1 @c Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001 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 2001-11-20
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 @command{g77} that are visible
44 to the programmers who actually write and maintain Fortran
45 code they compile with @command{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
49 users use @command{g77}.
50 @xref{News,,News About GNU Fortran}, for information on
51 such changes to @command{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 @command{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 @command{g77},
111 see @ref{Known Bugs,,Known Bugs In GNU Fortran}.
115 For information on bugs in the @value{which-g77} version of @command{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 @command{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} @command{g77}.
128 In particular, while this information is intended to apply to
129 the @value{which-g77} version of @command{g77},
130 only an official @emph{release} of that version
131 is expected to contain documentation that is
132 most consistent with the @command{g77} product in that version.
134 Nevertheless, information on @emph{previous} releases of @command{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 @command{g77} versions up-to-date.
146 An online, ``live'' version of this document
147 (derived directly from the mainline, development version
148 of @command{g77} within @command{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:
158 @command{g77} now has its man page generated from the texinfo documentation,
159 to guarantee that it remains up to date.
162 @command{g77} used to reject the following program on 32-bit targets:
165 DIMENSION A(140 000 000)
170 prog.f: In program `prog':
172 DIMENSION A(140 000 000)
174 Array `a' at (^) is too large to handle
176 because 140 000 000 reals is larger than the largest bit-extent that can be
177 expressed in 32 bits. However, bit-sizes never play a role after offsets
178 have been converted to byte addresses. Therefore this check has been removed.
179 Note: On GNU/Linux systems one has to compile programs that occupy more
180 than 1 Gbyte statically, i.e.@: @code{g77 -static ...}.
183 Based on work done by Juergen Pfeifer (@email{juergen.pfeifer@@gmx.net})
184 libf2c is now a shared library. One can still link in all objects with
185 the program by specifying the @option{-static} option.
188 Robert Anderson (@email{rwa@@alumni.princeton.edu}) thought up a two
189 line change that enables g77 to compile such code as:
193 DIMENSION A(N(1),N(2))
197 Note the use of array elements in the bounds of the adjustable array A.
200 George Helffrich (@email{george@@geo.titech.ac.jp}) implemented a change
201 in substring index checking (when specifying @option{-fbounds-check})
202 that permits the use of zero length substrings of the form
206 Based on code developed by Pedro Vazquez (@email{vazquez@@penelope.iqm.unicamp.br}),
207 the @code{libf2c} library is now able to read and write files larger than
208 2 Gbyte on 32-bit target machines, if the operating system supports this.
211 @heading In 0.5.26, @code{GCC} 3.0 versus @code{GCC} 2.95:
214 When a REWIND was issued after a WRITE statement on an unformatted
215 file, the implicit truncation was performed by copying the truncated
216 file to /tmp and copying the result back. This has been fixed by using
217 the @code{ftruncate} OS function. Thanks go to the GAMESS developers
218 for bringing this to our attention.
221 Using options @option{-g}, @option{-ggdb} or @option{-gdwarf[-2]} (where
222 appropriate for your target) now also enables debugging information
223 for COMMON BLOCK and EQUIVALENCE items to be emitted.
224 Thanks go to Andrew Vaught (@email{andy@@xena.eas.asu.edu}) and
225 George Helffrich (@email{george@@geology.bristol.ac.uk}) for
226 fixing this longstanding problem.
229 It is not necessary anymore to use the option @option{-femulate-complex}
230 to compile Fortran code using COMPLEX arithmetic, even on 64-bit machines
231 (like the Alpha). This will improve code generation.
234 INTRINSIC arithmetic functions are now treated as routines that do not
235 depend on anything but their argument(s). This enables further instruction
236 scheduling, because it is known that they cannot read or modify arbitrary
241 Upgrade to @code{libf2c} as of 2000-12-05.
243 This fixes a bug where a namelist containing initialization of LOGICAL
244 items and a variable starting with T or F would be read incorrectly.
247 The @code{TtyNam} intrinsics now set @var{Name} to all spaces (at run time)
248 if the system has no @code{ttyname} implementation available.
251 Upgrade to @code{libf2c} as of 1999-06-28.
253 This fixes a bug whereby
254 input to a @code{NAMELIST} read involving a repeat count,
255 such as @samp{K(5)=10*3},
256 was not properly handled by @code{libf2c}.
257 The first item was written to @samp{K(5)},
258 but the remaining nine were written elsewhere (still within the array),
259 not necessarily starting at @samp{K(6)}.
263 @heading In 0.5.25, @code{GCC} 2.95 (@code{EGCS} 1.2) versus @code{EGCS} 1.1.2:
267 @command{g77} no longer generates bad code for assignments,
268 or other conversions,
269 of @code{REAL} or @code{COMPLEX} constant expressions
270 to type @code{INTEGER(KIND=2)}
271 (often referred to as @code{INTEGER*8}).
273 For example, @samp{INTEGER*8 J; J = 4E10} now works as documented.
278 @command{g77} no longer truncates @code{INTEGER(KIND=2)}
279 (usually @code{INTEGER*8})
280 subscript expressions when evaluating array references
281 on systems with pointers widers than @code{INTEGER(KIND=1)}
287 @command{g77} no longer generates bad code
288 for an assignment to a @code{COMPLEX} variable or array
289 that partially overlaps one or more of the sources
290 of the same assignment
291 (a very rare construction).
292 It now assigns through a temporary,
293 in cases where such partial overlap is deemed possible.
298 @code{libg2c} (@code{libf2c}) no longer loses track
299 of the file being worked on
300 during a @code{BACKSPACE} operation.
305 @code{libg2c} (@code{libf2c}) fixes a bug whereby
306 input to a @code{NAMELIST} read involving a repeat count,
307 such as @samp{K(5)=10*3},
308 was not properly handled by @code{libf2c}.
309 The first item was written to @samp{K(5)},
310 but the remaining nine were written elsewhere (still within the array),
311 not necessarily starting at @samp{K(6)}.
316 @c Tim Prince reported this, regarding the TEST_FPU benchmark.
317 Automatic arrays now seem to be working on HP-UX systems.
322 The @code{Date} intrinsic now returns the correct result
323 on big-endian systems.
328 Fix @command{g77} so it no longer crashes when compiling
329 I/O statements using keywords that define @code{INTEGER} values,
330 such as @samp{IOSTAT=@var{j}},
331 where @var{j} is other than default @code{INTEGER}
332 (such as @code{INTEGER*2}).
333 Instead, it issues a diagnostic.
338 Fix @command{g77} so it properly handles @samp{DATA A/@var{rpt}*@var{val}/},
339 where @var{rpt} is not default @code{INTEGER}, such as @code{INTEGER*2},
340 instead of producing a spurious diagnostic.
341 Also fix @samp{DATA (A(I),I=1,N)},
342 where @samp{N} is not default @code{INTEGER}
343 to work instead of crashing @command{g77}.
348 The @option{-ax} option is now obeyed when compiling Fortran programs.
349 (It is passed to the @file{f771} driver.)
353 The new @option{-fbounds-check} option
354 causes @command{g77} to compile run-time bounds checks
355 of array subscripts, as well as of substring start and end points.
358 @code{libg2c} now supports building as multilibbed library,
359 which provides better support for systems
360 that require options such as @option{-mieee}
364 Source file names with the suffixes @samp{.FOR} and @samp{.FPP}
365 now are recognized by @command{g77}
366 as if they ended in @samp{.for} and @samp{.fpp}, respectively.
369 The order of arguments to the @emph{subroutine} forms of the
370 @code{CTime}, @code{DTime}, @code{ETime}, and @code{TtyNam}
371 intrinsics has been swapped.
372 The argument serving as the returned value
373 for the corresponding function forms
374 now is the @emph{second} argument,
375 making these consistent with the other subroutine forms
376 of @code{libU77} intrinsics.
379 @command{g77} now warns about a reference to an intrinsic
380 that has an interface that is not Year 2000 (Y2K) compliant.
381 Also, @code{libg2c} has been changed to increase the likelihood
382 of catching references to the implementations of these intrinsics
383 using the @code{EXTERNAL} mechanism
384 (which would avoid the new warnings).
387 @xref{Year 2000 (Y2K) Problems}, for more information.
392 @command{g77} now warns about a reference to a function
393 when the corresponding @emph{subsequent} function program unit
394 disagrees with the reference concerning the type of the function.
398 @option{-fno-emulate-complex} is now the default option.
399 This should result in improved performance
400 of code that uses the @code{COMPLEX} data type.
403 @cindex double-precision performance
404 @cindex -malign-double
406 The @option{-malign-double} option
407 now reliably aligns @emph{all} double-precision variables and arrays
408 on Intel x86 targets.
412 Even without the @option{-malign-double} option,
413 @command{g77} reliably aligns local double-precision variables
414 that are not in @code{EQUIVALENCE} areas
415 and not @code{SAVE}'d.
420 @command{g77} now open-codes (``inlines'') division of @code{COMPLEX} operands
421 instead of generating a run-time call to
422 the @code{libf2c} routines @code{c_div} or @code{z_div},
423 unless the @option{-Os} option is specified.
427 @command{g77} no longer generates code to maintain @code{errno},
428 a C-language concept,
429 when performing operations such as the @code{SqRt} intrinsic.
433 @command{g77} developers can temporarily use
434 the @option{-fflatten-arrays} option
435 to compare how the compiler handles code generation
436 using C-like constructs as compared to the
437 Fortran-like method constructs normally used.
442 A substantial portion of the @command{g77} front end's code-generation component
444 It now generates code using facilities more robustly supported
445 by the @command{gcc} back end.
446 One effect of this rewrite is that some codes no longer produce
447 a spurious ``label @var{lab} used before containing binding contour''
452 Support for the @option{-fugly} option has been removed.
456 Improve documentation and indexing,
457 including information on Year 2000 (Y2K) compliance,
458 and providing more information on internals of the front end.
463 Upgrade to @code{libf2c} as of 1999-05-10.
467 @heading In 0.5.24 versus 0.5.23:
469 There is no @command{g77} version 0.5.24 at this time,
471 0.5.24 is the version number designated for bug fixes and,
472 perhaps, some new features added,
474 Version 0.5.23 requires @command{gcc} 2.8.1,
475 as 0.5.24 was planned to require.
477 Due to @code{EGCS} becoming @code{GCC}
478 (which is now an acronym for ``GNU Compiler Collection''),
479 and @code{EGCS} 1.2 becoming officially designated @code{GCC} 2.95,
480 there seems to be no need for an actual 0.5.24 release.
482 To reduce the confusion already resulting from use of 0.5.24
483 to designate @command{g77} versions within @code{EGCS} versions 1.0 and 1.1,
484 as well as in versions of @command{g77} documentation and notices
486 ``mainline'' @command{g77} version numbering resumes
487 at 0.5.25 with @code{GCC} 2.95 (@code{EGCS} 1.2),
488 skipping over 0.5.24 as a placeholder version number.
490 To repeat, there is no @command{g77} 0.5.24, but there is now a 0.5.25.
491 Please remain calm and return to your keypunch units.
493 @c 1999-03-15: EGCS 1.1.2 released.
