1 Thu Oct 1 10:43:45 1998 Nick Clifton <nickc@cygnus.com>
3 * lex.c: Replace occurances of HANDLE_SYSV_PRAGMA with
4 HANDLE_GENERIC_PRAGMAS.
6 Mon Sep 28 04:22:00 1998 Jeffrey A Law (law@cygnus.com)
8 * news.texi: Update from Craig.
10 1998-09-23 Dave Love <d.love@dl.ac.uk>
12 * g77.texi: Additions about `/*', trailing comments and cpp.
14 1998-09-18 Dave Love <d.love@dl.ac.uk>
16 * g77.texi: Various additions and some small fixes.
18 Thu Sep 10 14:55:44 1998 Kamil Iskra <iskra@student.uci.agh.edu.pl>
20 * Make-lang.in (f77.install-common): Add missing "else true;".
22 1998-09-07 Dave Love <d.love@dl.ac.uk>
24 * ChangeLog.egcs: Deleted. Entries merged here.
26 1998-09-05 Dave Love <d.love@dl.ac.uk>
28 * Makefile.in (LDFLAGS): Set from BOOT_LDFLAGS.
29 (F771_LDFLAGS): Variable dispensed with.
31 Fri Sep 4 19:53:34 1998 Craig Burley <burley@gnu.org>
33 * intdoc.in: Minor editorial tweaks.
35 Fri Sep 4 18:35:52 1998 Craig Burley <burley@gnu.org>
37 * lang-options.h: Convert to wrap option and doc string
38 in a new macro invocation, FTNOPT, so the nearly identical
39 list can be used in FSF-g77.
41 Fri Sep 4 18:35:52 1998 Craig Burley <burley@gnu.org>
43 * Makefile.in (fini.o): Don't define USE_HCONFIG here.
44 * fini.c: Define USE_HCONFIG here instead, so deps-kinda
45 picks up correct dependency.
47 * Makefile.in (proj-h.o): Fix dependencies list.
49 Wed Sep 02 09:25:29 1998 Nick Clifton <nickc@cygnus.com>
51 * lex.c (ffe_lex_hash): Change how HANDLE_PRAGMA and
52 HANDLE_SYSV_PRAGMA would be called if they pragma parsing was
54 Generate warning messages if unknown pragmas are encountered.
55 (pragma_getc): New function: retrieves characters from the
56 input stream. Defined when HANDLE_PRAGMA is defined.
57 (pragma_ungetc): New function: replaces characters back into the
58 input stream. Defined when HANDLE_PRAGMA is defined.
60 Tue Sep 1 10:00:21 1998 Craig Burley <burley@gnu.org>
62 * bugs.texi, g77.1, g77.texi, intdoc.in, news.texi: Doc updates
65 1998-08-23 Dave Love <d.love@dl.ac.uk>
67 * g77.texi: Increment `version-g77' and fix a few typos.
69 Tue Aug 18 21:41:31 1998 Jeffrey A Law (law@cygnus.com)
71 * Make-lang.in: Add several "else true" clauses to deal with lame
74 Tue Aug 11 08:12:14 1998 H.J. Lu (hjl@gnu.org)
76 * Make-lang.in (g77.o): Touch lang-f77 before checking it.
78 1998-08-09 Dave Love <d.love@dl.ac.uk>
80 * Make-lang.in (f/g77.dvi): Replace non-working use of texi2dvi
81 with explicit use of tex.
82 (f77.mostlyclean): Remove TeX index files.
84 * g77install.texi (Prerequisites): Kluge round TeX lossage with
85 hyphen in @value in @code.
87 Tue Aug 4 16:59:39 1998 Craig Burley <burley@gnu.org>
89 * com.c (ffecom_convert_narrow_, ffecom_convert_widen_):
90 Allow conversion from pointer to same-sized integer,
91 to fix invoking SIGNAL as a function.
93 1998-07-26 Dave Love <d.love@dl.ac.uk>
95 * BUGS, INSTALL, NEWS: Rebuilt.
97 Sat Jul 25 17:23:55 1998 Craig Burley <burley@gnu.org>
100 * stc.c (ffestc_R1229_start): Set info to ANY as well.
102 Tue Jul 21 04:33:37 1998 Craig Burley <burley@gnu.org>
104 * g77spec.c (lang_specific_driver): Return unmolested
105 command line when --help seen.
106 Comment out code that printed g77-specific --help info.
108 Sat Jul 18 19:16:48 1998 Craig Burley <burley@gnu.org>
110 * lang-options.h: Fix up doc strings.
111 Remove the unimplemented -fdcp-intrinsics-* options.
113 * str-1t.fin: Change mixed-case spelling of `GoTo' from
116 Thu Jul 16 13:26:36 1998 Craig Burley <burley@gnu.org>
118 * com.c (ffecom_finish_symbol_transform_): Revert change
119 of 1998-05-23, as it was too aggressive, in that it
120 prevented transformation of (used) functions before
121 primary code generation.
123 1998-07-15 Dave Love <d.love@dl.ac.uk>
125 * intdoc.texi: Regenerated.
127 Mon Jul 13 18:45:06 1998 Craig Burley <burley@gnu.org>
129 * Make-lang.in (f77.rebuilt): Fix to depend on
130 build-dir-based, not source-based, g77.info.
132 * g77.texi: Merge docs with 0.5.24.
133 * g77install.texi: Ditto.
135 Mon Jul 13 18:02:29 1998 Craig Burley <burley@gnu.org>
137 Cleanups vis-a-vis g77-0.5.24:
138 * g77spec.c (lang_specific_driver): Tabify source.
139 * top.c (ffe_decode_option): Use fixed macro to set
140 internal-checking flag.
141 * top.h (ffe_set_is_do_internal_checks): Fix macro.
143 Mon Jul 13 17:33:44 1998 Craig Burley <burley@gnu.org>
145 Cleanups vis-a-vis system.h cutover and g77-0.5.24:
146 * Makefile.in (fini.o): Define USE_HCONFIG macro
147 so source code doesn't have to.
148 * fini.c: Don't define USE_HCONFIG here, since
149 source code usually shouldn't care about this.
150 * ansify.c: Include stddef.h only if we have it.
154 Mon Jul 13 17:30:29 1998 Nick Clifton <nickc@cygnus.com>
156 * lang-options.h: Format changed to work with --help support added
159 Mon Jul 13 11:54:03 1998 Craig Burley <burley@gnu.org>
161 * com.c (ffecom_push_tempvar): Replace kludge that
162 munged back-end globals directly with proper calls
163 to push_topmost_sequence and pop_topmost_sequence.
165 1998-07-12 Dave Love <d.love@dl.ac.uk>
167 * version.c: Bump version.
169 Sat Jul 11 19:24:32 1998 Craig Burley <burley@gnu.org>
172 * equiv.c (ffeequiv_offset_): Don't crash on various
173 possible ANY operands.
175 Sat Jul 11 18:24:37 1998 Craig Burley <burley@gnu.org>
177 * com.c (ffecom_expr_) [FFEBLD_opCONTER]: Die if padding
178 for constant is non-zero.
180 * com.c (__eprintf): Delete this function, it is obsolete.
182 1998-07-09 Dave Love <d.love@dl.ac.uk>
184 * intdoc.in (HOSTNM_func, HOSTNM_subr): Update last change.
186 Thu Jul 9 00:45:59 1998 Craig Burley <burley@gnu.org>
188 Fix debugging of CHARACTER*(*), etc., which requires
189 emitting debug info on types like `ftnlen':
190 * com.c (ffecom_start_progunit_): Don't bother
191 resetting "invented" flag for identifier.
192 (ffecom_transform_equiv_): Don't bother zeroing
193 "ignored" flag for decl.
194 (pushdecl): No longer set "ignored", "used", or
195 "suppressed debug" flags for decls having "invented"
198 1998-07-06 Mike Stump <mrs@wrs.com>
200 * Make-lang.in (f77.stage?): Use mv -f instead of just mv so that
203 1998-07-06 Dave Love <d.love@dl.ac.uk>
205 * intdoc.in (HOSTNM_func, HOSTNM_subr): Note possible need for
208 1998-07-05 Dave Love <d.love@dl.ac.uk>
210 * intdoc.in: Add entry for DATE_AND_TIME.
212 * intrin.def: Add implementation for DATE_AND_TIME. Make second
213 and third args of SYSTEM_CLOCK optional.
215 * com.c (ffecom_expr_intrinsic_): New case for DATE_AND_TIME.
217 * com-rt.def (FFECOM_gfrtSYSTEM_CLOCK): Call G77_system_clock_0,
219 (FFECOM_gfrtDATE_AND_TIME): New DEFGFRT.
221 Wed Jul 1 11:19:13 1998 Craig Burley <burley@gnu.org>
223 Fix 980701-1.f (which was producing "unaligned trap"
224 on an Alpha running GNU/Linux, as predicted):
225 * equiv.c (ffeequiv_layout_local_): Don't bother
226 coping with pre-padding of entire area while building
227 it; do that instead after the building is done, and
228 do it by modifying only the modulo field. This covers
229 the case of alignment stringency being increased without
230 lowering the starting offset, unlike the previous changes,
231 and even more elegantly than those.
233 * target.c (ffetarget_align): Make sure alignments
234 are non-zero, just in case.
236 Mon Jun 29 09:47:33 1998 Craig Burley <burley@gnu.org>
239 * bld.h: New `pad' field and accessor macros for
240 ACCTER, ARRTER, and CONTER ops.
241 * bld.c (ffebld_new_accter, ffebld_new_arrter,
242 ffebld_new_conter_with_orig): Initialize `pad' field
244 * com.c (ffecom_transform_common_): Include initial
245 padding (aka modulo aka offset) in size calculation.
246 Copy initial padding value into FFE initialization expression
247 so the GBE transformation of that expression includes it.
248 Make array low bound 0 instead of 1, for consistency.
249 (ffecom_transform_equiv_): Include initial
250 padding (aka modulo aka offset) in size calculation.
251 Copy initial padding value into FFE initialization expression
252 so the GBE transformation of that expression includes it.
253 Make array low bound 0 instead of 1, for consistency.
254 (ffecom_expr_, case FFEBLD_opACCTER): Delete unused `size'
256 Track destination offset separately, allowing for
258 Don't bother setting initial PURPOSE offset if zero.
259 Include initial padding in size calculation.
260 (ffecom_expr_, case FFEBLD_opARRTER): Allow for
262 Include initial padding in size calculation.
263 Make array low bound 0 instead of 1, for consistency.
264 (ffecom_finish_global_): Make array low bound 0 instead
265 of 1, for consistency.
266 (ffecom_notify_init_storage): Copy `pad' field from old
267 ACCTER to new ARRTER.
268 (ffecom_notify_init_symbol): Ditto.
269 * data.c (ffedata_gather_): Initialize `pad' field in new
271 (ffedata_value_): Ditto.
272 * equiv.c (ffeequiv_layout_local_): When lowering start
273 of equiv area, extend lowering to maintain needed alignment.
274 * target.c (ffetarget_align): Handle negative offset correctly.
276 * global.c (ffeglobal_pad_common): Warn about non-zero
277 padding only the first time its seen.
278 If new padding larger than old, update old.
279 (ffeglobal_save_common): Use correct type for size throughout.
280 * global.h: Use correct type for size throughout.
281 (ffeglobal_common_pad): New macro.
282 (ffeglobal_pad): Delete this unused and broken macro.
284 Sat Jun 27 12:18:33 1998 Jeffrey A Law (law@cygnus.com)
286 * Make-lang.in (g77): Depend on mkstemp.o. Link in mkstemp.o.
288 Fri Jun 26 11:54:19 1998 Craig Burley <burley@gnu.org>
290 * g77spec.c (lang_specific_driver): Put `-lg2c' in
291 front of any `-lm' that is seen.
293 Wed Jun 24 01:01:23 1998 Jeffrey A Law (law@cygnus.com)
295 * g77spec.c (lang_specific_driver): Revert last change.
297 Mon Jun 22 23:12:05 1998 H.J. Lu (hjl@gnu.org)
299 * Make-lang.in (G77STAGESTUFF): Add g77.c.
301 Fri Jun 19 07:54:40 1998 H.J. Lu (hjl@gnu.org)
303 * g77spec.c (lang_specific_driver): Check n_infiles before
306 Mon Jun 15 23:39:24 1998 Craig Burley <burley@gnu.org>
308 * Make-lang.in (f/g77.info): Use -f when removing
309 pre-existing Info files, if any. (This rm command
310 can go away once makeinfo has been changed to delete
311 .info-N files beyond the last one it creates.)
313 * Make-lang.in ($(srcdir)/f/intdoc.texi): Compile
314 using $(INCLUDES) macro to get the new hconfig.h
315 and system.h headers.
317 Mon Jun 15 22:21:57 1998 Craig Burley <burley@gnu.org>
344 * system.j: New file.
346 Use toplev.h where appropriate:
355 * toplev.j: New file.
357 Conditionalize all dumping/reporting routines so they don't
358 get built for gcc/egcs:
370 Use hconfig.h instead of config.h where appropriate:
371 * Makefile.in (proj-h.o): Compile with -DUSE_HCONFIG.
372 * fini.c: Define USE_HCONFIG before including proj.h.
374 * Makefile.in (deps-kinda): Redirect stderr to stdout,
375 to eliminate diagnostics vis-a-vis g77spec.c.
377 * Makefile.in: Regenerate dependencies via deps-kinda.
379 * lex.c (ffelex_file_fixed, ffelex_file_free): Eliminate
380 apparently spurious warnings about uninitialized variables
381 `c', `column', and so on.
383 Sat Jun 13 03:13:18 1998 Craig Burley <burley@gnu.org>
385 * g77spec.c (lang_specific_driver): Print out egcs
386 version info first, to be compatible with what some
387 test facilities expect.
389 Wed Jun 10 13:17:32 1998 Dave Brolley <brolley@cygnus.com>
391 * top.h (ffe_decode_option): New argc/argv interface.
392 * top.c (ffe_decode_option): New argc/argv interface.
393 * parse.c (yyparse): New argc/argv interface for ffe_decode_option.
394 * com.c (lang_decode_option): New argc/argv interface.
396 Sun Jun 7 14:04:34 1998 Richard Henderson <rth@cygnus.com>
398 * com.c (lang_init_options): New function.
399 * top.c (ffe_decode_option): Remove all trace of -fset-g77-defaults.
400 Set ffe_is_do_internal_checks_ with -version.
401 * lang-options.h: Likewise.
402 * lang-specs.h: Likewise.
404 Fri Jun 5 15:53:17 1998 Per Bothner <bothner@cygnus.com>
406 * g77spec.c (lang_specific_pre_link, lang_specific_extra_ofiles):
407 Define - update needed by gcc.c change.
409 Mon Jun 1 19:37:42 1998 Craig Burley <burley@gnu.org>
411 * com.c (ffecom_init_0): Fix setup of INTEGER(KIND=7)
413 * info.c (ffeinfo_type): Don't crash on null type.
414 * expr.c (ffeexpr_fulfill_call_): Don't special-case
415 %LOC(expr) or LOC(expr).
416 Delete FFEGLOBAL_argsummaryPTR.
417 * global.c, global.h: Delete FFEGLOBAL_argsummaryPTR.
419 Thu May 28 21:32:18 1998 Craig Burley <burley@gnu.org>
421 Restore circa-0.5.22 capabilities of `g77' driver:
422 * Make-lang.in (g77spec.o): Depend on f/version.h.
423 (g77version.o): New rule to compile g77 version info.
424 (g77$(exeext)): Depend on and link in g77version.o.
425 * g77spec.c: Rewrite to be more like 0.5.22 version
426 of g77.c, making filtering of command line smarter
427 so mixed Fortran and C (etc.) can be compiled, verbose
428 version info can be obtained, etc.
429 * lang-specs.h (f77-version): New "language" to support
430 "g77 -v" command under new gcc 2.8 regime.
431 * lex.c (ffelex_file_fixed): If -fnull-version, just
432 substitute a "source file" that prints out version info.
433 * top.c, top.h: Support -fnull-version.
435 * lang-specs.h: Use "%O" instead of OO macro to specify
436 object extension. Remove old stringizing cruft.
438 * Make-lang.in (g77.c, g77spec.o, g77.o, g77$(exeext),
439 g77-cross$(exeext), f771,
440 $(srcdir)/f/g77.info, $(srcdir)/f/g77.dvi,
441 $(srcdir)/f/intdoc.texi,
442 f77.install-common, f77.install-info, f77.install-man,
443 f77.uninstall, $(G77STAGESTUFF), f77.stage1, f77.stage2,
444 f77.stage3, f77.stage4, f77.distdir): Don't do anything
445 unless user specified "f77" or "F77" in $LANGUAGES either
446 during configuration or explicitly. For convenience of
447 various tests and to work around lack of the assignment
448 "LANGUAGES=$(BOOT_LANGUAGES)" in the "make stage1" command
449 of "make bootstrap" in gcc, use a touch file named "lang-f77"
450 to communicate whether this is the case.
452 * Make-lang.in (F77_FLAGS_TO_PASS): Delete this macro,
453 replace with minimal expansion of its former self in
454 each of the two instances where it was used.
456 * Makefile.in (HOST_CC): Delete this definition.
458 * com.c (index, rindex): Delete these declarations.
460 * proj.h: (isascii): Delete this.
462 * Make-lang.in (f77.install-common): Warn if `f77-install-ok'
463 flag-file exists, since it no longer triggers any activity.
465 Rename libf2c.a and f2c.h to libg2c.a and g2c.h,
466 normalize and simplify g77/libg2c build process:
467 * Make-lang.in: Remove all support for overwriting
468 /usr/bin/f77 etc., or whatever the actual names are
469 via $(prefix) and $(local_prefix). (g++ overwrites
470 /usr/bin/c++, but then it's often the only C++ compiler
471 on the system; f77 often exists on systems that are
473 (f77.realclean): Remove obsolete target.
474 (g77.c, g77$(exeext)): Minor changes to look more like g++'s
476 (f771): Now built with srcdir=gcc/f, not srcdir=gcc, to be
477 more like g++ and such.
478 (f/Makefile): Removed, as g++ doesn't need this rule.
479 (f77.install-common): No longer install f77, etc.
480 (f77.install-man): No longer install f77.1.
481 (f77.uninstall): No longer uninstall f77, f77.1, etc.
482 (f77.stage1, f77.stage2, f77.stage3, f77.stage4): Do work
483 only if "f77" appears in $(LANGUAGES).
484 (Note: gcc's Makefile.in's bootstrap target should set
485 LANGUAGES=$(BOOT_LANGUAGES) when making the stage1 target.)
486 * Makefile.in: Update vis-a-vis gcc/cp/Makefile.in.
488 (g77-only): Relocate.
489 (all.indirect, f771, *.o): Now assumes current directory
490 is this dir (gcc/f), not the parent directory.
491 (TAGS): Remove "echo 'parse.y,0' >> TAGS ;" line.
492 * config-lang.in: Delete commented-out code.
493 Fix stagestuff definition. Add more stuff to
494 diff_excludes definition. Don't create any directories.
495 Set outputs to f/Makefile, to get variable substition
496 to happen (what does that really do, anyway?!).
497 * g77spec.c: Rename libf2c to libg2c.
499 * com.h: Remove all of the gcc back-end decls,
500 since egcs should have all of them correct.
502 * com.c: Include "proj.h" before anything else,
503 as that's how things are supposed to work.
506 * bad.c: Include "flags.j" here, since some diagnostics
507 check flag_pedantic_errors.
509 * Makefile.in (f/*.o): Rebuild dependencies via
512 * output.j: New source file.
513 * Make-lang.in (F77_SRCS): Update accordingly.
514 * Makefile.in (OUTPUT_H): Ditto.
516 * com.c: Include "output.j" here.
519 Mon May 25 03:34:42 1998 Craig Burley <burley@gnu.org>
521 * com.c (ffecom_expr_): Fix D**I and Z**I cases to
522 not convert (DOUBLE PRECISION) D and (DOUBLE COMPLEX) Z
523 to INTEGER. (This is dead code here anyway.)
525 Sat May 23 06:32:52 1998 Craig Burley <burley@gnu.org>
527 * com.c (ffecom_finish_symbol_transform_): Don't transform
528 statement (nested) functions, to avoid gcc compiling them
529 and thus producing linker errors if they refer to undefined
530 external functions. But warn if they're unused and -Wunused.
531 * bad.def (FFEBAD_SFUNC_UNUSED): New diagnostic.
533 Wed May 20 12:12:55 1998 Craig Burley <burley@gnu.org>
535 * Version 0.5.23 released.
537 Tue May 19 14:52:41 1998 Craig Burley <burley@gnu.org>
539 * bad.def (FFEBAD_OPEN_UNSUPPORTED, FFEBAD_INQUIRE_UNSUPPORTED,
540 FFEBAD_READ_UNSUPPORTED, FFEBAD_WRITE_UNSUPPORTED,
541 FFEBAD_QUAD_UNSUPPORTED, FFEBAD_BLOCKDATA_STMT,
542 FFEBAD_TRUNCATING_CHARACTER, FFEBAD_TRUNCATING_HOLLERITH,
543 FFEBAD_TRUNCATING_NUMERIC, FFEBAD_TRUNCATING_TYPELESS,
544 FFEBAD_TYPELESS_OVERFLOW): Change these from warnings
547 Tue May 19 14:51:59 1998 Craig Burley <burley@gnu.org>
549 * Make-lang.in (f77.install-info, f77.uninstall):
550 Use install-info as appropriate.
552 Tue May 19 12:56:54 1998 Craig Burley <burley@gnu.org>
554 * com.c (ffecom_init_0): Rename xargc to f__xargc,
555 in accord with same-dated change to f/runtime.
557 Fri May 15 10:52:49 1998 Craig Burley <burley@gnu.org>
559 * com.c (ffecom_convert_narrow_, ffecom_convert_widen_):
560 Be even more persnickety in checking for internal bugs.
561 Also, if precision isn't changing, just return the expr.
563 * expr.c (ffeexpr_token_number_): Call
564 ffeexpr_make_float_const_ to make an integer.
