1 @c Copyright (C) 1995-1999 Free Software Foundation, Inc.
2 @c This is part of the G77 manual.
3 @c For copying conditions, see the file g77.texi.
5 @c The text of this file appears in the file NEWS
6 @c in the G77 distribution, as well as in the G77 manual.
8 @c Keep this the same as the dates above, since it's used
9 @c in the standalone derivations of this file (e.g. NEWS).
10 @set copyrights-news 1995-1999
12 @set last-update-news 1999-04-23
17 @c The immediately following lines apply to the NEWS file
18 @c which is derived from this file.
19 @emph{Note:} This file is automatically generated from the files
20 @file{news0.texi} and @file{news.texi}.
21 @file{NEWS} is @emph{not} a source file,
22 although it is normally included within source distributions.
24 This file lists news about the @value{which-g77} version
25 (and some other versions) of the GNU Fortran compiler.
26 Copyright (C) @value{copyrights-news} Free Software Foundation, Inc.
27 You may copy, distribute, and modify it freely as long as you preserve
28 this copyright notice and permission notice.
31 @chapter News About GNU Fortran
37 @chapter User-visible Changes
38 @cindex versions, recent
39 @cindex recent versions
40 @cindex changes, user-visible
41 @cindex user-visible changes
43 This chapter describes changes to @code{g77} that are visible
44 to the programmers who actually write and maintain Fortran
45 code they compile with @code{g77}.
46 Information on changes to installation procedures,
47 changes to the documentation, and bug fixes is
48 not provided here, unless it is likely to affect how
50 @xref{News,,News About GNU Fortran}, for information on
51 such changes to @code{g77}.
56 @chapter News About GNU Fortran
57 @cindex versions, recent
58 @cindex recent versions
63 Changes made to recent versions of GNU Fortran are listed
64 below, with the most recent version first.
66 The changes are generally listed in order:
70 Code-generation and run-time-library bug-fixes
73 Compiler and run-time-library crashes involving valid code
80 Fixes and enhancements to existing features
92 This order is not strict---for example, some items
93 involve a combination of these elements.
96 Note that two variants of @code{g77} are tracked below.
97 The @code{egcs} variant is described vis-a-vis
98 previous versions of @code{egcs} and/or
99 an official FSF version,
102 Therefore, @code{egcs} versions sometimes have multiple listings
103 to help clarify how they differ from other versions,
104 though this can make getting a complete picture
105 of what a particular @code{egcs} version contains
106 somewhat more difficult.
109 For information on bugs in the @value{which-g77} version of @code{g77},
110 see @ref{Known Bugs,,Known Bugs In GNU Fortran}.
114 For information on bugs in the @value{which-g77} version of @code{g77},
115 see @file{@value{path-g77}/BUGS}.
119 @emph{Warning:} The information below is still under development,
120 and might not accurately reflect the @code{g77} code base
121 of which it is a part.
122 Efforts are made to keep it somewhat up-to-date,
123 but they are particularly concentrated
124 on any version of this information
125 that is distributed as part of a @emph{released} @code{g77}.
127 In particular, while this information is intended to apply to
128 the @value{which-g77} version of @code{g77},
129 only an official @emph{release} of that version
130 is expected to contain documentation that is
131 most consistent with the @code{g77} product in that version.
133 Nevertheless, information on @emph{previous} releases of @code{g77}, below,
134 is likely to be more up-to-date and accurate
135 than the equivalent information that accompanied
137 assuming the last-updated date of the information below
138 is later than the dates of those releases.
140 That's due to attempts to keep this development version
141 of news about previous @code{g77} versions up-to-date.
145 An online, ``live'' version of this document
146 (derived directly from the mainline, development version
147 of @code{g77} within @code{egcs})
149 @uref{http://egcs.cygnus.com/onlinedocs/g77_news.html}.
152 The following information was last updated on @value{last-update-news}:
154 @heading In @code{egcs} 1.2 (versus 1.1.2):
158 @code{g77} no longer generates bad code for assignments,
159 or other conversions,
160 of @code{REAL} or @code{COMPLEX} constant expressions
161 to type @code{INTEGER(KIND=2)}
162 (often referred to as @code{INTEGER*8}).
164 For example, @samp{INTEGER*8 J; J = 4E10} now works as documented.
169 Fix @code{g77} so it no longer crashes when compiling
170 I/O statements using keywords that define @code{INTEGER} values,
171 such as @samp{IOSTAT=@var{j}},
172 where @var{j} is other than default @code{INTEGER}
173 (such as @code{INTEGER*2}).
174 Instead, it issues a diagnostic.
179 Fix @code{g77} so it properly handles @samp{DATA A/@var{rpt}*@var{val}/},
180 where @var{rpt} is not default @code{INTEGER}, such as @code{INTEGER*2},
181 instead of producing a spurious diagnostic.
186 The @samp{-ax} option is now obeyed when compiling Fortran programs.
187 (It is passed to the @file{f771} driver.)
191 The new @samp{-fsubscript-check} option
192 causes @code{g77} to compile run-time bounds checks
193 of array subscripts, as well as of substring start and end points.
195 The current implementation uses the @code{libf2c}
196 library routine @code{s_rnge} to print the diagnostic.
197 Since a future version of @code{g77} might use a different implementation,
198 use the new @samp{-ff2c-subscript-check} option
199 if your application requires use of @code{s_rnge} or a compile-time diagnostic.
202 Source file names with the suffixes @samp{.FOR} and @samp{.FPP}
203 now are recognized by @code{g77}
204 as if they ended in @samp{.for} and @samp{.fpp}, respectively.
207 @code{g77} now warns about a reference to an intrinsic
208 that has an interface that is not Year 2000 (Y2K) compliant.
209 Also, the @code{libg2c} has been changed to increase the likelihood
210 of catching references to the implementations of these intrinsics
211 using the @code{EXTERNAL} mechanism
212 (which would avoid the new warnings).
215 @xref{Year 2000 (Y2K) Problems}, for more information.
220 @code{g77} now warns about a reference to a function
221 when the corresponding @emph{subsequent} function program unit
222 disagrees with the reference concerning the type of the function.
226 @samp{-fno-emulate-complex} is now the default option.
227 This should result in improved performance
228 of code that uses the @code{COMPLEX} data type.
231 @cindex double-precision performance
232 @cindex -malign-double
234 The @samp{-malign-double} option
235 now reliably aligns @emph{all} double-precision variables and arrays
236 on Intel x86 targets.
240 Even without the @samp{-malign-double} option,
241 @code{g77} reliably aligns local double-precision variables
242 that are not in @code{EQUIVALENCE} areas
243 and not @code{SAVE}'d.
248 A substantial portion of the @code{g77} front end's code-generation component
250 It now generates code using facilities more robustly supported
251 by the @code{gcc} back end.
252 One effect of this rewrite is that some codes no longer produce
253 a spurious ``label @var{lab} used before containing binding contour''
258 Support for the @samp{-fugly} option has been removed.
262 Improve documentation and indexing,
263 including information on Year 2000 (Y2K) compliance,
264 and providing more information on internals of the front end.
269 Upgrade to @code{libf2c} as of 1999-03-17.
273 @heading In 0.5.24 and @code{egcs} 1.1.2 (versus 0.5.23 and 1.1.1):
277 Fix the @code{IDate} intrinsic (VXT) (in @code{libg2c})
278 so the returned year is in the documented, non-Y2K-compliant range
280 instead of being returned as 100 in the year 2000.
283 @xref{IDate Intrinsic (VXT)},
284 for more information.
288 Fix the @code{Date_and_Time} intrinsic (in @code{libg2c})
289 to return the milliseconds value properly
293 Fix the @code{LStat} intrinsic (in @code{libg2c})
294 to return device-ID information properly
298 Improve documentation.
302 @heading In 0.5.24 and @code{egcs} 1.1.1 (versus 0.5.23 and 1.1):
306 Fix @code{libg2c} so it performs an implicit @code{ENDFILE} operation
308 whenever a @code{REWIND} is done.
310 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
311 @code{g77}'s version of @code{libf2c}.)
314 Fix @code{libg2c} so it no longer crashes with a spurious diagnostic
315 upon doing any I/O following a direct formatted write.
317 (This bug was introduced in 0.5.23 and @code{egcs} 1.1 in
318 @code{g77}'s version of @code{libf2c}.)
321 Fix @code{g77} so it no longer crashes compiling references
322 to the @code{Rand} intrinsic on some systems.
325 Fix @code{g77} portion of installation process so it works
326 better on some systems
327 (those with shells requiring @samp{else true} clauses
328 on @code{if} constructs
329 for the completion code to be set properly).
333 @heading In @code{egcs} 1.1 (versus 0.5.24):
337 Fix @code{g77} crash compiling code
338 containing the construct @samp{CMPLX(0.)} or similar.
344 (or apparently infinite run-time)
345 when compiling certain complicated expressions
346 involving @code{COMPLEX} arithmetic
347 (especially multiplication).
352 @cindex stack, 387 coprocessor
356 Fix a code-generation bug that afflicted
357 Intel x86 targets when @samp{-O2} was specified
358 compiling, for example, an old version of
359 the @code{DNRM2} routine.
361 The x87 coprocessor stack was being
362 mismanaged in cases involving assigned @code{GOTO}
367 @cindex double-precision performance
368 @cindex -malign-double
370 Align static double-precision variables and arrays
372 regardless of whether @samp{-malign-double} is specified.
374 Generally, this affects only local variables and arrays
375 having the @code{SAVE} attribute
376 or given initial values via @code{DATA}.
379 @c 1998-09-01: egcs-1.1 released.
380 @heading In @code{egcs} 1.1 (versus @code{egcs} 1.0.3):
384 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
385 that wrote one byte beyond the end of its @code{CHARACTER}
387 and in the @code{libU77} intrinsics
388 @code{GMTime} and @code{LTime}
389 that overwrote their arguments.
394 Assumed arrays with negative bounds
395 (such as @samp{REAL A(-1:*)})
396 no longer elicit spurious diagnostics from @code{g77},
397 even on systems with pointers having
398 different sizes than integers.
400 This bug is not known to have existed in any
401 recent version of @code{gcc}.
402 It was introduced in an early release of @code{egcs}.
407 Valid combinations of @code{EXTERNAL},
408 passing that external as a dummy argument
409 without explicitly giving it a type,
410 and, in a subsequent program unit,
411 referencing that external as
412 an external function with a different type
413 no longer crash @code{g77}.
418 @code{CASE DEFAULT} no longer crashes @code{g77}.
423 The @samp{-Wunused} option no longer issues a spurious
424 warning about the ``master'' procedure generated by
425 @code{g77} for procedures containing @code{ENTRY} statements.
