369 and @file{.f03} extensions as Fortran source code, and compile it
370 accordingly. A @command{gfortran} driver program is also provided,
371 which is identical to @command{gcc} except that it automatically links
372 the Fortran runtime libraries into the compiled program.
374 This manual specifically documents the Fortran front end, which handles
375 the programming language's syntax and semantics. The aspects of GCC
376 which relate to the optimization passes and the back-end code generation
377 are documented in the GCC manual; see
378 @ref{Top,,Introduction,gcc,Using the GNU Compiler Collection (GCC)}.
379 The two manuals together provide a complete reference for the GNU
383 @c ---------------------------------------------------------------------
384 @c GNU Fortran and G77
385 @c ---------------------------------------------------------------------
387 @node GNU Fortran and G77
388 @section GNU Fortran and G77
392 The GNU Fortran compiler is the successor to G77, the Fortran 77 front
393 end included in GCC prior to version 4. It is an entirely new program
394 that has been designed to provide Fortran 95 support and extensibility
395 for future Fortran language standards, as well as providing backwards
396 compatibility for Fortran 77 and nearly all of the GNU language
397 extensions supported by G77.
400 @c ---------------------------------------------------------------------
402 @c ---------------------------------------------------------------------
405 @section Project Status
408 As soon as @command{gfortran} can parse all of the statements correctly,
409 it will be in the ``larva'' state.
410 When we generate code, the ``puppa'' state.
411 When @command{gfortran} is done,
412 we'll see if it will be a beautiful butterfly,
413 or just a big bug....
415 --Andy Vaught, April 2000
418 The start of the GNU Fortran 95 project was announced on
419 the GCC homepage in March 18, 2000
420 (even though Andy had already been working on it for a while,
423 The GNU Fortran compiler is able to compile nearly all
424 standard-compliant Fortran 95, Fortran 90, and Fortran 77 programs,
425 including a number of standard and non-standard extensions, and can be
426 used on real-world programs. In particular, the supported extensions
427 include OpenMP, Cray-style pointers, and several Fortran 2003 features
428 such as enumeration, stream I/O, and some of the enhancements to
429 allocatable array support from TR 15581. However, it is still under
430 development and has a few remaining rough edges.
432 At present, the GNU Fortran compiler passes the
433 @uref{http://www.fortran-2000.com/ArnaudRecipes/fcvs21_f95.html,
434 NIST Fortran 77 Test Suite}, and produces acceptable results on the
435 @uref{http://www.netlib.org/lapack/faq.html#1.21, LAPACK Test Suite}.
436 It also provides respectable performance on
437 the @uref{http://www.polyhedron.com/pb05.html, Polyhedron Fortran
438 compiler benchmarks} and the
439 @uref{http://www.llnl.gov/asci_benchmarks/asci/limited/lfk/README.html,
440 Livermore Fortran Kernels test}. It has been used to compile a number of
441 large real-world programs, including
442 @uref{http://mysite.verizon.net/serveall/moene.pdf, the HIRLAM
443 weather-forecasting code} and
444 @uref{http://www.theochem.uwa.edu.au/tonto/, the Tonto quantum
445 chemistry package}; see @url{http://gcc.gnu.org/wiki/GfortranApps} for an
448 Among other things, the GNU Fortran compiler is intended as a replacement
449 for G77. At this point, nearly all programs that could be compiled with
450 G77 can be compiled with GNU Fortran, although there are a few minor known
453 The primary work remaining to be done on GNU Fortran falls into three
454 categories: bug fixing (primarily regarding the treatment of invalid code
455 and providing useful error messages), improving the compiler optimizations
456 and the performance of compiled code, and extending the compiler to support
457 future standards---in particular, Fortran 2003.
460 @c ---------------------------------------------------------------------
462 @c ---------------------------------------------------------------------
468 The GNU Fortran compiler implements
469 ISO/IEC 1539:1997 (Fortran 95). As such, it can also compile essentially all
470 standard-compliant Fortran 90 and Fortran 77 programs. It also supports
471 the ISO/IEC TR-15581 enhancements to allocatable arrays, and
472 the @uref{http://www.openmp.org/drupal/mp-documents/spec25.pdf,
473 OpenMP Application Program Interface v2.5} specification.
475 In the future, the GNU Fortran compiler may also support other standard
476 variants of and extensions to the Fortran language. These include
477 ISO/IEC 1539-1:2004 (Fortran 2003).
480 @c =====================================================================
481 @c PART I: INVOCATION REFERENCE
482 @c =====================================================================
485 \part{I}{Invoking GNU Fortran}
488 @c ---------------------------------------------------------------------
490 @c ---------------------------------------------------------------------
495 @c ---------------------------------------------------------------------
497 @c ---------------------------------------------------------------------
500 @chapter Runtime: Influencing runtime behavior with environment variables
503 The behavior of the @command{gfortran} can be influenced by
504 environment variables.
