1 @c Copyright (C) 1995-1999 Free Software Foundation, Inc.
2 @c This is part of the G77 manual.
3 @c For copying conditions, see the file g77.texi.
5 @c The text of this file appears in the file NEWS
6 @c in the G77 distribution, as well as in the G77 manual.
8 @c Keep this the same as the dates above, since it's used
9 @c in the standalone derivations of this file (e.g. NEWS).
10 @set copyrights-news 1995-1999
12 @set last-update-news 1999-05-03
17 @c The immediately following lines apply to the NEWS file
18 @c which is derived from this file.
19 @emph{Note:} This file is automatically generated from the files
20 @file{news0.texi} and @file{news.texi}.
21 @file{NEWS} is @emph{not} a source file,
22 although it is normally included within source distributions.
24 This file lists news about the @value{which-g77} version
25 (and some other versions) of the GNU Fortran compiler.
26 Copyright (C) @value{copyrights-news} Free Software Foundation, Inc.
27 You may copy, distribute, and modify it freely as long as you preserve
28 this copyright notice and permission notice.
31 @chapter News About GNU Fortran
37 @chapter User-visible Changes
38 @cindex versions, recent
39 @cindex recent versions
40 @cindex changes, user-visible
41 @cindex user-visible changes
43 This chapter describes changes to @code{g77} that are visible
44 to the programmers who actually write and maintain Fortran
45 code they compile with @code{g77}.
46 Information on changes to installation procedures,
47 changes to the documentation, and bug fixes is
48 not provided here, unless it is likely to affect how
50 @xref{News,,News About GNU Fortran}, for information on
51 such changes to @code{g77}.
56 @chapter News About GNU Fortran
57 @cindex versions, recent
58 @cindex recent versions
63 Changes made to recent versions of GNU Fortran are listed
64 below, with the most recent version first.
66 The changes are generally listed in order:
70 Code-generation and run-time-library bug-fixes
73 Compiler and run-time-library crashes involving valid code
80 Fixes and enhancements to existing features
92 This order is not strict---for example, some items
93 involve a combination of these elements.
96 Note that two variants of @code{g77} are tracked below.
97 The @code{egcs} variant is described vis-a-vis
98 previous versions of @code{egcs} and/or
99 an official FSF version,
102 Therefore, @code{egcs} versions sometimes have multiple listings
103 to help clarify how they differ from other versions,
104 though this can make getting a complete picture
105 of what a particular @code{egcs} version contains
106 somewhat more difficult.
109 For information on bugs in the @value{which-g77} version of @code{g77},
110 see @ref{Known Bugs,,Known Bugs In GNU Fortran}.
114 For information on bugs in the @value{which-g77} version of @code{g77},
115 see @file{@value{path-g77}/BUGS}.
119 @emph{Warning:} The information below is still under development,
120 and might not accurately reflect the @code{g77} code base
121 of which it is a part.
122 Efforts are made to keep it somewhat up-to-date,
123 but they are particularly concentrated
124 on any version of this information
125 that is distributed as part of a @emph{released} @code{g77}.
127 In particular, while this information is intended to apply to
128 the @value{which-g77} version of @code{g77},
129 only an official @emph{release} of that version
130 is expected to contain documentation that is
131 most consistent with the @code{g77} product in that version.
133 Nevertheless, information on @emph{previous} releases of @code{g77}, below,
134 is likely to be more up-to-date and accurate
135 than the equivalent information that accompanied
137 assuming the last-updated date of the information below
138 is later than the dates of those releases.
140 That's due to attempts to keep this development version
141 of news about previous @code{g77} versions up-to-date.
145 An online, ``live'' version of this document
146 (derived directly from the mainline, development version
147 of @code{g77} within @code{egcs})
149 @uref{http://egcs.cygnus.com/onlinedocs/g77_news.html}.
152 The following information was last updated on @value{last-update-news}:
154 @heading In @code{egcs} 1.2 (versus 1.1.2):
158 @code{g77} no longer generates bad code for assignments,
159 or other conversions,
160 of @code{REAL} or @code{COMPLEX} constant expressions
161 to type @code{INTEGER(KIND=2)}
162 (often referred to as @code{INTEGER*8}).
164 For example, @samp{INTEGER*8 J; J = 4E10} now works as documented.
169 @code{g77} no longer generates bad code
170 for an assignment to a @code{COMPLEX} variable or array
171 that partially overlaps one or more of the sources
172 of the same assignment
173 (a very rare construction).
174 It now assigns through a temporary,
175 in cases where such partial overlap is deemed possible.
180 Fix @code{g77} so it no longer crashes when compiling
181 I/O statements using keywords that define @code{INTEGER} values,
182 such as @samp{IOSTAT=@var{j}},
183 where @var{j} is other than default @code{INTEGER}
184 (such as @code{INTEGER*2}).
185 Instead, it issues a diagnostic.
190 Fix @code{g77} so it properly handles @samp{DATA A/@var{rpt}*@var{val}/},
191 where @var{rpt} is not default @code{INTEGER}, such as @code{INTEGER*2},
192 instead of producing a spurious diagnostic.
193 Also fix @samp{DATA (A(I),I=1,N)},
194 where @samp{N} is not default @code{INTEGER}
195 to work instead of crashing @code{g77}.
200 The @samp{-ax} option is now obeyed when compiling Fortran programs.
201 (It is passed to the @file{f771} driver.)
205 The new @samp{-fsubscript-check} option
206 causes @code{g77} to compile run-time bounds checks
207 of array subscripts, as well as of substring start and end points.
209 The current implementation uses the @code{libf2c}
210 library routine @code{s_rnge} to print the diagnostic.
211 Since a future version of @code{g77} might use a different implementation,
212 use the new @samp{-ff2c-subscript-check} option
213 if your application requires use of @code{s_rnge} or a compile-time diagnostic.
216 Source file names with the suffixes @samp{.FOR} and @samp{.FPP}
217 now are recognized by @code{g77}
218 as if they ended in @samp{.for} and @samp{.fpp}, respectively.
221 @code{g77} now warns about a reference to an intrinsic
222 that has an interface that is not Year 2000 (Y2K) compliant.
223 Also, the @code{libg2c} has been changed to increase the likelihood
224 of catching references to the implementations of these intrinsics
225 using the @code{EXTERNAL} mechanism
226 (which would avoid the new warnings).
229 @xref{Year 2000 (Y2K) Problems}, for more information.
234 @code{g77} now warns about a reference to a function
235 when the corresponding @emph{subsequent} function program unit
236 disagrees with the reference concerning the type of the function.
240 @samp{-fno-emulate-complex} is now the default option.
241 This should result in improved performance
242 of code that uses the @code{COMPLEX} data type.
245 @cindex double-precision performance
246 @cindex -malign-double
248 The @samp{-malign-double} option
249 now reliably aligns @emph{all} double-precision variables and arrays
250 on Intel x86 targets.
254 Even without the @samp{-malign-double} option,
255 @code{g77} reliably aligns local double-precision variables
256 that are not in @code{EQUIVALENCE} areas
257 and not @code{SAVE}'d.
262 A substantial portion of the @code{g77} front end's code-generation component
264 It now generates code using facilities more robustly supported
265 by the @code{gcc} back end.
266 One effect of this rewrite is that some codes no longer produce
267 a spurious ``label @var{lab} used before containing binding contour''
272 Support for the @samp{-fugly} option has been removed.
276 Improve documentation and indexing,
277 including information on Year 2000 (Y2K) compliance,
278 and providing more information on internals of the front end.
283 Upgrade to @code{libf2c} as of 1999-05-03.
287 @heading In 0.5.24 and @code{egcs} 1.1.2 (versus 0.5.23 and 1.1.1):
291 Fix the @code{IDate} intrinsic (VXT) (in @code{libg2c})
292 so the returned year is in the documented, non-Y2K-compliant range
294 instead of being returned as 100 in the year 2000.
297 @xref{IDate Intrinsic (VXT)},
298 for more information.
302 Fix the @code{Date_and_Time} intrinsic (in @code{libg2c})
303 to return the milliseconds value properly
307 Fix the @code{LStat} intrinsic (in @code{libg2c})
308 to return device-ID information properly
312 Improve documentation.
316 @heading In 0.5.24 and @code{egcs} 1.1.1 (versus 0.5.23 and 1.1):
320 Fix @code{libg2c} so it performs an implicit @code{ENDFILE} operation
322 whenever a @code{REWIND} is done.
324 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
325 @code{g77}'s version of @code{libf2c}.)
328 Fix @code{libg2c} so it no longer crashes with a spurious diagnostic
329 upon doing any I/O following a direct formatted write.
331 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
332 @code{g77}'s version of @code{libf2c}.)
335 Fix @code{g77} so it no longer crashes compiling references
336 to the @code{Rand} intrinsic on some systems.
339 Fix @code{g77} portion of installation process so it works
340 better on some systems
341 (those with shells requiring @samp{else true} clauses
342 on @code{if} constructs
343 for the completion code to be set properly).
347 @heading In @code{egcs} 1.1 (versus 0.5.24):
351 Fix @code{g77} crash compiling code
352 containing the construct @samp{CMPLX(0.)} or similar.
358 (or apparently infinite run-time)
359 when compiling certain complicated expressions
360 involving @code{COMPLEX} arithmetic
361 (especially multiplication).
366 @cindex stack, 387 coprocessor
370 Fix a code-generation bug that afflicted
371 Intel x86 targets when @samp{-O2} was specified
372 compiling, for example, an old version of
373 the @code{DNRM2} routine.
375 The x87 coprocessor stack was being
376 mismanaged in cases involving assigned @code{GOTO}
381 @cindex double-precision performance
382 @cindex -malign-double
384 Align static double-precision variables and arrays
386 regardless of whether @samp{-malign-double} is specified.
388 Generally, this affects only local variables and arrays
389 having the @code{SAVE} attribute
390 or given initial values via @code{DATA}.
393 @c 1998-09-01: egcs-1.1 released.
394 @heading In @code{egcs} 1.1 (versus @code{egcs} 1.0.3):
398 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
399 that wrote one byte beyond the end of its @code{CHARACTER}
401 and in the @code{libU77} intrinsics
402 @code{GMTime} and @code{LTime}
403 that overwrote their arguments.
408 Assumed arrays with negative bounds
409 (such as @samp{REAL A(-1:*)})
410 no longer elicit spurious diagnostics from @code{g77},
411 even on systems with pointers having
412 different sizes than integers.
414 This bug is not known to have existed in any
415 recent version of @code{gcc}.
