1 @c Copyright (C) 2004, 2005, 2006, 2007
2 @c Free Software Foundation, Inc.
3 @c This is part of the GNU Fortran manual.
4 @c For copying conditions, see the file gfortran.texi.
8 Copyright @copyright{} 2004, 2005, 2006, 2007
9 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.2 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU Fortran compiler.
33 gfortran [@option{-c}|@option{-S}|@option{-E}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}]
39 [@option{-m}@var{machine-option}@dots{}]
40 [@option{-o} @var{outfile}] @var{infile}@dots{}
42 Only the most useful options are listed here; see below for the
46 gpl(7), gfdl(7), fsf-funding(7),
47 cpp(1), gcov(1), gcc(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
48 and the Info entries for @file{gcc}, @file{cpp}, @file{gfortran}, @file{as},
49 @file{ld}, @file{binutils} and @file{gdb}.
52 For instructions on reporting bugs, see
56 See the Info entry for @command{gfortran} for contributors to GCC and
61 @node Invoking GNU Fortran
62 @chapter GNU Fortran Command Options
63 @cindex GNU Fortran command options
64 @cindex command options
65 @cindex options, @command{gfortran} command
67 @c man begin DESCRIPTION
69 The @command{gfortran} command supports all the options supported by the
70 @command{gcc} command. Only options specific to GNU Fortran are documented here.
72 @xref{Invoking GCC,,GCC Command Options,gcc,Using the GNU Compiler
73 Collection (GCC)}, for information
74 on the non-Fortran-specific aspects of the @command{gcc} command (and,
75 therefore, the @command{gfortran} command).
77 @cindex options, negative forms
78 All GCC and GNU Fortran options
79 are accepted both by @command{gfortran} and by @command{gcc}
80 (as well as any other drivers built at the same time,
81 such as @command{g++}),
82 since adding GNU Fortran to the GCC distribution
83 enables acceptance of GNU Fortran options
84 by all of the relevant drivers.
86 In some cases, options have positive and negative forms;
87 the negative form of @option{-ffoo} would be @option{-fno-foo}.
88 This manual documents only one of these two forms, whichever
89 one is not the default.
93 * Option Summary:: Brief list of all @command{gfortran} options,
95 * Fortran Dialect Options:: Controlling the variant of Fortran language
97 * Error and Warning Options:: How picky should the compiler be?
98 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
99 * Directory Options:: Where to find module files
100 * Link Options :: Influencing the linking step
101 * Runtime Options:: Influencing runtime behavior
102 * Code Gen Options:: Specifying conventions for function calls, data layout
104 * Environment Variables:: Environment variables that affect @command{gfortran}.
108 @section Option summary
112 Here is a summary of all the options specific to GNU Fortran, grouped
113 by type. Explanations are in the following sections.
116 @item Fortran Language Options
117 @xref{Fortran Dialect Options,,Options controlling Fortran dialect}.
118 @gccoptlist{-fall-intrinsics -ffree-form -fno-fixed-form @gol
119 -fdollar-ok -fimplicit-none -fmax-identifier-length @gol
120 -std=@var{std} -fd-lines-as-code -fd-lines-as-comments @gol
121 -ffixed-line-length-@var{n} -ffixed-line-length-none @gol
122 -ffree-line-length-@var{n} -ffree-line-length-none @gol
123 -fdefault-double-8 -fdefault-integer-8 -fdefault-real-8 @gol
124 -fcray-pointer -fopenmp -fno-range-check -fbackslash -fmodule-private}
126 @item Error and Warning Options
127 @xref{Error and Warning Options,,Options to request or suppress errors
129 @gccoptlist{-fmax-errors=@var{n} @gol
130 -fsyntax-only -pedantic -pedantic-errors @gol
131 -Wall -Waliasing -Wampersand -Wcharacter-truncation -Wconversion @gol
132 -Wimplicit-interface -Wline-truncation -Wnonstd-intrinsics -Wsurprising @gol
133 -Wno-tabs -Wunderflow -Wunused-parameter}
135 @item Debugging Options
136 @xref{Debugging Options,,Options for debugging your program or GNU Fortran}.
137 @gccoptlist{-fdump-parse-tree -ffpe-trap=@var{list} @gol
138 -fdump-core -fbacktrace}
140 @item Directory Options
141 @xref{Directory Options,,Options for directory search}.
142 @gccoptlist{-I@var{dir} -J@var{dir} -M@var{dir} -fintrinsic-modules-path @var{dir}}
145 @xref{Link Options,,Options for influencing the linking step}.
146 @gccoptlist{-static-libgfortran}
148 @item Runtime Options
149 @xref{Runtime Options,,Options for influencing runtime behavior}.