429 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
430 compile-time constant @code{INTEGER} expression.
433 Fix @code{g77} @samp{-g} option so procedures that
434 use @code{ENTRY} can be stepped through, line by line,
438 Allow any @code{REAL} argument to intrinsics
439 @code{Second} and @code{CPU_Time}.
442 Use @code{tempnam}, if available, to open scratch files
443 (as in @samp{OPEN(STATUS='SCRATCH')})
444 so that the @code{TMPDIR} environment variable,
448 @code{g77}'s version of @code{libf2c} separates out
449 the setting of global state
450 (such as command-line arguments and signal handling)
451 from @file{main.o} into distinct, new library
454 This should make it easier to write portable applications
455 that have their own (non-Fortran) @code{main()} routine
456 properly set up the @code{libf2c} environment, even
457 when @code{libf2c} (now @code{libg2c}) is a shared library.
461 @code{g77} no longer installs the @file{f77} command
462 and @file{f77.1} man page
463 in the @file{/usr} or @file{/usr/local} heirarchy,
464 even if the @file{f77-install-ok} file exists
465 in the source or build directory.
466 See the installation documentation for more information.
471 @code{g77} no longer installs the @file{libf2c.a} library
472 and @file{f2c.h} include file
473 in the @file{/usr} or @file{/usr/local} heirarchy,
474 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
475 in the source or build directory.
476 See the installation documentation for more information.
481 The @file{libf2c.a} library produced by @code{g77} has been
482 renamed to @file{libg2c.a}.
483 It is installed only in the @code{gcc} ``private''
484 directory heirarchy, @file{gcc-lib}.
485 This allows system administrators and users to choose which
486 version of the @code{libf2c} library from @code{netlib} they
487 wish to use on a case-by-case basis.
488 See the installation documentation for more information.
493 The @file{f2c.h} include (header) file produced by @code{g77}
494 has been renamed to @file{g2c.h}.
495 It is installed only in the @code{gcc} ``private''
496 directory heirarchy, @file{gcc-lib}.
497 This allows system administrators and users to choose which
498 version of the include file from @code{netlib} they
499 wish to use on a case-by-case basis.
500 See the installation documentation for more information.
504 The @code{g77} command now expects the run-time library
505 to be named @code{libg2c.a} instead of @code{libf2c.a},
506 to ensure that a version other than the one built and
507 installed as part of the same @code{g77} version is picked up.
511 During the configuration and build process,
512 @code{g77} creates subdirectories it needs only as it
514 Other cleaning up of the configuration and build process
515 has been performed as well.
520 @code{install-info} now used to update the directory of
521 Info documentation to contain an entry for @code{g77}
522 (during installation).
526 Some diagnostics have been changed from warnings to errors,
527 to prevent inadvertent use of the resulting, probably buggy,
529 These mostly include diagnostics about use of unsupported features
530 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
531 @code{WRITE} statements,
532 and about truncations of various sorts of constants.
536 Improve compilation of @code{FORMAT} expressions so that
537 a null byte is appended to the last operand if it
539 This provides a cleaner run-time diagnostic as provided
540 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
545 Improve documentation and indexing.
550 The upgrade to @code{libf2c} as of 1998-06-18
551 should fix a variety of problems, including
552 those involving some uses of the @code{T} format
553 specifier, and perhaps some build (porting) problems
558 @heading In 0.5.24 and @code{egcs} 1.1 (versus 0.5.23):
562 @code{g77} no longer produces incorrect code
564 for @code{EQUIVALENCE} and @code{COMMON}
565 aggregates that, due to ``unnatural'' ordering of members
566 vis-a-vis their types, require initial padding.
571 @code{g77} no longer crashes when compiling code
572 containing specification statements such as
573 @samp{INTEGER(KIND=7) PTR}.
578 @code{g77} no longer crashes when compiling code
579 such as @samp{J = SIGNAL(1, 2)}.
583 @code{g77} now treats @samp{%LOC(@var{expr})} and
584 @samp{LOC(@var{expr})} as ``ordinary'' expressions
585 when they are used as arguments in procedure calls.
586 This change applies only to global (filewide) analysis,
587 making it consistent with
588 how @code{g77} actually generates code
591 Previously, @code{g77} treated these expressions
592 as denoting special ``pointer'' arguments
593 for the purposes of filewide analysis.
596 The @code{g77} driver now ensures that @samp{-lg2c}
597 is specified in the link phase prior to any
598 occurrence of @samp{-lm}.
599 This prevents accidentally linking to a routine
600 in the SunOS4 @samp{-lm} library
601 when the generated code wants to link to the one
602 in @code{libf2c} (@code{libg2c}).
605 @code{g77} emits more debugging information when
608 This new information allows, for example,
609 @kbd{which __g77_length_a} to be used in @code{gdb}
610 to determine the type of the phantom length argument
611 supplied with @code{CHARACTER} variables.
613 This information pertains to internally-generated
614 type, variable, and other information,
615 not to the longstanding deficiencies vis-a-vis
616 @code{COMMON} and @code{EQUIVALENCE}.
619 The F90 @code{Date_and_Time} intrinsic now is
623 The F90 @code{System_Clock} intrinsic allows
624 the optional arguments (except for the @code{Count}
625 argument) to be omitted.
629 Upgrade to @code{libf2c} as of 1998-06-18.
634 Improve documentation and indexing.
639 @heading In previous versions:
641 Information on previous versions is not provided
642 in this @file{@value{path-g77}/NEWS} file,
644 See @file{@value{path-g77}/news.texi},
645 or any of its other derivations
646 (Info, HTML, dvi forms)
647 for such information.
651 @c 1998-05-20: 0.5.23 released.
652 @heading In 0.5.23 (versus 0.5.22):
655 This release contains several regressions against
656 version 0.5.22 of @code{g77}, due to using the
657 ``vanilla'' @code{gcc} back end instead of patching
658 it to fix a few bugs and improve performance in a
661 Features that have been dropped from this version
662 of @code{g77} due to their being implemented
663 via @code{g77}-specific patches to the @code{gcc}
664 back end in previous releases include:
668 Support for @code{__restrict__} keyword,
669 the options @samp{-fargument-alias}, @samp{-fargument-noalias},
670 and @samp{-fargument-noalias-global},
671 and the corresponding alias-analysis code.
673 (@code{egcs} has the alias-analysis
674 code, but not the @code{__restrict__} keyword.
675 @code{egcs} @code{g77} users benefit from the alias-analysis
676 code despite the lack of the @code{__restrict__} keyword,
677 which is a C-language construct.)
680 Support for the GNU compiler options
681 @samp{-fmove-all-movables},
682 @samp{-freduce-all-givs},
683 and @samp{-frerun-loop-opt}.
685 (@code{egcs} supports these options.
686 @code{g77} users of @code{egcs} benefit from them even if
687 they are not explicitly specified,
688 because the defaults are optimized for @code{g77} users.)
691 Support for the @samp{-W} option warning about
692 integer division by zero.
695 The Intel x86-specific option @samp{-malign-double}
696 applying to stack-allocated data
697 as well as statically-allocate data.
701 Note that the @file{gcc/f/gbe/} subdirectory has been removed
702 from this distribution as a result of @code{g77} no longer
703 including patches for the @code{gcc} back end.
708 Fix bugs in the @code{libU77} intrinsic @code{HostNm}
709 that wrote one byte beyond the end of its @code{CHARACTER}
711 and in the @code{libU77} intrinsics
712 @code{GMTime} and @code{LTime}
713 that overwrote their arguments.
717 Support @code{gcc} version 2.8,
718 and remove support for prior versions of @code{gcc}.
720 @cindex -@w{}-driver option
721 @cindex @code{g77} options, -@w{}-driver
722 @cindex options, -@w{}-driver
724 Remove support for the @samp{--driver} option,
725 as @code{g77} now does all the driving,
726 just like @code{gcc}.
730 @code{CASE DEFAULT} no longer crashes @code{g77}.
735 Valid combinations of @code{EXTERNAL},
736 passing that external as a dummy argument
737 without explicitly giving it a type,
738 and, in a subsequent program unit,
739 referencing that external as
740 an external function with a different type
741 no longer crash @code{g77}.
746 @code{g77} no longer installs the @file{f77} command
747 and @file{f77.1} man page
748 in the @file{/usr} or @file{/usr/local} heirarchy,
749 even if the @file{f77-install-ok} file exists
750 in the source or build directory.
751 See the installation documentation for more information.
756 @code{g77} no longer installs the @file{libf2c.a} library
757 and @file{f2c.h} include file
758 in the @file{/usr} or @file{/usr/local} heirarchy,
759 even if the @file{f2c-install-ok} or @file{f2c-exists-ok} files exist
760 in the source or build directory.
761 See the installation documentation for more information.
766 The @file{libf2c.a} library produced by @code{g77} has been
767 renamed to @file{libg2c.a}.
768 It is installed only in the @code{gcc} ``private''
769 directory heirarchy, @file{gcc-lib}.
770 This allows system administrators and users to choose which
771 version of the @code{libf2c} library from @code{netlib} they
772 wish to use on a case-by-case basis.
773 See the installation documentation for more information.
778 The @file{f2c.h} include (header) file produced by @code{g77}
779 has been renamed to @file{g2c.h}.
780 It is installed only in the @code{gcc} ``private''
781 directory heirarchy, @file{gcc-lib}.
782 This allows system administrators and users to choose which
783 version of the include file from @code{netlib} they
784 wish to use on a case-by-case basis.
785 See the installation documentation for more information.
789 The @code{g77} command now expects the run-time library
790 to be named @code{libg2c.a} instead of @code{libf2c.a},
791 to ensure that a version other than the one built and
792 installed as part of the same @code{g77} version is picked up.
796 The @samp{-Wunused} option no longer issues a spurious
797 warning about the ``master'' procedure generated by
798 @code{g77} for procedures containing @code{ENTRY} statements.
802 @code{g77}'s version of @code{libf2c} separates out
803 the setting of global state
804 (such as command-line arguments and signal handling)
805 from @file{main.o} into distinct, new library
808 This should make it easier to write portable applications
809 that have their own (non-Fortran) @code{main()} routine
810 properly set up the @code{libf2c} environment, even
811 when @code{libf2c} (now @code{libg2c}) is a shared library.
815 During the configuration and build process,
816 @code{g77} creates subdirectories it needs only as it
817 needs them, thus avoiding unnecessary creation of, for example,
818 @file{stage1/f/runtime} when doing a non-bootstrap build.
819 Other cleaning up of the configuration and build process
820 has been performed as well.