506 Malformed environment variables are silently ignored.
509 * GFORTRAN_STDIN_UNIT:: Unit number for standard input
510 * GFORTRAN_STDOUT_UNIT:: Unit number for standard output
511 * GFORTRAN_STDERR_UNIT:: Unit number for standard error
512 * GFORTRAN_USE_STDERR:: Send library output to standard error
513 * GFORTRAN_TMPDIR:: Directory for scratch files
514 * GFORTRAN_UNBUFFERED_ALL:: Don't buffer output
515 * GFORTRAN_SHOW_LOCUS:: Show location for runtime errors
516 * GFORTRAN_OPTIONAL_PLUS:: Print leading + where permitted
517 * GFORTRAN_DEFAULT_RECL:: Default record length for new files
518 * GFORTRAN_LIST_SEPARATOR:: Separator for list output
519 * GFORTRAN_CONVERT_UNIT:: Set endianness for unformatted I/O
522 @node GFORTRAN_STDIN_UNIT
523 @section @env{GFORTRAN_STDIN_UNIT}---Unit number for standard input
525 This environment variable can be used to select the unit number
526 preconnected to standard input. This must be a positive integer.
527 The default value is 5.
529 @node GFORTRAN_STDOUT_UNIT
530 @section @env{GFORTRAN_STDOUT_UNIT}---Unit number for standard output
532 This environment variable can be used to select the unit number
533 preconnected to standard output. This must be a positive integer.
534 The default value is 6.
536 @node GFORTRAN_STDERR_UNIT
537 @section @env{GFORTRAN_STDERR_UNIT}---Unit number for standard error
539 This environment variable can be used to select the unit number
540 preconnected to standard error. This must be a positive integer.
541 The default value is 0.
543 @node GFORTRAN_USE_STDERR
544 @section @env{GFORTRAN_USE_STDERR}---Send library output to standard error
546 This environment variable controls where library output is sent.
547 If the first letter is @samp{y}, @samp{Y} or @samp{1}, standard
548 error is used. If the first letter is @samp{n}, @samp{N} or
549 @samp{0}, standard output is used.
551 @node GFORTRAN_TMPDIR
552 @section @env{GFORTRAN_TMPDIR}---Directory for scratch files
554 This environment variable controls where scratch files are
555 created. If this environment variable is missing,
556 GNU Fortran searches for the environment variable @env{TMP}. If
557 this is also missing, the default is @file{/tmp}.
559 @node GFORTRAN_UNBUFFERED_ALL
560 @section @env{GFORTRAN_UNBUFFERED_ALL}---Don't buffer output
562 This environment variable controls whether all output is unbuffered.
563 If the first letter is @samp{y}, @samp{Y} or @samp{1}, all output is
564 unbuffered. This will slow down large writes. If the first letter is
565 @samp{n}, @samp{N} or @samp{0}, output is buffered. This is the
568 @node GFORTRAN_SHOW_LOCUS
569 @section @env{GFORTRAN_SHOW_LOCUS}---Show location for runtime errors
571 If the first letter is @samp{y}, @samp{Y} or @samp{1}, filename and
572 line numbers for runtime errors are printed. If the first letter is
573 @samp{n}, @samp{N} or @samp{0}, don't print filename and line numbers
574 for runtime errors. The default is to print the location.
576 @node GFORTRAN_OPTIONAL_PLUS
577 @section @env{GFORTRAN_OPTIONAL_PLUS}---Print leading + where permitted
579 If the first letter is @samp{y}, @samp{Y} or @samp{1},
580 a plus sign is printed
581 where permitted by the Fortran standard. If the first letter
582 is @samp{n}, @samp{N} or @samp{0}, a plus sign is not printed
583 in most cases. Default is not to print plus signs.
585 @node GFORTRAN_DEFAULT_RECL
586 @section @env{GFORTRAN_DEFAULT_RECL}---Default record length for new files
588 This environment variable specifies the default record length, in
589 bytes, for files which are opened without a @code{RECL} tag in the
590 @code{OPEN} statement. This must be a positive integer. The
591 default value is 1073741824 bytes (1 GB).
593 @node GFORTRAN_LIST_SEPARATOR
594 @section @env{GFORTRAN_LIST_SEPARATOR}---Separator for list output
596 This environment variable specifies the separator when writing
597 list-directed output. It may contain any number of spaces and
598 at most one comma. If you specify this on the command line,
599 be sure to quote spaces, as in
601 $ GFORTRAN_LIST_SEPARATOR=' , ' ./a.out
603 when @command{a.out} is the compiled Fortran program that you want to run.
