1 \input texinfo @c -*-texinfo-*-
3 @setfilename gfortran.info
4 @set copyrights-gfortran 1999-2013
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 1.3 or
84 any later version published by the Free Software Foundation; with the
85 Invariant Sections being ``Funding Free Software'', the Front-Cover
86 Texts being (a) (see below), and with the Back-Cover Texts being (b)
87 (see below). A copy of the license is included in the section entitled
88 ``GNU Free Documentation License''.
90 (a) The FSF's Front-Cover Text is:
94 (b) The FSF's Back-Cover Text is:
96 You have freedom to copy and modify this GNU Manual, like GNU
97 software. Copies published by the Free Software Foundation raise
98 funds for GNU development.
102 @dircategory Software development
104 * gfortran: (gfortran). The GNU Fortran Compiler.
106 This file documents the use and the internals of
107 the GNU Fortran compiler, (@command{gfortran}).
109 Published by the Free Software Foundation
110 51 Franklin Street, Fifth Floor
111 Boston, MA 02110-1301 USA
117 @setchapternewpage odd
119 @title Using GNU Fortran
121 @author The @t{gfortran} team
123 @vskip 0pt plus 1filll
124 Published by the Free Software Foundation@*
125 51 Franklin Street, Fifth Floor@*
126 Boston, MA 02110-1301, USA@*
127 @c Last printed ??ber, 19??.@*
128 @c Printed copies are available for $? each.@*
134 @c TODO: The following "Part" definitions are included here temporarily
135 @c until they are incorporated into the official Texinfo distribution.
138 \global\let\partentry=\dosmallpartentry
139 \global\let\blankpartentry=\dosmallblankpartentry
144 \global\let\partentry=\dopartentry
145 \global\let\blankpartentry=\doblankpartentry
151 @c ---------------------------------------------------------------------
152 @c TexInfo table of contents.
153 @c ---------------------------------------------------------------------
160 This manual documents the use of @command{gfortran},
161 the GNU Fortran compiler. You can find in this manual how to invoke
162 @command{gfortran}, as well as its features and incompatibilities.
165 @emph{Warning:} This document, and the compiler it describes, are still
166 under development. While efforts are made to keep it up-to-date, it might
167 not accurately reflect the status of the most recent GNU Fortran compiler.
171 @comment When you add a new menu item, please keep the right hand
172 @comment aligned to the same column. Do not use tabs. This provides
173 @comment better formatting.
178 Part I: Invoking GNU Fortran
179 * Invoking GNU Fortran:: Command options supported by @command{gfortran}.
180 * Runtime:: Influencing runtime behavior with environment variables.
182 Part II: Language Reference
183 * Fortran 2003 and 2008 status:: Fortran 2003 and 2008 features supported by GNU Fortran.
184 * Compiler Characteristics:: User-visible implementation details.
185 * Extensions:: Language extensions implemented by GNU Fortran.
186 * Mixed-Language Programming:: Interoperability with C
187 * Intrinsic Procedures:: Intrinsic procedures supported by GNU Fortran.
188 * Intrinsic Modules:: Intrinsic modules supported by GNU Fortran.
190 * Contributing:: How you can help.
191 * Copying:: GNU General Public License says
192 how you can copy and share GNU Fortran.
193 * GNU Free Documentation License::
194 How you can copy and share this manual.
195 * Funding:: How to help assure continued work for free software.
196 * Option Index:: Index of command line options
197 * Keyword Index:: Index of concepts
201 @c ---------------------------------------------------------------------
203 @c ---------------------------------------------------------------------
206 @chapter Introduction
208 @c The following duplicates the text on the TexInfo table of contents.
210 This manual documents the use of @command{gfortran}, the GNU Fortran
211 compiler. You can find in this manual how to invoke @command{gfortran},
212 as well as its features and incompatibilities.
215 @emph{Warning:} This document, and the compiler it describes, are still
216 under development. While efforts are made to keep it up-to-date, it
217 might not accurately reflect the status of the most recent GNU Fortran
222 The GNU Fortran compiler front end was
223 designed initially as a free replacement for,
224 or alternative to, the Unix @command{f95} command;
225 @command{gfortran} is the command you will use to invoke the compiler.
228 * About GNU Fortran:: What you should know about the GNU Fortran compiler.
229 * GNU Fortran and GCC:: You can compile Fortran, C, or other programs.
230 * Preprocessing and conditional compilation:: The Fortran preprocessor
231 * GNU Fortran and G77:: Why we chose to start from scratch.
232 * Project Status:: Status of GNU Fortran, roadmap, proposed extensions.
233 * Standards:: Standards supported by GNU Fortran.
237 @c ---------------------------------------------------------------------
239 @c ---------------------------------------------------------------------
241 @node About GNU Fortran
242 @section About GNU Fortran
244 The GNU Fortran compiler supports the Fortran 77, 90 and 95 standards
245 completely, parts of the Fortran 2003 and Fortran 2008 standards, and
246 several vendor extensions. The development goal is to provide the
251 Read a user's program,
252 stored in a file and containing instructions written
253 in Fortran 77, Fortran 90, Fortran 95, Fortran 2003 or Fortran 2008.
254 This file contains @dfn{source code}.
257 Translate the user's program into instructions a computer
258 can carry out more quickly than it takes to translate the
259 instructions in the first
260 place. The result after compilation of a program is
262 code designed to be efficiently translated and processed
263 by a machine such as your computer.
264 Humans usually are not as good writing machine code
265 as they are at writing Fortran (or C++, Ada, or Java),
266 because it is easy to make tiny mistakes writing machine code.
269 Provide the user with information about the reasons why
270 the compiler is unable to create a binary from the source code.
271 Usually this will be the case if the source code is flawed.
272 The Fortran 90 standard requires that the compiler can point out
273 mistakes to the user.
274 An incorrect usage of the language causes an @dfn{error message}.
276 The compiler will also attempt to diagnose cases where the
277 user's program contains a correct usage of the language,
278 but instructs the computer to do something questionable.
279 This kind of diagnostics message is called a @dfn{warning message}.
282 Provide optional information about the translation passes
283 from the source code to machine code.
284 This can help a user of the compiler to find the cause of
285 certain bugs which may not be obvious in the source code,
286 but may be more easily found at a lower level compiler output.
287 It also helps developers to find bugs in the compiler itself.
290 Provide information in the generated machine code that can
291 make it easier to find bugs in the program (using a debugging tool,
292 called a @dfn{debugger}, such as the GNU Debugger @command{gdb}).
295 Locate and gather machine code already generated to
296 perform actions requested by statements in the user's program.
297 This machine code is organized into @dfn{modules} and is located
298 and @dfn{linked} to the user program.
301 The GNU Fortran compiler consists of several components:
305 A version of the @command{gcc} command
306 (which also might be installed as the system's @command{cc} command)
307 that also understands and accepts Fortran source code.
308 The @command{gcc} command is the @dfn{driver} program for
309 all the languages in the GNU Compiler Collection (GCC);
311 you can compile the source code of any language for
312 which a front end is available in GCC.
315 The @command{gfortran} command itself,
316 which also might be installed as the
317 system's @command{f95} command.
318 @command{gfortran} is just another driver program,
319 but specifically for the Fortran compiler only.
320 The difference with @command{gcc} is that @command{gfortran}
321 will automatically link the correct libraries to your program.
324 A collection of run-time libraries.
325 These libraries contain the machine code needed to support
326 capabilities of the Fortran language that are not directly
327 provided by the machine code generated by the
328 @command{gfortran} compilation phase,
329 such as intrinsic functions and subroutines,
330 and routines for interaction with files and the operating system.
331 @c and mechanisms to spawn,
332 @c unleash and pause threads in parallelized code.
335 The Fortran compiler itself, (@command{f951}).
336 This is the GNU Fortran parser and code generator,
337 linked to and interfaced with the GCC backend library.
338 @command{f951} ``translates'' the source code to
339 assembler code. You would typically not use this
341 instead, the @command{gcc} or @command{gfortran} driver
342 programs will call it for you.
346 @c ---------------------------------------------------------------------
347 @c GNU Fortran and GCC
348 @c ---------------------------------------------------------------------
350 @node GNU Fortran and GCC
351 @section GNU Fortran and GCC
352 @cindex GNU Compiler Collection
355 GNU Fortran is a part of GCC, the @dfn{GNU Compiler Collection}. GCC
356 consists of a collection of front ends for various languages, which
357 translate the source code into a language-independent form called
358 @dfn{GENERIC}. This is then processed by a common middle end which
359 provides optimization, and then passed to one of a collection of back
360 ends which generate code for different computer architectures and
363 Functionally, this is implemented with a driver program (@command{gcc})
364 which provides the command-line interface for the compiler. It calls
365 the relevant compiler front-end program (e.g., @command{f951} for
366 Fortran) for each file in the source code, and then calls the assembler
367 and linker as appropriate to produce the compiled output. In a copy of
368 GCC which has been compiled with Fortran language support enabled,
369 @command{gcc} will recognize files with @file{.f}, @file{.for}, @file{.ftn},
370 @file{.f90}, @file{.f95}, @file{.f03} and @file{.f08} extensions as
371 Fortran source code, and compile it accordingly. A @command{gfortran}
372 driver program is also provided, which is identical to @command{gcc}
373 except that it automatically links the Fortran runtime libraries into the
376 Source files with @file{.f}, @file{.for}, @file{.fpp}, @file{.ftn}, @file{.F},
377 @file{.FOR}, @file{.FPP}, and @file{.FTN} extensions are treated as fixed form.
378 Source files with @file{.f90}, @file{.f95}, @file{.f03}, @file{.f08},
379 @file{.F90}, @file{.F95}, @file{.F03} and @file{.F08} extensions are
380 treated as free form. The capitalized versions of either form are run
381 through preprocessing. Source files with the lower case @file{.fpp}
382 extension are also run through preprocessing.
384 This manual specifically documents the Fortran front end, which handles
385 the programming language's syntax and semantics. The aspects of GCC
386 which relate to the optimization passes and the back-end code generation
387 are documented in the GCC manual; see
388 @ref{Top,,Introduction,gcc,Using the GNU Compiler Collection (GCC)}.
389 The two manuals together provide a complete reference for the GNU
393 @c ---------------------------------------------------------------------
394 @c Preprocessing and conditional compilation
395 @c ---------------------------------------------------------------------
397 @node Preprocessing and conditional compilation
398 @section Preprocessing and conditional compilation
401 @cindex Conditional compilation
402 @cindex Preprocessing
403 @cindex preprocessor, include file handling
405 Many Fortran compilers including GNU Fortran allow passing the source code
406 through a C preprocessor (CPP; sometimes also called the Fortran preprocessor,
407 FPP) to allow for conditional compilation. In the case of GNU Fortran,
408 this is the GNU C Preprocessor in the traditional mode. On systems with
409 case-preserving file names, the preprocessor is automatically invoked if the
410 filename extension is @file{.F}, @file{.FOR}, @file{.FTN}, @file{.fpp},
411 @file{.FPP}, @file{.F90}, @file{.F95}, @file{.F03} or @file{.F08}. To manually
412 invoke the preprocessor on any file, use @option{-cpp}, to disable
413 preprocessing on files where the preprocessor is run automatically, use
416 If a preprocessed file includes another file with the Fortran @code{INCLUDE}
417 statement, the included file is not preprocessed. To preprocess included
418 files, use the equivalent preprocessor statement @code{#include}.
420 If GNU Fortran invokes the preprocessor, @code{__GFORTRAN__}
421 is defined and @code{__GNUC__}, @code{__GNUC_MINOR__} and
422 @code{__GNUC_PATCHLEVEL__} can be used to determine the version of the
423 compiler. See @ref{Top,,Overview,cpp,The C Preprocessor} for details.
425 While CPP is the de-facto standard for preprocessing Fortran code,
426 Part 3 of the Fortran 95 standard (ISO/IEC 1539-3:1998) defines
427 Conditional Compilation, which is not widely used and not directly
428 supported by the GNU Fortran compiler. You can use the program coco
429 to preprocess such files (@uref{http://www.daniellnagle.com/coco.html}).
432 @c ---------------------------------------------------------------------
433 @c GNU Fortran and G77
434 @c ---------------------------------------------------------------------
436 @node GNU Fortran and G77
437 @section GNU Fortran and G77
439 @cindex @command{g77}
441 The GNU Fortran compiler is the successor to @command{g77}, the Fortran
442 77 front end included in GCC prior to version 4. It is an entirely new
443 program that has been designed to provide Fortran 95 support and
444 extensibility for future Fortran language standards, as well as providing
445 backwards compatibility for Fortran 77 and nearly all of the GNU language
446 extensions supported by @command{g77}.
449 @c ---------------------------------------------------------------------
451 @c ---------------------------------------------------------------------
454 @section Project Status
457 As soon as @command{gfortran} can parse all of the statements correctly,
458 it will be in the ``larva'' state.
459 When we generate code, the ``puppa'' state.
460 When @command{gfortran} is done,
461 we'll see if it will be a beautiful butterfly,
462 or just a big bug....
464 --Andy Vaught, April 2000
467 The start of the GNU Fortran 95 project was announced on
468 the GCC homepage in March 18, 2000
469 (even though Andy had already been working on it for a while,
472 The GNU Fortran compiler is able to compile nearly all
473 standard-compliant Fortran 95, Fortran 90, and Fortran 77 programs,
474 including a number of standard and non-standard extensions, and can be
475 used on real-world programs. In particular, the supported extensions
476 include OpenMP, Cray-style pointers, and several Fortran 2003 and Fortran
477 2008 features, including TR 15581. However, it is still under
478 development and has a few remaining rough edges.
480 At present, the GNU Fortran compiler passes the
481 @uref{http://www.fortran-2000.com/ArnaudRecipes/fcvs21_f95.html,
482 NIST Fortran 77 Test Suite}, and produces acceptable results on the
483 @uref{http://www.netlib.org/lapack/faq.html#1.21, LAPACK Test Suite}.
484 It also provides respectable performance on
485 the @uref{http://www.polyhedron.com/pb05.html, Polyhedron Fortran
486 compiler benchmarks} and the
487 @uref{http://www.llnl.gov/asci_benchmarks/asci/limited/lfk/README.html,
488 Livermore Fortran Kernels test}. It has been used to compile a number of
489 large real-world programs, including
490 @uref{http://mysite.verizon.net/serveall/moene.pdf, the HIRLAM
491 weather-forecasting code} and
492 @uref{http://www.theochem.uwa.edu.au/tonto/, the Tonto quantum
493 chemistry package}; see @url{http://gcc.gnu.org/@/wiki/@/GfortranApps} for an
496 Among other things, the GNU Fortran compiler is intended as a replacement
497 for G77. At this point, nearly all programs that could be compiled with
498 G77 can be compiled with GNU Fortran, although there are a few minor known
501 The primary work remaining to be done on GNU Fortran falls into three
502 categories: bug fixing (primarily regarding the treatment of invalid code
503 and providing useful error messages), improving the compiler optimizations
504 and the performance of compiled code, and extending the compiler to support
505 future standards---in particular, Fortran 2003 and Fortran 2008.