494 @heading In @code{EGCS} 1.1.2 versus @code{EGCS} 1.1.1:
498 Fix the @code{IDate} intrinsic (VXT) (in @code{libg2c})
499 so the returned year is in the documented, non-Y2K-compliant range
501 instead of being returned as 100 in the year 2000.
504 @xref{IDate Intrinsic (VXT)},
505 for more information.
509 Fix the @code{Date_and_Time} intrinsic (in @code{libg2c})
510 to return the milliseconds value properly
514 Fix the @code{LStat} intrinsic (in @code{libg2c})
515 to return device-ID information properly
519 Improve documentation.
523 @c 1998-12-04: EGCS 1.1.1 released.
524 @heading In @code{EGCS} 1.1.1 versus @code{EGCS} 1.1:
528 Fix @code{libg2c} so it performs an implicit @code{ENDFILE} operation
530 whenever a @code{REWIND} is done.
532 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
533 @command{g77}'s version of @code{libf2c}.)
536 Fix @code{libg2c} so it no longer crashes with a spurious diagnostic
537 upon doing any I/O following a direct formatted write.
539 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
540 @command{g77}'s version of @code{libf2c}.)
543 Fix @command{g77} so it no longer crashes compiling references
544 to the @code{Rand} intrinsic on some systems.
547 Fix @command{g77} portion of installation process so it works
548 better on some systems
549 (those with shells requiring @samp{else true} clauses
550 on @code{if} constructs
551 for the completion code to be set properly).
555 @c 1998-09-03: EGCS 1.1 released.
556 @heading In @code{EGCS} 1.1 versus @code{EGCS} 1.0.3:
560 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
561 that wrote one byte beyond the end of its @code{CHARACTER}
563 and in the @code{libU77} intrinsics
564 @code{GMTime} and @code{LTime}
565 that overwrote their arguments.
570 Assumed arrays with negative bounds
571 (such as @samp{REAL A(-1:*)})
572 no longer elicit spurious diagnostics from @command{g77},
573 even on systems with pointers having
574 different sizes than integers.
576 This bug is not known to have existed in any
577 recent version of @command{gcc}.
578 It was introduced in an early release of @code{egcs}.
583 Valid combinations of @code{EXTERNAL},
584 passing that external as a dummy argument
585 without explicitly giving it a type,
586 and, in a subsequent program unit,
587 referencing that external as
588 an external function with a different type
589 no longer crash @command{g77}.
594 @code{CASE DEFAULT} no longer crashes @command{g77}.
599 The @option{-Wunused} option no longer issues a spurious
600 warning about the ``master'' procedure generated by
601 @command{g77} for procedures containing @code{ENTRY} statements.
605 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
606 compile-time constant @code{INTEGER} expression.
609 Fix @command{g77} @option{-g} option so procedures that
610 use @code{ENTRY} can be stepped through, line by line,
614 Allow any @code{REAL} argument to intrinsics
615 @code{Second} and @code{CPU_Time}.
618 Use @code{tempnam}, if available, to open scratch files
619 (as in @samp{OPEN(STATUS='SCRATCH')})
620 so that the @code{TMPDIR} environment variable,
624 @command{g77}'s version of @code{libf2c} separates out
625 the setting of global state
626 (such as command-line arguments and signal handling)
627 from @file{main.o} into distinct, new library
630 This should make it easier to write portable applications
631 that have their own (non-Fortran) @code{main()} routine
632 properly set up the @code{libf2c} environment, even
633 when @code{libf2c} (now @code{libg2c}) is a shared library.
637 @command{g77} no longer installs the @file{f77} command
638 and @file{f77.1} man page
639 in the @file{/usr} or @file{/usr/local} hierarchy,
640 even if the @file{f77-install-ok} file exists
641 in the source or build directory.
642 See the installation documentation for more information.
647 @command{g77} no longer installs the @file{libf2c.a} library
648 and @file{f2c.h} include file
649 in the @file{/usr} or @file{/usr/local} hierarchy,
650 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
651 in the source or build directory.
652 See the installation documentation for more information.
657 The @file{libf2c.a} library produced by @command{g77} has been
658 renamed to @file{libg2c.a}.
659 It is installed only in the @command{gcc} ``private''
660 directory hierarchy, @file{gcc-lib}.
661 This allows system administrators and users to choose which
662 version of the @code{libf2c} library from @code{netlib} they
663 wish to use on a case-by-case basis.
664 See the installation documentation for more information.
669 The @file{f2c.h} include (header) file produced by @command{g77}
670 has been renamed to @file{g2c.h}.
671 It is installed only in the @command{gcc} ``private''
672 directory hierarchy, @file{gcc-lib}.
673 This allows system administrators and users to choose which
674 version of the include file from @code{netlib} they
675 wish to use on a case-by-case basis.
676 See the installation documentation for more information.
680 The @command{g77} command now expects the run-time library
681 to be named @code{libg2c.a} instead of @code{libf2c.a},
682 to ensure that a version other than the one built and
683 installed as part of the same @command{g77} version is picked up.
687 During the configuration and build process,
688 @command{g77} creates subdirectories it needs only as it
690 Other cleaning up of the configuration and build process
691 has been performed as well.
696 @code{install-info} now used to update the directory of
697 Info documentation to contain an entry for @command{g77}
698 (during installation).
702 Some diagnostics have been changed from warnings to errors,
703 to prevent inadvertent use of the resulting, probably buggy,
705 These mostly include diagnostics about use of unsupported features
706 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
707 @code{WRITE} statements,
708 and about truncations of various sorts of constants.
712 Improve compilation of @code{FORMAT} expressions so that
713 a null byte is appended to the last operand if it
715 This provides a cleaner run-time diagnostic as provided
716 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
721 Improve documentation and indexing.
726 The upgrade to @code{libf2c} as of 1998-06-18
727 should fix a variety of problems, including
728 those involving some uses of the @code{T} format
729 specifier, and perhaps some build (porting) problems
734 @c 1998-09-03: EGCS 1.1 released.
735 @heading In @code{EGCS} 1.1 versus @command{g77} 0.5.23:
739 @cindex stack, 387 coprocessor
743 Fix a code-generation bug that afflicted
744 Intel x86 targets when @option{-O2} was specified
745 compiling, for example, an old version of
746 the @code{DNRM2} routine.
748 The x87 coprocessor stack was being
749 mismanaged in cases involving assigned @code{GOTO}
755 @command{g77} no longer produces incorrect code
757 for @code{EQUIVALENCE} and @code{COMMON}
758 aggregates that, due to ``unnatural'' ordering of members
759 vis-a-vis their types, require initial padding.
764 Fix @command{g77} crash compiling code
765 containing the construct @samp{CMPLX(0.)} or similar.
770 @command{g77} no longer crashes when compiling code
771 containing specification statements such as
772 @samp{INTEGER(KIND=7) PTR}.
777 @command{g77} no longer crashes when compiling code
778 such as @samp{J = SIGNAL(1, 2)}.
782 @command{g77} now treats @samp{%LOC(@var{expr})} and
783 @samp{LOC(@var{expr})} as ``ordinary'' expressions
784 when they are used as arguments in procedure calls.
785 This change applies only to global (filewide) analysis,
786 making it consistent with
787 how @command{g77} actually generates code
790 Previously, @command{g77} treated these expressions
791 as denoting special ``pointer'' arguments
792 for the purposes of filewide analysis.
796 Fix @command{g77} crash
797 (or apparently infinite run-time)
798 when compiling certain complicated expressions
799 involving @code{COMPLEX} arithmetic
800 (especially multiplication).
804 @cindex double-precision performance
805 @cindex -malign-double
807 Align static double-precision variables and arrays
809 regardless of whether @option{-malign-double} is specified.
811 Generally, this affects only local variables and arrays
812 having the @code{SAVE} attribute
813 or given initial values via @code{DATA}.
816 The @command{g77} driver now ensures that @option{-lg2c}
817 is specified in the link phase prior to any
818 occurrence of @option{-lm}.
819 This prevents accidentally linking to a routine
820 in the SunOS4 @option{-lm} library
821 when the generated code wants to link to the one
822 in @code{libf2c} (@code{libg2c}).
825 @command{g77} emits more debugging information when
828 This new information allows, for example,
829 @kbd{which __g77_length_a} to be used in @command{gdb}
830 to determine the type of the phantom length argument
831 supplied with @code{CHARACTER} variables.
833 This information pertains to internally-generated
834 type, variable, and other information,
835 not to the longstanding deficiencies vis-a-vis
836 @code{COMMON} and @code{EQUIVALENCE}.
839 The F90 @code{Date_and_Time} intrinsic now is
843 The F90 @code{System_Clock} intrinsic allows
844 the optional arguments (except for the @code{Count}
845 argument) to be omitted.
849 Upgrade to @code{libf2c} as of 1998-06-18.
854 Improve documentation and indexing.
859 @heading In previous versions:
861 Information on previous versions is not provided
862 in this @file{@value{path-g77}/NEWS} file,
864 See @file{@value{path-g77}/news.texi},
865 or any of its other derivations
866 (Info, HTML, dvi forms)
867 for such information.
871 @c 1998-05-20: 0.5.23 released.
872 @heading In 0.5.23 versus 0.5.22:
875 This release contains several regressions against
876 version 0.5.22 of @command{g77}, due to using the
877 ``vanilla'' @command{gcc} back end instead of patching
878 it to fix a few bugs and improve performance in a
881 Features that have been dropped from this version
882 of @command{g77} due to their being implemented
883 via @command{g77}-specific patches to the @command{gcc}
884 back end in previous releases include:
888 Support for @code{__restrict__} keyword,
889 the options @option{-fargument-alias}, @option{-fargument-noalias},
890 and @option{-fargument-noalias-global},
891 and the corresponding alias-analysis code.
893 (@code{egcs} has the alias-analysis
894 code, but not the @code{__restrict__} keyword.
895 @code{egcs} @command{g77} users benefit from the alias-analysis
896 code despite the lack of the @code{__restrict__} keyword,
897 which is a C-language construct.)
900 Support for the GNU compiler options
901 @option{-fmove-all-movables},
902 @option{-freduce-all-givs},
903 and @option{-frerun-loop-opt}.
905 (@code{egcs} supports these options.
906 @command{g77} users of @code{egcs} benefit from them even if
907 they are not explicitly specified,
908 because the defaults are optimized for @command{g77} users.)
911 Support for the @option{-W} option warning about
912 integer division by zero.
915 The Intel x86-specific option @option{-malign-double}
916 applying to stack-allocated data
917 as well as statically-allocate data.
921 Note that the @file{gcc/f/gbe/} subdirectory has been removed
922 from this distribution as a result of @command{g77} no longer
923 including patches for the @command{gcc} back end.
928 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
929 that wrote one byte beyond the end of its @code{CHARACTER}
931 and in the @code{libU77} intrinsics
932 @code{GMTime} and @code{LTime}
933 that overwrote their arguments.
937 Support @command{gcc} version 2.8,
938 and remove support for prior versions of @command{gcc}.
940 @cindex -@w{}-driver option
941 @cindex @command{g77} options, -@w{}-driver
942 @cindex options, -@w{}-driver
944 Remove support for the @option{--driver} option,
945 as @command{g77} now does all the driving,
946 just like @command{gcc}.