604 Default is a single space.
606 @node GFORTRAN_CONVERT_UNIT
607 @section @env{GFORTRAN_CONVERT_UNIT}---Set endianness for unformatted I/O
609 By setting the @env{GFORTRAN_CONVERT_UNIT} variable, it is possible
610 to change the representation of data for unformatted files.
611 The syntax for the @env{GFORTRAN_CONVERT_UNIT} variable is:
613 GFORTRAN_CONVERT_UNIT: mode | mode ';' exception ;
614 mode: 'native' | 'swap' | 'big_endian' | 'little_endian' ;
615 exception: mode ':' unit_list | unit_list ;
616 unit_list: unit_spec | unit_list unit_spec ;
617 unit_spec: INTEGER | INTEGER '-' INTEGER ;
619 The variable consists of an optional default mode, followed by
620 a list of optional exceptions, which are separated by semicolons
621 from the preceding default and each other. Each exception consists
622 of a format and a comma-separated list of units. Valid values for
623 the modes are the same as for the @code{CONVERT} specifier:
626 @item @code{NATIVE} Use the native format. This is the default.
627 @item @code{SWAP} Swap between little- and big-endian.
628 @item @code{LITTLE_ENDIAN} Use the little-endian format
629 for unformatted files.
630 @item @code{BIG_ENDIAN} Use the big-endian format for unformatted files.
632 A missing mode for an exception is taken to mean @code{BIG_ENDIAN}.
633 Examples of values for @env{GFORTRAN_CONVERT_UNIT} are:
635 @item @code{'big_endian'} Do all unformatted I/O in big_endian mode.
636 @item @code{'little_endian;native:10-20,25'} Do all unformatted I/O
637 in little_endian mode, except for units 10 to 20 and 25, which are in
639 @item @code{'10-20'} Units 10 to 20 are big-endian, the rest is native.
642 Setting the environment variables should be done on the command
643 line or via the @command{export}
644 command for @command{sh}-compatible shells and via @command{setenv}
645 for @command{csh}-compatible shells.
647 Example for @command{sh}:
650 $ GFORTRAN_CONVERT_UNIT='big_endian;native:10-20' ./a.out
653 Example code for @command{csh}:
656 % setenv GFORTRAN_CONVERT_UNIT 'big_endian;native:10-20'
660 Using anything but the native representation for unformatted data
661 carries a significant speed overhead. If speed in this area matters
662 to you, it is best if you use this only for data that needs to be
665 @xref{CONVERT specifier}, for an alternative way to specify the
666 data representation for unformatted files. @xref{Runtime Options}, for
667 setting a default data representation for the whole program. The
668 @code{CONVERT} specifier overrides the @option{-fconvert} compile options.
671 @c =====================================================================
672 @c PART II: LANGUAGE REFERENCE
673 @c =====================================================================
676 \part{II}{Language Reference}
679 @c ---------------------------------------------------------------------
680 @c Fortran 2003 Status
681 @c ---------------------------------------------------------------------
683 @node Fortran 2003 status
684 @chapter Fortran 2003 Status
686 Although GNU Fortran focuses on implementing the Fortran 95
687 standard for the time being, a few Fortran 2003 features are currently
692 Intrinsics @code{command_argument_count}, @code{get_command},
693 @code{get_command_argument}, @code{get_environment_variable}, and
697 @cindex Array constructors
699 Array constructors using square brackets. That is, @code{[...]} rather
703 @cindex @code{FLUSH} statement
704 @code{FLUSH} statement.
707 @cindex @code{IOMSG=} specifier
708 @code{IOMSG=} specifier for I/O statements.
711 @cindex @code{ENUM} statement
712 @cindex @code{ENUMERATOR} statement
713 @cindex @code{-fshort-enums} option
714 Support for the declaration of enumeration constants via the
715 @code{ENUM} and @code{ENUMERATOR} statements. Interoperability with
716 @command{gcc} is guaranteed also for the case where the
717 @command{-fshort-enums} command line option is given.
724 @cindex @code{ALLOCATABLE} dummy arguments
725 @code{ALLOCATABLE} dummy arguments.
727 @cindex @code{ALLOCATABLE} function results
728 @code{ALLOCATABLE} function results
730 @cindex @code{ALLOCATABLE} components of derived types
731 @code{ALLOCATABLE} components of derived types
735 @cindex @code{STREAM} I/O
736 @cindex @code{ACCESS='STREAM'} I/O
737 The @code{OPEN} statement supports the @code{ACCESS='STREAM'} specifier,
738 allowing I/O without any record structure.
741 Namelist input/output for internal files.
744 @cindex @code{PROTECTED}
745 The @code{PROTECTED} statement and attribute.
749 The @code{VALUE} statement and attribute.
752 @cindex @code{VOLATILE}
753 The @code{VOLATILE} statement and attribute.
756 @cindex @code{IMPORT}
757 The @code{IMPORT} statement, allowing to import
758 host-associated derived types.