508 @c ---------------------------------------------------------------------
510 @c ---------------------------------------------------------------------
517 * Varying Length Character Strings::
520 The GNU Fortran compiler implements
521 ISO/IEC 1539:1997 (Fortran 95). As such, it can also compile essentially all
522 standard-compliant Fortran 90 and Fortran 77 programs. It also supports
523 the ISO/IEC TR-15581 enhancements to allocatable arrays.
525 GNU Fortran also have a partial support for ISO/IEC 1539-1:2004 (Fortran
526 2003), ISO/IEC 1539-1:2010 (Fortran 2008), the Technical Specification
527 @code{Further Interoperability of Fortran with C} (ISO/IEC TS 29113:2012).
528 Full support of those standards and future Fortran standards is planned.
529 The current status of the support is can be found in the
530 @ref{Fortran 2003 status}, @ref{Fortran 2008 status} and
531 @ref{TS 29113 status} sections of the documentation.
533 Additionally, the GNU Fortran compilers supports the OpenMP specification
534 (version 3.1, @url{http://openmp.org/@/wp/@/openmp-specifications/}).
536 @node Varying Length Character Strings
537 @subsection Varying Length Character Strings
538 @cindex Varying length character strings
539 @cindex Varying length strings
540 @cindex strings, varying length
542 The Fortran 95 standard specifies in Part 2 (ISO/IEC 1539-2:2000)
543 varying length character strings. While GNU Fortran currently does not
544 support such strings directly, there exist two Fortran implementations
545 for them, which work with GNU Fortran. They can be found at
546 @uref{http://www.fortran.com/@/iso_varying_string.f95} and at
547 @uref{ftp://ftp.nag.co.uk/@/sc22wg5/@/ISO_VARYING_STRING/}.
549 Deferred-length character strings of Fortran 2003 supports part of
550 the features of @code{ISO_VARYING_STRING} and should be considered as
551 replacement. (Namely, allocatable or pointers of the type
552 @code{character(len=:)}.)
555 @c =====================================================================
556 @c PART I: INVOCATION REFERENCE
557 @c =====================================================================
560 \part{I}{Invoking GNU Fortran}
563 @c ---------------------------------------------------------------------
565 @c ---------------------------------------------------------------------
570 @c ---------------------------------------------------------------------
572 @c ---------------------------------------------------------------------
575 @chapter Runtime: Influencing runtime behavior with environment variables
576 @cindex environment variable
578 The behavior of the @command{gfortran} can be influenced by
579 environment variables.
581 Malformed environment variables are silently ignored.
584 * TMPDIR:: Directory for scratch files
585 * GFORTRAN_STDIN_UNIT:: Unit number for standard input
586 * GFORTRAN_STDOUT_UNIT:: Unit number for standard output
587 * GFORTRAN_STDERR_UNIT:: Unit number for standard error
588 * GFORTRAN_UNBUFFERED_ALL:: Do not buffer I/O for all units.
589 * GFORTRAN_UNBUFFERED_PRECONNECTED:: Do not buffer I/O for preconnected units.
590 * GFORTRAN_SHOW_LOCUS:: Show location for runtime errors
591 * GFORTRAN_OPTIONAL_PLUS:: Print leading + where permitted
592 * GFORTRAN_DEFAULT_RECL:: Default record length for new files
593 * GFORTRAN_LIST_SEPARATOR:: Separator for list output
594 * GFORTRAN_CONVERT_UNIT:: Set endianness for unformatted I/O
595 * GFORTRAN_ERROR_BACKTRACE:: Show backtrace on run-time errors
599 @section @env{TMPDIR}---Directory for scratch files
601 When opening a file with @code{STATUS='SCRATCH'}, GNU Fortran tries to
602 create the file in one of the potential directories by testing each
603 directory in the order below.
607 The environment variable @env{TMPDIR}, if it exists.
610 On the MinGW target, the directory returned by the @code{GetTempPath}
611 function. Alternatively, on the Cygwin target, the @env{TMP} and
612 @env{TEMP} environment variables, if they exist, in that order.
615 The @code{P_tmpdir} macro if it is defined, otherwise the directory
619 @node GFORTRAN_STDIN_UNIT
620 @section @env{GFORTRAN_STDIN_UNIT}---Unit number for standard input
622 This environment variable can be used to select the unit number
623 preconnected to standard input. This must be a positive integer.
624 The default value is 5.
626 @node GFORTRAN_STDOUT_UNIT
627 @section @env{GFORTRAN_STDOUT_UNIT}---Unit number for standard output
629 This environment variable can be used to select the unit number
630 preconnected to standard output. This must be a positive integer.
631 The default value is 6.
633 @node GFORTRAN_STDERR_UNIT
634 @section @env{GFORTRAN_STDERR_UNIT}---Unit number for standard error
636 This environment variable can be used to select the unit number
637 preconnected to standard error. This must be a positive integer.
638 The default value is 0.
640 @node GFORTRAN_UNBUFFERED_ALL
641 @section @env{GFORTRAN_UNBUFFERED_ALL}---Do not buffer I/O on all units
643 This environment variable controls whether all I/O is unbuffered. If
644 the first letter is @samp{y}, @samp{Y} or @samp{1}, all I/O is
645 unbuffered. This will slow down small sequential reads and writes. If
646 the first letter is @samp{n}, @samp{N} or @samp{0}, I/O is buffered.
649 @node GFORTRAN_UNBUFFERED_PRECONNECTED
650 @section @env{GFORTRAN_UNBUFFERED_PRECONNECTED}---Do not buffer I/O on preconnected units
652 The environment variable named @env{GFORTRAN_UNBUFFERED_PRECONNECTED} controls
653 whether I/O on a preconnected unit (i.e.@: STDOUT or STDERR) is unbuffered. If
654 the first letter is @samp{y}, @samp{Y} or @samp{1}, I/O is unbuffered. This
655 will slow down small sequential reads and writes. If the first letter
656 is @samp{n}, @samp{N} or @samp{0}, I/O is buffered. This is the default.
658 @node GFORTRAN_SHOW_LOCUS
659 @section @env{GFORTRAN_SHOW_LOCUS}---Show location for runtime errors
661 If the first letter is @samp{y}, @samp{Y} or @samp{1}, filename and
662 line numbers for runtime errors are printed. If the first letter is
663 @samp{n}, @samp{N} or @samp{0}, do not print filename and line numbers
664 for runtime errors. The default is to print the location.
666 @node GFORTRAN_OPTIONAL_PLUS
667 @section @env{GFORTRAN_OPTIONAL_PLUS}---Print leading + where permitted
669 If the first letter is @samp{y}, @samp{Y} or @samp{1},
670 a plus sign is printed
671 where permitted by the Fortran standard. If the first letter
672 is @samp{n}, @samp{N} or @samp{0}, a plus sign is not printed
673 in most cases. Default is not to print plus signs.
675 @node GFORTRAN_DEFAULT_RECL
676 @section @env{GFORTRAN_DEFAULT_RECL}---Default record length for new files
678 This environment variable specifies the default record length, in
679 bytes, for files which are opened without a @code{RECL} tag in the
680 @code{OPEN} statement. This must be a positive integer. The
681 default value is 1073741824 bytes (1 GB).
683 @node GFORTRAN_LIST_SEPARATOR
684 @section @env{GFORTRAN_LIST_SEPARATOR}---Separator for list output
686 This environment variable specifies the separator when writing
687 list-directed output. It may contain any number of spaces and
688 at most one comma. If you specify this on the command line,
689 be sure to quote spaces, as in
691 $ GFORTRAN_LIST_SEPARATOR=' , ' ./a.out
693 when @command{a.out} is the compiled Fortran program that you want to run.
694 Default is a single space.
696 @node GFORTRAN_CONVERT_UNIT
697 @section @env{GFORTRAN_CONVERT_UNIT}---Set endianness for unformatted I/O
699 By setting the @env{GFORTRAN_CONVERT_UNIT} variable, it is possible
700 to change the representation of data for unformatted files.
701 The syntax for the @env{GFORTRAN_CONVERT_UNIT} variable is:
703 GFORTRAN_CONVERT_UNIT: mode | mode ';' exception | exception ;
704 mode: 'native' | 'swap' | 'big_endian' | 'little_endian' ;
705 exception: mode ':' unit_list | unit_list ;
706 unit_list: unit_spec | unit_list unit_spec ;
707 unit_spec: INTEGER | INTEGER '-' INTEGER ;
709 The variable consists of an optional default mode, followed by
710 a list of optional exceptions, which are separated by semicolons
711 from the preceding default and each other. Each exception consists
712 of a format and a comma-separated list of units. Valid values for
713 the modes are the same as for the @code{CONVERT} specifier:
716 @item @code{NATIVE} Use the native format. This is the default.
717 @item @code{SWAP} Swap between little- and big-endian.
718 @item @code{LITTLE_ENDIAN} Use the little-endian format
719 for unformatted files.
720 @item @code{BIG_ENDIAN} Use the big-endian format for unformatted files.
722 A missing mode for an exception is taken to mean @code{BIG_ENDIAN}.
723 Examples of values for @env{GFORTRAN_CONVERT_UNIT} are:
725 @item @code{'big_endian'} Do all unformatted I/O in big_endian mode.
726 @item @code{'little_endian;native:10-20,25'} Do all unformatted I/O
727 in little_endian mode, except for units 10 to 20 and 25, which are in
729 @item @code{'10-20'} Units 10 to 20 are big-endian, the rest is native.
732 Setting the environment variables should be done on the command
733 line or via the @command{export}
734 command for @command{sh}-compatible shells and via @command{setenv}
735 for @command{csh}-compatible shells.
737 Example for @command{sh}:
740 $ GFORTRAN_CONVERT_UNIT='big_endian;native:10-20' ./a.out
743 Example code for @command{csh}:
746 % setenv GFORTRAN_CONVERT_UNIT 'big_endian;native:10-20'
750 Using anything but the native representation for unformatted data
751 carries a significant speed overhead. If speed in this area matters
752 to you, it is best if you use this only for data that needs to be
755 @xref{CONVERT specifier}, for an alternative way to specify the
756 data representation for unformatted files. @xref{Runtime Options}, for
757 setting a default data representation for the whole program. The
758 @code{CONVERT} specifier overrides the @option{-fconvert} compile options.
760 @emph{Note that the values specified via the GFORTRAN_CONVERT_UNIT
761 environment variable will override the CONVERT specifier in the
762 open statement}. This is to give control over data formats to
763 users who do not have the source code of their program available.
765 @node GFORTRAN_ERROR_BACKTRACE
766 @section @env{GFORTRAN_ERROR_BACKTRACE}---Show backtrace on run-time errors
768 If the @env{GFORTRAN_ERROR_BACKTRACE} variable is set to @samp{y},
769 @samp{Y} or @samp{1} (only the first letter is relevant) then a
770 backtrace is printed when a serious run-time error occurs. To disable
771 the backtracing, set the variable to @samp{n}, @samp{N}, @samp{0}.
772 Default is to print a backtrace unless the @option{-fno-backtrace}
773 compile option was used.
775 @c =====================================================================
776 @c PART II: LANGUAGE REFERENCE
777 @c =====================================================================
780 \part{II}{Language Reference}
783 @c ---------------------------------------------------------------------
784 @c Fortran 2003 and 2008 Status
785 @c ---------------------------------------------------------------------
787 @node Fortran 2003 and 2008 status
788 @chapter Fortran 2003 and 2008 Status
791 * Fortran 2003 status::
792 * Fortran 2008 status::
796 @node Fortran 2003 status
797 @section Fortran 2003 status
799 GNU Fortran supports several Fortran 2003 features; an incomplete
800 list can be found below. See also the
801 @uref{http://gcc.gnu.org/wiki/Fortran2003, wiki page} about Fortran 2003.
804 @item Procedure pointers including procedure-pointer components with
805 @code{PASS} attribute.
807 @item Procedures which are bound to a derived type (type-bound procedures)
808 including @code{PASS}, @code{PROCEDURE} and @code{GENERIC}, and
809 operators bound to a type.
811 @item Abstract interfaces and type extension with the possibility to
812 override type-bound procedures or to have deferred binding.
814 @item Polymorphic entities (``@code{CLASS}'') for derived types -- including
815 @code{SAME_TYPE_AS}, @code{EXTENDS_TYPE_OF} and @code{SELECT TYPE} for
816 scalars and arrays, including unlimited polymorphism.
818 @item Generic interface names, which have the same name as derived types,
819 are now supported. This allows one to write constructor functions. Note
820 that Fortran does not support static constructor functions. For static
821 variables, only default initialization or structure-constructor
822 initialization are available.
824 @item The @code{ASSOCIATE} construct.
826 @item Interoperability with C including enumerations,
828 @item In structure constructors the components with default values may be
831 @item Extensions to the @code{ALLOCATE} statement, allowing for a
832 type-specification with type parameter and for allocation and initialization
833 from a @code{SOURCE=} expression; @code{ALLOCATE} and @code{DEALLOCATE}
834 optionally return an error message string via @code{ERRMSG=}.
836 @item Reallocation on assignment: If an intrinsic assignment is
837 used, an allocatable variable on the left-hand side is automatically allocated
838 (if unallocated) or reallocated (if the shape is different). Currently, scalar
839 deferred character length left-hand sides are correctly handled but arrays
840 are not yet fully implemented.
842 @item Transferring of allocations via @code{MOVE_ALLOC}.
844 @item The @code{PRIVATE} and @code{PUBLIC} attributes may be given individually
845 to derived-type components.
847 @item In pointer assignments, the lower bound may be specified and
848 the remapping of elements is supported.
850 @item For pointers an @code{INTENT} may be specified which affect the
851 association status not the value of the pointer target.
853 @item Intrinsics @code{command_argument_count}, @code{get_command},
854 @code{get_command_argument}, and @code{get_environment_variable}.
856 @item Support for Unicode characters (ISO 10646) and UTF-8, including
857 the @code{SELECTED_CHAR_KIND} and @code{NEW_LINE} intrinsic functions.
859 @item Support for binary, octal and hexadecimal (BOZ) constants in the
860 intrinsic functions @code{INT}, @code{REAL}, @code{CMPLX} and @code{DBLE}.
862 @item Support for namelist variables with allocatable and pointer
863 attribute and nonconstant length type parameter.
866 @cindex array, constructors
868 Array constructors using square brackets. That is, @code{[...]} rather
869 than @code{(/.../)}. Type-specification for array constructors like
870 @code{(/ some-type :: ... /)}.
872 @item Extensions to the specification and initialization expressions,
873 including the support for intrinsics with real and complex arguments.
875 @item Support for the asynchronous input/output syntax; however, the
876 data transfer is currently always synchronously performed.
879 @cindex @code{FLUSH} statement
880 @cindex statement, @code{FLUSH}
881 @code{FLUSH} statement.
884 @cindex @code{IOMSG=} specifier
885 @code{IOMSG=} specifier for I/O statements.
