1 @c Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000 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 1995,1996,1997,1998,1999,2000,2001
12 @set last-update-news 2001-04-21
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, as appropriate.
100 Note that all such variants are obsolete @emph{as of July 1999} -
101 the information is retained here only for its historical value.
103 Therefore, @code{egcs} versions sometimes have multiple listings
104 to help clarify how they differ from other versions,
105 though this can make getting a complete picture
106 of what a particular @code{egcs} version contains
107 somewhat more difficult.
110 For information on bugs in the @value{which-g77} version of @code{g77},
111 see @ref{Known Bugs,,Known Bugs In GNU Fortran}.
115 For information on bugs in the @value{which-g77} version of @code{g77},
116 see @file{@value{path-g77}/BUGS}.
120 @emph{Warning:} The information below is still under development,
121 and might not accurately reflect the @code{g77} code base
122 of which it is a part.
123 Efforts are made to keep it somewhat up-to-date,
124 but they are particularly concentrated
125 on any version of this information
126 that is distributed as part of a @emph{released} @code{g77}.
128 In particular, while this information is intended to apply to
129 the @value{which-g77} version of @code{g77},
130 only an official @emph{release} of that version
131 is expected to contain documentation that is
132 most consistent with the @code{g77} product in that version.
134 Nevertheless, information on @emph{previous} releases of @code{g77}, below,
135 is likely to be more up-to-date and accurate
136 than the equivalent information that accompanied
138 assuming the last-updated date of the information below
139 is later than the dates of those releases.
141 That's due to attempts to keep this development version
142 of news about previous @code{g77} versions up-to-date.
146 An online, ``live'' version of this document
147 (derived directly from the mainline, development version
148 of @code{g77} within @code{gcc})
150 @uref{http://www.gnu.org/software/gcc/onlinedocs/g77_news.html}.
153 The following information was last updated on @value{last-update-news}:
155 @heading In development, 0.5.27, @code{GCC} 3.1 versus @code{GCC} 3.0:
163 @heading In prerelease, 0.5.26, @code{GCC} 3.0 versus @code{GCC} 2.95:
167 When a REWIND was issued after a WRITE statement on an unformatted
168 file, the implicit truncation was performed by copying the truncated
169 file to /tmp and copying the result back. This has been fixed by using
170 the @code{ftruncate} OS function. Thanks go to the GAMESS developers
171 for bringing this to our attention.
176 It is not necessary anymore to use the option @code{-femulate-complex}
177 to compile Fortran code using COMPLEX arithmetic, even on 64-bit machines
178 (like the Alpha). This will improve code generation.
183 INTRINSIC arithmetic functions are now treated as routines that do not
184 depend on anything but their argument(s). This enables further instruction
185 scheduling, because it is known that they cannot read or modify arbitrary
191 Upgrade to @code{libf2c} as of 2000-12-05.
193 This fixes a bug where a namelist containing initialization of LOGICAL
194 items and a variable starting with T or F would be read incorrectly.
199 Using options @code{-g}, @code{-ggdb} or @code{-gdwarf[-2]} (where
200 appropriate for your target) now also enables debugging information
201 for COMMON BLOCK and EQUIVALENCE items to be emitted.
206 The @code{TtyNam} intrinsics now set @var{Name} to all spaces (at run time)
207 if the system has no @code{ttyname} implementation available.
212 Upgrade to @code{libf2c} as of 1999-06-28.
214 This fixes a bug whereby
215 input to a @code{NAMELIST} read involving a repeat count,
216 such as @samp{K(5)=10*3},
217 was not properly handled by @code{libf2c}.
218 The first item was written to @samp{K(5)},
219 but the remaining nine were written elsewhere (still within the array),
220 not necessarily starting at @samp{K(6)}.
224 @heading In 0.5.25, @code{GCC} 2.95 (@code{EGCS} 1.2) versus @code{EGCS} 1.1.2:
228 @code{g77} no longer generates bad code for assignments,
229 or other conversions,
230 of @code{REAL} or @code{COMPLEX} constant expressions
231 to type @code{INTEGER(KIND=2)}
232 (often referred to as @code{INTEGER*8}).
234 For example, @samp{INTEGER*8 J; J = 4E10} now works as documented.
239 @code{g77} no longer truncates @code{INTEGER(KIND=2)}
240 (usually @code{INTEGER*8})
241 subscript expressions when evaluating array references
242 on systems with pointers widers than @code{INTEGER(KIND=1)}
248 @code{g77} no longer generates bad code
249 for an assignment to a @code{COMPLEX} variable or array
250 that partially overlaps one or more of the sources
251 of the same assignment
252 (a very rare construction).
253 It now assigns through a temporary,
254 in cases where such partial overlap is deemed possible.
259 @code{libg2c} (@code{libf2c}) no longer loses track
260 of the file being worked on
261 during a @code{BACKSPACE} operation.
266 @code{libg2c} (@code{libf2c}) fixes a bug whereby
267 input to a @code{NAMELIST} read involving a repeat count,
268 such as @samp{K(5)=10*3},
269 was not properly handled by @code{libf2c}.
270 The first item was written to @samp{K(5)},
271 but the remaining nine were written elsewhere (still within the array),
272 not necessarily starting at @samp{K(6)}.
277 @c Tim Prince reported this, regarding the TEST_FPU benchmark.
278 Automatic arrays now seem to be working on HP-UX systems.
283 The @code{Date} intrinsic now returns the correct result
284 on big-endian systems.
289 Fix @code{g77} so it no longer crashes when compiling
290 I/O statements using keywords that define @code{INTEGER} values,
291 such as @samp{IOSTAT=@var{j}},
292 where @var{j} is other than default @code{INTEGER}
293 (such as @code{INTEGER*2}).
294 Instead, it issues a diagnostic.
299 Fix @code{g77} so it properly handles @samp{DATA A/@var{rpt}*@var{val}/},
300 where @var{rpt} is not default @code{INTEGER}, such as @code{INTEGER*2},
301 instead of producing a spurious diagnostic.
302 Also fix @samp{DATA (A(I),I=1,N)},
303 where @samp{N} is not default @code{INTEGER}
304 to work instead of crashing @code{g77}.
309 The @samp{-ax} option is now obeyed when compiling Fortran programs.
310 (It is passed to the @file{f771} driver.)
314 The new @samp{-fbounds-check} option
315 causes @code{g77} to compile run-time bounds checks
316 of array subscripts, as well as of substring start and end points.
319 @code{libg2c} now supports building as multilibbed library,
320 which provides better support for systems
321 that require options such as @samp{-mieee}
325 Source file names with the suffixes @samp{.FOR} and @samp{.FPP}
326 now are recognized by @code{g77}
327 as if they ended in @samp{.for} and @samp{.fpp}, respectively.
330 The order of arguments to the @emph{subroutine} forms of the
331 @code{CTime}, @code{DTime}, @code{ETime}, and @code{TtyNam}
332 intrinsics has been swapped.
333 The argument serving as the returned value
334 for the corresponding function forms
335 now is the @emph{second} argument,
336 making these consistent with the other subroutine forms
337 of @code{libU77} intrinsics.
340 @code{g77} now warns about a reference to an intrinsic
341 that has an interface that is not Year 2000 (Y2K) compliant.
342 Also, @code{libg2c} has been changed to increase the likelihood
343 of catching references to the implementations of these intrinsics
344 using the @code{EXTERNAL} mechanism
345 (which would avoid the new warnings).
348 @xref{Year 2000 (Y2K) Problems}, for more information.
353 @code{g77} now warns about a reference to a function
354 when the corresponding @emph{subsequent} function program unit
355 disagrees with the reference concerning the type of the function.
359 @samp{-fno-emulate-complex} is now the default option.
360 This should result in improved performance
361 of code that uses the @code{COMPLEX} data type.
364 @cindex double-precision performance
365 @cindex -malign-double
367 The @samp{-malign-double} option
368 now reliably aligns @emph{all} double-precision variables and arrays
369 on Intel x86 targets.
373 Even without the @samp{-malign-double} option,
374 @code{g77} reliably aligns local double-precision variables
375 that are not in @code{EQUIVALENCE} areas
376 and not @code{SAVE}'d.
381 @code{g77} now open-codes (``inlines'') division of @code{COMPLEX} operands
382 instead of generating a run-time call to
383 the @code{libf2c} routines @code{c_div} or @code{z_div},
384 unless the @samp{-Os} option is specified.
388 @code{g77} no longer generates code to maintain @code{errno},
389 a C-language concept,
390 when performing operations such as the @code{SqRt} intrinsic.
394 @code{g77} developers can temporarily use
395 the @samp{-fflatten-arrays} option
396 to compare how the compiler handles code generation
397 using C-like constructs as compared to the
398 Fortran-like method constructs normally used.
403 A substantial portion of the @code{g77} front end's code-generation component
405 It now generates code using facilities more robustly supported
406 by the @code{gcc} back end.
407 One effect of this rewrite is that some codes no longer produce
408 a spurious ``label @var{lab} used before containing binding contour''
413 Support for the @samp{-fugly} option has been removed.
417 Improve documentation and indexing,
418 including information on Year 2000 (Y2K) compliance,
419 and providing more information on internals of the front end.
424 Upgrade to @code{libf2c} as of 1999-05-10.
428 @heading In 0.5.24 versus 0.5.23:
430 There is no @code{g77} version 0.5.24 at this time,
432 0.5.24 is the version number designated for bug fixes and,
433 perhaps, some new features added,
435 Version 0.5.23 requires @code{gcc} 2.8.1,
436 as 0.5.24 was planned to require.
438 Due to @code{EGCS} becoming @code{GCC}
439 (which is now an acronym for ``GNU Compiler Collection''),
440 and @code{EGCS} 1.2 becoming officially designated @code{GCC} 2.95,
441 there seems to be no need for an actual 0.5.24 release.
443 To reduce the confusion already resulting from use of 0.5.24
444 to designate @code{g77} versions within @code{EGCS} versions 1.0 and 1.1,
445 as well as in versions of @code{g77} documentation and notices
447 ``mainline'' @code{g77} version numbering resumes
448 at 0.5.25 with @code{GCC} 2.95 (@code{EGCS} 1.2),
449 skipping over 0.5.24 as a placeholder version number.
451 To repeat, there is no @code{g77} 0.5.24, but there is now a 0.5.25.
452 Please remain calm and return to your keypunch units.
454 @c 1999-03-15: EGCS 1.1.2 released.
455 @heading In @code{EGCS} 1.1.2 versus @code{EGCS} 1.1.1:
459 Fix the @code{IDate} intrinsic (VXT) (in @code{libg2c})
460 so the returned year is in the documented, non-Y2K-compliant range
462 instead of being returned as 100 in the year 2000.
465 @xref{IDate Intrinsic (VXT)},
466 for more information.
470 Fix the @code{Date_and_Time} intrinsic (in @code{libg2c})
471 to return the milliseconds value properly
475 Fix the @code{LStat} intrinsic (in @code{libg2c})
476 to return device-ID information properly
480 Improve documentation.
484 @c 1998-12-04: EGCS 1.1.1 released.