761 @cindex @code{USE, INTRINSIC}
762 @cindex @code{ISO_FORTRAN_ENV}
763 @code{USE} statement with @code{INTRINSIC} and @code{NON_INTRINSIC}
764 attribute; supported intrinsic modules: @code{ISO_FORTRAN_ENV},
765 @code{OMP_LIB} and @code{OMP_LIB_KINDS}.
770 @c ---------------------------------------------------------------------
772 @c ---------------------------------------------------------------------
774 @c Maybe this chapter should be merged with the 'Standards' section,
775 @c whenever that is written :-)
781 GNU Fortran implements a number of extensions over standard
782 Fortran. This chapter contains information on their syntax and
783 meaning. There are currently two categories of GNU Fortran
784 extensions, those that provide functionality beyond that provided
785 by any standard, and those that are supported by GNU Fortran
786 purely for backward compatibility with legacy compilers. By default,
787 @option{-std=gnu} allows the compiler to accept both types of
788 extensions, but to warn about the use of the latter. Specifying
789 either @option{-std=f95} or @option{-std=f2003} disables both types
790 of extensions, and @option{-std=legacy} allows both without warning.
793 * Old-style kind specifications::
794 * Old-style variable initialization::
795 * Extensions to namelist::
796 * X format descriptor without count field::
797 * Commas in FORMAT specifications::
798 * Missing period in FORMAT specifications::
800 * BOZ literal constants::
801 * Real array indices::
803 * Implicitly convert LOGICAL and INTEGER values::
804 * Hollerith constants support::
806 * CONVERT specifier::
810 @node Old-style kind specifications
811 @section Old-style kind specifications
812 @cindex Kind specifications
814 GNU Fortran allows old-style kind specifications in declarations. These
820 where @code{TYPESPEC} is a basic type (@code{INTEGER}, @code{REAL},
821 etc.), and where @code{size} is a byte count corresponding to the
822 storage size of a valid kind for that type. (For @code{COMPLEX}
823 variables, @code{size} is the total size of the real and imaginary
824 parts.) The statement then declares @code{x}, @code{y} and @code{z} to
825 be of type @code{TYPESPEC} with the appropriate kind. This is
826 equivalent to the standard-conforming declaration
831 where @code{k} is equal to @code{size} for most types, but is equal to
832 @code{size/2} for the @code{COMPLEX} type.
834 @node Old-style variable initialization
835 @section Old-style variable initialization
836 @cindex Initialization
838 GNU Fortran allows old-style initialization of variables of the
842 REAL x(2,2) /3*0.,1./
844 The syntax for the initializers is as for the @code{DATA} statement, but
845 unlike in a @code{DATA} statement, an initializer only applies to the
846 variable immediately preceding the initialization. In other words,
847 something like @code{INTEGER I,J/2,3/} is not valid. This style of
848 initialization is only allowed in declarations without double colons
849 (@code{::}); the double colons were introduced in Fortran 90, which also
850 introduced a standard syntax for initializing variables in type
853 Examples of standard-conforming code equivalent to the above example
857 INTEGER :: i = 1, j = 2
858 REAL :: x(2,2) = RESHAPE((/0.,0.,0.,1./),SHAPE(x))
862 DATA i/1/, j/2/, x/3*0.,1./
865 Note that variables which are explicitly initialized in declarations
866 or in @code{DATA} statements automatically acquire the @code{SAVE}
869 @node Extensions to namelist
870 @section Extensions to namelist
873 GNU Fortran fully supports the Fortran 95 standard for namelist I/O
874 including array qualifiers, substrings and fully qualified derived types.
875 The output from a namelist write is compatible with namelist read. The
876 output has all names in upper case and indentation to column 1 after the
877 namelist name. Two extensions are permitted:
879 Old-style use of @samp{$} instead of @samp{&}
882 X(:)%Y(2) = 1.0 2.0 3.0
887 It should be noted that the default terminator is @samp{/} rather than
890 Querying of the namelist when inputting from stdin. After at least
891 one space, entering @samp{?} sends to stdout the namelist name and the names of
892 the variables in the namelist:
903 Entering @samp{=?} outputs the namelist to stdout, as if
904 @code{WRITE(*,NML = mynml)} had been called:
909 X(1)%Y= 0.000000 , 1.000000 , 0.000000 ,
910 X(2)%Y= 0.000000 , 2.000000 , 0.000000 ,
911 X(3)%Y= 0.000000 , 3.000000 , 0.000000 ,
915 To aid this dialog, when input is from stdin, errors send their
916 messages to stderr and execution continues, even if @code{IOSTAT} is set.