416 It was introduced in an early release of @code{egcs}.
421 Valid combinations of @code{EXTERNAL},
422 passing that external as a dummy argument
423 without explicitly giving it a type,
424 and, in a subsequent program unit,
425 referencing that external as
426 an external function with a different type
427 no longer crash @code{g77}.
432 @code{CASE DEFAULT} no longer crashes @code{g77}.
437 The @samp{-Wunused} option no longer issues a spurious
438 warning about the ``master'' procedure generated by
439 @code{g77} for procedures containing @code{ENTRY} statements.
443 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
444 compile-time constant @code{INTEGER} expression.
447 Fix @code{g77} @samp{-g} option so procedures that
448 use @code{ENTRY} can be stepped through, line by line,
452 Allow any @code{REAL} argument to intrinsics
453 @code{Second} and @code{CPU_Time}.
456 Use @code{tempnam}, if available, to open scratch files
457 (as in @samp{OPEN(STATUS='SCRATCH')})
458 so that the @code{TMPDIR} environment variable,
462 @code{g77}'s version of @code{libf2c} separates out
463 the setting of global state
464 (such as command-line arguments and signal handling)
465 from @file{main.o} into distinct, new library
468 This should make it easier to write portable applications
469 that have their own (non-Fortran) @code{main()} routine
470 properly set up the @code{libf2c} environment, even
471 when @code{libf2c} (now @code{libg2c}) is a shared library.
475 @code{g77} no longer installs the @file{f77} command
476 and @file{f77.1} man page
477 in the @file{/usr} or @file{/usr/local} heirarchy,
478 even if the @file{f77-install-ok} file exists
479 in the source or build directory.
480 See the installation documentation for more information.
485 @code{g77} no longer installs the @file{libf2c.a} library
486 and @file{f2c.h} include file
487 in the @file{/usr} or @file{/usr/local} heirarchy,
488 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
489 in the source or build directory.
490 See the installation documentation for more information.
495 The @file{libf2c.a} library produced by @code{g77} has been
496 renamed to @file{libg2c.a}.
497 It is installed only in the @code{gcc} ``private''
498 directory heirarchy, @file{gcc-lib}.
499 This allows system administrators and users to choose which
500 version of the @code{libf2c} library from @code{netlib} they
501 wish to use on a case-by-case basis.
502 See the installation documentation for more information.
507 The @file{f2c.h} include (header) file produced by @code{g77}
508 has been renamed to @file{g2c.h}.
509 It is installed only in the @code{gcc} ``private''
510 directory heirarchy, @file{gcc-lib}.
511 This allows system administrators and users to choose which
512 version of the include file from @code{netlib} they
513 wish to use on a case-by-case basis.
514 See the installation documentation for more information.
518 The @code{g77} command now expects the run-time library
519 to be named @code{libg2c.a} instead of @code{libf2c.a},
520 to ensure that a version other than the one built and
521 installed as part of the same @code{g77} version is picked up.
525 During the configuration and build process,
526 @code{g77} creates subdirectories it needs only as it
528 Other cleaning up of the configuration and build process
529 has been performed as well.
534 @code{install-info} now used to update the directory of
535 Info documentation to contain an entry for @code{g77}
536 (during installation).
540 Some diagnostics have been changed from warnings to errors,
541 to prevent inadvertent use of the resulting, probably buggy,
543 These mostly include diagnostics about use of unsupported features
544 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
545 @code{WRITE} statements,
546 and about truncations of various sorts of constants.
550 Improve compilation of @code{FORMAT} expressions so that
551 a null byte is appended to the last operand if it
553 This provides a cleaner run-time diagnostic as provided
554 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
559 Improve documentation and indexing.
564 The upgrade to @code{libf2c} as of 1998-06-18
565 should fix a variety of problems, including
566 those involving some uses of the @code{T} format
567 specifier, and perhaps some build (porting) problems
572 @heading In 0.5.24 and @code{egcs} 1.1 (versus 0.5.23):
576 @code{g77} no longer produces incorrect code
578 for @code{EQUIVALENCE} and @code{COMMON}
579 aggregates that, due to ``unnatural'' ordering of members
580 vis-a-vis their types, require initial padding.
585 @code{g77} no longer crashes when compiling code
586 containing specification statements such as
587 @samp{INTEGER(KIND=7) PTR}.
592 @code{g77} no longer crashes when compiling code
593 such as @samp{J = SIGNAL(1, 2)}.
597 @code{g77} now treats @samp{%LOC(@var{expr})} and
598 @samp{LOC(@var{expr})} as ``ordinary'' expressions
599 when they are used as arguments in procedure calls.
600 This change applies only to global (filewide) analysis,
601 making it consistent with
602 how @code{g77} actually generates code
605 Previously, @code{g77} treated these expressions
606 as denoting special ``pointer'' arguments
607 for the purposes of filewide analysis.
610 The @code{g77} driver now ensures that @samp{-lg2c}
611 is specified in the link phase prior to any
612 occurrence of @samp{-lm}.
613 This prevents accidentally linking to a routine
614 in the SunOS4 @samp{-lm} library
615 when the generated code wants to link to the one
616 in @code{libf2c} (@code{libg2c}).
619 @code{g77} emits more debugging information when
622 This new information allows, for example,
623 @kbd{which __g77_length_a} to be used in @code{gdb}
624 to determine the type of the phantom length argument
625 supplied with @code{CHARACTER} variables.
627 This information pertains to internally-generated
628 type, variable, and other information,
629 not to the longstanding deficiencies vis-a-vis
630 @code{COMMON} and @code{EQUIVALENCE}.
633 The F90 @code{Date_and_Time} intrinsic now is
637 The F90 @code{System_Clock} intrinsic allows
638 the optional arguments (except for the @code{Count}
639 argument) to be omitted.
643 Upgrade to @code{libf2c} as of 1998-06-18.
648 Improve documentation and indexing.
653 @heading In previous versions:
655 Information on previous versions is not provided
656 in this @file{@value{path-g77}/NEWS} file,
658 See @file{@value{path-g77}/news.texi},
659 or any of its other derivations
660 (Info, HTML, dvi forms)
661 for such information.
665 @c 1998-05-20: 0.5.23 released.
666 @heading In 0.5.23 (versus 0.5.22):
669 This release contains several regressions against
670 version 0.5.22 of @code{g77}, due to using the
671 ``vanilla'' @code{gcc} back end instead of patching
672 it to fix a few bugs and improve performance in a
675 Features that have been dropped from this version
676 of @code{g77} due to their being implemented
677 via @code{g77}-specific patches to the @code{gcc}
678 back end in previous releases include:
682 Support for @code{__restrict__} keyword,
683 the options @samp{-fargument-alias}, @samp{-fargument-noalias},
684 and @samp{-fargument-noalias-global},
685 and the corresponding alias-analysis code.
687 (@code{egcs} has the alias-analysis
688 code, but not the @code{__restrict__} keyword.
689 @code{egcs} @code{g77} users benefit from the alias-analysis
690 code despite the lack of the @code{__restrict__} keyword,
691 which is a C-language construct.)
694 Support for the GNU compiler options
695 @samp{-fmove-all-movables},
696 @samp{-freduce-all-givs},
697 and @samp{-frerun-loop-opt}.
699 (@code{egcs} supports these options.
700 @code{g77} users of @code{egcs} benefit from them even if
701 they are not explicitly specified,
702 because the defaults are optimized for @code{g77} users.)
705 Support for the @samp{-W} option warning about
706 integer division by zero.
709 The Intel x86-specific option @samp{-malign-double}
710 applying to stack-allocated data
711 as well as statically-allocate data.
715 Note that the @file{gcc/f/gbe/} subdirectory has been removed
716 from this distribution as a result of @code{g77} no longer
717 including patches for the @code{gcc} back end.
722 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
723 that wrote one byte beyond the end of its @code{CHARACTER}
725 and in the @code{libU77} intrinsics
726 @code{GMTime} and @code{LTime}
727 that overwrote their arguments.
731 Support @code{gcc} version 2.8,
732 and remove support for prior versions of @code{gcc}.
734 @cindex -@w{}-driver option
735 @cindex @code{g77} options, -@w{}-driver
736 @cindex options, -@w{}-driver
738 Remove support for the @samp{--driver} option,
739 as @code{g77} now does all the driving,
740 just like @code{gcc}.
744 @code{CASE DEFAULT} no longer crashes @code{g77}.
749 Valid combinations of @code{EXTERNAL},
750 passing that external as a dummy argument
751 without explicitly giving it a type,
752 and, in a subsequent program unit,
753 referencing that external as
754 an external function with a different type
755 no longer crash @code{g77}.
760 @code{g77} no longer installs the @file{f77} command
761 and @file{f77.1} man page
762 in the @file{/usr} or @file{/usr/local} heirarchy,
763 even if the @file{f77-install-ok} file exists
764 in the source or build directory.
765 See the installation documentation for more information.
770 @code{g77} no longer installs the @file{libf2c.a} library
771 and @file{f2c.h} include file
772 in the @file{/usr} or @file{/usr/local} heirarchy,
773 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
774 in the source or build directory.
775 See the installation documentation for more information.
780 The @file{libf2c.a} library produced by @code{g77} has been
781 renamed to @file{libg2c.a}.
782 It is installed only in the @code{gcc} ``private''
783 directory heirarchy, @file{gcc-lib}.
784 This allows system administrators and users to choose which
785 version of the @code{libf2c} library from @code{netlib} they
786 wish to use on a case-by-case basis.
787 See the installation documentation for more information.
792 The @file{f2c.h} include (header) file produced by @code{g77}
793 has been renamed to @file{g2c.h}.
794 It is installed only in the @code{gcc} ``private''
795 directory heirarchy, @file{gcc-lib}.
796 This allows system administrators and users to choose which
797 version of the include file from @code{netlib} they
798 wish to use on a case-by-case basis.
799 See the installation documentation for more information.
803 The @code{g77} command now expects the run-time library
804 to be named @code{libg2c.a} instead of @code{libf2c.a},
805 to ensure that a version other than the one built and
806 installed as part of the same @code{g77} version is picked up.
810 The @samp{-Wunused} option no longer issues a spurious
811 warning about the ``master'' procedure generated by
812 @code{g77} for procedures containing @code{ENTRY} statements.
816 @code{g77}'s version of @code{libf2c} separates out
817 the setting of global state
818 (such as command-line arguments and signal handling)
819 from @file{main.o} into distinct, new library
822 This should make it easier to write portable applications
823 that have their own (non-Fortran) @code{main()} routine
824 properly set up the @code{libf2c} environment, even
825 when @code{libf2c} (now @code{libg2c}) is a shared library.