825 @code{install-info} now used to update the directory of
826 Info documentation to contain an entry for @code{g77}
827 (during installation).
831 Some diagnostics have been changed from warnings to errors,
832 to prevent inadvertent use of the resulting, probably buggy,
834 These mostly include diagnostics about use of unsupported features
835 in the @code{OPEN}, @code{INQUIRE}, @code{READ}, and
836 @code{WRITE} statements,
837 and about truncations of various sorts of constants.
841 Improve documentation and indexing.
846 Upgrade to @code{libf2c} as of 1998-04-20.
848 This should fix a variety of problems, including
849 those involving some uses of the @code{T} format
850 specifier, and perhaps some build (porting) problems
855 @c 1998-03-16: 0.5.22 released.
856 @heading In 0.5.22 (versus 0.5.21):
860 Fix code generation for iterative @code{DO} loops that
861 have one or more references to the iteration variable,
862 or to aliases of it, in their control expressions.
863 For example, @samp{DO 10 J=2,J} now is compiled correctly.
868 @cindex stack, 387 coprocessor
872 Fix a code-generation bug that afflicted
873 Intel x86 targets when @samp{-O2} was specified
874 compiling, for example, an old version of
875 the @code{DNRM2} routine.
877 The x87 coprocessor stack was being
878 mismanaged in cases involving assigned @code{GOTO}
884 Fix @code{DTime} intrinsic so as not to truncate
885 results to integer values (on some systems).
889 Fix @code{Signal} intrinsic so it offers portable
890 support for 64-bit systems (such as Digital Alphas
895 Fix run-time crash involving @code{NAMELIST} on 64-bit
896 machines such as Alphas.
901 Fix @code{g77} version of @code{libf2c} so it no longer
902 produces a spurious @samp{I/O recursion} diagnostic at run time
903 when an I/O operation (such as @samp{READ *,I}) is interrupted
904 in a manner that causes the program to be terminated
905 via the @code{f_exit} routine (such as via @kbd{C-c}).
910 Fix @code{g77} crash triggered by @code{CASE} statement with
911 an omitted lower or upper bound.
916 Fix @code{g77} crash compiling references to @code{CPU_Time}
923 (or apparently infinite run-time)
924 when compiling certain complicated expressions
925 involving @code{COMPLEX} arithmetic
926 (especially multiplication).
931 Fix @code{g77} crash on statements such as
932 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
933 @samp{Z} is @code{DOUBLE COMPLEX}.
938 Fix a @code{g++} crash.
942 Support @samp{FORMAT(I<@var{expr}>)} when @var{expr} is a
943 compile-time constant @code{INTEGER} expression.
946 Fix @code{g77} @samp{-g} option so procedures that
947 use @code{ENTRY} can be stepped through, line by line,
952 Fix a profiling-related bug in @code{gcc} back end for
953 Intel x86 architecture.
957 Allow any @code{REAL} argument to intrinsics
958 @code{Second} and @code{CPU_Time}.
961 Allow any numeric argument to intrinsics
962 @code{Int2} and @code{Int8}.
965 Use @code{tempnam}, if available, to open scratch files
966 (as in @samp{OPEN(STATUS='SCRATCH')})
967 so that the @code{TMPDIR} environment variable,
971 Rename the @code{gcc} keyword @code{restrict} to
972 @code{__restrict__}, to avoid rejecting valid, existing,
974 Support for @code{restrict} is now more like support
979 Fix @samp{-fpedantic} to not reject procedure invocations
980 such as @samp{I=J()} and @samp{CALL FOO()}.
984 Fix @samp{-fugly-comma} to affect invocations of
985 only external procedures.
986 Restore rejection of gratuitous trailing omitted
987 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
990 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
991 @samp{-fbadu77-intrinsics-*} options.
995 Improve diagnostic messages from @code{libf2c}
996 so it is more likely that the printing of the
997 active format string is limited to the string,
998 with no trailing garbage being printed.
1000 (Unlike @code{f2c}, @code{g77} did not append
1001 a null byte to its compiled form of every
1002 format string specified via a @code{FORMAT} statement.
1003 However, @code{f2c} would exhibit the problem
1004 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1005 by printing @samp{(I)garbage} as the format string.)
1008 @ifclear USERVISONLY
1010 Improve compilation of @code{FORMAT} expressions so that
1011 a null byte is appended to the last operand if it
1013 This provides a cleaner run-time diagnostic as provided
1014 by @code{libf2c} for statements like @samp{PRINT '(I1', 42}.
1017 @ifclear USERVISONLY
1019 Fix various crashes involving code with diagnosed errors.
1022 @ifclear USERVISONLY
1024 Fix cross-compilation bug when configuring @code{libf2c}.
1027 @ifclear USERVISONLY
1029 Improve diagnostics.
1032 @ifclear USERVISONLY
1034 Improve documentation and indexing.
1037 @ifclear USERVISONLY
1039 Upgrade to @code{libf2c} as of 1997-09-23.
1040 This fixes a formatted-I/O bug that afflicted
1041 64-bit systems with 32-bit integers
1042 (such as Digital Alpha running GNU/Linux).
1046 @c 1998-03-15: egcs-1.0.2 released.
1047 @heading In @code{egcs} 1.0.2 (versus @code{egcs} 1.0.1):
1049 @ifclear USERVISONLY
1051 Fix @code{g77} crash triggered by @code{CASE} statement with
1052 an omitted lower or upper bound.
1055 @ifclear USERVISONLY
1057 Fix @code{g77} crash on statements such as
1058 @samp{PRINT *, (REAL(Z(I)),I=1,2)}, where
1059 @samp{Z} is @code{DOUBLE COMPLEX}.
1062 @ifclear USERVISONLY
1064 @cindex support, ELF
1065 @cindex -fPIC option
1066 @cindex options, -fPIC
1068 Fix @samp{-fPIC} (such as compiling for ELF targets)
1069 on the Intel x86 architecture target
1070 so invalid assembler code is no longer produced.
1073 @ifclear USERVISONLY
1075 Fix @samp{-fpedantic} to not reject procedure invocations
1076 such as @samp{I=J()} and @samp{CALL FOO()}.
1079 @ifclear USERVISONLY
1081 Fix @samp{-fugly-comma} to affect invocations of
1082 only external procedures.
1083 Restore rejection of gratuitous trailing omitted
1084 arguments to intrinsics, as in @samp{I=MAX(3,4,,)}.
1088 Fix compiler so it accepts @samp{-fgnu-intrinsics-*} and
1089 @samp{-fbadu77-intrinsics-*} options.
1092 @c 1998-01-02: egcs-1.0.1 released.
1093 @heading In @code{egcs} 1.0.1 (versus @code{egcs} 1.0):
1094 @ifclear USERVISONLY
1097 Fix run-time crash involving @code{NAMELIST} on 64-bit
1098 machines such as Alphas.
1102 @c 1997-12-03: egcs-1.0 released.
1103 @heading In @code{egcs} 1.0 (versus 0.5.21):
1106 Version 1.0 of @code{egcs}
1107 contains several regressions against
1108 version 0.5.21 of @code{g77},
1110 ``vanilla'' @code{gcc} back end instead of patching
1111 it to fix a few bugs and improve performance in a
1114 Features that have been dropped from this version
1115 of @code{g77} due to their being implemented
1116 via @code{g77}-specific patches to the @code{gcc}
1117 back end in previous releases include:
1121 Support for the C-language @code{restrict} keyword.
1124 Support for the @samp{-W} option warning about
1125 integer division by zero.
1128 The Intel x86-specific option @samp{-malign-double}
1129 applying to stack-allocated data
1130 as well as statically-allocate data.
1133 @ifclear USERVISONLY
1134 Note that the @file{gcc/f/gbe/} subdirectory has been removed
1135 from this distribution as a result of @code{g77}
1136 being fully integrated with
1137 the @code{egcs} variant of the @code{gcc} back end.
1140 @ifclear USERVISONLY
1142 Fix code generation for iterative @code{DO} loops that
1143 have one or more references to the iteration variable,
1144 or to aliases of it, in their control expressions.
1145 For example, @samp{DO 10 J=2,J} now is compiled correctly.
1148 @ifclear USERVISONLY
1150 Fix @code{DTime} intrinsic so as not to truncate
1151 results to integer values (on some systems).
1154 @ifclear USERVISONLY
1156 @c Toon Moene discovered these.
1157 Some Fortran code, miscompiled
1158 by @code{g77} built on @code{gcc} version 2.8.1
1159 on m68k-next-nextstep3 configurations
1160 when using the @samp{-O2} option,
1161 is now compiled correctly.
1162 It is believed that a C function known to miscompile
1163 on that configuration
1164 when using the @samp{-O2 -funroll-loops} options
1165 also is now compiled correctly.
1168 @ifclear USERVISONLY
1170 Remove support for non-@code{egcs} versions of @code{gcc}.
1173 @cindex -@w{}-driver option
1174 @cindex @code{g77} options, -@w{}-driver
1175 @cindex options, -@w{}-driver
1177 Remove support for the @samp{--driver} option,
1178 as @code{g77} now does all the driving,
1179 just like @code{gcc}.
1182 Allow any numeric argument to intrinsics
1183 @code{Int2} and @code{Int8}.
1185 @ifclear USERVISONLY
1187 Improve diagnostic messages from @code{libf2c}
1188 so it is more likely that the printing of the
1189 active format string is limited to the string,
1190 with no trailing garbage being printed.
1192 (Unlike @code{f2c}, @code{g77} did not append
1193 a null byte to its compiled form of every
1194 format string specified via a @code{FORMAT} statement.
1195 However, @code{f2c} would exhibit the problem
1196 anyway for a statement like @samp{PRINT '(I)garbage', 1}
1197 by printing @samp{(I)garbage} as the format string.)
1200 @ifclear USERVISONLY
1202 Upgrade to @code{libf2c} as of 1997-09-23.
1203 This fixes a formatted-I/O bug that afflicted
1204 64-bit systems with 32-bit integers
1205 (such as Digital Alpha running GNU/Linux).
1209 @c 1997-09-09: 0.5.21 released.
1212 @ifclear USERVISONLY
1214 Fix a code-generation bug introduced by 0.5.20
1215 caused by loop unrolling (by specifying
1216 @samp{-funroll-loops} or similar).
1217 This bug afflicted all code compiled by
1218 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1219 Fortran, and so on).
1222 @ifclear USERVISONLY
1224 Fix a code-generation bug manifested when
1225 combining local @code{EQUIVALENCE} with a
1226 @code{DATA} statement that follows
1227 the first executable statement (or is
1228 treated as an executable-context statement
1229 as a result of using the @samp{-fpedantic}
1233 @ifclear USERVISONLY
1235 Fix a compiler crash that occured when an
1236 integer division by a constant zero is detected.