950 @code{CASE DEFAULT} no longer crashes @command{g77}.
955 Valid combinations of @code{EXTERNAL},
956 passing that external as a dummy argument
957 without explicitly giving it a type,
958 and, in a subsequent program unit,
959 referencing that external as
960 an external function with a different type
961 no longer crash @command{g77}.
966 @command{g77} no longer installs the @file{f77} command
967 and @file{f77.1} man page
968 in the @file{/usr} or @file{/usr/local} hierarchy,
969 even if the @file{f77-install-ok} file exists
970 in the source or build directory.
971 See the installation documentation for more information.
976 @command{g77} no longer installs the @file{libf2c.a} library
977 and @file{f2c.h} include file
978 in the @file{/usr} or @file{/usr/local} hierarchy,
979 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
980 in the source or build directory.
981 See the installation documentation for more information.
986 The @file{libf2c.a} library produced by @command{g77} has been
987 renamed to @file{libg2c.a}.
988 It is installed only in the @command{gcc} ``private''
989 directory hierarchy, @file{gcc-lib}.
990 This allows system administrators and users to choose which
991 version of the @code{libf2c} library from @code{netlib} they
992 wish to use on a case-by-case basis.
993 See the installation documentation for more information.
998 The @file{f2c.h} include (header) file produced by @command{g77}
999 has been renamed to @file{g2c.h}.
1000 It is installed only in the @command{gcc} ``private''
1001 directory hierarchy, @file{gcc-lib}.
1002 This allows system administrators and users to choose which
1003 version of the include file from @code{netlib} they
1004 wish to use on a case-by-case basis.
1005 See the installation documentation for more information.
1009 The @command{g77} command now expects the run-time library
1010 to be named @code{libg2c.a} instead of @code{libf2c.a},
1011 to ensure that a version other than the one built and
1012 installed as part of the same @command{g77} version is picked up.
1014 @ifclear USERVISONLY
1016 The @option{-Wunused} option no longer issues a spurious
1017 warning about the ``master'' procedure generated by
1018 @command{g77} for procedures containing @code{ENTRY} statements.
1022 @command{g77}'s version of @code{libf2c} separates out
1023 the setting of global state
1024 (such as command-line arguments and signal handling)
1025 from @file{main.o} into distinct, new library
1028 This should make it easier to write portable applications
1029 that have their own (non-Fortran) @code{main()} routine
1030 properly set up the @code{libf2c} environment, even
1031 when @code{libf2c} (now @code{libg2c}) is a shared library.
1033 @ifclear USERVISONLY
1035 During the configuration and build process,
1036 @command{g77} creates subdirectories it needs only as it
1037 needs them, thus avoiding unnecessary creation of, for example,
1038 @file{stage1/f/runtime} when doing a non-bootstrap build.
1039 Other cleaning up of the configuration and build process
1040 has been performed as well.
1043 @ifclear USERVISONLY
1045 @code{install-info} now used to update the directory of
1046 Info documentation to contain an entry for @command{g77}
1047 (during installation).
1051 Some diagnostics have been changed from warnings to errors,
1052 to prevent inadvertent use of the resulting, probably buggy,
1054 These mostly include diagnostics about use of unsupported features
1055 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
1056 @code{WRITE} statements,
1057 and about truncations of various sorts of constants.
1059 @ifclear USERVISONLY
1061 Improve documentation and indexing.
1064 @ifclear USERVISONLY
1066 Upgrade to @code{libf2c} as of 1998-04-20.
1068 This should fix a variety of problems, including
1069 those involving some uses of the @code{T} format
1070 specifier, and perhaps some build (porting) problems
1075 @c 1998-03-16: 0.5.22 released.
1076 @heading In 0.5.22 versus 0.5.21:
1078 @ifclear USERVISONLY
1080 Fix code generation for iterative @code{DO} loops that
1081 have one or more references to the iteration variable,
1082 or to aliases of it, in their control expressions.
1083 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1086 @ifclear USERVISONLY
1088 @cindex stack, 387 coprocessor
1092 Fix a code-generation bug that afflicted
1093 Intel x86 targets when @option{-O2} was specified
1094 compiling, for example, an old version of
1095 the @code{DNRM2} routine.
1097 The x87 coprocessor stack was being
1098 mismanaged in cases involving assigned @code{GOTO}
1102 @ifclear USERVISONLY
1104 Fix @code{DTime} intrinsic so as not to truncate
1105 results to integer values (on some systems).
1109 Fix @code{Signal} intrinsic so it offers portable
1110 support for 64-bit systems (such as Digital Alphas
1113 @ifclear USERVISONLY
1115 Fix run-time crash involving @code{NAMELIST} on 64-bit
1116 machines such as Alphas.
1119 @ifclear USERVISONLY
1121 Fix @command{g77} version of @code{libf2c} so it no longer
1122 produces a spurious @samp{I/O recursion} diagnostic at run time
1123 when an I/O operation (such as @samp{READ *,I}) is interrupted
1124 in a manner that causes the program to be terminated
1125 via the @code{f_exit} routine (such as via @kbd{C-c}).
1128 @ifclear USERVISONLY
1130 Fix @command{g77} crash triggered by @code{CASE} statement with
1131 an omitted lower or upper bound.
1134 @ifclear USERVISONLY
1136 Fix @command{g77} crash compiling references to @code{CPU_Time}
1140 @ifclear USERVISONLY
1142 Fix @command{g77} crash
1143 (or apparently infinite run-time)
1144 when compiling certain complicated expressions
1145 involving @code{COMPLEX} arithmetic
1146 (especially multiplication).
1149 @ifclear USERVISONLY
1151 Fix @command{g77} crash on statements such as
1152 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1153 @samp{Z} is @code{DOUBLE COMPLEX}.
1156 @ifclear USERVISONLY
1158 Fix a @command{g++} crash.
1162 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
1163 compile-time constant @code{INTEGER} expression.
1166 Fix @command{g77} @option{-g} option so procedures that
1167 use @code{ENTRY} can be stepped through, line by line,
1170 @ifclear USERVISONLY
1172 Fix a profiling-related bug in @command{gcc} back end for
1173 Intel x86 architecture.
1177 Allow any @code{REAL} argument to intrinsics
1178 @code{Second} and @code{CPU_Time}.
1181 Allow any numeric argument to intrinsics
1182 @code{Int2} and @code{Int8}.
1185 Use @code{tempnam}, if available, to open scratch files
1186 (as in @samp{OPEN(STATUS='SCRATCH')})
1187 so that the @code{TMPDIR} environment variable,
1188 if present, is used.
1191 Rename the @command{gcc} keyword @code{restrict} to
1192 @code{__restrict__}, to avoid rejecting valid, existing,
1194 Support for @code{restrict} is now more like support
1197 @ifclear USERVISONLY
1199 Fix @option{-fpedantic} to not reject procedure invocations
1200 such as @samp{I=J()} and @samp{CALL FOO()}.
1204 Fix @option{-fugly-comma} to affect invocations of
1205 only external procedures.
1206 Restore rejection of gratuitous trailing omitted
1207 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1210 Fix compiler so it accepts @option{-fgnu-intrinsics-*} and
1211 @option{-fbadu77-intrinsics-*} options.
1213 @ifclear USERVISONLY
1215 Improve diagnostic messages from @code{libf2c}
1216 so it is more likely that the printing of the
1217 active format string is limited to the string,
1218 with no trailing garbage being printed.
1220 (Unlike @command{f2c}, @command{g77} did not append
1221 a null byte to its compiled form of every
1222 format string specified via a @code{FORMAT} statement.
1223 However, @command{f2c} would exhibit the problem
1224 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1225 by printing @samp{(I)garbage} as the format string.)
1228 @ifclear USERVISONLY
1230 Improve compilation of @code{FORMAT} expressions so that
1231 a null byte is appended to the last operand if it
1233 This provides a cleaner run-time diagnostic as provided
1234 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
1237 @ifclear USERVISONLY
1239 Fix various crashes involving code with diagnosed errors.
1242 @ifclear USERVISONLY
1244 Fix cross-compilation bug when configuring @code{libf2c}.
1247 @ifclear USERVISONLY
1249 Improve diagnostics.
1252 @ifclear USERVISONLY
1254 Improve documentation and indexing.
1257 @ifclear USERVISONLY
1259 Upgrade to @code{libf2c} as of 1997-09-23.
1260 This fixes a formatted-I/O bug that afflicted
1261 64-bit systems with 32-bit integers
1262 (such as Digital Alpha running GNU/Linux).
1266 @c 1998-03-18: EGCS 1.0.2 released.
1267 @heading In @code{EGCS} 1.0.2 versus @code{EGCS} 1.0.1:
1269 @ifclear USERVISONLY
1271 Fix @command{g77} crash triggered by @code{CASE} statement with
1272 an omitted lower or upper bound.
1275 @ifclear USERVISONLY
1277 Fix @command{g77} crash on statements such as
1278 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1279 @samp{Z} is @code{DOUBLE COMPLEX}.
1282 @ifclear USERVISONLY
1284 @cindex support, ELF
1285 @cindex -fPIC option
1286 @cindex options, -fPIC
1288 Fix @option{-fPIC} (such as compiling for ELF targets)
1289 on the Intel x86 architecture target
1290 so invalid assembler code is no longer produced.
1293 @ifclear USERVISONLY
1295 Fix @option{-fpedantic} to not reject procedure invocations
1296 such as @samp{I=J()} and @samp{CALL FOO()}.
1299 @ifclear USERVISONLY
1301 Fix @option{-fugly-comma} to affect invocations of
1302 only external procedures.
1303 Restore rejection of gratuitous trailing omitted
1304 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1308 Fix compiler so it accepts @option{-fgnu-intrinsics-*} and
1309 @option{-fbadu77-intrinsics-*} options.
1312 @c 1998-01-06: EGCS 1.0.1 released.
1313 @heading In @code{EGCS} 1.0.1 versus @code{EGCS} 1.0:
1314 @ifclear USERVISONLY
1317 Fix run-time crash involving @code{NAMELIST} on 64-bit
1318 machines such as Alphas.
1322 @c 1997-12-03: EGCS 1.0 released.
1323 @heading In @code{EGCS} 1.0 versus @command{g77} 0.5.21:
1326 Version 1.0 of @code{egcs}
1327 contains several regressions against
1328 version 0.5.21 of @command{g77},
1330 ``vanilla'' @command{gcc} back end instead of patching
1331 it to fix a few bugs and improve performance in a
1334 Features that have been dropped from this version
1335 of @command{g77} due to their being implemented
1336 via @command{g77}-specific patches to the @command{gcc}
1337 back end in previous releases include:
1341 Support for the C-language @code{restrict} keyword.
1344 Support for the @option{-W} option warning about
1345 integer division by zero.
1348 The Intel x86-specific option @option{-malign-double}
1349 applying to stack-allocated data
1350 as well as statically-allocate data.