829 During the configuration and build process,
830 @code{g77} creates subdirectories it needs only as it
831 needs them, thus avoiding unnecessary creation of, for example,
832 @file{stage1/f/runtime} when doing a non-bootstrap build.
833 Other cleaning up of the configuration and build process
834 has been performed as well.
839 @code{install-info} now used to update the directory of
840 Info documentation to contain an entry for @code{g77}
841 (during installation).
845 Some diagnostics have been changed from warnings to errors,
846 to prevent inadvertent use of the resulting, probably buggy,
848 These mostly include diagnostics about use of unsupported features
849 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
850 @code{WRITE} statements,
851 and about truncations of various sorts of constants.
855 Improve documentation and indexing.
860 Upgrade to @code{libf2c} as of 1998-04-20.
862 This should fix a variety of problems, including
863 those involving some uses of the @code{T} format
864 specifier, and perhaps some build (porting) problems
869 @c 1998-03-16: 0.5.22 released.
870 @heading In 0.5.22 (versus 0.5.21):
874 Fix code generation for iterative @code{DO} loops that
875 have one or more references to the iteration variable,
876 or to aliases of it, in their control expressions.
877 For example, @samp{DO 10 J=2,J} now is compiled correctly.
882 @cindex stack, 387 coprocessor
886 Fix a code-generation bug that afflicted
887 Intel x86 targets when @samp{-O2} was specified
888 compiling, for example, an old version of
889 the @code{DNRM2} routine.
891 The x87 coprocessor stack was being
892 mismanaged in cases involving assigned @code{GOTO}
898 Fix @code{DTime} intrinsic so as not to truncate
899 results to integer values (on some systems).
903 Fix @code{Signal} intrinsic so it offers portable
904 support for 64-bit systems (such as Digital Alphas
909 Fix run-time crash involving @code{NAMELIST} on 64-bit
910 machines such as Alphas.
915 Fix @code{g77} version of @code{libf2c} so it no longer
916 produces a spurious @samp{I/O recursion} diagnostic at run time
917 when an I/O operation (such as @samp{READ *,I}) is interrupted
918 in a manner that causes the program to be terminated
919 via the @code{f_exit} routine (such as via @kbd{C-c}).
924 Fix @code{g77} crash triggered by @code{CASE} statement with
925 an omitted lower or upper bound.
930 Fix @code{g77} crash compiling references to @code{CPU_Time}
937 (or apparently infinite run-time)
938 when compiling certain complicated expressions
939 involving @code{COMPLEX} arithmetic
940 (especially multiplication).
945 Fix @code{g77} crash on statements such as
946 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
947 @samp{Z} is @code{DOUBLE COMPLEX}.
952 Fix a @code{g++} crash.
956 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
957 compile-time constant @code{INTEGER} expression.
960 Fix @code{g77} @samp{-g} option so procedures that
961 use @code{ENTRY} can be stepped through, line by line,
966 Fix a profiling-related bug in @code{gcc} back end for
967 Intel x86 architecture.
971 Allow any @code{REAL} argument to intrinsics
972 @code{Second} and @code{CPU_Time}.
975 Allow any numeric argument to intrinsics
976 @code{Int2} and @code{Int8}.
979 Use @code{tempnam}, if available, to open scratch files
980 (as in @samp{OPEN(STATUS='SCRATCH')})
981 so that the @code{TMPDIR} environment variable,
985 Rename the @code{gcc} keyword @code{restrict} to
986 @code{__restrict__}, to avoid rejecting valid, existing,
988 Support for @code{restrict} is now more like support
993 Fix @samp{-fpedantic} to not reject procedure invocations
994 such as @samp{I=J()} and @samp{CALL FOO()}.
998 Fix @samp{-fugly-comma} to affect invocations of
999 only external procedures.
1000 Restore rejection of gratuitous trailing omitted
1001 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1004 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
1005 @samp{-fbadu77-intrinsics-*} options.
1007 @ifclear USERVISONLY
1009 Improve diagnostic messages from @code{libf2c}
1010 so it is more likely that the printing of the
1011 active format string is limited to the string,
1012 with no trailing garbage being printed.
1014 (Unlike @code{f2c}, @code{g77} did not append
1015 a null byte to its compiled form of every
1016 format string specified via a @code{FORMAT} statement.
1017 However, @code{f2c} would exhibit the problem
1018 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1019 by printing @samp{(I)garbage} as the format string.)
1022 @ifclear USERVISONLY
1024 Improve compilation of @code{FORMAT} expressions so that
1025 a null byte is appended to the last operand if it
1027 This provides a cleaner run-time diagnostic as provided
1028 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
1031 @ifclear USERVISONLY
1033 Fix various crashes involving code with diagnosed errors.
1036 @ifclear USERVISONLY
1038 Fix cross-compilation bug when configuring @code{libf2c}.
1041 @ifclear USERVISONLY
1043 Improve diagnostics.
1046 @ifclear USERVISONLY
1048 Improve documentation and indexing.
1051 @ifclear USERVISONLY
1053 Upgrade to @code{libf2c} as of 1997-09-23.
1054 This fixes a formatted-I/O bug that afflicted
1055 64-bit systems with 32-bit integers
1056 (such as Digital Alpha running GNU/Linux).
1060 @c 1998-03-15: egcs-1.0.2 released.
1061 @heading In @code{egcs} 1.0.2 (versus @code{egcs} 1.0.1):
1063 @ifclear USERVISONLY
1065 Fix @code{g77} crash triggered by @code{CASE} statement with
1066 an omitted lower or upper bound.
1069 @ifclear USERVISONLY
1071 Fix @code{g77} crash on statements such as
1072 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1073 @samp{Z} is @code{DOUBLE COMPLEX}.
1076 @ifclear USERVISONLY
1078 @cindex support, ELF
1079 @cindex -fPIC option
1080 @cindex options, -fPIC
1082 Fix @samp{-fPIC} (such as compiling for ELF targets)
1083 on the Intel x86 architecture target
1084 so invalid assembler code is no longer produced.
1087 @ifclear USERVISONLY
1089 Fix @samp{-fpedantic} to not reject procedure invocations
1090 such as @samp{I=J()} and @samp{CALL FOO()}.
1093 @ifclear USERVISONLY
1095 Fix @samp{-fugly-comma} to affect invocations of
1096 only external procedures.
1097 Restore rejection of gratuitous trailing omitted
1098 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1102 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
1103 @samp{-fbadu77-intrinsics-*} options.
1106 @c 1998-01-02: egcs-1.0.1 released.
1107 @heading In @code{egcs} 1.0.1 (versus @code{egcs} 1.0):
1108 @ifclear USERVISONLY
1111 Fix run-time crash involving @code{NAMELIST} on 64-bit
1112 machines such as Alphas.
1116 @c 1997-12-03: egcs-1.0 released.
1117 @heading In @code{egcs} 1.0 (versus 0.5.21):
1120 Version 1.0 of @code{egcs}
1121 contains several regressions against
1122 version 0.5.21 of @code{g77},
1124 ``vanilla'' @code{gcc} back end instead of patching
1125 it to fix a few bugs and improve performance in a
1128 Features that have been dropped from this version
1129 of @code{g77} due to their being implemented
1130 via @code{g77}-specific patches to the @code{gcc}
1131 back end in previous releases include:
1135 Support for the C-language @code{restrict} keyword.
1138 Support for the @samp{-W} option warning about
1139 integer division by zero.
1142 The Intel x86-specific option @samp{-malign-double}
1143 applying to stack-allocated data
1144 as well as statically-allocate data.
1147 @ifclear USERVISONLY
1148 Note that the @file{gcc/f/gbe/} subdirectory has been removed
1149 from this distribution as a result of @code{g77}
1150 being fully integrated with
1151 the @code{egcs} variant of the @code{gcc} back end.
1154 @ifclear USERVISONLY
1156 Fix code generation for iterative @code{DO} loops that
1157 have one or more references to the iteration variable,
1158 or to aliases of it, in their control expressions.
1159 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1162 @ifclear USERVISONLY
1164 Fix @code{DTime} intrinsic so as not to truncate
1165 results to integer values (on some systems).
1168 @ifclear USERVISONLY
1170 @c Toon Moene discovered these.
1171 Some Fortran code, miscompiled
1172 by @code{g77} built on @code{gcc} version 2.8.1
1173 on m68k-next-nextstep3 configurations
1174 when using the @samp{-O2} option,
1175 is now compiled correctly.
1176 It is believed that a C function known to miscompile
1177 on that configuration
1178 when using the @samp{-O2 -funroll-loops} options
1179 also is now compiled correctly.
1182 @ifclear USERVISONLY
1184 Remove support for non-@code{egcs} versions of @code{gcc}.
1187 @cindex -@w{}-driver option
1188 @cindex @code{g77} options, -@w{}-driver
1189 @cindex options, -@w{}-driver
1191 Remove support for the @samp{--driver} option,
1192 as @code{g77} now does all the driving,
1193 just like @code{gcc}.
1196 Allow any numeric argument to intrinsics
1197 @code{Int2} and @code{Int8}.
1199 @ifclear USERVISONLY
1201 Improve diagnostic messages from @code{libf2c}
1202 so it is more likely that the printing of the
1203 active format string is limited to the string,
1204 with no trailing garbage being printed.
1206 (Unlike @code{f2c}, @code{g77} did not append
1207 a null byte to its compiled form of every
1208 format string specified via a @code{FORMAT} statement.
1209 However, @code{f2c} would exhibit the problem
1210 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1211 by printing @samp{(I)garbage} as the format string.)
1214 @ifclear USERVISONLY
1216 Upgrade to @code{libf2c} as of 1997-09-23.
1217 This fixes a formatted-I/O bug that afflicted
1218 64-bit systems with 32-bit integers
1219 (such as Digital Alpha running GNU/Linux).
1223 @c 1997-09-09: 0.5.21 released.
1226 @ifclear USERVISONLY
1228 Fix a code-generation bug introduced by 0.5.20
1229 caused by loop unrolling (by specifying
1230 @samp{-funroll-loops} or similar).
1231 This bug afflicted all code compiled by
1232 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1233 Fortran, and so on).
1236 @ifclear USERVISONLY
1238 Fix a code-generation bug manifested when
1239 combining local @code{EQUIVALENCE} with a
1240 @code{DATA} statement that follows
1241 the first executable statement (or is
1242 treated as an executable-context statement
1243 as a result of using the @samp{-fpedantic}
1247 @ifclear USERVISONLY
1249 Fix a compiler crash that occured when an
1250 integer division by a constant zero is detected.