150 @gccoptlist{-fconvert=@var{conversion} -frecord-marker=@var{length} @gol
151 -fmax-subrecord-length=@var{length} -fsign-zero}
153 @item Code Generation Options
154 @xref{Code Gen Options,,Options for code generation conventions}.
155 @gccoptlist{-fno-automatic -ff2c -fno-underscoring
156 -fsecond-underscore @gol
157 -fbounds-check -fmax-stack-var-size=@var{n} @gol
158 -fpack-derived -frepack-arrays -fshort-enums -fexternal-blas @gol
159 -fblas-matmul-limit=@var{n} -frecursive -finit-local-zero @gol
160 -finit-integer=@var{n} -finit-real=@var{<zero|inf|-inf|nan>} @gol
161 -finit-logical=@var{<true|false>} -finit-character=@var{n}}
165 * Fortran Dialect Options:: Controlling the variant of Fortran language
167 * Error and Warning Options:: How picky should the compiler be?
168 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
169 * Directory Options:: Where to find module files
170 * Link Options :: Influencing the linking step
171 * Runtime Options:: Influencing runtime behavior
172 * Code Gen Options:: Specifying conventions for function calls, data layout
176 @node Fortran Dialect Options
177 @section Options controlling Fortran dialect
178 @cindex dialect options
179 @cindex language, dialect options
180 @cindex options, dialect
182 The following options control the details of the Fortran dialect
183 accepted by the compiler:
188 @opindex @code{ffree-form}
189 @opindex @code{fno-fixed-form}
190 @cindex options, fortran dialect
191 @cindex file format, free
192 @cindex file format, fixed
193 Specify the layout used by the source file. The free form layout
194 was introduced in Fortran 90. Fixed form was traditionally used in
195 older Fortran programs. When neither option is specified, the source
196 form is determined by the file extension.
198 @item -fall-intrinsics
199 @opindex @code{fall-intrinsics}
200 Accept all of the intrinsic procedures provided in libgfortran
201 without regard to the setting of @option{-std}. In particular,
202 this option can be quite useful with @option{-std=f95}. Additionally,
203 @command{gfortran} will ignore @option{-Wnonstd-intrinsics}.
205 @item -fd-lines-as-code
206 @item -fd-lines-as-comments
207 @opindex @code{fd-lines-as-code}
208 @opindex @code{fd-lines-as-comments}
209 Enable special treatment for lines beginning with @code{d} or @code{D}
210 in fixed form sources. If the @option{-fd-lines-as-code} option is
211 given they are treated as if the first column contained a blank. If the
212 @option{-fd-lines-as-comments} option is given, they are treated as
215 @item -fdefault-double-8
216 @opindex @code{fdefault-double-8}
217 Set the @code{DOUBLE PRECISION} type to an 8 byte wide type.
219 @item -fdefault-integer-8
220 @opindex @code{fdefault-integer-8}
221 Set the default integer and logical types to an 8 byte wide type.
222 Do nothing if this is already the default.
224 @item -fdefault-real-8
225 @opindex @code{fdefault-real-8}
226 Set the default real type to an 8 byte wide type.
227 Do nothing if this is already the default.
230 @opindex @code{fdollar-ok}
233 @cindex character set
234 Allow @samp{$} as a valid character in a symbol name.
237 @opindex @code{backslash}
239 @cindex escape characters
240 Change the interpretation of backslashes in string literals
241 from a single backslash character to ``C-style'' escape characters.
242 The following combinations are expanded \a, \b, \f, \n, \r, \t,
243 \v, \\, and \0 to the ASCII characters alert, backspace, form feed,
244 newline, carriage return, horizontal tab, vertical tab, backslash,
245 and NUL, respectively. All other combinations of a character preceded
248 @item -fmodule-private
249 @opindex @code{fmodule-private}
250 @cindex module entities
252 Set the default accessibility of module entities to @code{PRIVATE}.
253 Use-associated entities will not be accessible unless they are explicitly
254 declared as @code{PUBLIC}.
256 @item -ffixed-line-length-@var{n}
257 @opindex @code{ffixed-line-length-}@var{n}
258 @cindex file format, fixed
259 Set column after which characters are ignored in typical fixed-form
260 lines in the source file, and through which spaces are assumed (as
261 if padded to that length) after the ends of short fixed-form lines.
263 Popular values for @var{n} include 72 (the
264 standard and the default), 80 (card image), and 132 (corresponding
265 to ``extended-source'' options in some popular compilers).
266 @var{n} may also be @samp{none}, meaning that the entire line is meaningful
267 and that continued character constants never have implicit spaces appended
268 to them to fill out the line.