1353 @ifclear USERVISONLY
1354 Note that the @file{gcc/f/gbe/} subdirectory has been removed
1355 from this distribution as a result of @command{g77}
1356 being fully integrated with
1357 the @code{egcs} variant of the @command{gcc} back end.
1360 @ifclear USERVISONLY
1362 Fix code generation for iterative @code{DO} loops that
1363 have one or more references to the iteration variable,
1364 or to aliases of it, in their control expressions.
1365 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1368 @ifclear USERVISONLY
1370 Fix @code{DTime} intrinsic so as not to truncate
1371 results to integer values (on some systems).
1374 @ifclear USERVISONLY
1376 @c Toon Moene discovered these.
1377 Some Fortran code, miscompiled
1378 by @command{g77} built on @command{gcc} version 2.8.1
1379 on m68k-next-nextstep3 configurations
1380 when using the @option{-O2} option,
1381 is now compiled correctly.
1382 It is believed that a C function known to miscompile
1383 on that configuration
1384 when using the @samp{-O2 -funroll-loops} options
1385 also is now compiled correctly.
1388 @ifclear USERVISONLY
1390 Remove support for non-@code{egcs} versions of @command{gcc}.
1393 @cindex -@w{}-driver option
1394 @cindex @command{g77} options, -@w{}-driver
1395 @cindex options, -@w{}-driver
1397 Remove support for the @option{--driver} option,
1398 as @command{g77} now does all the driving,
1399 just like @command{gcc}.
1402 Allow any numeric argument to intrinsics
1403 @code{Int2} and @code{Int8}.
1405 @ifclear USERVISONLY
1407 Improve diagnostic messages from @code{libf2c}
1408 so it is more likely that the printing of the
1409 active format string is limited to the string,
1410 with no trailing garbage being printed.
1412 (Unlike @command{f2c}, @command{g77} did not append
1413 a null byte to its compiled form of every
1414 format string specified via a @code{FORMAT} statement.
1415 However, @code{f2c} would exhibit the problem
1416 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1417 by printing @samp{(I)garbage} as the format string.)
1420 @ifclear USERVISONLY
1422 Upgrade to @code{libf2c} as of 1997-09-23.
1423 This fixes a formatted-I/O bug that afflicted
1424 64-bit systems with 32-bit integers
1425 (such as Digital Alpha running GNU/Linux).
1429 @c 1997-09-09: 0.5.21 released.
1432 @ifclear USERVISONLY
1434 Fix a code-generation bug introduced by 0.5.20
1435 caused by loop unrolling (by specifying
1436 @option{-funroll-loops} or similar).
1437 This bug afflicted all code compiled by
1438 version 2.7.2.2.f.2 of @command{gcc} (C, C++,
1439 Fortran, and so on).
1442 @ifclear USERVISONLY
1444 Fix a code-generation bug manifested when
1445 combining local @code{EQUIVALENCE} with a
1446 @code{DATA} statement that follows
1447 the first executable statement (or is
1448 treated as an executable-context statement
1449 as a result of using the @option{-fpedantic}
1453 @ifclear USERVISONLY
1455 Fix a compiler crash that occurred when an
1456 integer division by a constant zero is detected.
1457 Instead, when the @option{-W} option is specified,
1458 the @command{gcc} back end issues a warning about such a case.
1459 This bug afflicted all code compiled by
1460 version 2.7.2.2.f.2 of @command{gcc} (C, C++,
1461 Fortran, and so on).
1465 When the @option{-W} option is specified, @command{gcc}, @command{g77},
1466 and other GNU compilers that incorporate the @command{gcc}
1467 back end as modified by @command{g77}, issue
1468 a warning about integer division by constant zero.
1471 @ifclear USERVISONLY
1473 Fix a compiler crash that occurred in some cases
1474 of procedure inlining.
1475 (Such cases became more frequent in 0.5.20.)
1478 @ifclear USERVISONLY
1480 Fix a compiler crash resulting from using @code{DATA}
1481 or similar to initialize a @code{COMPLEX} variable or
1485 @ifclear USERVISONLY
1487 Fix compiler crashes involving use of @code{AND}, @code{OR},
1488 or @code{XOR} intrinsics.
1491 @ifclear USERVISONLY
1493 Fix compiler bug triggered when using a @code{COMMON}
1494 or @code{EQUIVALENCE} variable
1495 as the target of an @code{ASSIGN}
1496 or assigned-@code{GOTO} statement.
1499 @ifclear USERVISONLY
1501 Fix compiler crashes due to using the name of a some
1502 non-standard intrinsics (such as @code{FTell} or
1503 @code{FPutC}) as such and as the name of a procedure
1505 Such dual use of a name in a program is allowed by
1509 @c @command{g77}'s version of @code{libf2c} has been modified
1510 @c so that the external names of library's procedures do not
1511 @c conflict with names used for Fortran procedures compiled
1512 @c by @command{g77}.
1513 @c An additional layer of jacket procedures has been added
1514 @c to @code{libf2c} to map the old names to the new names,
1515 @c for automatic use by programs that interface to the
1516 @c library procedures via the external-procedure mechanism.
1518 @c For example, the intrinsic @code{FPUTC} previously was
1519 @c implemented by @command{g77} as a call to the @code{libf2c}
1520 @c routine @code{fputc_}.
1521 @c This would conflict with a Fortran procedure named @code{FPUTC}
1522 @c (using default compiler options), and this conflict
1523 @c would cause a crash under certain circumstances.
1525 @c Now, the intrinsic @code{FPUTC} calls @code{G77_fputc_0},
1526 @c which does not conflict with the @code{fputc_} external
1527 @c that implements a Fortran procedure named @code{FPUTC}.
1529 @c Programs that refer to @code{FPUTC} as an external procedure
1530 @c without supplying their own implementation will link to
1531 @c the new @code{libf2c} routine @code{fputc_}, which is
1532 @c simply a jacket routine that calls @code{G77_fputc_0}.
1534 @ifclear USERVISONLY
1536 Place automatic arrays on the stack, even if
1537 @code{SAVE} or the @option{-fno-automatic} option
1539 This avoids a compiler crash in some cases.
1542 @ifclear USERVISONLY
1544 The @option{-malign-double} option now reliably aligns
1545 @code{DOUBLE PRECISION} optimally on Pentium and
1546 Pentium Pro architectures (586 and 686 in @command{gcc}).
1550 New option @option{-Wno-globals} disables warnings
1551 about ``suspicious'' use of a name both as a global
1552 name and as the implicit name of an intrinsic, and
1553 warnings about disagreements over the number or natures of
1554 arguments passed to global procedures, or the
1555 natures of the procedures themselves.
1557 The default is to issue such warnings, which are
1558 new as of this version of @command{g77}.
1561 New option @option{-fno-globals} disables diagnostics
1562 about potentially fatal disagreements
1563 analysis problems, such as disagreements over the
1564 number or natures of arguments passed to global
1565 procedures, or the natures of those procedures themselves.
1567 The default is to issue such diagnostics and flag
1568 the compilation as unsuccessful.
1569 With this option, the diagnostics are issued as
1570 warnings, or, if @option{-Wno-globals} is specified,
1571 are not issued at all.
1573 This option also disables inlining of global procedures,
1574 to avoid compiler crashes resulting from coding errors
1575 that these diagnostics normally would identify.
1577 @ifclear USERVISONLY
1579 Diagnose cases where a reference to a procedure
1580 disagrees with the type of that procedure, or
1581 where disagreements about the number or nature
1583 This avoids a compiler crash.
1586 @ifclear USERVISONLY
1588 Fix parsing bug whereby @command{g77} rejected a
1589 second initialization specification immediately
1590 following the first's closing @samp{/} without
1591 an intervening comma in a @code{DATA} statement,
1592 and the second specification was an implied-DO list.
1595 @ifclear USERVISONLY
1597 Improve performance of the @command{gcc} back end so
1598 certain complicated expressions involving @code{COMPLEX}
1599 arithmetic (especially multiplication) don't appear to
1600 take forever to compile.
1603 @ifclear USERVISONLY
1605 Fix a couple of profiling-related bugs in @command{gcc}
1609 @ifclear USERVISONLY
1611 Integrate GNU Ada's (GNAT's) changes to the back end,
1612 which consist almost entirely of bug fixes.
1613 These fixes are circa version 3.10p of GNAT.
1616 @ifclear USERVISONLY
1618 Include some other @command{gcc} fixes that seem useful in
1619 @command{g77}'s version of @command{gcc}.
1620 (See @file{gcc/ChangeLog} for details---compare it
1621 to that file in the vanilla @code{gcc-2.7.2.3.tar.gz}
1626 Fix @code{libU77} routines that accept file and other names
1627 to strip trailing blanks from them, for consistency
1628 with other implementations.
1629 Blanks may be forcibly appended to such names by
1630 appending a single null character (@samp{CHAR(0)})
1631 to the significant trailing blanks.
1634 Fix @code{CHMOD} intrinsic to work with file names
1635 that have embedded blanks, commas, and so on.
1638 Fix @code{SIGNAL} intrinsic so it accepts an
1639 optional third @code{Status} argument.
1641 @ifclear USERVISONLY
1643 Fix @code{IDATE()} intrinsic subroutine (VXT form)
1644 so it accepts arguments in the correct order.
1645 Documentation fixed accordingly, and for
1646 @code{GMTIME()} and @code{LTIME()} as well.
1650 Make many changes to @code{libU77} intrinsics to
1651 support existing code more directly.
1653 Such changes include allowing both subroutine and
1654 function forms of many routines, changing @code{MCLOCK()}
1655 and @code{TIME()} to return @code{INTEGER(KIND=1)} values,
1656 introducing @code{MCLOCK8()} and @code{TIME8()} to
1657 return @code{INTEGER(KIND=2)} values,
1658 and placing functions that are intended to perform
1659 side effects in a new intrinsic group, @code{badu77}.
1661 @ifclear USERVISONLY
1663 Improve @code{libU77} so it is more portable.
1667 Add options @option{-fbadu77-intrinsics-delete},
1668 @option{-fbadu77-intrinsics-hide}, and so on.
1670 @ifclear USERVISONLY
1672 Fix crashes involving diagnosed or invalid code.
1675 @ifclear USERVISONLY
1677 @command{g77} and @command{gcc} now do a somewhat better
1678 job detecting and diagnosing arrays that are too
1679 large to handle before these cause diagnostics
1680 during the assembler or linker phase, a compiler
1681 crash, or generation of incorrect code.
1684 @ifclear USERVISONLY
1686 Make some fixes to alias analysis code.
1689 @ifclear USERVISONLY
1691 Add support for @code{restrict} keyword in @command{gcc}
1695 @ifclear USERVISONLY
1697 Support @command{gcc} version 2.7.2.3
1698 (modified by @command{g77} into version 2.7.2.3.f.1),
1700 support for prior versions of @command{gcc}.
1703 @ifclear USERVISONLY
1705 Incorporate GNAT's patches to the @command{gcc} back
1706 end into @command{g77}'s, so GNAT users do not need
1707 to apply GNAT's patches to build both GNAT and @command{g77}
1708 from the same source tree.
1711 @ifclear USERVISONLY
1713 Modify @command{make} rules and related code so that
1714 generation of Info documentation doesn't require
1715 compilation using @command{gcc}.