888 @cindex @code{ENUM} statement
889 @cindex @code{ENUMERATOR} statement
890 @cindex statement, @code{ENUM}
891 @cindex statement, @code{ENUMERATOR}
892 @opindex @code{fshort-enums}
893 Support for the declaration of enumeration constants via the
894 @code{ENUM} and @code{ENUMERATOR} statements. Interoperability with
895 @command{gcc} is guaranteed also for the case where the
896 @command{-fshort-enums} command line option is given.
903 @cindex @code{ALLOCATABLE} dummy arguments
904 @code{ALLOCATABLE} dummy arguments.
906 @cindex @code{ALLOCATABLE} function results
907 @code{ALLOCATABLE} function results
909 @cindex @code{ALLOCATABLE} components of derived types
910 @code{ALLOCATABLE} components of derived types
914 @cindex @code{STREAM} I/O
915 @cindex @code{ACCESS='STREAM'} I/O
916 The @code{OPEN} statement supports the @code{ACCESS='STREAM'} specifier,
917 allowing I/O without any record structure.
920 Namelist input/output for internal files.
922 @item Further I/O extensions: Rounding during formatted output, using of
923 a decimal comma instead of a decimal point, setting whether a plus sign
924 should appear for positive numbers.
927 @cindex @code{PROTECTED} statement
928 @cindex statement, @code{PROTECTED}
929 The @code{PROTECTED} statement and attribute.
932 @cindex @code{VALUE} statement
933 @cindex statement, @code{VALUE}
934 The @code{VALUE} statement and attribute.
937 @cindex @code{VOLATILE} statement
938 @cindex statement, @code{VOLATILE}
939 The @code{VOLATILE} statement and attribute.
942 @cindex @code{IMPORT} statement
943 @cindex statement, @code{IMPORT}
944 The @code{IMPORT} statement, allowing to import
945 host-associated derived types.
947 @item The intrinsic modules @code{ISO_FORTRAN_ENVIRONMENT} is supported,
948 which contains parameters of the I/O units, storage sizes. Additionally,
949 procedures for C interoperability are available in the @code{ISO_C_BINDING}
953 @cindex @code{USE, INTRINSIC} statement
954 @cindex statement, @code{USE, INTRINSIC}
955 @cindex @code{ISO_FORTRAN_ENV} statement
956 @cindex statement, @code{ISO_FORTRAN_ENV}
957 @code{USE} statement with @code{INTRINSIC} and @code{NON_INTRINSIC}
958 attribute; supported intrinsic modules: @code{ISO_FORTRAN_ENV},
959 @code{ISO_C_BINDING}, @code{OMP_LIB} and @code{OMP_LIB_KINDS}.
962 Renaming of operators in the @code{USE} statement.
967 @node Fortran 2008 status
968 @section Fortran 2008 status
970 The latest version of the Fortran standard is ISO/IEC 1539-1:2010, informally
971 known as Fortran 2008. The official version is available from International
972 Organization for Standardization (ISO) or its national member organizations.
973 The the final draft (FDIS) can be downloaded free of charge from
974 @url{http://www.nag.co.uk/@/sc22wg5/@/links.html}. Fortran is developed by the
975 Working Group 5 of Sub-Committee 22 of the Joint Technical Committee 1 of the
976 International Organization for Standardization and the International
977 Electrotechnical Commission (IEC). This group is known as
978 @uref{http://www.nag.co.uk/sc22wg5/, WG5}.
980 The GNU Fortran compiler supports several of the new features of Fortran 2008;
981 the @uref{http://gcc.gnu.org/wiki/Fortran2008Status, wiki} has some information
982 about the current Fortran 2008 implementation status. In particular, the
983 following is implemented.
986 @item The @option{-std=f2008} option and support for the file extensions
987 @file{.f08} and @file{.F08}.
989 @item The @code{OPEN} statement now supports the @code{NEWUNIT=} option,
990 which returns a unique file unit, thus preventing inadvertent use of the
991 same unit in different parts of the program.
993 @item The @code{g0} format descriptor and unlimited format items.
995 @item The mathematical intrinsics @code{ASINH}, @code{ACOSH}, @code{ATANH},
996 @code{ERF}, @code{ERFC}, @code{GAMMA}, @code{LOG_GAMMA}, @code{BESSEL_J0},
997 @code{BESSEL_J1}, @code{BESSEL_JN}, @code{BESSEL_Y0}, @code{BESSEL_Y1},
998 @code{BESSEL_YN}, @code{HYPOT}, @code{NORM2}, and @code{ERFC_SCALED}.
1000 @item Using complex arguments with @code{TAN}, @code{SINH}, @code{COSH},
1001 @code{TANH}, @code{ASIN}, @code{ACOS}, and @code{ATAN} is now possible;
1002 @code{ATAN}(@var{Y},@var{X}) is now an alias for @code{ATAN2}(@var{Y},@var{X}).
1004 @item Support of the @code{PARITY} intrinsic functions.
1006 @item The following bit intrinsics: @code{LEADZ} and @code{TRAILZ} for
1007 counting the number of leading and trailing zero bits, @code{POPCNT} and
1008 @code{POPPAR} for counting the number of one bits and returning the parity;
1009 @code{BGE}, @code{BGT}, @code{BLE}, and @code{BLT} for bitwise comparisons;
1010 @code{DSHIFTL} and @code{DSHIFTR} for combined left and right shifts,
1011 @code{MASKL} and @code{MASKR} for simple left and right justified masks,
1012 @code{MERGE_BITS} for a bitwise merge using a mask, @code{SHIFTA},
1013 @code{SHIFTL} and @code{SHIFTR} for shift operations, and the
1014 transformational bit intrinsics @code{IALL}, @code{IANY} and @code{IPARITY}.
1016 @item Support of the @code{EXECUTE_COMMAND_LINE} intrinsic subroutine.
1018 @item Support for the @code{STORAGE_SIZE} intrinsic inquiry function.
1020 @item The @code{INT@{8,16,32@}} and @code{REAL@{32,64,128@}} kind type
1021 parameters and the array-valued named constants @code{INTEGER_KINDS},
1022 @code{LOGICAL_KINDS}, @code{REAL_KINDS} and @code{CHARACTER_KINDS} of
1023 the intrinsic module @code{ISO_FORTRAN_ENV}.
1025 @item The module procedures @code{C_SIZEOF} of the intrinsic module
1026 @code{ISO_C_BINDINGS} and @code{COMPILER_VERSION} and @code{COMPILER_OPTIONS}
1027 of @code{ISO_FORTRAN_ENV}.
1029 @item Coarray support for serial programs with @option{-fcoarray=single} flag
1030 and experimental support for multiple images with the @option{-fcoarray=lib}
1033 @item The @code{DO CONCURRENT} construct is supported.
1035 @item The @code{BLOCK} construct is supported.
1037 @item The @code{STOP} and the new @code{ERROR STOP} statements now
1038 support all constant expressions.
1040 @item Support for the @code{CONTIGUOUS} attribute.
1042 @item Support for @code{ALLOCATE} with @code{MOLD}.
1044 @item Support for the @code{IMPURE} attribute for procedures, which
1045 allows for @code{ELEMENTAL} procedures without the restrictions of
1048 @item Null pointers (including @code{NULL()}) and not-allocated variables
1049 can be used as actual argument to optional non-pointer, non-allocatable
1050 dummy arguments, denoting an absent argument.
1052 @item Non-pointer variables with @code{TARGET} attribute can be used as
1053 actual argument to @code{POINTER} dummies with @code{INTENT(IN)}.
1055 @item Pointers including procedure pointers and those in a derived
1056 type (pointer components) can now be initialized by a target instead
1057 of only by @code{NULL}.
1059 @item The @code{EXIT} statement (with construct-name) can be now be
1060 used to leave not only the @code{DO} but also the @code{ASSOCIATE},
1061 @code{BLOCK}, @code{IF}, @code{SELECT CASE} and @code{SELECT TYPE}
1064 @item Internal procedures can now be used as actual argument.
1066 @item Minor features: obsolesce diagnostics for @code{ENTRY} with
1067 @option{-std=f2008}; a line may start with a semicolon; for internal
1068 and module procedures @code{END} can be used instead of
1069 @code{END SUBROUTINE} and @code{END FUNCTION}; @code{SELECTED_REAL_KIND}
1070 now also takes a @code{RADIX} argument; intrinsic types are supported
1071 for @code{TYPE}(@var{intrinsic-type-spec}); multiple type-bound procedures
1072 can be declared in a single @code{PROCEDURE} statement; implied-shape
1073 arrays are supported for named constants (@code{PARAMETER}).
1078 @node TS 29113 status
1079 @section Technical Specification 29113 Status
1081 GNU Fortran supports some of the new features of the Technical
1082 Specification (TS) 29113 on Further Interoperability of Fortran with C.
1083 The @uref{http://gcc.gnu.org/wiki/TS29113Status, wiki} has some information
1084 about the current TS 29113 implementation status. In particular, the
1085 following is implemented.
1087 See also @ref{Further Interoperability of Fortran with C}.
1090 @item The @option{-std=f2008ts} option.
1092 @item The @code{OPTIONAL} attribute is allowed for dummy arguments
1093 of @code{BIND(C) procedures.}
1095 @item The @code{RANK} intrinsic is supported.
1097 @item GNU Fortran's implementation for variables with @code{ASYNCHRONOUS}
1098 attribute is compatible with TS 29113.
1100 @item Assumed types (@code{TYPE(*)}.
1102 @item Assumed-rank (@code{DIMENSION(..)}). However, the array descriptor
1103 of the TS is not yet supported.
1108 @c ---------------------------------------------------------------------
1109 @c Compiler Characteristics
1110 @c ---------------------------------------------------------------------
1112 @node Compiler Characteristics
1113 @chapter Compiler Characteristics
1115 This chapter describes certain characteristics of the GNU Fortran
1116 compiler, that are not specified by the Fortran standard, but which
1117 might in some way or another become visible to the programmer.
1120 * KIND Type Parameters::
1121 * Internal representation of LOGICAL variables::
1122 * Thread-safety of the runtime library::
1123 * Data consistency and durability::
1127 @node KIND Type Parameters
1128 @section KIND Type Parameters
1131 The @code{KIND} type parameters supported by GNU Fortran for the primitive
1137 1, 2, 4, 8*, 16*, default: 4**
1140 1, 2, 4, 8*, 16*, default: 4**
1143 4, 8, 10*, 16*, default: 4***
1146 4, 8, 10*, 16*, default: 4***
1148 @item DOUBLE PRECISION
1149 4, 8, 10*, 16*, default: 8***
1157 * not available on all systems @*
1158 ** unless @option{-fdefault-integer-8} is used @*
1159 *** unless @option{-fdefault-real-8} is used (see @ref{Fortran Dialect Options})
1162 The @code{KIND} value matches the storage size in bytes, except for
1163 @code{COMPLEX} where the storage size is twice as much (or both real and
1164 imaginary part are a real value of the given size). It is recommended to use
1165 the @ref{SELECTED_CHAR_KIND}, @ref{SELECTED_INT_KIND} and
1166 @ref{SELECTED_REAL_KIND} intrinsics or the @code{INT8}, @code{INT16},
1167 @code{INT32}, @code{INT64}, @code{REAL32}, @code{REAL64}, and @code{REAL128}
1168 parameters of the @code{ISO_FORTRAN_ENV} module instead of the concrete values.
1169 The available kind parameters can be found in the constant arrays
1170 @code{CHARACTER_KINDS}, @code{INTEGER_KINDS}, @code{LOGICAL_KINDS} and
1171 @code{REAL_KINDS} in the @ref{ISO_FORTRAN_ENV} module. For C interoperability,
1172 the kind parameters of the @ref{ISO_C_BINDING} module should be used.
1175 @node Internal representation of LOGICAL variables
1176 @section Internal representation of LOGICAL variables
1177 @cindex logical, variable representation
1179 The Fortran standard does not specify how variables of @code{LOGICAL}
1180 type are represented, beyond requiring that @code{LOGICAL} variables
1181 of default kind have the same storage size as default @code{INTEGER}
1182 and @code{REAL} variables. The GNU Fortran internal representation is
1185 A @code{LOGICAL(KIND=N)} variable is represented as an
1186 @code{INTEGER(KIND=N)} variable, however, with only two permissible
1187 values: @code{1} for @code{.TRUE.} and @code{0} for
1188 @code{.FALSE.}. Any other integer value results in undefined behavior.
1190 See also @ref{Argument passing conventions} and @ref{Interoperability with C}.
1193 @node Thread-safety of the runtime library
1194 @section Thread-safety of the runtime library
1195 @cindex thread-safety, threads
1197 GNU Fortran can be used in programs with multiple threads, e.g.@: by
1198 using OpenMP, by calling OS thread handling functions via the
1199 @code{ISO_C_BINDING} facility, or by GNU Fortran compiled library code
1200 being called from a multi-threaded program.
1202 The GNU Fortran runtime library, (@code{libgfortran}), supports being
1203 called concurrently from multiple threads with the following
1206 During library initialization, the C @code{getenv} function is used,
1207 which need not be thread-safe. Similarly, the @code{getenv}
1208 function is used to implement the @code{GET_ENVIRONMENT_VARIABLE} and
1209 @code{GETENV} intrinsics. It is the responsibility of the user to
1210 ensure that the environment is not being updated concurrently when any
1211 of these actions are taking place.
1213 The @code{EXECUTE_COMMAND_LINE} and @code{SYSTEM} intrinsics are
1214 implemented with the @code{system} function, which need not be
1215 thread-safe. It is the responsibility of the user to ensure that
1216 @code{system} is not called concurrently.
1218 Finally, for platforms not supporting thread-safe POSIX functions,
1219 further functionality might not be thread-safe. For details, please
1220 consult the documentation for your operating system.
1223 @node Data consistency and durability
1224 @section Data consistency and durability
1225 @cindex consistency, durability
1227 This section contains a brief overview of data and metadata
1228 consistency and durability issues when doing I/O.
1230 With respect to durability, GNU Fortran makes no effort to ensure that
1231 data is committed to stable storage. If this is required, the GNU
1232 Fortran programmer can use the intrinsic @code{FNUM} to retrieve the
1233 low level file descriptor corresponding to an open Fortran unit. Then,
1234 using e.g. the @code{ISO_C_BINDING} feature, one can call the
1235 underlying system call to flush dirty data to stable storage, such as
1236 @code{fsync} on POSIX, @code{_commit} on MingW, or @code{fcntl(fd,
1237 F_FULLSYNC, 0)} on Mac OS X. The following example shows how to call
1241 ! Declare the interface for POSIX fsync function
1243 function fsync (fd) bind(c,name="fsync")
1244 use iso_c_binding, only: c_int
1245 integer(c_int), value :: fd
1246 integer(c_int) :: fsync
1250 ! Variable declaration
1254 open (10,file="foo")
1257 ! Perform I/O on unit 10
1262 ret = fsync(fnum(10))
1264 ! Handle possible error
1265 if (ret /= 0) stop "Error calling FSYNC"
1268 With respect to consistency, for regular files GNU Fortran uses
1269 buffered I/O in order to improve performance. This buffer is flushed
1270 automatically when full and in some other situations, e.g. when
1271 closing a unit. It can also be explicitly flushed with the
1272 @code{FLUSH} statement. Also, the buffering can be turned off with the
1273 @code{GFORTRAN_UNBUFFERED_ALL} and
1274 @code{GFORTRAN_UNBUFFERED_PRECONNECTED} environment variables. Special
1275 files, such as terminals and pipes, are always unbuffered. Sometimes,
1276 however, further things may need to be done in order to allow other
1277 processes to see data that GNU Fortran has written, as follows.