485 @heading In @code{EGCS} 1.1.1 versus @code{EGCS} 1.1:
489 Fix @code{libg2c} so it performs an implicit @code{ENDFILE} operation
491 whenever a @code{REWIND} is done.
493 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
494 @code{g77}'s version of @code{libf2c}.)
497 Fix @code{libg2c} so it no longer crashes with a spurious diagnostic
498 upon doing any I/O following a direct formatted write.
500 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
501 @code{g77}'s version of @code{libf2c}.)
504 Fix @code{g77} so it no longer crashes compiling references
505 to the @code{Rand} intrinsic on some systems.
508 Fix @code{g77} portion of installation process so it works
509 better on some systems
510 (those with shells requiring @samp{else true} clauses
511 on @code{if} constructs
512 for the completion code to be set properly).
516 @c 1998-09-03: EGCS 1.1 released.
517 @heading In @code{EGCS} 1.1 versus @code{EGCS} 1.0.3:
521 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
522 that wrote one byte beyond the end of its @code{CHARACTER}
524 and in the @code{libU77} intrinsics
525 @code{GMTime} and @code{LTime}
526 that overwrote their arguments.
531 Assumed arrays with negative bounds
532 (such as @samp{REAL A(-1:*)})
533 no longer elicit spurious diagnostics from @code{g77},
534 even on systems with pointers having
535 different sizes than integers.
537 This bug is not known to have existed in any
538 recent version of @code{gcc}.
539 It was introduced in an early release of @code{egcs}.
544 Valid combinations of @code{EXTERNAL},
545 passing that external as a dummy argument
546 without explicitly giving it a type,
547 and, in a subsequent program unit,
548 referencing that external as
549 an external function with a different type
550 no longer crash @code{g77}.
555 @code{CASE DEFAULT} no longer crashes @code{g77}.
560 The @samp{-Wunused} option no longer issues a spurious
561 warning about the ``master'' procedure generated by
562 @code{g77} for procedures containing @code{ENTRY} statements.
566 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
567 compile-time constant @code{INTEGER} expression.
570 Fix @code{g77} @samp{-g} option so procedures that
571 use @code{ENTRY} can be stepped through, line by line,
575 Allow any @code{REAL} argument to intrinsics
576 @code{Second} and @code{CPU_Time}.
579 Use @code{tempnam}, if available, to open scratch files
580 (as in @samp{OPEN(STATUS='SCRATCH')})
581 so that the @code{TMPDIR} environment variable,
585 @code{g77}'s version of @code{libf2c} separates out
586 the setting of global state
587 (such as command-line arguments and signal handling)
588 from @file{main.o} into distinct, new library
591 This should make it easier to write portable applications
592 that have their own (non-Fortran) @code{main()} routine
593 properly set up the @code{libf2c} environment, even
594 when @code{libf2c} (now @code{libg2c}) is a shared library.
598 @code{g77} no longer installs the @file{f77} command
599 and @file{f77.1} man page
600 in the @file{/usr} or @file{/usr/local} heirarchy,
601 even if the @file{f77-install-ok} file exists
602 in the source or build directory.
603 See the installation documentation for more information.
608 @code{g77} no longer installs the @file{libf2c.a} library
609 and @file{f2c.h} include file
610 in the @file{/usr} or @file{/usr/local} heirarchy,
611 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
612 in the source or build directory.
613 See the installation documentation for more information.
618 The @file{libf2c.a} library produced by @code{g77} has been
619 renamed to @file{libg2c.a}.
620 It is installed only in the @code{gcc} ``private''
621 directory heirarchy, @file{gcc-lib}.
622 This allows system administrators and users to choose which
623 version of the @code{libf2c} library from @code{netlib} they
624 wish to use on a case-by-case basis.
625 See the installation documentation for more information.
630 The @file{f2c.h} include (header) file produced by @code{g77}
631 has been renamed to @file{g2c.h}.
632 It is installed only in the @code{gcc} ``private''
633 directory heirarchy, @file{gcc-lib}.
634 This allows system administrators and users to choose which
635 version of the include file from @code{netlib} they
636 wish to use on a case-by-case basis.
637 See the installation documentation for more information.
641 The @code{g77} command now expects the run-time library
642 to be named @code{libg2c.a} instead of @code{libf2c.a},
643 to ensure that a version other than the one built and
644 installed as part of the same @code{g77} version is picked up.
648 During the configuration and build process,
649 @code{g77} creates subdirectories it needs only as it
651 Other cleaning up of the configuration and build process
652 has been performed as well.
657 @code{install-info} now used to update the directory of
658 Info documentation to contain an entry for @code{g77}
659 (during installation).
663 Some diagnostics have been changed from warnings to errors,
664 to prevent inadvertent use of the resulting, probably buggy,
666 These mostly include diagnostics about use of unsupported features
667 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
668 @code{WRITE} statements,
669 and about truncations of various sorts of constants.
673 Improve compilation of @code{FORMAT} expressions so that
674 a null byte is appended to the last operand if it
676 This provides a cleaner run-time diagnostic as provided
677 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
682 Improve documentation and indexing.
687 The upgrade to @code{libf2c} as of 1998-06-18
688 should fix a variety of problems, including
689 those involving some uses of the @code{T} format
690 specifier, and perhaps some build (porting) problems
695 @c 1998-09-03: EGCS 1.1 released.
696 @heading In @code{EGCS} 1.1 versus @code{g77} 0.5.23:
700 @cindex stack, 387 coprocessor
704 Fix a code-generation bug that afflicted
705 Intel x86 targets when @samp{-O2} was specified
706 compiling, for example, an old version of
707 the @code{DNRM2} routine.
709 The x87 coprocessor stack was being
710 mismanaged in cases involving assigned @code{GOTO}
716 @code{g77} no longer produces incorrect code
718 for @code{EQUIVALENCE} and @code{COMMON}
719 aggregates that, due to ``unnatural'' ordering of members
720 vis-a-vis their types, require initial padding.
725 Fix @code{g77} crash compiling code
726 containing the construct @samp{CMPLX(0.)} or similar.
731 @code{g77} no longer crashes when compiling code
732 containing specification statements such as
733 @samp{INTEGER(KIND=7) PTR}.
738 @code{g77} no longer crashes when compiling code
739 such as @samp{J = SIGNAL(1, 2)}.
743 @code{g77} now treats @samp{%LOC(@var{expr})} and
744 @samp{LOC(@var{expr})} as ``ordinary'' expressions
745 when they are used as arguments in procedure calls.
746 This change applies only to global (filewide) analysis,
747 making it consistent with
748 how @code{g77} actually generates code
751 Previously, @code{g77} treated these expressions
752 as denoting special ``pointer'' arguments
753 for the purposes of filewide analysis.
758 (or apparently infinite run-time)
759 when compiling certain complicated expressions
760 involving @code{COMPLEX} arithmetic
761 (especially multiplication).
765 @cindex double-precision performance
766 @cindex -malign-double
768 Align static double-precision variables and arrays
770 regardless of whether @samp{-malign-double} is specified.
772 Generally, this affects only local variables and arrays
773 having the @code{SAVE} attribute
774 or given initial values via @code{DATA}.
777 The @code{g77} driver now ensures that @samp{-lg2c}
778 is specified in the link phase prior to any
779 occurrence of @samp{-lm}.
780 This prevents accidentally linking to a routine
781 in the SunOS4 @samp{-lm} library
782 when the generated code wants to link to the one
783 in @code{libf2c} (@code{libg2c}).
786 @code{g77} emits more debugging information when
789 This new information allows, for example,
790 @kbd{which __g77_length_a} to be used in @code{gdb}
791 to determine the type of the phantom length argument
792 supplied with @code{CHARACTER} variables.
794 This information pertains to internally-generated
795 type, variable, and other information,
796 not to the longstanding deficiencies vis-a-vis
797 @code{COMMON} and @code{EQUIVALENCE}.
800 The F90 @code{Date_and_Time} intrinsic now is
804 The F90 @code{System_Clock} intrinsic allows
805 the optional arguments (except for the @code{Count}
806 argument) to be omitted.
810 Upgrade to @code{libf2c} as of 1998-06-18.
815 Improve documentation and indexing.
820 @heading In previous versions:
822 Information on previous versions is not provided
823 in this @file{@value{path-g77}/NEWS} file,
825 See @file{@value{path-g77}/news.texi},
826 or any of its other derivations
827 (Info, HTML, dvi forms)
828 for such information.
832 @c 1998-05-20: 0.5.23 released.
833 @heading In 0.5.23 versus 0.5.22:
836 This release contains several regressions against
837 version 0.5.22 of @code{g77}, due to using the
838 ``vanilla'' @code{gcc} back end instead of patching
839 it to fix a few bugs and improve performance in a
842 Features that have been dropped from this version
843 of @code{g77} due to their being implemented
844 via @code{g77}-specific patches to the @code{gcc}
845 back end in previous releases include:
849 Support for @code{__restrict__} keyword,
850 the options @samp{-fargument-alias}, @samp{-fargument-noalias},
851 and @samp{-fargument-noalias-global},
852 and the corresponding alias-analysis code.
854 (@code{egcs} has the alias-analysis
855 code, but not the @code{__restrict__} keyword.
856 @code{egcs} @code{g77} users benefit from the alias-analysis
857 code despite the lack of the @code{__restrict__} keyword,
858 which is a C-language construct.)
861 Support for the GNU compiler options
862 @samp{-fmove-all-movables},
863 @samp{-freduce-all-givs},
864 and @samp{-frerun-loop-opt}.
866 (@code{egcs} supports these options.
867 @code{g77} users of @code{egcs} benefit from them even if
868 they are not explicitly specified,
869 because the defaults are optimized for @code{g77} users.)
872 Support for the @samp{-W} option warning about
873 integer division by zero.
876 The Intel x86-specific option @samp{-malign-double}
877 applying to stack-allocated data
878 as well as statically-allocate data.
882 Note that the @file{gcc/f/gbe/} subdirectory has been removed
883 from this distribution as a result of @code{g77} no longer
884 including patches for the @code{gcc} back end.
889 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
890 that wrote one byte beyond the end of its @code{CHARACTER}
892 and in the @code{libU77} intrinsics
893 @code{GMTime} and @code{LTime}
894 that overwrote their arguments.
898 Support @code{gcc} version 2.8,
899 and remove support for prior versions of @code{gcc}.
901 @cindex -@w{}-driver option
902 @cindex @code{g77} options, -@w{}-driver
903 @cindex options, -@w{}-driver
905 Remove support for the @samp{--driver} option,
906 as @code{g77} now does all the driving,
907 just like @code{gcc}.
911 @code{CASE DEFAULT} no longer crashes @code{g77}.
916 Valid combinations of @code{EXTERNAL},
917 passing that external as a dummy argument
918 without explicitly giving it a type,
919 and, in a subsequent program unit,
920 referencing that external as
921 an external function with a different type
922 no longer crash @code{g77}.