565 (ffeexpr_make_float_const_): Handle making an integer.
567 * intrin.c (ffeintrin_init_0): Distinguish between
568 crashes on bad arg base and kind types.
570 Fri May 15 01:44:22 1998 Mumit Khan <khan@xraylith.wisc.edu>
572 * Make-lang.in (f77.mostlyclean): Add missing exeext.
574 Thu May 14 13:30:59 1998 Craig Burley <burley@gnu.org>
576 * Make-lang.in (f/expr.c): Now depends on f/stamp-str.
577 * expr.c: Use ffestrOther in place of ffeexprDotdot_.
578 * str-ot.fin: Add more keywords for expr.c.
580 * intdoc.c (dumpimp): Trivial fix.
582 * com.c (ffecom_expr_): Add ltkt variable for clarity.
584 Wed May 13 13:05:34 1998 Craig Burley <burley@gnu.org>
586 * Make-lang.in (G77STAGESTUFF): Add g77.o, g77spec.o,
588 (f77.clean): Add removal of g77.c, g77.o, g77spec.o,
590 (f77.distclean): Delete removal of g77.c.
592 Thu Apr 30 18:59:43 1998 Jim Wilson <wilson@cygnus.com>
594 * Make-lang.in (g77.info, g77.dvi, BUGS, INSTALL, NEWS): Put -o
595 option before input file.
597 Tue Apr 28 09:23:10 1998 Craig Burley <burley@gnu.org>
600 * global.c (ffeglobal_ref_progunit_): When transitioning
601 from EXT to FUNC, discard hook, since the decl, if any, is
604 Sun Apr 26 09:05:50 1998 Craig Burley <burley@gnu.org>
606 * com.c (ffecom_char_enhance_arg_): Wrap the upper bound
607 (the PARM_DECL specifying the length of the CHARACTER*(*)
608 dummy arg) in a variable_size invocation, to prevent
609 dwarf2out.c crashing when compiling code with -g.
611 Sat Apr 18 15:26:57 1998 Jim Wilson <wilson@cygnus.com>
613 * g77spec.c (lang_specific_driver): New argument in_added_libraries.
614 New local added_libraries. Increment count when add library to
617 Sat Apr 18 05:03:21 1998 Craig Burley <burley@gnu.org>
619 * com.c (ffecom_check_size_overflow_): Ignore overflow
620 as well if dummy argument.
622 Fri Apr 17 17:18:04 1998 Craig Burley <burley@gnu.org>
624 * version.h: Get rid of the overly large headers
625 here too, as done in version.c.
627 Tue Apr 14 15:51:37 1998 Dave Brolley <brolley@cygnus.com>
629 * com.c (init_parse): Now returns char* containing filename;
631 Tue Apr 14 14:40:40 1998 Craig Burley <burley@gnu.org>
633 * com.c (ffecom_start_progunit_): Mark function decl
634 as used, to avoid spurious warning (-Wunused) for ENTRY.
636 Tue Apr 14 14:19:34 1998 Craig Burley <burley@gnu.org>
638 * sta.c (ffesta_second_): Check for CASE DEFAULT
639 as well as CASE, or it won't be recognized.
641 Thu Apr 9 00:18:44 1998 Dave Brolley (brolley@cygnus.com)
643 * com.c (finput): New variable.
644 (init_parse): Handle !USE_CPPLIB.
645 (finish_parse): New function.
646 (lang_init): No longer declare finput.
648 Sat Apr 4 17:45:01 1998 Richard Henderson <rth@cygnus.com>
650 * com.c (ffecom_expr_): Revert Oct 22 change. Instead take a WIDENP
651 argument so that we can respect the signedness of the original type.
652 (ffecom_init_0): Do sizetype initialization first.
654 1998-03-28 Dave Love <d.love@dl.ac.uk>
656 * Make-lang.in (f771$(exeext)): Fix typo.
658 1998-03-24 Martin von Loewis <loewis@informatik.hu-berlin.de>
660 * com.c (lang_print_xnode): New function.
662 Mon Mar 23 21:20:35 1998 Craig Burley <burley@gnu.org>
664 * version.c: Reduce to a one-line file, like
665 gcc's version.c, since there's really no content
668 Mon Mar 23 11:58:43 1998 Craig Burley <burley@gnu.org>
670 * bugs.texi: Various updates.
672 * com.c (ffecom_tree_canonize_ptr_): Fix up spacing a bit.
674 Sun Mar 22 00:50:42 1998 Nick Clifton <nickc@cygnus.com>
675 Geoff Noer <noer@cygnus.com>
677 * Makefile.in: Various fixes for building cygwin32 native toolchains.
678 * Make-lang.in: Likewise.
680 Mon Mar 16 21:20:35 1998 Craig Burley <burley@gnu.org>
682 * expr.c (ffeexpr_sym_impdoitem_): Don't blindly
683 reset symbol info after calling ffesymbol_error,
686 Mon Mar 16 15:38:50 1998 Craig Burley <burley@gnu.org>
688 * Version 0.5.22 released.
690 Mon Mar 16 14:36:02 1998 Craig Burley <burley@gnu.org>
692 Make -g work better for ENTRY:
693 * com.c (ffecom_start_progunit_): Master function
694 for ENTRY-laden procedure is not really invented,
695 so it can be debugged.
696 (ffecom_do_entry_): Push/set/pop lineno for each
699 Sun Mar 15 05:48:49 1998 Craig Burley <burley@gnu.org>
701 * intrin.def: Fix spelling of mixed-case form
702 of `CPU_Time' (was `Cpu_Time').
704 Thu Mar 12 13:50:21 1998 Craig Burley <burley@gnu.org>
706 * lang-options.h: Sort all -f*-intrinsics-* options,
707 for consistency with other g77 versions.
709 Thu Mar 12 09:39:40 1998 Manfred Hollstein <manfred@s-direktnet.de>
711 * lang-specs.h: Properly put brackets around array elements in initializer.
713 1998-03-09 Dave Love <d.love@dl.ac.uk>
715 * Make-lang.in: Set CONFIG_SITE to a non-existent file since
716 /dev/null loses with bash 2.0/autoconf 2.12. Put
717 F77_FLAGS_TO_PASS before CC.
719 Sun Mar 8 16:35:34 1998 Craig Burley <burley@gnu.org>
721 * intrin.def: Use tabs instead of blanks more
722 consistently (excepting DEFGEN section for now).
724 Wed Mar 4 17:38:21 1998 Jeffrey A Law (law@cygnus.com)
726 * Make-lang.in: Remove more references to libf77.
728 Tue Mar 3 10:52:35 1998 Manfred Hollstein <manfred@s-direktnet.de>
730 * g77.texi: Use @url for citing URLs.
732 Sat Feb 28 15:24:38 1998 Craig Burley <burley@gnu.org>
734 * intrin.def: Make CPU_TIME's arg generic real to be just
737 Fri Feb 20 12:45:53 1998 Craig Burley <burley@gnu.org>
739 * expr.c (ffeexpr_token_arguments_): Make sure
740 outer exprstack isn't null.
742 1998-02-16 Dave Love <d.love@dl.ac.uk>
744 * Makefile.in (f/fini): Don't use -W -Wall with HOST_CC.
746 Fri Feb 13 00:14:56 1998 Kaveh R. Ghazi <ghazi@caip.rutgers.edu>
748 * com.c (type_for_mode): Add explicit braces to avoid ambiguous `else'.
750 * expr.c (ffeexpr_type_combine): Likewise.
751 (ffeexpr_reduce_): Likewise.
752 (ffeexpr_declare_parenthesized_): Likewise.
754 * src.c (ffesrc_strcmp_1ns2i): Likewise.
755 (ffesrc_strcmp_2c): Likewise.
756 (ffesrc_strncmp_2c): Likewise.
758 * stb.c (ffestb_halt1_): Likewise.
759 (ffestb_R90910_): Likewise.
760 (ffestb_R9109_): Likewise.
762 * stc.c (ffestc_R544_equiv_): Likewise.
764 * std.c (ffestd_subr_copy_easy_): Likewise.
765 (ffestd_R1001dump_): Likewise.
766 (ffestd_R1001dump_1005_1_): Likewise.
767 (ffestd_R1001dump_1005_2_): Likewise.
768 (ffestd_R1001dump_1005_3_): Likewise.
769 (ffestd_R1001dump_1005_4_): Likewise.
770 (ffestd_R1001dump_1005_5_): Likewise.
771 (ffestd_R1001dump_1010_2_): Likewise.
773 * ste.c (ffeste_R840): Likewise.
775 * sts.c (ffests_puttext): Likewise.
777 * symbol.c (ffesymbol_check_token_): Likewise.
779 * target.c (ffetarget_real1): Likewise.
780 (ffetarget_real2): Likewise.
782 Wed Feb 11 01:44:48 1998 Richard Henderson (rth@cygnus.com)
784 * com.c (ffecom_ptr_to_expr) [FFEBLD_opARRAYREF]: Do upper - lower
785 in the native type, so as to properly handle negative indices.
787 Tue Feb 3 20:13:05 1998 Richard Henderson <rth@cygnus.com>
789 * config-lang.in: Remove references to runtime/.
791 Sun Feb 1 12:43:49 1998 J"orn Rennecke <amylaar@cygnus.co.uk>
793 * com.c (ffecom_tree_canonize_ptr_): Place bitsizetype typed expr
794 as first agument in MULT_EXPR.
795 Use bitsize_int (0L, 0L) as zero for bitsizes.
796 (ffecom_tree_canonize_ref_):
797 Use bitsize_int (0L, 0L) as zero for bitsizes.
798 (ffecom_init_0): Use set_sizetype.
800 Sun Feb 1 02:26:58 1998 Richard Henderson <rth@cygnus.com>
802 * runtime directory -- moved into "libf2c" in the toplevel
804 * Make-lang.in: Remove all runtime related stuff.
806 Sun Jan 25 12:32:15 1998 Kaveh R. Ghazi <ghazi@caip.rutgers.edu>
808 * Make-lang.in (f77.stage1): Depend on stage1-start so parallel
810 * (f77.stage2): Likewise for stage2-start.
811 * (f77.stage3): Likewise for stage3-start.
812 * (f77.stage4): Likewise for stage4-start.
814 Sat Jan 17 21:28:08 1998 Pieter Nagel <pnagel@epiuse.co.za>
816 * Makefile.in (FLAGS_TO_PASS): Pass down gcc_include_dir and
817 local_prefix to sub-make invocations.
819 Tue Jan 13 22:07:54 1998 Jeffrey A Law (law@cygnus.com)
821 * lang-options.h: Add missing options.
823 Sun Jan 11 02:14:47 1998 Craig Burley <burley@gnu.org>
825 Support FORMAT(I<1+2>) (constant variable-FORMAT
827 * bad.def (FFEBAD_FORMAT_VARIABLE): New diagnostic.
828 * std.c (ffestd_R1001rtexpr_): New function.
829 (ffestd_R1001dump_, ffestd_R1001dump_1005_1_,
830 ffestd_R1001dump_1005_2_, ffestd_R1001dump_1005_3_,
831 ffestd_R1001dump_1005_4_, ffestd_R1001dump_1005_5_,
832 ffestd_R1001dump_1010_2_, ffestd_R1001dump_1010_3_,
833 ffestd_R1001dump_1010_4_, ffestd_R1001dump_1010_5_):
834 Use new function instead of ffestd_R1001error_.
836 * stb.c (ffestb_R10014_, ffestb_R10016_, ffestb_R10018_,
837 ffestb_R100110_): Restructure `for' loop for style.
839 Fix 970626-2.f by not doing most back-end processing
840 when current_function_decl is an ERROR_MARK, and by
841 making that the case when its type would be an ERROR_MARK:
842 * com.c (ffecom_start_progunit_, finish_function,
843 lang_printable_name, start_function,
844 ffecom_finish_symbol_transform_): Test for ERROR_MARK.
845 * std.c (ffestd_stmt_pass_): Don't do any downstream
846 processing if ERROR_MARK.
848 * Make-lang.in (f77.install-common): Don't install, and
849 don't uninstall existing, Info files if f/g77.info
850 doesn't exit. (This is a somewhat modified version
851 of an egcs patch on 1998-01-07 12:05:51 by Bruno Haible
852 <bruno@linuix.mathematik.uni-karlsruhe.de>.)
854 Fri Jan 9 19:09:07 1998 Craig Burley <burley@gnu.org>
856 Fix -fpedantic combined with `F()' invocation,
857 also -fugly-comma combined with `IARGC()' invocation:
858 * bad.def (FFEBAD_NULL_ARGUMENT_W): New diagnostic.
859 * expr.c (ffeexpr_finished_): Don't reject null expressions
860 in the argument-expression context -- let outer context
862 (ffeexpr_token_arguments_): Warn about null expressions
863 here if -fpedantic (as appropriate).
864 Obey -fugly-comma for only external-procedure invocations.
865 * intrin.c (ffeintrin_check_): No longer ignore explicit
866 omitted trailing args.
868 Tue Dec 23 14:58:04 1997 Craig Burley <burley@gnu.org>
870 * intrin.c (ffeintrin_fulfill_generic): Don't generate
871 FFEBAD_INTRINSIC_TYPE for CHARACTER*(*) intrinsic.
873 * com.c (ffecom_gfrt_basictype):
874 (ffecom_gfrt_kindtype):
876 (FFECOM_rttypeVOIDSTAR_): New return type `void *', for
877 the SIGNAL intrinsic.
878 * com-rt.def (FFECOM_rttypeSIGNAL): Now returns `void *'.
879 * intdoc.c: Replace `p' kind specifier with `7'.
880 * intrin.c (ffeintrin_check_, ffeintrin_init_0): Replace
881 `p' kind specifier with `7'.
882 * intrin.def (FFEINTRIN_impLOC, FFEINTRIN_impSIGNAL_func,
883 FFEINTRIN_impSIGNAL_subr): Replace `p' specifier with `7'.
884 Also, SIGNAL now returns a `void *' status, not `int'.
886 Improve run-time diagnostic for "PRINT '(I1', 42":
887 * com.c (ffecom_char_args_x_): Renamed from ffecom_char_args_,
888 which is now a macro (to avoid lots of changes to other code)
889 with new arg, ffecom_char_args_with_null_ being another new
890 macro to call same function with different value for new arg.
891 This function now appends a null byte to opCONTER expression
892 if the new arg is TRUE.
893 (ffecom_arg_ptr_to_expr): Support NULL length pointer.
894 * ste.c (ffeste_io_cilist_):
895 (ffeste_io_icilist_): Pass NULL length ptr for
896 FORMAT expression, so null byte gets appended where
898 * target.c (ffetarget_character1):
899 (ffetarget_concatenate_character1):
900 (ffetarget_substr_character1):
901 (ffetarget_convert_character1_character1):
902 (ffetarget_convert_character1_hollerith):
903 (ffetarget_convert_character1_integer4):
904 (ffetarget_convert_character1_logical4):
905 (ffetarget_convert_character1_typeless):
906 (ffetarget_hollerith): Append extra phantom null byte as
907 part of FFETARGET-NULL-BYTE kludge.
909 * intrin.def (FFEINTRIN_impCPU_TIME): Point to
910 FFECOM_gfrtSECOND as primary run-time routine.
912 Mon Dec 22 12:41:07 1997 Craig Burley <burley@gnu.org>
914 * intrin.c (ffeintrin_init_0): Remove duplicate
917 Fri Dec 19 00:12:01 1997 Richard Henderson <rth@cygnus.com>
919 * com.c (ffecom_sym_transform_): Assumed arrays have no upper bound.
921 Mon Dec 15 17:35:35 1997 Richard Henderson <rth@cygnus.com>
923 * com.c (ffecom_type_vardesc_): Vardesc.dims is a `ftnlen*'.
925 Sun Dec 14 02:49:58 1997 Craig Burley <burley@gnu.org>
927 * intrin.c (ffeintrin_init_0): Fix up indentation a bit.
928 Fix bug that prevented checking of arguments other
931 * intdoc.c: Fix up indentation a bit.
933 Tue Dec 9 16:20:57 1997 Richard Henderson <rth@cygnus.com>
935 * com.c (ffecom_type_vardesc_): Vardesc.dims is a `ftnlen*'.
937 Tue Dec 2 09:57:16 1997 Jeffrey A Law (law@cygnus.com)
939 * Make-lang.in (f77.clean): Remove g77.c.
941 Mon Dec 1 19:12:36 1997 Craig Burley <burley@gnu.org>
943 * intrin.c (ffeintrin_check_): Fix up indentation a bit more.
945 Mon Dec 1 16:21:08 1997 Craig Burley <burley@gnu.org>
947 * com.c (ffecom_arglist_expr_): Crash if non-supplied
948 optional arg isn't passed as an address.
949 Pass null pointer explicitly, instead of via ffecom routine.
950 If incoming argstring is NULL, substitute pointer to "0".
951 Recognize '0' as ending the usual arg stuff, just like '\0'.
953 Sun Nov 30 22:22:22 1997 Craig Burley <burley@gnu.org>
955 * intdoc.c: Minor fix-ups.
957 * intrin.c (ffeintrin_check_): Fix up indentation a bit.
959 * intrin.def: Fix up spacing a bit.
961 Tue Nov 25 15:33:28 1997 Jeffrey A Law (law@cygnus.com)
963 * Make-lang.in (f77.all.build): Add $(exeext) to binary files.
964 (f77.all.cross, f77.start.encap): Simliarly.
966 Fri Nov 21 09:35:20 1997 Fred Fish <fnf@cygnus.com>
968 * Make-lang.in (stmp-f2c.h): Move inclusion of F77_FLAGS_TO_PASS
969 to before override of CC so that the override works.
971 Thu Nov 20 00:58:14 1997 H.J. Lu (hjl@gnu.ai.mit.edu)
973 * Make-lang.in (f77.install-info): Depend on f77.info.
975 1997-11-17 Dave Love <d.love@dl.ac.uk>
977 * com.c (ffecom_arglist_expr_): Pass null pointers for optional
978 args which aren't supplied.
980 Sun Nov 16 21:45:43 1997 H.J. Lu (hjl@gnu.ai.mit.edu)
982 * Make-lang.in (f77.install-info): Depend on f77.info.
984 1997-11-14 Dave Love <d.love@dl.ac.uk>
986 * intrin.def: Supply gfrt for CPU_TIME. Generalize arg types of
989 1997-11-06 Dave Love <d.love@dl.ac.uk>
991 * intrin.def: Allow non-integer args for INT2 and INT8 (per
994 Sun Nov 2 19:49:51 1997 Richard Henderson <rth@cygnus.com>
996 * com.c (ffecom_expr_): Only use TREE_TYPE argument for simple
997 arithmetic; convert types as necessary; recurse with target tree type.
999 Tue Oct 28 02:21:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
1001 * lang-options.h: Add -fgnu-intrinsics-* and
1002 -fbadu77-intrinsics-* options.
1004 Sun Oct 26 02:36:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
1006 * com.c (lang_print_error_function): Fix to more
1007 reliably notice when the diagnosed region changes.
1009 Sat Oct 25 23:43:36 1997 Craig Burley <burley@gnu.ai.mit.edu>
1012 * sta.c, sta.h (ffesta_outpooldisp): New function.
1013 * std.c (ffestd_stmt_pass_): Don't kill NULL pool.
1014 (ffestd_R842): If pool already preserved, save NULL
1015 for pool, because it should be killed only once.
1017 * malloc.c [MALLOC_DEBUG]: Put initializer for `name'
1018 component in braces, to avoid compiler warning.
1020 Wed Oct 22 11:37:41 1997 Richard Henderson <rth@cygnus.com>
1022 * com.c (ffecom_expr_): Take an new arg TREE_TYPE that if non-null
1023 specifies the type in which to do the calculation. Change all callers.
1024 [FFEBLD_opARRAYREF]: Force the index expr to use sizetype.
1026 Thu Oct 16 02:04:08 1997 Paul Koning <pkoning@xedia.com>
1028 * Make-lang.in (stmp-f2c.h): Don't configure the runtime
1029 directory if LANGUAGES does not include f77.
1031 Mon Oct 13 12:12:41 1997 Richard Henderson <rth@cygnus.com>
1033 * Make-lang.in (g77*): Copied from cp/Make-lang.in g++*.
1034 * g77spec.c: New file, mostly copied from g++spec.c
1037 Fri Oct 10 13:00:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
1039 * ste.c (ffeste_begin_iterdo_): Fix loop setup so iteration
1040 variable is modified only after the #iterations is calculated;
1041 otherwise if the iteration variable is aliased to any of the
1042 operands in the start, end, or increment expressions, the
1043 wrong #iterations might be calculated.
1045 * com.c (ffecom_save_tree): Fix indentation.
1047 Mon Oct 6 14:15:03 1997 Jeffrey A Law (law@cygnus.com)
1049 * Make-lang.in (f77.mostlyclean): Clean up stuff in the
1051 (f77.clean, f77.distclean, f77.maintainer-clean): Likewise.
1053 1997-10-05 Dave Love <d.love@dl.ac.uk>
1055 * intrin.def: Make SECOND_subr's arg generic real for people
1056 porting from Cray and making everything double precision.
1058 Wed Oct 1 01:45:36 1997 Philippe De Muyter <phdm@info.ucl.ac.be>
1060 * g77.c (pexecute, main): Use unlink, not remove.
1062 Mon Sep 29 16:18:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
1064 * stu.c (ffestu_list_exec_transition_,
1065 ffestu_dummies_transition_): Specify `bool' type for
1066 `in_progress' variables.
1068 * com.h (assemble_string): Declare this routine (instead
1069 of #include'ing "output.h" from gcc) to eliminate warnings
1072 Mon Sep 29 10:37:07 1997 Jeffrey A Law (law@cygnus.com)
1074 * intdoc.c (main): Remove unused attribute for main's arguments.
1076 Sun Sep 28 01:47:17 1997 Jeffrey A Law (law@cygnus.com)
1078 * Make-lang.in (G77_FLAGS_TO_PASS): Pass down RANLIB, RANLIB_TEST
1079 and AR instead of the _FOR_TARGET versions.