1251 Instead, when the @samp{-W} option is specified,
1252 the @code{gcc} back end issues a warning about such a case.
1253 This bug afflicted all code compiled by
1254 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1255 Fortran, and so on).
1259 When the @samp{-W} option is specified, @code{gcc}, @code{g77},
1260 and other GNU compilers that incorporate the @code{gcc}
1261 back end as modified by @code{g77}, issue
1262 a warning about integer division by constant zero.
1265 @ifclear USERVISONLY
1267 Fix a compiler crash that occurred in some cases
1268 of procedure inlining.
1269 (Such cases became more frequent in 0.5.20.)
1272 @ifclear USERVISONLY
1274 Fix a compiler crash resulting from using @code{DATA}
1275 or similar to initialize a @code{COMPLEX} variable or
1279 @ifclear USERVISONLY
1281 Fix compiler crashes involving use of @code{AND}, @code{OR},
1282 or @code{XOR} intrinsics.
1285 @ifclear USERVISONLY
1287 Fix compiler bug triggered when using a @code{COMMON}
1288 or @code{EQUIVALENCE} variable
1289 as the target of an @code{ASSIGN}
1290 or assigned-@code{GOTO} statement.
1293 @ifclear USERVISONLY
1295 Fix compiler crashes due to using the name of a some
1296 non-standard intrinsics (such as @code{FTell} or
1297 @code{FPutC}) as such and as the name of a procedure
1299 Such dual use of a name in a program is allowed by
1303 @c @code{g77}'s version of @code{libf2c} has been modified
1304 @c so that the external names of library's procedures do not
1305 @c conflict with names used for Fortran procedures compiled
1307 @c An additional layer of jacket procedures has been added
1308 @c to @code{libf2c} to map the old names to the new names,
1309 @c for automatic use by programs that interface to the
1310 @c library procedures via the external-procedure mechanism.
1312 @c For example, the intrinsic @code{FPUTC} previously was
1313 @c implemented by @code{g77} as a call to the @code{libf2c}
1314 @c routine @code{fputc_}.
1315 @c This would conflict with a Fortran procedure named @code{FPUTC}
1316 @c (using default compiler options), and this conflict
1317 @c would cause a crash under certain circumstances.
1319 @c Now, the intrinsic @code{FPUTC} calls @code{G77_fputc_0},
1320 @c which does not conflict with the @code{fputc_} external
1321 @c that implements a Fortran procedure named @code{FPUTC}.
1323 @c Programs that refer to @code{FPUTC} as an external procedure
1324 @c without supplying their own implementation will link to
1325 @c the new @code{libf2c} routine @code{fputc_}, which is
1326 @c simply a jacket routine that calls @code{G77_fputc_0}.
1328 @ifclear USERVISONLY
1330 Place automatic arrays on the stack, even if
1331 @code{SAVE} or the @samp{-fno-automatic} option
1333 This avoids a compiler crash in some cases.
1336 @ifclear USERVISONLY
1338 The @samp{-malign-double} option now reliably aligns
1339 @code{DOUBLE PRECISION} optimally on Pentium and
1340 Pentium Pro architectures (586 and 686 in @code{gcc}).
1344 New option @samp{-Wno-globals} disables warnings
1345 about ``suspicious'' use of a name both as a global
1346 name and as the implicit name of an intrinsic, and
1347 warnings about disagreements over the number or natures of
1348 arguments passed to global procedures, or the
1349 natures of the procedures themselves.
1351 The default is to issue such warnings, which are
1352 new as of this version of @code{g77}.
1355 New option @samp{-fno-globals} disables diagnostics
1356 about potentially fatal disagreements
1357 analysis problems, such as disagreements over the
1358 number or natures of arguments passed to global
1359 procedures, or the natures of those procedures themselves.
1361 The default is to issue such diagnostics and flag
1362 the compilation as unsuccessful.
1363 With this option, the diagnostics are issued as
1364 warnings, or, if @samp{-Wno-globals} is specified,
1365 are not issued at all.
1367 This option also disables inlining of global procedures,
1368 to avoid compiler crashes resulting from coding errors
1369 that these diagnostics normally would identify.
1371 @ifclear USERVISONLY
1373 Diagnose cases where a reference to a procedure
1374 disagrees with the type of that procedure, or
1375 where disagreements about the number or nature
1377 This avoids a compiler crash.
1380 @ifclear USERVISONLY
1382 Fix parsing bug whereby @code{g77} rejected a
1383 second initialization specification immediately
1384 following the first's closing @samp{/} without
1385 an intervening comma in a @code{DATA} statement,
1386 and the second specification was an implied-DO list.
1389 @ifclear USERVISONLY
1391 Improve performance of the @code{gcc} back end so
1392 certain complicated expressions involving @code{COMPLEX}
1393 arithmetic (especially multiplication) don't appear to
1394 take forever to compile.
1397 @ifclear USERVISONLY
1399 Fix a couple of profiling-related bugs in @code{gcc}
1403 @ifclear USERVISONLY
1405 Integrate GNU Ada's (GNAT's) changes to the back end,
1406 which consist almost entirely of bug fixes.
1407 These fixes are circa version 3.10p of GNAT.
1410 @ifclear USERVISONLY
1412 Include some other @code{gcc} fixes that seem useful in
1413 @code{g77}'s version of @code{gcc}.
1414 (See @file{gcc/ChangeLog} for details---compare it
1415 to that file in the vanilla @code{gcc-2.7.2.3.tar.gz}
1420 Fix @code{libU77} routines that accept file and other names
1421 to strip trailing blanks from them, for consistency
1422 with other implementations.
1423 Blanks may be forcibly appended to such names by
1424 appending a single null character (@samp{CHAR(0)})
1425 to the significant trailing blanks.
1428 Fix @code{CHMOD} intrinsic to work with file names
1429 that have embedded blanks, commas, and so on.
1432 Fix @code{SIGNAL} intrinsic so it accepts an
1433 optional third @code{Status} argument.
1435 @ifclear USERVISONLY
1437 Fix @code{IDATE()} intrinsic subroutine (VXT form)
1438 so it accepts arguments in the correct order.
1439 Documentation fixed accordingly, and for
1440 @code{GMTIME()} and @code{LTIME()} as well.
1444 Make many changes to @code{libU77} intrinsics to
1445 support existing code more directly.
1447 Such changes include allowing both subroutine and
1448 function forms of many routines, changing @code{MCLOCK()}
1449 and @code{TIME()} to return @code{INTEGER(KIND=1)} values,
1450 introducing @code{MCLOCK8()} and @code{TIME8()} to
1451 return @code{INTEGER(KIND=2)} values,
1452 and placing functions that are intended to perform
1453 side effects in a new intrinsic group, @code{badu77}.
1455 @ifclear USERVISONLY
1457 Improve @code{libU77} so it is more portable.
1461 Add options @samp{-fbadu77-intrinsics-delete},
1462 @samp{-fbadu77-intrinsics-hide}, and so on.
1464 @ifclear USERVISONLY
1466 Fix crashes involving diagnosed or invalid code.
1469 @ifclear USERVISONLY
1471 @code{g77} and @code{gcc} now do a somewhat better
1472 job detecting and diagnosing arrays that are too
1473 large to handle before these cause diagnostics
1474 during the assembler or linker phase, a compiler
1475 crash, or generation of incorrect code.
1478 @ifclear USERVISONLY
1480 Make some fixes to alias analysis code.
1483 @ifclear USERVISONLY
1485 Add support for @code{restrict} keyword in @code{gcc}
1489 @ifclear USERVISONLY
1491 Support @code{gcc} version 2.7.2.3
1492 (modified by @code{g77} into version 2.7.2.3.f.1),
1494 support for prior versions of @code{gcc}.
1497 @ifclear USERVISONLY
1499 Incorporate GNAT's patches to the @code{gcc} back
1500 end into @code{g77}'s, so GNAT users do not need
1501 to apply GNAT's patches to build both GNAT and @code{g77}
1502 from the same source tree.
1505 @ifclear USERVISONLY
1507 Modify @code{make} rules and related code so that
1508 generation of Info documentation doesn't require
1509 compilation using @code{gcc}.
1510 Now, any ANSI C compiler should be adequate to
1511 produce the @code{g77} documentation (in particular,
1512 the tables of intrinsics) from scratch.
1516 Add @code{INT2} and @code{INT8} intrinsics.
1519 Add @code{CPU_TIME} intrinsic.
1522 Add @code{ALARM} intrinsic.
1525 @code{CTIME} intrinsic now accepts any @code{INTEGER}
1526 argument, not just @code{INTEGER(KIND=2)}.
1528 @ifclear USERVISONLY
1530 Warn when explicit type declaration disagrees with
1531 the type of an intrinsic invocation.
1534 @ifclear USERVISONLY
1536 Support @samp{*f771} entry in @code{gcc} @file{specs} file.
1539 @ifclear USERVISONLY
1541 Fix typo in @code{make} rule @code{g77-cross}, used only for
1545 @ifclear USERVISONLY
1547 Fix @code{libf2c} build procedure to re-archive library
1548 if previous attempt to archive was interrupted.
1551 @ifclear USERVISONLY
1553 Change @code{gcc} to unroll loops only during the last
1554 invocation (of as many as two invocations) of loop
1558 @ifclear USERVISONLY
1560 Improve handling of @samp{-fno-f2c} so that code that
1561 attempts to pass an intrinsic as an actual argument,
1562 such as @samp{CALL FOO(ABS)}, is rejected due to the fact
1563 that the run-time-library routine is, effectively,
1564 compiled with @samp{-ff2c} in effect.
1567 @ifclear USERVISONLY
1569 Fix @code{g77} driver to recognize @samp{-fsyntax-only}
1570 as an option that inhibits linking, just like @samp{-c} or
1571 @samp{-S}, and to recognize and properly handle the
1572 @samp{-nostdlib}, @samp{-M}, @samp{-MM}, @samp{-nodefaultlibs},
1573 and @samp{-Xlinker} options.
1576 @ifclear USERVISONLY
1578 Upgrade to @code{libf2c} as of 1997-08-16.
1581 @ifclear USERVISONLY
1583 Modify @code{libf2c} to consistently and clearly diagnose
1584 recursive I/O (at run time).
1588 @code{g77} driver now prints version information (such as produced
1589 by @kbd{g77 -v}) to @code{stderr} instead of @code{stdout}.
1592 The @samp{.r} suffix now designates a Ratfor source file,
1593 to be preprocessed via the @code{ratfor} command, available
1596 @ifclear USERVISONLY
1598 Fix some aspects of how @code{gcc} determines what kind of
1599 system is being configured and what kinds are supported.