927 @code{g77} no longer installs the @file{f77} command
928 and @file{f77.1} man page
929 in the @file{/usr} or @file{/usr/local} heirarchy,
930 even if the @file{f77-install-ok} file exists
931 in the source or build directory.
932 See the installation documentation for more information.
937 @code{g77} no longer installs the @file{libf2c.a} library
938 and @file{f2c.h} include file
939 in the @file{/usr} or @file{/usr/local} heirarchy,
940 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
941 in the source or build directory.
942 See the installation documentation for more information.
947 The @file{libf2c.a} library produced by @code{g77} has been
948 renamed to @file{libg2c.a}.
949 It is installed only in the @code{gcc} ``private''
950 directory heirarchy, @file{gcc-lib}.
951 This allows system administrators and users to choose which
952 version of the @code{libf2c} library from @code{netlib} they
953 wish to use on a case-by-case basis.
954 See the installation documentation for more information.
959 The @file{f2c.h} include (header) file produced by @code{g77}
960 has been renamed to @file{g2c.h}.
961 It is installed only in the @code{gcc} ``private''
962 directory heirarchy, @file{gcc-lib}.
963 This allows system administrators and users to choose which
964 version of the include file from @code{netlib} they
965 wish to use on a case-by-case basis.
966 See the installation documentation for more information.
970 The @code{g77} command now expects the run-time library
971 to be named @code{libg2c.a} instead of @code{libf2c.a},
972 to ensure that a version other than the one built and
973 installed as part of the same @code{g77} version is picked up.
977 The @samp{-Wunused} option no longer issues a spurious
978 warning about the ``master'' procedure generated by
979 @code{g77} for procedures containing @code{ENTRY} statements.
983 @code{g77}'s version of @code{libf2c} separates out
984 the setting of global state
985 (such as command-line arguments and signal handling)
986 from @file{main.o} into distinct, new library
989 This should make it easier to write portable applications
990 that have their own (non-Fortran) @code{main()} routine
991 properly set up the @code{libf2c} environment, even
992 when @code{libf2c} (now @code{libg2c}) is a shared library.
996 During the configuration and build process,
997 @code{g77} creates subdirectories it needs only as it
998 needs them, thus avoiding unnecessary creation of, for example,
999 @file{stage1/f/runtime} when doing a non-bootstrap build.
1000 Other cleaning up of the configuration and build process
1001 has been performed as well.
1004 @ifclear USERVISONLY
1006 @code{install-info} now used to update the directory of
1007 Info documentation to contain an entry for @code{g77}
1008 (during installation).
1012 Some diagnostics have been changed from warnings to errors,
1013 to prevent inadvertent use of the resulting, probably buggy,
1015 These mostly include diagnostics about use of unsupported features
1016 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
1017 @code{WRITE} statements,
1018 and about truncations of various sorts of constants.
1020 @ifclear USERVISONLY
1022 Improve documentation and indexing.
1025 @ifclear USERVISONLY
1027 Upgrade to @code{libf2c} as of 1998-04-20.
1029 This should fix a variety of problems, including
1030 those involving some uses of the @code{T} format
1031 specifier, and perhaps some build (porting) problems
1036 @c 1998-03-16: 0.5.22 released.
1037 @heading In 0.5.22 versus 0.5.21:
1039 @ifclear USERVISONLY
1041 Fix code generation for iterative @code{DO} loops that
1042 have one or more references to the iteration variable,
1043 or to aliases of it, in their control expressions.
1044 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1047 @ifclear USERVISONLY
1049 @cindex stack, 387 coprocessor
1053 Fix a code-generation bug that afflicted
1054 Intel x86 targets when @samp{-O2} was specified
1055 compiling, for example, an old version of
1056 the @code{DNRM2} routine.
1058 The x87 coprocessor stack was being
1059 mismanaged in cases involving assigned @code{GOTO}
1063 @ifclear USERVISONLY
1065 Fix @code{DTime} intrinsic so as not to truncate
1066 results to integer values (on some systems).
1070 Fix @code{Signal} intrinsic so it offers portable
1071 support for 64-bit systems (such as Digital Alphas
1074 @ifclear USERVISONLY
1076 Fix run-time crash involving @code{NAMELIST} on 64-bit
1077 machines such as Alphas.
1080 @ifclear USERVISONLY
1082 Fix @code{g77} version of @code{libf2c} so it no longer
1083 produces a spurious @samp{I/O recursion} diagnostic at run time
1084 when an I/O operation (such as @samp{READ *,I}) is interrupted
1085 in a manner that causes the program to be terminated
1086 via the @code{f_exit} routine (such as via @kbd{C-c}).
1089 @ifclear USERVISONLY
1091 Fix @code{g77} crash triggered by @code{CASE} statement with
1092 an omitted lower or upper bound.
1095 @ifclear USERVISONLY
1097 Fix @code{g77} crash compiling references to @code{CPU_Time}
1101 @ifclear USERVISONLY
1103 Fix @code{g77} crash
1104 (or apparently infinite run-time)
1105 when compiling certain complicated expressions
1106 involving @code{COMPLEX} arithmetic
1107 (especially multiplication).
1110 @ifclear USERVISONLY
1112 Fix @code{g77} crash on statements such as
1113 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1114 @samp{Z} is @code{DOUBLE COMPLEX}.
1117 @ifclear USERVISONLY
1119 Fix a @code{g++} crash.
1123 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
1124 compile-time constant @code{INTEGER} expression.
1127 Fix @code{g77} @samp{-g} option so procedures that
1128 use @code{ENTRY} can be stepped through, line by line,
1131 @ifclear USERVISONLY
1133 Fix a profiling-related bug in @code{gcc} back end for
1134 Intel x86 architecture.
1138 Allow any @code{REAL} argument to intrinsics
1139 @code{Second} and @code{CPU_Time}.
1142 Allow any numeric argument to intrinsics
1143 @code{Int2} and @code{Int8}.
1146 Use @code{tempnam}, if available, to open scratch files
1147 (as in @samp{OPEN(STATUS='SCRATCH')})
1148 so that the @code{TMPDIR} environment variable,
1149 if present, is used.
1152 Rename the @code{gcc} keyword @code{restrict} to
1153 @code{__restrict__}, to avoid rejecting valid, existing,
1155 Support for @code{restrict} is now more like support
1158 @ifclear USERVISONLY
1160 Fix @samp{-fpedantic} to not reject procedure invocations
1161 such as @samp{I=J()} and @samp{CALL FOO()}.
1165 Fix @samp{-fugly-comma} to affect invocations of
1166 only external procedures.
1167 Restore rejection of gratuitous trailing omitted
1168 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1171 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
1172 @samp{-fbadu77-intrinsics-*} options.
1174 @ifclear USERVISONLY
1176 Improve diagnostic messages from @code{libf2c}
1177 so it is more likely that the printing of the
1178 active format string is limited to the string,
1179 with no trailing garbage being printed.
1181 (Unlike @code{f2c}, @code{g77} did not append
1182 a null byte to its compiled form of every
1183 format string specified via a @code{FORMAT} statement.
1184 However, @code{f2c} would exhibit the problem
1185 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1186 by printing @samp{(I)garbage} as the format string.)
1189 @ifclear USERVISONLY
1191 Improve compilation of @code{FORMAT} expressions so that
1192 a null byte is appended to the last operand if it
1194 This provides a cleaner run-time diagnostic as provided
1195 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
1198 @ifclear USERVISONLY
1200 Fix various crashes involving code with diagnosed errors.
1203 @ifclear USERVISONLY
1205 Fix cross-compilation bug when configuring @code{libf2c}.
1208 @ifclear USERVISONLY
1210 Improve diagnostics.
1213 @ifclear USERVISONLY
1215 Improve documentation and indexing.
1218 @ifclear USERVISONLY
1220 Upgrade to @code{libf2c} as of 1997-09-23.
1221 This fixes a formatted-I/O bug that afflicted
1222 64-bit systems with 32-bit integers
1223 (such as Digital Alpha running GNU/Linux).
1227 @c 1998-03-18: EGCS 1.0.2 released.
1228 @heading In @code{EGCS} 1.0.2 versus @code{EGCS} 1.0.1:
1230 @ifclear USERVISONLY
1232 Fix @code{g77} crash triggered by @code{CASE} statement with
1233 an omitted lower or upper bound.
1236 @ifclear USERVISONLY
1238 Fix @code{g77} crash on statements such as
1239 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1240 @samp{Z} is @code{DOUBLE COMPLEX}.
1243 @ifclear USERVISONLY
1245 @cindex support, ELF
1246 @cindex -fPIC option
1247 @cindex options, -fPIC
1249 Fix @samp{-fPIC} (such as compiling for ELF targets)
1250 on the Intel x86 architecture target
1251 so invalid assembler code is no longer produced.
1254 @ifclear USERVISONLY
1256 Fix @samp{-fpedantic} to not reject procedure invocations
1257 such as @samp{I=J()} and @samp{CALL FOO()}.
1260 @ifclear USERVISONLY
1262 Fix @samp{-fugly-comma} to affect invocations of
1263 only external procedures.
1264 Restore rejection of gratuitous trailing omitted
1265 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1269 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
1270 @samp{-fbadu77-intrinsics-*} options.
1273 @c 1998-01-06: EGCS 1.0.1 released.
1274 @heading In @code{EGCS} 1.0.1 versus @code{EGCS} 1.0:
1275 @ifclear USERVISONLY
1278 Fix run-time crash involving @code{NAMELIST} on 64-bit
1279 machines such as Alphas.
1283 @c 1997-12-03: EGCS 1.0 released.
1284 @heading In @code{EGCS} 1.0 versus @code{g77} 0.5.21:
1287 Version 1.0 of @code{egcs}
1288 contains several regressions against
1289 version 0.5.21 of @code{g77},
1291 ``vanilla'' @code{gcc} back end instead of patching
1292 it to fix a few bugs and improve performance in a
1295 Features that have been dropped from this version
1296 of @code{g77} due to their being implemented
1297 via @code{g77}-specific patches to the @code{gcc}
1298 back end in previous releases include:
1302 Support for the C-language @code{restrict} keyword.
1305 Support for the @samp{-W} option warning about
1306 integer division by zero.
1309 The Intel x86-specific option @samp{-malign-double}
1310 applying to stack-allocated data
1311 as well as statically-allocate data.
1314 @ifclear USERVISONLY
1315 Note that the @file{gcc/f/gbe/} subdirectory has been removed
1316 from this distribution as a result of @code{g77}
1317 being fully integrated with
1318 the @code{egcs} variant of the @code{gcc} back end.
1321 @ifclear USERVISONLY
1323 Fix code generation for iterative @code{DO} loops that
1324 have one or more references to the iteration variable,
1325 or to aliases of it, in their control expressions.
1326 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1329 @ifclear USERVISONLY
1331 Fix @code{DTime} intrinsic so as not to truncate
1332 results to integer values (on some systems).
1335 @ifclear USERVISONLY
1337 @c Toon Moene discovered these.