1279 The Windows platform supports a relaxed metadata consistency model,
1280 where file metadata is written to the directory lazily. This means
1281 that, for instance, the @code{dir} command can show a stale size for a
1282 file. One can force a directory metadata update by closing the unit,
1283 or by calling @code{_commit} on the file descriptor. Note, though,
1284 that @code{_commit} will force all dirty data to stable storage, which
1285 is often a very slow operation.
1287 The Network File System (NFS) implements a relaxed consistency model
1288 called open-to-close consistency. Closing a file forces dirty data and
1289 metadata to be flushed to the server, and opening a file forces the
1290 client to contact the server in order to revalidate cached
1291 data. @code{fsync} will also force a flush of dirty data and metadata
1292 to the server. Similar to @code{open} and @code{close}, acquiring and
1293 releasing @code{fcntl} file locks, if the server supports them, will
1294 also force cache validation and flushing dirty data and metadata.
1297 @c ---------------------------------------------------------------------
1299 @c ---------------------------------------------------------------------
1301 @c Maybe this chapter should be merged with the 'Standards' section,
1302 @c whenever that is written :-)
1308 The two sections below detail the extensions to standard Fortran that are
1309 implemented in GNU Fortran, as well as some of the popular or
1310 historically important extensions that are not (or not yet) implemented.
1311 For the latter case, we explain the alternatives available to GNU Fortran
1312 users, including replacement by standard-conforming code or GNU
1316 * Extensions implemented in GNU Fortran::
1317 * Extensions not implemented in GNU Fortran::
1321 @node Extensions implemented in GNU Fortran
1322 @section Extensions implemented in GNU Fortran
1323 @cindex extensions, implemented
1325 GNU Fortran implements a number of extensions over standard
1326 Fortran. This chapter contains information on their syntax and
1327 meaning. There are currently two categories of GNU Fortran
1328 extensions, those that provide functionality beyond that provided
1329 by any standard, and those that are supported by GNU Fortran
1330 purely for backward compatibility with legacy compilers. By default,
1331 @option{-std=gnu} allows the compiler to accept both types of
1332 extensions, but to warn about the use of the latter. Specifying
1333 either @option{-std=f95}, @option{-std=f2003} or @option{-std=f2008}
1334 disables both types of extensions, and @option{-std=legacy} allows both
1338 * Old-style kind specifications::
1339 * Old-style variable initialization::
1340 * Extensions to namelist::
1341 * X format descriptor without count field::
1342 * Commas in FORMAT specifications::
1343 * Missing period in FORMAT specifications::
1345 * @code{Q} exponent-letter::
1346 * BOZ literal constants::
1347 * Real array indices::
1349 * Implicitly convert LOGICAL and INTEGER values::
1350 * Hollerith constants support::
1352 * CONVERT specifier::
1354 * Argument list functions::
1357 @node Old-style kind specifications
1358 @subsection Old-style kind specifications
1359 @cindex kind, old-style
1361 GNU Fortran allows old-style kind specifications in declarations. These
1367 where @code{TYPESPEC} is a basic type (@code{INTEGER}, @code{REAL},
1368 etc.), and where @code{size} is a byte count corresponding to the
1369 storage size of a valid kind for that type. (For @code{COMPLEX}
1370 variables, @code{size} is the total size of the real and imaginary
1371 parts.) The statement then declares @code{x}, @code{y} and @code{z} to
1372 be of type @code{TYPESPEC} with the appropriate kind. This is
1373 equivalent to the standard-conforming declaration
1378 where @code{k} is the kind parameter suitable for the intended precision. As
1379 kind parameters are implementation-dependent, use the @code{KIND},
1380 @code{SELECTED_INT_KIND} and @code{SELECTED_REAL_KIND} intrinsics to retrieve
1381 the correct value, for instance @code{REAL*8 x} can be replaced by:
1383 INTEGER, PARAMETER :: dbl = KIND(1.0d0)
1387 @node Old-style variable initialization
1388 @subsection Old-style variable initialization
1390 GNU Fortran allows old-style initialization of variables of the
1394 REAL x(2,2) /3*0.,1./
1396 The syntax for the initializers is as for the @code{DATA} statement, but
1397 unlike in a @code{DATA} statement, an initializer only applies to the
1398 variable immediately preceding the initialization. In other words,
1399 something like @code{INTEGER I,J/2,3/} is not valid. This style of
1400 initialization is only allowed in declarations without double colons
1401 (@code{::}); the double colons were introduced in Fortran 90, which also
1402 introduced a standard syntax for initializing variables in type
1405 Examples of standard-conforming code equivalent to the above example
1409 INTEGER :: i = 1, j = 2
1410 REAL :: x(2,2) = RESHAPE((/0.,0.,0.,1./),SHAPE(x))
1414 DATA i/1/, j/2/, x/3*0.,1./
1417 Note that variables which are explicitly initialized in declarations
1418 or in @code{DATA} statements automatically acquire the @code{SAVE}
1421 @node Extensions to namelist
1422 @subsection Extensions to namelist
1425 GNU Fortran fully supports the Fortran 95 standard for namelist I/O
1426 including array qualifiers, substrings and fully qualified derived types.
1427 The output from a namelist write is compatible with namelist read. The
1428 output has all names in upper case and indentation to column 1 after the
1429 namelist name. Two extensions are permitted:
1431 Old-style use of @samp{$} instead of @samp{&}
1434 X(:)%Y(2) = 1.0 2.0 3.0
1439 It should be noted that the default terminator is @samp{/} rather than
1442 Querying of the namelist when inputting from stdin. After at least
1443 one space, entering @samp{?} sends to stdout the namelist name and the names of
1444 the variables in the namelist:
1455 Entering @samp{=?} outputs the namelist to stdout, as if
1456 @code{WRITE(*,NML = mynml)} had been called:
1461 X(1)%Y= 0.000000 , 1.000000 , 0.000000 ,
1462 X(2)%Y= 0.000000 , 2.000000 , 0.000000 ,
1463 X(3)%Y= 0.000000 , 3.000000 , 0.000000 ,
1467 To aid this dialog, when input is from stdin, errors send their
1468 messages to stderr and execution continues, even if @code{IOSTAT} is set.
1470 @code{PRINT} namelist is permitted. This causes an error if
1471 @option{-std=f95} is used.
1474 REAL, dimension (4) :: x = (/1.0, 2.0, 3.0, 4.0/)
1477 END PROGRAM test_print
1480 Expanded namelist reads are permitted. This causes an error if
1481 @option{-std=f95} is used. In the following example, the first element
1482 of the array will be given the value 0.00 and the two succeeding
1483 elements will be given the values 1.00 and 2.00.
1486 X(1,1) = 0.00 , 1.00 , 2.00
1490 @node X format descriptor without count field
1491 @subsection @code{X} format descriptor without count field
1493 To support legacy codes, GNU Fortran permits the count field of the
1494 @code{X} edit descriptor in @code{FORMAT} statements to be omitted.
1495 When omitted, the count is implicitly assumed to be one.
1499 10 FORMAT (I1, X, I1)
1502 @node Commas in FORMAT specifications
1503 @subsection Commas in @code{FORMAT} specifications
1505 To support legacy codes, GNU Fortran allows the comma separator
1506 to be omitted immediately before and after character string edit
1507 descriptors in @code{FORMAT} statements.
1511 10 FORMAT ('FOO='I1' BAR='I2)
1515 @node Missing period in FORMAT specifications
1516 @subsection Missing period in @code{FORMAT} specifications
1518 To support legacy codes, GNU Fortran allows missing periods in format
1519 specifications if and only if @option{-std=legacy} is given on the
1520 command line. This is considered non-conforming code and is
1529 @node I/O item lists
1530 @subsection I/O item lists
1531 @cindex I/O item lists
1533 To support legacy codes, GNU Fortran allows the input item list
1534 of the @code{READ} statement, and the output item lists of the
1535 @code{WRITE} and @code{PRINT} statements, to start with a comma.
1537 @node @code{Q} exponent-letter
1538 @subsection @code{Q} exponent-letter
1539 @cindex @code{Q} exponent-letter
1541 GNU Fortran accepts real literal constants with an exponent-letter
1542 of @code{Q}, for example, @code{1.23Q45}. The constant is interpreted
1543 as a @code{REAL(16)} entity on targets that support this type. If
1544 the target does not support @code{REAL(16)} but has a @code{REAL(10)}
1545 type, then the real-literal-constant will be interpreted as a
1546 @code{REAL(10)} entity. In the absence of @code{REAL(16)} and
1547 @code{REAL(10)}, an error will occur.
1549 @node BOZ literal constants
1550 @subsection BOZ literal constants
1551 @cindex BOZ literal constants
1553 Besides decimal constants, Fortran also supports binary (@code{b}),
1554 octal (@code{o}) and hexadecimal (@code{z}) integer constants. The
1555 syntax is: @samp{prefix quote digits quote}, were the prefix is
1556 either @code{b}, @code{o} or @code{z}, quote is either @code{'} or
1557 @code{"} and the digits are for binary @code{0} or @code{1}, for
1558 octal between @code{0} and @code{7}, and for hexadecimal between
1559 @code{0} and @code{F}. (Example: @code{b'01011101'}.)
1561 Up to Fortran 95, BOZ literals were only allowed to initialize
1562 integer variables in DATA statements. Since Fortran 2003 BOZ literals
1563 are also allowed as argument of @code{REAL}, @code{DBLE}, @code{INT}
1564 and @code{CMPLX}; the result is the same as if the integer BOZ
1565 literal had been converted by @code{TRANSFER} to, respectively,
1566 @code{real}, @code{double precision}, @code{integer} or @code{complex}.
1567 As GNU Fortran extension the intrinsic procedures @code{FLOAT},
1568 @code{DFLOAT}, @code{COMPLEX} and @code{DCMPLX} are treated alike.
1570 As an extension, GNU Fortran allows hexadecimal BOZ literal constants to
1571 be specified using the @code{X} prefix, in addition to the standard
1572 @code{Z} prefix. The BOZ literal can also be specified by adding a
1573 suffix to the string, for example, @code{Z'ABC'} and @code{'ABC'Z} are
1576 Furthermore, GNU Fortran allows using BOZ literal constants outside
1577 DATA statements and the four intrinsic functions allowed by Fortran 2003.
1578 In DATA statements, in direct assignments, where the right-hand side
1579 only contains a BOZ literal constant, and for old-style initializers of
1580 the form @code{integer i /o'0173'/}, the constant is transferred
1581 as if @code{TRANSFER} had been used; for @code{COMPLEX} numbers, only
1582 the real part is initialized unless @code{CMPLX} is used. In all other
1583 cases, the BOZ literal constant is converted to an @code{INTEGER} value with
1584 the largest decimal representation. This value is then converted
1585 numerically to the type and kind of the variable in question.
1586 (For instance, @code{real :: r = b'0000001' + 1} initializes @code{r}
1587 with @code{2.0}.) As different compilers implement the extension
1588 differently, one should be careful when doing bitwise initialization
1589 of non-integer variables.
1591 Note that initializing an @code{INTEGER} variable with a statement such
1592 as @code{DATA i/Z'FFFFFFFF'/} will give an integer overflow error rather
1593 than the desired result of @math{-1} when @code{i} is a 32-bit integer
1594 on a system that supports 64-bit integers. The @samp{-fno-range-check}
1595 option can be used as a workaround for legacy code that initializes
1596 integers in this manner.
1598 @node Real array indices
1599 @subsection Real array indices
1600 @cindex array, indices of type real
1602 As an extension, GNU Fortran allows the use of @code{REAL} expressions
1603 or variables as array indices.
1605 @node Unary operators
1606 @subsection Unary operators
1607 @cindex operators, unary
1609 As an extension, GNU Fortran allows unary plus and unary minus operators
1610 to appear as the second operand of binary arithmetic operators without
1611 the need for parenthesis.
1617 @node Implicitly convert LOGICAL and INTEGER values
1618 @subsection Implicitly convert @code{LOGICAL} and @code{INTEGER} values
1619 @cindex conversion, to integer
1620 @cindex conversion, to logical
1622 As an extension for backwards compatibility with other compilers, GNU
1623 Fortran allows the implicit conversion of @code{LOGICAL} values to
1624 @code{INTEGER} values and vice versa. When converting from a
1625 @code{LOGICAL} to an @code{INTEGER}, @code{.FALSE.} is interpreted as
1626 zero, and @code{.TRUE.} is interpreted as one. When converting from
1627 @code{INTEGER} to @code{LOGICAL}, the value zero is interpreted as
1628 @code{.FALSE.} and any nonzero value is interpreted as @code{.TRUE.}.
1639 However, there is no implicit conversion of @code{INTEGER} values in
1640 @code{if}-statements, nor of @code{LOGICAL} or @code{INTEGER} values
1643 @node Hollerith constants support
1644 @subsection Hollerith constants support
1645 @cindex Hollerith constants
1647 GNU Fortran supports Hollerith constants in assignments, function
1648 arguments, and @code{DATA} and @code{ASSIGN} statements. A Hollerith
1649 constant is written as a string of characters preceded by an integer
1650 constant indicating the character count, and the letter @code{H} or
1651 @code{h}, and stored in bytewise fashion in a numeric (@code{INTEGER},
1652 @code{REAL}, or @code{complex}) or @code{LOGICAL} variable. The
1653 constant will be padded or truncated to fit the size of the variable in
1656 Examples of valid uses of Hollerith constants:
1659 data x /16Habcdefghijklmnop, 16Hqrstuvwxyz012345/
1660 x(1) = 16HABCDEFGHIJKLMNOP
1664 Invalid Hollerith constants examples:
1667 a = 8H12345678 ! Valid, but the Hollerith constant will be truncated.
1668 a = 0H ! At least one character is needed.
1671 In general, Hollerith constants were used to provide a rudimentary
1672 facility for handling character strings in early Fortran compilers,
1673 prior to the introduction of @code{CHARACTER} variables in Fortran 77;
1674 in those cases, the standard-compliant equivalent is to convert the
1675 program to use proper character strings. On occasion, there may be a
1676 case where the intent is specifically to initialize a numeric variable
1677 with a given byte sequence. In these cases, the same result can be
1678 obtained by using the @code{TRANSFER} statement, as in this example.
1680 INTEGER(KIND=4) :: a
1681 a = TRANSFER ("abcd", a) ! equivalent to: a = 4Habcd
1686 @subsection Cray pointers
1687 @cindex pointer, Cray
1689 Cray pointers are part of a non-standard extension that provides a
1690 C-like pointer in Fortran. This is accomplished through a pair of
1691 variables: an integer "pointer" that holds a memory address, and a
1692 "pointee" that is used to dereference the pointer.