918 @code{PRINT} namelist is permitted. This causes an error if
919 @option{-std=f95} is used.
922 REAL, dimension (4) :: x = (/1.0, 2.0, 3.0, 4.0/)
925 END PROGRAM test_print
928 Expanded namelist reads are permitted. This causes an error if
929 @option{-std=f95} is used. In the following example, the first element
930 of the array will be given the value 0.00 and the two succeeding
931 elements will be given the values 1.00 and 2.00.
934 X(1,1) = 0.00 , 1.00 , 2.00
938 @node X format descriptor without count field
939 @section @code{X} format descriptor without count field
940 @cindex @code{X} format descriptor without count field
942 To support legacy codes, GNU Fortran permits the count field of the
943 @code{X} edit descriptor in @code{FORMAT} statements to be omitted.
944 When omitted, the count is implicitly assumed to be one.
948 10 FORMAT (I1, X, I1)
951 @node Commas in FORMAT specifications
952 @section Commas in @code{FORMAT} specifications
953 @cindex Commas in @code{FORMAT} specifications
955 To support legacy codes, GNU Fortran allows the comma separator
956 to be omitted immediately before and after character string edit
957 descriptors in @code{FORMAT} statements.
961 10 FORMAT ('FOO='I1' BAR='I2)
965 @node Missing period in FORMAT specifications
966 @section Missing period in @code{FORMAT} specifications
967 @cindex Missing period in @code{FORMAT} specifications
969 To support legacy codes, GNU Fortran allows missing periods in format
970 specifications if and only if @option{-std=legacy} is given on the
971 command line. This is considered non-conforming code and is
981 @section I/O item lists
982 @cindex I/O item lists
984 To support legacy codes, GNU Fortran allows the input item list
985 of the @code{READ} statement, and the output item lists of the
986 @code{WRITE} and @code{PRINT} statements, to start with a comma.
988 @node BOZ literal constants
989 @section BOZ literal constants
990 @cindex BOZ literal constants
992 As an extension, GNU Fortran allows hexadecimal BOZ literal constants to
993 be specified using the X prefix, in addition to the standard Z prefix.
994 BOZ literal constants can also be specified by adding a suffix to the
995 string. For example, @code{Z'ABC'} and @code{'ABC'Z} are equivalent.
997 The Fortran standard restricts the appearance of a BOZ literal constant
998 to the @code{DATA} statement, and it is expected to be assigned to an
999 @code{INTEGER} variable. GNU Fortran permits a BOZ literal to appear in
1000 any initialization expression as well as assignment statements.
1002 Attempts to use a BOZ literal constant to do a bitwise initialization of
1003 a variable can lead to confusion. A BOZ literal constant is converted
1004 to an @code{INTEGER} value with the kind type with the largest decimal
1005 representation, and this value is then converted numerically to the type
1006 and kind of the variable in question. Thus, one should not expect a
1007 bitwise copy of the BOZ literal constant to be assigned to a @code{REAL}
1010 Similarly, initializing an @code{INTEGER} variable with a statement such
1011 as @code{DATA i/Z'FFFFFFFF'/} will produce an integer overflow rather
1012 than the desired result of @math{-1} when @code{i} is a 32-bit integer
1013 on a system that supports 64-bit integers. The @samp{-fno-range-check}
1014 option can be used as a workaround for legacy code that initializes
1015 integers in this manner.
1017 @node Real array indices
1018 @section Real array indices
1019 @cindex Real array indices
1021 As an extension, GNU Fortran allows the use of @code{REAL} expressions
1022 or variables as array indices.
1024 @node Unary operators
1025 @section Unary operators
1026 @cindex Unary operators
1028 As an extension, GNU Fortran allows unary plus and unary minus operators
1029 to appear as the second operand of binary arithmetic operators without
1030 the need for parenthesis.
1036 @node Implicitly convert LOGICAL and INTEGER values
1037 @section Implicitly convert @code{LOGICAL} and @code{INTEGER} values
1038 @cindex Implicitly convert @code{LOGICAL} and @code{INTEGER} values
1040 As an extension for backwards compatibility with other compilers, GNU
1041 Fortran allows the implicit conversion of @code{LOGICAL} values to
1042 @code{INTEGER} values and vice versa. When converting from a
1043 @code{LOGICAL} to an @code{INTEGER}, @code{.FALSE.} is interpreted as
1044 zero, and @code{.TRUE.} is interpreted as one. When converting from
1045 @code{INTEGER} to @code{LOGICAL}, the value zero is interpreted as
1046 @code{.FALSE.} and any nonzero value is interpreted as @code{.TRUE.}.