1237 Instead, when the @samp{-W} option is specified,
1238 the @code{gcc} back end issues a warning about such a case.
1239 This bug afflicted all code compiled by
1240 version 2.7.2.2.f.2 of @code{gcc} (C, C++,
1241 Fortran, and so on).
1245 When the @samp{-W} option is specified, @code{gcc}, @code{g77},
1246 and other GNU compilers that incorporate the @code{gcc}
1247 back end as modified by @code{g77}, issue
1248 a warning about integer division by constant zero.
1251 @ifclear USERVISONLY
1253 Fix a compiler crash that occurred in some cases
1254 of procedure inlining.
1255 (Such cases became more frequent in 0.5.20.)
1258 @ifclear USERVISONLY
1260 Fix a compiler crash resulting from using @code{DATA}
1261 or similar to initialize a @code{COMPLEX} variable or
1265 @ifclear USERVISONLY
1267 Fix compiler crashes involving use of @code{AND}, @code{OR},
1268 or @code{XOR} intrinsics.
1271 @ifclear USERVISONLY
1273 Fix compiler bug triggered when using a @code{COMMON}
1274 or @code{EQUIVALENCE} variable
1275 as the target of an @code{ASSIGN}
1276 or assigned-@code{GOTO} statement.
1279 @ifclear USERVISONLY
1281 Fix compiler crashes due to using the name of a some
1282 non-standard intrinsics (such as @code{FTell} or
1283 @code{FPutC}) as such and as the name of a procedure
1285 Such dual use of a name in a program is allowed by
1289 @c @code{g77}'s version of @code{libf2c} has been modified
1290 @c so that the external names of library's procedures do not
1291 @c conflict with names used for Fortran procedures compiled
1293 @c An additional layer of jacket procedures has been added
1294 @c to @code{libf2c} to map the old names to the new names,
1295 @c for automatic use by programs that interface to the
1296 @c library procedures via the external-procedure mechanism.
1298 @c For example, the intrinsic @code{FPUTC} previously was
1299 @c implemented by @code{g77} as a call to the @code{libf2c}
1300 @c routine @code{fputc_}.
1301 @c This would conflict with a Fortran procedure named @code{FPUTC}
1302 @c (using default compiler options), and this conflict
1303 @c would cause a crash under certain circumstances.
1305 @c Now, the intrinsic @code{FPUTC} calls @code{G77_fputc_0},
1306 @c which does not conflict with the @code{fputc_} external
1307 @c that implements a Fortran procedure named @code{FPUTC}.
1309 @c Programs that refer to @code{FPUTC} as an external procedure
1310 @c without supplying their own implementation will link to
1311 @c the new @code{libf2c} routine @code{fputc_}, which is
1312 @c simply a jacket routine that calls @code{G77_fputc_0}.
1314 @ifclear USERVISONLY
1316 Place automatic arrays on the stack, even if
1317 @code{SAVE} or the @samp{-fno-automatic} option
1319 This avoids a compiler crash in some cases.
1322 @ifclear USERVISONLY
1324 The @samp{-malign-double} option now reliably aligns
1325 @code{DOUBLE PRECISION} optimally on Pentium and
1326 Pentium Pro architectures (586 and 686 in @code{gcc}).
1330 New option @samp{-Wno-globals} disables warnings
1331 about ``suspicious'' use of a name both as a global
1332 name and as the implicit name of an intrinsic, and
1333 warnings about disagreements over the number or natures of
1334 arguments passed to global procedures, or the
1335 natures of the procedures themselves.
1337 The default is to issue such warnings, which are
1338 new as of this version of @code{g77}.
1341 New option @samp{-fno-globals} disables diagnostics
1342 about potentially fatal disagreements
1343 analysis problems, such as disagreements over the
1344 number or natures of arguments passed to global
1345 procedures, or the natures of those procedures themselves.
1347 The default is to issue such diagnostics and flag
1348 the compilation as unsuccessful.
1349 With this option, the diagnostics are issued as
1350 warnings, or, if @samp{-Wno-globals} is specified,
1351 are not issued at all.
1353 This option also disables inlining of global procedures,
1354 to avoid compiler crashes resulting from coding errors
1355 that these diagnostics normally would identify.
1357 @ifclear USERVISONLY
1359 Diagnose cases where a reference to a procedure
1360 disagrees with the type of that procedure, or
1361 where disagreements about the number or nature
1363 This avoids a compiler crash.
1366 @ifclear USERVISONLY
1368 Fix parsing bug whereby @code{g77} rejected a
1369 second initialization specification immediately
1370 following the first's closing @samp{/} without
1371 an intervening comma in a @code{DATA} statement,
1372 and the second specification was an implied-DO list.
1375 @ifclear USERVISONLY
1377 Improve performance of the @code{gcc} back end so
1378 certain complicated expressions involving @code{COMPLEX}
1379 arithmetic (especially multiplication) don't appear to
1380 take forever to compile.
1383 @ifclear USERVISONLY
1385 Fix a couple of profiling-related bugs in @code{gcc}
1389 @ifclear USERVISONLY
1391 Integrate GNU Ada's (GNAT's) changes to the back end,
1392 which consist almost entirely of bug fixes.
1393 These fixes are circa version 3.10p of GNAT.
1396 @ifclear USERVISONLY
1398 Include some other @code{gcc} fixes that seem useful in
1399 @code{g77}'s version of @code{gcc}.
1400 (See @file{gcc/ChangeLog} for details---compare it
1401 to that file in the vanilla @code{gcc-2.7.2.3.tar.gz}
1406 Fix @code{libU77} routines that accept file and other names
1407 to strip trailing blanks from them, for consistency
1408 with other implementations.
1409 Blanks may be forcibly appended to such names by
1410 appending a single null character (@samp{CHAR(0)})
1411 to the significant trailing blanks.
1414 Fix @code{CHMOD} intrinsic to work with file names
1415 that have embedded blanks, commas, and so on.
1418 Fix @code{SIGNAL} intrinsic so it accepts an
1419 optional third @code{Status} argument.
1421 @ifclear USERVISONLY
1423 Fix @code{IDATE()} intrinsic subroutine (VXT form)
1424 so it accepts arguments in the correct order.
1425 Documentation fixed accordingly, and for
1426 @code{GMTIME()} and @code{LTIME()} as well.
1430 Make many changes to @code{libU77} intrinsics to
1431 support existing code more directly.
1433 Such changes include allowing both subroutine and
1434 function forms of many routines, changing @code{MCLOCK()}
1435 and @code{TIME()} to return @code{INTEGER(KIND=1)} values,
1436 introducing @code{MCLOCK8()} and @code{TIME8()} to
1437 return @code{INTEGER(KIND=2)} values,
1438 and placing functions that are intended to perform
1439 side effects in a new intrinsic group, @code{badu77}.
1441 @ifclear USERVISONLY
1443 Improve @code{libU77} so it is more portable.
1447 Add options @samp{-fbadu77-intrinsics-delete},
1448 @samp{-fbadu77-intrinsics-hide}, and so on.
1450 @ifclear USERVISONLY
1452 Fix crashes involving diagnosed or invalid code.
1455 @ifclear USERVISONLY
1457 @code{g77} and @code{gcc} now do a somewhat better
1458 job detecting and diagnosing arrays that are too
1459 large to handle before these cause diagnostics
1460 during the assembler or linker phase, a compiler
1461 crash, or generation of incorrect code.
1464 @ifclear USERVISONLY
1466 Make some fixes to alias analysis code.
1469 @ifclear USERVISONLY
1471 Add support for @code{restrict} keyword in @code{gcc}
1475 @ifclear USERVISONLY
1477 Support @code{gcc} version 2.7.2.3
1478 (modified by @code{g77} into version 2.7.2.3.f.1),
1480 support for prior versions of @code{gcc}.
1483 @ifclear USERVISONLY
1485 Incorporate GNAT's patches to the @code{gcc} back
1486 end into @code{g77}'s, so GNAT users do not need
1487 to apply GNAT's patches to build both GNAT and @code{g77}
1488 from the same source tree.
1491 @ifclear USERVISONLY
1493 Modify @code{make} rules and related code so that
1494 generation of Info documentation doesn't require
1495 compilation using @code{gcc}.
1496 Now, any ANSI C compiler should be adequate to
1497 produce the @code{g77} documentation (in particular,
1498 the tables of intrinsics) from scratch.
1502 Add @code{INT2} and @code{INT8} intrinsics.
1505 Add @code{CPU_TIME} intrinsic.
1508 Add @code{ALARM} intrinsic.
1511 @code{CTIME} intrinsic now accepts any @code{INTEGER}
1512 argument, not just @code{INTEGER(KIND=2)}.
1514 @ifclear USERVISONLY
1516 Warn when explicit type declaration disagrees with
1517 the type of an intrinsic invocation.
1520 @ifclear USERVISONLY
1522 Support @samp{*f771} entry in @code{gcc} @file{specs} file.
1525 @ifclear USERVISONLY
1527 Fix typo in @code{make} rule @code{g77-cross}, used only for
1531 @ifclear USERVISONLY
1533 Fix @code{libf2c} build procedure to re-archive library
1534 if previous attempt to archive was interrupted.
1537 @ifclear USERVISONLY
1539 Change @code{gcc} to unroll loops only during the last
1540 invocation (of as many as two invocations) of loop
1544 @ifclear USERVISONLY
1546 Improve handling of @samp{-fno-f2c} so that code that
1547 attempts to pass an intrinsic as an actual argument,
1548 such as @samp{CALL FOO(ABS)}, is rejected due to the fact
1549 that the run-time-library routine is, effectively,
1550 compiled with @samp{-ff2c} in effect.
1553 @ifclear USERVISONLY
1555 Fix @code{g77} driver to recognize @samp{-fsyntax-only}
1556 as an option that inhibits linking, just like @samp{-c} or
1557 @samp{-S}, and to recognize and properly handle the
1558 @samp{-nostdlib}, @samp{-M}, @samp{-MM}, @samp{-nodefaultlibs},
1559 and @samp{-Xlinker} options.
1562 @ifclear USERVISONLY
1564 Upgrade to @code{libf2c} as of 1997-08-16.
1567 @ifclear USERVISONLY
1569 Modify @code{libf2c} to consistently and clearly diagnose
1570 recursive I/O (at run time).
1574 @code{g77} driver now prints version information (such as produced
1575 by @kbd{g77 -v}) to @code{stderr} instead of @code{stdout}.