1694 Pointer/pointee pairs are declared in statements of the form:
1696 pointer ( <pointer> , <pointee> )
1700 pointer ( <pointer1> , <pointee1> ), ( <pointer2> , <pointee2> ), ...
1702 The pointer is an integer that is intended to hold a memory address.
1703 The pointee may be an array or scalar. A pointee can be an assumed
1704 size array---that is, the last dimension may be left unspecified by
1705 using a @code{*} in place of a value---but a pointee cannot be an
1706 assumed shape array. No space is allocated for the pointee.
1708 The pointee may have its type declared before or after the pointer
1709 statement, and its array specification (if any) may be declared
1710 before, during, or after the pointer statement. The pointer may be
1711 declared as an integer prior to the pointer statement. However, some
1712 machines have default integer sizes that are different than the size
1713 of a pointer, and so the following code is not portable:
1718 If a pointer is declared with a kind that is too small, the compiler
1719 will issue a warning; the resulting binary will probably not work
1720 correctly, because the memory addresses stored in the pointers may be
1721 truncated. It is safer to omit the first line of the above example;
1722 if explicit declaration of ipt's type is omitted, then the compiler
1723 will ensure that ipt is an integer variable large enough to hold a
1726 Pointer arithmetic is valid with Cray pointers, but it is not the same
1727 as C pointer arithmetic. Cray pointers are just ordinary integers, so
1728 the user is responsible for determining how many bytes to add to a
1729 pointer in order to increment it. Consider the following example:
1733 pointer (ipt, pointee)
1737 The last statement does not set @code{ipt} to the address of
1738 @code{target(1)}, as it would in C pointer arithmetic. Adding @code{1}
1739 to @code{ipt} just adds one byte to the address stored in @code{ipt}.
1741 Any expression involving the pointee will be translated to use the
1742 value stored in the pointer as the base address.
1744 To get the address of elements, this extension provides an intrinsic
1745 function @code{LOC()}. The @code{LOC()} function is equivalent to the
1746 @code{&} operator in C, except the address is cast to an integer type:
1749 pointer(ipt, arpte(10))
1751 ipt = loc(ar) ! Makes arpte is an alias for ar
1752 arpte(1) = 1.0 ! Sets ar(1) to 1.0
1754 The pointer can also be set by a call to the @code{MALLOC} intrinsic
1757 Cray pointees often are used to alias an existing variable. For
1765 As long as @code{ipt} remains unchanged, @code{iarr} is now an alias for
1766 @code{target}. The optimizer, however, will not detect this aliasing, so
1767 it is unsafe to use @code{iarr} and @code{target} simultaneously. Using
1768 a pointee in any way that violates the Fortran aliasing rules or
1769 assumptions is illegal. It is the user's responsibility to avoid doing
1770 this; the compiler works under the assumption that no such aliasing
1773 Cray pointers will work correctly when there is no aliasing (i.e., when
1774 they are used to access a dynamically allocated block of memory), and
1775 also in any routine where a pointee is used, but any variable with which
1776 it shares storage is not used. Code that violates these rules may not
1777 run as the user intends. This is not a bug in the optimizer; any code
1778 that violates the aliasing rules is illegal. (Note that this is not
1779 unique to GNU Fortran; any Fortran compiler that supports Cray pointers
1780 will ``incorrectly'' optimize code with illegal aliasing.)
1782 There are a number of restrictions on the attributes that can be applied
1783 to Cray pointers and pointees. Pointees may not have the
1784 @code{ALLOCATABLE}, @code{INTENT}, @code{OPTIONAL}, @code{DUMMY},
1785 @code{TARGET}, @code{INTRINSIC}, or @code{POINTER} attributes. Pointers
1786 may not have the @code{DIMENSION}, @code{POINTER}, @code{TARGET},
1787 @code{ALLOCATABLE}, @code{EXTERNAL}, or @code{INTRINSIC} attributes, nor
1788 may they be function results. Pointees may not occur in more than one
1789 pointer statement. A pointee cannot be a pointer. Pointees cannot occur
1790 in equivalence, common, or data statements.
1792 A Cray pointer may also point to a function or a subroutine. For
1793 example, the following excerpt is valid:
1797 pointer (subptr,subpte)
1807 A pointer may be modified during the course of a program, and this
1808 will change the location to which the pointee refers. However, when
1809 pointees are passed as arguments, they are treated as ordinary
1810 variables in the invoked function. Subsequent changes to the pointer
1811 will not change the base address of the array that was passed.
1813 @node CONVERT specifier
1814 @subsection @code{CONVERT} specifier
1815 @cindex @code{CONVERT} specifier
1817 GNU Fortran allows the conversion of unformatted data between little-
1818 and big-endian representation to facilitate moving of data
1819 between different systems. The conversion can be indicated with
1820 the @code{CONVERT} specifier on the @code{OPEN} statement.
1821 @xref{GFORTRAN_CONVERT_UNIT}, for an alternative way of specifying
1822 the data format via an environment variable.
1824 Valid values for @code{CONVERT} are:
1826 @item @code{CONVERT='NATIVE'} Use the native format. This is the default.
1827 @item @code{CONVERT='SWAP'} Swap between little- and big-endian.
1828 @item @code{CONVERT='LITTLE_ENDIAN'} Use the little-endian representation
1829 for unformatted files.
1830 @item @code{CONVERT='BIG_ENDIAN'} Use the big-endian representation for
1834 Using the option could look like this:
1836 open(file='big.dat',form='unformatted',access='sequential', &
1837 convert='big_endian')
1840 The value of the conversion can be queried by using
1841 @code{INQUIRE(CONVERT=ch)}. The values returned are
1842 @code{'BIG_ENDIAN'} and @code{'LITTLE_ENDIAN'}.
1844 @code{CONVERT} works between big- and little-endian for
1845 @code{INTEGER} values of all supported kinds and for @code{REAL}
1846 on IEEE systems of kinds 4 and 8. Conversion between different
1847 ``extended double'' types on different architectures such as
1848 m68k and x86_64, which GNU Fortran
1849 supports as @code{REAL(KIND=10)} and @code{REAL(KIND=16)}, will
1852 @emph{Note that the values specified via the GFORTRAN_CONVERT_UNIT
1853 environment variable will override the CONVERT specifier in the
1854 open statement}. This is to give control over data formats to
1855 users who do not have the source code of their program available.
1857 Using anything but the native representation for unformatted data
1858 carries a significant speed overhead. If speed in this area matters
1859 to you, it is best if you use this only for data that needs to be
1866 OpenMP (Open Multi-Processing) is an application programming
1867 interface (API) that supports multi-platform shared memory
1868 multiprocessing programming in C/C++ and Fortran on many
1869 architectures, including Unix and Microsoft Windows platforms.
1870 It consists of a set of compiler directives, library routines,
1871 and environment variables that influence run-time behavior.
1873 GNU Fortran strives to be compatible to the
1874 @uref{http://www.openmp.org/mp-documents/spec31.pdf,
1875 OpenMP Application Program Interface v3.1}.
1877 To enable the processing of the OpenMP directive @code{!$omp} in
1878 free-form source code; the @code{c$omp}, @code{*$omp} and @code{!$omp}
1879 directives in fixed form; the @code{!$} conditional compilation sentinels
1880 in free form; and the @code{c$}, @code{*$} and @code{!$} sentinels
1881 in fixed form, @command{gfortran} needs to be invoked with the
1882 @option{-fopenmp}. This also arranges for automatic linking of the
1883 GNU OpenMP runtime library @ref{Top,,libgomp,libgomp,GNU OpenMP
1886 The OpenMP Fortran runtime library routines are provided both in a
1887 form of a Fortran 90 module named @code{omp_lib} and in a form of
1888 a Fortran @code{include} file named @file{omp_lib.h}.
1890 An example of a parallelized loop taken from Appendix A.1 of
1891 the OpenMP Application Program Interface v2.5:
1893 SUBROUTINE A1(N, A, B)
1896 !$OMP PARALLEL DO !I is private by default
1898 B(I) = (A(I) + A(I-1)) / 2.0
1900 !$OMP END PARALLEL DO
1907 @option{-fopenmp} implies @option{-frecursive}, i.e., all local arrays
1908 will be allocated on the stack. When porting existing code to OpenMP,
1909 this may lead to surprising results, especially to segmentation faults
1910 if the stacksize is limited.
1913 On glibc-based systems, OpenMP enabled applications cannot be statically
1914 linked due to limitations of the underlying pthreads-implementation. It
1915 might be possible to get a working solution if
1916 @command{-Wl,--whole-archive -lpthread -Wl,--no-whole-archive} is added
1917 to the command line. However, this is not supported by @command{gcc} and
1918 thus not recommended.
1921 @node Argument list functions
1922 @subsection Argument list functions @code{%VAL}, @code{%REF} and @code{%LOC}
1923 @cindex argument list functions
1928 GNU Fortran supports argument list functions @code{%VAL}, @code{%REF}
1929 and @code{%LOC} statements, for backward compatibility with g77.
1930 It is recommended that these should be used only for code that is
1931 accessing facilities outside of GNU Fortran, such as operating system
1932 or windowing facilities. It is best to constrain such uses to isolated
1933 portions of a program--portions that deal specifically and exclusively
1934 with low-level, system-dependent facilities. Such portions might well
1935 provide a portable interface for use by the program as a whole, but are
1936 themselves not portable, and should be thoroughly tested each time they
1937 are rebuilt using a new compiler or version of a compiler.
1939 @code{%VAL} passes a scalar argument by value, @code{%REF} passes it by
1940 reference and @code{%LOC} passes its memory location. Since gfortran
1941 already passes scalar arguments by reference, @code{%REF} is in effect
1942 a do-nothing. @code{%LOC} has the same effect as a Fortran pointer.
1944 An example of passing an argument by value to a C subroutine foo.:
1947 C prototype void foo_ (float x);
1956 For details refer to the g77 manual
1957 @uref{http://gcc.gnu.org/@/onlinedocs/@/gcc-3.4.6/@/g77/@/index.html#Top}.
1959 Also, @code{c_by_val.f} and its partner @code{c_by_val.c} of the
1960 GNU Fortran testsuite are worth a look.
1963 @node Extensions not implemented in GNU Fortran
1964 @section Extensions not implemented in GNU Fortran
1965 @cindex extensions, not implemented
1967 The long history of the Fortran language, its wide use and broad
1968 userbase, the large number of different compiler vendors and the lack of
1969 some features crucial to users in the first standards have lead to the
1970 existence of a number of important extensions to the language. While
1971 some of the most useful or popular extensions are supported by the GNU
1972 Fortran compiler, not all existing extensions are supported. This section
1973 aims at listing these extensions and offering advice on how best make
1974 code that uses them running with the GNU Fortran compiler.
1976 @c More can be found here:
1977 @c -- http://gcc.gnu.org/onlinedocs/gcc-3.4.6/g77/Missing-Features.html
1978 @c -- the list of Fortran and libgfortran bugs closed as WONTFIX:
1979 @c http://tinyurl.com/2u4h5y
1982 * STRUCTURE and RECORD::
1983 @c * UNION and MAP::
1984 * ENCODE and DECODE statements::
1985 * Variable FORMAT expressions::
1986 @c * Q edit descriptor::
1987 @c * AUTOMATIC statement::
1988 @c * TYPE and ACCEPT I/O Statements::
1989 @c * .XOR. operator::
1990 @c * CARRIAGECONTROL, DEFAULTFILE, DISPOSE and RECORDTYPE I/O specifiers::
1991 @c * Omitted arguments in procedure call::
1992 * Alternate complex function syntax::
1996 @node STRUCTURE and RECORD
1997 @subsection @code{STRUCTURE} and @code{RECORD}
1998 @cindex @code{STRUCTURE}
1999 @cindex @code{RECORD}
2001 Record structures are a pre-Fortran-90 vendor extension to create
2002 user-defined aggregate data types. GNU Fortran does not support
2003 record structures, only Fortran 90's ``derived types'', which have
2006 In many cases, record structures can easily be converted to derived types.
2007 To convert, replace @code{STRUCTURE /}@var{structure-name}@code{/}
2008 by @code{TYPE} @var{type-name}. Additionally, replace
2009 @code{RECORD /}@var{structure-name}@code{/} by
2010 @code{TYPE(}@var{type-name}@code{)}. Finally, in the component access,
2011 replace the period (@code{.}) by the percent sign (@code{%}).
2013 Here is an example of code using the non portable record structure syntax:
2016 ! Declaring a structure named ``item'' and containing three fields:
2017 ! an integer ID, an description string and a floating-point price.
2020 CHARACTER(LEN=200) description
2024 ! Define two variables, an single record of type ``item''
2025 ! named ``pear'', and an array of items named ``store_catalog''
2026 RECORD /item/ pear, store_catalog(100)
2028 ! We can directly access the fields of both variables
2030 pear.description = "juicy D'Anjou pear"
2032 store_catalog(7).id = 7831
2033 store_catalog(7).description = "milk bottle"
2034 store_catalog(7).price = 1.2
2036 ! We can also manipulate the whole structure
2037 store_catalog(12) = pear
2038 print *, store_catalog(12)
2042 This code can easily be rewritten in the Fortran 90 syntax as following:
2045 ! ``STRUCTURE /name/ ... END STRUCTURE'' becomes
2046 ! ``TYPE name ... END TYPE''
2049 CHARACTER(LEN=200) description
2053 ! ``RECORD /name/ variable'' becomes ``TYPE(name) variable''
2054 TYPE(item) pear, store_catalog(100)
2056 ! Instead of using a dot (.) to access fields of a record, the
2057 ! standard syntax uses a percent sign (%)
2059 pear%description = "juicy D'Anjou pear"
2061 store_catalog(7)%id = 7831
2062 store_catalog(7)%description = "milk bottle"
2063 store_catalog(7)%price = 1.2
2065 ! Assignments of a whole variable do not change
2066 store_catalog(12) = pear
2067 print *, store_catalog(12)
2071 @c @node UNION and MAP
2072 @c @subsection @code{UNION} and @code{MAP}
2073 @c @cindex @code{UNION}
2074 @c @cindex @code{MAP}
2076 @c For help writing this one, see
2077 @c http://www.eng.umd.edu/~nsw/ench250/fortran1.htm#UNION and
2078 @c http://www.tacc.utexas.edu/services/userguides/pgi/pgiws_ug/pgi32u06.htm
2081 @node ENCODE and DECODE statements
2082 @subsection @code{ENCODE} and @code{DECODE} statements
2083 @cindex @code{ENCODE}
2084 @cindex @code{DECODE}
2086 GNU Fortran does not support the @code{ENCODE} and @code{DECODE}
2087 statements. These statements are best replaced by @code{READ} and
2088 @code{WRITE} statements involving internal files (@code{CHARACTER}
2089 variables and arrays), which have been part of the Fortran standard since
2090 Fortran 77. For example, replace a code fragment like
2095 c ... Code that sets LINE
2096 DECODE (80, 9000, LINE) A, B, C
2097 9000 FORMAT (1X, 3(F10.5))
2104 CHARACTER(LEN=80) LINE
2106 c ... Code that sets LINE
2107 READ (UNIT=LINE, FMT=9000) A, B, C
2108 9000 FORMAT (1X, 3(F10.5))
2111 Similarly, replace a code fragment like
2116 c ... Code that sets A, B and C
2117 ENCODE (80, 9000, LINE) A, B, C
2118 9000 FORMAT (1X, 'OUTPUT IS ', 3(F10.5))
2125 CHARACTER(LEN=80) LINE
2127 c ... Code that sets A, B and C
2128 WRITE (UNIT=LINE, FMT=9000) A, B, C
2129 9000 FORMAT (1X, 'OUTPUT IS ', 3(F10.5))
2133 @node Variable FORMAT expressions
2134 @subsection Variable @code{FORMAT} expressions
2135 @cindex @code{FORMAT}
2137 A variable @code{FORMAT} expression is format statement which includes
2138 angle brackets enclosing a Fortran expression: @code{FORMAT(I<N>)}. GNU
2139 Fortran does not support this legacy extension. The effect of variable
2140 format expressions can be reproduced by using the more powerful (and
2141 standard) combination of internal output and string formats. For example,
2142 replace a code fragment like this:
2153 c Variable declaration
2154 CHARACTER(LEN=20) FMT
2156 c Other code here...