1050 IF (i) PRINT *, 'True'
1053 @node Hollerith constants support
1054 @section Hollerith constants support
1055 @cindex Hollerith constants
1057 GNU Fortran supports Hollerith constants in assignments, function
1058 arguments, and @code{DATA} and @code{ASSIGN} statements. A Hollerith
1059 constant is written as a string of characters preceded by an integer
1060 constant indicating the character count, and the letter @code{H} or
1061 @code{h}, and stored in bytewise fashion in a numeric (@code{INTEGER},
1062 @code{REAL}, or @code{complex}) or @code{LOGICAL} variable. The
1063 constant will be padded or truncated to fit the size of the variable in
1066 Examples of valid uses of Hollerith constants:
1069 data x /16Habcdefghijklmnop, 16Hqrstuvwxyz012345/
1070 x(1) = 16HABCDEFGHIJKLMNOP
1074 Invalid Hollerith constants examples:
1077 a = 8H12345678 ! Valid, but the Hollerith constant will be truncated.
1078 a = 0H ! At least one character is needed.
1081 In general, Hollerith constants were used to provide a rudimentary
1082 facility for handling character strings in early Fortran compilers,
1083 prior to the introduction of @code{CHARACTER} variables in Fortran 77;
1084 in those cases, the standard-compliant equivalent is to convert the
1085 program to use proper character strings. On occasion, there may be a
1086 case where the intent is specifically to initialize a numeric variable
1087 with a given byte sequence. In these cases, the same result can be
1088 obtained by using the @code{TRANSFER} statement, as in this example.
1090 INTEGER(KIND=4) :: a
1091 a = TRANSFER ("abcd", a) ! equivalent to: a = 4Habcd
1096 @section Cray pointers
1097 @cindex Cray pointers
1099 Cray pointers are part of a non-standard extension that provides a
1100 C-like pointer in Fortran. This is accomplished through a pair of
1101 variables: an integer "pointer" that holds a memory address, and a
1102 "pointee" that is used to dereference the pointer.
1104 Pointer/pointee pairs are declared in statements of the form:
1106 pointer ( <pointer> , <pointee> )
1110 pointer ( <pointer1> , <pointee1> ), ( <pointer2> , <pointee2> ), ...
1112 The pointer is an integer that is intended to hold a memory address.
1113 The pointee may be an array or scalar. A pointee can be an assumed
1114 size array---that is, the last dimension may be left unspecified by
1115 using a @code{*} in place of a value---but a pointee cannot be an
1116 assumed shape array. No space is allocated for the pointee.
1118 The pointee may have its type declared before or after the pointer
1119 statement, and its array specification (if any) may be declared
1120 before, during, or after the pointer statement. The pointer may be
1121 declared as an integer prior to the pointer statement. However, some
1122 machines have default integer sizes that are different than the size
1123 of a pointer, and so the following code is not portable:
1128 If a pointer is declared with a kind that is too small, the compiler
1129 will issue a warning; the resulting binary will probably not work
1130 correctly, because the memory addresses stored in the pointers may be
1131 truncated. It is safer to omit the first line of the above example;
1132 if explicit declaration of ipt's type is omitted, then the compiler
1133 will ensure that ipt is an integer variable large enough to hold a
1136 Pointer arithmetic is valid with Cray pointers, but it is not the same
1137 as C pointer arithmetic. Cray pointers are just ordinary integers, so
1138 the user is responsible for determining how many bytes to add to a
1139 pointer in order to increment it. Consider the following example:
1143 pointer (ipt, pointee)
1147 The last statement does not set @code{ipt} to the address of
1148 @code{target(1)}, as it would in C pointer arithmetic. Adding @code{1}
1149 to @code{ipt} just adds one byte to the address stored in @code{ipt}.
1151 Any expression involving the pointee will be translated to use the
1152 value stored in the pointer as the base address.
1154 To get the address of elements, this extension provides an intrinsic
1155 function @code{LOC()}. The @code{LOC()} function is equivalent to the
1156 @code{&} operator in C, except the address is cast to an integer type:
1159 pointer(ipt, arpte(10))
1161 ipt = loc(ar) ! Makes arpte is an alias for ar
1162 arpte(1) = 1.0 ! Sets ar(1) to 1.0
1164 The pointer can also be set by a call to the @code{MALLOC} intrinsic
1167 Cray pointees often are used to alias an existing variable. For
1175 As long as @code{ipt} remains unchanged, @code{iarr} is now an alias for
1176 @code{target}. The optimizer, however, will not detect this aliasing, so
1177 it is unsafe to use @code{iarr} and @code{target} simultaneously. Using
1178 a pointee in any way that violates the Fortran aliasing rules or
1179 assumptions is illegal. It is the user's responsibility to avoid doing
1180 this; the compiler works under the assumption that no such aliasing
1183 Cray pointers will work correctly when there is no aliasing (i.e., when
1184 they are used to access a dynamically allocated block of memory), and
1185 also in any routine where a pointee is used, but any variable with which
1186 it shares storage is not used. Code that violates these rules may not
1187 run as the user intends. This is not a bug in the optimizer; any code
1188 that violates the aliasing rules is illegal. (Note that this is not
1189 unique to GNU Fortran; any Fortran compiler that supports Cray pointers
1190 will ``incorrectly'' optimize code with illegal aliasing.)