1338 Some Fortran code, miscompiled
1339 by @code{g77} built on @code{gcc} version 2.8.1
1340 on m68k-next-nextstep3 configurations
1341 when using the @samp{-O2} option,
1342 is now compiled correctly.
1343 It is believed that a C function known to miscompile
1344 on that configuration
1345 when using the @samp{-O2 -funroll-loops} options
1346 also is now compiled correctly.
1349 @ifclear USERVISONLY
1351 Remove support for non-@code{egcs} versions of @code{gcc}.
1354 @cindex -@w{}-driver option
1355 @cindex @code{g77} options, -@w{}-driver
1356 @cindex options, -@w{}-driver
1358 Remove support for the @samp{--driver} option,
1359 as @code{g77} now does all the driving,
1360 just like @code{gcc}.
1363 Allow any numeric argument to intrinsics
1364 @code{Int2} and @code{Int8}.
1366 @ifclear USERVISONLY
1368 Improve diagnostic messages from @code{libf2c}
1369 so it is more likely that the printing of the
1370 active format string is limited to the string,
1371 with no trailing garbage being printed.
1373 (Unlike @code{f2c}, @code{g77} did not append
1374 a null byte to its compiled form of every
1375 format string specified via a @code{FORMAT} statement.
1376 However, @code{f2c} would exhibit the problem
1377 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1378 by printing @samp{(I)garbage} as the format string.)
1381 @ifclear USERVISONLY
1383 Upgrade to @code{libf2c} as of 1997-09-23.
1384 This fixes a formatted-I/O bug that afflicted
1385 64-bit systems with 32-bit integers
1386 (such as Digital Alpha running GNU/Linux).
1390 @c 1997-09-09: 0.5.21 released.
1393 @ifclear USERVISONLY
1395 Fix a code-generation bug introduced by 0.5.20
1396 caused by loop unrolling (by specifying
1397 @samp{-funroll-loops} or similar).
1398 This bug afflicted all code compiled by
1399 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1400 Fortran, and so on).
1403 @ifclear USERVISONLY
1405 Fix a code-generation bug manifested when
1406 combining local @code{EQUIVALENCE} with a
1407 @code{DATA} statement that follows
1408 the first executable statement (or is
1409 treated as an executable-context statement
1410 as a result of using the @samp{-fpedantic}
1414 @ifclear USERVISONLY
1416 Fix a compiler crash that occured when an
1417 integer division by a constant zero is detected.
1418 Instead, when the @samp{-W} option is specified,
1419 the @code{gcc} back end issues a warning about such a case.
1420 This bug afflicted all code compiled by
1421 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1422 Fortran, and so on).
1426 When the @samp{-W} option is specified, @code{gcc}, @code{g77},
1427 and other GNU compilers that incorporate the @code{gcc}
1428 back end as modified by @code{g77}, issue
1429 a warning about integer division by constant zero.
1432 @ifclear USERVISONLY
1434 Fix a compiler crash that occurred in some cases
1435 of procedure inlining.
1436 (Such cases became more frequent in 0.5.20.)
1439 @ifclear USERVISONLY
1441 Fix a compiler crash resulting from using @code{DATA}
1442 or similar to initialize a @code{COMPLEX} variable or
1446 @ifclear USERVISONLY
1448 Fix compiler crashes involving use of @code{AND}, @code{OR},
1449 or @code{XOR} intrinsics.
1452 @ifclear USERVISONLY
1454 Fix compiler bug triggered when using a @code{COMMON}
1455 or @code{EQUIVALENCE} variable
1456 as the target of an @code{ASSIGN}
1457 or assigned-@code{GOTO} statement.
1460 @ifclear USERVISONLY
1462 Fix compiler crashes due to using the name of a some
1463 non-standard intrinsics (such as @code{FTell} or
1464 @code{FPutC}) as such and as the name of a procedure
1466 Such dual use of a name in a program is allowed by
1470 @c @code{g77}'s version of @code{libf2c} has been modified
1471 @c so that the external names of library's procedures do not
1472 @c conflict with names used for Fortran procedures compiled
1474 @c An additional layer of jacket procedures has been added
1475 @c to @code{libf2c} to map the old names to the new names,
1476 @c for automatic use by programs that interface to the
1477 @c library procedures via the external-procedure mechanism.
1479 @c For example, the intrinsic @code{FPUTC} previously was
1480 @c implemented by @code{g77} as a call to the @code{libf2c}
1481 @c routine @code{fputc_}.
1482 @c This would conflict with a Fortran procedure named @code{FPUTC}
1483 @c (using default compiler options), and this conflict
1484 @c would cause a crash under certain circumstances.
1486 @c Now, the intrinsic @code{FPUTC} calls @code{G77_fputc_0},
1487 @c which does not conflict with the @code{fputc_} external
1488 @c that implements a Fortran procedure named @code{FPUTC}.
1490 @c Programs that refer to @code{FPUTC} as an external procedure
1491 @c without supplying their own implementation will link to
1492 @c the new @code{libf2c} routine @code{fputc_}, which is
1493 @c simply a jacket routine that calls @code{G77_fputc_0}.
1495 @ifclear USERVISONLY
1497 Place automatic arrays on the stack, even if
1498 @code{SAVE} or the @samp{-fno-automatic} option
1500 This avoids a compiler crash in some cases.
1503 @ifclear USERVISONLY
1505 The @samp{-malign-double} option now reliably aligns
1506 @code{DOUBLE PRECISION} optimally on Pentium and
1507 Pentium Pro architectures (586 and 686 in @code{gcc}).
1511 New option @samp{-Wno-globals} disables warnings
1512 about ``suspicious'' use of a name both as a global
1513 name and as the implicit name of an intrinsic, and
1514 warnings about disagreements over the number or natures of
1515 arguments passed to global procedures, or the
1516 natures of the procedures themselves.
1518 The default is to issue such warnings, which are
1519 new as of this version of @code{g77}.
1522 New option @samp{-fno-globals} disables diagnostics
1523 about potentially fatal disagreements
1524 analysis problems, such as disagreements over the
1525 number or natures of arguments passed to global
1526 procedures, or the natures of those procedures themselves.
1528 The default is to issue such diagnostics and flag
1529 the compilation as unsuccessful.
1530 With this option, the diagnostics are issued as
1531 warnings, or, if @samp{-Wno-globals} is specified,
1532 are not issued at all.
1534 This option also disables inlining of global procedures,
1535 to avoid compiler crashes resulting from coding errors
1536 that these diagnostics normally would identify.
1538 @ifclear USERVISONLY
1540 Diagnose cases where a reference to a procedure
1541 disagrees with the type of that procedure, or
1542 where disagreements about the number or nature
1544 This avoids a compiler crash.
1547 @ifclear USERVISONLY
1549 Fix parsing bug whereby @code{g77} rejected a
1550 second initialization specification immediately
1551 following the first's closing @samp{/} without
1552 an intervening comma in a @code{DATA} statement,
1553 and the second specification was an implied-DO list.
1556 @ifclear USERVISONLY
1558 Improve performance of the @code{gcc} back end so
1559 certain complicated expressions involving @code{COMPLEX}
1560 arithmetic (especially multiplication) don't appear to
1561 take forever to compile.
1564 @ifclear USERVISONLY
1566 Fix a couple of profiling-related bugs in @code{gcc}
1570 @ifclear USERVISONLY
1572 Integrate GNU Ada's (GNAT's) changes to the back end,
1573 which consist almost entirely of bug fixes.
1574 These fixes are circa version 3.10p of GNAT.
1577 @ifclear USERVISONLY
1579 Include some other @code{gcc} fixes that seem useful in
1580 @code{g77}'s version of @code{gcc}.
1581 (See @file{gcc/ChangeLog} for details---compare it
1582 to that file in the vanilla @code{gcc-2.7.2.3.tar.gz}
1587 Fix @code{libU77} routines that accept file and other names
1588 to strip trailing blanks from them, for consistency
1589 with other implementations.
1590 Blanks may be forcibly appended to such names by
1591 appending a single null character (@samp{CHAR(0)})
1592 to the significant trailing blanks.
1595 Fix @code{CHMOD} intrinsic to work with file names
1596 that have embedded blanks, commas, and so on.
1599 Fix @code{SIGNAL} intrinsic so it accepts an
1600 optional third @code{Status} argument.
1602 @ifclear USERVISONLY
1604 Fix @code{IDATE()} intrinsic subroutine (VXT form)
1605 so it accepts arguments in the correct order.
1606 Documentation fixed accordingly, and for
1607 @code{GMTIME()} and @code{LTIME()} as well.
1611 Make many changes to @code{libU77} intrinsics to
1612 support existing code more directly.
1614 Such changes include allowing both subroutine and
1615 function forms of many routines, changing @code{MCLOCK()}
1616 and @code{TIME()} to return @code{INTEGER(KIND=1)} values,
1617 introducing @code{MCLOCK8()} and @code{TIME8()} to
1618 return @code{INTEGER(KIND=2)} values,
1619 and placing functions that are intended to perform
1620 side effects in a new intrinsic group, @code{badu77}.
1622 @ifclear USERVISONLY
1624 Improve @code{libU77} so it is more portable.
1628 Add options @samp{-fbadu77-intrinsics-delete},
1629 @samp{-fbadu77-intrinsics-hide}, and so on.
1631 @ifclear USERVISONLY
1633 Fix crashes involving diagnosed or invalid code.
1636 @ifclear USERVISONLY
1638 @code{g77} and @code{gcc} now do a somewhat better
1639 job detecting and diagnosing arrays that are too
1640 large to handle before these cause diagnostics
1641 during the assembler or linker phase, a compiler
1642 crash, or generation of incorrect code.
1645 @ifclear USERVISONLY
1647 Make some fixes to alias analysis code.
1650 @ifclear USERVISONLY
1652 Add support for @code{restrict} keyword in @code{gcc}
1656 @ifclear USERVISONLY
1658 Support @code{gcc} version 2.7.2.3
1659 (modified by @code{g77} into version 2.7.2.3.f.1),
1661 support for prior versions of @code{gcc}.
1664 @ifclear USERVISONLY
1666 Incorporate GNAT's patches to the @code{gcc} back
1667 end into @code{g77}'s, so GNAT users do not need
1668 to apply GNAT's patches to build both GNAT and @code{g77}
1669 from the same source tree.
1672 @ifclear USERVISONLY
1674 Modify @code{make} rules and related code so that
1675 generation of Info documentation doesn't require
1676 compilation using @code{gcc}.
1677 Now, any ANSI C compiler should be adequate to
1678 produce the @code{g77} documentation (in particular,
1679 the tables of intrinsics) from scratch.
1683 Add @code{INT2} and @code{INT8} intrinsics.
1686 Add @code{CPU_TIME} intrinsic.
1689 Add @code{ALARM} intrinsic.
1692 @code{CTIME} intrinsic now accepts any @code{INTEGER}
1693 argument, not just @code{INTEGER(KIND=2)}.
1695 @ifclear USERVISONLY
1697 Warn when explicit type declaration disagrees with
1698 the type of an intrinsic invocation.