2158 WRITE(FMT,'("(I", I0, ")")') N+1
2166 c Variable declaration
2167 CHARACTER(LEN=20) FMT
2169 c Other code here...
2172 WRITE(6,"(I" // ADJUSTL(FMT) // ")") INT1
2176 @node Alternate complex function syntax
2177 @subsection Alternate complex function syntax
2178 @cindex Complex function
2180 Some Fortran compilers, including @command{g77}, let the user declare
2181 complex functions with the syntax @code{COMPLEX FUNCTION name*16()}, as
2182 well as @code{COMPLEX*16 FUNCTION name()}. Both are non-standard, legacy
2183 extensions. @command{gfortran} accepts the latter form, which is more
2184 common, but not the former.
2188 @c ---------------------------------------------------------------------
2189 @c Mixed-Language Programming
2190 @c ---------------------------------------------------------------------
2192 @node Mixed-Language Programming
2193 @chapter Mixed-Language Programming
2194 @cindex Interoperability
2195 @cindex Mixed-language programming
2198 * Interoperability with C::
2199 * GNU Fortran Compiler Directives::
2200 * Non-Fortran Main Program::
2201 * Naming and argument-passing conventions::
2204 This chapter is about mixed-language interoperability, but also applies
2205 if one links Fortran code compiled by different compilers. In most cases,
2206 use of the C Binding features of the Fortran 2003 standard is sufficient,
2207 and their use is highly recommended.
2210 @node Interoperability with C
2211 @section Interoperability with C
2215 * Derived Types and struct::
2216 * Interoperable Global Variables::
2217 * Interoperable Subroutines and Functions::
2218 * Working with Pointers::
2219 * Further Interoperability of Fortran with C::
2222 Since Fortran 2003 (ISO/IEC 1539-1:2004(E)) there is a
2223 standardized way to generate procedure and derived-type
2224 declarations and global variables which are interoperable with C
2225 (ISO/IEC 9899:1999). The @code{bind(C)} attribute has been added
2226 to inform the compiler that a symbol shall be interoperable with C;
2227 also, some constraints are added. Note, however, that not
2228 all C features have a Fortran equivalent or vice versa. For instance,
2229 neither C's unsigned integers nor C's functions with variable number
2230 of arguments have an equivalent in Fortran.
2232 Note that array dimensions are reversely ordered in C and that arrays in
2233 C always start with index 0 while in Fortran they start by default with
2234 1. Thus, an array declaration @code{A(n,m)} in Fortran matches
2235 @code{A[m][n]} in C and accessing the element @code{A(i,j)} matches
2236 @code{A[j-1][i-1]}. The element following @code{A(i,j)} (C: @code{A[j-1][i-1]};
2237 assuming @math{i < n}) in memory is @code{A(i+1,j)} (C: @code{A[j-1][i]}).
2239 @node Intrinsic Types
2240 @subsection Intrinsic Types
2242 In order to ensure that exactly the same variable type and kind is used
2243 in C and Fortran, the named constants shall be used which are defined in the
2244 @code{ISO_C_BINDING} intrinsic module. That module contains named constants
2245 for kind parameters and character named constants for the escape sequences
2246 in C. For a list of the constants, see @ref{ISO_C_BINDING}.
2248 For logical types, please note that the Fortran standard only guarantees
2249 interoperability between C99's @code{_Bool} and Fortran's @code{C_Bool}-kind
2250 logicals and C99 defines that @code{true} has the value 1 and @code{false}
2251 the value 0. Using any other integer value with GNU Fortran's @code{LOGICAL}
2252 (with any kind parameter) gives an undefined result. (Passing other integer
2253 values than 0 and 1 to GCC's @code{_Bool} is also undefined, unless the
2254 integer is explicitly or implicitly casted to @code{_Bool}.)
2258 @node Derived Types and struct
2259 @subsection Derived Types and struct
2261 For compatibility of derived types with @code{struct}, one needs to use
2262 the @code{BIND(C)} attribute in the type declaration. For instance, the
2263 following type declaration
2267 TYPE, BIND(C) :: myType
2268 INTEGER(C_INT) :: i1, i2
2269 INTEGER(C_SIGNED_CHAR) :: i3
2270 REAL(C_DOUBLE) :: d1
2271 COMPLEX(C_FLOAT_COMPLEX) :: c1
2272 CHARACTER(KIND=C_CHAR) :: str(5)
2276 matches the following @code{struct} declaration in C
2281 /* Note: "char" might be signed or unsigned. */
2289 Derived types with the C binding attribute shall not have the @code{sequence}
2290 attribute, type parameters, the @code{extends} attribute, nor type-bound
2291 procedures. Every component must be of interoperable type and kind and may not
2292 have the @code{pointer} or @code{allocatable} attribute. The names of the
2293 components are irrelevant for interoperability.
2295 As there exist no direct Fortran equivalents, neither unions nor structs
2296 with bit field or variable-length array members are interoperable.
2298 @node Interoperable Global Variables
2299 @subsection Interoperable Global Variables
2301 Variables can be made accessible from C using the C binding attribute,
2302 optionally together with specifying a binding name. Those variables
2303 have to be declared in the declaration part of a @code{MODULE},
2304 be of interoperable type, and have neither the @code{pointer} nor
2305 the @code{allocatable} attribute.
2311 integer(C_INT), bind(C, name="_MyProject_flags") :: global_flag
2312 type(myType), bind(C) :: tp
2316 Here, @code{_MyProject_flags} is the case-sensitive name of the variable
2317 as seen from C programs while @code{global_flag} is the case-insensitive
2318 name as seen from Fortran. If no binding name is specified, as for
2319 @var{tp}, the C binding name is the (lowercase) Fortran binding name.
2320 If a binding name is specified, only a single variable may be after the
2321 double colon. Note of warning: You cannot use a global variable to
2322 access @var{errno} of the C library as the C standard allows it to be
2323 a macro. Use the @code{IERRNO} intrinsic (GNU extension) instead.
2325 @node Interoperable Subroutines and Functions
2326 @subsection Interoperable Subroutines and Functions
2328 Subroutines and functions have to have the @code{BIND(C)} attribute to
2329 be compatible with C. The dummy argument declaration is relatively
2330 straightforward. However, one needs to be careful because C uses
2331 call-by-value by default while Fortran behaves usually similar to
2332 call-by-reference. Furthermore, strings and pointers are handled
2333 differently. Note that in Fortran 2003 and 2008 only explicit size
2334 and assumed-size arrays are supported but not assumed-shape or
2335 deferred-shape (i.e. allocatable or pointer) arrays. However, those
2336 are allowed since the Technical Specification 29113, see
2337 @ref{Further Interoperability of Fortran with C}
2339 To pass a variable by value, use the @code{VALUE} attribute.
2340 Thus, the following C prototype
2343 @code{int func(int i, int *j)}
2346 matches the Fortran declaration
2349 integer(c_int) function func(i,j)
2350 use iso_c_binding, only: c_int
2351 integer(c_int), VALUE :: i
2355 Note that pointer arguments also frequently need the @code{VALUE} attribute,
2356 see @ref{Working with Pointers}.
2358 Strings are handled quite differently in C and Fortran. In C a string
2359 is a @code{NUL}-terminated array of characters while in Fortran each string
2360 has a length associated with it and is thus not terminated (by e.g.
2361 @code{NUL}). For example, if one wants to use the following C function,
2365 void print_C(char *string) /* equivalent: char string[] */
2367 printf("%s\n", string);
2371 to print ``Hello World'' from Fortran, one can call it using
2374 use iso_c_binding, only: C_CHAR, C_NULL_CHAR
2376 subroutine print_c(string) bind(C, name="print_C")
2377 use iso_c_binding, only: c_char
2378 character(kind=c_char) :: string(*)
2379 end subroutine print_c
2381 call print_c(C_CHAR_"Hello World"//C_NULL_CHAR)
2384 As the example shows, one needs to ensure that the
2385 string is @code{NUL} terminated. Additionally, the dummy argument
2386 @var{string} of @code{print_C} is a length-one assumed-size
2387 array; using @code{character(len=*)} is not allowed. The example
2388 above uses @code{c_char_"Hello World"} to ensure the string
2389 literal has the right type; typically the default character
2390 kind and @code{c_char} are the same and thus @code{"Hello World"}
2391 is equivalent. However, the standard does not guarantee this.
2393 The use of strings is now further illustrated using the C library
2394 function @code{strncpy}, whose prototype is
2397 char *strncpy(char *restrict s1, const char *restrict s2, size_t n);
2400 The function @code{strncpy} copies at most @var{n} characters from
2401 string @var{s2} to @var{s1} and returns @var{s1}. In the following
2402 example, we ignore the return value:
2407 character(len=30) :: str,str2
2409 ! Ignore the return value of strncpy -> subroutine
2410 ! "restrict" is always assumed if we do not pass a pointer
2411 subroutine strncpy(dest, src, n) bind(C)
2413 character(kind=c_char), intent(out) :: dest(*)
2414 character(kind=c_char), intent(in) :: src(*)
2415 integer(c_size_t), value, intent(in) :: n
2416 end subroutine strncpy
2418 str = repeat('X',30) ! Initialize whole string with 'X'
2419 call strncpy(str, c_char_"Hello World"//C_NULL_CHAR, &
2420 len(c_char_"Hello World",kind=c_size_t))
2421 print '(a)', str ! prints: "Hello WorldXXXXXXXXXXXXXXXXXXX"
2425 The intrinsic procedures are described in @ref{Intrinsic Procedures}.
2427 @node Working with Pointers
2428 @subsection Working with Pointers
2430 C pointers are represented in Fortran via the special opaque derived type
2431 @code{type(c_ptr)} (with private components). Thus one needs to
2432 use intrinsic conversion procedures to convert from or to C pointers.
2434 For some applications, using an assumed type (@code{TYPE(*)}) can be an
2435 alternative to a C pointer; see
2436 @ref{Further Interoperability of Fortran with C}.
2442 type(c_ptr) :: cptr1, cptr2
2443 integer, target :: array(7), scalar
2444 integer, pointer :: pa(:), ps
2445 cptr1 = c_loc(array(1)) ! The programmer needs to ensure that the
2446 ! array is contiguous if required by the C
2448 cptr2 = c_loc(scalar)
2449 call c_f_pointer(cptr2, ps)
2450 call c_f_pointer(cptr2, pa, shape=[7])
2453 When converting C to Fortran arrays, the one-dimensional @code{SHAPE} argument
2456 If a pointer is a dummy-argument of an interoperable procedure, it usually
2457 has to be declared using the @code{VALUE} attribute. @code{void*}
2458 matches @code{TYPE(C_PTR), VALUE}, while @code{TYPE(C_PTR)} alone
2459 matches @code{void**}.
2461 Procedure pointers are handled analogously to pointers; the C type is
2462 @code{TYPE(C_FUNPTR)} and the intrinsic conversion procedures are
2463 @code{C_F_PROCPOINTER} and @code{C_FUNLOC}.
2465 Let us consider two examples of actually passing a procedure pointer from
2466 C to Fortran and vice versa. Note that these examples are also very
2467 similar to passing ordinary pointers between both languages. First,
2468 consider this code in C:
2471 /* Procedure implemented in Fortran. */
2472 void get_values (void (*)(double));
2474 /* Call-back routine we want called from Fortran. */
2478 printf ("Number is %f.\n", x);
2481 /* Call Fortran routine and pass call-back to it. */
2485 get_values (&print_it);
2489 A matching implementation for @code{get_values} in Fortran, that correctly
2490 receives the procedure pointer from C and is able to call it, is given
2491 in the following @code{MODULE}:
2497 ! Define interface of call-back routine.
2499 SUBROUTINE callback (x)
2500 USE, INTRINSIC :: ISO_C_BINDING
2501 REAL(KIND=C_DOUBLE), INTENT(IN), VALUE :: x
2502 END SUBROUTINE callback
2507 ! Define C-bound procedure.
2508 SUBROUTINE get_values (cproc) BIND(C)
2509 USE, INTRINSIC :: ISO_C_BINDING
2510 TYPE(C_FUNPTR), INTENT(IN), VALUE :: cproc
2512 PROCEDURE(callback), POINTER :: proc
2514 ! Convert C to Fortran procedure pointer.
2515 CALL C_F_PROCPOINTER (cproc, proc)
2518 CALL proc (1.0_C_DOUBLE)
2519 CALL proc (-42.0_C_DOUBLE)
2520 CALL proc (18.12_C_DOUBLE)
2521 END SUBROUTINE get_values
2526 Next, we want to call a C routine that expects a procedure pointer argument
2527 and pass it a Fortran procedure (which clearly must be interoperable!).
2528 Again, the C function may be:
2532 call_it (int (*func)(int), int arg)
2538 It can be used as in the following Fortran code:
2542 USE, INTRINSIC :: ISO_C_BINDING
2545 ! Define interface of C function.
2547 INTEGER(KIND=C_INT) FUNCTION call_it (func, arg) BIND(C)
2548 USE, INTRINSIC :: ISO_C_BINDING
2549 TYPE(C_FUNPTR), INTENT(IN), VALUE :: func
2550 INTEGER(KIND=C_INT), INTENT(IN), VALUE :: arg
2551 END FUNCTION call_it
2556 ! Define procedure passed to C function.
2557 ! It must be interoperable!
2558 INTEGER(KIND=C_INT) FUNCTION double_it (arg) BIND(C)
2559 INTEGER(KIND=C_INT), INTENT(IN), VALUE :: arg
2560 double_it = arg + arg
2561 END FUNCTION double_it
2564 SUBROUTINE foobar ()
2565 TYPE(C_FUNPTR) :: cproc
2566 INTEGER(KIND=C_INT) :: i
2568 ! Get C procedure pointer.