269 @option{-ffixed-line-length-0} means the same thing as
270 @option{-ffixed-line-length-none}.
272 @item -ffree-line-length-@var{n}
273 @opindex @code{ffree-line-length-}@var{n}
274 @cindex file format, free
275 Set column after which characters are ignored in typical free-form
276 lines in the source file. The default value is 132.
277 @var{n} may be @samp{none}, meaning that the entire line is meaningful.
278 @option{-ffree-line-length-0} means the same thing as
279 @option{-ffree-line-length-none}.
281 @item -fmax-identifier-length=@var{n}
282 @opindex @code{fmax-identifier-length=}@var{n}
283 Specify the maximum allowed identifier length. Typical values are
284 31 (Fortran 95) and 63 (Fortran 2003).
286 @item -fimplicit-none
287 @opindex @code{fimplicit-none}
288 Specify that no implicit typing is allowed, unless overridden by explicit
289 @code{IMPLICIT} statements. This is the equivalent of adding
290 @code{implicit none} to the start of every procedure.
293 @opindex @code{fcray-pointer}
294 Enable the Cray pointer extension, which provides C-like pointer
298 @opindex @code{fopenmp}
300 Enable the OpenMP extensions. This includes OpenMP @code{!$omp} directives
302 and @code{c$omp}, @code{*$omp} and @code{!$omp} directives in fixed form,
303 @code{!$} conditional compilation sentinels in free form
304 and @code{c$}, @code{*$} and @code{!$} sentinels in fixed form,
305 and when linking arranges for the OpenMP runtime library to be linked
306 in. The option @option{-fopenmp} implies @option{-frecursive}.
308 @item -fno-range-check
309 @opindex @code{frange-check}
310 Disable range checking on results of simplification of constant
311 expressions during compilation. For example, GNU Fortran will give
312 an error at compile time when simplifying @code{a = 1. / 0}.
313 With this option, no error will be given and @code{a} will be assigned
314 the value @code{+Infinity}. If an expression evaluates to a value
315 outside of the relevant range of [@code{-HUGE()}:@code{HUGE()}],
316 then the expression will be replaced by @code{-Inf} or @code{+Inf}
318 Similarly, @code{DATA i/Z'FFFFFFFF'/} will result in an integer overflow
319 on most systems, but with @option{-fno-range-check} the value will
320 ``wrap around'' and @code{i} will be initialized to @math{-1} instead.
323 @opindex @code{std=}@var{std} option
324 Specify the standard to which the program is expected to conform, which
325 may be one of @samp{f95}, @samp{f2003}, @samp{gnu}, or @samp{legacy}.
326 The default value for @var{std} is @samp{gnu}, which specifies a
327 superset of the Fortran 95 standard that includes all of the extensions
328 supported by GNU Fortran, although warnings will be given for obsolete
329 extensions not recommended for use in new code. The @samp{legacy} value
330 is equivalent but without the warnings for obsolete extensions, and may
331 be useful for old non-standard programs. The @samp{f95} and
332 @samp{f2003} values specify strict conformance to the Fortran 95 and
333 Fortran 2003 standards, respectively; errors are given for all
334 extensions beyond the relevant language standard, and warnings are given
335 for the Fortran 77 features that are permitted but obsolescent in later
340 @node Error and Warning Options
341 @section Options to request or suppress errors and warnings
342 @cindex options, warnings
343 @cindex options, errors
344 @cindex warnings, suppressing
345 @cindex messages, error
346 @cindex messages, warning
347 @cindex suppressing warnings
349 Errors are diagnostic messages that report that the GNU Fortran compiler
350 cannot compile the relevant piece of source code. The compiler will
351 continue to process the program in an attempt to report further errors
352 to aid in debugging, but will not produce any compiled output.
354 Warnings are diagnostic messages that report constructions which
355 are not inherently erroneous but which are risky or suggest there is
356 likely to be a bug in the program. Unless @option{-Werror} is specified,
357 they do not prevent compilation of the program.
359 You can request many specific warnings with options beginning @option{-W},
360 for example @option{-Wimplicit} to request warnings on implicit
361 declarations. Each of these specific warning options also has a
362 negative form beginning @option{-Wno-} to turn off warnings;
363 for example, @option{-Wno-implicit}. This manual lists only one of the
364 two forms, whichever is not the default.