1081 Tue Sep 23 00:39:57 1997 Alexandre Oliva <oliva@dcc.unicamp.br>
1083 * Make-lang.in: install.texi was renamed to g77install.texi
1084 * install0.texi: Likewise.
1086 Fri Sep 19 01:12:27 1997 Craig Burley <burley@gnu.ai.mit.edu>
1088 * expr.c (ffeexpr_reduced_eqop2_):
1089 (ffeexpr_reduced_relop2_): Minor fixes to diagnostic code.
1091 * fini.c (main): Change return type to `int'.
1093 Thu Sep 18 17:31:38 1997 Jeffrey A Law (law@cygnus.com)
1095 * proj.h (FFEPROJ_BSEARCH): Delete all references.
1096 (FFEPROJ_STRTOUL): Likewise.
1097 * proj.c (bsearch): Compile this if no bsearch is provided by the
1099 (strtoul): Similarly.
1101 * g77install.texi: Renamed from install.texi
1102 * g77.texi: Corresponding changes.
1104 * fini.c (main): Return type is int.
1106 * com.c (lang_printable_name): Use verbosity argument.
1108 Thu Sep 18 16:08:40 1997 Jeffrey A Law (law@cygnus.com)
1110 * Make-lang.in: Fix merge problems.
1112 Wed Sep 17 10:47:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1114 * com-rt.def (FFECOM_gfrtDSIGN, FFECOM_gfrtISIGN,
1115 FFECOM_gfrtSIGN): Add second argument.
1117 * expr.c (ffeexpr_cb_comma_c_): Trivial fixes.
1119 Sun Sep 14 21:01:23 1997 Jeffrey A Law (law@cygnus.com)
1121 * Make-lang.in: Various changes to build info files
1122 in the object tree rather than the source tree.
1124 * proj.h: Include ctype.h.
1126 Sun Sep 14 12:35:20 1997 Fred Fish (fnf@ninemoons.com)
1128 * proj.h (isascii): Provide a default definition if none is available.
1130 Thu Sep 11 19:26:10 1997 Dave Love <d.love@dl.ac.uk>
1132 * config-lang.in: Remove the messages about possible build problems.
1134 Wed Sep 10 16:39:47 1997 Jim Wilson <wilson@cygnus.com>
1136 * Make-lang.in (LN, LN_S): New macros, use where appropriate.
1138 Tue Sep 9 13:20:40 1997 Jim Wilson <wilson@cygnus.com>
1140 * g77.c (pexecute, doit): Add checks for __CYGWIN32__.
1142 Tue Sep 9 01:59:35 1997 Craig Burley <burley@gnu.ai.mit.edu>
1144 * Version 0.5.21 released.
1146 Tue Sep 9 00:31:01 1997 Craig Burley <burley@gnu.ai.mit.edu>
1148 * intdoc.c (dumpem): Put appropriate commentary in
1149 output file, so readers know it isn't source.
1151 Wed Aug 27 20:32:03 1997 Jeffrey A Law (law@cygnus.com)
1153 * top.c (ffe_decode_option): Turn on flag_move_all_moveables
1154 and flag_reduce_all_givs.
1156 Wed Aug 27 08:08:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
1158 * proj.h: Always #include "config.j" first, to pick up
1159 gcc's configuration.
1160 * com.c: Change bcopy() and bzero() calls to memcpy()
1161 and memset() calls, to make more of g77 ANSI C.
1163 1997-08-26 Dave Love <d.love@dl.ac.uk>
1165 * Make-lang.in ($(srcdir)/f/runtime/configure,
1166 $(srcdir)/f/runtime/libU77/configure): Fix for when srcdir isn't
1169 Tue Aug 26 05:59:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
1171 * ansify.c (main): Make sure readers of stdout know
1172 it's derived from stdin; omit comment text; get source
1173 line numbers in future stderr output to be correct.
1175 Tue Aug 26 01:36:01 1997 Craig Burley <burley@gnu.ai.mit.edu>
1178 * stb.c (ffestb_R5284_): Allow OPEN_PAREN after closing
1179 SLASH as well as NAME.
1181 Mon Aug 25 23:48:17 1997 Craig Burley <burley@gnu.ai.mit.edu>
1183 Changes to allow g77 docs to be built entirely from scratch
1184 using any ANSI C compiler, not requiring GNU C:
1185 * Make-lang.in ($(srcdir)/f/intdoc.texi): "Pipe" new
1186 location of intrinsic documentation data base, f/intdoc.in,
1187 through new `ansify' program to append `\n\' to quoted
1188 newlines, into f/intdoc.h0. Do appropriate cleanups. Explain.
1189 (f77.mostlyclean): Add f/ansify and f/intdoc.h0 to cleanups.
1190 * f/ansify.c: New program.
1191 * f/intdoc.c: Fix so it conforms to ANSI C.
1192 #include f/intdoc.h0 instead of f/intdoc.h.
1193 Avoid some warnings.
1194 * f/intdoc.h, f/intdoc.in: Rename the former to the latter; no
1195 changes made to the content in this patch!
1196 * f/intrin.h (ffeintrinFamily): Fix to conform to ANSI C.
1198 Mon Aug 25 23:24:32 1997 H.J. Lu (hjl@gnu.ai.mit.edu)
1200 * Make-lang.in ($(srcdir)/f/runtime/configure,
1201 $(srcdir)/f/runtime/libU77/configure, f77.mostlyclean,
1202 f77.clean, f77.distclean, f77.maintainer-clean, f77.realclean):
1203 Handle absolute pathname of $(srcdir).
1205 (include/f2c.h, f/runtime/Makefile, f/runtime/libF77/Makefile,
1206 f/runtime/libI77/Makefile, f/runtime/libU77/Makefile): Only
1207 depend on stmp-f2c.h.
1208 (f77.maintainer-clean): Don't make itself.
1210 Sun Aug 24 17:00:27 1997 Jim Wilson <wilson@cygnus.com>
1212 * Make-lang.in (f77.install-info): Don't cd into srcdir. Add srcdir
1213 to filenames. Use sed to extract base filename for install.
1215 Sun Aug 24 06:52:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
1217 Fix up g77 compiler data base for libf2c routines:
1218 * com-rt.def (FFECOM_gfrtSIGNAL): Change return type to
1219 FTNINT to match actual code.
1221 * com.c (ffecomRttype_): Replace FFECOM_rttypeINT_ with
1222 FFECOM_rttypeFTNINT_.
1223 Add and fix up comments.
1224 (ffecom_make_gfrt_, ffecom_gfrt_basictype,
1225 ffecom_gfrt_kindtype): Replace FFECOM_rttypeINT_ with
1226 FFECOM_rttypeFTNINT_; add FFECOM_rttypeDOUBLEREAL_.
1228 Thu Aug 21 13:15:29 1997 Jim Wilson <wilson@cygnus.com>
1230 * Make-lang.in (f77): Delete f77-runtime.
1231 (f77.all.build, f77.all.cross, f77.rest.encap): Add f77-runtime.
1233 Wed Aug 20 17:18:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
1235 * global.c (ffeglobal_ref_progunit_): It's okay to have
1236 a different CHARACTER*n length for a reference if the
1237 existing length is for another reference, not a definition.
1239 Wed Aug 20 16:36:59 1997 Jim Wilson <wilson@cygnus.com>
1241 * intdoc.texi: Readd generated file.
1243 Mon Aug 18 14:27:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
1246 * global.c (ffeglobal_new_progunit_): Distinguish
1247 between previously defined, versus inferred, filewide
1248 when it comes to diagnostics.
1251 * global.c (ffeglobal_ref_progunit_): Change BDATA into EXT
1252 right at the beginning, so EXTERNAL FOO followed later
1253 by SUBROUTINE FOO is not diagnosed.
1256 * com-rt.def (FFECOM_gfrtALARM): Returns `integer', not
1259 Mon Aug 18 09:01:54 1997 Jeffrey A Law (law@cygnus.com)
1261 * Makefile.in (F77_OBJS): Re-alphabetize.
1262 * Make-lang.in (F77_SRCS): Likewise.
1264 Sun Aug 17 08:35:11 1997 Jeffrey A Law (law@cygnus.com)
1267 * install.texi: Remove "Object File Differences" section. Remove
1268 all references to zzz.o failing comparison tests.
1269 * version.c, version.h: Renamed from zzz.c and zzz.h. Remove
1270 date and time stamps so a 3 stage build reports no differences.
1271 * Make-lang.in: Corresponding changes.
1272 * Makefile.in: Likewise.
1273 * g77.c, parse.c: Likewise.
1275 * intdoc.texi: Remove generated file from distribution.
1277 Sun Aug 17 03:32:44 1997 Craig Burley <burley@gnu.ai.mit.edu>
1279 Fix up problems when virtual memory exhausted:
1280 * malloc.c (malloc_new_): Use gcc's xmalloc(), so we
1281 print a nicer message when malloc returns no memory.
1282 (malloc_resize_): Ditto for xrealloc().
1284 * Make-lang.in, Makefile.in: Comment out lines containing
1287 Sat Aug 16 19:41:33 1997 Craig Burley <burley@gnu.ai.mit.edu>
1289 * com.c (ffecom_make_gfrt_): For rttypeREAL_F2C_, return
1290 double_type_node; for rttypeREAL_GNU_, return
1293 1997-08-13 Dave Love <d.love@dl.ac.uk>
1295 * config-lang.in (diff_excludes): Add some hints about known
1296 problematic platforms.
1298 1997-08-13 Dave Love <d.love@dl.ac.uk>
1300 * intdoc.h: Document `alarm'.
1302 Tue Aug 12 10:23:02 1997 Jeffrey A Law (law@cygnus.com)
1304 * config-lang.in: Don't demand the backend patch.
1305 * com.c (lang_printable_name): Second argument is now an int. Don't
1306 store into the value of the second argument.
1307 * top.c (ffe_decode_option): Temporarily disable setting
1308 of "Toon" loop options until we figure out how to address
1311 Mon Aug 11 23:18:35 1997 Jeffrey A Law (law@cygnus.com)
1313 * g77-0.5.21-19970811 Imported.
1314 This file describes changes to the front end necessary to make
1317 Mon Aug 11 21:19:22 1997 Craig Burley <burley@gnu.ai.mit.edu>
1319 * Make-lang.in ($(RUNTIMESTAGESTUFF)): Add
1320 f/runtime/stamp-lib.
1322 Mon Aug 11 01:52:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
1324 * com.c (ffecom_build_complex_constant_): Go with the
1325 new build_complex() approach used in gcc-2.8.
1327 * com.c (ffecom_sym_transform_): Don't set
1328 DECL_IN_SYSTEM_HEADER for a tree node that isn't
1329 a VAR_DECL, which happens when var is in common!
1331 * com.c (ffecom_expr_intrinsic_) (case FFEINTRIN_impALARM):
1332 No need to test codegen_imp -- there's only one valid here.
1334 * intrin.def (FFEINTRIN_impALARM): Specify `Status' argument
1337 Fri Aug 8 05:40:23 1997 Craig Burley <burley@gnu.ai.mit.edu>
1339 Substantial changes to accommodate distinctions among
1340 run-time routines that support intrinsics, and between
1341 routines that compute and return the same type vs. those
1342 that compute one type and return another (or `void'):
1343 * com-rt.def: Specify new return type REAL_F2C_ instead
1344 of many DOUBLE_, COMPLEX_F2C_ instead of COMPLEX_, and
1346 Clear up the *BES* routines "once and for all".
1347 * com.c: New return types.
1348 (ffecom_convert_narrow_, ffecom_convert_widen_):
1349 New functions that are "safe" variants of convert(),
1350 to catch errors that ffecom_expr_intrinsic_() now
1352 (ffecom_arglist_expr_): Ensure arguments are not
1353 converted to narrower types.
1354 (ffecom_call_): Ensure return value is not converted
1356 (ffecom_char_args_): Use new ffeintrin_gfrt_direct()
1358 (ffecom_expr_intrinsic_): Simplify how run-time
1359 routine is selected (via `gfrt' only now; lose the
1360 redundant `ix' variable).
1361 Eliminate the `library' label; any code that doesn't
1362 return directly just `break's out now with `gfrt'
1364 Set `gfrt' to default choice initially, either a
1365 fast direct form or, if not available, a slower
1366 indirect-callable form.
1367 (ffecom_make_gfrt_): No longer need to do special
1368 check for complex; it's built into the new return-type
1370 (ffecom_ptr_to_expr): Use new ffeintrin_gfrt_indirect()
1372 * intrin.c, intrin.h: `gfrt' field replaced with three fields,
1373 so it is easier to provide faster direct-callable and
1374 GNU-convention indirect-callable routines in the future.
1375 DEFIMP macro adjusted accordingly, along with all its uses.
1376 (ffeintrin_gfrt_direct): New function.
1377 (ffeintrin_gfrt_indirect): Ditto.
1378 (ffeintrin_is_actualarg): If `-fno-f2c' is in effect,
1379 require a GNU-callable version of intrinsic instead of
1380 an f2c-callable version, so indirect calling is still checked.
1381 * intrin.def: Replace one GFRT field with the three new fields,
1382 as appropriate for each DEFIMP intrinsic.
1384 * com.c (ffecom_stabilize_aggregate_,
1385 ffecom_convert_to_complex_): Make these `static'.
1387 Thu Aug 7 11:24:34 1997 Craig Burley <burley@gnu.ai.mit.edu>
1389 Provide means for front end to determine actual
1390 "standard" return type for an intrinsic if it is
1391 passed as an actual argument:
1392 * com.h, com.c (ffecom_gfrt_basictype,
1393 ffecom_gfrt_kindtype): New functions.
1394 (ffecom_gfrt_kind_type_): Replaced with new function.
1395 All callers updated.
1396 (ffecom_make_gfrt_): No longer need do anything
1399 * intrin.c (ffeintrin_basictype, ffeintrin_kindtype):
1400 Now returns correct type info for specific intrinsic
1401 (based on type of run-time-library implementation).
1403 Wed Aug 6 23:08:46 1997 Craig Burley <burley@gnu.ai.mit.edu>
1405 * global.c (ffeglobal_ref_progunit_): Don't reset
1406 number of arguments just due to new type info,
1407 so useful warnings can be issued.
1409 1997-08-06 Dave Love <d.love@dl.ac.uk>
1411 * intrin.def: Fix IDATE_vxt argument order.
1412 * intdoc.h: Likewise.
1414 Thu Jul 31 22:22:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
1416 * global.c (ffeglobal_proc_ref_arg): If REF/DESCR
1417 disagreement, DESCR is CHARACTER, and types disagree,
1418 pretend the argsummary agrees so the message ends up
1419 being about type disagreement.
1420 (ffeglobal_proc_def_arg): Ditto.
1422 * expr.c (ffeexpr_token_first_rhs_3_): Set info for LABTOK
1423 to NONE of everything, to avoid misdiagnosing filewide
1424 usage of alternate returns.
1426 Sun Jul 20 23:07:47 1997 Craig Burley <burley@gnu.ai.mit.edu>
1428 * com.c (ffecom_sym_transform_): If type gets set
1429 to error_mark_node, just return that for transformed symbol.
1430 (ffecom_member_phase2_): If type gets set to error_mark_node,
1432 (ffecom_check_size_overflow_): Add `dummy' argument to
1433 flag that type is for a dummy, update all callers.
1435 Sun Jul 13 17:40:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
1438 * where.c (ffewhere_set_from_track): If start point
1439 is too large, just use initial start point. 0.6 should
1440 fix all this properly.
1443 * com.c (ffecom_sym_transform_): Preserve error_mark_node for type.
1444 (ffecom_type_localvar_): Ditto.
1445 (ffecom_sym_transform_): If type is error_mark_node,
1446 don't error-check decl size, because back end responds by
1447 setting that to an integer 0 instead of error_mark_node.
1448 (ffecom_transform_common_): Same as earlier fix to _transform_
1449 in that size is checked by dividing BITS_PER_UNIT instead of
1451 (ffecom_transform_equiv_): Ditto.
1454 * stb.c (ffestb_R10014_): Fix flaky fall-through in error
1455 test for FFELEX_typeCONCAT by just replicating the code,
1456 and do FFELEX_typeCOLONCOLON while at it.
1458 1997-07-07 Dave Love <d.love@dl.ac.uk>
1460 * intdoc.h: Add various missing pieces; correct GMTIME, LTIME
1463 * intrin.def, com-rt.def: Add alarm.
1465 * com.c (ffecom_expr_intrinsic_): Add case for alarm.
1467 Thu Jun 26 04:19:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
1470 * com.c (ffecom_sym_transform_): For sanity-check compare
1471 of gbe size of local variable to g77 expectation,
1472 use varasm.c/assemble_variable technique of dividing
1473 BITS_PER_UNIT out of gbe info instead of multiplying
1474 g77 info up, to avoid crash when size in bytes is very
1475 large, and overflows an `int' or similar when multiplied.
1478 * com.c (ffecom_finish_symbol_transform_): Don't bother
1479 transforming a dummy argument, to avoid a crash.
1480 * ste.c (ffeste_R1227): Don't return a value if the
1481 result decl, or its type, is error_mark_node.
1484 * lex.c (ffelex_splice_tokens): `-fdollar-ok' is
1485 irrelevant to whether a DOLLAR token should be made
1486 from an initial character of `$'.
1489 * stb.c (ffestb_do3_): DO iteration variable is an
1490 lhs, not rhs, expression.
1492 Fix 970626-7.f and 970626-8.f:
1493 * expr.c (ffeexpr_cb_comma_i_1_): Set IMPDO expression
1494 to have clean info, because undefined rank, for example,
1495 caused crash on mangled source on UltraSPARC but not
1496 on Alpha for a series of weird reasons.
1497 (ffeexpr_cb_close_paren_): If not CLOSE_PAREN, push
1498 opANY expression onto stack instead of attempting
1499 to mimic what program might have wanted.
1500 (ffeexpr_cb_close_paren_): Don't wrap opPAREN around
1501 opIMPDO, just warn that it's gratuitous.
1502 * bad.def (FFEBAD_IMPDO_PAREN): New warning.
1505 * expr.c (ffeexpr_declare_parenthesized_): Must shut down
1506 parsing in kindANY case, otherwise the parsing engine might
1507 decide there's an ambiguity.
1508 (ffeexpr_token_name_rhs_): Eliminate parentypeSUBROUTINE_
1509 case, so we crash right away if it comes through.
1510 * st.c, st.h, sta.c, sta.h (ffest_shutdown, ffesta_shutdown):
1513 Tue Jun 24 19:47:29 1997 Craig Burley <burley@gnu.ai.mit.edu>
1515 * com.c (ffecom_check_size_overflow_): New function
1516 catches some cases of the size of a type getting
1517 too large. varasm.c must catch the rest.
1518 (ffecom_sym_transform_): Use new function.
1519 (ffecom_type_localvar_): Ditto.
1521 Mon Jun 23 01:09:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
1523 * global.c (ffeglobal_proc_def_arg): Fix comparison
1525 (ffeglobal_proc_ref_arg): Ditto.
1527 * lang-options.h, top.c: Rename `-fdebug' to `-fxyzzy',
1528 since it's an unsupported internals option and some
1529 poor user might guess that it does something.
1531 * bad.def: Make a warning for each filewide diagnostic.
1532 Put all filewides together.
1533 * com.c (ffecom_sym_transform_): Don't substitute
1534 known global tree for global entities when `-fno-globals'.
1535 * global.c (ffeglobal_new_progunit_): Don't produce
1536 fatal diagnostics about globals when `-fno-globals'.
1537 Instead, produce equivalent warning when `-Wglobals'.
1538 (ffeglobal_proc_ref_arg): Ditto.
1539 (ffeglobal_proc_ref_nargs): Ditto.
1540 (ffeglobal_ref_progunit_): Ditto.
1541 * lang-options.h, top.c, top.h: New `-fno-globals' option.
1543 Sat Jun 21 12:32:54 1997 Craig Burley <burley@gnu.ai.mit.edu>
1545 * expr.c (ffeexpr_fulfill_call_): Set array variable
1546 to avoid warning about uninitialized variable.
1548 * Make-lang.in: Get rid of any setting of HOST_* macros,
1549 since these will break gcc's build!
1550 * makefile: New file to make building derived files
1553 Thu Jun 19 18:19:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
1555 * g77.c (main): Install Emilio Lopes' patch to support
1556 Ratfor, and to fix the printing of the version string
1557 to go to stderr, not stdout.
1558 * lang-specs.h: Install Emilio Lopes' patch to support
1559 Ratfor, and patch the result to support picking up
1560 `*f771' from the `specs' file.
1562 Thu Jun 12 14:36:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
1564 * storag.c (ffestorag_update_init, ffestorag_update_save):
1565 Also update parent, in case equivalence processing
1566 has already eliminated pointers to it via the
1567 local equivalence info.
1569 Tue Jun 10 14:08:26 1997 Craig Burley <burley@gnu.ai.mit.edu>
1571 * intdoc.c: Add cross-reference to end of description
1572 of any generic intrinsic pointing to other intrinsics
1575 Warn about explicit type declaration for intrinsic
1576 that disagrees with invocation:
1577 * expr.c (ffeexpr_paren_rhs_let_): Preserve type info
1578 for intrinsic functions.
1579 (ffeexpr_token_funsubstr_): Ditto.
1580 * intrin.c (ffeintrin_fulfill_generic): Warn if type
1581 info of fulfilled intrinsic invocation disagrees with
1582 explicit type info given symbol.
1583 (ffeintrin_fulfill_specific): Ditto.
1584 * stc.c (ffestc_R1208_item): Preserve type info
1586 (ffestc_R501_item): Ditto.
1588 Mon Jun 9 17:45:44 1997 Craig Burley <burley@gnu.ai.mit.edu>
1590 * com.c (ffecom_expr_intrinsic_): Fix several of the
1591 libU77/libF77-unix handlers to properly convert their
1594 * com-rt.def (FFECOM_gfrtFSTAT): Append missing "i" to
1597 Fri Jun 6 14:37:30 1997 Craig Burley <burley@gnu.ai.mit.edu>
1599 * com.c (ffecom_expr_intrinsic_): Have a case statement
1600 for every intrinsic implementation, so missing ones
1601 are caught via gcc warnings.