1701 @ifclear USERVISONLY
1703 Support @samp{*f771} entry in @code{gcc} @file{specs} file.
1706 @ifclear USERVISONLY
1708 Fix typo in @code{make} rule @code{g77-cross}, used only for
1712 @ifclear USERVISONLY
1714 Fix @code{libf2c} build procedure to re-archive library
1715 if previous attempt to archive was interrupted.
1718 @ifclear USERVISONLY
1720 Change @code{gcc} to unroll loops only during the last
1721 invocation (of as many as two invocations) of loop
1725 @ifclear USERVISONLY
1727 Improve handling of @samp{-fno-f2c} so that code that
1728 attempts to pass an intrinsic as an actual argument,
1729 such as @samp{CALL FOO(ABS)}, is rejected due to the fact
1730 that the run-time-library routine is, effectively,
1731 compiled with @samp{-ff2c} in effect.
1734 @ifclear USERVISONLY
1736 Fix @code{g77} driver to recognize @samp{-fsyntax-only}
1737 as an option that inhibits linking, just like @samp{-c} or
1738 @samp{-S}, and to recognize and properly handle the
1739 @samp{-nostdlib}, @samp{-M}, @samp{-MM}, @samp{-nodefaultlibs},
1740 and @samp{-Xlinker} options.
1743 @ifclear USERVISONLY
1745 Upgrade to @code{libf2c} as of 1997-08-16.
1748 @ifclear USERVISONLY
1750 Modify @code{libf2c} to consistently and clearly diagnose
1751 recursive I/O (at run time).
1755 @code{g77} driver now prints version information (such as produced
1756 by @kbd{g77 -v}) to @code{stderr} instead of @code{stdout}.
1759 The @samp{.r} suffix now designates a Ratfor source file,
1760 to be preprocessed via the @code{ratfor} command, available
1763 @ifclear USERVISONLY
1765 Fix some aspects of how @code{gcc} determines what kind of
1766 system is being configured and what kinds are supported.
1767 For example, GNU Linux/Alpha ELF systems now are directly
1771 @ifclear USERVISONLY
1773 Improve diagnostics.
1776 @ifclear USERVISONLY
1778 Improve documentation and indexing.
1781 @ifclear USERVISONLY
1783 Include all pertinent files for @code{libf2c} that come
1784 from @code{netlib.bell-labs.com}; give any such files
1785 that aren't quite accurate in @code{g77}'s version of
1786 @code{libf2c} the suffix @samp{.netlib}.
1789 @ifclear USERVISONLY
1791 Reserve @code{INTEGER(KIND=0)} for future use.
1795 @c 1997-02-28: 0.5.20 released.
1799 The @samp{-fno-typeless-boz} option is now the default.
1801 This option specifies that non-decimal-radix
1802 constants using the prefixed-radix form (such as @samp{Z'1234'})
1803 are to be interpreted as @code{INTEGER(KIND=1)} constants.
1804 Specify @samp{-ftypeless-boz} to cause such
1805 constants to be interpreted as typeless.
1807 (Version 0.5.19 introduced @samp{-fno-typeless-boz} and
1811 @xref{Fortran Dialect Options,,Options Controlling Fortran Dialect},
1812 for information on the @samp{-ftypeless-boz} option.
1816 Options @samp{-ff90-intrinsics-enable} and
1817 @samp{-fvxt-intrinsics-enable} now are the
1820 Some programs might use names that clash with
1821 intrinsic names defined (and now enabled) by these
1822 options or by the new @code{libU77} intrinsics.
1823 Users of such programs might need to compile them
1824 differently (using, for example, @samp{-ff90-intrinsics-disable})
1825 or, better yet, insert appropriate @code{EXTERNAL}
1826 statements specifying that these names are not intended
1827 to be names of intrinsics.
1830 The @code{ALWAYS_FLUSH} macro is no longer defined when
1831 building @code{libf2c}, which should result in improved
1832 I/O performance, especially over NFS.
1834 @emph{Note:} If you have code that depends on the behavior
1835 of @code{libf2c} when built with @code{ALWAYS_FLUSH} defined,
1836 you will have to modify @code{libf2c} accordingly before
1837 building it from this and future versions of @code{g77}.
1840 @xref{Output Assumed To Flush}, for more information.
1844 Dave Love's implementation of @code{libU77} has been
1845 added to the version of @code{libf2c} distributed with
1846 and built as part of @code{g77}.
1847 @code{g77} now knows about the routines in this library
1851 New option @samp{-fvxt} specifies that the
1852 source file is written in VXT Fortran, instead of GNU Fortran.
1855 @xref{VXT Fortran}, for more information on the constructs
1856 recognized when the @samp{-fvxt} option is specified.
1860 The @samp{-fvxt-not-f90} option has been deleted,
1861 along with its inverse, @samp{-ff90-not-vxt}.
1863 If you used one of these deleted options, you should
1864 re-read the pertinent documentation to determine which
1865 options, if any, are appropriate for compiling your
1866 code with this version of @code{g77}.
1869 @xref{Other Dialects}, for more information.
1873 The @samp{-fugly} option now issues a warning, as it
1874 likely will be removed in a future version.
1876 (Enabling all the @samp{-fugly-*} options is unlikely
1877 to be feasible, or sensible, in the future,
1878 so users should learn to specify only those
1879 @samp{-fugly-*} options they really need for a
1880 particular source file.)
1883 The @samp{-fugly-assumed} option, introduced in
1884 version 0.5.19, has been changed to
1885 better accommodate old and new code.
1888 @xref{Ugly Assumed-Size Arrays}, for more information.
1891 @ifclear USERVISONLY
1893 Make a number of fixes to the @code{g77} front end and
1894 the @code{gcc} back end to better support Alpha (AXP)
1896 This includes providing at least one bug-fix to the
1897 @code{gcc} back end for Alphas.
1901 Related to supporting Alpha (AXP) machines, the @code{LOC()}
1902 intrinsic and @code{%LOC()} construct now return
1903 values of @code{INTEGER(KIND=0)} type,
1904 as defined by the GNU Fortran language.
1906 This type is wide enough
1907 (holds the same number of bits)
1908 as the character-pointer type on the machine.
1910 On most machines, this won't make a difference,
1911 whereas, on Alphas and other systems with 64-bit pointers,
1912 the @code{INTEGER(KIND=0)} type is equivalent to @code{INTEGER(KIND=2)}
1913 (often referred to as @code{INTEGER*8})
1914 instead of the more common @code{INTEGER(KIND=1)}
1915 (often referred to as @code{INTEGER*4}).
1918 Emulate @code{COMPLEX} arithmetic in the @code{g77} front
1919 end, to avoid bugs in @code{complex} support in the
1920 @code{gcc} back end.
1921 New option @samp{-fno-emulate-complex}
1922 causes @code{g77} to revert the 0.5.19 behavior.
1924 @ifclear USERVISONLY
1926 Fix bug whereby @samp{REAL A(1)}, for example, caused
1927 a compiler crash if @samp{-fugly-assumed} was in effect
1928 and @var{A} was a local (automatic) array.
1929 That case is no longer affected by the new
1930 handling of @samp{-fugly-assumed}.
1933 @ifclear USERVISONLY
1935 Fix @code{g77} command driver so that @samp{g77 -o foo.f}
1936 no longer deletes @file{foo.f} before issuing other
1937 diagnostics, and so the @samp{-x} option is properly
1941 @ifclear USERVISONLY
1943 Enable inlining of subroutines and functions by the @code{gcc}
1945 This works as it does for @code{gcc} itself---program units
1946 may be inlined for invocations that follow them in the same
1947 program unit, as long as the appropriate compile-time
1948 options are specified.
1952 Dummy arguments are no longer assumed to potentially alias
1954 other dummy arguments or @code{COMMON} areas when any of
1955 these are defined (assigned to) by Fortran code.
1957 This can result in faster and/or smaller programs when
1958 compiling with optimization enabled, though on some
1959 systems this effect is observed only when @samp{-fforce-addr}
1962 New options @samp{-falias-check}, @samp{-fargument-alias},
1963 @samp{-fargument-noalias},
1964 and @samp{-fno-argument-noalias-global} control the
1965 way @code{g77} handles potential aliasing.
1968 @xref{Aliasing Assumed To Work}, for detailed information on why the
1969 new defaults might result in some programs no longer working the way they
1970 did when compiled by previous versions of @code{g77}.
1973 @ifclear USERVISONLY
1975 The @code{CONJG()} and @code{DCONJG()} intrinsics now
1976 are compiled in-line.
1979 @ifclear USERVISONLY
1981 The bug-fix for 0.5.19.1 has been re-done.
1982 The @code{g77} compiler has been changed back to
1983 assume @code{libf2c} has no aliasing problems in
1984 its implementations of the @code{COMPLEX} (and
1985 @code{DOUBLE COMPLEX}) intrinsics.
1986 The @code{libf2c} has been changed to have no such
1989 As a result, 0.5.20 is expected to offer improved performance
1990 over 0.5.19.1, perhaps as good as 0.5.19 in most
1991 or all cases, due to this change alone.
1993 @emph{Note:} This change requires version 0.5.20 of
1994 @code{libf2c}, at least, when linking code produced
1995 by any versions of @code{g77} other than 0.5.19.1.
1996 Use @samp{g77 -v} to determine the version numbers
1997 of the @code{libF77}, @code{libI77}, and @code{libU77}
1998 components of the @code{libf2c} library.
1999 (If these version numbers are not printed---in
2000 particular, if the linker complains about unresolved
2001 references to names like @samp{g77__fvers__}---that
2002 strongly suggests your installation has an obsolete
2003 version of @code{libf2c}.)
2007 New option @samp{-fugly-assign} specifies that the
2008 same memory locations are to be used to hold the
2009 values assigned by both statements @samp{I = 3} and
2010 @samp{ASSIGN 10 TO I}, for example.
2011 (Normally, @code{g77} uses a separate memory location
2012 to hold assigned statement labels.)
2015 @xref{Ugly Assigned Labels}, for more information.
2019 @code{FORMAT} and @code{ENTRY} statements now are allowed to
2020 precede @code{IMPLICIT NONE} statements.
2022 @ifclear USERVISONLY
2024 Produce diagnostic for unsupported @code{SELECT CASE} on
2025 @code{CHARACTER} type, instead of crashing, at compile time.
2028 @ifclear USERVISONLY
2030 Fix crashes involving diagnosed or invalid code.
2033 @ifclear USERVISONLY
2035 Change approach to building @code{libf2c} archive
2036 (@file{libf2c.a}) so that members are added to it
2037 only when truly necessary, so the user that installs
2038 an already-built @code{g77} doesn't need to have write
2039 access to the build tree (whereas the user doing the
2040 build might not have access to install new software
2044 @ifclear USERVISONLY
2046 Support @code{gcc} version 2.7.2.2
2047 (modified by @code{g77} into version 2.7.2.2.f.2),
2049 support for prior versions of @code{gcc}.
2052 @ifclear USERVISONLY
2054 Upgrade to @code{libf2c} as of 1997-02-08, and
2055 fix up some of the build procedures.