1716 Now, any ANSI C compiler should be adequate to
1717 produce the @command{g77} documentation (in particular,
1718 the tables of intrinsics) from scratch.
1722 Add @code{INT2} and @code{INT8} intrinsics.
1725 Add @code{CPU_TIME} intrinsic.
1728 Add @code{ALARM} intrinsic.
1731 @code{CTIME} intrinsic now accepts any @code{INTEGER}
1732 argument, not just @code{INTEGER(KIND=2)}.
1734 @ifclear USERVISONLY
1736 Warn when explicit type declaration disagrees with
1737 the type of an intrinsic invocation.
1740 @ifclear USERVISONLY
1742 Support @samp{*f771} entry in @command{gcc} @file{specs} file.
1745 @ifclear USERVISONLY
1747 Fix typo in @command{make} rule @command{g77-cross}, used only for
1751 @ifclear USERVISONLY
1753 Fix @code{libf2c} build procedure to re-archive library
1754 if previous attempt to archive was interrupted.
1757 @ifclear USERVISONLY
1759 Change @command{gcc} to unroll loops only during the last
1760 invocation (of as many as two invocations) of loop
1764 @ifclear USERVISONLY
1766 Improve handling of @option{-fno-f2c} so that code that
1767 attempts to pass an intrinsic as an actual argument,
1768 such as @samp{CALL FOO(ABS)}, is rejected due to the fact
1769 that the run-time-library routine is, effectively,
1770 compiled with @option{-ff2c} in effect.
1773 @ifclear USERVISONLY
1775 Fix @command{g77} driver to recognize @option{-fsyntax-only}
1776 as an option that inhibits linking, just like @option{-c} or
1777 @option{-S}, and to recognize and properly handle the
1778 @option{-nostdlib}, @option{-M}, @option{-MM}, @option{-nodefaultlibs},
1779 and @option{-Xlinker} options.
1782 @ifclear USERVISONLY
1784 Upgrade to @code{libf2c} as of 1997-08-16.
1787 @ifclear USERVISONLY
1789 Modify @code{libf2c} to consistently and clearly diagnose
1790 recursive I/O (at run time).
1794 @command{g77} driver now prints version information (such as produced
1795 by @kbd{g77 -v}) to @code{stderr} instead of @code{stdout}.
1798 The @samp{.r} suffix now designates a Ratfor source file,
1799 to be preprocessed via the @command{ratfor} command, available
1802 @ifclear USERVISONLY
1804 Fix some aspects of how @command{gcc} determines what kind of
1805 system is being configured and what kinds are supported.
1806 For example, GNU Linux/Alpha ELF systems now are directly
1810 @ifclear USERVISONLY
1812 Improve diagnostics.
1815 @ifclear USERVISONLY
1817 Improve documentation and indexing.
1820 @ifclear USERVISONLY
1822 Include all pertinent files for @code{libf2c} that come
1823 from @code{netlib.bell-labs.com}; give any such files
1824 that aren't quite accurate in @command{g77}'s version of
1825 @code{libf2c} the suffix @samp{.netlib}.
1828 @ifclear USERVISONLY
1830 Reserve @code{INTEGER(KIND=0)} for future use.
1834 @c 1997-02-28: 0.5.20 released.
1838 The @option{-fno-typeless-boz} option is now the default.
1840 This option specifies that non-decimal-radix
1841 constants using the prefixed-radix form (such as @samp{Z'1234'})
1842 are to be interpreted as @code{INTEGER(KIND=1)} constants.
1843 Specify @option{-ftypeless-boz} to cause such
1844 constants to be interpreted as typeless.
1846 (Version 0.5.19 introduced @option{-fno-typeless-boz} and
1850 @xref{Fortran Dialect Options,,Options Controlling Fortran Dialect},
1851 for information on the @option{-ftypeless-boz} option.
1855 Options @option{-ff90-intrinsics-enable} and
1856 @option{-fvxt-intrinsics-enable} now are the
1859 Some programs might use names that clash with
1860 intrinsic names defined (and now enabled) by these
1861 options or by the new @code{libU77} intrinsics.
1862 Users of such programs might need to compile them
1863 differently (using, for example, @option{-ff90-intrinsics-disable})
1864 or, better yet, insert appropriate @code{EXTERNAL}
1865 statements specifying that these names are not intended
1866 to be names of intrinsics.
1869 The @code{ALWAYS_FLUSH} macro is no longer defined when
1870 building @code{libf2c}, which should result in improved
1871 I/O performance, especially over NFS.
1873 @emph{Note:} If you have code that depends on the behavior
1874 of @code{libf2c} when built with @code{ALWAYS_FLUSH} defined,
1875 you will have to modify @code{libf2c} accordingly before
1876 building it from this and future versions of @command{g77}.
1879 @xref{Output Assumed To Flush}, for more information.
1883 Dave Love's implementation of @code{libU77} has been
1884 added to the version of @code{libf2c} distributed with
1885 and built as part of @command{g77}.
1886 @command{g77} now knows about the routines in this library
1890 New option @option{-fvxt} specifies that the
1891 source file is written in VXT Fortran, instead of GNU Fortran.
1894 @xref{VXT Fortran}, for more information on the constructs
1895 recognized when the @option{-fvxt} option is specified.
1899 The @option{-fvxt-not-f90} option has been deleted,
1900 along with its inverse, @option{-ff90-not-vxt}.
1902 If you used one of these deleted options, you should
1903 re-read the pertinent documentation to determine which
1904 options, if any, are appropriate for compiling your
1905 code with this version of @command{g77}.
1908 @xref{Other Dialects}, for more information.
1912 The @option{-fugly} option now issues a warning, as it
1913 likely will be removed in a future version.
1915 (Enabling all the @option{-fugly-*} options is unlikely
1916 to be feasible, or sensible, in the future,
1917 so users should learn to specify only those
1918 @option{-fugly-*} options they really need for a
1919 particular source file.)
1922 The @option{-fugly-assumed} option, introduced in
1923 version 0.5.19, has been changed to
1924 better accommodate old and new code.
1927 @xref{Ugly Assumed-Size Arrays}, for more information.
1930 @ifclear USERVISONLY
1932 Make a number of fixes to the @command{g77} front end and
1933 the @command{gcc} back end to better support Alpha (AXP)
1935 This includes providing at least one bug-fix to the
1936 @command{gcc} back end for Alphas.
1940 Related to supporting Alpha (AXP) machines, the @code{LOC()}
1941 intrinsic and @code{%LOC()} construct now return
1942 values of @code{INTEGER(KIND=0)} type,
1943 as defined by the GNU Fortran language.
1945 This type is wide enough
1946 (holds the same number of bits)
1947 as the character-pointer type on the machine.
1949 On most machines, this won't make a difference,
1950 whereas, on Alphas and other systems with 64-bit pointers,
1951 the @code{INTEGER(KIND=0)} type is equivalent to @code{INTEGER(KIND=2)}
1952 (often referred to as @code{INTEGER*8})
1953 instead of the more common @code{INTEGER(KIND=1)}
1954 (often referred to as @code{INTEGER*4}).
1957 Emulate @code{COMPLEX} arithmetic in the @command{g77} front
1958 end, to avoid bugs in @code{complex} support in the
1959 @command{gcc} back end.
1960 New option @option{-fno-emulate-complex}
1961 causes @command{g77} to revert the 0.5.19 behavior.
1963 @ifclear USERVISONLY
1965 Fix bug whereby @samp{REAL A(1)}, for example, caused
1966 a compiler crash if @option{-fugly-assumed} was in effect
1967 and @var{A} was a local (automatic) array.
1968 That case is no longer affected by the new
1969 handling of @option{-fugly-assumed}.
1972 @ifclear USERVISONLY
1974 Fix @command{g77} command driver so that @samp{g77 -o foo.f}
1975 no longer deletes @file{foo.f} before issuing other
1976 diagnostics, and so the @option{-x} option is properly
1980 @ifclear USERVISONLY
1982 Enable inlining of subroutines and functions by the @command{gcc}
1984 This works as it does for @command{gcc} itself---program units
1985 may be inlined for invocations that follow them in the same
1986 program unit, as long as the appropriate compile-time
1987 options are specified.
1991 Dummy arguments are no longer assumed to potentially alias
1993 other dummy arguments or @code{COMMON} areas when any of
1994 these are defined (assigned to) by Fortran code.
1996 This can result in faster and/or smaller programs when
1997 compiling with optimization enabled, though on some
1998 systems this effect is observed only when @option{-fforce-addr}
2001 New options @option{-falias-check}, @option{-fargument-alias},
2002 @option{-fargument-noalias},
2003 and @option{-fno-argument-noalias-global} control the
2004 way @command{g77} handles potential aliasing.
2007 @xref{Aliasing Assumed To Work}, for detailed information on why the
2008 new defaults might result in some programs no longer working the way they
2009 did when compiled by previous versions of @command{g77}.
2012 @ifclear USERVISONLY
2014 The @code{CONJG()} and @code{DCONJG()} intrinsics now
2015 are compiled in-line.
2018 @ifclear USERVISONLY
2020 The bug-fix for 0.5.19.1 has been re-done.
2021 The @command{g77} compiler has been changed back to
2022 assume @code{libf2c} has no aliasing problems in
2023 its implementations of the @code{COMPLEX} (and
2024 @code{DOUBLE COMPLEX}) intrinsics.
2025 The @code{libf2c} has been changed to have no such
2028 As a result, 0.5.20 is expected to offer improved performance
2029 over 0.5.19.1, perhaps as good as 0.5.19 in most
2030 or all cases, due to this change alone.
2032 @emph{Note:} This change requires version 0.5.20 of
2033 @code{libf2c}, at least, when linking code produced
2034 by any versions of @command{g77} other than 0.5.19.1.
2035 Use @samp{g77 -v} to determine the version numbers
2036 of the @code{libF77}, @code{libI77}, and @code{libU77}
2037 components of the @code{libf2c} library.
2038 (If these version numbers are not printed---in
2039 particular, if the linker complains about unresolved
2040 references to names like @samp{g77__fvers__}---that
2041 strongly suggests your installation has an obsolete
2042 version of @code{libf2c}.)
2046 New option @option{-fugly-assign} specifies that the
2047 same memory locations are to be used to hold the
2048 values assigned by both statements @samp{I = 3} and
2049 @samp{ASSIGN 10 TO I}, for example.
2050 (Normally, @command{g77} uses a separate memory location
2051 to hold assigned statement labels.)
2054 @xref{Ugly Assigned Labels}, for more information.
2058 @code{FORMAT} and @code{ENTRY} statements now are allowed to
2059 precede @code{IMPLICIT NONE} statements.
2061 @ifclear USERVISONLY
2063 Produce diagnostic for unsupported @code{SELECT CASE} on
2064 @code{CHARACTER} type, instead of crashing, at compile time.