1602 Don't call ffeintrin_codegen_imp anymore.
1603 * intrin.c (ffeintrin_fulfill_generic): Remove cg_imp
1605 (ffeintrin_codegen_imp): Delete this function.
1606 * intrin.def, intrin.h: Remove DEFIMQ stuff from here
1609 Thu Jun 5 13:03:07 1997 Craig Burley <burley@gnu.ai.mit.edu>
1611 * top.c (ffe_decode_option): New -fbadu77-intrinsics-*
1614 * intrin.h: New BADU77 family.
1615 * intrin.c (ffeintrin_state_family): Ditto.
1617 Implement new scheme to track intrinsic names vs. forms:
1618 * intrin.c (ffeintrin_fulfill_generic),
1619 (ffeintrin_fulfill_specific), (ffeintrin_is_intrinsic),
1620 intrin.def: The documented name is now either in the
1621 generic info or, if no generic, in the specific info.
1622 For a generic, the specific info contains merely the
1623 distinguishing form (usually "function" or "subroutine"),
1624 used for diagnostics about ambiguous references and
1625 in the documentation.
1627 * intrin.def: Clean up formatting of DEFNAME block.
1628 Convert many libU77 intrinsics into generics that
1629 support both subroutine and function forms.
1630 Put the function forms of side-effect routines into
1631 the new BADU77 family.
1632 Make MCLOCK and TIME return INTEGER*4 again, and add
1633 INTEGER*8 equivalents called MCLOCK8 and TIME8.
1634 Fix up more status return values to be written and
1635 insist on them being I1 as well.
1636 * com.c (ffecom_expr_intrinsic_): Lots of changes to
1637 support new libU77 intrinsic interfaces.
1639 Mon Jun 2 00:37:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
1641 * com.c (ffecom_init_0): Pointer type is now INTEGER(KIND=7),
1642 not INTEGER(KIND=0), since we want to reserve KIND=0 for
1645 Thu May 29 14:30:33 1997 Craig Burley <burley@gnu.ai.mit.edu>
1647 Fix bugs preventing CTIME(I*4) from working correctly:
1648 * com.c (ffecom_char_args_): For FUNCREF case, process
1649 args to intrinsic just as they would be in
1650 ffecom_expr_intrinsic_.
1651 * com-rt.def (FFECOM_gfrtCTIME, FFECOM_gfrtTTYNAM): Fix
1652 argument decls to specify `&'.
1654 Wed May 28 22:19:49 1997 Craig Burley <burley@gnu.ai.mit.edu>
1656 Fix gratuitous warnings exposed by dophot aka 970528-1:
1657 * global.c (ffeglobal_proc_def_arg, ffeglobal_proc_ref_arg):
1658 Support distinct function/subroutine arguments instead of
1661 * expr.c (ffeexpr_fulfill_call_): A SYMTER with kindNONE
1662 also is a procedure (either function or subroutine).
1664 Mon May 26 20:25:31 1997 Craig Burley <burley@gnu.ai.mit.edu>
1666 * bad.def: Have several lexer diagnostics refer to
1667 documentation for people who need more info on what Fortran
1668 source code is supposed to look like.
1670 * expr.c (ffeexpr_reduced_bool1_), bad.def: New diagnostics
1671 specific to .NOT. now mention only one operand instead
1674 * g77.c: Recognize -fsyntax-only, similar to -c etc.
1675 (lookup_option): Fix bug that prevented non-`--' options
1676 from being recognized.
1678 Sun May 25 04:29:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
1680 * intrin.def (FFEINTRIN_impCTIME): Accept `I*' expression
1681 for STime instead of requiring `I2'.
1683 Tue May 20 16:14:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
1685 * symbol.c (ffesymbol_reference): All references to
1686 standard intrinsics are considered explicit, so as
1687 to avoid generating basically useless warnings.
1688 * intrin.c, intrin.h (ffeintrin_is_standard): Returns TRUE
1689 if intrinsic is standard.
1691 Sun May 18 21:14:59 1997 Craig Burley <burley@gnu.ai.mit.edu>
1693 * com-rt.def: Changed all external names of the
1694 form `"\([a-z0-9]*\)_' to `"G77_\1_0"' so as to
1695 allow any name valid as an intrinsic to be used
1696 as such and as a user-defined external procedure
1697 name or common block as well.
1699 Thu May 8 13:07:10 1997 Craig Burley <burley@gnu.ai.mit.edu>
1701 * expr.c (ffeexpr_cb_end_notloc_): For %VAL, %REF, and
1702 %DESCR, copy arg info into new node.
1704 Mon May 5 14:42:17 1997 Craig Burley <burley@gnu.ai.mit.edu>
1706 From Uwe F. Mayer <mayer@math.Vanderbilt.Edu>:
1707 * Make-lang.in (g77-cross): Fix typo in g77.c path.
1709 From Brian McIlwrath <bkm@star.rl.ac.uk>:
1710 * lang-specs.h: Have g77 pick up options from a section
1711 labeled `*f771' of the `specs' file.
1713 Sat May 3 02:46:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1715 * intrin.def (FFEINTRIN_defSIGNAL): Add optional `Status'
1716 argument that com.c already expects (per Dave Love).
1718 More changes to support better tracking of (filewide)
1719 globals, in particular, the arguments to procedures:
1720 * bad.def (FFEBAD_FILEWIDE_NARGS, FFEBAD_FILEWIDE_NARGS_W,
1721 FFEBAD_FILEWIDE_ARG, FFEBAD_FILEWIDE_ARG_W): New diagnostics.
1722 * expr.c (ffebad_fulfill_call_): Provide info on each
1723 argument to ffeglobal.
1724 * global.c, global.h (ffeglobal_proc_def_arg,
1725 ffeglobal_proc_def_nargs, ffeglobal_proc_ref_arg,
1726 ffeglobal_proc_ref_args): New functions.
1727 (ffeglobalArgSummary, ffeglobalArgInfo_): New types.
1729 Tue Apr 29 18:35:41 1997 Craig Burley <burley@gnu.ai.mit.edu>
1731 More changes to support better tracking of (filewide)
1733 * expr.c (ffeexpr_fulfill_call_): New function.
1734 (ffeexpr_token_name_lhs_): Call after building procedure
1735 reference expression. Also leave info field for ANY-ized
1737 (ffeexpr_token_arguments_): Ditto.
1739 Mon Apr 28 20:04:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
1741 Changes to support better tracking of (filewide)
1742 globals, mainly to avoid crashes due to inlining:
1743 * bad.def: Go back to quoting intrinsic names,
1744 (FFEBAD_FILEWIDE_DISAGREEMENT, FFEBAD_FILEWIDE_TIFF,
1745 FFEBAD_FILEWIDE_TYPE_MISMATCH): New diagnostics.
1746 (FFEBAD_INTRINSIC_EXPIMP, FFEBAD_INTRINSIC_GLOBAL): Reword
1748 * com.c (ffecom_do_entry_, ffecom_start_progunit_,
1749 ffecom_sym_transform_): Accommodate new FFEGLOBAL_typeEXT
1751 * expr.c (ffeexpr_sym_lhs_call_, ffeexpr_sym_lhs_extfunc_,
1752 ffeexpr_sym_rhs_actualarg_, ffeexpr_declare_parenthesized_,
1753 ffeexpr_paren_rhs_let_, ffeexpr_token_funsubstr_):
1754 Fill in real kind info instead of leaving NONE where
1756 Register references to intrinsics and globals with ffesymbol
1757 using new ffesymbol_reference function instead of
1758 ffesymbol_globalize.
1759 * global.c (ffeglobal_type_string_): New array for
1761 * global.h, global.c:
1762 Replace ->init mechanism with ->tick mechanism.
1763 Move other common-related members into a substructure of
1764 a union, so the proc substructure can be introduced
1765 to include members related to externals other than commons.
1766 Don't complain about ANY-ized globals; ANY-ize globals
1767 once they're complained about, in any case where code
1768 generation could become a problem.
1769 Handle global entries that have NONE type (seen as
1770 intrinsics), EXT type (seen as EXTERNAL), and so on.
1771 Keep track of kind and type of externals, both via
1772 definition and via reference.
1773 Diagnose disagreements about kind or type of externals
1774 (such as functions).
1775 (ffeglobal_ref_intrinsic, ffeglobal_ref_progunit_): New
1777 * stc.c (ffestc_R1207_item, ffestc_R1208_item,
1778 ffestc_R1219, ffestc_R1226):
1779 Call ffesymbol_reference, not ffesymbol_globalize.
1780 * stu.c (ffestu_sym_end_transition,
1781 ffestu_sym_exec_transition):
1782 Call ffesymbol_reference, not ffesymbol_globalize.
1783 * symbol.c (ffesymbol_globalize): Removed...
1784 (ffesymbol_reference): ...to this new function,
1785 which more generally registers references to symbols,
1786 globalizes globals, and calls on the ffeglobal module
1787 to check globals filewide.
1789 * global.h, global.c: Rename some macros and functions
1790 to more clearly distinguish common from other globals.
1791 All callers changed.
1793 * com.c (ffecom_sym_transform_): Trees describing
1794 filewide globals must be allocated on permanent obstack.
1796 * expr.c (ffeexpr_token_name_lhs_): Don't generate
1797 gratuitous diagnostics for FFEINFO_whereANY case.
1799 Thu Apr 17 03:27:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
1801 * global.c: Add support for flagging intrinsic/global
1802 confusion via warnings.
1803 * bad.def (FFEBAD_INTRINSIC_EXPIMP,
1804 FFEBAD_INTRINSIC_GLOBAL): New diagnostics.
1805 * expr.c (ffeexpr_token_funsubstr_): Ditto.
1806 (ffeexpr_sym_lhs_call_): Ditto.
1807 (ffeexpr_paren_rhs_let_): Ditto.
1808 * stc.c (ffestc_R1208_item): Ditto.
1810 Wed Apr 16 22:40:56 1997 Craig Burley <burley@gnu.ai.mit.edu>
1812 * expr.c (ffeexpr_declare_parenthesized_): INCLUDE
1813 context can't be an intrinsic invocation either.
1815 Fri Mar 28 10:43:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
1817 * expr.c (ffeexpr_token_arguments_): Make sure top of
1818 exprstack is operand before dereferencing operand field.
1820 * lex.c (ffelex_prepare_eos_): Fill up truncated
1821 hollerith token, so crash on null ->text field doesn't
1824 * stb.c (ffestb_R10014_): If NAMES isn't recognized (or
1825 the recognized part is followed in the token by a
1826 non-digit), don't try and collect digits, as there
1827 might be more than FFEWHERE_indexMAX letters to skip
1828 past to do so -- and the code is diagnosed anyway.
1830 Thu Mar 27 00:02:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
1832 * com.c (ffecom_sym_transform_): Force local
1833 adjustable array onto stack.
1835 * stc.c (ffestc_R547_item_object): Don't actually put
1836 the symbol in COMMON if the symbol has already been
1837 EQUIVALENCE'd to a different COMMON area.
1839 * equiv.c (ffeequiv_add): Don't actually do anything
1840 if there's a disagreement over which COMMON area is
1843 Tue Mar 25 03:35:19 1997 Craig Burley <burley@gnu.ai.mit.edu>
1845 * com.c (ffecom_transform_common_): If no explicit init
1846 of COMMON area, don't actually init it even though
1847 storage area suggests it.
1849 Mon Mar 24 12:10:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1851 * lex.c (ffelex_image_char_): Avoid overflowing the
1852 column counter itself, as well as the card image.
1854 * where.c (ffewhere_line_new): Cast ffelex_line_length()
1855 to (size_t) so 255 doesn't overflow to 0!
1857 * stc.c (ffestc_labeldef_notloop_begin_): Don't gratuitously
1858 terminate loop before processing statement, so block
1859 doesn't disappear out from under EXIT/CYCLE processing.
1860 (ffestc_labeldef_notloop_): Has old code from above
1861 function, instead of just calling it.
1863 * expr.c (ffeexpr_cb_comma_i_4_): Don't skip over
1864 arbitrary token (such as EOS).
1866 * com.c (ffecom_init_zero_): Handle RECORD_TYPE and
1867 UNION_TYPE so -fno-zeros works with -femulated-complex.
1869 1997-03-12 Dave Love <d.love@dl.ac.uk>
1871 * intrin.def: New intrinsics INT2, INT8, CPU_TIME. Fix AND, OR,
1872 XOR. [Integrated by burley, AND/OR/XOR already fixed, INT8
1873 implementation changed/fixed.]
1875 Wed Mar 12 10:40:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1877 * Make-lang.in ($(srcdir)/f/intdoc.texi): Simplify rules
1878 so building f/intdoc is not always necessary; remove
1879 f/intdoc after running it if it is built.
1881 Tue Mar 11 23:42:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
1883 * intrin.def (FFEINTRIN_impAND, FFEINTRIN_impOR,
1884 FFEINTRIN_impXOR): Use the IAND, IOR, and IEOR implementations
1885 of these, instead of crashing in ffecom_expr_intrinsic_
1886 or adding case labels there.
1888 Mon Mar 10 22:51:23 1997 Craig Burley <burley@gnu.ai.mit.edu>
1890 * intdoc.c: Fix so any C compiler can compile this.
1892 Fri Feb 28 13:16:50 1997 Craig Burley <burley@gnu.ai.mit.edu>
1894 * Version 0.5.20 released.
1896 Fri Feb 28 01:45:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
1898 * Make-lang.in (RUNTIMESTAGESTUFF, LIBU77STAGESTUFF):
1899 Move some files incorrectly in the former to the latter,
1900 and add another file or two to the latter.
1902 New meanings for (KIND=n), and new denotations in the
1903 little language describing intrinsics:
1904 * com.c (ffecom_init_0): Assign new meanings.
1905 * intdoc.c: Document new meanings.
1906 Support the new denotations.
1907 * intrin.c: Employ new meanings, mapping them to internal
1908 values (which are the same as they ever were for now).
1909 Support the new denotations.
1910 * intrin.def: Switch DEFIMP table to the new denotations.
1912 * intrin.c (ffeintrin_check_): Fix bug that was leaving
1913 LOC() and %LOC() returning INTEGER*4 on systems where
1914 it should return INTEGER*8.
1916 * type.c: Canonicalize function definitions, for etags
1919 Wed Feb 26 20:43:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
1921 * com.c (ffecom_init_0): Choose INTEGER(KIND=n) types,
1922 where n is 2, 3, and 4, according to the new docs
1923 instead of according to the old C correspondences
1924 (which seem less useful at this point).
1926 * equiv.c (ffeequiv_destroy_): New function.
1927 (ffeequiv_layout_local_): Use this new function
1928 whenever the laying out of a local equivalence chain
1929 is aborted for any reason.
1930 Otherwise ensure that symbols no longer reference
1931 the stale ffeequiv entries that result when they
1932 are killed off in this procedure.
1933 Also, the rooted symbol is one that has storage,
1934 it really is irrelevant whether it has an equiv entry
1935 at this point (though the code to remove the equiv
1936 entry was put in at the end, just in case).
1937 (ffeequiv_kill): When doing internal checks, make
1938 sure the victim isn't named by any symbols it points
1939 to. Not as complete a check as looking through the
1940 entire symbol table (which does matter, since some
1941 code in equiv.c used to remove symbols from the lists
1942 for an ffeequiv victim but not remove that victim as the
1943 symbol's equiv info), but this check did find some
1944 real bugs in the code (that were fixed).
1946 Mon Feb 24 16:42:13 1997 Craig Burley <burley@gnu.ai.mit.edu>
1948 * com.c (ffecom_expr_intrinsic_): Fix a couple of
1949 warnings about uninitialized variables.
1950 * intrin.c (ffeintrin_check_): Ditto, but there were
1951 a couple of _real_ uninitialized-variable _bugs_ here!
1952 (ffeintrin_fulfill_specific): Ditto, no real bug here.
1954 Sun Feb 23 15:01:20 1997 Craig Burley <burley@gnu.ai.mit.edu>
1956 Clean up diagnostics (especially about intrinsics):
1957 * bad.def (FFEBAD_UNIMPL_STMT): Remove.
1958 (FFEBAD_INTRINSIC_*, FFEBAD_NEED_INTRINSIC): Clean these
1959 up so they're friendlier.
1960 (FFEBAD_INTRINSIC_CMPAMBIG): New.
1961 * intrin.c (ffeintrin_fulfill_generic,
1962 ffeintrin_fulfill_specific, ffeintrin_is_intrinsic):
1964 generic or specific name text (which is for doc purposes
1965 anyway) over implementation name text (which is for
1967 * intrin.def: Use more descriptive name texts for generics
1968 and specifics in cases where the names themselves are not
1969 enough (e.g. IDATE, which has two forms).
1971 Fix some intrinsic mappings:
1972 * intrin.def (FFEINTRIN_specIDINT, FFEINTRIN_specAND,
1973 FFEINTRIN_specDFLOAT, FFEINTRIN_specDREAL, FFEINTRIN_specOR,
1974 FFEINTRIN_specXOR): Now have their own implementations,
1975 instead of borrowing from others.
1976 (FFEINTRIN_specAJMAX0, FFEINTRIN_specAJMIN0, FFEINTRIN_specBJTEST,
1977 FFEINTRIN_specDFLOTJ, FFEINTRIN_specFLOATJ, FFEINTRIN_specJIABS,
1978 FFEINTRIN_specJIAND, FFEINTRIN_specJIBCLR, FFEINTRIN_specJIBITS,
1979 FFEINTRIN_specJIBSET, FFEINTRIN_specJIDIM, FFEINTRIN_specJIDINT,
1980 FFEINTRIN_specJIDNNT, FFEINTRIN_specJIEOR, FFEINTRIN_specJIFIX,
1981 FFEINTRIN_specJINT, FFEINTRIN_specJIOR, FFEINTRIN_specJISHFT,
1982 FFEINTRIN_specJISHFTC, FFEINTRIN_specJISIN, FFEINTRIN_specJMAX0,
1983 FFEINTRIN_specJMAX1, FFEINTRIN_specJMIN0, FFEINTRIN_specJMIN1,
1984 FFEINTRIN_specJMOD, FFEINTRIN_specJNINT, FFEINTRIN_specJNOT,):
1985 Turn these implementations off, since it's not clear
1986 just what types they expect in the context of portable Fortran.
1987 (DFLOAT): Now in FVZ family, since f2c supports them
1989 Support intrinsic inquiry functions (BIT_SIZE, LEN):
1990 * intrin.c: Allow `i' in <arg_extra>.
1991 * intrin.def (FFEINTRIN_impBIT_SIZE, FFEINTRIN_impLEN):
1994 Sat Feb 22 13:34:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
1996 Only warn, don't error, for reference to unimplemented
1998 * bad.def (FFEBAD_INTRINSIC_UNIMPLW): Warning version
2000 * intrin.c (ffeintrin_is_intrinsic): Use new warning
2001 version of _UNIMPL (FFEBAD_INTRINSIC_UNIMPLW).
2003 Complain about REAL(Z) and AIMAG(Z) (Z is DOUBLE COMPLEX):
2004 * bad.def (FFEBAD_INTRINSIC_CMPAMBIG): New diagnostic.
2005 * expr.c: Needed #include "intrin.h" anyway.
2006 (ffeexpr_token_intrincheck_): New function handles delayed
2007 diagnostic for "REAL(REAL(expr)" if next token isn't ")".
2008 (ffeexpr_token_arguments_): Do most of the actual checking here.
2009 * intrin.h, intrin.c (ffeintrin_fulfill_specific): New
2010 argument, check_intrin, to tell caller that intrin is REAL(Z)
2011 or AIMAG(Z). All callers updated, mostly to pass NULL in
2013 (ffeintrin_check_): Also has new arg check_intrin for same
2014 purpose. All callers updated the same way.
2015 * intrin.def (FFEINTRIN_impAIMAG): Change return type
2016 from "R0" to "RC", to accommodate f2c (and perhaps other
2017 non-F90 F77 compilers).
2018 * top.h, top.c: New option -fugly-complex.
2020 New GNU intrinsics REALPART, IMAGPART, and COMPLEX:
2021 * com.c (ffecom_expr_intrinsic_): Implement impCOMPLEX
2022 and impREALPART here. (specIMAGPART => specAIMAG.)
2023 * intrin.def: Add the intrinsics here.
2025 Rename implementations of VXTIDATE and VXTTIME to IDATEVXT
2026 and TIMEVXT, so they sort more consistently:
2027 * com.c (ffecom_expr_intrinsic_):
2030 Delete intrinsic group `dcp', add `gnu', etc.:
2031 * intrin.c (ffeintrin_state_family): FFEINTRIN_familyGNU
2032 replaces FFEINTRIN_familyDCP, and gets state from `gnu'
2034 Get rid of FFEINTRIN_familyF2Z, nobody needs it.
2035 Move FFEINTRIN_specDCMPLX from DCP family to FVZ family,
2037 Move FFEINTRIN_specDFLOAT from F2C family to FVZ family.
2038 (FFEINTRIN_specZABS, FFEINTRIN_specZCOS, FFEINTRIN_specZEXP,
2039 FFEINTRIN_specZLOG, FFEINTRIN_specZSIN, FFEINTRIN_specZSQRT):
2040 Move these from F2Z family to F2C family.
2041 * intrin.h (FFEINTRIN_familyF2Z, FFEINTRIN_familyDCP): Remove.
2042 (FFEINTRIN_familyGNU): Add.
2043 * top.h, top.c: Replace `dcp' with `gnu'.
2045 * com.c (ffecom_expr_intrinsic_): Clean up by collecting
2046 simple conversions into one nice, conceptual place.
2047 Fix up some intrinsic subroutines (MVBITS, KILL, UMASK) to
2048 properly push and pop call temps, to avoid wasting temp
2051 * g77.c (doit): Toon says variables should be defined
2052 before being referenced. Spoilsport.