2058 @ifclear USERVISONLY
2060 Improve general build procedures for @code{g77},
2061 fixing minor bugs (such as deletion of any file
2062 named @file{f771} in the parent directory of @code{gcc/}).
2066 Enable full support of @code{INTEGER(KIND=2)}
2067 (often referred to as @code{INTEGER*8})
2069 @code{libf2c} and @file{f2c.h} so that @code{f2c} users
2070 may make full use of its features via the @code{g77}
2071 version of @file{f2c.h} and the @code{INTEGER(KIND=2)}
2072 support routines in the @code{g77} version of @code{libf2c}.
2075 Improve @code{g77} driver and @code{libf2c} so that @samp{g77 -v}
2076 yields version information on the library.
2079 The @code{SNGL} and @code{FLOAT} intrinsics now are
2080 specific intrinsics, instead of synonyms for the
2081 generic intrinsic @code{REAL}.
2084 New intrinsics have been added.
2085 These are @code{REALPART}, @code{IMAGPART},
2087 @code{LONG}, and @code{SHORT}.
2090 A new group of intrinsics, @code{gnu}, has been added
2091 to contain the new @code{REALPART}, @code{IMAGPART},
2092 and @code{COMPLEX} intrinsics.
2093 An old group, @code{dcp}, has been removed.
2096 Complain about industry-wide ambiguous references
2097 @samp{REAL(@var{expr})} and @samp{AIMAG(@var{expr})},
2098 where @var{expr} is @code{DOUBLE COMPLEX} (or any
2099 complex type other than @code{COMPLEX}), unless
2100 @samp{-ff90} option specifies Fortran 90 interpretation
2101 or new @samp{-fugly-complex} option, in conjunction with
2102 @samp{-fnot-f90}, specifies @code{f2c} interpretation.
2104 @ifclear USERVISONLY
2106 Make improvements to diagnostics.
2109 @ifclear USERVISONLY
2111 Speed up compiler a bit.
2114 @ifclear USERVISONLY
2116 Improvements to documentation and indexing, including
2117 a new chapter containing information on one, later
2118 more, diagnostics that users are directed to pull
2119 up automatically via a message in the diagnostic itself.
2121 (Hence the menu item @code{M} for the node
2122 @code{Diagnostics} in the top-level menu of
2123 the Info documentation.)
2127 @ifclear DOC-OLDNEWS
2128 @heading In previous versions:
2130 Information on previous versions is archived
2131 in @file{@value{path-g77}/news.texi}
2132 following the test of the @code{DOC-OLDNEWS} macro.
2136 @c 1997-02-01: 0.5.19.1 released.
2137 @heading In 0.5.19.1:
2140 Code-generation bugs afflicting operations on complex
2141 data have been fixed.
2143 These bugs occurred when assigning the result of an
2144 operation to a complex variable (or array element)
2145 that also served as an input to that operation.
2147 The operations affected by this bug were: @code{CONJG()},
2148 @code{DCONJG()}, @code{CCOS()}, @code{CDCOS()},
2149 @code{CLOG()}, @code{CDLOG()}, @code{CSIN()}, @code{CDSIN()},
2150 @code{CSQRT()}, @code{CDSQRT()}, complex division, and
2151 raising a @code{DOUBLE COMPLEX} operand to an @code{INTEGER}
2153 (The related generic and @samp{Z}-prefixed intrinsics,
2154 such as @code{ZSIN()}, also were affected.)
2156 For example, @samp{C = CSQRT(C)}, @samp{Z = Z/C}, and @samp{Z = Z**I}
2157 (where @samp{C} is @code{COMPLEX} and @samp{Z} is
2158 @code{DOUBLE COMPLEX}) have been fixed.
2161 @c 1996-12-07: 0.5.19 released.
2165 Fix @code{FORMAT} statement parsing so negative values for
2166 specifiers such as @code{P} (e.g. @samp{FORMAT(-1PF8.1)})
2167 are correctly processed as negative.
2170 Fix @code{SIGNAL} intrinsic so it once again accepts a
2171 procedure as its second argument.
2174 A temporary kludge option provides bare-bones information on
2175 @code{COMMON} and @code{EQUIVALENCE} members at debug time.
2178 New @samp{-fonetrip} option specifies FORTRAN-66-style
2179 one-trip @code{DO} loops.
2182 New @samp{-fno-silent} option causes names of program units
2183 to be printed as they are compiled, in a fashion similar to
2184 UNIX @code{f77} and @code{f2c}.
2187 New @samp{-fugly-assumed} option specifies that arrays
2188 dimensioned via @samp{DIMENSION X(1)}, for example, are to be
2189 treated as assumed-size.
2192 New @samp{-fno-typeless-boz} option specifies that non-decimal-radix
2193 constants using the prefixed-radix form (such as @samp{Z'1234'})
2194 are to be interpreted as @code{INTEGER} constants.
2197 New @samp{-ff66} option is a ``shorthand'' option that specifies
2198 behaviors considered appropriate for FORTRAN 66 programs.
2201 New @samp{-ff77} option is a ``shorthand'' option that specifies
2202 behaviors considered appropriate for UNIX @code{f77} programs.
2205 New @samp{-fugly-comma} and @samp{-fugly-logint} options provided
2206 to perform some of what @samp{-fugly} used to do.
2207 @samp{-fugly} and @samp{-fno-ugly} are now ``shorthand'' options,
2208 in that they do nothing more than enable (or disable) other
2209 @samp{-fugly-*} options.
2212 Fix parsing of assignment statements involving targets that
2213 are substrings of elements of @code{CHARACTER} arrays having
2214 names such as @samp{READ}, @samp{WRITE}, @samp{GOTO}, and
2215 @samp{REALFUNCTIONFOO}.
2218 Fix crashes involving diagnosed code.
2221 Fix handling of local @code{EQUIVALENCE} areas so certain cases
2222 of valid Fortran programs are not misdiagnosed as improperly
2223 extending the area backwards.
2226 Support @code{gcc} version 2.7.2.1.
2229 Upgrade to @code{libf2c} as of 1996-09-26, and
2230 fix up some of the build procedures.
2233 Change code generation for list-directed I/O so it allows
2234 for new versions of @code{libf2c} that might return non-zero
2235 status codes for some operations previously assumed to always
2238 This change not only affects how @code{IOSTAT=} variables
2239 are set by list-directed I/O, it also affects whether
2240 @code{END=} and @code{ERR=} labels are reached by these
2244 Add intrinsic support for new @code{FTELL} and @code{FSEEK}
2245 procedures in @code{libf2c}.
2248 Modify @code{fseek_()} in @code{libf2c} to be more portable
2249 (though, in practice, there might be no systems where this
2250 matters) and to catch invalid @code{whence} arguments.
2253 Some useless warnings from the @samp{-Wunused} option have
2257 Fix a problem building the @file{f771} executable
2258 on AIX systems by linking with the @samp{-bbigtoc} option.
2261 Abort configuration if @code{gcc} has not been patched
2262 using the patch file provided in the @file{gcc/f/gbe/}
2266 Add options @samp{--help} and @samp{--version} to the
2267 @code{g77} command, to conform to GNU coding guidelines.
2268 Also add printing of @code{g77} version number when
2269 the @samp{--verbose} (@samp{-v}) option is used.
2272 Change internally generated name for local @code{EQUIVALENCE}
2273 areas to one based on the alphabetically sorted first name
2274 in the list of names for entities placed at the beginning
2278 Improvements to documentation and indexing.
2281 @c 1996-04-01: 0.5.18 released.
2285 Add some rudimentary support for @code{INTEGER*1},
2286 @code{INTEGER*2}, @code{INTEGER*8},
2287 and their @code{LOGICAL} equivalents.
2288 (This support works on most, maybe all, @code{gcc} targets.)
2290 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2291 for providing the patch for this!
2293 Among the missing elements from the support for these
2294 features are full intrinsic support and constants.
2297 Add some rudimentary support for the @code{BYTE} and
2298 @code{WORD} type-declaration statements.
2299 @code{BYTE} corresponds to @code{INTEGER*1},
2300 while @code{WORD} corresponds to @code{INTEGER*2}.
2302 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2303 for providing the patch for this!
2306 The compiler code handling intrinsics has been largely
2307 rewritten to accommodate the new types.
2308 No new intrinsics or arguments for existing
2309 intrinsics have been added, so there is, at this
2310 point, no intrinsic to convert to @code{INTEGER*8},
2314 Support automatic arrays in procedures.
2317 Reduce space/time requirements for handling large
2318 @emph{sparsely} initialized aggregate arrays.
2319 This improvement applies to only a subset of
2320 the general problem to be addressed in 0.6.
2323 Treat initial values of zero as if they weren't
2324 specified (in DATA and type-declaration statements).
2325 The initial values will be set to zero anyway, but the amount
2326 of compile time processing them will be reduced,
2327 in some cases significantly (though, again, this
2328 is only a subset of the general problem to be
2331 A new option, @samp{-fzeros}, is introduced to
2332 enable the traditional treatment of zeros as any
2336 With @samp{-ff90} in force, @code{g77} incorrectly
2337 interpreted @samp{REAL(Z)} as returning a @code{REAL}
2338 result, instead of as a @code{DOUBLE PRECISION}
2340 (Here, @samp{Z} is @code{DOUBLE COMPLEX}.)
2342 With @samp{-fno-f90} in force, the interpretation remains
2343 unchanged, since this appears to be how at least some
2344 F77 code using the @code{DOUBLE COMPLEX} extension expected
2347 Essentially, @samp{REAL(Z)} in F90 is the same as
2348 @samp{DBLE(Z)}, while in extended F77, it appears to
2349 be the same as @samp{REAL(REAL(Z))}.
2352 An expression involving exponentiation, where both operands
2353 were type @code{INTEGER} and the right-hand operand
2354 was negative, was erroneously evaluated.
2357 Fix bugs involving @code{DATA} implied-@code{DO} constructs
2358 (these involved an errant diagnostic and a crash, both on good
2359 code, one involving subsequent statement-function definition).
2362 Close @code{INCLUDE} files after processing them, so compiling source
2363 files with lots of @code{INCLUDE} statements does not result in
2364 being unable to open @code{INCLUDE} files after all the available
2365 file descriptors are used up.
2368 Speed up compiling, especially of larger programs, and perhaps
2369 slightly reduce memory utilization while compiling (this is
2370 @emph{not} the improvement planned for 0.6 involving large aggregate
2371 areas)---these improvements result from simply turning
2372 off some low-level code to do self-checking that hasn't been
2373 triggered in a long time.
2376 Introduce three new options that
2377 implement optimizations in the @code{gcc} back end (GBE).
2378 These options are @samp{-fmove-all-movables}, @samp{-freduce-all-givs},
2379 and @samp{-frerun-loop-opt}, which are enabled, by default,
2380 for Fortran compilations.
2381 These optimizations are intended to help toon Fortran programs.
2384 Patch the GBE to do a better job optimizing certain
2385 kinds of references to array elements.