2067 @ifclear USERVISONLY
2069 Fix crashes involving diagnosed or invalid code.
2072 @ifclear USERVISONLY
2074 Change approach to building @code{libf2c} archive
2075 (@file{libf2c.a}) so that members are added to it
2076 only when truly necessary, so the user that installs
2077 an already-built @command{g77} doesn't need to have write
2078 access to the build tree (whereas the user doing the
2079 build might not have access to install new software
2083 @ifclear USERVISONLY
2085 Support @command{gcc} version 2.7.2.2
2086 (modified by @command{g77} into version 2.7.2.2.f.2),
2088 support for prior versions of @command{gcc}.
2091 @ifclear USERVISONLY
2093 Upgrade to @code{libf2c} as of 1997-02-08, and
2094 fix up some of the build procedures.
2097 @ifclear USERVISONLY
2099 Improve general build procedures for @command{g77},
2100 fixing minor bugs (such as deletion of any file
2101 named @file{f771} in the parent directory of @code{gcc/}).
2105 Enable full support of @code{INTEGER(KIND=2)}
2106 (often referred to as @code{INTEGER*8})
2108 @code{libf2c} and @file{f2c.h} so that @command{f2c} users
2109 may make full use of its features via the @command{g77}
2110 version of @file{f2c.h} and the @code{INTEGER(KIND=2)}
2111 support routines in the @command{g77} version of @code{libf2c}.
2114 Improve @command{g77} driver and @code{libf2c} so that @samp{g77 -v}
2115 yields version information on the library.
2118 The @code{SNGL} and @code{FLOAT} intrinsics now are
2119 specific intrinsics, instead of synonyms for the
2120 generic intrinsic @code{REAL}.
2123 New intrinsics have been added.
2124 These are @code{REALPART}, @code{IMAGPART},
2126 @code{LONG}, and @code{SHORT}.
2129 A new group of intrinsics, @code{gnu}, has been added
2130 to contain the new @code{REALPART}, @code{IMAGPART},
2131 and @code{COMPLEX} intrinsics.
2132 An old group, @code{dcp}, has been removed.
2135 Complain about industry-wide ambiguous references
2136 @samp{REAL(@var{expr})} and @samp{AIMAG(@var{expr})},
2137 where @var{expr} is @code{DOUBLE COMPLEX} (or any
2138 complex type other than @code{COMPLEX}), unless
2139 @option{-ff90} option specifies Fortran 90 interpretation
2140 or new @option{-fugly-complex} option, in conjunction with
2141 @option{-fnot-f90}, specifies @command{f2c} interpretation.
2143 @ifclear USERVISONLY
2145 Make improvements to diagnostics.
2148 @ifclear USERVISONLY
2150 Speed up compiler a bit.
2153 @ifclear USERVISONLY
2155 Improvements to documentation and indexing, including
2156 a new chapter containing information on one, later
2157 more, diagnostics that users are directed to pull
2158 up automatically via a message in the diagnostic itself.
2160 (Hence the menu item @code{M} for the node
2161 @code{Diagnostics} in the top-level menu of
2162 the Info documentation.)
2166 @ifclear DOC-OLDNEWS
2167 @heading In previous versions:
2169 Information on previous versions is archived
2170 in @file{@value{path-g77}/news.texi}
2171 following the test of the @code{DOC-OLDNEWS} macro.
2175 @c 1997-02-01: 0.5.19.1 released.
2176 @heading In 0.5.19.1:
2179 Code-generation bugs afflicting operations on complex
2180 data have been fixed.
2182 These bugs occurred when assigning the result of an
2183 operation to a complex variable (or array element)
2184 that also served as an input to that operation.
2186 The operations affected by this bug were: @code{CONJG()},
2187 @code{DCONJG()}, @code{CCOS()}, @code{CDCOS()},
2188 @code{CLOG()}, @code{CDLOG()}, @code{CSIN()}, @code{CDSIN()},
2189 @code{CSQRT()}, @code{CDSQRT()}, complex division, and
2190 raising a @code{DOUBLE COMPLEX} operand to an @code{INTEGER}
2192 (The related generic and @samp{Z}-prefixed intrinsics,
2193 such as @code{ZSIN()}, also were affected.)
2195 For example, @samp{C = CSQRT(C)}, @samp{Z = Z/C}, and @samp{Z = Z**I}
2196 (where @samp{C} is @code{COMPLEX} and @samp{Z} is
2197 @code{DOUBLE COMPLEX}) have been fixed.
2200 @c 1996-12-07: 0.5.19 released.
2204 Fix @code{FORMAT} statement parsing so negative values for
2205 specifiers such as @code{P} (e.g. @samp{FORMAT(-1PF8.1)})
2206 are correctly processed as negative.
2209 Fix @code{SIGNAL} intrinsic so it once again accepts a
2210 procedure as its second argument.
2213 A temporary kludge option provides bare-bones information on
2214 @code{COMMON} and @code{EQUIVALENCE} members at debug time.
2217 New @option{-fonetrip} option specifies FORTRAN-66-style
2218 one-trip @code{DO} loops.
2221 New @option{-fno-silent} option causes names of program units
2222 to be printed as they are compiled, in a fashion similar to
2223 UNIX @command{f77} and @command{f2c}.
2226 New @option{-fugly-assumed} option specifies that arrays
2227 dimensioned via @samp{DIMENSION X(1)}, for example, are to be
2228 treated as assumed-size.
2231 New @option{-fno-typeless-boz} option specifies that non-decimal-radix
2232 constants using the prefixed-radix form (such as @samp{Z'1234'})
2233 are to be interpreted as @code{INTEGER} constants.
2236 New @option{-ff66} option is a ``shorthand'' option that specifies
2237 behaviors considered appropriate for FORTRAN 66 programs.
2240 New @option{-ff77} option is a ``shorthand'' option that specifies
2241 behaviors considered appropriate for UNIX @command{f77} programs.
2244 New @option{-fugly-comma} and @option{-fugly-logint} options provided
2245 to perform some of what @option{-fugly} used to do.
2246 @option{-fugly} and @option{-fno-ugly} are now ``shorthand'' options,
2247 in that they do nothing more than enable (or disable) other
2248 @option{-fugly-*} options.
2251 Fix parsing of assignment statements involving targets that
2252 are substrings of elements of @code{CHARACTER} arrays having
2253 names such as @samp{READ}, @samp{WRITE}, @samp{GOTO}, and
2254 @samp{REALFUNCTIONFOO}.
2257 Fix crashes involving diagnosed code.
2260 Fix handling of local @code{EQUIVALENCE} areas so certain cases
2261 of valid Fortran programs are not misdiagnosed as improperly
2262 extending the area backwards.
2265 Support @command{gcc} version 2.7.2.1.
2268 Upgrade to @code{libf2c} as of 1996-09-26, and
2269 fix up some of the build procedures.
2272 Change code generation for list-directed I/O so it allows
2273 for new versions of @code{libf2c} that might return non-zero
2274 status codes for some operations previously assumed to always
2277 This change not only affects how @code{IOSTAT=} variables
2278 are set by list-directed I/O, it also affects whether
2279 @code{END=} and @code{ERR=} labels are reached by these
2283 Add intrinsic support for new @code{FTELL} and @code{FSEEK}
2284 procedures in @code{libf2c}.
2287 Modify @code{fseek_()} in @code{libf2c} to be more portable
2288 (though, in practice, there might be no systems where this
2289 matters) and to catch invalid @code{whence} arguments.
2292 Some useless warnings from the @option{-Wunused} option have
2296 Fix a problem building the @file{f771} executable
2297 on AIX systems by linking with the @option{-bbigtoc} option.
2300 Abort configuration if @command{gcc} has not been patched
2301 using the patch file provided in the @file{gcc/f/gbe/}
2305 Add options @option{--help} and @option{--version} to the
2306 @command{g77} command, to conform to GNU coding guidelines.
2307 Also add printing of @command{g77} version number when
2308 the @option{--verbose} (@option{-v}) option is used.
2311 Change internally generated name for local @code{EQUIVALENCE}
2312 areas to one based on the alphabetically sorted first name
2313 in the list of names for entities placed at the beginning
2317 Improvements to documentation and indexing.
2320 @c 1996-04-01: 0.5.18 released.
2324 Add some rudimentary support for @code{INTEGER*1},
2325 @code{INTEGER*2}, @code{INTEGER*8},
2326 and their @code{LOGICAL} equivalents.
2327 (This support works on most, maybe all, @command{gcc} targets.)
2329 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2330 for providing the patch for this!
2332 Among the missing elements from the support for these
2333 features are full intrinsic support and constants.
2336 Add some rudimentary support for the @code{BYTE} and
2337 @code{WORD} type-declaration statements.
2338 @code{BYTE} corresponds to @code{INTEGER*1},
2339 while @code{WORD} corresponds to @code{INTEGER*2}.
2341 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2342 for providing the patch for this!
2345 The compiler code handling intrinsics has been largely
2346 rewritten to accommodate the new types.
2347 No new intrinsics or arguments for existing
2348 intrinsics have been added, so there is, at this
2349 point, no intrinsic to convert to @code{INTEGER*8},
2353 Support automatic arrays in procedures.
2356 Reduce space/time requirements for handling large
2357 @emph{sparsely} initialized aggregate arrays.
2358 This improvement applies to only a subset of
2359 the general problem to be addressed in 0.6.
2362 Treat initial values of zero as if they weren't
2363 specified (in DATA and type-declaration statements).
2364 The initial values will be set to zero anyway, but the amount
2365 of compile time processing them will be reduced,
2366 in some cases significantly (though, again, this
2367 is only a subset of the general problem to be
2370 A new option, @option{-fzeros}, is introduced to
2371 enable the traditional treatment of zeros as any
2375 With @option{-ff90} in force, @command{g77} incorrectly
2376 interpreted @samp{REAL(Z)} as returning a @code{REAL}
2377 result, instead of as a @code{DOUBLE PRECISION}
2379 (Here, @samp{Z} is @code{DOUBLE COMPLEX}.)
2381 With @option{-fno-f90} in force, the interpretation remains
2382 unchanged, since this appears to be how at least some
2383 F77 code using the @code{DOUBLE COMPLEX} extension expected
2386 Essentially, @samp{REAL(Z)} in F90 is the same as
2387 @samp{DBLE(Z)}, while in extended F77, it appears to
2388 be the same as @samp{REAL(REAL(Z))}.
2391 An expression involving exponentiation, where both operands
2392 were type @code{INTEGER} and the right-hand operand
2393 was negative, was erroneously evaluated.
2396 Fix bugs involving @code{DATA} implied-@code{DO} constructs
2397 (these involved an errant diagnostic and a crash, both on good
2398 code, one involving subsequent statement-function definition).
2401 Close @code{INCLUDE} files after processing them, so compiling source
2402 files with lots of @code{INCLUDE} statements does not result in
2403 being unable to open @code{INCLUDE} files after all the available
2404 file descriptors are used up.
2407 Speed up compiling, especially of larger programs, and perhaps
2408 slightly reduce memory utilization while compiling (this is
2409 @emph{not} the improvement planned for 0.6 involving large aggregate
2410 areas)---these improvements result from simply turning
2411 off some low-level code to do self-checking that hasn't been
2412 triggered in a long time.
2415 Introduce three new options that
2416 implement optimizations in the @command{gcc} back end (GBE).