1600 For example, GNU Linux/Alpha ELF systems now are directly
1604 @ifclear USERVISONLY
1606 Improve diagnostics.
1609 @ifclear USERVISONLY
1611 Improve documentation and indexing.
1614 @ifclear USERVISONLY
1616 Include all pertinent files for @code{libf2c} that come
1617 from @code{netlib.bell-labs.com}; give any such files
1618 that aren't quite accurate in @code{g77}'s version of
1619 @code{libf2c} the suffix @samp{.netlib}.
1622 @ifclear USERVISONLY
1624 Reserve @code{INTEGER(KIND=0)} for future use.
1628 @c 1997-02-28: 0.5.20 released.
1632 The @samp{-fno-typeless-boz} option is now the default.
1634 This option specifies that non-decimal-radix
1635 constants using the prefixed-radix form (such as @samp{Z'1234'})
1636 are to be interpreted as @code{INTEGER(KIND=1)} constants.
1637 Specify @samp{-ftypeless-boz} to cause such
1638 constants to be interpreted as typeless.
1640 (Version 0.5.19 introduced @samp{-fno-typeless-boz} and
1644 @xref{Fortran Dialect Options,,Options Controlling Fortran Dialect},
1645 for information on the @samp{-ftypeless-boz} option.
1649 Options @samp{-ff90-intrinsics-enable} and
1650 @samp{-fvxt-intrinsics-enable} now are the
1653 Some programs might use names that clash with
1654 intrinsic names defined (and now enabled) by these
1655 options or by the new @code{libU77} intrinsics.
1656 Users of such programs might need to compile them
1657 differently (using, for example, @samp{-ff90-intrinsics-disable})
1658 or, better yet, insert appropriate @code{EXTERNAL}
1659 statements specifying that these names are not intended
1660 to be names of intrinsics.
1663 The @code{ALWAYS_FLUSH} macro is no longer defined when
1664 building @code{libf2c}, which should result in improved
1665 I/O performance, especially over NFS.
1667 @emph{Note:} If you have code that depends on the behavior
1668 of @code{libf2c} when built with @code{ALWAYS_FLUSH} defined,
1669 you will have to modify @code{libf2c} accordingly before
1670 building it from this and future versions of @code{g77}.
1673 @xref{Output Assumed To Flush}, for more information.
1677 Dave Love's implementation of @code{libU77} has been
1678 added to the version of @code{libf2c} distributed with
1679 and built as part of @code{g77}.
1680 @code{g77} now knows about the routines in this library
1684 New option @samp{-fvxt} specifies that the
1685 source file is written in VXT Fortran, instead of GNU Fortran.
1688 @xref{VXT Fortran}, for more information on the constructs
1689 recognized when the @samp{-fvxt} option is specified.
1693 The @samp{-fvxt-not-f90} option has been deleted,
1694 along with its inverse, @samp{-ff90-not-vxt}.
1696 If you used one of these deleted options, you should
1697 re-read the pertinent documentation to determine which
1698 options, if any, are appropriate for compiling your
1699 code with this version of @code{g77}.
1702 @xref{Other Dialects}, for more information.
1706 The @samp{-fugly} option now issues a warning, as it
1707 likely will be removed in a future version.
1709 (Enabling all the @samp{-fugly-*} options is unlikely
1710 to be feasible, or sensible, in the future,
1711 so users should learn to specify only those
1712 @samp{-fugly-*} options they really need for a
1713 particular source file.)
1716 The @samp{-fugly-assumed} option, introduced in
1717 version 0.5.19, has been changed to
1718 better accommodate old and new code.
1721 @xref{Ugly Assumed-Size Arrays}, for more information.
1724 @ifclear USERVISONLY
1726 Make a number of fixes to the @code{g77} front end and
1727 the @code{gcc} back end to better support Alpha (AXP)
1729 This includes providing at least one bug-fix to the
1730 @code{gcc} back end for Alphas.
1734 Related to supporting Alpha (AXP) machines, the @code{LOC()}
1735 intrinsic and @code{%LOC()} construct now return
1736 values of @code{INTEGER(KIND=0)} type,
1737 as defined by the GNU Fortran language.
1739 This type is wide enough
1740 (holds the same number of bits)
1741 as the character-pointer type on the machine.
1743 On most machines, this won't make a difference,
1744 whereas, on Alphas and other systems with 64-bit pointers,
1745 the @code{INTEGER(KIND=0)} type is equivalent to @code{INTEGER(KIND=2)}
1746 (often referred to as @code{INTEGER*8})
1747 instead of the more common @code{INTEGER(KIND=1)}
1748 (often referred to as @code{INTEGER*4}).
1751 Emulate @code{COMPLEX} arithmetic in the @code{g77} front
1752 end, to avoid bugs in @code{complex} support in the
1753 @code{gcc} back end.
1754 New option @samp{-fno-emulate-complex}
1755 causes @code{g77} to revert the 0.5.19 behavior.
1757 @ifclear USERVISONLY
1759 Fix bug whereby @samp{REAL A(1)}, for example, caused
1760 a compiler crash if @samp{-fugly-assumed} was in effect
1761 and @var{A} was a local (automatic) array.
1762 That case is no longer affected by the new
1763 handling of @samp{-fugly-assumed}.
1766 @ifclear USERVISONLY
1768 Fix @code{g77} command driver so that @samp{g77 -o foo.f}
1769 no longer deletes @file{foo.f} before issuing other
1770 diagnostics, and so the @samp{-x} option is properly
1774 @ifclear USERVISONLY
1776 Enable inlining of subroutines and functions by the @code{gcc}
1778 This works as it does for @code{gcc} itself---program units
1779 may be inlined for invocations that follow them in the same
1780 program unit, as long as the appropriate compile-time
1781 options are specified.
1785 Dummy arguments are no longer assumed to potentially alias
1787 other dummy arguments or @code{COMMON} areas when any of
1788 these are defined (assigned to) by Fortran code.
1790 This can result in faster and/or smaller programs when
1791 compiling with optimization enabled, though on some
1792 systems this effect is observed only when @samp{-fforce-addr}
1795 New options @samp{-falias-check}, @samp{-fargument-alias},
1796 @samp{-fargument-noalias},
1797 and @samp{-fno-argument-noalias-global} control the
1798 way @code{g77} handles potential aliasing.
1801 @xref{Aliasing Assumed To Work}, for detailed information on why the
1802 new defaults might result in some programs no longer working the way they
1803 did when compiled by previous versions of @code{g77}.
1806 @ifclear USERVISONLY
1808 The @code{CONJG()} and @code{DCONJG()} intrinsics now
1809 are compiled in-line.
1812 @ifclear USERVISONLY
1814 The bug-fix for 0.5.19.1 has been re-done.
1815 The @code{g77} compiler has been changed back to
1816 assume @code{libf2c} has no aliasing problems in
1817 its implementations of the @code{COMPLEX} (and
1818 @code{DOUBLE COMPLEX}) intrinsics.
1819 The @code{libf2c} has been changed to have no such
1822 As a result, 0.5.20 is expected to offer improved performance
1823 over 0.5.19.1, perhaps as good as 0.5.19 in most
1824 or all cases, due to this change alone.
1826 @emph{Note:} This change requires version 0.5.20 of
1827 @code{libf2c}, at least, when linking code produced
1828 by any versions of @code{g77} other than 0.5.19.1.
1829 Use @samp{g77 -v} to determine the version numbers
1830 of the @code{libF77}, @code{libI77}, and @code{libU77}
1831 components of the @code{libf2c} library.
1832 (If these version numbers are not printed---in
1833 particular, if the linker complains about unresolved
1834 references to names like @samp{g77__fvers__}---that
1835 strongly suggests your installation has an obsolete
1836 version of @code{libf2c}.)
1840 New option @samp{-fugly-assign} specifies that the
1841 same memory locations are to be used to hold the
1842 values assigned by both statements @samp{I = 3} and
1843 @samp{ASSIGN 10 TO I}, for example.
1844 (Normally, @code{g77} uses a separate memory location
1845 to hold assigned statement labels.)
1848 @xref{Ugly Assigned Labels}, for more information.
1852 @code{FORMAT} and @code{ENTRY} statements now are allowed to
1853 precede @code{IMPLICIT NONE} statements.
1855 @ifclear USERVISONLY
1857 Produce diagnostic for unsupported @code{SELECT CASE} on
1858 @code{CHARACTER} type, instead of crashing, at compile time.
1861 @ifclear USERVISONLY
1863 Fix crashes involving diagnosed or invalid code.
1866 @ifclear USERVISONLY
1868 Change approach to building @code{libf2c} archive
1869 (@file{libf2c.a}) so that members are added to it
1870 only when truly necessary, so the user that installs
1871 an already-built @code{g77} doesn't need to have write
1872 access to the build tree (whereas the user doing the
1873 build might not have access to install new software
1877 @ifclear USERVISONLY
1879 Support @code{gcc} version 2.7.2.2
1880 (modified by @code{g77} into version 2.7.2.2.f.2),
1882 support for prior versions of @code{gcc}.
1885 @ifclear USERVISONLY
1887 Upgrade to @code{libf2c} as of 1997-02-08, and
1888 fix up some of the build procedures.
1891 @ifclear USERVISONLY
1893 Improve general build procedures for @code{g77},
1894 fixing minor bugs (such as deletion of any file
1895 named @file{f771} in the parent directory of @code{gcc/}).
1899 Enable full support of @code{INTEGER(KIND=2)}
1900 (often referred to as @code{INTEGER*8})
1902 @code{libf2c} and @file{f2c.h} so that @code{f2c} users
1903 may make full use of its features via the @code{g77}
1904 version of @file{f2c.h} and the @code{INTEGER(KIND=2)}
1905 support routines in the @code{g77} version of @code{libf2c}.
1908 Improve @code{g77} driver and @code{libf2c} so that @samp{g77 -v}
1909 yields version information on the library.
1912 The @code{SNGL} and @code{FLOAT} intrinsics now are
1913 specific intrinsics, instead of synonyms for the
1914 generic intrinsic @code{REAL}.
1917 New intrinsics have been added.
1918 These are @code{REALPART}, @code{IMAGPART},
1920 @code{LONG}, and @code{SHORT}.
1923 A new group of intrinsics, @code{gnu}, has been added
1924 to contain the new @code{REALPART}, @code{IMAGPART},
1925 and @code{COMPLEX} intrinsics.