1192 There are a number of restrictions on the attributes that can be applied
1193 to Cray pointers and pointees. Pointees may not have the
1194 @code{ALLOCATABLE}, @code{INTENT}, @code{OPTIONAL}, @code{DUMMY},
1195 @code{TARGET}, @code{INTRINSIC}, or @code{POINTER} attributes. Pointers
1196 may not have the @code{DIMENSION}, @code{POINTER}, @code{TARGET},
1197 @code{ALLOCATABLE}, @code{EXTERNAL}, or @code{INTRINSIC} attributes.
1198 Pointees may not occur in more than one pointer statement. A pointee
1199 cannot be a pointer. Pointees cannot occur in equivalence, common, or
1202 A Cray pointer may also point to a function or a subroutine. For
1203 example, the following excerpt is valid:
1207 pointer (subptr,subpte)
1217 A pointer may be modified during the course of a program, and this
1218 will change the location to which the pointee refers. However, when
1219 pointees are passed as arguments, they are treated as ordinary
1220 variables in the invoked function. Subsequent changes to the pointer
1221 will not change the base address of the array that was passed.
1223 @node CONVERT specifier
1224 @section CONVERT specifier
1225 @cindex CONVERT specifier
1227 GNU Fortran allows the conversion of unformatted data between little-
1228 and big-endian representation to facilitate moving of data
1229 between different systems. The conversion can be indicated with
1230 the @code{CONVERT} specifier on the @code{OPEN} statement.
1231 @xref{GFORTRAN_CONVERT_UNIT}, for an alternative way of specifying
1232 the data format via an environment variable.
1234 Valid values for @code{CONVERT} are:
1236 @item @code{CONVERT='NATIVE'} Use the native format. This is the default.
1237 @item @code{CONVERT='SWAP'} Swap between little- and big-endian.
1238 @item @code{CONVERT='LITTLE_ENDIAN'} Use the little-endian representation
1239 for unformatted files.
1240 @item @code{CONVERT='BIG_ENDIAN'} Use the big-endian representation for
1244 Using the option could look like this:
1246 open(file='big.dat',form='unformatted',access='sequential', &
1247 convert='big_endian')
1250 The value of the conversion can be queried by using
1251 @code{INQUIRE(CONVERT=ch)}. The values returned are
1252 @code{'BIG_ENDIAN'} and @code{'LITTLE_ENDIAN'}.
1254 @code{CONVERT} works between big- and little-endian for
1255 @code{INTEGER} values of all supported kinds and for @code{REAL}
1256 on IEEE systems of kinds 4 and 8. Conversion between different
1257 ``extended double'' types on different architectures such as
1258 m68k and x86_64, which GNU Fortran
1259 supports as @code{REAL(KIND=10)} and @code{REAL(KIND=16)}, will
1262 @emph{Note that the values specified via the GFORTRAN_CONVERT_UNIT
1263 environment variable will override the CONVERT specifier in the
1264 open statement}. This is to give control over data formats to
1265 users who do not have the source code of their program available.
1267 Using anything but the native representation for unformatted data
1268 carries a significant speed overhead. If speed in this area matters
1269 to you, it is best if you use this only for data that needs to be
1276 GNU Fortran attempts to be OpenMP Application Program Interface v2.5
1277 compatible when invoked with the @option{-fopenmp} option. GNU Fortran
1278 then generates parallelized code according to the OpenMP directives
1279 used in the source. The OpenMP Fortran runtime library
1280 routines are provided both in a form of a Fortran 90 module named
1281 @code{omp_lib} and in a form of a Fortran @code{include} file named
1284 For details refer to the actual
1285 @uref{http://www.openmp.org/drupal/mp-documents/spec25.pdf,
1286 OpenMP Application Program Interface v2.5} specification.
1288 @c ---------------------------------------------------------------------
1289 @c Intrinsic Procedures
1290 @c ---------------------------------------------------------------------
1292 @include intrinsic.texi
1299 @c ---------------------------------------------------------------------
1301 @c ---------------------------------------------------------------------
1304 @unnumbered Contributing
1305 @cindex Contributing
1307 Free software is only possible if people contribute to efforts
1309 We're always in need of more people helping out with ideas
1310 and comments, writing documentation and contributing code.