2417 These options are @option{-fmove-all-movables}, @option{-freduce-all-givs},
2418 and @option{-frerun-loop-opt}, which are enabled, by default,
2419 for Fortran compilations.
2420 These optimizations are intended to help toon Fortran programs.
2423 Patch the GBE to do a better job optimizing certain
2424 kinds of references to array elements.
2427 Due to patches to the GBE, the version number of @command{gcc}
2428 also is patched to make it easier to manage installations,
2429 especially useful if it turns out a @command{g77} change to the
2432 The @command{g77}-modified version number is the @command{gcc}
2433 version number with the string @samp{.f.@var{n}} appended,
2434 where @samp{f} identifies the version as enhanced for
2435 Fortran, and @var{n} is @samp{1} for the first Fortran
2436 patch for that version of @command{gcc}, @samp{2} for the
2439 So, this introduces version 2.7.2.f.1 of @command{gcc}.
2442 Make several improvements and fixes to diagnostics, including
2443 the removal of two that were inappropriate or inadequate.
2446 Warning about two successive arithmetic operators, produced
2447 by @option{-Wsurprising}, now produced @emph{only} when both
2448 operators are, indeed, arithmetic (not relational/boolean).
2451 @option{-Wsurprising} now warns about the remaining cases
2452 of using non-integral variables for implied-@code{DO}
2453 loops, instead of these being rejected unless @option{-fpedantic}
2454 or @option{-fugly} specified.
2457 Allow @code{SAVE} of a local variable or array, even after
2458 it has been given an initial value via @code{DATA}, for example.
2461 Introduce an Info version of @command{g77} documentation, which
2462 supersedes @file{gcc/f/CREDITS}, @file{gcc/f/DOC}, and
2463 @file{gcc/f/PROJECTS}.
2464 These files will be removed in a future release.
2465 The files @file{gcc/f/BUGS}, @file{gcc/f/INSTALL}, and
2466 @file{gcc/f/NEWS} now are automatically built from
2467 the texinfo source when distributions are made.
2469 This effort was inspired by a first pass at translating
2470 @file{g77-0.5.16/f/DOC} that was contributed to Craig by
2471 David Ronis (@email{ronis@@onsager.chem.mcgill.ca}).
2474 New @option{-fno-second-underscore} option to specify
2475 that, when @option{-funderscoring} is in effect, a second
2476 underscore is not to be appended to Fortran names already
2477 containing an underscore.
2480 Change the way iterative @code{DO} loops work to follow
2482 In particular, calculation of the iteration count is
2483 still done by converting the start, end, and increment
2484 parameters to the type of the @code{DO} variable, but
2485 the result of the calculation is always converted to
2486 the default @code{INTEGER} type.
2488 (This should have no effect on existing code compiled
2489 by @command{g77}, but code written to assume that use
2490 of a @emph{wider} type for the @code{DO} variable
2491 will result in an iteration count being fully calculated
2492 using that wider type (wider
2493 than default @code{INTEGER}) must be rewritten.)
2496 Support @command{gcc} version 2.7.2.
2499 Upgrade to @code{libf2c} as of 1996-03-23, and
2500 fix up some of the build procedures.
2502 Note that the email addresses related to @command{f2c}
2503 have changed---the distribution site now is
2504 named @code{netlib.bell-labs.com}, and the
2505 maintainer's new address is @email{dmg@@bell-labs.com}.
2508 @c 1995-11-18: 0.5.17 released.
2512 @strong{Fix serious bug} in @samp{g77 -v} command that can cause removal of a
2513 system's @file{/dev/null} special file if run by user @code{root}.
2515 @strong{All users} of version 0.5.16 should ensure that
2516 they have not removed @file{/dev/null} or replaced it with an ordinary
2517 file (e.g. by comparing the output of @samp{ls -l /dev/null} with
2518 @samp{ls -l /dev/zero}.
2519 If the output isn't basically the
2520 same, contact your system
2521 administrator about restoring @file{/dev/null} to its proper status).
2523 This bug is particularly insidious because removing @file{/dev/null} as
2524 a special file can go undetected for quite a while, aside from
2525 various applications and programs exhibiting sudden, strange
2528 I sincerely apologize for not realizing the
2529 implications of the fact that when @samp{g77 -v} runs the @command{ld} command
2530 with @samp{-o /dev/null} that @command{ld} tries to @emph{remove} the executable
2531 it is supposed to build (especially if it reports unresolved
2532 references, which it should in this case)!
2535 Fix crash on @samp{CHARACTER*(*) FOO} in a main or block data program unit.
2538 Fix crash that can occur when diagnostics given outside of any
2539 program unit (such as when input file contains @samp{@@foo}).
2542 Fix crashes, infinite loops (hangs), and such involving diagnosed code.
2545 Fix @code{ASSIGN}'ed variables so they can be @code{SAVE}'d or dummy arguments,
2546 and issue clearer error message in cases where target of @code{ASSIGN}
2547 or @code{ASSIGN}ed @code{GOTO}/@code{FORMAT} is too small (which should
2551 Make @code{libf2c} build procedures work on more systems again by
2552 eliminating unnecessary invocations of @samp{ld -r -x} and @command{mv}.
2555 Fix omission of @option{-funix-intrinsics-@dots{}} options in list of permitted
2556 options to compiler.
2559 Fix failure to always diagnose missing type declaration for
2560 @code{IMPLICIT NONE}.
2563 Fix compile-time performance problem (which could sometimes
2564 crash the compiler, cause a hang, or whatever, due to a bug
2565 in the back end) involving exponentiation with a large @code{INTEGER}
2566 constant for the right-hand operator (e.g. @samp{I**32767}).
2569 Fix build procedures so cross-compiling @command{g77} (the @command{fini}
2570 utility in particular) is properly built using the host compiler.
2573 Add new @option{-Wsurprising} option to warn about constructs that are
2574 interpreted by the Fortran standard (and @command{g77}) in ways that
2575 are surprising to many programmers.
2578 Add @code{ERF()} and @code{ERFC()} as generic intrinsics mapping to existing
2579 @code{ERF}/@code{DERF} and @code{ERFC}/@code{DERFC} specific intrinsics.
2581 @emph{Note:} You should
2582 specify @samp{INTRINSIC ERF,ERFC} in any code where you might use
2583 these as generic intrinsics, to improve likelihood of diagnostics
2584 (instead of subtle run-time bugs) when using a compiler that
2585 doesn't support these as intrinsics (e.g. @command{f2c}).
2588 Remove from @option{-fno-pedantic} the diagnostic about @code{DO}
2589 with non-@code{INTEGER} index variable; issue that under
2590 @option{-Wsurprising} instead.
2593 Clarify some diagnostics that say things like ``ignored'' when that's
2597 Clarify diagnostic on use of @code{.EQ.}/@code{.NE.} on @code{LOGICAL}
2601 Minor improvements to code generation for various operations on
2602 @code{LOGICAL} operands.
2605 Minor improvement to code generation for some @code{DO} loops on some
2609 Support @command{gcc} version 2.7.1.
2612 Upgrade to @code{libf2c} as of 1995-11-15.
2615 @c 1995-08-30: 0.5.16 released.
2619 Fix a code-generation bug involving complicated @code{EQUIVALENCE} statements
2620 not involving @code{COMMON}.
2623 Fix code-generation bugs involving invoking ``gratis'' library procedures
2624 in @code{libf2c} from code compiled with @option{-fno-f2c} by making these
2625 procedures known to @command{g77} as intrinsics (not affected by -fno-f2c).
2626 This is known to fix code invoking @code{ERF()}, @code{ERFC()},
2627 @code{DERF()}, and @code{DERFC()}.
2630 Update @code{libf2c} to include netlib patches through 1995-08-16, and
2631 @code{#define} @code{WANT_LEAD_0} to 1 to make @command{g77}-compiled code more
2632 consistent with other Fortran implementations by outputting
2633 leading zeros in formatted and list-directed output.
2636 Fix a code-generation bug involving adjustable dummy arrays with high
2637 bounds whose primaries are changed during procedure execution, and
2638 which might well improve code-generation performance for such arrays
2639 compared to @command{f2c} plus @command{gcc} (but apparently only when using
2640 @file{gcc-2.7.0} or later).
2643 Fix a code-generation bug involving invocation of @code{COMPLEX} and
2644 @code{DOUBLE COMPLEX} @code{FUNCTION}s and doing @code{COMPLEX} and
2645 @code{DOUBLE COMPLEX} divides, when the result
2646 of the invocation or divide is assigned directly to a variable
2647 that overlaps one or more of the arguments to the invocation or divide.
2650 Fix crash by not generating new optimal code for @samp{X**I} if @samp{I} is
2651 nonconstant and the expression is used to dimension a dummy
2652 array, since the @command{gcc} back end does not support the necessary
2653 mechanics (and the @command{gcc} front end rejects the equivalent
2654 construct, as it turns out).
2657 Fix crash on expressions like @samp{COMPLEX**INTEGER}.
2660 Fix crash on expressions like @samp{(1D0,2D0)**2}, i.e. raising a
2661 @code{DOUBLE COMPLEX} constant to an @code{INTEGER} constant power.
2664 Fix crashes and such involving diagnosed code.
2667 Diagnose, instead of crashing on, statement function definitions
2668 having duplicate dummy argument names.
2671 Fix bug causing rejection of good code involving statement function
2675 Fix bug resulting in debugger not knowing size of local equivalence
2676 area when any member of area has initial value (via @code{DATA},
2680 Fix installation bug that prevented installation of @command{g77} driver.
2681 Provide for easy selection of whether to install copy of @command{g77}
2682 as @command{f77} to replace the broken code.
2685 Fix @command{gcc} driver (affects @command{g77} thereby) to not
2686 gratuitously invoke the
2687 @code{f771} program (e.g. when @option{-E} is specified).
2690 Fix diagnostic to point to correct source line when it immediately
2691 follows an @code{INCLUDE} statement.
2694 Support more compiler options in @command{gcc}/@command{g77} when
2695 compiling Fortran files.
2696 These options include @option{-p}, @option{-pg}, @option{-aux-info}, @option{-P},
2697 correct setting of version-number macros for preprocessing, full
2698 recognition of @option{-O0}, and
2699 automatic insertion of configuration-specific linker specs.
2702 Add new intrinsics that interface to existing routines in @code{libf2c}:
2703 @code{ABORT}, @code{DERF}, @code{DERFC}, @code{ERF}, @code{ERFC}, @code{EXIT},
2704 @code{FLUSH}, @code{GETARG}, @code{GETENV}, @code{IARGC},
2705 @code{SIGNAL}, and @code{SYSTEM}.
2706 Note that @code{ABORT}, @code{EXIT}, @code{FLUSH}, @code{SIGNAL}, and
2707 @code{SYSTEM} are intrinsic subroutines, not functions (since they
2708 have side effects), so to get the return values from @code{SIGNAL}
2709 and @code{SYSTEM}, append a final argument specifying an @code{INTEGER}
2710 variable or array element (e.g. @samp{CALL SYSTEM('rm foo',ISTAT)}).
2713 Add new intrinsic group named @code{unix} to contain the new intrinsics,
2714 and by default enable this new group.