1926 An old group, @code{dcp}, has been removed.
1929 Complain about industry-wide ambiguous references
1930 @samp{REAL(@var{expr})} and @samp{AIMAG(@var{expr})},
1931 where @var{expr} is @code{DOUBLE COMPLEX} (or any
1932 complex type other than @code{COMPLEX}), unless
1933 @samp{-ff90} option specifies Fortran 90 interpretation
1934 or new @samp{-fugly-complex} option, in conjunction with
1935 @samp{-fnot-f90}, specifies @code{f2c} interpretation.
1937 @ifclear USERVISONLY
1939 Make improvements to diagnostics.
1942 @ifclear USERVISONLY
1944 Speed up compiler a bit.
1947 @ifclear USERVISONLY
1949 Improvements to documentation and indexing, including
1950 a new chapter containing information on one, later
1951 more, diagnostics that users are directed to pull
1952 up automatically via a message in the diagnostic itself.
1954 (Hence the menu item @code{M} for the node
1955 @code{Diagnostics} in the top-level menu of
1956 the Info documentation.)
1960 @ifclear DOC-OLDNEWS
1961 @heading In previous versions:
1963 Information on previous versions is archived
1964 in @file{@value{path-g77}/news.texi}
1965 following the test of the @code{DOC-OLDNEWS} macro.
1969 @c 1997-02-01: 0.5.19.1 released.
1970 @heading In 0.5.19.1:
1973 Code-generation bugs afflicting operations on complex
1974 data have been fixed.
1976 These bugs occurred when assigning the result of an
1977 operation to a complex variable (or array element)
1978 that also served as an input to that operation.
1980 The operations affected by this bug were: @code{CONJG()},
1981 @code{DCONJG()}, @code{CCOS()}, @code{CDCOS()},
1982 @code{CLOG()}, @code{CDLOG()}, @code{CSIN()}, @code{CDSIN()},
1983 @code{CSQRT()}, @code{CDSQRT()}, complex division, and
1984 raising a @code{DOUBLE COMPLEX} operand to an @code{INTEGER}
1986 (The related generic and @samp{Z}-prefixed intrinsics,
1987 such as @code{ZSIN()}, also were affected.)
1989 For example, @samp{C = CSQRT(C)}, @samp{Z = Z/C}, and @samp{Z = Z**I}
1990 (where @samp{C} is @code{COMPLEX} and @samp{Z} is
1991 @code{DOUBLE COMPLEX}) have been fixed.
1994 @c 1996-12-07: 0.5.19 released.
1998 Fix @code{FORMAT} statement parsing so negative values for
1999 specifiers such as @code{P} (e.g. @samp{FORMAT(-1PF8.1)})
2000 are correctly processed as negative.
2003 Fix @code{SIGNAL} intrinsic so it once again accepts a
2004 procedure as its second argument.
2007 A temporary kludge option provides bare-bones information on
2008 @code{COMMON} and @code{EQUIVALENCE} members at debug time.
2011 New @samp{-fonetrip} option specifies FORTRAN-66-style
2012 one-trip @code{DO} loops.
2015 New @samp{-fno-silent} option causes names of program units
2016 to be printed as they are compiled, in a fashion similar to
2017 UNIX @code{f77} and @code{f2c}.
2020 New @samp{-fugly-assumed} option specifies that arrays
2021 dimensioned via @samp{DIMENSION X(1)}, for example, are to be
2022 treated as assumed-size.
2025 New @samp{-fno-typeless-boz} option specifies that non-decimal-radix
2026 constants using the prefixed-radix form (such as @samp{Z'1234'})
2027 are to be interpreted as @code{INTEGER} constants.
2030 New @samp{-ff66} option is a ``shorthand'' option that specifies
2031 behaviors considered appropriate for FORTRAN 66 programs.
2034 New @samp{-ff77} option is a ``shorthand'' option that specifies
2035 behaviors considered appropriate for UNIX @code{f77} programs.
2038 New @samp{-fugly-comma} and @samp{-fugly-logint} options provided
2039 to perform some of what @samp{-fugly} used to do.
2040 @samp{-fugly} and @samp{-fno-ugly} are now ``shorthand'' options,
2041 in that they do nothing more than enable (or disable) other
2042 @samp{-fugly-*} options.
2045 Fix parsing of assignment statements involving targets that
2046 are substrings of elements of @code{CHARACTER} arrays having
2047 names such as @samp{READ}, @samp{WRITE}, @samp{GOTO}, and
2048 @samp{REALFUNCTIONFOO}.
2051 Fix crashes involving diagnosed code.
2054 Fix handling of local @code{EQUIVALENCE} areas so certain cases
2055 of valid Fortran programs are not misdiagnosed as improperly
2056 extending the area backwards.
2059 Support @code{gcc} version 2.7.2.1.
2062 Upgrade to @code{libf2c} as of 1996-09-26, and
2063 fix up some of the build procedures.
2066 Change code generation for list-directed I/O so it allows
2067 for new versions of @code{libf2c} that might return non-zero
2068 status codes for some operations previously assumed to always
2071 This change not only affects how @code{IOSTAT=} variables
2072 are set by list-directed I/O, it also affects whether
2073 @code{END=} and @code{ERR=} labels are reached by these
2077 Add intrinsic support for new @code{FTELL} and @code{FSEEK}
2078 procedures in @code{libf2c}.
2081 Modify @code{fseek_()} in @code{libf2c} to be more portable
2082 (though, in practice, there might be no systems where this
2083 matters) and to catch invalid @code{whence} arguments.
2086 Some useless warnings from the @samp{-Wunused} option have
2090 Fix a problem building the @file{f771} executable
2091 on AIX systems by linking with the @samp{-bbigtoc} option.
2094 Abort configuration if @code{gcc} has not been patched
2095 using the patch file provided in the @file{gcc/f/gbe/}
2099 Add options @samp{--help} and @samp{--version} to the
2100 @code{g77} command, to conform to GNU coding guidelines.
2101 Also add printing of @code{g77} version number when
2102 the @samp{--verbose} (@samp{-v}) option is used.
2105 Change internally generated name for local @code{EQUIVALENCE}
2106 areas to one based on the alphabetically sorted first name
2107 in the list of names for entities placed at the beginning
2111 Improvements to documentation and indexing.
2114 @c 1996-04-01: 0.5.18 released.
2118 Add some rudimentary support for @code{INTEGER*1},
2119 @code{INTEGER*2}, @code{INTEGER*8},
2120 and their @code{LOGICAL} equivalents.
2121 (This support works on most, maybe all, @code{gcc} targets.)
2123 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2124 for providing the patch for this!
2126 Among the missing elements from the support for these
2127 features are full intrinsic support and constants.
2130 Add some rudimentary support for the @code{BYTE} and
2131 @code{WORD} type-declaration statements.
2132 @code{BYTE} corresponds to @code{INTEGER*1},
2133 while @code{WORD} corresponds to @code{INTEGER*2}.
2135 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2136 for providing the patch for this!
2139 The compiler code handling intrinsics has been largely
2140 rewritten to accommodate the new types.
2141 No new intrinsics or arguments for existing
2142 intrinsics have been added, so there is, at this
2143 point, no intrinsic to convert to @code{INTEGER*8},
2147 Support automatic arrays in procedures.
2150 Reduce space/time requirements for handling large
2151 @emph{sparsely} initialized aggregate arrays.
2152 This improvement applies to only a subset of
2153 the general problem to be addressed in 0.6.
2156 Treat initial values of zero as if they weren't
2157 specified (in DATA and type-declaration statements).
2158 The initial values will be set to zero anyway, but the amount
2159 of compile time processing them will be reduced,
2160 in some cases significantly (though, again, this
2161 is only a subset of the general problem to be
2164 A new option, @samp{-fzeros}, is introduced to
2165 enable the traditional treatment of zeros as any
2169 With @samp{-ff90} in force, @code{g77} incorrectly
2170 interpreted @samp{REAL(Z)} as returning a @code{REAL}
2171 result, instead of as a @code{DOUBLE PRECISION}
2173 (Here, @samp{Z} is @code{DOUBLE COMPLEX}.)
2175 With @samp{-fno-f90} in force, the interpretation remains
2176 unchanged, since this appears to be how at least some
2177 F77 code using the @code{DOUBLE COMPLEX} extension expected
2180 Essentially, @samp{REAL(Z)} in F90 is the same as
2181 @samp{DBLE(Z)}, while in extended F77, it appears to
2182 be the same as @samp{REAL(REAL(Z))}.
2185 An expression involving exponentiation, where both operands
2186 were type @code{INTEGER} and the right-hand operand
2187 was negative, was erroneously evaluated.
2190 Fix bugs involving @code{DATA} implied-@code{DO} constructs
2191 (these involved an errant diagnostic and a crash, both on good
2192 code, one involving subsequent statement-function definition).
2195 Close @code{INCLUDE} files after processing them, so compiling source
2196 files with lots of @code{INCLUDE} statements does not result in
2197 being unable to open @code{INCLUDE} files after all the available
2198 file descriptors are used up.
2201 Speed up compiling, especially of larger programs, and perhaps
2202 slightly reduce memory utilization while compiling (this is
2203 @emph{not} the improvement planned for 0.6 involving large aggregate
2204 areas)---these improvements result from simply turning
2205 off some low-level code to do self-checking that hasn't been
2206 triggered in a long time.
2209 Introduce three new options that
2210 implement optimizations in the @code{gcc} back end (GBE).
2211 These options are @samp{-fmove-all-movables}, @samp{-freduce-all-givs},
2212 and @samp{-frerun-loop-opt}, which are enabled, by default,
2213 for Fortran compilations.
2214 These optimizations are intended to help toon Fortran programs.
2217 Patch the GBE to do a better job optimizing certain
2218 kinds of references to array elements.
2221 Due to patches to the GBE, the version number of @code{gcc}
2222 also is patched to make it easier to manage installations,
2223 especially useful if it turns out a @code{g77} change to the
2226 The @code{g77}-modified version number is the @code{gcc}
2227 version number with the string @samp{.f.@var{n}} appended,
2228 where @samp{f} identifies the version as enhanced for
2229 Fortran, and @var{n} is @samp{1} for the first Fortran
2230 patch for that version of @code{gcc}, @samp{2} for the
2233 So, this introduces version 2.7.2.f.1 of @code{gcc}.
2236 Make several improvements and fixes to diagnostics, including
2237 the removal of two that were inappropriate or inadequate.
2240 Warning about two successive arithmetic operators, produced
2241 by @samp{-Wsurprising}, now produced @emph{only} when both
2242 operators are, indeed, arithmetic (not relational/boolean).