366 These options control the amount and kinds of errors and warnings produced
370 @item -fmax-errors=@var{n}
371 @opindex @code{fmax-errors=}@var{n}
372 @cindex errors, limiting
373 Limits the maximum number of error messages to @var{n}, at which point
374 GNU Fortran bails out rather than attempting to continue processing the
375 source code. If @var{n} is 0, there is no limit on the number of error
379 @opindex @code{fsyntax-only}
380 @cindex syntax checking
381 Check the code for syntax errors, but don't actually compile it. This
382 will generate module files for each module present in the code, but no
386 @opindex @code{pedantic}
387 Issue warnings for uses of extensions to Fortran 95.
388 @option{-pedantic} also applies to C-language constructs where they
389 occur in GNU Fortran source files, such as use of @samp{\e} in a
390 character constant within a directive like @code{#include}.
392 Valid Fortran 95 programs should compile properly with or without
394 However, without this option, certain GNU extensions and traditional
395 Fortran features are supported as well.
396 With this option, many of them are rejected.
398 Some users try to use @option{-pedantic} to check programs for conformance.
399 They soon find that it does not do quite what they want---it finds some
400 nonstandard practices, but not all.
401 However, improvements to GNU Fortran in this area are welcome.
403 This should be used in conjunction with @option{-std=f95} or
406 @item -pedantic-errors
407 @opindex @code{pedantic-errors}
408 Like @option{-pedantic}, except that errors are produced rather than
414 @cindex warnings, all
415 Enables commonly used warning options pertaining to usage that
416 we recommend avoiding and that we believe are easy to avoid.
417 This currently includes @option{-Waliasing},
418 @option{-Wampersand}, @option{-Wsurprising}, @option{-Wnonstd-intrinsics},
419 @option{-Wno-tabs}, and @option{-Wline-truncation}.
422 @opindex @code{Waliasing}
424 @cindex warnings, aliasing
425 Warn about possible aliasing of dummy arguments. Specifically, it warns
426 if the same actual argument is associated with a dummy argument with
427 @code{INTENT(IN)} and a dummy argument with @code{INTENT(OUT)} in a call
428 with an explicit interface.
430 The following example will trigger the warning.
434 integer, intent(in) :: a
435 integer, intent(out) :: b
444 @opindex @code{Wampersand}
445 @cindex warnings, ampersand
447 Warn about missing ampersand in continued character constants. The warning is
448 given with @option{-Wampersand}, @option{-pedantic}, @option{-std=f95}, and
449 @option{-std=f2003}. Note: With no ampersand given in a continued character
450 constant, GNU Fortran assumes continuation at the first non-comment,
451 non-whitespace character after the ampersand that initiated the continuation.
453 @item -Wcharacter-truncation
454 @opindex @code{Wcharacter-truncation}
455 @cindex warnings, character truncation
456 Warn when a character assignment will truncate the assigned string.
459 @opindex @code{Wconversion}
460 @cindex warnings, conversion
462 Warn about implicit conversions between different types.
464 @item -Wimplicit-interface
465 @opindex @code{Wimplicit-interface}
466 @cindex warnings, implicit interface
467 Warn if a procedure is called without an explicit interface.
468 Note this only checks that an explicit interface is present. It does not
469 check that the declared interfaces are consistent across program units.
471 @item -Wnonstd-intrinsics
472 @opindex @code{Wnonstd-intrinsics}
473 @cindex warnings, non-standard intrinsics
474 Warn if the user tries to use an intrinsic that does not belong to the
475 standard the user has chosen via the @option{-std} option.
478 @opindex @code{Wsurprising}
479 @cindex warnings, suspicious code
480 Produce a warning when ``suspicious'' code constructs are encountered.
481 While technically legal these usually indicate that an error has been made.
483 This currently produces a warning under the following circumstances:
487 An INTEGER SELECT construct has a CASE that can never be matched as its
488 lower value is greater than its upper value.
491 A LOGICAL SELECT construct has three CASE statements.
494 A TRANSFER specifies a source that is shorter than the destination.
498 @opindex @code{Wtabs}
499 @cindex warnings, tabs
501 By default, tabs are accepted as whitespace, but tabs are not members
502 of the Fortran Character Set. For continuation lines, a tab followed
503 by a digit between 1 and 9 is supported. @option{-Wno-tabs} will cause
504 a warning to be issued if a tab is encountered. Note, @option{-Wno-tabs}
505 is active for @option{-pedantic}, @option{-std=f95}, @option{-std=f2003},
509 @opindex @code{Wunderflow}
510 @cindex warnings, underflow
512 Produce a warning when numerical constant expressions are
513 encountered, which yield an UNDERFLOW during compilation.
515 @item -Wunused-parameter
516 @opindex @code{Wunused-parameter}
517 @cindex warnings, unused parameter
518 @cindex unused parameter
519 Contrary to @command{gcc}'s meaning of @option{-Wunused-parameter},
520 @command{gfortran}'s implementation of this option does not warn
521 about unused dummy arguments, but about unused @code{PARAMETER} values.