2717 Move @code{LOC()} intrinsic out of the @code{vxt} group to the new
2721 Improve @command{g77} so that @samp{g77 -v} by itself (or with
2722 certain other options, including @option{-B}, @option{-b}, @option{-i},
2723 @option{-nostdlib}, and @option{-V}) reports lots more useful
2724 version info, and so that long-form options @command{gcc} accepts are
2725 understood by @command{g77} as well (even in truncated, unambiguous forms).
2728 Add new @command{g77} option @option{--driver=name} to specify driver when
2729 default, @command{gcc}, isn't appropriate.
2732 Add support for @samp{#} directives (as output by the preprocessor) in the
2733 compiler, and enable generation of those directives by the
2734 preprocessor (when compiling @samp{.F} files) so diagnostics and debugging
2735 info are more useful to users of the preprocessor.
2738 Produce better diagnostics, more like @command{gcc}, with info such as
2739 @samp{In function `foo':} and @samp{In file included from...:}.
2742 Support @command{gcc}'s @option{-fident} and @option{-fno-ident} options.
2745 When @option{-Wunused} in effect, don't warn about local variables used as
2746 statement-function dummy arguments or @code{DATA} implied-@code{DO} iteration
2747 variables, even though, strictly speaking, these are not uses
2748 of the variables themselves.
2751 When @samp{-W -Wunused} in effect, don't warn about unused dummy arguments
2752 at all, since there's no way to turn this off for individual
2753 cases (@command{g77} might someday start warning about these)---applies
2754 to @command{gcc} versions 2.7.0 and later, since earlier versions didn't
2755 warn about unused dummy arguments.
2758 New option @option{-fno-underscoring} that inhibits transformation of names
2759 (by appending one or two underscores) so users may experiment
2760 with implications of such an environment.
2763 Minor improvement to @file{gcc/f/info} module to make it easier to build
2764 @command{g77} using the native (non-@command{gcc}) compiler on certain machines
2765 (but definitely not all machines nor all non-@command{gcc} compilers).
2767 do not report bugs showing problems compilers have with
2768 macros defined in @file{gcc/f/target.h} and used in places like
2769 @file{gcc/f/expr.c}.
2772 Add warning to be printed for each invocation of the compiler
2773 if the target machine @code{INTEGER}, @code{REAL}, or @code{LOGICAL} size
2775 since @command{g77} is known to not work well for such cases.
2778 Lots of new documentation (though work is still needed to put it into
2779 canonical GNU format).
2782 Build @code{libf2c} with @option{-g0}, not @option{-g2}, in effect
2783 (by default), to produce
2784 smaller library without lots of debugging clutter.
2787 @c 1995-05-19: 0.5.15 released.
2791 Fix bad code generation involving @samp{X**I} and temporary, internal variables
2792 generated by @command{g77} and the back end (such as for @code{DO} loops).
2795 Fix crash given @samp{CHARACTER A;DATA A/.TRUE./}.
2798 Replace crash with diagnostic given @samp{CHARACTER A;DATA A/1.0/}.
2801 Fix crash or other erratic behavior when null character constant
2802 (@samp{''}) is encountered.
2805 Fix crash or other erratic behavior involving diagnosed code.
2808 Fix code generation for external functions returning type @code{REAL} when
2809 the @option{-ff2c} option is in force (which it is by default) so that
2810 @command{f2c} compatibility is indeed provided.
2813 Disallow @samp{COMMON I(10)} if @samp{I} has previously been specified
2814 with an array declarator.
2817 New @option{-ffixed-line-length-@var{n}} option, where @var{n} is the
2819 of a typical fixed-form line, defaulting to 72 columns, such
2820 that characters beyond column @var{n} are ignored, or @var{n} is @samp{none},
2821 meaning no characters are ignored.
2822 does not affect lines
2823 with @samp{&} in column 1, which are always processed as if
2824 @option{-ffixed-line-length-none} was in effect.
2827 No longer generate better code for some kinds of array references,
2828 as @command{gcc} back end is to be fixed to do this even better, and it
2829 turned out to slow down some code in some cases after all.
2832 In @code{COMMON} and @code{EQUIVALENCE} areas with any members given initial
2833 values (e.g. via @code{DATA}), uninitialized members now always
2834 initialized to binary zeros (though this is not required by
2835 the standard, and might not be done in future versions
2837 Previously, in some @code{COMMON}/@code{EQUIVALENCE} areas
2838 (essentially those with members of more than one type), the
2839 uninitialized members were initialized to spaces, to
2840 cater to @code{CHARACTER} types, but it seems no existing code expects
2841 that, while much existing code expects binary zeros.
2847 Don't emit bad code when low bound of adjustable array is nonconstant
2848 and thus might vary as an expression at run time.
2851 Emit correct code for calculation of number of trips in @code{DO} loops
2853 where the loop should not execute at all.
2854 (This bug affected cases
2855 where the difference between the begin and end values was less
2856 than the step count, though probably not for floating-point cases.)
2859 Fix crash when extra parentheses surround item in
2860 @code{DATA} implied-@code{DO} list.
2863 Fix crash over minor internal inconsistencies in handling diagnostics,
2864 just substitute dummy strings where necessary.
2867 Fix crash on some systems when compiling call to @code{MVBITS()} intrinsic.
2870 Fix crash on array assignment @samp{TYPE@var{ddd}(@dots{})=@dots{}}, where @var{ddd}
2871 is a string of one or more digits.
2874 Fix crash on @code{DCMPLX()} with a single @code{INTEGER} argument.
2877 Fix various crashes involving code with diagnosed errors.
2880 Support @option{-I} option for @code{INCLUDE} statement, plus @command{gcc}'s
2881 @file{header.gcc} facility for handling systems like MS-DOS.
2884 Allow @code{INCLUDE} statement to be continued across multiple lines,
2885 even allow it to coexist with other statements on the same line.
2888 Incorporate Bellcore fixes to @code{libf2c} through 1995-03-15---this
2889 fixes a bug involving infinite loops reading EOF with empty list-directed
2893 Remove all the @command{g77}-specific auto-configuration scripts, code,
2895 except for temporary substitutes for bsearch() and strtoul(), as
2896 too many configure/build problems were reported in these areas.
2897 People will have to fix their systems' problems themselves, or at
2898 least somewhere other than @command{g77}, which expects a working ANSI C
2899 environment (and, for now, a GNU C compiler to compile @command{g77} itself).
2902 Complain if initialized common redeclared as larger in subsequent program
2906 Warn if blank common initialized, since its size can vary and hence
2907 related warnings that might be helpful won't be seen.
2910 New @option{-fbackslash} option, on by default, that causes @samp{\}
2911 within @code{CHARACTER}
2912 and Hollerith constants to be interpreted a la GNU C.
2914 this behavior is somewhat different from @command{f2c}'s, which supports only
2915 a limited subset of backslash (escape) sequences.
2918 Make @option{-fugly-args} the default.
2921 New @option{-fugly-init} option, on by default, that allows typeless/Hollerith
2922 to be specified as initial values for variables or named constants
2923 (@code{PARAMETER}), and also allows character<->numeric conversion in
2924 those contexts---turn off via @option{-fno-ugly-init}.
2927 New @option{-finit-local-zero} option to initialize
2928 local variables to binary zeros.
2929 This does not affect whether they are @code{SAVE}d, i.e. made
2930 automatic or static.
2933 New @option{-Wimplicit} option to warn about implicitly typed variables, arrays,
2935 (Basically causes all program units to default to @code{IMPLICIT NONE}.)
2938 @option{-Wall} now implies @option{-Wuninitialized} as with @command{gcc}
2939 (i.e. unless @option{-O} not specified, since @option{-Wuninitialized}
2940 requires @option{-O}), and implies @option{-Wunused} as well.
2943 @option{-Wunused} no longer gives spurious messages for unused
2944 @code{EXTERNAL} names (since they are assumed to refer to block data
2945 program units, to make use of libraries more reliable).
2948 Support @code{%LOC()} and @code{LOC()} of character arguments.
2951 Support null (zero-length) character constants and expressions.
2954 Support @command{f2c}'s @code{IMAG()} generic intrinsic.
2957 Support @code{ICHAR()}, @code{IACHAR()}, and @code{LEN()} of
2958 character expressions that are valid in assignments but
2959 not normally as actual arguments.
2962 Support @command{f2c}-style @samp{&} in column 1 to mean continuation line.
2965 Allow @code{NAMELIST}, @code{EXTERNAL}, @code{INTRINSIC}, and @code{VOLATILE}
2966 in @code{BLOCK DATA}, even though these are not allowed by the standard.
2969 Allow @code{RETURN} in main program unit.
2972 Changes to Hollerith-constant support to obey Appendix C of the
2977 Now padded on the right with zeros, not spaces.
2980 Hollerith ``format specifications'' in the form of arrays of
2981 non-character allowed.
2984 Warnings issued when non-space truncation occurs when converting
2988 When specified as actual argument, now passed
2989 by reference to @code{INTEGER} (padded on right with spaces if constant
2990 too small, otherwise fully intact if constant wider the @code{INTEGER}
2991 type) instead of by value.
2994 @strong{Warning:} @command{f2c} differs on the
2995 interpretation of @samp{CALL FOO(1HX)}, which it treats exactly the
2996 same as @samp{CALL FOO('X')}, but which the standard and @command{g77} treat
2997 as @samp{CALL FOO(%REF('X '))} (padded with as many spaces as necessary
2998 to widen to @code{INTEGER}), essentially.
3001 Changes and fixes to typeless-constant support:
3005 Now treated as a typeless double-length @code{INTEGER} value.
3008 Warnings issued when overflow occurs.
3011 Padded on the left with zeros when converting
3015 Should be properly aligned and ordered on
3016 the target machine for whatever type it is turned into.
3019 When specified as actual argument, now passed as reference to
3020 a default @code{INTEGER} constant.
3024 @code{%DESCR()} of a non-@code{CHARACTER} expression now passes a pointer to
3025 the expression plus a length for the expression just as if
3026 it were a @code{CHARACTER} expression.
3027 For example, @samp{CALL FOO(%DESCR(D))}, where
3028 @samp{D} is @code{REAL*8}, is the same as @samp{CALL FOO(D,%VAL(8)))}.
3031 Name of multi-entrypoint master function changed to incorporate
3032 the name of the primary entry point instead of a decimal
3033 value, so the name of the master function for @samp{SUBROUTINE X}
3034 with alternate entry points is now @samp{__g77_masterfun_x}.
3037 Remove redundant message about zero-step-count @code{DO} loops.
3040 Clean up diagnostic messages, shortening many of them.
3043 Fix typo in @command{g77} man page.
3046 Clarify implications of constant-handling bugs in @file{f/BUGS}.
3049 Generate better code for @samp{**} operator with a right-hand operand of
3050 type @code{INTEGER}.
3053 Generate better code for @code{SQRT()} and @code{DSQRT()},
3054 also when @option{-ffast-math}
3055 specified, enable better code generation for @code{SIN()} and @code{COS()}.
3058 Generate better code for some kinds of array references.
3061 Speed up lexing somewhat (this makes the compilation phase noticeably