2054 * intrin.c (ffeintrin_check_): Now Dave's worried about
2055 warnings about uninitialized variables. Okay, so for
2056 basic return values 'g' and 's', they _were_
2057 uninitialized -- is determinism really _that_ useful?
2059 * intrin.def (FFEINTRIN_impFGETC): Fix STATUS argument
2060 so that it is INTENT(OUT) instead of INTENT(IN).
2062 1997-02-21 Dave Love <d.love@dl.ac.uk>
2064 * intrin.def, com.c: Support Sun-type `short' and `long'
2065 intrinsics. Perhaps should also do Microcruft-style `int2'.
2067 Thu Feb 20 15:16:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
2069 * com.c (ffecom_expr_intrinsic_): Clean up indentation.
2070 Support SECONDSUBR intrinsic implementation.
2071 Rename SECOND to SECONDFUNC for direct support via library.
2073 * g77.c: Fix to return proper status value to shell,
2074 by obtaining it from processes it spawns.
2076 * intdoc.c: Fix minor typo.
2078 * intrin.def: Turn SECOND into generic that maps into
2079 function and subroutine forms.
2081 * intrin.def: Make FLOAT and SNGL into specific intrinsics.
2083 * intrin.def, intrin.h: Change the way DEFGEN and DEFSPEC
2084 macros work, to save on verbage.
2086 Mon Feb 17 02:08:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
2088 New subsystem to automatically generate documentation
2090 * Make-lang.in ($(srcdir)/f/g77.info,
2091 $(srcdir)/f/g77.dvi): Move g77 doc rules around.
2092 Add to g77 doc rules the new subsystem.
2093 (f77.mostlyclean, f77.maintainer-clean): Also clean up
2094 after new doc subsystem.
2095 * intdoc.c, intdoc.h: New doc subsystem code.
2096 * intrin.h [FFEINTRIN_DOC]: When 1, don't pull in
2097 stuff not needed by doc subsystem.
2099 Improve on intrinsics mechanism to both be more
2100 self-documenting and to catch more user errors:
2101 * intrin.c (ffeintrin_check_): Recognize new arg-len
2102 and arg-rank information, and check it.
2103 Move goto and signal indicators to the basic type.
2104 Permit reference to arbitrary argument number, not
2105 just first argument (for BESJN and BESYN).
2106 (ffeintrin_init_0): Check and accept new notations.
2107 * intrin.c, intrin.def: Value in COL now identifies
2108 arguments starting with number 0 being the first.
2110 Some minor intrinsics cleanups (resulting from doc work):
2111 * com.c (ffecom_expr_intrinsic_): Implement FLUSH
2112 directly once again, handle its optional argument,
2113 so it need not be a generic (awkward to handle in docs).
2114 * intrin.def (BESJ0, BESJ1, BESJN, BESY0, BESY1, BESYN,
2115 CHDIR, CHMOD, CTIME, DBESJ0, DBESJ1, DBESJN, DBESY0,
2116 DBESY1, DBESYN, DDIM, ETIME, FGETC, FNUM, FPUTC, FSTAT,
2117 GERROR, GETCWD, GETGID, GETLOG, GETPID, GETUID, GMTIME,
2118 HOSTNM, IDATE, IERRNO, IIDINT, IRAND, ISATTY, ITIME, JIDINT,
2119 LNBLNK, LSTAT, LTIME, MCLOCK, PERROR, SRAND, SYMLNK, TTYNAM,
2120 UMASK): Change capitalization of initcaps (official) name
2121 to be consistent with Burley's somewhat arbitrary rules.
2122 (BESJN, BESYN): These have return arguments of same type
2123 as their _second_ argument.
2124 (FLUSH): Now a specific, not generic, intrinsic, with one
2126 (FLUSH1): Eliminated.
2127 Add arg-len and arg-rank info to several intrinsics.
2128 (ITIME): Change argument type from REAL to INTEGER.
2130 Tue Feb 11 14:04:42 1997 Craig Burley <burley@gnu.ai.mit.edu>
2132 * Make-lang.in (f771): Invocation of Makefile now done
2133 with $(srcdir)=gcc to go along with $(VPATH)=gcc.
2134 ($(srcdir)/f/runtime/configure,
2135 $(srcdir)/f/runtime/libU77/configure): Break these out
2136 so spurious triggers of this rule don't happen (as when
2137 configure.in is more recent than libU77/configure).
2138 (f77.rebuilt): Distinguish source versus build files,
2139 so this target can be invoked from build directory and
2141 * Makefile.in: This now expects $(srcdir) to be the gcc
2142 source directory, not gcc/f, to agree with $(VPATH).
2143 Accordingly, $(INCLUDES) has been fixed, various cruft
2144 removed, the removal of f771 has been fixed to remove
2145 the _real_ f771 (not the one in gcc's parent directory),
2148 * lex.c: Part of ffelex_finish_statement_() now done
2149 by new function ffelex_prepare_eos_(), so that, in one
2150 popular case, the EOS can be prepared while the pointer
2151 is at the end of the non-continued line instead of the
2152 end of the line that marks no continuation. This improves
2153 the appearance of diagnostics substantially.
2155 Mon Feb 10 12:44:06 1997 Craig Burley <burley@gnu.ai.mit.edu>
2157 * Make-lang.in: runtime Makefile's, and include/f2c.h,
2158 also depend on f/runtime/configure and f/runtime/libU77/configure.
2160 Fix various libU77 routines:
2161 * com-rt.def (FFECOM_gfrtCTIME, FFECOM_gfrtMCLOCK,
2162 FFECOM_gfrtTIME): These now use INTEGER*8 for time values,
2163 for compatibility with systems like Alpha.
2164 (FFECOM_gfrtSYSTEM_CLOCK, FFECOM_gfrtTTYNAM): Delete incorrect
2165 trailing underscore in routine names.
2166 * intrin.c, intrin.def: Support INTEGER*8 return values and
2167 arguments ('4'). Change FFEINTRIN_impCTIME, FFEINTRIN_impMCLOCK,
2168 and FFEINTRIN_impTIME accordingly.
2169 (ffeintrin_is_intrinsic): Don't give caller a clue about
2170 form of intrinsic -- shouldn't be needed at this point.
2172 Cope with generic intrinsics that are subroutines and functions:
2173 * com.c (ffecom_finish_symbol_transform_, ffecom_expr_transform_):
2174 Don't transform an intrinsic that is not known to be a subroutine
2175 or a function. (Maybe someday have to avoid transforming
2176 any intrinsic with an undecided or unknown implementation.)
2177 * expr.c (ffeexpr_declare_unadorned_,
2178 ffeexpr_declare_parenthesized_): Ok to invoke generic
2179 intrinsic that has at least one subroutine form as a
2181 Ok to pass intrinsic as actual arg if it has a known specific
2182 intrinsic form that is valid as actual arg.
2183 (ffeexpr_declare_parenthesized_): An unknown kind of
2184 intrinsic has a paren_type chosen based on context.
2185 (ffeexpr_token_arguments_): Build funcref/subrref based
2186 on context, not on kind of procedure being called.
2187 * intrin.h, intrin.c (ffeintrin_is_intrinsic): Undo changes of
2188 Tue Feb 4 23:12:04 1997 by me, change all callers to leave
2189 intrinsics as FFEINFO_kindNONE at this point. (Some callers
2190 also had unused variables deleted as a result.)
2192 Enable all intrinsic groups (especially f90 and vxt):
2193 * target.h (FFETARGET_defaultSTATE_DCP, FFETARGET_defaultSTATE_F2C,
2194 FFETARGET_defaultSTATE_F90, FFETARGET_defaultSTATE_MIL,
2195 FFETARGET_defaultSTATE_UNIX, FFETARGET_defaultSTATE_VXT):
2196 Delete these macros, let top.c set them directly.
2197 * top.c (ffeintrinsic_state_dcp_, ffe_intrinsic_state_f2c_,
2198 ffe_intrinsic_state_f90_, ffe_intrinsic_state_mil_,
2199 ffe_intrinsic_state_unix_, ffe_intrinsic_state_vxt_):
2200 Enable all these directly.
2202 Sat Feb 8 03:21:50 1997 Craig Burley <burley@gnu.ai.mit.edu>
2204 * g77.c: Incorporate recent changes to ../gcc.c.
2205 For version magic (e.g. `g77 -v'), instead of compiling
2206 /dev/null, write, compile, run, and then delete a small
2207 program that prints the version numbers of the three
2208 components of libf2c (libF77, libI77, and libU77),
2209 so we get this info with bug reports.
2210 Also, this change reduces the chances of accidentally
2211 linking to an old (complex-alias-problem) libf2c.
2212 Fix `-L' so the argument is expected in `-Larg'.
2214 * com.h (FFECOM_f2cLONGINT): For INTEGER*8 support in f2c.h,
2215 dynamically determine proper type here, instead of
2216 assuming `long long int' is correct.
2218 Tue Feb 4 23:12:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
2220 Add libU77 library from Dave Love <d.love@dl.ac.uk>:
2221 * Make-lang.in (f77-runtime): Depend on new Makefile.
2222 (f/runtime/libU77/Makefile): New rule.
2223 Also configure libU77.
2224 ($(srcdir)/f/runtime/configure: Use Makefile.in,
2225 so configuration doesn't have to have happened.
2226 (f77.mostlyclean, f77.clean, f77.distclean,
2227 f77.maintainer-clean): Some fixups here, but more work
2229 (RUNTIMESTAGESTUFF): Add libU77's config.status.
2230 (LIBU77STAGESTUFF, f77.stage1, f77.stage2, f77.stage3,
2231 f77.stage4): New macro, appropriate uses added.
2232 * com-rt.def: Add libU77 procedures.
2233 * com.c (ffecom_f2c_ptr_to_integer_type_node,
2234 ffecom_f2c_ptr_to_real_type_node): New type nodes.
2235 (FFECOM_rttypeCHARACTER_): New type of run-time function.
2236 (ffecom_char_args_): Handle CHARACTER*n intrinsics
2237 where n != 1 here, instead of in ffecom_expr_intrinsic_.
2238 (ffecom_expr_intrinsic_): New code to handle new
2240 In particular, change how FFEINTRIN_impFLUSH is handled.
2241 (ffecom_make_gfrt_): Handle new type of run-time function.
2242 (ffecom_init_0): Initialize new type nodes.
2243 * config-lang.in: New libU77 directory.
2244 * intrin.h, intrin.c (ffeintrin_is_intrinsic): Handle
2245 potential generic for subroutine _and_ function
2246 specifics via two new arguments. All callers changed.
2247 Properly ignore deleted/disabled intrinsics in resolving
2249 (ffeintrin_check_, ffeintrin_init_0): Handle CHARACTER intrinsics of (*)
2251 * intrin.def: Permission granted by FSF to place this in
2252 public domain, which will allow it to serve as source
2253 for both g77 program and its documentation.
2254 Add libU77 intrinsics.
2255 (FLUSH): Now a generic, not specific, intrinsic.
2256 (DEFIMP): Now support return modifier for CHARACTER intrinsics.
2258 * com-rt.def (FFECOM_gfrtDIM, FFECOM_gfrtERF,
2259 FFECOM_gfrtERFC, FFECOM_gfrtEXP, FFECOM_gfrtSIGN,
2260 FFECOM_gfrtSIN, FFECOM_gfrtSINH, FFECOM_gfrtTAN,
2261 FFECOM_gfrtTANH, FFECOM_gfrtPOW_RI): Change "&r" to "&f".
2263 Sat Feb 1 12:15:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
2265 * Version 0.5.19.1 released.
2267 * com.c (ffecom_expr_, ffecom_expr_intrinsic_,
2268 ffecom_tree_divide_): FFECOM_gfrtPOW_ZI,
2269 FFECOM_gfrtCONJG, FFECOM_gfrtDCONJG,
2270 FFECOM_gfrtCCOS, FFECOM_gfrtCDCOS,
2271 FFECOM_gfrtCLOG, FFECOM_gfrtCDLOG,
2272 FFECOM_gfrtCSIN, FFECOM_gfrtCDSIN,
2273 FFECOM_gfrtCSQRT, FFECOM_gfrtCDSQRT,
2274 FFECOM_gfrtDIV_CC, FFECOM_gfrtDIV_ZZ: These all require
2275 result to _not_ overlap one or more inputs.
2277 Sat Feb 1 00:25:55 1997 Craig Burley <burley@gnu.ai.mit.edu>
2279 * com.c (ffecom_init_0): Do internal checks only if
2280 -fset-g77-defaults not specified.
2282 Fix %LOC(), LOC() to return sufficiently wide type:
2283 * com.h, com.c (ffecom_pointer_kind_, ffecom_label_kind_,
2284 ffecom_pointer_kind(), ffecom_label_kind()): New globals
2285 and accessor macros hold kind for integer pointers on target
2287 (ffecom_init_0): Determine narrowest INTEGER type that
2288 can hold a pointer (usually INTEGER*4 or INTEGER*8),
2289 store it in ffecom_pointer_kind_, etc.
2290 * expr.c (ffeexpr_cb_end_loc_): Use right type for %LOC().
2291 * intrin.c (ffeintrin_check_, ffeintrin_init_0): Support
2292 new 'p' kind for type of intrinsic.
2293 * intrin.def (FFEINTRIN_impLOC): Returns "Ip" instead of "I1",
2294 so LOC() type is correct for target machine.
2296 Support -fugly-assign:
2297 * lang-options.h, top.h, top.c (ffe_decode_option):
2298 Accept -fugly-assign and -fno-ugly-assign.
2299 * com.c (ffecom_expr_): Handle -fugly-assign.
2300 * expr.c (ffeexpr_finished_): Check right type for ASSIGN
2303 Fri Jan 31 14:30:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
2305 Remove last vestiges of -fvxt-not-f90:
2306 * stb.c (ffestb_R10012_, ffestb_R10014_, ffestb_V0201_):
2309 Fri Jan 31 02:13:54 1997 Craig Burley <burley@gnu.ai.mit.edu>
2311 * top.c (ffe_decode_option): Warn if -fugly is specified,
2314 * symbol.h: No need to #include "bad.h".
2316 Reorganize features from -fvxt-not-f90 to -fvxt:
2317 * lang-options.h, top.h, top.c:
2318 Accept -fvxt and -fno-vxt, but not -fvxt-not-f90 or -ff90-not-vxt.
2319 Warn if the latter two are used.
2320 * expr.c (ffeexpr_nil_rhs_): Double-quote means octal constant.
2321 (ffeexpr_token_rhs_): Double-quote means octal constant.
2322 * target.h (FFETARGET_defaultIS_VXT_NOT_90): Delete macro
2323 definition, no longer needed.
2325 Make some -ff90 features the default:
2326 * data.c (ffedata_value): DATA implies SAVE.
2327 * src.h (ffesrc_is_name_noninit): Underscores always okay.
2329 Fix up some more #error directives by quoting their text:
2330 * bld.c (ffebld_constant_is_zero):
2333 Sat Jan 18 18:22:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
2335 * g77.c (lookup_option, main): Recognize `-Xlinker',
2336 `-Wl,', `-l', `-L', `--library-directory', `-o',
2338 (lookup_option): Don't depend on SWITCH_TAKES_ARG
2339 being correct, it might or might not have `-x' in
2340 it depending on host.
2341 Return NULL argument if it would be an empty string.
2342 (main): If no input files (by gcc.c's definition)
2343 but `-o' or `--output' specified, produce diagnostic
2344 to avoid overwriting output via gcc.
2345 Recognize C++ `+e' options.
2346 Treat -L as another non-magical option (like -B).
2347 Don't append_arg `-x' twice.
2349 Fri Jan 10 23:36:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
2351 * top.c [BUILT_FOR_270] (ffe_decode_option): Make
2352 -fargument-noalias-global the default.
2354 Fri Jan 10 07:42:27 1997 Craig Burley <burley@gnu.ai.mit.edu>
2356 Enable inlining of previously-compiled program units:
2357 * com.c (ffecom_do_entry_, ffecom_start_progunit_):
2358 Register new public function in ffeglobal database.
2359 (ffecom_sym_transform_): Any GLOBAL or potentially GLOBAL
2360 symbol should be looked up in ffeglobal database and
2361 that tree node used, if found. That way, gcc knows
2362 the references are to those earlier definitions, so it
2363 can emit shorter branches/calls, inline, etc.
2364 (ffecom_transform_common_): Minor change for clarity.
2365 * expr.c (ffeexpr_sym-lhs_call_, ffeexpr_sym_lhs_extfunc_,
2366 ffeexpr_sym_rhs_actualarg_, ffeexpr_paren_rhs_let_,
2367 ffeexpr_token_funsubstr_): Globalize symbol as needed.
2368 * global.c (ffeglobal_promoted): New function to look up
2369 existing local symbol in ffeglobal database.
2370 * global.h: Declare new function.
2371 * name.h (ffename_token): New macro, plus alphabetize.
2372 * stc.c (ffestc_R1207_item): Globalize EXTERNAL symbol.
2373 * stu.c (ffestu_sym_end_transition, ffestu_sym_exec_transition):
2374 Globalize symbol as needed.
2375 * symbol.h, symbol.c (ffesymbol_globalize): New function.
2377 Thu Jan 9 14:20:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
2379 * ste.c (ffeste_R809): Produce a diagnostic for SELECT CASE
2380 on CHARACTER type, instead of crashing.
2382 Thu Jan 9 00:52:45 1997 Craig Burley <burley@gnu.ai.mit.edu>
2384 * stc.c (ffestc_order_entry_, ffestc_order_format_,
2385 ffestc_R1226): Allow ENTRY and FORMAT before IMPLICIT
2386 NONE, by having them transition only to state 1 instead
2387 of state 2 (which is disallowed by IMPLICIT NONE).
2389 Mon Jan 6 22:44:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
2391 Fix AXP bug found by Rick Niles (961201-1.f):
2392 * com.c (ffecom_init_0): Undo my 1996-05-14 change, as
2393 it is incorrect and prevented easily finding this bug.
2394 * target.h [__alpha__] (ffetargetReal1, ffetargetReal2):
2395 Use int instead of long.
2396 (ffetarget_cvt_r1_to_rv_, ffetarget_cvt_rv_to_r1_,
2397 ffetarget_cvt_r2_to_rv_, ffetarget_cvt_rv_to_r2_):
2398 New functions that intercede for callers of
2399 REAL_VALUE_(TO|UNTO)_TARGET_(SINGLE|DOUBLE).
2400 All callers changed, and damaging casts to (long *) removed.
2402 Sun Jan 5 03:26:11 1997 Craig Burley <burley@gnu.ai.mit.edu>
2404 * Make-lang.in (g77, g77-cross): Depend on both g77.c and
2405 zzz.c, in $(srcdir)/f/.
2407 Better design for -fugly-assumed:
2408 * stc.c (ffestc_R501_item, ffestc_R524_item,
2409 ffestc_R547_item_object): Pass new is_ugly_assumed flag.
2410 * stt.c, stt.h (ffestt_dimlist_as_expr,
2411 ffestt_dimlist_type): New is_ugly_assumed flag now
2412 controls whether "1" is treated as "*".
2413 Don't treat "2-1" or other collapsed constants as "*".
2415 Sat Jan 4 15:26:22 1997 Craig Burley <burley@gnu.ai.mit.edu>
2417 * stb.c (ffestb_R10012_): Don't confirm on FORMAT(A,)
2418 or even FORMAT(A,,B), as R1229 only warns about the
2419 former currently, and this seems reasonable.
2421 Improvements to diagnostics:
2422 * sta.c (ffesta_second_): Don't add any ffestb parsers
2423 unless they're specifically called for.
2424 Set up ffesta_tokens[0] before calling ffestc_exec_transition,
2425 else stale info might get used.
2426 (ffesta_save_): Do a better job picking which parser to run
2427 after running all parsers with no confirmed possibles.
2428 (FFESTA_maxPOSSIBLES_): Decrease from 100 now that so few
2429 possibles are ever on the list at a given time.
2430 (struct _ffesta_possible): Add named attribute.
2431 (ffesta_add_possible_exec_, ffesta_add_possible_nonexec_):
2432 Make these into macros that call a single function that now
2433 sets the named attribute.
2434 (ffesta_add_possible_unnamed_exec_,
2435 ffeseta_add_possible_unnamed_nonexec_): New macros.
2436 (ffesta_second_): Designate unnamed possibles as
2438 * stb.c (ffestb_R1229, ffestb_R12291_): Use more general
2439 diagnostic, so things like "POINTER (FOO, BAR)" are
2440 diagnosed as unrecognized statements, not invalid statement
2442 * stb.h, stb.c (ffestb_unimplemented): Remove function.
2444 1996-12-30 Dave Love <d.love@dl.ac.uk>
2446 * com.c: #include libU77/config.h
2447 (ffecom_f2c_ptr_to_integer_type_node,
2448 ffecom_f2c_ptr_to_integer_type_node): New variables.
2449 (ffecom_init_0): Use them.
2450 (ffecom_expr_intrinsic_): Many news cases for libU77 intrinsics.
2452 * com-rt.def: New definitions for libU77.
2453 * intrin.def: Likewise. Also correct ftell arg spec.
2455 * Makefile.in (f/runtime/libU77/config.h): New target for com.c
2457 * Make-lang.in (f771): Depend on f/runtime/Makefile for the above.
2459 Sat Dec 28 12:28:29 1996 Craig Burley <burley@gnu.ai.mit.edu>
2461 * stt.c (ffestt_dimlist_type): Treat ([...,]1) in dimlist
2462 as ([...,]*) if -fugly-assumed, so assumed-size array
2463 detected early enough.
2465 Thu Dec 19 14:01:57 1996 Craig Burley <burley@gnu.ai.mit.edu>
2467 * target.h (FFETARGET_REAL_VALUE_FROM_INT_): Conditionalize
2468 definition on BUILT_FOR_280, not BUILT_WITH_280, since
2469 the name of the macro was (properly) changed since 0.5.19.