522 @option{-Wunused-parameter} is not included in @option{-Wall} but is
523 implied by @option{-Wall -Wextra}.
526 @opindex @code{Werror}
527 @cindex warnings, to errors
528 Turns all warnings into errors.
531 @xref{Error and Warning Options,,Options to Request or Suppress Errors and
532 Warnings, gcc,Using the GNU Compiler Collection (GCC)}, for information on
533 more options offered by the GBE shared by @command{gfortran}, @command{gcc}
534 and other GNU compilers.
536 Some of these have no effect when compiling programs written in Fortran.
538 @node Debugging Options
539 @section Options for debugging your program or GNU Fortran
540 @cindex options, debugging
541 @cindex debugging information options
543 GNU Fortran has various special options that are used for debugging
544 either your program or the GNU Fortran compiler.
547 @item -fdump-parse-tree
548 @opindex @code{fdump-parse-tree}
549 Output the internal parse tree before starting code generation. Only
550 really useful for debugging the GNU Fortran compiler itself.
552 @item -ffpe-trap=@var{list}
553 @opindex @code{ffpe-trap=}@var{list}
554 Specify a list of IEEE exceptions when a Floating Point Exception
555 (FPE) should be raised. On most systems, this will result in a SIGFPE
556 signal being sent and the program being interrupted, producing a core
557 file useful for debugging. @var{list} is a (possibly empty) comma-separated
558 list of the following IEEE exceptions: @samp{invalid} (invalid floating
559 point operation, such as @code{SQRT(-1.0)}), @samp{zero} (division by
560 zero), @samp{overflow} (overflow in a floating point operation),
561 @samp{underflow} (underflow in a floating point operation),
562 @samp{precision} (loss of precision during operation) and @samp{denormal}
563 (operation produced a denormal value).
565 Some of the routines in the Fortran runtime library, like
566 @samp{CPU_TIME}, are likely to to trigger floating point exceptions when
567 @code{ffpe-trap=precision} is used. For this reason, the use of
568 @code{ffpe-trap=precision} is not recommended.
571 @opindex @code{fbacktrace}
574 Specify that, when a runtime error is encountered or a deadly signal is
575 emitted (segmentation fault, illegal instruction, bus error or
576 floating-point exception), the Fortran runtime
577 library should output a backtrace of the error. This option
578 only has influence for compilation of the Fortran main program.
582 @opindex @code{fdump-core}
583 Request that a core-dump file is written to disk when a runtime error
584 is encountered on systems that support core dumps. This option is
585 only effective for the compilation of the Fortran main program.
588 @xref{Debugging Options,,Options for Debugging Your Program or GCC,
589 gcc,Using the GNU Compiler Collection (GCC)}, for more information on
592 @node Directory Options
593 @section Options for directory search
594 @cindex directory, options
595 @cindex options, directory search
597 @cindex INCLUDE directive
598 @cindex directive, INCLUDE
599 These options affect how GNU Fortran searches
600 for files specified by the @code{INCLUDE} directive and where it searches
601 for previously compiled modules.
603 It also affects the search paths used by @command{cpp} when used to preprocess
608 @opindex @code{I}@var{dir}
609 @cindex directory, search paths for inclusion
610 @cindex inclusion, directory search paths for
611 @cindex search paths, for included files
612 @cindex paths, search
613 @cindex module search path
614 These affect interpretation of the @code{INCLUDE} directive
615 (as well as of the @code{#include} directive of the @command{cpp}
618 Also note that the general behavior of @option{-I} and
619 @code{INCLUDE} is pretty much the same as of @option{-I} with
620 @code{#include} in the @command{cpp} preprocessor, with regard to
621 looking for @file{header.gcc} files and other such things.
623 This path is also used to search for @file{.mod} files when previously
624 compiled modules are required by a @code{USE} statement.
626 @xref{Directory Options,,Options for Directory Search,
627 gcc,Using the GNU Compiler Collection (GCC)}, for information on the
632 @opindex @code{M}@var{dir}
633 @opindex @code{J}@var{dir}
634 @cindex paths, search
635 @cindex module search path
636 This option specifies where to put @file{.mod} files for compiled modules.
637 It is also added to the list of directories to searched by an @code{USE}
640 The default is the current directory.
642 @option{-J} is an alias for @option{-M} to avoid conflicts with existing
645 @item -fintrinsic-modules-path @var{dir}
646 @opindex @code{fintrinsic-modules-path} @var{dir}
647 @cindex paths, search
648 @cindex module search path
649 This option specifies the location of pre-compiled intrinsic modules, if
650 they are not in the default location expected by the compiler.