2245 @samp{-Wsurprising} now warns about the remaining cases
2246 of using non-integral variables for implied-@code{DO}
2247 loops, instead of these being rejected unless @samp{-fpedantic}
2248 or @samp{-fugly} specified.
2251 Allow @code{SAVE} of a local variable or array, even after
2252 it has been given an initial value via @code{DATA}, for example.
2255 Introduce an Info version of @code{g77} documentation, which
2256 supercedes @file{gcc/f/CREDITS}, @file{gcc/f/DOC}, and
2257 @file{gcc/f/PROJECTS}.
2258 These files will be removed in a future release.
2259 The files @file{gcc/f/BUGS}, @file{gcc/f/INSTALL}, and
2260 @file{gcc/f/NEWS} now are automatically built from
2261 the texinfo source when distributions are made.
2263 This effort was inspired by a first pass at translating
2264 @file{g77-0.5.16/f/DOC} that was contributed to Craig by
2265 David Ronis (@email{ronis@@onsager.chem.mcgill.ca}).
2268 New @samp{-fno-second-underscore} option to specify
2269 that, when @samp{-funderscoring} is in effect, a second
2270 underscore is not to be appended to Fortran names already
2271 containing an underscore.
2274 Change the way iterative @code{DO} loops work to follow
2276 In particular, calculation of the iteration count is
2277 still done by converting the start, end, and increment
2278 parameters to the type of the @code{DO} variable, but
2279 the result of the calculation is always converted to
2280 the default @code{INTEGER} type.
2282 (This should have no effect on existing code compiled
2283 by @code{g77}, but code written to assume that use
2284 of a @emph{wider} type for the @code{DO} variable
2285 will result in an iteration count being fully calculated
2286 using that wider type (wider
2287 than default @code{INTEGER}) must be rewritten.)
2290 Support @code{gcc} version 2.7.2.
2293 Upgrade to @code{libf2c} as of 1996-03-23, and
2294 fix up some of the build procedures.
2296 Note that the email addresses related to @code{f2c}
2297 have changed---the distribution site now is
2298 named @code{netlib.bell-labs.com}, and the
2299 maintainer's new address is @email{dmg@@bell-labs.com}.
2302 @c 1995-11-18: 0.5.17 released.
2306 @strong{Fix serious bug} in @samp{g77 -v} command that can cause removal of a
2307 system's @file{/dev/null} special file if run by user @code{root}.
2309 @strong{All users} of version 0.5.16 should ensure that
2310 they have not removed @file{/dev/null} or replaced it with an ordinary
2311 file (e.g. by comparing the output of @samp{ls -l /dev/null} with
2312 @samp{ls -l /dev/zero}.
2313 If the output isn't basically the
2314 same, contact your system
2315 administrator about restoring @file{/dev/null} to its proper status).
2317 This bug is particularly insidious because removing @file{/dev/null} as
2318 a special file can go undetected for quite a while, aside from
2319 various applications and programs exhibiting sudden, strange
2322 I sincerely apologize for not realizing the
2323 implications of the fact that when @samp{g77 -v} runs the @code{ld} command
2324 with @samp{-o /dev/null} that @code{ld} tries to @emph{remove} the executable
2325 it is supposed to build (especially if it reports unresolved
2326 references, which it should in this case)!
2329 Fix crash on @samp{CHARACTER*(*) FOO} in a main or block data program unit.
2332 Fix crash that can occur when diagnostics given outside of any
2333 program unit (such as when input file contains @samp{@@foo}).
2336 Fix crashes, infinite loops (hangs), and such involving diagnosed code.
2339 Fix @code{ASSIGN}'ed variables so they can be @code{SAVE}'d or dummy arguments,
2340 and issue clearer error message in cases where target of @code{ASSIGN}
2341 or @code{ASSIGN}ed @code{GOTO}/@code{FORMAT} is too small (which should
2345 Make @code{libf2c} build procedures work on more systems again by
2346 eliminating unnecessary invocations of @samp{ld -r -x} and @code{mv}.
2349 Fix omission of @samp{-funix-intrinsics-@dots{}} options in list of permitted
2350 options to compiler.
2353 Fix failure to always diagnose missing type declaration for
2354 @code{IMPLICIT NONE}.
2357 Fix compile-time performance problem (which could sometimes
2358 crash the compiler, cause a hang, or whatever, due to a bug
2359 in the back end) involving exponentiation with a large @code{INTEGER}
2360 constant for the right-hand operator (e.g. @samp{I**32767}).
2363 Fix build procedures so cross-compiling @code{g77} (the @code{fini}
2364 utility in particular) is properly built using the host compiler.
2367 Add new @samp{-Wsurprising} option to warn about constructs that are
2368 interpreted by the Fortran standard (and @code{g77}) in ways that
2369 are surprising to many programmers.
2372 Add @code{ERF()} and @code{ERFC()} as generic intrinsics mapping to existing
2373 @code{ERF}/@code{DERF} and @code{ERFC}/@code{DERFC} specific intrinsics.
2375 @emph{Note:} You should
2376 specify @samp{INTRINSIC ERF,ERFC} in any code where you might use
2377 these as generic intrinsics, to improve likelihood of diagnostics
2378 (instead of subtle run-time bugs) when using a compiler that
2379 doesn't support these as intrinsics (e.g. @code{f2c}).
2382 Remove from @samp{-fno-pedantic} the diagnostic about @code{DO}
2383 with non-@code{INTEGER} index variable; issue that under
2384 @samp{-Wsurprising} instead.
2387 Clarify some diagnostics that say things like ``ignored'' when that's
2391 Clarify diagnostic on use of @code{.EQ.}/@code{.NE.} on @code{LOGICAL}
2395 Minor improvements to code generation for various operations on
2396 @code{LOGICAL} operands.
2399 Minor improvement to code generation for some @code{DO} loops on some
2403 Support @code{gcc} version 2.7.1.
2406 Upgrade to @code{libf2c} as of 1995-11-15.
2409 @c 1995-08-30: 0.5.16 released.
2413 Fix a code-generation bug involving complicated @code{EQUIVALENCE} statements
2414 not involving @code{COMMON}.
2417 Fix code-generation bugs involving invoking ``gratis'' library procedures
2418 in @code{libf2c} from code compiled with @samp{-fno-f2c} by making these
2419 procedures known to @code{g77} as intrinsics (not affected by -fno-f2c).
2420 This is known to fix code invoking @code{ERF()}, @code{ERFC()},
2421 @code{DERF()}, and @code{DERFC()}.
2424 Update @code{libf2c} to include netlib patches through 1995-08-16, and
2425 @code{#define} @code{WANT_LEAD_0} to 1 to make @code{g77}-compiled code more
2426 consistent with other Fortran implementations by outputting
2427 leading zeros in formatted and list-directed output.
2430 Fix a code-generation bug involving adjustable dummy arrays with high
2431 bounds whose primaries are changed during procedure execution, and
2432 which might well improve code-generation performance for such arrays
2433 compared to @code{f2c} plus @code{gcc} (but apparently only when using
2434 @file{gcc-2.7.0} or later).
2437 Fix a code-generation bug involving invocation of @code{COMPLEX} and
2438 @code{DOUBLE COMPLEX} @code{FUNCTION}s and doing @code{COMPLEX} and
2439 @code{DOUBLE COMPLEX} divides, when the result
2440 of the invocation or divide is assigned directly to a variable
2441 that overlaps one or more of the arguments to the invocation or divide.
2444 Fix crash by not generating new optimal code for @samp{X**I} if @samp{I} is
2445 nonconstant and the expression is used to dimension a dummy
2446 array, since the @code{gcc} back end does not support the necessary
2447 mechanics (and the @code{gcc} front end rejects the equivalent
2448 construct, as it turns out).
2451 Fix crash on expressions like @samp{COMPLEX**INTEGER}.
2454 Fix crash on expressions like @samp{(1D0,2D0)**2}, i.e. raising a
2455 @code{DOUBLE COMPLEX} constant to an @code{INTEGER} constant power.
2458 Fix crashes and such involving diagnosed code.
2461 Diagnose, instead of crashing on, statement function definitions
2462 having duplicate dummy argument names.
2465 Fix bug causing rejection of good code involving statement function
2469 Fix bug resulting in debugger not knowing size of local equivalence
2470 area when any member of area has initial value (via @code{DATA},
2474 Fix installation bug that prevented installation of @code{g77} driver.
2475 Provide for easy selection of whether to install copy of @code{g77}
2476 as @code{f77} to replace the broken code.
2479 Fix @code{gcc} driver (affects @code{g77} thereby) to not
2480 gratuitously invoke the
2481 @code{f771} program (e.g. when @samp{-E} is specified).
2484 Fix diagnostic to point to correct source line when it immediately
2485 follows an @code{INCLUDE} statement.
2488 Support more compiler options in @code{gcc}/@code{g77} when
2489 compiling Fortran files.
2490 These options include @samp{-p}, @samp{-pg}, @samp{-aux-info}, @samp{-P},
2491 correct setting of version-number macros for preprocessing, full
2492 recognition of @samp{-O0}, and
2493 automatic insertion of configuration-specific linker specs.
2496 Add new intrinsics that interface to existing routines in @code{libf2c}:
2497 @code{ABORT}, @code{DERF}, @code{DERFC}, @code{ERF}, @code{ERFC}, @code{EXIT},
2498 @code{FLUSH}, @code{GETARG}, @code{GETENV}, @code{IARGC},
2499 @code{SIGNAL}, and @code{SYSTEM}.
2500 Note that @code{ABORT}, @code{EXIT}, @code{FLUSH}, @code{SIGNAL}, and
2501 @code{SYSTEM} are intrinsic subroutines, not functions (since they
2502 have side effects), so to get the return values from @code{SIGNAL}
2503 and @code{SYSTEM}, append a final argument specifying an @code{INTEGER}
2504 variable or array element (e.g. @samp{CALL SYSTEM('rm foo',ISTAT)}).
2507 Add new intrinsic group named @code{unix} to contain the new intrinsics,
2508 and by default enable this new group.
2511 Move @code{LOC()} intrinsic out of the @code{vxt} group to the new
2515 Improve @code{g77} so that @samp{g77 -v} by itself (or with
2516 certain other options, including @samp{-B}, @samp{-b}, @samp{-i},
2517 @samp{-nostdlib}, and @samp{-V}) reports lots more useful
2518 version info, and so that long-form options @code{gcc} accepts are
2519 understood by @code{g77} as well (even in truncated, unambiguous forms).
2522 Add new @code{g77} option @samp{--driver=name} to specify driver when
2523 default, @code{gcc}, isn't appropriate.