1312 If you want to contribute to GNU Fortran,
1313 have a look at the long lists of projects you can take on.
1314 Some of these projects are small,
1315 some of them are large;
1316 some are completely orthogonal to the rest of what is
1317 happening on GNU Fortran,
1318 but others are ``mainstream'' projects in need of enthusiastic hackers.
1319 All of these projects are important!
1320 We'll eventually get around to the things here,
1321 but they are also things doable by someone who is willing and able.
1326 * Proposed Extensions::
1331 @section Contributors to GNU Fortran
1332 @cindex Contributors
1336 Most of the parser was hand-crafted by @emph{Andy Vaught}, who is
1337 also the initiator of the whole project. Thanks Andy!
1338 Most of the interface with GCC was written by @emph{Paul Brook}.
1340 The following individuals have contributed code and/or
1341 ideas and significant help to the GNU Fortran project
1342 (in no particular order):
1346 @item Katherine Holcomb
1347 @item Tobias Schl@"uter
1348 @item Steven Bosscher
1351 @item Niels Kristian Bech Jensen
1352 @item Steven Johnson
1357 @item Fran@,{c}ois-Xavier Coudert
1358 @item Steven G. Kargl
1360 @item Janne Blomqvist
1367 @item Richard Henderson
1368 @item Richard Sandiford
1369 @item Richard Guenther
1370 @item Bernhard Fischer
1373 The following people have contributed bug reports,
1374 smaller or larger patches,
1375 and much needed feedback and encouragement for the
1376 GNU Fortran project:
1379 @item Erik Schnetter
1384 Many other individuals have helped debug,
1385 test and improve the GNU Fortran compiler over the past few years,
1386 and we welcome you to do the same!
1387 If you already have done so,
1388 and you would like to see your name listed in the
1389 list above, please contact us.
1397 @item Help build the test suite
1398 Solicit more code for donation to the test suite.
1399 We can keep code private on request.
1401 @item Bug hunting/squishing
1402 Find bugs and write more test cases!
1403 Test cases are especially very welcome,
1404 because it allows us to concentrate on fixing bugs
1405 instead of isolating them.
1407 @item Smaller projects (``bug'' fixes):
1409 @item Allow init exprs to be numbers raised to integer powers.
1410 @item Implement correct rounding.
1411 @item Implement F restrictions on Fortran 95 syntax.
1412 @item See about making Emacs-parsable error messages.
1416 If you wish to work on the runtime libraries,
1417 please contact a project maintainer.
1421 @node Proposed Extensions
1422 @section Proposed Extensions
1424 Here's a list of proposed extensions for the GNU Fortran compiler, in no particular
1425 order. Most of these are necessary to be fully compatible with
1426 existing Fortran compilers, but they are not part of the official
1427 J3 Fortran 95 standard.
1429 @subsection Compiler extensions:
1432 User-specified alignment rules for structures.
1435 Flag to generate @code{Makefile} info.
1438 Automatically extend single precision constants to double.
1441 Compile code that conserves memory by dynamically allocating common and
1442 module storage either on stack or heap.
1445 Compile flag to generate code for array conformance checking (suggest -CC).
1448 User control of symbol names (underscores, etc).
1451 Compile setting for maximum size of stack frame size before spilling
1452 parts to static or heap.
1455 Flag to force local variables into static space.
1458 Flag to force local variables onto stack.
1461 Flag for maximum errors before ending compile.
1464 Option to initialize otherwise uninitialized integer and floating
1469 @subsection Environment Options
1472 Pluggable library modules for random numbers, linear algebra.
1473 LA should use BLAS calling conventions.
1476 Environment variables controlling actions on arithmetic exceptions like
1477 overflow, underflow, precision loss---Generate NaN, abort, default.
1481 Set precision for fp units that support it (i387).
1484 Variable for setting fp rounding mode.
1487 Variable to fill uninitialized variables with a user-defined bit
1491 Environment variable controlling filename that is opened for that unit
1495 Environment variable to clear/trash memory being freed.
1498 Environment variable to control tracing of allocations and frees.
1501 Environment variable to display allocated memory at normal program end.
1504 Environment variable for filename for * IO-unit.
1507 Environment variable for temporary file directory.
1510 Environment variable forcing standard output to be line buffered (unix).
1515 @c ---------------------------------------------------------------------
1516 @c GNU General Public License
1517 @c ---------------------------------------------------------------------
1523 @c ---------------------------------------------------------------------
1524 @c GNU Free Documentation License
1525 @c ---------------------------------------------------------------------
1531 @c ---------------------------------------------------------------------
1532 @c Funding Free Software
1533 @c ---------------------------------------------------------------------
1535 @include funding.texi
1537 @c ---------------------------------------------------------------------
1539 @c ---------------------------------------------------------------------