654 @section Influencing the linking step
655 @cindex options, linking
656 @cindex linking, static
658 These options come into play when the compiler links object files into an
659 executable output file. They are meaningless if the compiler is not doing
663 @item -static-libgfortran
664 @opindex @code{static-libgfortran}
665 On systems that provide @file{libgfortran} as a shared and a static
666 library, this option forces the use of the static version. If no
667 shared version of @file{libgfortran} was built when the compiler was
668 configured, this option has no effect.
672 @node Runtime Options
673 @section Influencing runtime behavior
674 @cindex options, runtime
676 These options affect the runtime behavior of programs compiled with GNU Fortran.
678 @item -fconvert=@var{conversion}
679 @opindex @code{fconvert=}@var{conversion}
680 Specify the representation of data for unformatted files. Valid
681 values for conversion are: @samp{native}, the default; @samp{swap},
682 swap between big- and little-endian; @samp{big-endian}, use big-endian
683 representation for unformatted files; @samp{little-endian}, use little-endian
684 representation for unformatted files.
686 @emph{This option has an effect only when used in the main program.
687 The @code{CONVERT} specifier and the GFORTRAN_CONVERT_UNIT environment
688 variable override the default specified by @option{-fconvert}.}
690 @item -frecord-marker=@var{length}
691 @opindex @code{frecord-marker=}@var{length}
692 Specify the length of record markers for unformatted files.
693 Valid values for @var{length} are 4 and 8. Default is 4.
694 @emph{This is different from previous versions of @command{gfortran}},
695 which specified a default record marker length of 8 on most
696 systems. If you want to read or write files compatible
697 with earlier versions of @command{gfortran}, use @option{-frecord-marker=8}.
699 @item -fmax-subrecord-length=@var{length}
700 @opindex @code{fmax-subrecord-length=}@var{length}
701 Specify the maximum length for a subrecord. The maximum permitted
702 value for length is 2147483639, which is also the default. Only
703 really useful for use by the gfortran testsuite.
706 @opindex @code{fsign-zero}
707 When writing zero values, show the negative sign if the sign bit is set.
708 @code{fno-sign-zero} does not print the negative sign of zero values for
709 compatibility with F77. Default behavior is to show the negative sign.
712 @node Code Gen Options
713 @section Options for code generation conventions
714 @cindex code generation, conventions
715 @cindex options, code generation
716 @cindex options, run-time
718 These machine-independent options control the interface conventions
719 used in code generation.
721 Most of them have both positive and negative forms; the negative form
722 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
723 one of the forms is listed---the one which is not the default. You
724 can figure out the other form by either removing @option{no-} or adding
729 @opindex @code{fno-automatic}
730 @cindex @code{SAVE} statement
731 @cindex statement, @code{SAVE}
732 Treat each program unit (except those marked as RECURSIVE) as if the
733 @code{SAVE} statement were specified for every local variable and array
734 referenced in it. Does not affect common blocks. (Some Fortran compilers
735 provide this option under the name @option{-static} or @option{-save}.)
736 The default, which is @option{-fautomatic}, uses the stack for local
737 variables smaller than the value given by @option{-fmax-stack-var-size}.
738 Use the option @option{-frecursive} to use no static memory.
742 @cindex calling convention
743 @cindex @command{f2c} calling convention
744 @cindex @command{g77} calling convention
745 @cindex libf2c calling convention
746 Generate code designed to be compatible with code generated
747 by @command{g77} and @command{f2c}.
749 The calling conventions used by @command{g77} (originally implemented
750 in @command{f2c}) require functions that return type
751 default @code{REAL} to actually return the C type @code{double}, and
752 functions that return type @code{COMPLEX} to return the values via an
753 extra argument in the calling sequence that points to where to
754 store the return value. Under the default GNU calling conventions, such
755 functions simply return their results as they would in GNU
756 C---default @code{REAL} functions return the C type @code{float}, and
757 @code{COMPLEX} functions return the GNU C type @code{complex}.
758 Additionally, this option implies the @option{-fsecond-underscore}
759 option, unless @option{-fno-second-underscore} is explicitly requested.
761 This does not affect the generation of code that interfaces with
762 the @command{libgfortran} library.
764 @emph{Caution:} It is not a good idea to mix Fortran code compiled with
765 @option{-ff2c} with code compiled with the default @option{-fno-f2c}
766 calling conventions as, calling @code{COMPLEX} or default @code{REAL}
767 functions between program parts which were compiled with different
768 calling conventions will break at execution time.