2569 cproc = C_FUNLOC (double_it)
2572 DO i = 1_C_INT, 10_C_INT
2573 PRINT *, call_it (cproc, i)
2575 END SUBROUTINE foobar
2580 @node Further Interoperability of Fortran with C
2581 @subsection Further Interoperability of Fortran with C
2583 The Technical Specification ISO/IEC TS 29113:2012 on further
2584 interoperability of Fortran with C extends the interoperability support
2585 of Fortran 2003 and Fortran 2008. Besides removing some restrictions
2586 and constraints, it adds assumed-type (@code{TYPE(*)}) and assumed-rank
2587 (@code{dimension}) variables and allows for interoperability of
2588 assumed-shape, assumed-rank and deferred-shape arrays, including
2589 allocatables and pointers.
2591 Note: Currently, GNU Fortran does not support the array descriptor
2592 (dope vector) as specified in the Technical Specification, but uses
2593 an array descriptor with different fields. The Chasm Language
2594 Interoperability Tools, @url{http://chasm-interop.sourceforge.net/},
2595 provide an interface to GNU Fortran's array descriptor.
2597 The Technical Specification adds the following new features, which
2598 are supported by GNU Fortran:
2602 @item The @code{ASYNCHRONOUS} attribute has been clarified and
2603 extended to allow its use with asynchronous communication in
2604 user-provided libraries such as in implementations of the
2605 Message Passing Interface specification.
2607 @item Many constraints have been relaxed, in particular for
2608 the @code{C_LOC} and @code{C_F_POINTER} intrinsics.
2610 @item The @code{OPTIONAL} attribute is now allowed for dummy
2611 arguments; an absent argument matches a @code{NULL} pointer.
2613 @item Assumed types (@code{TYPE(*)}) have been added, which may
2614 only be used for dummy arguments. They are unlimited polymorphic
2615 but contrary to @code{CLASS(*)} they do not contain any type
2616 information, similar to C's @code{void *} pointers. Expressions
2617 of any type and kind can be passed; thus, it can be used as
2618 replacement for @code{TYPE(C_PTR)}, avoiding the use of
2619 @code{C_LOC} in the caller.
2621 Note, however, that @code{TYPE(*)} only accepts scalar arguments,
2622 unless the @code{DIMENSION} is explicitly specified. As
2623 @code{DIMENSION(*)} only supports array (including array elements) but
2624 no scalars, it is not a full replacement for @code{C_LOC}. On the
2625 other hand, assumed-type assumed-rank dummy arguments
2626 (@code{TYPE(*), DIMENSION(..)}) allow for both scalars and arrays, but
2627 require special code on the callee side to handle the array descriptor.
2629 @item Assumed-rank arrays (@code{DIMENSION(..)}) as dummy argument
2630 allow that scalars and arrays of any rank can be passed as actual
2631 argument. As the Technical Specification does not provide for direct
2632 means to operate with them, they have to be used either from the C side
2633 or be converted using @code{C_LOC} and @code{C_F_POINTER} to scalars
2634 or arrays of a specific rank. The rank can be determined using the
2635 @code{RANK} intrinisic.
2639 Currently unimplemented:
2643 @item GNU Fortran always uses an array descriptor, which does not
2644 match the one of the Technical Specification. The
2645 @code{ISO_Fortran_binding.h} header file and the C functions it
2646 specifies are not available.
2648 @item Using assumed-shape, assumed-rank and deferred-shape arrays in
2649 @code{BIND(C)} procedures is not fully supported. In particular,
2650 C interoperable strings of other length than one are not supported
2651 as this requires the new array descriptor.
2655 @node GNU Fortran Compiler Directives
2656 @section GNU Fortran Compiler Directives
2658 The Fortran standard describes how a conforming program shall
2659 behave; however, the exact implementation is not standardized. In order
2660 to allow the user to choose specific implementation details, compiler
2661 directives can be used to set attributes of variables and procedures
2662 which are not part of the standard. Whether a given attribute is
2663 supported and its exact effects depend on both the operating system and
2664 on the processor; see
2665 @ref{Top,,C Extensions,gcc,Using the GNU Compiler Collection (GCC)}
2668 For procedures and procedure pointers, the following attributes can
2669 be used to change the calling convention:
2672 @item @code{CDECL} -- standard C calling convention
2673 @item @code{STDCALL} -- convention where the called procedure pops the stack
2674 @item @code{FASTCALL} -- part of the arguments are passed via registers
2675 instead using the stack
2678 Besides changing the calling convention, the attributes also influence
2679 the decoration of the symbol name, e.g., by a leading underscore or by
2680 a trailing at-sign followed by the number of bytes on the stack. When
2681 assigning a procedure to a procedure pointer, both should use the same
2684 On some systems, procedures and global variables (module variables and
2685 @code{COMMON} blocks) need special handling to be accessible when they
2686 are in a shared library. The following attributes are available:
2689 @item @code{DLLEXPORT} -- provide a global pointer to a pointer in the DLL
2690 @item @code{DLLIMPORT} -- reference the function or variable using a global pointer
2693 For dummy arguments, the @code{NO_ARG_CHECK} attribute can be used; in
2694 other compilers, it is also known as @code{IGNORE_TKR}. For dummy arguments
2695 with this attribute actual arguments of any type and kind (similar to
2696 @code{TYPE(*)}), scalars and arrays of any rank (no equivalent
2697 in Fortran standard) are accepted. As with @code{TYPE(*)}, the argument
2698 is unlimited polymorphic and no type information is available.
2699 Additionally, the argument may only be passed to dummy arguments
2700 with the @code{NO_ARG_CHECK} attribute and as argument to the
2701 @code{PRESENT} intrinsic function and to @code{C_LOC} of the
2702 @code{ISO_C_BINDING} module.
2704 Variables with @code{NO_ARG_CHECK} attribute shall be of assumed-type
2705 (@code{TYPE(*)}; recommended) or of type @code{INTEGER}, @code{LOGICAL},
2706 @code{REAL} or @code{COMPLEX}. They shall not have the @code{ALLOCATE},
2707 @code{CODIMENSION}, @code{INTENT(OUT)}, @code{POINTER} or @code{VALUE}
2708 attribute; furthermore, they shall be either scalar or of assumed-size
2709 (@code{dimension(*)}). As @code{TYPE(*)}, the @code{NO_ARG_CHECK} attribute
2710 requires an explicit interface.
2713 @item @code{NO_ARG_CHECK} -- disable the type, kind and rank checking
2717 The attributes are specified using the syntax
2719 @code{!GCC$ ATTRIBUTES} @var{attribute-list} @code{::} @var{variable-list}
2721 where in free-form source code only whitespace is allowed before @code{!GCC$}
2722 and in fixed-form source code @code{!GCC$}, @code{cGCC$} or @code{*GCC$} shall
2723 start in the first column.
2725 For procedures, the compiler directives shall be placed into the body
2726 of the procedure; for variables and procedure pointers, they shall be in
2727 the same declaration part as the variable or procedure pointer.
2731 @node Non-Fortran Main Program
2732 @section Non-Fortran Main Program
2735 * _gfortran_set_args:: Save command-line arguments
2736 * _gfortran_set_options:: Set library option flags
2737 * _gfortran_set_convert:: Set endian conversion
2738 * _gfortran_set_record_marker:: Set length of record markers
2739 * _gfortran_set_fpe:: Set when a Floating Point Exception should be raised
2740 * _gfortran_set_max_subrecord_length:: Set subrecord length
2743 Even if you are doing mixed-language programming, it is very
2744 likely that you do not need to know or use the information in this
2745 section. Since it is about the internal structure of GNU Fortran,
2746 it may also change in GCC minor releases.
2748 When you compile a @code{PROGRAM} with GNU Fortran, a function
2749 with the name @code{main} (in the symbol table of the object file)
2750 is generated, which initializes the libgfortran library and then
2751 calls the actual program which uses the name @code{MAIN__}, for
2752 historic reasons. If you link GNU Fortran compiled procedures
2753 to, e.g., a C or C++ program or to a Fortran program compiled by
2754 a different compiler, the libgfortran library is not initialized
2755 and thus a few intrinsic procedures do not work properly, e.g.
2756 those for obtaining the command-line arguments.
2758 Therefore, if your @code{PROGRAM} is not compiled with
2759 GNU Fortran and the GNU Fortran compiled procedures require
2760 intrinsics relying on the library initialization, you need to
2761 initialize the library yourself. Using the default options,
2762 gfortran calls @code{_gfortran_set_args} and
2763 @code{_gfortran_set_options}. The initialization of the former
2764 is needed if the called procedures access the command line
2765 (and for backtracing); the latter sets some flags based on the
2766 standard chosen or to enable backtracing. In typical programs,
2767 it is not necessary to call any initialization function.
2769 If your @code{PROGRAM} is compiled with GNU Fortran, you shall
2770 not call any of the following functions. The libgfortran
2771 initialization functions are shown in C syntax but using C
2772 bindings they are also accessible from Fortran.
2775 @node _gfortran_set_args
2776 @subsection @code{_gfortran_set_args} --- Save command-line arguments
2777 @fnindex _gfortran_set_args
2778 @cindex libgfortran initialization, set_args
2781 @item @emph{Description}:
2782 @code{_gfortran_set_args} saves the command-line arguments; this
2783 initialization is required if any of the command-line intrinsics
2784 is called. Additionally, it shall be called if backtracing is
2785 enabled (see @code{_gfortran_set_options}).
2787 @item @emph{Syntax}:
2788 @code{void _gfortran_set_args (int argc, char *argv[])}
2790 @item @emph{Arguments}:
2791 @multitable @columnfractions .15 .70
2792 @item @var{argc} @tab number of command line argument strings
2793 @item @var{argv} @tab the command-line argument strings; argv[0]
2794 is the pathname of the executable itself.
2797 @item @emph{Example}:
2799 int main (int argc, char *argv[])
2801 /* Initialize libgfortran. */
2802 _gfortran_set_args (argc, argv);
2809 @node _gfortran_set_options
2810 @subsection @code{_gfortran_set_options} --- Set library option flags
2811 @fnindex _gfortran_set_options
2812 @cindex libgfortran initialization, set_options
2815 @item @emph{Description}:
2816 @code{_gfortran_set_options} sets several flags related to the Fortran
2817 standard to be used, whether backtracing should be enabled
2818 and whether range checks should be performed. The syntax allows for
2819 upward compatibility since the number of passed flags is specified; for
2820 non-passed flags, the default value is used. See also
2821 @pxref{Code Gen Options}. Please note that not all flags are actually
2824 @item @emph{Syntax}:
2825 @code{void _gfortran_set_options (int num, int options[])}
2827 @item @emph{Arguments}:
2828 @multitable @columnfractions .15 .70
2829 @item @var{num} @tab number of options passed
2830 @item @var{argv} @tab The list of flag values
2833 @item @emph{option flag list}:
2834 @multitable @columnfractions .15 .70
2835 @item @var{option}[0] @tab Allowed standard; can give run-time errors
2836 if e.g. an input-output edit descriptor is invalid in a given standard.
2837 Possible values are (bitwise or-ed) @code{GFC_STD_F77} (1),
2838 @code{GFC_STD_F95_OBS} (2), @code{GFC_STD_F95_DEL} (4), @code{GFC_STD_F95}
2839 (8), @code{GFC_STD_F2003} (16), @code{GFC_STD_GNU} (32),
2840 @code{GFC_STD_LEGACY} (64), @code{GFC_STD_F2008} (128),
2841 @code{GFC_STD_F2008_OBS} (256) and GFC_STD_F2008_TS (512). Default:
2842 @code{GFC_STD_F95_OBS | GFC_STD_F95_DEL | GFC_STD_F95 | GFC_STD_F2003
2843 | GFC_STD_F2008 | GFC_STD_F2008_TS | GFC_STD_F2008_OBS | GFC_STD_F77
2844 | GFC_STD_GNU | GFC_STD_LEGACY}.
2845 @item @var{option}[1] @tab Standard-warning flag; prints a warning to
2846 standard error. Default: @code{GFC_STD_F95_DEL | GFC_STD_LEGACY}.
2847 @item @var{option}[2] @tab If non zero, enable pedantic checking.
2849 @item @var{option}[3] @tab Unused.
2850 @item @var{option}[4] @tab If non zero, enable backtracing on run-time
2851 errors. Default: off. (Default in the compiler: on.)
2852 Note: Installs a signal handler and requires command-line
2853 initialization using @code{_gfortran_set_args}.
2854 @item @var{option}[5] @tab If non zero, supports signed zeros.
2856 @item @var{option}[6] @tab Enables run-time checking. Possible values
2857 are (bitwise or-ed): GFC_RTCHECK_BOUNDS (1), GFC_RTCHECK_ARRAY_TEMPS (2),
2858 GFC_RTCHECK_RECURSION (4), GFC_RTCHECK_DO (16), GFC_RTCHECK_POINTER (32).
2860 @item @var{option}[7] @tab Unused.
2861 @item @var{option}[8] @tab Show a warning when invoking @code{STOP} and
2862 @code{ERROR STOP} if a floating-point exception occurred. Possible values
2863 are (bitwise or-ed) @code{GFC_FPE_INVALID} (1), @code{GFC_FPE_DENORMAL} (2),
2864 @code{GFC_FPE_ZERO} (4), @code{GFC_FPE_OVERFLOW} (8),
2865 @code{GFC_FPE_UNDERFLOW} (16), @code{GFC_FPE_INEXACT} (32). Default: None (0).
2866 (Default in the compiler: @code{GFC_FPE_INVALID | GFC_FPE_DENORMAL |
2867 GFC_FPE_ZERO | GFC_FPE_OVERFLOW | GFC_FPE_UNDERFLOW}.)
2870 @item @emph{Example}:
2872 /* Use gfortran 4.9 default options. */
2873 static int options[] = @{68, 511, 0, 0, 1, 1, 0, 0, 31@};
2874 _gfortran_set_options (9, &options);
2879 @node _gfortran_set_convert
2880 @subsection @code{_gfortran_set_convert} --- Set endian conversion
2881 @fnindex _gfortran_set_convert
2882 @cindex libgfortran initialization, set_convert
2885 @item @emph{Description}:
2886 @code{_gfortran_set_convert} set the representation of data for
2889 @item @emph{Syntax}:
2890 @code{void _gfortran_set_convert (int conv)}
2892 @item @emph{Arguments}:
2893 @multitable @columnfractions .15 .70
2894 @item @var{conv} @tab Endian conversion, possible values:
2895 GFC_CONVERT_NATIVE (0, default), GFC_CONVERT_SWAP (1),
2896 GFC_CONVERT_BIG (2), GFC_CONVERT_LITTLE (3).
2899 @item @emph{Example}:
2901 int main (int argc, char *argv[])
2903 /* Initialize libgfortran. */
2904 _gfortran_set_args (argc, argv);
2905 _gfortran_set_convert (1);
2912 @node _gfortran_set_record_marker
2913 @subsection @code{_gfortran_set_record_marker} --- Set length of record markers
2914 @fnindex _gfortran_set_record_marker
2915 @cindex libgfortran initialization, set_record_marker
2918 @item @emph{Description}:
2919 @code{_gfortran_set_record_marker} sets the length of record markers
2920 for unformatted files.