1578 The @samp{.r} suffix now designates a Ratfor source file,
1579 to be preprocessed via the @code{ratfor} command, available
1582 @ifclear USERVISONLY
1584 Fix some aspects of how @code{gcc} determines what kind of
1585 system is being configured and what kinds are supported.
1586 For example, GNU Linux/Alpha ELF systems now are directly
1590 @ifclear USERVISONLY
1592 Improve diagnostics.
1595 @ifclear USERVISONLY
1597 Improve documentation and indexing.
1600 @ifclear USERVISONLY
1602 Include all pertinent files for @code{libf2c} that come
1603 from @code{netlib.bell-labs.com}; give any such files
1604 that aren't quite accurate in @code{g77}'s version of
1605 @code{libf2c} the suffix @samp{.netlib}.
1608 @ifclear USERVISONLY
1610 Reserve @code{INTEGER(KIND=0)} for future use.
1614 @c 1997-02-28: 0.5.20 released.
1618 The @samp{-fno-typeless-boz} option is now the default.
1620 This option specifies that non-decimal-radix
1621 constants using the prefixed-radix form (such as @samp{Z'1234'})
1622 are to be interpreted as @code{INTEGER(KIND=1)} constants.
1623 Specify @samp{-ftypeless-boz} to cause such
1624 constants to be interpreted as typeless.
1626 (Version 0.5.19 introduced @samp{-fno-typeless-boz} and
1630 @xref{Fortran Dialect Options,,Options Controlling Fortran Dialect},
1631 for information on the @samp{-ftypeless-boz} option.
1635 Options @samp{-ff90-intrinsics-enable} and
1636 @samp{-fvxt-intrinsics-enable} now are the
1639 Some programs might use names that clash with
1640 intrinsic names defined (and now enabled) by these
1641 options or by the new @code{libU77} intrinsics.
1642 Users of such programs might need to compile them
1643 differently (using, for example, @samp{-ff90-intrinsics-disable})
1644 or, better yet, insert appropriate @code{EXTERNAL}
1645 statements specifying that these names are not intended
1646 to be names of intrinsics.
1649 The @code{ALWAYS_FLUSH} macro is no longer defined when
1650 building @code{libf2c}, which should result in improved
1651 I/O performance, especially over NFS.
1653 @emph{Note:} If you have code that depends on the behavior
1654 of @code{libf2c} when built with @code{ALWAYS_FLUSH} defined,
1655 you will have to modify @code{libf2c} accordingly before
1656 building it from this and future versions of @code{g77}.
1659 @xref{Output Assumed To Flush}, for more information.
1663 Dave Love's implementation of @code{libU77} has been
1664 added to the version of @code{libf2c} distributed with
1665 and built as part of @code{g77}.
1666 @code{g77} now knows about the routines in this library
1670 New option @samp{-fvxt} specifies that the
1671 source file is written in VXT Fortran, instead of GNU Fortran.
1674 @xref{VXT Fortran}, for more information on the constructs
1675 recognized when the @samp{-fvxt} option is specified.
1679 The @samp{-fvxt-not-f90} option has been deleted,
1680 along with its inverse, @samp{-ff90-not-vxt}.
1682 If you used one of these deleted options, you should
1683 re-read the pertinent documentation to determine which
1684 options, if any, are appropriate for compiling your
1685 code with this version of @code{g77}.
1688 @xref{Other Dialects}, for more information.
1692 The @samp{-fugly} option now issues a warning, as it
1693 likely will be removed in a future version.
1695 (Enabling all the @samp{-fugly-*} options is unlikely
1696 to be feasible, or sensible, in the future,
1697 so users should learn to specify only those
1698 @samp{-fugly-*} options they really need for a
1699 particular source file.)
1702 The @samp{-fugly-assumed} option, introduced in
1703 version 0.5.19, has been changed to
1704 better accommodate old and new code.
1707 @xref{Ugly Assumed-Size Arrays}, for more information.
1710 @ifclear USERVISONLY
1712 Make a number of fixes to the @code{g77} front end and
1713 the @code{gcc} back end to better support Alpha (AXP)
1715 This includes providing at least one bug-fix to the
1716 @code{gcc} back end for Alphas.
1720 Related to supporting Alpha (AXP) machines, the @code{LOC()}
1721 intrinsic and @code{%LOC()} construct now return
1722 values of @code{INTEGER(KIND=0)} type,
1723 as defined by the GNU Fortran language.
1725 This type is wide enough
1726 (holds the same number of bits)
1727 as the character-pointer type on the machine.
1729 On most machines, this won't make a difference,
1730 whereas, on Alphas and other systems with 64-bit pointers,
1731 the @code{INTEGER(KIND=0)} type is equivalent to @code{INTEGER(KIND=2)}
1732 (often referred to as @code{INTEGER*8})
1733 instead of the more common @code{INTEGER(KIND=1)}
1734 (often referred to as @code{INTEGER*4}).
1737 Emulate @code{COMPLEX} arithmetic in the @code{g77} front
1738 end, to avoid bugs in @code{complex} support in the
1739 @code{gcc} back end.
1740 New option @samp{-fno-emulate-complex}
1741 causes @code{g77} to revert the 0.5.19 behavior.
1743 @ifclear USERVISONLY
1745 Fix bug whereby @samp{REAL A(1)}, for example, caused
1746 a compiler crash if @samp{-fugly-assumed} was in effect
1747 and @var{A} was a local (automatic) array.
1748 That case is no longer affected by the new
1749 handling of @samp{-fugly-assumed}.
1752 @ifclear USERVISONLY
1754 Fix @code{g77} command driver so that @samp{g77 -o foo.f}
1755 no longer deletes @file{foo.f} before issuing other
1756 diagnostics, and so the @samp{-x} option is properly
1760 @ifclear USERVISONLY
1762 Enable inlining of subroutines and functions by the @code{gcc}
1764 This works as it does for @code{gcc} itself---program units
1765 may be inlined for invocations that follow them in the same
1766 program unit, as long as the appropriate compile-time
1767 options are specified.
1771 Dummy arguments are no longer assumed to potentially alias
1773 other dummy arguments or @code{COMMON} areas when any of
1774 these are defined (assigned to) by Fortran code.
1776 This can result in faster and/or smaller programs when
1777 compiling with optimization enabled, though on some
1778 systems this effect is observed only when @samp{-fforce-addr}
1781 New options @samp{-falias-check}, @samp{-fargument-alias},
1782 @samp{-fargument-noalias},
1783 and @samp{-fno-argument-noalias-global} control the
1784 way @code{g77} handles potential aliasing.
1787 @xref{Aliasing Assumed To Work}, for detailed information on why the
1788 new defaults might result in some programs no longer working the way they
1789 did when compiled by previous versions of @code{g77}.
1792 @ifclear USERVISONLY
1794 The @code{CONJG()} and @code{DCONJG()} intrinsics now
1795 are compiled in-line.
1798 @ifclear USERVISONLY
1800 The bug-fix for 0.5.19.1 has been re-done.
1801 The @code{g77} compiler has been changed back to
1802 assume @code{libf2c} has no aliasing problems in
1803 its implementations of the @code{COMPLEX} (and
1804 @code{DOUBLE COMPLEX}) intrinsics.
1805 The @code{libf2c} has been changed to have no such
1808 As a result, 0.5.20 is expected to offer improved performance
1809 over 0.5.19.1, perhaps as good as 0.5.19 in most
1810 or all cases, due to this change alone.
1812 @emph{Note:} This change requires version 0.5.20 of
1813 @code{libf2c}, at least, when linking code produced
1814 by any versions of @code{g77} other than 0.5.19.1.
1815 Use @samp{g77 -v} to determine the version numbers
1816 of the @code{libF77}, @code{libI77}, and @code{libU77}
1817 components of the @code{libf2c} library.
1818 (If these version numbers are not printed---in
1819 particular, if the linker complains about unresolved
1820 references to names like @samp{g77__fvers__}---that
1821 strongly suggests your installation has an obsolete
1822 version of @code{libf2c}.)
1826 New option @samp{-fugly-assign} specifies that the
1827 same memory locations are to be used to hold the
1828 values assigned by both statements @samp{I = 3} and
1829 @samp{ASSIGN 10 TO I}, for example.
1830 (Normally, @code{g77} uses a separate memory location
1831 to hold assigned statement labels.)
1834 @xref{Ugly Assigned Labels}, for more information.
1838 @code{FORMAT} and @code{ENTRY} statements now are allowed to
1839 precede @code{IMPLICIT NONE} statements.
1841 @ifclear USERVISONLY
1843 Produce diagnostic for unsupported @code{SELECT CASE} on
1844 @code{CHARACTER} type, instead of crashing, at compile time.
1847 @ifclear USERVISONLY
1849 Fix crashes involving diagnosed or invalid code.
1852 @ifclear USERVISONLY
1854 Change approach to building @code{libf2c} archive
1855 (@file{libf2c.a}) so that members are added to it
1856 only when truly necessary, so the user that installs
1857 an already-built @code{g77} doesn't need to have write
1858 access to the build tree (whereas the user doing the
1859 build might not have access to install new software
1863 @ifclear USERVISONLY
1865 Support @code{gcc} version 2.7.2.2
1866 (modified by @code{g77} into version 2.7.2.2.f.2),
1868 support for prior versions of @code{gcc}.
1871 @ifclear USERVISONLY
1873 Upgrade to @code{libf2c} as of 1997-02-08, and
1874 fix up some of the build procedures.
1877 @ifclear USERVISONLY
1879 Improve general build procedures for @code{g77},
1880 fixing minor bugs (such as deletion of any file
1881 named @file{f771} in the parent directory of @code{gcc/}).
1885 Enable full support of @code{INTEGER(KIND=2)}
1886 (often referred to as @code{INTEGER*8})
1888 @code{libf2c} and @file{f2c.h} so that @code{f2c} users
1889 may make full use of its features via the @code{g77}
1890 version of @file{f2c.h} and the @code{INTEGER(KIND=2)}
1891 support routines in the @code{g77} version of @code{libf2c}.
1894 Improve @code{g77} driver and @code{libf2c} so that @samp{g77 -v}
1895 yields version information on the library.
1898 The @code{SNGL} and @code{FLOAT} intrinsics now are
1899 specific intrinsics, instead of synonyms for the
1900 generic intrinsic @code{REAL}.
1903 New intrinsics have been added.
1904 These are @code{REALPART}, @code{IMAGPART},
1906 @code{LONG}, and @code{SHORT}.
1909 A new group of intrinsics, @code{gnu}, has been added
1910 to contain the new @code{REALPART}, @code{IMAGPART},
1911 and @code{COMPLEX} intrinsics.