770 @emph{Caution:} This will break code which passes intrinsic functions
771 of type default @code{REAL} or @code{COMPLEX} as actual arguments, as
772 the library implementations use the @option{-fno-f2c} calling conventions.
774 @item -fno-underscoring
775 @opindex @code{fno-underscoring}
777 @cindex symbol names, underscores
778 @cindex transforming symbol names
779 @cindex symbol names, transforming
780 Do not transform names of entities specified in the Fortran
781 source file by appending underscores to them.
783 With @option{-funderscoring} in effect, GNU Fortran appends one
784 underscore to external names with no underscores. This is done to ensure
785 compatibility with code produced by many UNIX Fortran compilers.
787 @emph{Caution}: The default behavior of GNU Fortran is
788 incompatible with @command{f2c} and @command{g77}, please use the
789 @option{-ff2c} option if you want object files compiled with
790 GNU Fortran to be compatible with object code created with these
793 Use of @option{-fno-underscoring} is not recommended unless you are
794 experimenting with issues such as integration of GNU Fortran into
795 existing system environments (vis-@`{a}-vis existing libraries, tools,
798 For example, with @option{-funderscoring}, and assuming other defaults like
799 @option{-fcase-lower} and that @code{j()} and @code{max_count()} are
800 external functions while @code{my_var} and @code{lvar} are local variables,
803 I = J() + MAX_COUNT (MY_VAR, LVAR)
806 is implemented as something akin to:
808 i = j_() + max_count__(&my_var__, &lvar);
811 With @option{-fno-underscoring}, the same statement is implemented as:
814 i = j() + max_count(&my_var, &lvar);
817 Use of @option{-fno-underscoring} allows direct specification of
818 user-defined names while debugging and when interfacing GNU Fortran
819 code with other languages.
821 Note that just because the names match does @emph{not} mean that the
822 interface implemented by GNU Fortran for an external name matches the
823 interface implemented by some other language for that same name.
824 That is, getting code produced by GNU Fortran to link to code produced
825 by some other compiler using this or any other method can be only a
826 small part of the overall solution---getting the code generated by
827 both compilers to agree on issues other than naming can require
828 significant effort, and, unlike naming disagreements, linkers normally
829 cannot detect disagreements in these other areas.
831 Also, note that with @option{-fno-underscoring}, the lack of appended
832 underscores introduces the very real possibility that a user-defined
833 external name will conflict with a name in a system library, which
834 could make finding unresolved-reference bugs quite difficult in some
835 cases---they might occur at program run time, and show up only as
836 buggy behavior at run time.
838 In future versions of GNU Fortran we hope to improve naming and linking
839 issues so that debugging always involves using the names as they appear
840 in the source, even if the names as seen by the linker are mangled to
841 prevent accidental linking between procedures with incompatible
844 @item -fsecond-underscore
845 @opindex @code{fsecond-underscore}
847 @cindex symbol names, underscores
848 @cindex transforming symbol names
849 @cindex symbol names, transforming
850 @cindex @command{f2c} calling convention
851 @cindex @command{g77} calling convention
852 @cindex libf2c calling convention
853 By default, GNU Fortran appends an underscore to external
854 names. If this option is used GNU Fortran appends two
855 underscores to names with underscores and one underscore to external names
856 with no underscores. GNU Fortran also appends two underscores to
857 internal names with underscores to avoid naming collisions with external
860 This option has no effect if @option{-fno-underscoring} is
861 in effect. It is implied by the @option{-ff2c} option.
863 Otherwise, with this option, an external name such as @code{MAX_COUNT}
864 is implemented as a reference to the link-time external symbol
865 @code{max_count__}, instead of @code{max_count_}. This is required
866 for compatibility with @command{g77} and @command{f2c}, and is implied
867 by use of the @option{-ff2c} option.
870 @opindex @code{fbounds-check}
871 @cindex array, bounds checking
872 @cindex bounds checking
873 @cindex range checking
874 @cindex subscript checking
875 @cindex checking subscripts
876 Enable generation of run-time checks for array subscripts
877 and against the declared minimum and maximum values. It also
878 checks array indices for assumed and deferred
879 shape arrays against the actual allocated bounds.
881 Some checks require that @option{-fbounds-check} is set for
882 the compilation of the main program.
884 In the future this may also include other forms of checking, e.g., checking
885 substring references.
887 @item -fmax-stack-var-size=@var{n}
888 @opindex @code{fmax-stack-var-size}
889 This option specifies the size in bytes of the largest array that will be put
890 on the stack; if the size is exceeded static memory is used (except in
891 procedures marked as RECURSIVE). Use the option @option{-frecursive} to
892 allow for recursive procedures which do not have a RECURSIVE attribute or
893 for parallel programs. Use @option{-fno-automatic} to never use the stack.