2922 @item @emph{Syntax}:
2923 @code{void _gfortran_set_record_marker (int val)}
2925 @item @emph{Arguments}:
2926 @multitable @columnfractions .15 .70
2927 @item @var{val} @tab Length of the record marker; valid values
2928 are 4 and 8. Default is 4.
2931 @item @emph{Example}:
2933 int main (int argc, char *argv[])
2935 /* Initialize libgfortran. */
2936 _gfortran_set_args (argc, argv);
2937 _gfortran_set_record_marker (8);
2944 @node _gfortran_set_fpe
2945 @subsection @code{_gfortran_set_fpe} --- Enable floating point exception traps
2946 @fnindex _gfortran_set_fpe
2947 @cindex libgfortran initialization, set_fpe
2950 @item @emph{Description}:
2951 @code{_gfortran_set_fpe} enables floating point exception traps for
2952 the specified exceptions. On most systems, this will result in a
2953 SIGFPE signal being sent and the program being aborted.
2955 @item @emph{Syntax}:
2956 @code{void _gfortran_set_fpe (int val)}
2958 @item @emph{Arguments}:
2959 @multitable @columnfractions .15 .70
2960 @item @var{option}[0] @tab IEEE exceptions. Possible values are
2961 (bitwise or-ed) zero (0, default) no trapping,
2962 @code{GFC_FPE_INVALID} (1), @code{GFC_FPE_DENORMAL} (2),
2963 @code{GFC_FPE_ZERO} (4), @code{GFC_FPE_OVERFLOW} (8),
2964 @code{GFC_FPE_UNDERFLOW} (16), and @code{GFC_FPE_INEXACT} (32).
2967 @item @emph{Example}:
2969 int main (int argc, char *argv[])
2971 /* Initialize libgfortran. */
2972 _gfortran_set_args (argc, argv);
2973 /* FPE for invalid operations such as SQRT(-1.0). */
2974 _gfortran_set_fpe (1);
2981 @node _gfortran_set_max_subrecord_length
2982 @subsection @code{_gfortran_set_max_subrecord_length} --- Set subrecord length
2983 @fnindex _gfortran_set_max_subrecord_length
2984 @cindex libgfortran initialization, set_max_subrecord_length
2987 @item @emph{Description}:
2988 @code{_gfortran_set_max_subrecord_length} set the maximum length
2989 for a subrecord. This option only makes sense for testing and
2990 debugging of unformatted I/O.
2992 @item @emph{Syntax}:
2993 @code{void _gfortran_set_max_subrecord_length (int val)}
2995 @item @emph{Arguments}:
2996 @multitable @columnfractions .15 .70
2997 @item @var{val} @tab the maximum length for a subrecord;
2998 the maximum permitted value is 2147483639, which is also
3002 @item @emph{Example}:
3004 int main (int argc, char *argv[])
3006 /* Initialize libgfortran. */
3007 _gfortran_set_args (argc, argv);
3008 _gfortran_set_max_subrecord_length (8);
3015 @node Naming and argument-passing conventions
3016 @section Naming and argument-passing conventions
3018 This section gives an overview about the naming convention of procedures
3019 and global variables and about the argument passing conventions used by
3020 GNU Fortran. If a C binding has been specified, the naming convention
3021 and some of the argument-passing conventions change. If possible,
3022 mixed-language and mixed-compiler projects should use the better defined
3023 C binding for interoperability. See @pxref{Interoperability with C}.
3026 * Naming conventions::
3027 * Argument passing conventions::
3031 @node Naming conventions
3032 @subsection Naming conventions
3034 According the Fortran standard, valid Fortran names consist of a letter
3035 between @code{A} to @code{Z}, @code{a} to @code{z}, digits @code{0},
3036 @code{1} to @code{9} and underscores (@code{_}) with the restriction
3037 that names may only start with a letter. As vendor extension, the
3038 dollar sign (@code{$}) is additionally permitted with the option
3039 @option{-fdollar-ok}, but not as first character and only if the
3040 target system supports it.
3042 By default, the procedure name is the lower-cased Fortran name with an
3043 appended underscore (@code{_}); using @option{-fno-underscoring} no
3044 underscore is appended while @code{-fsecond-underscore} appends two
3045 underscores. Depending on the target system and the calling convention,
3046 the procedure might be additionally dressed; for instance, on 32bit
3047 Windows with @code{stdcall}, an at-sign @code{@@} followed by an integer
3048 number is appended. For the changing the calling convention, see
3049 @pxref{GNU Fortran Compiler Directives}.
3051 For common blocks, the same convention is used, i.e. by default an
3052 underscore is appended to the lower-cased Fortran name. Blank commons
3053 have the name @code{__BLNK__}.
3055 For procedures and variables declared in the specification space of a
3056 module, the name is formed by @code{__}, followed by the lower-cased
3057 module name, @code{_MOD_}, and the lower-cased Fortran name. Note that
3058 no underscore is appended.
3061 @node Argument passing conventions
3062 @subsection Argument passing conventions
3064 Subroutines do not return a value (matching C99's @code{void}) while
3065 functions either return a value as specified in the platform ABI or
3066 the result variable is passed as hidden argument to the function and
3067 no result is returned. A hidden result variable is used when the
3068 result variable is an array or of type @code{CHARACTER}.
3070 Arguments are passed according to the platform ABI. In particular,
3071 complex arguments might not be compatible to a struct with two real
3072 components for the real and imaginary part. The argument passing
3073 matches the one of C99's @code{_Complex}. Functions with scalar
3074 complex result variables return their value and do not use a
3075 by-reference argument. Note that with the @option{-ff2c} option,
3076 the argument passing is modified and no longer completely matches
3077 the platform ABI. Some other Fortran compilers use @code{f2c}
3078 semantic by default; this might cause problems with
3081 GNU Fortran passes most arguments by reference, i.e. by passing a
3082 pointer to the data. Note that the compiler might use a temporary
3083 variable into which the actual argument has been copied, if required
3084 semantically (copy-in/copy-out).
3086 For arguments with @code{ALLOCATABLE} and @code{POINTER}
3087 attribute (including procedure pointers), a pointer to the pointer
3088 is passed such that the pointer address can be modified in the
3091 For dummy arguments with the @code{VALUE} attribute: Scalar arguments
3092 of the type @code{INTEGER}, @code{LOGICAL}, @code{REAL} and
3093 @code{COMPLEX} are passed by value according to the platform ABI.
3094 (As vendor extension and not recommended, using @code{%VAL()} in the
3095 call to a procedure has the same effect.) For @code{TYPE(C_PTR)} and
3096 procedure pointers, the pointer itself is passed such that it can be
3097 modified without affecting the caller.
3098 @c FIXME: Document how VALUE is handled for CHARACTER, TYPE,
3099 @c CLASS and arrays, i.e. whether the copy-in is done in the caller
3100 @c or in the callee.
3102 For Boolean (@code{LOGICAL}) arguments, please note that GCC expects
3103 only the integer value 0 and 1. If a GNU Fortran @code{LOGICAL}
3104 variable contains another integer value, the result is undefined.
3105 As some other Fortran compilers use @math{-1} for @code{.TRUE.},
3106 extra care has to be taken -- such as passing the value as
3107 @code{INTEGER}. (The same value restriction also applies to other
3108 front ends of GCC, e.g. to GCC's C99 compiler for @code{_Bool}
3109 or GCC's Ada compiler for @code{Boolean}.)
3111 For arguments of @code{CHARACTER} type, the character length is passed
3112 as hidden argument. For deferred-length strings, the value is passed
3113 by reference, otherwise by value. The character length has the type
3114 @code{INTEGER(kind=4)}. Note with C binding, @code{CHARACTER(len=1)}
3115 result variables are returned according to the platform ABI and no
3116 hidden length argument is used for dummy arguments; with @code{VALUE},
3117 those variables are passed by value.
3119 For @code{OPTIONAL} dummy arguments, an absent argument is denoted
3120 by a NULL pointer, except for scalar dummy arguments of type
3121 @code{INTEGER}, @code{LOGICAL}, @code{REAL} and @code{COMPLEX}
3122 which have the @code{VALUE} attribute. For those, a hidden Boolean
3123 argument (@code{logical(kind=C_bool),value}) is used to indicate
3124 whether the argument is present.
3126 Arguments which are assumed-shape, assumed-rank or deferred-rank
3127 arrays or, with @option{-fcoarray=lib}, allocatable scalar coarrays use
3128 an array descriptor. All other arrays pass the address of the
3129 first element of the array. With @option{-fcoarray=lib}, the token
3130 and the offset belonging to nonallocatable coarrays dummy arguments
3131 are passed as hidden argument along the character length hidden
3132 arguments. The token is an oparque pointer identifying the coarray
3133 and the offset is a passed-by-value integer of kind @code{C_PTRDIFF_T},
3134 denoting the byte offset between the base address of the coarray and
3135 the passed scalar or first element of the passed array.
3137 The arguments are passed in the following order
3139 @item Result variable, when the function result is passed by reference
3140 @item Character length of the function result, if it is a of type
3141 @code{CHARACTER} and no C binding is used
3142 @item The arguments in the order in which they appear in the Fortran
3144 @item The the present status for optional arguments with value attribute,
3145 which are internally passed by value
3146 @item The character length and/or coarray token and offset for the first
3147 argument which is a @code{CHARACTER} or a nonallocatable coarray dummy
3148 argument, followed by the hidden arguments of the next dummy argument
3154 @c Intrinsic Procedures
3155 @c ---------------------------------------------------------------------
3157 @include intrinsic.texi
3164 @c ---------------------------------------------------------------------
3166 @c ---------------------------------------------------------------------
3169 @unnumbered Contributing
3170 @cindex Contributing
3172 Free software is only possible if people contribute to efforts
3174 We're always in need of more people helping out with ideas
3175 and comments, writing documentation and contributing code.
3177 If you want to contribute to GNU Fortran,
3178 have a look at the long lists of projects you can take on.
3179 Some of these projects are small,
3180 some of them are large;
3181 some are completely orthogonal to the rest of what is
3182 happening on GNU Fortran,
3183 but others are ``mainstream'' projects in need of enthusiastic hackers.
3184 All of these projects are important!
3185 We will eventually get around to the things here,
3186 but they are also things doable by someone who is willing and able.
3191 * Proposed Extensions::
3196 @section Contributors to GNU Fortran
3197 @cindex Contributors
3201 Most of the parser was hand-crafted by @emph{Andy Vaught}, who is
3202 also the initiator of the whole project. Thanks Andy!
3203 Most of the interface with GCC was written by @emph{Paul Brook}.
3205 The following individuals have contributed code and/or
3206 ideas and significant help to the GNU Fortran project
3207 (in alphabetical order):
3210 @item Janne Blomqvist
3211 @item Steven Bosscher
3214 @item Fran@,{c}ois-Xavier Coudert
3218 @item Bernhard Fischer
3220 @item Richard Guenther
3221 @item Richard Henderson
3222 @item Katherine Holcomb
3224 @item Niels Kristian Bech Jensen
3225 @item Steven Johnson
3226 @item Steven G. Kargl
3234 @item Christopher D. Rickett
3235 @item Richard Sandiford
3236 @item Tobias Schl@"uter
3245 The following people have contributed bug reports,
3246 smaller or larger patches,
3247 and much needed feedback and encouragement for the
3248 GNU Fortran project:
3252 @item Dominique d'Humi@`eres
3254 @item Erik Schnetter
3255 @item Joost VandeVondele
3258 Many other individuals have helped debug,
3259 test and improve the GNU Fortran compiler over the past few years,
3260 and we welcome you to do the same!
3261 If you already have done so,
3262 and you would like to see your name listed in the
3263 list above, please contact us.
3271 @item Help build the test suite
3272 Solicit more code for donation to the test suite: the more extensive the
3273 testsuite, the smaller the risk of breaking things in the future! We can
3274 keep code private on request.
3276 @item Bug hunting/squishing
3277 Find bugs and write more test cases! Test cases are especially very
3278 welcome, because it allows us to concentrate on fixing bugs instead of
3279 isolating them. Going through the bugzilla database at
3280 @url{http://gcc.gnu.org/@/bugzilla/} to reduce testcases posted there and
3281 add more information (for example, for which version does the testcase
3282 work, for which versions does it fail?) is also very helpful.
3287 @node Proposed Extensions
3288 @section Proposed Extensions
3290 Here's a list of proposed extensions for the GNU Fortran compiler, in no particular
3291 order. Most of these are necessary to be fully compatible with
3292 existing Fortran compilers, but they are not part of the official
3293 J3 Fortran 95 standard.
3295 @subsection Compiler extensions:
3298 User-specified alignment rules for structures.
3301 Automatically extend single precision constants to double.
3304 Compile code that conserves memory by dynamically allocating common and
3305 module storage either on stack or heap.
3308 Compile flag to generate code for array conformance checking (suggest -CC).
3311 User control of symbol names (underscores, etc).
3314 Compile setting for maximum size of stack frame size before spilling
3315 parts to static or heap.
3318 Flag to force local variables into static space.
3321 Flag to force local variables onto stack.
3325 @subsection Environment Options
3328 Pluggable library modules for random numbers, linear algebra.
3329 LA should use BLAS calling conventions.
3332 Environment variables controlling actions on arithmetic exceptions like
3333 overflow, underflow, precision loss---Generate NaN, abort, default.
3337 Set precision for fp units that support it (i387).
3340 Variable for setting fp rounding mode.
3343 Variable to fill uninitialized variables with a user-defined bit
3347 Environment variable controlling filename that is opened for that unit
3351 Environment variable to clear/trash memory being freed.
3354 Environment variable to control tracing of allocations and frees.
3357 Environment variable to display allocated memory at normal program end.
3360 Environment variable for filename for * IO-unit.
3363 Environment variable for temporary file directory.
3366 Environment variable forcing standard output to be line buffered (Unix).
3371 @c ---------------------------------------------------------------------
3372 @c GNU General Public License
3373 @c ---------------------------------------------------------------------
3375 @include gpl_v3.texi
3379 @c ---------------------------------------------------------------------
3380 @c GNU Free Documentation License
3381 @c ---------------------------------------------------------------------
3387 @c ---------------------------------------------------------------------
3388 @c Funding Free Software
3389 @c ---------------------------------------------------------------------
3391 @include funding.texi
3393 @c ---------------------------------------------------------------------
3395 @c ---------------------------------------------------------------------
3398 @unnumbered Option Index
3399 @command{gfortran}'s command line options are indexed here without any
3400 initial @samp{-} or @samp{--}. Where an option has both positive and
3401 negative forms (such as -foption and -fno-option), relevant entries in
3402 the manual are indexed under the most appropriate form; it may sometimes
3403 be useful to look up both forms.
3407 @unnumbered Keyword Index