2471 Fix warnings/errors resulting from ffetargetOffset becoming
2472 `long long int' instead of `unsigned long' as of 0.5.19,
2473 while ffebitCount remains `unsigned long':
2474 * bld.c (ffebld_constantarray_dump): Avoid warnings by
2475 using loop var of appropriate type, and using casts.
2476 * com.c (ffecom_expr_): Use right type for loop var.
2477 (ffecom_sym_transform_, ffecom_transform_equiv_):
2478 Cast to right type in assertions.
2479 * data.c (ffedata_gather_, ffedata_value_): Cast to right
2480 type in assertions and comparisons.
2482 Wed Dec 18 12:07:11 1996 Craig Burley <burley@gnu.ai.mit.edu>
2484 Patch from Alexandre Oliva <oliva@dcc.unicamp.br>:
2485 * Makefile.in (all.indirect): Don't pass -bbigtoc option
2488 Cope with new versions of gcc:
2489 * com.h (BUILT_FOR_280): New macro.
2490 * com.c (ffecom_ptr_to_expr): Conditionalize test of
2492 (ffecom_build_complex_constant_): Conditionalize calling
2493 sequence for build_complex.
2495 Sat Dec 7 07:15:17 1996 Craig Burley <burley@gnu.ai.mit.edu>
2497 * Version 0.5.19 released.
2499 Fri Dec 6 12:23:55 1996 Craig Burley <burley@gnu.ai.mit.edu>
2501 * g77.c: Default to assuming "f77" is in $LANGUAGES, since
2502 the LANGUAGE_F77 macro isn't defined by anyone anymore (but
2503 might as well leave the no-f77 code in just in case).
2504 * Make-lang.in (g77, g77-cross): Don't define LANGUAGE_F77
2507 1996-12-06 Dave Love <d.love@dl.ac.uk>
2509 * Make-lang.in (g77, g77-cross): Revert to building `g77' or not
2510 conditional on `f77' in LANGUAGES.
2512 Wed Dec 4 13:08:44 1996 Craig Burley <burley@gnu.ai.mit.edu>
2514 * Make-lang.in (g77, g77-cross): No libs or lib dependencies
2515 in case where "f77" is not in $LANGUAGES.
2517 * lex.c (ffelex_image_char_, ffelex_file_fixed,
2518 ffelex_file_free): Fixes to properly handle lines with
2519 null character, and too-long lines as well.
2521 * lex.c: Call ffebad_start_msg_lex instead of
2522 ffebad_start_msg throughout.
2524 Sun Dec 1 21:19:55 1996 Craig Burley <burley@gnu.ai.mit.edu>
2526 Fix-up for 1996-11-25 changes:
2527 * com.c (ffecom_member_phase2_): Subtract out 0 offset for
2528 elegance and consistency with EQUIVALENCE aggregates.
2529 (ffecom_sym_transform_): Ditto for LOCAL/COMMON, and
2530 ensure we get the same parent storage area.
2531 * data.c (ffedata_gather_, ffedata_value_): Subtract out
2534 Wed Nov 27 13:55:57 1996 Craig Burley <burley@gnu.ai.mit.edu>
2536 * proj.h: Quote the text of the #error message, to avoid
2537 strange-looking diagnostics from non-gcc ANSI compilers.
2539 * top.c: Make -fno-debug-kludge the default.
2541 Mon Nov 25 20:13:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2543 Provide more info on EQUIVALENCE mismatches:
2544 * bad.def (FFEBAD_EQUIV_MISMATCH): More detailed message.
2545 * equiv.c (ffeequiv_layout_local_, ffeequiv_layout_cblock):
2546 More details for FFEBAD_EQUIV_MISMATCH.
2548 Fix problem with EQUIVALENCE handling:
2549 * equiv.c (ffeequiv_layout_local_): Redesign algorithm --
2550 old one was broken, resulting in rejection of good code.
2551 (ffeequiv_offset_): Add argument, change callers.
2552 Clean up the code, fix up the (probably unused) negative-value
2554 * com.c (ffecom_sym_transform_): For local EQUIVALENCE
2555 member, subtract out aggregate offset (which is <= 0).
2557 Thu Nov 21 12:44:56 1996 Craig Burley <burley@gnu.ai.mit.edu>
2559 Change type of ffetargetOffset from `unsigned long' to `long long':
2560 * bld.c (ffebld_constantarray_dump): Change printf formats.
2561 * storag.c (ffestorag_dump): Ditto.
2562 * symbol.c (ffesymbol_report): Ditto.
2563 * target.h (ffetargetOffset_f): Ditto and change type itself.
2565 Handle situation where list of languages does not include f77:
2566 * Make-lang.in: Define LANGUAGE_F77 to 1 only if `f77' is in
2567 the $LANGUAGES macro for the build.
2568 * g77.c: Compile to a (nearly) no-op program if LANGUAGE_F77
2569 is not defined to 1.
2571 Fixes to delay confirmation of READ, WRITE, and GOTO statements
2572 so the corresponding assignments to same-named CHAR*(*) arrays
2574 * stb.c (ffestb_R90915_, ffestb_91014_): New functions.
2575 (ffestb_goto3_, ffestb_goto5_): Move confirmation from 3 to 5
2576 for the OPEN_PAREN case.
2577 (ffestb_R9091_, ffestb_R9094_, ffestb_R90913_, ffestb_R90914_,
2578 ffestb_R91012_, ffestb_R91013_): Use new functions, and confirm
2579 except for the OPEN_PAREN case.
2581 Fixes to not confirm declarations with an open paren where
2582 an equal sign or other assignment-like token might be, so the
2583 corresponding assignments to same-named CHAR*(*) arrays work:
2584 (ffestb_decl_entsp_5_): Move assertion so we crash on that first,
2585 if it turns out to be wrong, before the less-debuggable crash
2586 on mistaken confirmation.
2587 (ffestb_decl_entsp_6_, ffestb_decl_entsp_7_, ffestb_decl_entsp_8_):
2588 Include OPEN_PAREN in list of assignment-only tokens.
2590 Fix more diagnosed-crash bugs:
2591 * stu.c (ffestu_sym_end_transition): ANY-ize an adjustable array
2592 with bad dimension expressions even if still stateUNCERTAIN.
2593 (ffestu_symter_end_transition_, ffestu_symter_exec_transition_):
2594 Return TRUE for opANY as well.
2595 For code elegance, move opSYMTER case into first switch.
2597 1996-11-17 Dave Love <d.love@dl.ac.uk>
2599 * lex.c: Fix last change.
2601 1996-11-14 Dave Love <d.love@dl.ac.uk>
2603 * Make-lang.in, config-lang.in: Remove the (broken) libU77 stuff,
2606 Thu Nov 14 15:40:59 1996 Craig Burley <burley@gnu.ai.mit.edu>
2608 * bad.def (FFEBAD_UNIMPL_STMT): Explain that invalid
2609 intrinsic references can trigger this message, too.
2611 1996-11-12 Dave Love <d.love@dl.ac.uk>
2613 * lex.c: Declare dwarfout routines.
2615 * config-lang.in: Sink grep o/p.
2617 Mon Nov 11 14:21:13 1996 Craig Burley <burley@gnu.ai.mit.edu>
2619 * g77.c (main): Might as well print version number
2620 for --verbose as well.
2622 Thu Nov 7 18:41:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
2624 * expr.c, lang-options.h, target.h, top.c, top.h: Split out
2625 remaining -fugly stuff into -fugly-logint and -fugly-comma,
2626 leaving -fugly as simply a `macro' that expands into other
2627 options, and eliminate defaults for some of the ugly stuff
2630 * Make-lang.in (gcc-cross): Compile zzz.c, not version.o (!),
2631 in to get version info for this target.
2633 * config-lang.in: Test for GBE patch application based
2634 on whether 2.6.x or 2.7.x GBE is detected.
2636 Wed Nov 6 14:19:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2638 * Make-lang.in (g77): Compile zzz.c in to get version info.
2639 * g77.c: Add support for --help and --version.
2641 * g77.c (lookup_option): Short-circuit long-winded tests
2642 when second char is not hyphen, just to save a spot of time.
2644 Sat Nov 2 13:50:31 1996 Craig Burley <burley@gnu.ai.mit.edu>
2646 * intrin.def: Add FTELL and FSEEK intrinsics, plus new
2647 `g' codes for alternate-return (GOTO) arguments.
2648 * intrin.c (ffeintrin_check_): Support `g' codes.
2649 * com-rt.def: Add ftell_() and fseek_() to database.
2650 * com.c (ffecom_expr_intrinsic_): Ditto. Also, let each
2651 subroutine intrinsic decide for itself what to do with
2652 tree_type, the default being NULL_TREE once again (so
2653 ffecom_call_ doesn't think it's supposed to cast the
2654 function call to the type in the fall-through case).
2656 * ste.c (ffeste_R909_finish): Don't special-case list-directed
2657 I/O, now that libf2c can return non-zero status codes.
2658 (ffeste_R910_finish): Ditto.
2659 (ffeste_io_call_): Simplify logic.
2661 (ffeste_subr_beru_):
2664 (ffeste_R909_start):
2666 (ffeste_R909_finish):
2667 (ffeste_R910_start):
2669 (ffeste_R910_finish):
2670 (ffeste_R911_start):
2671 (ffeste_R923A): Ditto all the above.
2673 Thu Oct 31 20:56:28 1996 Craig Burley <burley@gnu.ai.mit.edu>
2675 * config-lang.in, Make-lang.in: Rename flag file
2676 build-u77 to build-libu77, for consistency with
2677 install-libf2c and such.
2679 * config-lang.in: Don't complain about failure to patch
2680 if pre-2.7.0 gcc is involved (since our patch for that
2681 doesn't add support for tooning).
2683 Sat Oct 26 05:56:51 1996 Craig Burley <burley@gnu.ai.mit.edu>
2685 * bad.def (FFEBAD_TYPELESS_TOO_LARGE): Remove this
2686 unused and redundant diagnostic.
2688 Sat Oct 26 00:45:42 1996 Craig Burley <burley@gnu.ai.mit.edu>
2690 * target.c (ffetarget_integerhex): Fix dumb bug.
2692 1996-10-20 Dave Love <d.love@dl.ac.uk>
2694 * gbe/2.7.2.1.diff: New file.
2696 * Makefile.in (F771_LDFLAGS): Add -bbigtoc for AIX4.1 up, suggested by
2697 endo@material.tohoku.ac.jp [among others!].
2699 Sat Oct 19 03:11:14 1996 Craig Burley <burley@gnu.ai.mit.edu>
2701 * bad.def, bld.c, bld.h, expr.c, lang-options.h, target.c,
2702 target.h, top.c, top.h (ffebld_constant_new_integerbinary,
2703 ffebld_constant_new_integerhex, ffebld_constant_new_integeroctal,
2704 ffeexpr_token_name_apos_name_, ffetarget_integerbinary,
2705 ffetarget_integerhex, ffetarget_integeroctal): Support
2706 new -fno-typeless-boz option with new functions, mods to
2707 existing octal-handling functions, new macros, new error
2708 messages, and so on.
2710 * com.c, lang-options.h, top.c, top.h (ffecom_notify_primary_entry):
2711 Print program unit name on stderr if -fno-silent (new option).
2713 * lang-options.h, top.c, top.h, stt.c (ffestt_dimlist_as_expr):
2714 Treat ([...,]1) in dimlist as ([...,]*) if -fugly-assumed
2717 * lang-options.h: Comment out options duplicated in gcc/toplev.c,
2718 because, somehow, having them commented in and building on my
2719 DEC Alpha results in a cc1 that always segfaults, and gdb that
2720 also segfaults whenever it debugs it up to init_lex() calling
2723 Thu Oct 17 00:39:27 1996 Craig Burley <burley@gnu.ai.mit.edu>
2725 * stb.c (ffestb_R10013_): Don't change meaning of .sign until
2726 after previous meaning/value used to set sign of value
2729 Sun Oct 13 22:15:23 1996 Craig Burley <burley@gnu.ai.mit.edu>
2731 * top.c (ffe_decode_option): Don't set back-end flags
2732 that are nonexistent prior to gcc 2.7.0.
2734 Sun Oct 13 12:48:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2736 * com.c (convert): Don't convert emulated complex expr to
2737 real (via REALPART_EXPR) if the target type is (emulated)
2740 Wed Oct 2 21:57:12 1996 Craig Burley <burley@gnu.ai.mit.edu>
2742 * com.c (ffecom_debug_kludge_): Set DECL_IN_SYSTEM_HEADER so
2743 -Wunused doesn't complain about these manufactured decls.
2744 (ffecom_expr_): Ditto, for original (non-ASSIGN'ed) variable.
2745 (ffecom_transform_equiv_): Clear DECL_IGNORED_P for aggregate
2746 area so it shows up as a debug-accessible symbol.
2747 (pushdecl): Default for "invented" identifiers (a g77-specific
2748 concept for now) is that they are artificial, in system header,
2749 ignored for debugging purposes, used, and (for types) suppressed.
2750 This ought to be overkill.
2752 Fri Sep 27 23:13:07 1996 Craig Burley <burley@gnu.ai.mit.edu>
2754 * ste.c (ffeste_begin_iterdo_, ffeste_end_iterdo_): Support
2755 one-trip DO loops (F66-style).
2756 * lang-options.h, top.c, top.h (-fonetrip): New option.
2758 Thu Sep 26 00:18:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
2760 * com.c (ffecom_debug_kludge_): New function.
2761 (ffecom_sym_transform_): Use new function for COMMON and EQUIVALENCE
2764 * lang-options.h, top.c, top.h (-fno-debug-kludge):
2767 1996-09-24 Dave Love <d.love@dl.ac.uk>
2769 * Make-lang.in (include/f2c.h):
2770 Remove dependencies on xmake_file and tmake_file.
2771 They expand inconsistently in 2.8 c.f. 2.7; $(GCC_PARTS) depends on
2774 1996-09-22 Dave Love <d.love@dl.ac.uk>
2776 * config-lang.in: Add --enable-libu77 option handling.
2779 Conditionally add --enable-libu77 when running runtime configure.
2780 Define LIBU77STAGESTUFF and use it in relevant rules.
2782 1996-08-21 Dave Love <d.love@dl.ac.uk>
2784 * Make-lang.in (f77-runtime):
2785 `stmp-hdrs' should have been `stmp-headers'.
2787 1996-08-20 Dave Love <d.love@dl.ac.uk>
2789 * Make-lang.in (f77-runtime):
2790 Depend on stmp-hdrs, not stmp-int-hdrs, since libF77
2793 Sat Jun 22 18:17:11 1996 Craig Burley <burley@gnu.ai.mit.edu>
2795 * com.c (ffecom_tree_divide_): Fix RECORD_TYPE case to
2796 look at type of first field, properly, to determine
2797 whether to call c_div or z_div.
2799 Tue Jun 4 04:27:18 1996 Craig Burley <burley@gnu.ai.mit.edu>
2801 * com.c (ffecom_build_complex_constant_): Explicitly specify
2803 (ffecom_expr_): Fix thinko.
2804 (ffecom_2): For COMPLEX_EXPR, explicitly specify TREE_PURPOSE.
2806 Mon May 27 16:23:43 1996 Craig Burley <burley@gnu.ai.mit.edu>
2808 Changes to optionally avoid gcc's back-end complex support:
2809 * com.c (ffecom_stabilize_aggregate_): New function.
2810 (ffecom_convert_to_complex_): New function.
2811 (ffecom_make_complex_type_): New function.
2812 (ffecom_build_complex_constant_): New function.
2813 (ffecom_expr_): For opCONVERT of non-COMPLEX to COMPLEX,
2814 don't bother explicitly converting to the subtype first,
2815 because gcc does that anyway, and more code would have
2816 to be added to find the subtype for the emulated-complex
2818 (ffecom_f2c_make_type_): Use ffecom_make_complex_type_
2819 instead of make_node etc. to make a complex type.
2820 (ffecom_1, ffecom_2): Translate operations on COMPLEX operands
2821 to appropriate operations when emulating complex.
2822 (ffecom_constantunion): Use ffecom_build_complex_constant_
2823 instead of build_complex to build a complex constant.
2824 (ffecom_init_0): Change point at which types are laid out
2825 for improved consistency.
2826 Use ffecom_make_complex_type_ instead of make_node etc.
2827 to make a complex type.
2828 Always calculate storage sizes from TYPE_SIZE, never TYPE_PRECISION.
2829 (convert): Use e, not expr, since we've copied into that anyway.
2830 For RECORD_TYPE cases, do emulated-complex conversions.
2831 (ffecom_f2c_set_lio_code_): Always calculate storage sizes
2832 from TYPE_SIZE, never TYPE_PRECISION.
2833 (ffecom_tree_divide_): Allow RECORD_TYPE to also be handled
2834 by run-time library.
2835 (ffecom_expr_intrinsic_): Handle possible RECORD_TYPE as argument
2838 * top.h, top.c, lang-options.h: Support new -f(no-)emulate-complex option.
2840 * com.c (ffecom_sym_transform_): Clarify and fix typos in comments.
2842 Mon May 20 02:06:27 1996 Craig Burley <burley@gnu.ai.mit.edu>
2844 * target.h: Use new REAL_VALUE_UNTO_TARGET_* macros instead
2845 of REAL_VALUE_FROM_TARGET_DOUBLE and _SINGLE.
2846 Explicitly use long instead of HOST_WIDE_INT for emulation
2847 of ffetargetReal1 and ffetargetReal2.
2849 1996-05-20 Dave Love <d.love@dl.ac.uk>
2852 Test for patch being applied with flag_move_all_movables in toplev.c.
2854 * install.texi (Patching GNU Fortran):
2855 Mention overriding X_CFLAGS rather than
2856 editing proj.h on SunOS4.
2858 * Make-lang.in (F77_FLAGS_TO_PASS):
2859 Add X_CFLAGS (convenient for SunOS4 kluge, in
2861 (f77.{,mostly,dist}clean): Reorder things, in particular not to delete
2862 Makefiles too early.
2864 * g77.c (DEFAULT_SWITCH_TAKES_ARG): Define a la gcc.c in the
2865 current GCC snapshot.
2867 Tue May 14 00:24:07 1996 Craig Burley <burley@gnu.ai.mit.edu>
2869 Changes for DEC Alpha AXP support:
2870 * com.c (ffecom_init_0): REAL_ARITHMETIC means internal
2871 REAL/DOUBLE PRECISION might well have a different size
2872 than the compiled type, so don't crash if this is the
2874 * target.h: Use `int' for ffetargetInteger1,
2875 ffetargetLogical1, and magical tests. Set _f format
2876 strings accordingly.
2878 Tue Apr 16 14:08:28 1996 Craig Burley <burley@gnu.ai.mit.edu>
2880 * top.c (ffe_decode_option): -Wall no longer implies
2883 Sat Apr 13 14:50:06 1996 Craig Burley <burley@gnu.ai.mit.edu>
2885 * com.c (ffecom_char_args_): If item is error_mark_node,
2886 set *length that way, too.
2888 * com.c (ffecom_expr_power_integer_): If either operand
2889 is error_mark_node, return that.
2891 * com.c (ffecom_intrinsic_len_): If item is error_mark_node,
2892 return that for length.
2894 * expr.c (ffeexpr_declare_unadorned_,
2895 ffeexpr_declare_parenthesized_): Instead of crashing
2896 on unexpected contexts, produce a diagnostic.
2898 * intrin.c (ffeintrin_check_), intrin.def (impSIGNAL):
2899 Allow procedure as second arg to SIGNAL intrinsic.
2901 * stu.c (ffestu_symter_end_transition_): New function.
2902 (ffestu_symter_exec_transition_): Return bool arg.
2903 Always transition symbol (don't inhibit when !whereNONE).
2904 (ffestu_sym_end_transition): If DUMMY/LOCAL arg has any
2905 opANY exprs in its dimlist, diagnose it so it doesn't
2906 make it through to later stages that try to deal with
2908 (ffestu_sym_exec_transition): If sym has any opANY exprs
2909 in its dimlist, diagnose it so it becomes opANY itself.
2911 * symbol.c (ffesymbol_error): If token arg is NULL,
2912 just ANY-ize the symbol -- don't produce diagnostic.
2914 Mon Apr 1 10:14:02 1996 Craig Burley <burley@gnu.ai.mit.edu>
2916 * Version 0.5.18 released.
2918 Mon Mar 25 20:52:24 1996 Craig Burley <burley@gnu.ai.mit.edu>
2920 * com.c (ffecom_expr_power_integer_): Don't generate code
2921 that compares COMPLEX (or, as it happens, REAL) via "LT_EXPR",
2922 since the back end crashes on that. (This code would never
2923 be executed anyway, but the test that avoids it has now been
2924 translated to control whether the code gets generated at all.)
2927 * com.c (ffecom_type_localvar_): Handle variable-sized
2928 dimension bounds expressions here, so they get calculated
2929 and saved on procedure entry. Fixes 960323-4.f.
2931 * com.c (ffecom_notify_init_symbol): Symbol has no init
2932 info at all if only zeros have been used to initialize it.
2935 * expr.c, expr.h (ffeexpr_type_combine): Renamed from
2936 ffeexpr_type_combine_ and now a public procedure; last arg now
2937 a token, instead of an internal structure used to extract a token.
2938 Now allows the outputs to be aliased with the inputs.
2939 Now allows a NULL token to mean "don't report error".
2940 (ffeexpr_reduced_bool2_, ffeexpr_reduced_eqop2_,
2941 ffeexpr_reduced_math2_, ffeexpr_reduced_power_,
2942 ffeexpr_reduced_relop2_): Handle new calling sequence for
2943 ffeexpr_type_combine.
2944 * (ffeexpr_convert): Don't put an opCONVERT node
2945 in just because the size is unknown; all downstream code
2946 should be able to deal without it being there anyway, and
2947 getting rid of it allows new intrinsic code to more easily
2948 combine types and such without generating bad code.
2949 * info.c, info.h (ffeinfo_kindtype_max): Rewrite to do
2950 proper comparison of size of types, not just comparison
2951 of their internal kind numbers (so I2.eq.I1 doesn't promote
2952 I1 to I2, rather the other way around).
2953 * intrin.c (ffeintrin_check_): Combine types of arguments
2954 in COL a la expression handling, for greater flexibility
2955 and permissiveness (though, someday, -fpedantic should
2956 report use of this kind of thing).