895 This option currently only affects local arrays declared with constant
896 bounds, and may not apply to all character variables.
897 Future versions of GNU Fortran may improve this behavior.
899 The default value for @var{n} is 32768.
902 @opindex @code{fpack-derived}
903 @cindex structure packing
904 This option tells GNU Fortran to pack derived type members as closely as
905 possible. Code compiled with this option is likely to be incompatible
906 with code compiled without this option, and may execute slower.
908 @item -frepack-arrays
909 @opindex @code{frepack-arrays}
910 @cindex repacking arrays
911 In some circumstances GNU Fortran may pass assumed shape array
912 sections via a descriptor describing a noncontiguous area of memory.
913 This option adds code to the function prologue to repack the data into
914 a contiguous block at runtime.
916 This should result in faster accesses to the array. However it can introduce
917 significant overhead to the function call, especially when the passed data
921 @opindex @code{fshort-enums}
922 This option is provided for interoperability with C code that was
923 compiled with the @option{-fshort-enums} option. It will make
924 GNU Fortran choose the smallest @code{INTEGER} kind a given
925 enumerator set will fit in, and give all its enumerators this kind.
927 @item -fexternal-blas
928 @opindex @code{fexternal-blas}
929 This option will make @command{gfortran} generate calls to BLAS functions
930 for some matrix operations like @code{MATMUL}, instead of using our own
931 algorithms, if the size of the matrices involved is larger than a given
932 limit (see @option{-fblas-matmul-limit}). This may be profitable if an
933 optimized vendor BLAS library is available. The BLAS library will have
934 to be specified at link time.
936 @item -fblas-matmul-limit=@var{n}
937 @opindex @code{fblas-matmul-limit}
938 Only significant when @option{-fexternal-blas} is in effect.
939 Matrix multiplication of matrices with size larger than (or equal to) @var{n}
940 will be performed by calls to BLAS functions, while others will be
941 handled by @command{gfortran} internal algorithms. If the matrices
942 involved are not square, the size comparison is performed using the
943 geometric mean of the dimensions of the argument and result matrices.
945 The default value for @var{n} is 30.
948 @opindex @code{frecursive}
949 Allow indirect recursion by forcing all local arrays to be allocated
950 on the stack. This flag cannot be used together with
951 @option{-fmax-stack-var-size=} or @option{-fno-automatic}.
953 @item -finit-local-zero
954 @item -finit-integer=@var{n}
955 @item -finit-real=@var{<zero|inf|-inf|nan>}
956 @item -finit-logical=@var{<true|false>}
957 @item -finit-character=@var{n}
958 @opindex @code{finit-local-zero}
959 @opindex @code{finit-integer}
960 @opindex @code{finit-real}
961 @opindex @code{finit-logical}
962 @opindex @code{finit-character}
963 The @option{-finit-local-zero} option instructs the compiler to
964 initialize local @code{INTEGER}, @code{REAL}, and @code{COMPLEX}
965 variables to zero, @code{LOGICAL} variables to false, and
966 @code{CHARACTER} variables to a string of null bytes. Finer-grained
967 initialization options are provided by the
968 @option{-finit-integer=@var{n}},
969 @option{-finit-real=@var{<zero|inf|-inf|nan>}} (which also initializes
970 the real and imaginary parts of local @code{COMPLEX} variables),
971 @option{-finit-logical=@var{<true|false>}}, and
972 @option{-finit-character=@var{n}} (where @var{n} is an ASCII character
973 value) options. These options do not initialize components of derived
974 type variables, nor do they initialize variables that appear in an
975 @code{EQUIVALENCE} statement. (This limitation may be removed in
978 Note that the @option{-finit-real=nan} option initializes @code{REAL}
979 and @code{COMPLEX} variables with a quiet NaN.
982 @xref{Code Gen Options,,Options for Code Generation Conventions,
983 gcc,Using the GNU Compiler Collection (GCC)}, for information on more options
985 shared by @command{gfortran}, @command{gcc}, and other GNU compilers.
990 @node Environment Variables
991 @section Environment variables affecting @command{gfortran}
992 @cindex environment variable
994 @c man begin ENVIRONMENT
996 The @command{gfortran} compiler currently does not make use of any environment
997 variables to control its operation above and beyond those
998 that affect the operation of @command{gcc}.
1000 @xref{Environment Variables,,Environment Variables Affecting GCC,
1001 gcc,Using the GNU Compiler Collection (GCC)}, for information on environment
1004 @xref{Runtime}, for environment variables that affect the
1005 run-time behavior of programs compiled with GNU Fortran.