2388 Due to patches to the GBE, the version number of @code{gcc}
2389 also is patched to make it easier to manage installations,
2390 especially useful if it turns out a @code{g77} change to the
2393 The @code{g77}-modified version number is the @code{gcc}
2394 version number with the string @samp{.f.@var{n}} appended,
2395 where @samp{f} identifies the version as enhanced for
2396 Fortran, and @var{n} is @samp{1} for the first Fortran
2397 patch for that version of @code{gcc}, @samp{2} for the
2400 So, this introduces version 2.7.2.f.1 of @code{gcc}.
2403 Make several improvements and fixes to diagnostics, including
2404 the removal of two that were inappropriate or inadequate.
2407 Warning about two successive arithmetic operators, produced
2408 by @samp{-Wsurprising}, now produced @emph{only} when both
2409 operators are, indeed, arithmetic (not relational/boolean).
2412 @samp{-Wsurprising} now warns about the remaining cases
2413 of using non-integral variables for implied-@code{DO}
2414 loops, instead of these being rejected unless @samp{-fpedantic}
2415 or @samp{-fugly} specified.
2418 Allow @code{SAVE} of a local variable or array, even after
2419 it has been given an initial value via @code{DATA}, for example.
2422 Introduce an Info version of @code{g77} documentation, which
2423 supercedes @file{gcc/f/CREDITS}, @file{gcc/f/DOC}, and
2424 @file{gcc/f/PROJECTS}.
2425 These files will be removed in a future release.
2426 The files @file{gcc/f/BUGS}, @file{gcc/f/INSTALL}, and
2427 @file{gcc/f/NEWS} now are automatically built from
2428 the texinfo source when distributions are made.
2430 This effort was inspired by a first pass at translating
2431 @file{g77-0.5.16/f/DOC} that was contributed to Craig by
2432 David Ronis (@email{ronis@@onsager.chem.mcgill.ca}).
2435 New @samp{-fno-second-underscore} option to specify
2436 that, when @samp{-funderscoring} is in effect, a second
2437 underscore is not to be appended to Fortran names already
2438 containing an underscore.
2441 Change the way iterative @code{DO} loops work to follow
2443 In particular, calculation of the iteration count is
2444 still done by converting the start, end, and increment
2445 parameters to the type of the @code{DO} variable, but
2446 the result of the calculation is always converted to
2447 the default @code{INTEGER} type.
2449 (This should have no effect on existing code compiled
2450 by @code{g77}, but code written to assume that use
2451 of a @emph{wider} type for the @code{DO} variable
2452 will result in an iteration count being fully calculated
2453 using that wider type (wider
2454 than default @code{INTEGER}) must be rewritten.)
2457 Support @code{gcc} version 2.7.2.
2460 Upgrade to @code{libf2c} as of 1996-03-23, and
2461 fix up some of the build procedures.
2463 Note that the email addresses related to @code{f2c}
2464 have changed---the distribution site now is
2465 named @code{netlib.bell-labs.com}, and the
2466 maintainer's new address is @email{dmg@@bell-labs.com}.
2469 @c 1995-11-18: 0.5.17 released.
2473 @strong{Fix serious bug} in @samp{g77 -v} command that can cause removal of a
2474 system's @file{/dev/null} special file if run by user @code{root}.
2476 @strong{All users} of version 0.5.16 should ensure that
2477 they have not removed @file{/dev/null} or replaced it with an ordinary
2478 file (e.g. by comparing the output of @samp{ls -l /dev/null} with
2479 @samp{ls -l /dev/zero}.
2480 If the output isn't basically the
2481 same, contact your system
2482 administrator about restoring @file{/dev/null} to its proper status).
2484 This bug is particularly insidious because removing @file{/dev/null} as
2485 a special file can go undetected for quite a while, aside from
2486 various applications and programs exhibiting sudden, strange
2489 I sincerely apologize for not realizing the
2490 implications of the fact that when @samp{g77 -v} runs the @code{ld} command
2491 with @samp{-o /dev/null} that @code{ld} tries to @emph{remove} the executable
2492 it is supposed to build (especially if it reports unresolved
2493 references, which it should in this case)!
2496 Fix crash on @samp{CHARACTER*(*) FOO} in a main or block data program unit.
2499 Fix crash that can occur when diagnostics given outside of any
2500 program unit (such as when input file contains @samp{@@foo}).
2503 Fix crashes, infinite loops (hangs), and such involving diagnosed code.
2506 Fix @code{ASSIGN}'ed variables so they can be @code{SAVE}'d or dummy arguments,
2507 and issue clearer error message in cases where target of @code{ASSIGN}
2508 or @code{ASSIGN}ed @code{GOTO}/@code{FORMAT} is too small (which should
2512 Make @code{libf2c} build procedures work on more systems again by
2513 eliminating unnecessary invocations of @samp{ld -r -x} and @code{mv}.
2516 Fix omission of @samp{-funix-intrinsics-@dots{}} options in list of permitted
2517 options to compiler.
2520 Fix failure to always diagnose missing type declaration for
2521 @code{IMPLICIT NONE}.
2524 Fix compile-time performance problem (which could sometimes
2525 crash the compiler, cause a hang, or whatever, due to a bug
2526 in the back end) involving exponentiation with a large @code{INTEGER}
2527 constant for the right-hand operator (e.g. @samp{I**32767}).
2530 Fix build procedures so cross-compiling @code{g77} (the @code{fini}
2531 utility in particular) is properly built using the host compiler.
2534 Add new @samp{-Wsurprising} option to warn about constructs that are
2535 interpreted by the Fortran standard (and @code{g77}) in ways that
2536 are surprising to many programmers.
2539 Add @code{ERF()} and @code{ERFC()} as generic intrinsics mapping to existing
2540 @code{ERF}/@code{DERF} and @code{ERFC}/@code{DERFC} specific intrinsics.
2542 @emph{Note:} You should
2543 specify @samp{INTRINSIC ERF,ERFC} in any code where you might use
2544 these as generic intrinsics, to improve likelihood of diagnostics
2545 (instead of subtle run-time bugs) when using a compiler that
2546 doesn't support these as intrinsics (e.g. @code{f2c}).
2549 Remove from @samp{-fno-pedantic} the diagnostic about @code{DO}
2550 with non-@code{INTEGER} index variable; issue that under
2551 @samp{-Wsurprising} instead.
2554 Clarify some diagnostics that say things like ``ignored'' when that's
2558 Clarify diagnostic on use of @code{.EQ.}/@code{.NE.} on @code{LOGICAL}
2562 Minor improvements to code generation for various operations on
2563 @code{LOGICAL} operands.
2566 Minor improvement to code generation for some @code{DO} loops on some
2570 Support @code{gcc} version 2.7.1.
2573 Upgrade to @code{libf2c} as of 1995-11-15.
2576 @c 1995-08-30: 0.5.16 released.
2580 Fix a code-generation bug involving complicated @code{EQUIVALENCE} statements
2581 not involving @code{COMMON}.
2584 Fix code-generation bugs involving invoking ``gratis'' library procedures
2585 in @code{libf2c} from code compiled with @samp{-fno-f2c} by making these
2586 procedures known to @code{g77} as intrinsics (not affected by -fno-f2c).
2587 This is known to fix code invoking @code{ERF()}, @code{ERFC()},
2588 @code{DERF()}, and @code{DERFC()}.
2591 Update @code{libf2c} to include netlib patches through 1995-08-16, and
2592 @code{#define} @code{WANT_LEAD_0} to 1 to make @code{g77}-compiled code more
2593 consistent with other Fortran implementations by outputting
2594 leading zeros in formatted and list-directed output.
2597 Fix a code-generation bug involving adjustable dummy arrays with high
2598 bounds whose primaries are changed during procedure execution, and
2599 which might well improve code-generation performance for such arrays
2600 compared to @code{f2c} plus @code{gcc} (but apparently only when using
2601 @file{gcc-2.7.0} or later).
2604 Fix a code-generation bug involving invocation of @code{COMPLEX} and
2605 @code{DOUBLE COMPLEX} @code{FUNCTION}s and doing @code{COMPLEX} and
2606 @code{DOUBLE COMPLEX} divides, when the result
2607 of the invocation or divide is assigned directly to a variable
2608 that overlaps one or more of the arguments to the invocation or divide.
2611 Fix crash by not generating new optimal code for @samp{X**I} if @samp{I} is
2612 nonconstant and the expression is used to dimension a dummy
2613 array, since the @code{gcc} back end does not support the necessary
2614 mechanics (and the @code{gcc} front end rejects the equivalent
2615 construct, as it turns out).
2618 Fix crash on expressions like @samp{COMPLEX**INTEGER}.
2621 Fix crash on expressions like @samp{(1D0,2D0)**2}, i.e. raising a
2622 @code{DOUBLE COMPLEX} constant to an @code{INTEGER} constant power.
2625 Fix crashes and such involving diagnosed code.
2628 Diagnose, instead of crashing on, statement function definitions
2629 having duplicate dummy argument names.
2632 Fix bug causing rejection of good code involving statement function
2636 Fix bug resulting in debugger not knowing size of local equivalence
2637 area when any member of area has initial value (via @code{DATA},
2641 Fix installation bug that prevented installation of @code{g77} driver.
2642 Provide for easy selection of whether to install copy of @code{g77}
2643 as @code{f77} to replace the broken code.
2646 Fix @code{gcc} driver (affects @code{g77} thereby) to not
2647 gratuitously invoke the
2648 @code{f771} program (e.g. when @samp{-E} is specified).
2651 Fix diagnostic to point to correct source line when it immediately
2652 follows an @code{INCLUDE} statement.
2655 Support more compiler options in @code{gcc}/@code{g77} when
2656 compiling Fortran files.
2657 These options include @samp{-p}, @samp{-pg}, @samp{-aux-info}, @samp{-P},
2658 correct setting of version-number macros for preprocessing, full
2659 recognition of @samp{-O0}, and
2660 automatic insertion of configuration-specific linker specs.
2663 Add new intrinsics that interface to existing routines in @code{libf2c}:
2664 @code{ABORT}, @code{DERF}, @code{DERFC}, @code{ERF}, @code{ERFC}, @code{EXIT},
2665 @code{FLUSH}, @code{GETARG}, @code{GETENV}, @code{IARGC},
2666 @code{SIGNAL}, and @code{SYSTEM}.
2667 Note that @code{ABORT}, @code{EXIT}, @code{FLUSH}, @code{SIGNAL}, and
2668 @code{SYSTEM} are intrinsic subroutines, not functions (since they
2669 have side effects), so to get the return values from @code{SIGNAL}
2670 and @code{SYSTEM}, append a final argument specifying an @code{INTEGER}
2671 variable or array element (e.g. @samp{CALL SYSTEM('rm foo',ISTAT)}).
2674 Add new intrinsic group named @code{unix} to contain the new intrinsics,
2675 and by default enable this new group.