1912 An old group, @code{dcp}, has been removed.
1915 Complain about industry-wide ambiguous references
1916 @samp{REAL(@var{expr})} and @samp{AIMAG(@var{expr})},
1917 where @var{expr} is @code{DOUBLE COMPLEX} (or any
1918 complex type other than @code{COMPLEX}), unless
1919 @samp{-ff90} option specifies Fortran 90 interpretation
1920 or new @samp{-fugly-complex} option, in conjunction with
1921 @samp{-fnot-f90}, specifies @code{f2c} interpretation.
1923 @ifclear USERVISONLY
1925 Make improvements to diagnostics.
1928 @ifclear USERVISONLY
1930 Speed up compiler a bit.
1933 @ifclear USERVISONLY
1935 Improvements to documentation and indexing, including
1936 a new chapter containing information on one, later
1937 more, diagnostics that users are directed to pull
1938 up automatically via a message in the diagnostic itself.
1940 (Hence the menu item @code{M} for the node
1941 @code{Diagnostics} in the top-level menu of
1942 the Info documentation.)
1946 @ifclear DOC-OLDNEWS
1947 @heading In previous versions:
1949 Information on previous versions is archived
1950 in @file{@value{path-g77}/news.texi}
1951 following the test of the @code{DOC-OLDNEWS} macro.
1955 @c 1997-02-01: 0.5.19.1 released.
1956 @heading In 0.5.19.1:
1959 Code-generation bugs afflicting operations on complex
1960 data have been fixed.
1962 These bugs occurred when assigning the result of an
1963 operation to a complex variable (or array element)
1964 that also served as an input to that operation.
1966 The operations affected by this bug were: @code{CONJG()},
1967 @code{DCONJG()}, @code{CCOS()}, @code{CDCOS()},
1968 @code{CLOG()}, @code{CDLOG()}, @code{CSIN()}, @code{CDSIN()},
1969 @code{CSQRT()}, @code{CDSQRT()}, complex division, and
1970 raising a @code{DOUBLE COMPLEX} operand to an @code{INTEGER}
1972 (The related generic and @samp{Z}-prefixed intrinsics,
1973 such as @code{ZSIN()}, also were affected.)
1975 For example, @samp{C = CSQRT(C)}, @samp{Z = Z/C}, and @samp{Z = Z**I}
1976 (where @samp{C} is @code{COMPLEX} and @samp{Z} is
1977 @code{DOUBLE COMPLEX}) have been fixed.
1980 @c 1996-12-07: 0.5.19 released.
1984 Fix @code{FORMAT} statement parsing so negative values for
1985 specifiers such as @code{P} (e.g. @samp{FORMAT(-1PF8.1)})
1986 are correctly processed as negative.
1989 Fix @code{SIGNAL} intrinsic so it once again accepts a
1990 procedure as its second argument.
1993 A temporary kludge option provides bare-bones information on
1994 @code{COMMON} and @code{EQUIVALENCE} members at debug time.
1997 New @samp{-fonetrip} option specifies FORTRAN-66-style
1998 one-trip @code{DO} loops.
2001 New @samp{-fno-silent} option causes names of program units
2002 to be printed as they are compiled, in a fashion similar to
2003 UNIX @code{f77} and @code{f2c}.
2006 New @samp{-fugly-assumed} option specifies that arrays
2007 dimensioned via @samp{DIMENSION X(1)}, for example, are to be
2008 treated as assumed-size.
2011 New @samp{-fno-typeless-boz} option specifies that non-decimal-radix
2012 constants using the prefixed-radix form (such as @samp{Z'1234'})
2013 are to be interpreted as @code{INTEGER} constants.
2016 New @samp{-ff66} option is a ``shorthand'' option that specifies
2017 behaviors considered appropriate for FORTRAN 66 programs.
2020 New @samp{-ff77} option is a ``shorthand'' option that specifies
2021 behaviors considered appropriate for UNIX @code{f77} programs.
2024 New @samp{-fugly-comma} and @samp{-fugly-logint} options provided
2025 to perform some of what @samp{-fugly} used to do.
2026 @samp{-fugly} and @samp{-fno-ugly} are now ``shorthand'' options,
2027 in that they do nothing more than enable (or disable) other
2028 @samp{-fugly-*} options.
2031 Fix parsing of assignment statements involving targets that
2032 are substrings of elements of @code{CHARACTER} arrays having
2033 names such as @samp{READ}, @samp{WRITE}, @samp{GOTO}, and
2034 @samp{REALFUNCTIONFOO}.
2037 Fix crashes involving diagnosed code.
2040 Fix handling of local @code{EQUIVALENCE} areas so certain cases
2041 of valid Fortran programs are not misdiagnosed as improperly
2042 extending the area backwards.
2045 Support @code{gcc} version 2.7.2.1.
2048 Upgrade to @code{libf2c} as of 1996-09-26, and
2049 fix up some of the build procedures.
2052 Change code generation for list-directed I/O so it allows
2053 for new versions of @code{libf2c} that might return non-zero
2054 status codes for some operations previously assumed to always
2057 This change not only affects how @code{IOSTAT=} variables
2058 are set by list-directed I/O, it also affects whether
2059 @code{END=} and @code{ERR=} labels are reached by these
2063 Add intrinsic support for new @code{FTELL} and @code{FSEEK}
2064 procedures in @code{libf2c}.
2067 Modify @code{fseek_()} in @code{libf2c} to be more portable
2068 (though, in practice, there might be no systems where this
2069 matters) and to catch invalid @code{whence} arguments.
2072 Some useless warnings from the @samp{-Wunused} option have
2076 Fix a problem building the @file{f771} executable
2077 on AIX systems by linking with the @samp{-bbigtoc} option.
2080 Abort configuration if @code{gcc} has not been patched
2081 using the patch file provided in the @file{gcc/f/gbe/}
2085 Add options @samp{--help} and @samp{--version} to the
2086 @code{g77} command, to conform to GNU coding guidelines.
2087 Also add printing of @code{g77} version number when
2088 the @samp{--verbose} (@samp{-v}) option is used.
2091 Change internally generated name for local @code{EQUIVALENCE}
2092 areas to one based on the alphabetically sorted first name
2093 in the list of names for entities placed at the beginning
2097 Improvements to documentation and indexing.
2100 @c 1996-04-01: 0.5.18 released.
2104 Add some rudimentary support for @code{INTEGER*1},
2105 @code{INTEGER*2}, @code{INTEGER*8},
2106 and their @code{LOGICAL} equivalents.
2107 (This support works on most, maybe all, @code{gcc} targets.)
2109 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2110 for providing the patch for this!
2112 Among the missing elements from the support for these
2113 features are full intrinsic support and constants.
2116 Add some rudimentary support for the @code{BYTE} and
2117 @code{WORD} type-declaration statements.
2118 @code{BYTE} corresponds to @code{INTEGER*1},
2119 while @code{WORD} corresponds to @code{INTEGER*2}.
2121 Thanks to Scott Snyder (@email{snyder@@d0sgif.fnal.gov})
2122 for providing the patch for this!
2125 The compiler code handling intrinsics has been largely
2126 rewritten to accommodate the new types.
2127 No new intrinsics or arguments for existing
2128 intrinsics have been added, so there is, at this
2129 point, no intrinsic to convert to @code{INTEGER*8},
2133 Support automatic arrays in procedures.
2136 Reduce space/time requirements for handling large
2137 @emph{sparsely} initialized aggregate arrays.
2138 This improvement applies to only a subset of
2139 the general problem to be addressed in 0.6.
2142 Treat initial values of zero as if they weren't
2143 specified (in DATA and type-declaration statements).
2144 The initial values will be set to zero anyway, but the amount
2145 of compile time processing them will be reduced,
2146 in some cases significantly (though, again, this
2147 is only a subset of the general problem to be
2150 A new option, @samp{-fzeros}, is introduced to
2151 enable the traditional treatment of zeros as any
2155 With @samp{-ff90} in force, @code{g77} incorrectly
2156 interpreted @samp{REAL(Z)} as returning a @code{REAL}
2157 result, instead of as a @code{DOUBLE PRECISION}
2159 (Here, @samp{Z} is @code{DOUBLE COMPLEX}.)
2161 With @samp{-fno-f90} in force, the interpretation remains
2162 unchanged, since this appears to be how at least some
2163 F77 code using the @code{DOUBLE COMPLEX} extension expected
2166 Essentially, @samp{REAL(Z)} in F90 is the same as
2167 @samp{DBLE(Z)}, while in extended F77, it appears to
2168 be the same as @samp{REAL(REAL(Z))}.
2171 An expression involving exponentiation, where both operands
2172 were type @code{INTEGER} and the right-hand operand
2173 was negative, was erroneously evaluated.
2176 Fix bugs involving @code{DATA} implied-@code{DO} constructs
2177 (these involved an errant diagnostic and a crash, both on good
2178 code, one involving subsequent statement-function definition).
2181 Close @code{INCLUDE} files after processing them, so compiling source
2182 files with lots of @code{INCLUDE} statements does not result in
2183 being unable to open @code{INCLUDE} files after all the available
2184 file descriptors are used up.
2187 Speed up compiling, especially of larger programs, and perhaps
2188 slightly reduce memory utilization while compiling (this is
2189 @emph{not} the improvement planned for 0.6 involving large aggregate
2190 areas)---these improvements result from simply turning
2191 off some low-level code to do self-checking that hasn't been
2192 triggered in a long time.
2195 Introduce three new options that
2196 implement optimizations in the @code{gcc} back end (GBE).
2197 These options are @samp{-fmove-all-movables}, @samp{-freduce-all-givs},
2198 and @samp{-frerun-loop-opt}, which are enabled, by default,
2199 for Fortran compilations.
2200 These optimizations are intended to help toon Fortran programs.
2203 Patch the GBE to do a better job optimizing certain
2204 kinds of references to array elements.
2207 Due to patches to the GBE, the version number of @code{gcc}
2208 also is patched to make it easier to manage installations,
2209 especially useful if it turns out a @code{g77} change to the
2212 The @code{g77}-modified version number is the @code{gcc}
2213 version number with the string @samp{.f.@var{n}} appended,
2214 where @samp{f} identifies the version as enhanced for
2215 Fortran, and @var{n} is @samp{1} for the first Fortran
2216 patch for that version of @code{gcc}, @samp{2} for the
2219 So, this introduces version 2.7.2.f.1 of @code{gcc}.
2222 Make several improvements and fixes to diagnostics, including
2223 the removal of two that were inappropriate or inadequate.
2226 Warning about two successive arithmetic operators, produced
2227 by @samp{-Wsurprising}, now produced @emph{only} when both
2228 operators are, indeed, arithmetic (not relational/boolean).