2957 Make sure Hollerith/typeless where CHARACTER expected is
2958 rejected. This all fixes 960323-2.f.
2960 * ste.c (ffeste_begin_iterdo_): Fix some more type conversions
2961 so INTEGER*2-laden DO loops don't crash at compile time on
2962 certain machines. Believed to fix 960323-1.f.
2964 * stu.c (ffestu_sym_end_transition): Certainly reject
2965 whereDUMMY not in any dummy list, whether stateUNCERTAIN
2966 or stateUNDERSTOOD. Fixes 960323-0.f.
2968 Tue Mar 19 13:12:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
2970 * data.c (ffedata_value): Fix crash on opANY, and simplify
2971 the code at the same time.
2973 * Make-lang.in (f77-runtime): Also depends on lib[FI]77/Makefile...
2974 (include/f2c.h...): ...which in turn depend on */Makefile.in.
2975 (f77.rebuilt): Rebuild runtime stuff too.
2977 * intrin.c (ffeintrin_check_): Accommodate TYPELESS/HOLLERITH
2978 types, convert args as necessary, etc.
2980 * expr.c (ffeexpr_convert): Fix test for TYPELESS/HOLLERITH
2981 to obey the docs; crash if no source token when error.
2982 (ffeexpr_collapse_convert): Crash if no token when error.
2984 Mon Mar 18 15:51:30 1996 Craig Burley <burley@gnu.ai.mit.edu>
2986 * com.c (ffecom_init_zero_): Renamed from
2987 ffecom_init_local_zero_; now handles top-level
2988 (COMMON) initializations too.
2990 * bld.c (ffebld_constant_is_zero):
2991 * com.c (ffecom_symbol_transform_, ffecom_sym_transform_assign_,
2992 ffecom_transform_common_, ffecom_transform_equiv_):
3005 * top.h: All of this is mostly housekeeping-type changes
3006 to support -f(no-)zeros, i.e. not always stuff zero
3007 values into the initializer fields of symbol/storage objects,
3008 but still track that they have been given initial values.
3010 * bad.def: Fix wording for DATA-related diagnostics.
3012 * com.c (ffecom_sym_transform_assign_): Don't check
3013 any EQUIVALENCE stuff for local ASSIGN, the check was
3014 bad (crashing), and it's not necessary, anyway.
3016 * com.c (ffecom_expr_intrinsic_): For MAX and MIN,
3017 ignore null arguments as far arg[123], and fix handling
3018 of ANY arguments. (New intrinsic support now allows
3019 spurious trailing null arguments.)
3021 * com.c (ffecom_init_0): Add HOLLERITH (unsigned)
3022 equivalents for INTEGER*2, *4, and *8, so shift intrinsics
3023 and other things that need unsigned versions of signed
3026 Sat Mar 16 12:11:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
3028 * storag.c (ffestorag_exec_layout): Treat adjustable
3029 local array like dummy -- don't create storage object.
3030 * com.c (ffecom_sym_transform_): Allow for NULL storage
3031 object in LOCAL case (adjustable array).
3033 Fri Mar 15 13:09:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
3035 * com.c (ffecom_sym_transform_): Allow local symbols
3036 with nonconstant sizes (adjustable local arrays).
3037 (ffecom_type_localvar_): Allow dimensions with nonconstant
3038 component (adjustable local arrays).
3039 * expr.c: Various minor changes to handle adjustable
3040 local arrays (a new case of stateUNCERTAIN).
3041 * stu.c (ffestu_sym_end_transition,
3042 ffestu_sym_exec_transition): Ditto.
3043 * symbol.def: Update docs to reflect these changes.
3045 * com.c (ffecom_expr_): Reduce space/time needed for
3046 opACCTER case by handling it here instead of converting
3047 it to opARRTER earlier on.
3048 (ffecom_notify_init_storage): Don't convert ACCTER to ARRTER.
3049 (ffecom_notify_init_symbol): Ditto.
3051 * com.c (ffecom_init_0): Crash and burn if any of the types'
3052 sizes, according to the GBE, disagrees with the sizes of
3053 the FFE's internal implementation. This might catch
3054 Alpha/SGI bugs earlier.
3056 Fri Mar 15 01:09:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
3058 * com-rt.def, com.c, com.h: Changes for rewrite of intrinsic
3060 * com.c (ffecom_arglist_expr_): New function.
3061 (ffecom_widest_expr_type_): New function.
3062 (ffecom_expr_intrinsic_): Reorganize, some rewriting.
3063 (ffecom_f2c_make_type_): Layout complex types.
3064 (ffecom_gfrt_args_): New function.
3065 (ffecom_list_expr): Trivial change for consistency.
3067 * expr.c (ffeexpr_token_name_rhs_): Go back to getting
3068 type from specific, not implementation, info.
3069 (ffeexpr_token_funsubstr_): Set intrinsic implementation too!
3070 * intrin.c: Major rewrite of most portions.
3071 * intrin.def: Major rearchitecting of tables.
3072 * intrin.h (ffeintrin_basictype, ffeintrin_kindtype):
3073 Now (once again) take ffeintrinSpec as arg, not ffeintrinImp;
3074 for now, these return NONE, since they're not really needed
3075 and adding the necessary info to the tables is not trivial.
3076 (ffeintrin_codegen_imp): New function.
3077 * stc.c (ffestc_R1208_item): Change way ffeintrin funcs called,
3078 back to original per above; but comment out the code anyway.
3080 * intrin.c (ffe_init_0): Do internal checks only if
3081 -fset-g77-defaults not specified.
3083 * lang-options.h: Add -fset-g77-defaults option.
3084 * lang-specs.h: Always pass -fset-g77-defaults.
3085 * top.c, top.h: New option.
3087 Sat Mar 9 17:49:50 1996 Craig Burley <burley@gnu.ai.mit.edu>
3089 * Make-lang.in (stmp-int-hdrs): Use --no-validate when
3090 generating the f77.rebuilt files (BUGS, INSTALL, NEWS)
3091 so cross-references can work properly in g77.info
3092 without a lot of hassle. Users can probably deal with
3093 the way they end up looking in the f77.rebuilt files.
3095 * bld.c (ffebld_constant_new_integer4_val): INTEGER*8
3096 support -- new function.
3097 (ffebld_constant_new_logical4_val): New function.
3098 * com.c (ffecom_f2c_longint_type_node): New type.
3099 (FFECOM_rttypeLONGINT_): New return type code.
3100 (ffecom_expr_): Add code to invoke pow_qq instead
3101 of pow_ii for INTEGER4 (INTEGER*8) case.
3102 If ffecom_expr_power_integer_ returns NULL_TREE, just do
3104 (ffecom_make_gfrt_): Handle new type.
3105 (ffecom_expr_power_integer_): Let caller do the work if in
3106 dummy-transforming case, since
3107 caller now knows about INTEGER*8 and such, by returning
3109 * expr.c (ffeexpr_reduced_power_): Complain about non-INTEGER
3110 raised to INTEGER4 (INTEGER*8) power.
3112 * target.c (ffetarget_power_integerdefault_integerdefault):
3114 * com.c (ffecom_expr_power_integer_): Fix (-1)**(-8) and similar
3115 to ABS() the integral result if the exponent is negative
3118 * ste.c (ffeste_begin_iterdo_): Clean up a type ref.
3119 Always convert iteration count to _default_ INTEGER.
3121 * sta.c (ffesta_second_): Add BYTE and WORD type/stmts;
3122 changes by Scott Snyder <snyder@d0sgif.fnal.gov>.
3123 * stb.c (ffestb_decl_recursive): Ditto.
3124 (ffestb_decl_recursive): Ditto.
3125 (ffestb_decl_entsp_2_): Ditto.
3126 (ffestb_decl_entsp_3_): Ditto.
3127 (ffestb_decl_funcname_2_): Ditto.
3128 (ffestb_decl_R539): Ditto.
3129 (ffestb_decl_R5395_): Ditto.
3130 * stc.c (ffestc_establish_declstmt_): Ditto.
3131 * std.c (ffestd_R539item): Ditto.
3132 (ffestd_R1219): Ditto.
3134 * str-1t.fin: Ditto.
3135 * str-2t.fin: Ditto.
3137 * expr.c (ffeexpr_finished_): For DO loops, allow
3138 any INTEGER type; convert LOGICAL (assuming -fugly)
3139 to corresponding INTEGER type instead of always default
3140 INTEGER; let later phases do conversion of DO start,
3141 end, incr vars for implied-DO; change checks for non-integral
3142 DO vars to be -Wsurprising warnings.
3143 * ste.c (ffeste_io_impdo_): Convert start, end, and incr
3144 to type of DO variable.
3146 * com.c (ffecom_init_0): Add new types for [IL][234],
3147 much of which was done by Scott Snyder <snyder@d0sgif.fnal.gov>.
3151 Wed Mar 6 14:08:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
3153 * top.c (ffe_init_gbe_): Make -frerun-loop-opt the default.
3155 Mon Mar 4 12:27:00 1996 Craig Burley <burley@gnu.ai.mit.edu>
3157 * expr.c (ffeexpr_exprstack_push_unary_): Really warn only
3158 about two successive _arithmetic_ operators.
3160 * stc.c (ffestc_R522item_object): Allow SAVE of (understood)
3163 * top.c (ffe_decode_option): New -f(no-)second-underscore options.
3164 * top.h: New options.
3165 * com.c (ffecom_get_external_identifier_, ffecom_get_identifier_):
3168 * Make-lang.in (f77.maintainer-clean): Clean f/BUGS, f/INSTALL,
3170 ($(srcdir)/f/BUGS, $(srcdir)/f/INSTALL, $(srcdir)/f/NEWS):
3172 ($(srcdir)/f/g77.info, $(srcdir)/f/g77.dvi): Depend on
3173 f/bugs.texi and f/news.texi.
3174 (f77.install-man): Install f77 man pages (if enabled).
3175 (f77.uninstall): Uninstall info docs, f77 man pages (if enabled).
3177 * top.c (ffe_init_gbe_): New function.
3178 (ffe_decode_option, ffe_file): Call ffe_init_gbe_ to
3179 set defaults for gcc options.
3181 Sat Jan 20 13:57:19 1996 Craig Burley <burley@gnu.ai.mit.edu>
3183 * com.c (ffecom_get_identifier_): Eliminate needless
3184 comparison of results of strchr.
3186 Tue Dec 26 11:41:56 1995 Craig Burley <burley@gnu.ai.mit.edu>
3188 * Make-lang.in: Add rules for new files g77.texi, g77.info,
3190 Reorganize the *clean rules to more closely parallel gcc's.
3192 * config-lang.in: Exclude g77.info from diffs.
3194 Sun Dec 10 02:29:13 1995 Craig Burley <burley@gnu.ai.mit.edu>
3196 * expr.c (ffeexpr_declare_unadorned_,
3197 ffeexpr_declare_parenthesized_): Break out handling of
3198 contextDATAIMPDO[INDEX,CTRL] so it's independent of symbol state.
3199 Don't exec-transition these here (let ffeexpr_sym_impdoitem_
3200 handle that when appropriate). Don't "declare" them twice.
3202 Tue Dec 5 06:48:26 1995 Craig Burley <burley@gnu.ai.mit.edu>
3204 * stc.c (ffestc_promote_sfdummy_): Allow whereNONE parent
3205 symbol, since it is not necessarily known whether it will
3206 become LOCAL or DUMMY.
3208 Mon Dec 4 03:46:55 1995 Craig Burley <burley@gnu.ai.mit.edu>
3210 * lex.c (ffelex_display_token, ffelex_type_string_): Resurrect
3211 these from their old versions and update them for possible invocation
3213 * lex.h (ffelex_display_token): Declare this in case anyone
3214 else wants to call it.
3216 * lex.c (ffelex_total_tokens_): Have this reflect actual allocated
3217 tokens, no longer include outstanding "uses" of tokens.
3219 * malloc.c, malloc.h (MALLOC_DEBUG): New macro to control
3220 checking of whether callers follow rules, now defaults to 0
3221 for "no checking" to improve compile times.
3223 * malloc.c (malloc_pool_kill): Fix bug that could prevent
3224 subpool from actually being killed (wasn't setting its use
3227 * proj.h, *.c (dmpout): Replace all occurrences of `stdout'
3228 and some of `stderr' with `dmpout', so where to dump debugging
3229 output can be easily controlled during build; add default
3230 for `dmpout' of `stderr' to proj.h.
3232 Sun Dec 3 00:56:29 1995 Craig Burley <burley@gnu.ai.mit.edu>
3234 * com.c (ffecom_return_expr): Eliminate attempt at warning
3235 about unset return values, since the back end does this better,
3236 with better wording, and is not triggered by clearly working
3237 (but spaghetti) code as easily as this test.
3239 Sat Dec 2 08:28:56 1995 Craig Burley <burley@gnu.ai.mit.edu>
3241 * target.c (ffetarget_power_*_integerdefault): Raising 0 to
3242 integer constant power should not be an error condition;
3243 if so, other code should catch 0 to any power, etc.
3245 * bad.def (FFEBAD_BAD_POWER): 0**integer now a warning instead
3248 Fri Dec 1 00:12:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
3250 * bad.def: Clarify diagnostic regarding complex constant elements.
3251 * expr.c (ffeexpr_cb_comma_c_): Capitalize real/imaginary
3252 for clarified diagnostic.
3254 * com.c (ffecom_close_include_): Close the file!
3256 * lex.c (ffelex_file_fixed): Update line info if the line
3257 has any content, not just if it finishes a previous line
3259 (ffelex_file_free): Clarify switch statement code.
3261 Sat Nov 18 19:37:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
3263 * Version 0.5.17 released.
3265 Fri Nov 17 14:27:24 1995 Craig Burley (burley@gnu.ai.mit.edu)
3267 * Make-lang.in: Fix typo in comment.
3269 * Makefile.in (f/fini.o, f/proj-h.o): Don't use `$<' since
3270 not all makes support it (e.g. NeXT make), use explicit
3271 source name instead (with $(srcdir) and munging).
3272 (ASSERT_H): assert.h lives in source dir, not build dir.
3274 Thu Nov 16 12:47:50 1995 Craig Burley (burley@gnu.ai.mit.edu)
3276 * com.c (ffecom_init_0): Fix dumb bug in code to produce
3277 warning message about non-32-bit-systems.
3279 * stc.c (ffestc_R501_item): Parenthesize test to make
3280 warning go away (and perhaps fix bug).
3282 Thu Nov 16 03:43:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
3284 * g77.c: Upgrade to 2.7.0's gcc.c.
3285 Fix -v to pass a temp name instead of "/dev/null" for "-o".
3287 Fri Nov 10 19:16:05 1995 Craig Burley (burley@gnu.ai.mit.edu)
3289 * ste.c (ffeste_begin_iterdo_): Add Toon's change to
3290 make loops faster on some machines (implement termination
3291 condition as "--i >= 0" instead of "i-- > 0").
3293 Thu Nov 2 03:58:17 1995 Craig Burley (burley@gnu.ai.mit.edu)
3295 * Make-lang.in: Remove unnecessary $(exeext) a la cp/Make-lang.in.
3297 * com.c (ffecom_expr_): Restore old strategy for assignp variant
3298 of opSYMTER case...always return the ASSIGN version of var.
3299 That way, `-O -Wuninitialized' will catch "I=3;GOTO I;END"
3300 (though the diagnostic will refer to `__g77_ASSIGN_i').
3302 * com.c (ffecom_expr_power_integer_): For constant rhs case,
3303 wrap every new eval of lhs in save_expr() so it is clear to
3304 back end that MULT_EXPR(lhs,lhs) has identical operands,
3305 otherwise for an rhs like 32767 it generates around 65K pseudo
3306 registers, with which stupid_life_analysis cannot cope
3307 (due to reg_renumber in regs.h being `short *' instead of
3310 * com.c (ffecom_expr_): Speed up implementation of LOGICAL
3311 versions of opNOT, opAND, opOR, opXOR/opNEQV, and opEQV by
3312 assuming the values actually are kosher LOGICAL bit patterns.
3313 Also simplify code that implements some of the INTEGER versions
3316 * com.c (skip_redundant_dir_prefix, read_name_map,
3317 ffecom_open_include_, signed_type, unsigned_type): Fold in
3318 changes to cccp.c made from 2.7.0 through ss-950826.
3320 * equiv.c (ffeequiv_layout_local_): Kill the equiv list
3323 * expr.c (ffeexpr_reduced_eqop2_): Issue specific diagnostic
3324 regarding usage of .EQV./.NEQV. in preference to .EQ./.NE..
3326 * intrin.c: Add ERF and ERFC as generic intrinsics.
3329 * sta.c (ffesta_save_, ffesta_second_): Whoever calls
3330 ffestd_exec_begin must also set ffesta_seen_first_exec = TRUE,
3331 and anytime stc sees an exec transition, it must do both.
3332 stc.c (ffestc_eof): Same.
3334 * stc.c (ffestc_promote_sfdummy_): If failed implicit typing
3335 or CHARACTER*(*) arg, after calling ffesymbol_error, don't
3336 reset info to ENTITY/DUMMY, because ffecom_sym_transform_
3337 doesn't expect such a thing with ANY/ANY type.
3339 * target.h (*logical*): Change some of these so they parallel
3340 changes in com.c, e.g. for _eqv_, use (l)==(r) instead of
3341 !!(l)==!!(r), to get a more faithful result.
3343 Fri Oct 27 07:06:59 1995 Craig Burley (burley@gnu.ai.mit.edu)
3345 * com.c (ffecom_sym_transform_): Simplify code for local
3348 * expr.c (ffeexpr_exprstack_push_unary_): Warn about two
3349 successive operators.
3350 (ffeexpr_exprstack_push_binary_): Warn about "surprising"
3351 operator precedence, as in "-2**2".
3353 * lang-options.h: Add -W(no-)surprising options.
3355 * parse.c (yyparse): Don't reset -fpedantic if not -pedantic.
3357 * top.c (ffe_decode_option): Support new -Wsurprising option.
3360 Mon Oct 23 09:14:15 1995 Craig Burley (burley@gnu.ai.mit.edu)
3362 * com.c (ffecom_finish_symbol_transform_): Don't transform
3363 NONE/NONE (CHARACTER*(*)) vars, as these don't mean anything
3364 in debugging terms, and can't be turned into anything
3365 in the back end (so ffecom_sym_transform_ crashes on them).
3367 * com.c (ffecom_expr_): Change strategy for assignp variant
3368 of opSYMTER case...always return the original var unless
3369 it is not wide enough.
3371 * ste.c (ffeste_io_cilist_): Clarify diagnostic for ASSIGN
3372 involving too-narrow variable. This shouldn't happen, though.
3373 (ffeste_io_icilist_): Ditto.
3374 (ffeste_R838): Ditto.
3375 (ffeste_R839): Ditto.
3377 Thu Oct 19 03:21:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
3379 * com.c (ffecom_sym_transform_assign_): Set TREE_STATIC
3380 using the same decision-making process as used for their twin
3381 variables, so ASSIGN can last across RETURN/CALL as appropriate.
3383 Fri Sep 22 20:21:18 1995 Craig Burley (burley@gnu.ai.mit.edu)
3385 * Makefile.in: fini is a host program, so it needs a host-compiled
3386 version of proj.o, named proj-h.o. f/fini, f/fini.o, and
3387 f/proj-h.o targets updated accordingly.
3389 * com.c (__eprintf): New function.
3391 Wed Sep 20 02:26:36 1995 Craig Burley (burley@gnu.ai.mit.edu)
3393 * lang-options.h: Add omitted -funix-intrinsics-* options.
3395 * malloc.c (malloc_find_inpool_): Check for infinite
3396 loop, crash if detected (user reports encountering
3397 them in some large programs, this might help track
3400 Thu Sep 7 13:00:32 1995 Craig Burley (burley@gnu.ai.mit.edu)
3402 * com.c (lang_print_error_function): Don't dereference null
3403 pointer when outside any program unit.
3404 (ffecom_let_char_, ffecom_arg_ptr_to_expr): If catlist
3405 item or length ever error_mark_node, don't continue processing,
3406 since back-end functions like build_pointer_type crash on
3407 error_mark_node's (due to pushing bad obstacks, etc.).
3409 Wed Aug 30 15:58:35 1995 Craig Burley (burley@gnu.ai.mit.edu)
3411 * Version 0.5.16 released.
3413 Mon Aug 28 12:24:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
3415 * bad.c (ffebad_finish): Fix botched message when no places
3416 are printed (due to unknown line info, etc.).
3418 * std.c (ffestd_subr_labels_): Do a better job finding
3419 line info in the case of typeANY and diagnostics.
3421 Fri Aug 25 15:19:29 1995 Craig Burley (burley@gnu.ai.mit.edu)
3423 * com.c (DECL_ARTIFICIAL): Surround all references to this
3424 macro with #if !BUILT_FOR_270 and #endif.
3425 (init_lex): Surround print_error_function decl with
3426 #if !BUILT_FOR_270 and #endif.
3427 (lang_init): Call new ffelex_hash_kludge function to solve
3428 problem with preprocessed files that have INCLUDE statements.
3430 * lex.c (ffelex_getc_): New function.
3431 (ffelex_cfelex_): Use ffelex_getc_ instead of getc in any
3432 paths of code that can be affected by ffelex_hash_kludge.
3433 Don't make an EOF token for unrecognized token; set token
3434 to NULL instead, to avoid problems when not initialized.
3435 (ffelex_hash_): Use ffelex_getc_ instead of getc in any
3436 paths of code that can be affected by ffelex_hash_kludge.
3437 Test token returned by ffelex_cfelex_ for NULL, meaning
3439 Get rid of useless used_up variable.
3440 Don't do ffewhere stuff or kill any tokens if in
3442 (ffelex_file_fixed, ffelex_file_free): Use ffelex_getc_
3443 instead of getc in any paths of code that can be affected
3444 by ffelex_hash_kludge.