2678 Move @code{LOC()} intrinsic out of the @code{vxt} group to the new
2682 Improve @code{g77} so that @samp{g77 -v} by itself (or with
2683 certain other options, including @samp{-B}, @samp{-b}, @samp{-i},
2684 @samp{-nostdlib}, and @samp{-V}) reports lots more useful
2685 version info, and so that long-form options @code{gcc} accepts are
2686 understood by @code{g77} as well (even in truncated, unambiguous forms).
2689 Add new @code{g77} option @samp{--driver=name} to specify driver when
2690 default, @code{gcc}, isn't appropriate.
2693 Add support for @samp{#} directives (as output by the preprocessor) in the
2694 compiler, and enable generation of those directives by the
2695 preprocessor (when compiling @samp{.F} files) so diagnostics and debugging
2696 info are more useful to users of the preprocessor.
2699 Produce better diagnostics, more like @code{gcc}, with info such as
2700 @samp{In function `foo':} and @samp{In file included from...:}.
2703 Support @code{gcc}'s @samp{-fident} and @samp{-fno-ident} options.
2706 When @samp{-Wunused} in effect, don't warn about local variables used as
2707 statement-function dummy arguments or @code{DATA} implied-@code{DO} iteration
2708 variables, even though, strictly speaking, these are not uses
2709 of the variables themselves.
2712 When @samp{-W -Wunused} in effect, don't warn about unused dummy arguments
2713 at all, since there's no way to turn this off for individual
2714 cases (@code{g77} might someday start warning about these)---applies
2715 to @code{gcc} versions 2.7.0 and later, since earlier versions didn't
2716 warn about unused dummy arguments.
2719 New option @samp{-fno-underscoring} that inhibits transformation of names
2720 (by appending one or two underscores) so users may experiment
2721 with implications of such an environment.
2724 Minor improvement to @file{gcc/f/info} module to make it easier to build
2725 @code{g77} using the native (non-@code{gcc}) compiler on certain machines
2726 (but definitely not all machines nor all non-@code{gcc} compilers).
2728 do not report bugs showing problems compilers have with
2729 macros defined in @file{gcc/f/target.h} and used in places like
2730 @file{gcc/f/expr.c}.
2733 Add warning to be printed for each invocation of the compiler
2734 if the target machine @code{INTEGER}, @code{REAL}, or @code{LOGICAL} size
2736 since @code{g77} is known to not work well for such cases.
2739 Lots of new documentation (though work is still needed to put it into
2740 canonical GNU format).
2743 Build @code{libf2c} with @samp{-g0}, not @samp{-g2}, in effect
2744 (by default), to produce
2745 smaller library without lots of debugging clutter.
2748 @c 1995-05-19: 0.5.15 released.
2752 Fix bad code generation involving @samp{X**I} and temporary, internal variables
2753 generated by @code{g77} and the back end (such as for @code{DO} loops).
2756 Fix crash given @samp{CHARACTER A;DATA A/.TRUE./}.
2759 Replace crash with diagnostic given @samp{CHARACTER A;DATA A/1.0/}.
2762 Fix crash or other erratic behavior when null character constant
2763 (@samp{''}) is encountered.
2766 Fix crash or other erratic behavior involving diagnosed code.
2769 Fix code generation for external functions returning type @code{REAL} when
2770 the @samp{-ff2c} option is in force (which it is by default) so that
2771 @code{f2c} compatibility is indeed provided.
2774 Disallow @samp{COMMON I(10)} if @samp{I} has previously been specified
2775 with an array declarator.
2778 New @samp{-ffixed-line-length-@var{n}} option, where @var{n} is the
2780 of a typical fixed-form line, defaulting to 72 columns, such
2781 that characters beyond column @var{n} are ignored, or @var{n} is @samp{none},
2782 meaning no characters are ignored.
2783 does not affect lines
2784 with @samp{&} in column 1, which are always processed as if
2785 @samp{-ffixed-line-length-none} was in effect.
2788 No longer generate better code for some kinds of array references,
2789 as @code{gcc} back end is to be fixed to do this even better, and it
2790 turned out to slow down some code in some cases after all.
2793 In @code{COMMON} and @code{EQUIVALENCE} areas with any members given initial
2794 values (e.g. via @code{DATA}), uninitialized members now always
2795 initialized to binary zeros (though this is not required by
2796 the standard, and might not be done in future versions
2798 Previously, in some @code{COMMON}/@code{EQUIVALENCE} areas
2799 (essentially those with members of more than one type), the
2800 uninitialized members were initialized to spaces, to
2801 cater to @code{CHARACTER} types, but it seems no existing code expects
2802 that, while much existing code expects binary zeros.
2808 Don't emit bad code when low bound of adjustable array is nonconstant
2809 and thus might vary as an expression at run time.
2812 Emit correct code for calculation of number of trips in @code{DO} loops
2814 where the loop should not execute at all.
2815 (This bug affected cases
2816 where the difference between the begin and end values was less
2817 than the step count, though probably not for floating-point cases.)
2820 Fix crash when extra parentheses surround item in
2821 @code{DATA} implied-@code{DO} list.
2824 Fix crash over minor internal inconsistencies in handling diagnostics,
2825 just substitute dummy strings where necessary.
2828 Fix crash on some systems when compiling call to @code{MVBITS()} intrinsic.
2831 Fix crash on array assignment @samp{TYPE@var{ddd}(@dots{})=@dots{}}, where @var{ddd}
2832 is a string of one or more digits.
2835 Fix crash on @code{DCMPLX()} with a single @code{INTEGER} argument.
2838 Fix various crashes involving code with diagnosed errors.
2841 Support @samp{-I} option for @code{INCLUDE} statement, plus @code{gcc}'s
2842 @file{header.gcc} facility for handling systems like MS-DOS.
2845 Allow @code{INCLUDE} statement to be continued across multiple lines,
2846 even allow it to coexist with other statements on the same line.
2849 Incorporate Bellcore fixes to @code{libf2c} through 1995-03-15---this
2850 fixes a bug involving infinite loops reading EOF with empty list-directed
2854 Remove all the @code{g77}-specific auto-configuration scripts, code,
2856 except for temporary substitutes for bsearch() and strtoul(), as
2857 too many configure/build problems were reported in these areas.
2858 People will have to fix their systems' problems themselves, or at
2859 least somewhere other than @code{g77}, which expects a working ANSI C
2860 environment (and, for now, a GNU C compiler to compile @code{g77} itself).
2863 Complain if initialized common redeclared as larger in subsequent program
2867 Warn if blank common initialized, since its size can vary and hence
2868 related warnings that might be helpful won't be seen.
2871 New @samp{-fbackslash} option, on by default, that causes @samp{\}
2872 within @code{CHARACTER}
2873 and Hollerith constants to be interpreted a la GNU C.
2875 this behavior is somewhat different from @code{f2c}'s, which supports only
2876 a limited subset of backslash (escape) sequences.
2879 Make @samp{-fugly-args} the default.
2882 New @samp{-fugly-init} option, on by default, that allows typeless/Hollerith
2883 to be specified as initial values for variables or named constants
2884 (@code{PARAMETER}), and also allows character<->numeric conversion in
2885 those contexts---turn off via @samp{-fno-ugly-init}.
2888 New @samp{-finit-local-zero} option to initialize
2889 local variables to binary zeros.
2890 This does not affect whether they are @code{SAVE}d, i.e. made
2891 automatic or static.
2894 New @samp{-Wimplicit} option to warn about implicitly typed variables, arrays,
2896 (Basically causes all program units to default to @code{IMPLICIT NONE}.)
2899 @samp{-Wall} now implies @samp{-Wuninitialized} as with @code{gcc}
2900 (i.e. unless @samp{-O} not specified, since @samp{-Wuninitialized}
2901 requires @samp{-O}), and implies @samp{-Wunused} as well.
2904 @samp{-Wunused} no longer gives spurious messages for unused
2905 @code{EXTERNAL} names (since they are assumed to refer to block data
2906 program units, to make use of libraries more reliable).
2909 Support @code{%LOC()} and @code{LOC()} of character arguments.
2912 Support null (zero-length) character constants and expressions.
2915 Support @code{f2c}'s @code{IMAG()} generic intrinsic.
2918 Support @code{ICHAR()}, @code{IACHAR()}, and @code{LEN()} of
2919 character expressions that are valid in assignments but
2920 not normally as actual arguments.
2923 Support @code{f2c}-style @samp{&} in column 1 to mean continuation line.
2926 Allow @code{NAMELIST}, @code{EXTERNAL}, @code{INTRINSIC}, and @code{VOLATILE}
2927 in @code{BLOCK DATA}, even though these are not allowed by the standard.
2930 Allow @code{RETURN} in main program unit.
2933 Changes to Hollerith-constant support to obey Appendix C of the
2938 Now padded on the right with zeros, not spaces.
2941 Hollerith ``format specifications'' in the form of arrays of
2942 non-character allowed.
2945 Warnings issued when non-space truncation occurs when converting
2949 When specified as actual argument, now passed
2950 by reference to @code{INTEGER} (padded on right with spaces if constant
2951 too small, otherwise fully intact if constant wider the @code{INTEGER}
2952 type) instead of by value.
2955 @strong{Warning:} @code{f2c} differs on the
2956 interpretation of @samp{CALL FOO(1HX)}, which it treats exactly the
2957 same as @samp{CALL FOO('X')}, but which the standard and @code{g77} treat
2958 as @samp{CALL FOO(%REF('X '))} (padded with as many spaces as necessary
2959 to widen to @code{INTEGER}), essentially.
2962 Changes and fixes to typeless-constant support:
2966 Now treated as a typeless double-length @code{INTEGER} value.
2969 Warnings issued when overflow occurs.
2972 Padded on the left with zeros when converting
2976 Should be properly aligned and ordered on
2977 the target machine for whatever type it is turned into.
2980 When specified as actual argument, now passed as reference to
2981 a default @code{INTEGER} constant.
2985 @code{%DESCR()} of a non-@code{CHARACTER} expression now passes a pointer to
2986 the expression plus a length for the expression just as if
2987 it were a @code{CHARACTER} expression.
2988 For example, @samp{CALL FOO(%DESCR(D))}, where
2989 @samp{D} is @code{REAL*8}, is the same as @samp{CALL FOO(D,%VAL(8)))}.
2992 Name of multi-entrypoint master function changed to incorporate
2993 the name of the primary entry point instead of a decimal
2994 value, so the name of the master function for @samp{SUBROUTINE X}
2995 with alternate entry points is now @samp{__g77_masterfun_x}.
2998 Remove redundant message about zero-step-count @code{DO} loops.
3001 Clean up diagnostic messages, shortening many of them.
3004 Fix typo in @code{g77} man page.
3007 Clarify implications of constant-handling bugs in @file{f/BUGS}.
3010 Generate better code for @samp{**} operator with a right-hand operand of
3011 type @code{INTEGER}.
3014 Generate better code for @code{SQRT()} and @code{DSQRT()},
3015 also when @samp{-ffast-math}
3016 specified, enable better code generation for @code{SIN()} and @code{COS()}.
3019 Generate better code for some kinds of array references.
3022 Speed up lexing somewhat (this makes the compilation phase noticeably