2231 @samp{-Wsurprising} now warns about the remaining cases
2232 of using non-integral variables for implied-@code{DO}
2233 loops, instead of these being rejected unless @samp{-fpedantic}
2234 or @samp{-fugly} specified.
2237 Allow @code{SAVE} of a local variable or array, even after
2238 it has been given an initial value via @code{DATA}, for example.
2241 Introduce an Info version of @code{g77} documentation, which
2242 supercedes @file{gcc/f/CREDITS}, @file{gcc/f/DOC}, and
2243 @file{gcc/f/PROJECTS}.
2244 These files will be removed in a future release.
2245 The files @file{gcc/f/BUGS}, @file{gcc/f/INSTALL}, and
2246 @file{gcc/f/NEWS} now are automatically built from
2247 the texinfo source when distributions are made.
2249 This effort was inspired by a first pass at translating
2250 @file{g77-0.5.16/f/DOC} that was contributed to Craig by
2251 David Ronis (@email{ronis@@onsager.chem.mcgill.ca}).
2254 New @samp{-fno-second-underscore} option to specify
2255 that, when @samp{-funderscoring} is in effect, a second
2256 underscore is not to be appended to Fortran names already
2257 containing an underscore.
2260 Change the way iterative @code{DO} loops work to follow
2262 In particular, calculation of the iteration count is
2263 still done by converting the start, end, and increment
2264 parameters to the type of the @code{DO} variable, but
2265 the result of the calculation is always converted to
2266 the default @code{INTEGER} type.
2268 (This should have no effect on existing code compiled
2269 by @code{g77}, but code written to assume that use
2270 of a @emph{wider} type for the @code{DO} variable
2271 will result in an iteration count being fully calculated
2272 using that wider type (wider
2273 than default @code{INTEGER}) must be rewritten.)
2276 Support @code{gcc} version 2.7.2.
2279 Upgrade to @code{libf2c} as of 1996-03-23, and
2280 fix up some of the build procedures.
2282 Note that the email addresses related to @code{f2c}
2283 have changed---the distribution site now is
2284 named @code{netlib.bell-labs.com}, and the
2285 maintainer's new address is @email{dmg@@bell-labs.com}.
2288 @c 1995-11-18: 0.5.17 released.
2292 @strong{Fix serious bug} in @samp{g77 -v} command that can cause removal of a
2293 system's @file{/dev/null} special file if run by user @code{root}.
2295 @strong{All users} of version 0.5.16 should ensure that
2296 they have not removed @file{/dev/null} or replaced it with an ordinary
2297 file (e.g. by comparing the output of @samp{ls -l /dev/null} with
2298 @samp{ls -l /dev/zero}.
2299 If the output isn't basically the
2300 same, contact your system
2301 administrator about restoring @file{/dev/null} to its proper status).
2303 This bug is particularly insidious because removing @file{/dev/null} as
2304 a special file can go undetected for quite a while, aside from
2305 various applications and programs exhibiting sudden, strange
2308 I sincerely apologize for not realizing the
2309 implications of the fact that when @samp{g77 -v} runs the @code{ld} command
2310 with @samp{-o /dev/null} that @code{ld} tries to @emph{remove} the executable
2311 it is supposed to build (especially if it reports unresolved
2312 references, which it should in this case)!
2315 Fix crash on @samp{CHARACTER*(*) FOO} in a main or block data program unit.
2318 Fix crash that can occur when diagnostics given outside of any
2319 program unit (such as when input file contains @samp{@@foo}).
2322 Fix crashes, infinite loops (hangs), and such involving diagnosed code.
2325 Fix @code{ASSIGN}'ed variables so they can be @code{SAVE}'d or dummy arguments,
2326 and issue clearer error message in cases where target of @code{ASSIGN}
2327 or @code{ASSIGN}ed @code{GOTO}/@code{FORMAT} is too small (which should
2331 Make @code{libf2c} build procedures work on more systems again by
2332 eliminating unnecessary invocations of @samp{ld -r -x} and @code{mv}.
2335 Fix omission of @samp{-funix-intrinsics-@dots{}} options in list of permitted
2336 options to compiler.
2339 Fix failure to always diagnose missing type declaration for
2340 @code{IMPLICIT NONE}.
2343 Fix compile-time performance problem (which could sometimes
2344 crash the compiler, cause a hang, or whatever, due to a bug
2345 in the back end) involving exponentiation with a large @code{INTEGER}
2346 constant for the right-hand operator (e.g. @samp{I**32767}).
2349 Fix build procedures so cross-compiling @code{g77} (the @code{fini}
2350 utility in particular) is properly built using the host compiler.
2353 Add new @samp{-Wsurprising} option to warn about constructs that are
2354 interpreted by the Fortran standard (and @code{g77}) in ways that
2355 are surprising to many programmers.
2358 Add @code{ERF()} and @code{ERFC()} as generic intrinsics mapping to existing
2359 @code{ERF}/@code{DERF} and @code{ERFC}/@code{DERFC} specific intrinsics.
2361 @emph{Note:} You should
2362 specify @samp{INTRINSIC ERF,ERFC} in any code where you might use
2363 these as generic intrinsics, to improve likelihood of diagnostics
2364 (instead of subtle run-time bugs) when using a compiler that
2365 doesn't support these as intrinsics (e.g. @code{f2c}).
2368 Remove from @samp{-fno-pedantic} the diagnostic about @code{DO}
2369 with non-@code{INTEGER} index variable; issue that under
2370 @samp{-Wsurprising} instead.
2373 Clarify some diagnostics that say things like ``ignored'' when that's
2377 Clarify diagnostic on use of @code{.EQ.}/@code{.NE.} on @code{LOGICAL}
2381 Minor improvements to code generation for various operations on
2382 @code{LOGICAL} operands.
2385 Minor improvement to code generation for some @code{DO} loops on some
2389 Support @code{gcc} version 2.7.1.
2392 Upgrade to @code{libf2c} as of 1995-11-15.
2395 @c 1995-08-30: 0.5.16 released.
2399 Fix a code-generation bug involving complicated @code{EQUIVALENCE} statements
2400 not involving @code{COMMON}.
2403 Fix code-generation bugs involving invoking ``gratis'' library procedures
2404 in @code{libf2c} from code compiled with @samp{-fno-f2c} by making these
2405 procedures known to @code{g77} as intrinsics (not affected by -fno-f2c).
2406 This is known to fix code invoking @code{ERF()}, @code{ERFC()},
2407 @code{DERF()}, and @code{DERFC()}.
2410 Update @code{libf2c} to include netlib patches through 1995-08-16, and
2411 @code{#define} @code{WANT_LEAD_0} to 1 to make @code{g77}-compiled code more
2412 consistent with other Fortran implementations by outputting
2413 leading zeros in formatted and list-directed output.
2416 Fix a code-generation bug involving adjustable dummy arrays with high
2417 bounds whose primaries are changed during procedure execution, and
2418 which might well improve code-generation performance for such arrays
2419 compared to @code{f2c} plus @code{gcc} (but apparently only when using
2420 @file{gcc-2.7.0} or later).
2423 Fix a code-generation bug involving invocation of @code{COMPLEX} and
2424 @code{DOUBLE COMPLEX} @code{FUNCTION}s and doing @code{COMPLEX} and
2425 @code{DOUBLE COMPLEX} divides, when the result
2426 of the invocation or divide is assigned directly to a variable
2427 that overlaps one or more of the arguments to the invocation or divide.
2430 Fix crash by not generating new optimal code for @samp{X**I} if @samp{I} is
2431 nonconstant and the expression is used to dimension a dummy
2432 array, since the @code{gcc} back end does not support the necessary
2433 mechanics (and the @code{gcc} front end rejects the equivalent
2434 construct, as it turns out).
2437 Fix crash on expressions like @samp{COMPLEX**INTEGER}.
2440 Fix crash on expressions like @samp{(1D0,2D0)**2}, i.e. raising a
2441 @code{DOUBLE COMPLEX} constant to an @code{INTEGER} constant power.
2444 Fix crashes and such involving diagnosed code.
2447 Diagnose, instead of crashing on, statement function definitions
2448 having duplicate dummy argument names.
2451 Fix bug causing rejection of good code involving statement function
2455 Fix bug resulting in debugger not knowing size of local equivalence
2456 area when any member of area has initial value (via @code{DATA},
2460 Fix installation bug that prevented installation of @code{g77} driver.
2461 Provide for easy selection of whether to install copy of @code{g77}
2462 as @code{f77} to replace the broken code.
2465 Fix @code{gcc} driver (affects @code{g77} thereby) to not
2466 gratuitously invoke the
2467 @code{f771} program (e.g. when @samp{-E} is specified).
2470 Fix diagnostic to point to correct source line when it immediately
2471 follows an @code{INCLUDE} statement.
2474 Support more compiler options in @code{gcc}/@code{g77} when
2475 compiling Fortran files.
2476 These options include @samp{-p}, @samp{-pg}, @samp{-aux-info}, @samp{-P},
2477 correct setting of version-number macros for preprocessing, full
2478 recognition of @samp{-O0}, and
2479 automatic insertion of configuration-specific linker specs.
2482 Add new intrinsics that interface to existing routines in @code{libf2c}:
2483 @code{ABORT}, @code{DERF}, @code{DERFC}, @code{ERF}, @code{ERFC}, @code{EXIT},
2484 @code{FLUSH}, @code{GETARG}, @code{GETENV}, @code{IARGC},
2485 @code{SIGNAL}, and @code{SYSTEM}.
2486 Note that @code{ABORT}, @code{EXIT}, @code{FLUSH}, @code{SIGNAL}, and
2487 @code{SYSTEM} are intrinsic subroutines, not functions (since they
2488 have side effects), so to get the return values from @code{SIGNAL}
2489 and @code{SYSTEM}, append a final argument specifying an @code{INTEGER}
2490 variable or array element (e.g. @samp{CALL SYSTEM('rm foo',ISTAT)}).
2493 Add new intrinsic group named @code{unix} to contain the new intrinsics,
2494 and by default enable this new group.
2497 Move @code{LOC()} intrinsic out of the @code{vxt} group to the new
2501 Improve @code{g77} so that @samp{g77 -v} by itself (or with
2502 certain other options, including @samp{-B}, @samp{-b}, @samp{-i},
2503 @samp{-nostdlib}, and @samp{-V}) reports lots more useful
2504 version info, and so that long-form options @code{gcc} accepts are
2505 understood by @code{g77} as well (even in truncated, unambiguous forms).
2508 Add new @code{g77} option @samp{--driver=name} to specify driver when
2509 default, @code{gcc}, isn't appropriate.