2526 Add support for @samp{#} directives (as output by the preprocessor) in the
2527 compiler, and enable generation of those directives by the
2528 preprocessor (when compiling @samp{.F} files) so diagnostics and debugging
2529 info are more useful to users of the preprocessor.
2532 Produce better diagnostics, more like @code{gcc}, with info such as
2533 @samp{In function `foo':} and @samp{In file included from...:}.
2536 Support @code{gcc}'s @samp{-fident} and @samp{-fno-ident} options.
2539 When @samp{-Wunused} in effect, don't warn about local variables used as
2540 statement-function dummy arguments or @code{DATA} implied-@code{DO} iteration
2541 variables, even though, strictly speaking, these are not uses
2542 of the variables themselves.
2545 When @samp{-W -Wunused} in effect, don't warn about unused dummy arguments
2546 at all, since there's no way to turn this off for individual
2547 cases (@code{g77} might someday start warning about these)---applies
2548 to @code{gcc} versions 2.7.0 and later, since earlier versions didn't
2549 warn about unused dummy arguments.
2552 New option @samp{-fno-underscoring} that inhibits transformation of names
2553 (by appending one or two underscores) so users may experiment
2554 with implications of such an environment.
2557 Minor improvement to @file{gcc/f/info} module to make it easier to build
2558 @code{g77} using the native (non-@code{gcc}) compiler on certain machines
2559 (but definitely not all machines nor all non-@code{gcc} compilers).
2561 do not report bugs showing problems compilers have with
2562 macros defined in @file{gcc/f/target.h} and used in places like
2563 @file{gcc/f/expr.c}.
2566 Add warning to be printed for each invocation of the compiler
2567 if the target machine @code{INTEGER}, @code{REAL}, or @code{LOGICAL} size
2569 since @code{g77} is known to not work well for such cases.
2572 Lots of new documentation (though work is still needed to put it into
2573 canonical GNU format).
2576 Build @code{libf2c} with @samp{-g0}, not @samp{-g2}, in effect
2577 (by default), to produce
2578 smaller library without lots of debugging clutter.
2581 @c 1995-05-19: 0.5.15 released.
2585 Fix bad code generation involving @samp{X**I} and temporary, internal variables
2586 generated by @code{g77} and the back end (such as for @code{DO} loops).
2589 Fix crash given @samp{CHARACTER A;DATA A/.TRUE./}.
2592 Replace crash with diagnostic given @samp{CHARACTER A;DATA A/1.0/}.
2595 Fix crash or other erratic behavior when null character constant
2596 (@samp{''}) is encountered.
2599 Fix crash or other erratic behavior involving diagnosed code.
2602 Fix code generation for external functions returning type @code{REAL} when
2603 the @samp{-ff2c} option is in force (which it is by default) so that
2604 @code{f2c} compatibility is indeed provided.
2607 Disallow @samp{COMMON I(10)} if @samp{I} has previously been specified
2608 with an array declarator.
2611 New @samp{-ffixed-line-length-@var{n}} option, where @var{n} is the
2613 of a typical fixed-form line, defaulting to 72 columns, such
2614 that characters beyond column @var{n} are ignored, or @var{n} is @samp{none},
2615 meaning no characters are ignored.
2616 does not affect lines
2617 with @samp{&} in column 1, which are always processed as if
2618 @samp{-ffixed-line-length-none} was in effect.
2621 No longer generate better code for some kinds of array references,
2622 as @code{gcc} back end is to be fixed to do this even better, and it
2623 turned out to slow down some code in some cases after all.
2626 In @code{COMMON} and @code{EQUIVALENCE} areas with any members given initial
2627 values (e.g. via @code{DATA}), uninitialized members now always
2628 initialized to binary zeros (though this is not required by
2629 the standard, and might not be done in future versions
2631 Previously, in some @code{COMMON}/@code{EQUIVALENCE} areas
2632 (essentially those with members of more than one type), the
2633 uninitialized members were initialized to spaces, to
2634 cater to @code{CHARACTER} types, but it seems no existing code expects
2635 that, while much existing code expects binary zeros.
2641 Don't emit bad code when low bound of adjustable array is nonconstant
2642 and thus might vary as an expression at run time.
2645 Emit correct code for calculation of number of trips in @code{DO} loops
2647 where the loop should not execute at all.
2648 (This bug affected cases
2649 where the difference between the begin and end values was less
2650 than the step count, though probably not for floating-point cases.)
2653 Fix crash when extra parentheses surround item in
2654 @code{DATA} implied-@code{DO} list.
2657 Fix crash over minor internal inconsistencies in handling diagnostics,
2658 just substitute dummy strings where necessary.
2661 Fix crash on some systems when compiling call to @code{MVBITS()} intrinsic.
2664 Fix crash on array assignment @samp{TYPE@var{ddd}(@dots{})=@dots{}}, where @var{ddd}
2665 is a string of one or more digits.
2668 Fix crash on @code{DCMPLX()} with a single @code{INTEGER} argument.
2671 Fix various crashes involving code with diagnosed errors.
2674 Support @samp{-I} option for @code{INCLUDE} statement, plus @code{gcc}'s
2675 @file{header.gcc} facility for handling systems like MS-DOS.
2678 Allow @code{INCLUDE} statement to be continued across multiple lines,
2679 even allow it to coexist with other statements on the same line.
2682 Incorporate Bellcore fixes to @code{libf2c} through 1995-03-15---this
2683 fixes a bug involving infinite loops reading EOF with empty list-directed
2687 Remove all the @code{g77}-specific auto-configuration scripts, code,
2689 except for temporary substitutes for bsearch() and strtoul(), as
2690 too many configure/build problems were reported in these areas.
2691 People will have to fix their systems' problems themselves, or at
2692 least somewhere other than @code{g77}, which expects a working ANSI C
2693 environment (and, for now, a GNU C compiler to compile @code{g77} itself).
2696 Complain if initialized common redeclared as larger in subsequent program
2700 Warn if blank common initialized, since its size can vary and hence
2701 related warnings that might be helpful won't be seen.
2704 New @samp{-fbackslash} option, on by default, that causes @samp{\}
2705 within @code{CHARACTER}
2706 and Hollerith constants to be interpreted a la GNU C.
2708 this behavior is somewhat different from @code{f2c}'s, which supports only
2709 a limited subset of backslash (escape) sequences.
2712 Make @samp{-fugly-args} the default.
2715 New @samp{-fugly-init} option, on by default, that allows typeless/Hollerith
2716 to be specified as initial values for variables or named constants
2717 (@code{PARAMETER}), and also allows character<->numeric conversion in
2718 those contexts---turn off via @samp{-fno-ugly-init}.
2721 New @samp{-finit-local-zero} option to initialize
2722 local variables to binary zeros.
2723 This does not affect whether they are @code{SAVE}d, i.e. made
2724 automatic or static.
2727 New @samp{-Wimplicit} option to warn about implicitly typed variables, arrays,
2729 (Basically causes all program units to default to @code{IMPLICIT NONE}.)
2732 @samp{-Wall} now implies @samp{-Wuninitialized} as with @code{gcc}
2733 (i.e. unless @samp{-O} not specified, since @samp{-Wuninitialized}
2734 requires @samp{-O}), and implies @samp{-Wunused} as well.
2737 @samp{-Wunused} no longer gives spurious messages for unused
2738 @code{EXTERNAL} names (since they are assumed to refer to block data
2739 program units, to make use of libraries more reliable).
2742 Support @code{%LOC()} and @code{LOC()} of character arguments.
2745 Support null (zero-length) character constants and expressions.
2748 Support @code{f2c}'s @code{IMAG()} generic intrinsic.
2751 Support @code{ICHAR()}, @code{IACHAR()}, and @code{LEN()} of
2752 character expressions that are valid in assignments but
2753 not normally as actual arguments.
2756 Support @code{f2c}-style @samp{&} in column 1 to mean continuation line.
2759 Allow @code{NAMELIST}, @code{EXTERNAL}, @code{INTRINSIC}, and @code{VOLATILE}
2760 in @code{BLOCK DATA}, even though these are not allowed by the standard.
2763 Allow @code{RETURN} in main program unit.
2766 Changes to Hollerith-constant support to obey Appendix C of the
2771 Now padded on the right with zeros, not spaces.
2774 Hollerith ``format specifications'' in the form of arrays of
2775 non-character allowed.
2778 Warnings issued when non-space truncation occurs when converting
2782 When specified as actual argument, now passed
2783 by reference to @code{INTEGER} (padded on right with spaces if constant
2784 too small, otherwise fully intact if constant wider the @code{INTEGER}
2785 type) instead of by value.
2788 @strong{Warning:} @code{f2c} differs on the
2789 interpretation of @samp{CALL FOO(1HX)}, which it treats exactly the
2790 same as @samp{CALL FOO('X')}, but which the standard and @code{g77} treat
2791 as @samp{CALL FOO(%REF('X '))} (padded with as many spaces as necessary
2792 to widen to @code{INTEGER}), essentially.
2795 Changes and fixes to typeless-constant support:
2799 Now treated as a typeless double-length @code{INTEGER} value.
2802 Warnings issued when overflow occurs.
2805 Padded on the left with zeros when converting
2809 Should be properly aligned and ordered on
2810 the target machine for whatever type it is turned into.
2813 When specified as actual argument, now passed as reference to
2814 a default @code{INTEGER} constant.
2818 @code{%DESCR()} of a non-@code{CHARACTER} expression now passes a pointer to
2819 the expression plus a length for the expression just as if
2820 it were a @code{CHARACTER} expression.
2821 For example, @samp{CALL FOO(%DESCR(D))}, where
2822 @samp{D} is @code{REAL*8}, is the same as @samp{CALL FOO(D,%VAL(8)))}.
2825 Name of multi-entrypoint master function changed to incorporate
2826 the name of the primary entry point instead of a decimal
2827 value, so the name of the master function for @samp{SUBROUTINE X}
2828 with alternate entry points is now @samp{__g77_masterfun_x}.
2831 Remove redundant message about zero-step-count @code{DO} loops.
2834 Clean up diagnostic messages, shortening many of them.
2837 Fix typo in @code{g77} man page.
2840 Clarify implications of constant-handling bugs in @file{f/BUGS}.
2843 Generate better code for @samp{**} operator with a right-hand operand of
2844 type @code{INTEGER}.
2847 Generate better code for @code{SQRT()} and @code{DSQRT()},
2848 also when @samp{-ffast-math}
2849 specified, enable better code generation for @code{SIN()} and @code{COS()}.
2852 Generate better code for some kinds of array references.
2855 Speed up lexing somewhat (this makes the compilation phase noticeably