2512 Add support for @samp{#} directives (as output by the preprocessor) in the
2513 compiler, and enable generation of those directives by the
2514 preprocessor (when compiling @samp{.F} files) so diagnostics and debugging
2515 info are more useful to users of the preprocessor.
2518 Produce better diagnostics, more like @code{gcc}, with info such as
2519 @samp{In function `foo':} and @samp{In file included from...:}.
2522 Support @code{gcc}'s @samp{-fident} and @samp{-fno-ident} options.
2525 When @samp{-Wunused} in effect, don't warn about local variables used as
2526 statement-function dummy arguments or @code{DATA} implied-@code{DO} iteration
2527 variables, even though, strictly speaking, these are not uses
2528 of the variables themselves.
2531 When @samp{-W -Wunused} in effect, don't warn about unused dummy arguments
2532 at all, since there's no way to turn this off for individual
2533 cases (@code{g77} might someday start warning about these)---applies
2534 to @code{gcc} versions 2.7.0 and later, since earlier versions didn't
2535 warn about unused dummy arguments.
2538 New option @samp{-fno-underscoring} that inhibits transformation of names
2539 (by appending one or two underscores) so users may experiment
2540 with implications of such an environment.
2543 Minor improvement to @file{gcc/f/info} module to make it easier to build
2544 @code{g77} using the native (non-@code{gcc}) compiler on certain machines
2545 (but definitely not all machines nor all non-@code{gcc} compilers).
2547 do not report bugs showing problems compilers have with
2548 macros defined in @file{gcc/f/target.h} and used in places like
2549 @file{gcc/f/expr.c}.
2552 Add warning to be printed for each invocation of the compiler
2553 if the target machine @code{INTEGER}, @code{REAL}, or @code{LOGICAL} size
2555 since @code{g77} is known to not work well for such cases.
2558 Lots of new documentation (though work is still needed to put it into
2559 canonical GNU format).
2562 Build @code{libf2c} with @samp{-g0}, not @samp{-g2}, in effect
2563 (by default), to produce
2564 smaller library without lots of debugging clutter.
2567 @c 1995-05-19: 0.5.15 released.
2571 Fix bad code generation involving @samp{X**I} and temporary, internal variables
2572 generated by @code{g77} and the back end (such as for @code{DO} loops).
2575 Fix crash given @samp{CHARACTER A;DATA A/.TRUE./}.
2578 Replace crash with diagnostic given @samp{CHARACTER A;DATA A/1.0/}.
2581 Fix crash or other erratic behavior when null character constant
2582 (@samp{''}) is encountered.
2585 Fix crash or other erratic behavior involving diagnosed code.
2588 Fix code generation for external functions returning type @code{REAL} when
2589 the @samp{-ff2c} option is in force (which it is by default) so that
2590 @code{f2c} compatibility is indeed provided.
2593 Disallow @samp{COMMON I(10)} if @samp{I} has previously been specified
2594 with an array declarator.
2597 New @samp{-ffixed-line-length-@var{n}} option, where @var{n} is the
2599 of a typical fixed-form line, defaulting to 72 columns, such
2600 that characters beyond column @var{n} are ignored, or @var{n} is @samp{none},
2601 meaning no characters are ignored.
2602 does not affect lines
2603 with @samp{&} in column 1, which are always processed as if
2604 @samp{-ffixed-line-length-none} was in effect.
2607 No longer generate better code for some kinds of array references,
2608 as @code{gcc} back end is to be fixed to do this even better, and it
2609 turned out to slow down some code in some cases after all.
2612 In @code{COMMON} and @code{EQUIVALENCE} areas with any members given initial
2613 values (e.g. via @code{DATA}), uninitialized members now always
2614 initialized to binary zeros (though this is not required by
2615 the standard, and might not be done in future versions
2617 Previously, in some @code{COMMON}/@code{EQUIVALENCE} areas
2618 (essentially those with members of more than one type), the
2619 uninitialized members were initialized to spaces, to
2620 cater to @code{CHARACTER} types, but it seems no existing code expects
2621 that, while much existing code expects binary zeros.
2627 Don't emit bad code when low bound of adjustable array is nonconstant
2628 and thus might vary as an expression at run time.
2631 Emit correct code for calculation of number of trips in @code{DO} loops
2633 where the loop should not execute at all.
2634 (This bug affected cases
2635 where the difference between the begin and end values was less
2636 than the step count, though probably not for floating-point cases.)
2639 Fix crash when extra parentheses surround item in
2640 @code{DATA} implied-@code{DO} list.
2643 Fix crash over minor internal inconsistencies in handling diagnostics,
2644 just substitute dummy strings where necessary.
2647 Fix crash on some systems when compiling call to @code{MVBITS()} intrinsic.
2650 Fix crash on array assignment @samp{TYPE@var{ddd}(@dots{})=@dots{}}, where @var{ddd}
2651 is a string of one or more digits.
2654 Fix crash on @code{DCMPLX()} with a single @code{INTEGER} argument.
2657 Fix various crashes involving code with diagnosed errors.
2660 Support @samp{-I} option for @code{INCLUDE} statement, plus @code{gcc}'s
2661 @file{header.gcc} facility for handling systems like MS-DOS.
2664 Allow @code{INCLUDE} statement to be continued across multiple lines,
2665 even allow it to coexist with other statements on the same line.
2668 Incorporate Bellcore fixes to @code{libf2c} through 1995-03-15---this
2669 fixes a bug involving infinite loops reading EOF with empty list-directed
2673 Remove all the @code{g77}-specific auto-configuration scripts, code,
2675 except for temporary substitutes for bsearch() and strtoul(), as
2676 too many configure/build problems were reported in these areas.
2677 People will have to fix their systems' problems themselves, or at
2678 least somewhere other than @code{g77}, which expects a working ANSI C
2679 environment (and, for now, a GNU C compiler to compile @code{g77} itself).
2682 Complain if initialized common redeclared as larger in subsequent program
2686 Warn if blank common initialized, since its size can vary and hence
2687 related warnings that might be helpful won't be seen.
2690 New @samp{-fbackslash} option, on by default, that causes @samp{\}
2691 within @code{CHARACTER}
2692 and Hollerith constants to be interpreted a la GNU C.
2694 this behavior is somewhat different from @code{f2c}'s, which supports only
2695 a limited subset of backslash (escape) sequences.
2698 Make @samp{-fugly-args} the default.
2701 New @samp{-fugly-init} option, on by default, that allows typeless/Hollerith
2702 to be specified as initial values for variables or named constants
2703 (@code{PARAMETER}), and also allows character<->numeric conversion in
2704 those contexts---turn off via @samp{-fno-ugly-init}.
2707 New @samp{-finit-local-zero} option to initialize
2708 local variables to binary zeros.
2709 This does not affect whether they are @code{SAVE}d, i.e. made
2710 automatic or static.
2713 New @samp{-Wimplicit} option to warn about implicitly typed variables, arrays,
2715 (Basically causes all program units to default to @code{IMPLICIT NONE}.)
2718 @samp{-Wall} now implies @samp{-Wuninitialized} as with @code{gcc}
2719 (i.e. unless @samp{-O} not specified, since @samp{-Wuninitialized}
2720 requires @samp{-O}), and implies @samp{-Wunused} as well.
2723 @samp{-Wunused} no longer gives spurious messages for unused
2724 @code{EXTERNAL} names (since they are assumed to refer to block data
2725 program units, to make use of libraries more reliable).
2728 Support @code{%LOC()} and @code{LOC()} of character arguments.
2731 Support null (zero-length) character constants and expressions.
2734 Support @code{f2c}'s @code{IMAG()} generic intrinsic.
2737 Support @code{ICHAR()}, @code{IACHAR()}, and @code{LEN()} of
2738 character expressions that are valid in assignments but
2739 not normally as actual arguments.
2742 Support @code{f2c}-style @samp{&} in column 1 to mean continuation line.
2745 Allow @code{NAMELIST}, @code{EXTERNAL}, @code{INTRINSIC}, and @code{VOLATILE}
2746 in @code{BLOCK DATA}, even though these are not allowed by the standard.
2749 Allow @code{RETURN} in main program unit.
2752 Changes to Hollerith-constant support to obey Appendix C of the
2757 Now padded on the right with zeros, not spaces.
2760 Hollerith ``format specifications'' in the form of arrays of
2761 non-character allowed.
2764 Warnings issued when non-space truncation occurs when converting
2768 When specified as actual argument, now passed
2769 by reference to @code{INTEGER} (padded on right with spaces if constant
2770 too small, otherwise fully intact if constant wider the @code{INTEGER}
2771 type) instead of by value.
2774 @strong{Warning:} @code{f2c} differs on the
2775 interpretation of @samp{CALL FOO(1HX)}, which it treats exactly the
2776 same as @samp{CALL FOO('X')}, but which the standard and @code{g77} treat
2777 as @samp{CALL FOO(%REF('X '))} (padded with as many spaces as necessary
2778 to widen to @code{INTEGER}), essentially.
2781 Changes and fixes to typeless-constant support:
2785 Now treated as a typeless double-length @code{INTEGER} value.
2788 Warnings issued when overflow occurs.
2791 Padded on the left with zeros when converting
2795 Should be properly aligned and ordered on
2796 the target machine for whatever type it is turned into.
2799 When specified as actual argument, now passed as reference to
2800 a default @code{INTEGER} constant.
2804 @code{%DESCR()} of a non-@code{CHARACTER} expression now passes a pointer to
2805 the expression plus a length for the expression just as if
2806 it were a @code{CHARACTER} expression.
2807 For example, @samp{CALL FOO(%DESCR(D))}, where
2808 @samp{D} is @code{REAL*8}, is the same as @samp{CALL FOO(D,%VAL(8)))}.
2811 Name of multi-entrypoint master function changed to incorporate
2812 the name of the primary entry point instead of a decimal
2813 value, so the name of the master function for @samp{SUBROUTINE X}
2814 with alternate entry points is now @samp{__g77_masterfun_x}.
2817 Remove redundant message about zero-step-count @code{DO} loops.
2820 Clean up diagnostic messages, shortening many of them.
2823 Fix typo in @code{g77} man page.
2826 Clarify implications of constant-handling bugs in @file{f/BUGS}.
2829 Generate better code for @samp{**} operator with a right-hand operand of
2830 type @code{INTEGER}.
2833 Generate better code for @code{SQRT()} and @code{DSQRT()},
2834 also when @samp{-ffast-math}
2835 specified, enable better code generation for @code{SIN()} and @code{COS()}.
2838 Generate better code for some kinds of array references.
2841 Speed up lexing somewhat (this makes the compilation phase noticeably