3445 (ffelex_hash_kludge): New function.
3447 * lex.h (ffelex_hash_kludge): New function.
3449 Wed Aug 23 15:17:40 1995 Craig Burley (burley@gnu.ai.mit.edu)
3451 * com.c: Implement -f(no-)underscoring options by always
3452 compiling in code to do it, and having that code inhibit
3453 itself when -fno-underscoring is in effect. This option
3454 overrides -f(no-)f2c for this purpose; -f(no-)f2c returns
3455 to it's <=0.5.15 behavior of affecting only how code
3456 is generated, not how/whether names are mangled.
3458 * target.h: Redo specification of appending underscores so
3459 the macros are named "_default" instead of "_is" and the
3460 two-underscore macro defaults to 1.
3462 * top.c, top.h (underscoring): Add appropriate stuff
3463 for the -f(no-)underscoring options.
3465 Tue Aug 22 10:25:01 1995 Craig Burley (burley@gnu.ai.mit.edu)
3467 * bad.c (ffebad_finish): Call report_error_function (in toplev.c)
3468 to better identify location of problem.
3469 Say "(continued):" instead of "(continued:)" for consistency.
3471 * com.c (ffecom_gen_sfuncdef_): Set and reset new
3472 ffecom_nested_entry_ variable to hold ffesymbol being compiled.
3473 (lang_print_error_function): New function from toplev.c.
3474 Use ffecom_nested_entry_ to help determine which name
3475 and kind-string to print.
3476 (ffecom_expr_intrinsic_): Handle EXIT and FLUSH invocations
3477 with different calling sequences than library functions.
3478 Have SIGNAL and SYSTEM push and pop calltemps, and convert
3479 their return values to the destination type (just in case).
3480 (FFECOM_rttypeINT_): New return type for `int', in case
3481 gcc/f/runtime/libF77/system_.c(system_) is really supposed
3482 to return `int' instead of `ftnint'.
3484 * com.h (report_error_function): Declare this.
3486 * equiv.c (ffeequiv_layout_local_): Don't forget to consider
3487 root variable itself as possible "first rooted variable",
3488 else might never set symbol and then crash later.
3490 * intrin.c (ffeintrin_check_exit_): Change to allow no args
3491 and rename to ffeintrin_check_int_1_o_ for `optional'.
3492 #define ffeintrin_check_exit_ and _flush_ to this new
3493 function, so intrin.def can refer to the appropriate names.
3495 * intrin.def (FFEINTRIN_impFLUSH): Validate using
3496 ffeintrin_check_flush_ so passing an INTEGER arg is allowed.
3498 * lex.c (ffelex_file_push_, ffelex_file_pop_): New functions
3499 to manage input_file_stack in gbe.
3500 (ffelex_hash_): Call new functions (instead of doing code).
3501 (ffelex_include_): Call new functions to update stack for
3502 INCLUDE (_hash_ handles cpp output of #include).
3504 Mon Aug 21 08:09:04 1995 Craig Burley (burley@gnu.ai.mit.edu)
3506 * Makefile.in: Put `-W' in front of every `-Wall', since
3507 2.7.0 requires that to engage `-Wunused' for parameters.
3509 * com.c: Mark all parameters as artificial, so
3510 `-W -Wunused' doesn't complain about unused ones (since
3511 there's no way right not to individually specify attributes
3514 * proj.h: Don't #define UNUSED if already defined, regardless
3517 Sun Aug 20 16:03:56 1995 Craig Burley (burley@gnu.ai.mit.edu)
3519 * gbe/2.7.0.diff: Regenerate.
3521 * lang-options.h, lang-specs.h: If not __STDC__ (ANSI C),
3522 avoid doing anything, especially the stringizing in -specs.h.
3524 Thu Aug 17 03:36:12 1995 Craig Burley (burley@gnu.ai.mit.edu)
3526 * lang-specs.h: Remove useless optional settings of -traditional,
3527 since -traditional is always set anyway.
3529 Wed Aug 16 16:56:46 1995 Craig Burley (burley@gnu.ai.mit.edu)
3531 * Make-lang.in (F2C_INSTALL_FLAG, F2CLIBOK): More
3532 control over whether to install f2c-related stuff.
3533 (install-f2c-*): New targets to install f2c-related
3534 stuff in system, not just gcc, directories.
3536 * com.c: Change calls to ffecom_get_invented_identifier
3537 to use generally more predictable names.
3538 Change calls to build_range_type to ensure consistency
3539 of types of operands.
3540 (ffecom_get_external_identifier_): Change to accept
3541 symbol info, not just text, so it can use f2c flag for
3542 symbol to decide whether to append underscore(s).
3543 (ffecom_get_identifier_): Don't change names if f2c flag
3544 off for compilation.
3545 (ffecom_type_permanent_copy_): Use same type for new max as
3547 (ffecom_notify_init_storage): Offline fixups for stand-alone.
3549 * data.c (ffedata_gather): Explicitly test for common block,
3550 since it's no longer always the case that a local EQUIVALENCE
3551 group has no symbol ptr (it now can, if a user-predictable
3552 "rooted" symbol has been identified).
3554 * equiv.c: Add some debugging stuff.
3555 (ffeequiv_layout_local_): Set symbol ptr with user-predictable
3556 "rooted" symbol, for giving the invented aggregate a
3559 * g77.c (append_arg): Allow for 20 extra args instead of 10.
3560 (main): For version-only case, add `-fnull-version' and, unless
3561 explicitly omitted, `-lf2c -lm'.
3563 * lang-options.h: New "-fnull-version" option.
3565 * lang-specs.h: Support ".fpp" suffix for preprocessed source
3566 (useful for OS/2, MS-DOS, other case-insensitive systems).
3568 * stc.c (ffestc_R544_equiv_): Swap way lists are merged so this
3569 is consistent with the order in which lists are built, making
3570 user predictability of invented aggregate name much higher.
3572 * storag.c, storag.h (FFESTORAG_typeDUMMY): Delete this enum.
3574 * top.c: Accept, but otherwise ignore, `-fnull-version'.
3576 Tue Aug 15 07:01:07 1995 Craig Burley (burley@gnu.ai.mit.edu)
3578 * DOC, INSTALL, PROJECTS: Extensive improvements to documentation.
3580 Sun Aug 13 01:55:18 1995 Craig Burley (burley@gnu.ai.mit.edu)
3582 * INSTALL (f77-install-ok): Document the use of this file.
3584 * Make-lang.in (F77_INSTALL_FLAG): New flag to control
3585 whether to install an `f77' command (based on whether
3586 a file named `f77-install-ok' exists in the source or
3587 build directory) to replace the broken attempt to use
3588 comment lines to avoid installing `f77' (broken in the
3589 sense that it prevented installation of `g77').
3591 Mon Aug 7 06:14:26 1995 Craig Burley (burley@gnu.ai.mit.edu)
3593 * DOC: Add new sections for g77 & gcc compiler options,
3594 source code form, and types, sizes and precisions.
3595 Remove lots of old "delta-version" info, or at least
3598 * INSTALL: Add info here that used to be in DOC.
3601 * g77.c (lookup_option, main): Check for --print-* options,
3602 so we avoid adding version-determining stuff.
3604 Wed Jul 26 15:51:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
3606 * Make-lang.in, Makefile.in (input.j, INPUT_H): New file.
3607 Update dependencies accordingly.
3609 * bad.c (ffebad_here): Okay to use unknown line/col.
3611 * compilers.h (@f77-cpp-input): Remove -P option now that
3612 # directives are handled by f771. Update other options
3613 to be more consistent with @c in gcc/gcc.c. Don't run f771
3614 if -E specified, etc., a la @c.
3615 (@f77): Don't run f771 if -E specified, etc., a la @c.
3617 * config-lang.in: Avoid use of word "guaranteed".
3619 * input.j: New file to wrap around gcc/input.h.
3621 * lex.j: Add support for parsing # directives output by cpp.
3622 (ffelex_cfebackslash_): New function.
3623 (ffelex_cfelex_): New function.
3624 (ffelex_get_directive_line_): New function.
3625 (ffelex_hash_): New function.
3626 (ffelex_include_): Change to not use ffewhere_file_(begin|end).
3627 Also fix bug in pointing to next line (for diagnostics, &c)
3628 following successful INCLUDE.
3629 (ffelex_next_line_): New function that does chunk of code
3630 seen in several places elsewhere in the lexers.
3631 (ffelex_file_fixed): Delay finishing statement until source
3632 line is registered with ffewhere, so INCLUDE processing
3633 picks up the info correctly.
3634 Okay to kill or use unknown line/col objects now.
3635 Handle HASH (#) lines.
3636 Reorder tests for insubstantial lines to put most frequent
3637 occurrences at top, for possible minor speedup.
3638 Some general consolidation of code.
3639 (ffelex_file_free): Handle HASH (#) lines.
3640 Okay to kill or use unknown line/col objects now.
3641 Some general consolidation of code.
3642 (ffelex_init_1): Detect HASH (#) lines.
3643 (ffelex_set_expecting_hollerith): Okay to kill or use unknown
3644 line/col objects now.
3646 * lex.h (FFELEX_typeHASH): New enum.
3648 * options-lang.h (-fident, -fno-ident): New options.
3650 * stw.c (ffestw_update): Okay to kill unknown line/col objects
3653 * target.h (FFETARGET_okREALQUAD, FFETARGET_okCOMPLEXDOUBLE,
3654 FFETARGET_okCOMPLEXQUAD): #define these appropriately.
3656 * top.c: Include flag.j wrapper, not flags.h directly.
3657 (ffe_is_ident_): New flag.
3658 (ffe_decode_option): Handle -fident and -fno-ident.
3659 (ffe_file): Replace obsolete ffewhere_file_(begin|end) with
3662 * top.h (ffe_is_ident_, ffe_is_ident, ffe_set_is_ident):
3663 New flag and access functions.
3665 * where.c, where.h: Remove all tracking of parent file.
3666 (ffewhere_file_begin, ffewhere_file_end): Delete these.
3667 (ffewhere_line_use): Make it work with unknown line object.
3669 Mon Jul 17 03:04:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3671 * com.c (ffecom_sym_transform_): Set DECL_IN_SYSTEM_HEADER
3672 flag for any local vars used as stmtfunc dummies or DATA
3673 implied-DO iter vars, so no -Wunused warnings are produced
3674 for them (a la f2c).
3675 (ffecom_init_0): Do "extern int xargc;" for IARGC() intrinsic.
3676 Warn if target machine not 32 bits, since g77 isn't yet
3677 working on them at all well.
3679 * expr.c (ffeexpr_sym_lhs_call_, ffeexpr_sym_lhs_data_,
3680 ffeexpr_sym_lhs_extfunc_, ffeexpr_sym_rhs_actualarg_,
3681 ffeexpr_sym_rhs_let_, ffeexpr_paren_rhs_let_): Don't
3682 gratuitously set attr bits that don't apply just
3683 to avoid null set meaning error; instead, use explicit
3684 error flag, and allow null attr set, to
3685 fix certain bugs discovered by looking at this code.
3687 * g77.c: Major changes to improve support for gcc long options,
3688 to make `g77 -v' report more useful info, and so on.
3690 Mon Jul 3 14:49:16 1995 Craig Burley (burley@gnu.ai.mit.edu)
3692 * DOC, com.c, intrin.h, intrin.c, intrin.def, target.h, top.c,
3693 top.h: Add new `unix' group of intrinsics, which includes the
3694 newly added ERF, ERFC, EXIT, plus even newer ABORT, DERF, DERFC,
3695 FLUSH, GETARG, GETENV, SIGNAL, and SYSTEM.
3697 Tue Jun 27 23:01:05 1995 Craig Burley (burley@gnu.ai.mit.edu)
3699 * bld.c, bld.h (ffebld_constant_pool,
3700 ffebld_constant_character_pool): Use a single macro (the
3701 former) to access the pool for allocating constants, instead
3702 of latter in public and FFEBLD_CONSTANT_POOL_ internally
3703 in bld.c (which was the only one that was correct before
3704 these changes). Add verification of integrity of certain
3705 heap-allocated areas.
3707 * com.c (ffecom_overlap_, ffecom_args_overlap_,
3708 ffecom_tree_canonize_ptr_, ffecom_tree_canonize_ref_): New
3709 functions to optimize calling COMPLEX and, someday, CHARACTER
3710 functions requiring additional argument to be passed.
3711 (ffecom_call_, ffecom_call_binop_, ffecom_expr_,
3712 ffecom_expr_intrinsic_): Change calling
3713 sequences to include more info on possible destination.
3714 (ffecom_expr_intrinsic_): Add ERF(), ERFC(), and EXIT()
3716 (ffecom_sym_transform_): For assumed-size arrays, set high
3717 bound to highest possible value instead of low bound, to
3718 improve validity of overlap checking.
3719 (duplicate_decls): If olddecl and newdecl are the same,
3720 don't do any munging, just return affirmative.
3722 * expr.c: Change ffecom_constant_character_pool() to
3723 ffecom_constant_pool().
3725 * info.c (ffeinfo_new): Compile this version if not being
3728 * info.h (ffeinfo_new): Don't define macro if not being
3731 * intrin.c, intrin.def: Add ERF(), ERFC(), and EXIT() intrinsics.
3732 (ffeintrin_check_exit_): New for EXIT() subroutine intrinsic.
3734 * malloc.c, malloc.h (malloc_verify_*): New functions to verify
3735 integrity of heap-storage areas.
3737 * stc.c (ffestc_R834, ffestc_R835): Handle possibility that
3738 an enclosing DO won't have a construct name even when the
3739 CYCLE/EXIT does (i.e. without dereferencing NULL).
3741 * target.c, target.h (ffetarget_verify_character1): New function
3742 to verify integrity of heap storage used to hold character constant.
3744 Thu Jun 22 15:36:39 1995 Howard Gordon (flash@super.org)
3746 * stp.h (ffestpVxtcodeIx): Fix typo in typedef for this.
3748 Mon May 29 15:22:31 1995 Craig Burley (burley@gnu.ai.mit.edu)
3750 * *: Make all sorts of changes to accommodate upcoming gcc-2.7.0.
3751 I didn't keep track of them, nor just when I made them, nor
3752 when I (much later, probably in early August 1995) modified
3753 them so they could properly handle both 2.7.0 and 2.6.x.
3755 * com.c (ffecom_expr_power_integer_): Don't expand_start_stmt_expr
3756 if transforming dummy args, because the back end cannot handle
3757 that (it's rejected by the gcc front end), just generate
3758 call to run-time library.
3759 Back out changes in 0.5.15 because more temporaries might be
3760 needed anyway (for COMPLEX**INTEGER).
3761 (ffecom_push_tempvar): Remove inhibitor.
3762 Around start_decl and finish_decl (in particular, arround
3763 expand_decl, which is called by them), push NULL_TREE into
3764 sequence_rtl_expr, an external published by gcc/function.c.
3765 This makes sure the temporary is truly in the function's
3766 context, not the inner context of a statement-valued expression.
3767 (I think the back end is inconsistent here, but am not
3768 interested in convincing the gbe maintainers about this now.)
3769 (pushdecl): Make sure that when pushing PARM_DECLs, nothing
3770 other than them are pushed, as happened for 0.5.15 and which,
3771 if done for other reasons not fixed here, might well indicate
3772 some other problem -- so crash if it happens.
3774 * equiv.c (ffeequiv_layout_local_): If the local equiv group
3775 has a non-nil COMMON field, it should mean that an error has
3776 occurred and been reported, so just trash the local equiv
3777 group and do nothing.
3779 * stc.c (ffestc_promote_sfdummy_): Set sfdummy arg state to
3780 UNDERSTOOD so above checking for duplicate args actually
3781 works, and so we don't crash later in pushdecl.
3783 * ste.c (ffeste_R1001): Set initial value only for VAR_DECLs,
3784 not for, e.g., LABEL_DECLs, which the FORMAT label can be
3785 if it was previously treated as an executable label.
3787 Sat May 20 01:53:53 1995 Craig Burley (burley@gnu.ai.mit.edu)
3789 * com.c (ffecom_sym_transform_): For adjustable arrays,
3790 pass high bound through variable_size in case its primaries
3791 are changed (dumb0.f, and this might also improve
3792 performance so it approaches f2c|gcc).
3794 Fri May 19 11:00:36 1995 Craig Burley (burley@gnu.ai.mit.edu)
3796 * Version 0.5.15 released.
3798 * com.c (ffecom_expr_power_integer_): Push temp vars
3799 before expanding a statement expression, since that seems
3800 to cause temp vars to be "forgotten" after the end of the
3801 expansion in the back end. Disallow more temp-var
3802 pushing during such an expansion, just in case.
3803 (ffecom_push_tempvar): Crash if a new variable needs to be
3804 pushed but cannot be at this point (should never happen).
3806 Wed May 17 12:26:16 1995 Craig Burley (burley@gnu.ai.mit.edu)
3808 * expr.c (ffeexpr_collapse_convert): Add code to convert
3809 LOGICAL to CHARACTER. Reject conversion of REAL or COMPLEX
3810 to CHARACTER entirely, as it cannot be supported with all
3813 * target.h, target.c (ffetarget_convert_character1_logical1):
3816 Sun May 14 00:00:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3818 * com.c (ffecom_do_entry_, ffecom_gen_sfuncdef_,
3819 ffecom_start_progunit_, ffecom_sym_transform_,
3820 ffecom_init_0, start_function): Changes to have REAL
3821 external functions return same type as DOUBLE PRECISION
3822 external functions when -ff2c is in force; while at it,
3823 some code cleanups done.
3825 * stc.c (ffestc_R547_item_object): Disallow array declarator
3826 if one already exists for symbol.
3828 * ste.c (ffeste_R1227): Convert result variable to type
3829 of function result as seen by back end (e.g. for when REAL
3830 external function actually returns result as double).
3832 * target.h (FFETARGET_defaultFIXED_LINE_LENGTH): New
3833 macro for default for -ffixed-line-length-N option.
3835 * top.c (ffe_fixed_line_length_): Initialize this to new
3836 target.h macro instead of constant 72.
3838 Tue May 9 01:20:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
3840 * lex.c (ffelex_send_token_): If sending CHARACTER token with
3841 null text field, put a single '\0' in it and set length/size
3842 fields to 0 (to fix 950508-0.f).
3843 (ffelex_image_char_): When setting ffelex_bad_line_ to TRUE,
3844 always "close" card image by appending a null char and setting
3845 ffelex_card_length_. As part of this, append useful text
3846 to identify the two kinds of problems that involve this.
3847 (ffelex_file_fixed): Reset ffelex_bad_line_ to FALSE after
3848 seeing a line with invalid first character (fixes 950508-1.f).
3849 If final nontab column is zero, assume tab seen in line.
3850 (ffelex_card_image_): Always make this array 8 characters
3851 longer than reflected by ffelex_card_size_.
3852 (ffelex_init_1): Get final nontab column info from top instead
3855 * options-lang.h: Add -ffixed-line-length- prefix.
3857 * top.h: Add ffe_fixed_line_length() and _set_ version, plus
3858 corresponding extern.
3860 * top.c: Handle -ffixed-line-length- option prefix.
3862 Fri Apr 28 05:40:25 1995 Craig Burley (burley@gnu.ai.mit.edu)
3864 * Version 0.5.14 released.
3866 * Make-lang.in: Add assert.j.
3868 * Makefile.in: Add assert.j.
3870 * assert.j: New file.
3872 Thu Apr 27 16:24:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
3874 * bad.h (ffebad_severity): New function.
3876 * bad.c (ffebad_severity): New function.
3878 * bad.def (FFEBAD_OPEN_INCLUDE): Change severity from SEVERE
3879 to FATAL, since processing continues, and that seems fine.
3881 * com.c: Add facility to handle -I.
3882 (ffecom_file, ffecom_close_include, ffecom_open_include,
3883 ffecom_decode_include_option): New global functions for -I.
3884 (ffecom_file_, ffecom_initialize_char_syntax_,
3885 ffecom_close_include_, ffecom_decode_include_option_,
3886 ffecom_open_include_, append_include_chain, open_include_file,
3887 print_containing_files, read_filename_string, file_name_map,
3888 savestring): New internal functions for -I.
3890 * compilers.h: Pass -I flag(s) to f771 (via "%{I*}").
3892 * lex.c (ffelex_include_): Call ffecom_close_include
3893 to close include file, for its tracking needs for -I,
3894 instead of using fclose.
3896 * options-lang.h: Add -I prefix.
3898 * parse.c (yyparse): Call ffecom_file for main input file,
3899 so -I handling works (diagnostics).
3901 * std.c (ffestd_S3P4): Have ffecom_open_include handle
3902 opening and diagnosing errors with INCLUDE files.
3904 * ste.c (ffeste_begin_iterdo_): Use correct algorithm for
3905 calculating # of iterations -- mathematically similar but
3906 computationally different algorithm was not handling cases
3907 like "DO I=6,5,2" correctly, because (5-6)/2+1 => 1, not 0.
3909 * top.c (ffe_decode_option): Allow -I, restructure a bit
3910 for clarity and, maybe, speed.
3912 Mon Apr 17 13:31:11 1995 Craig Burley (burley@gnu.ai.mit.edu)
3914 * g77.c: Remove -lc, turns out not all systems has it, but
3915 leave other changes in for clarity of code.
3917 Sun Apr 16 21:50:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
3919 * com.c (ffecom_expr_): Implement ARRAY_EXPR as INDIRECT_REF
3920 of appropriate PLUS_EXPRs of ptr_to_expr of array, to see
3921 if this generates better code. (Conditional on
3922 FFECOM_FASTER_ARRAY_REFS.)
3924 Sun Apr 16 00:22:48 1995 Craig Burley (burley@gnu.ai.mit.edu)
3926 * Make-lang.in (F77_SRCS): Remove g77.c, since it doesn't
3927 contribute to building f771.
3929 * Makefile.in (dircheck): Remove/replace with f/Makefile, because
3930 phony targets that are referenced in other real targets get run
3931 when those targets are specified, which is a waste of time (e.g.
3932 when rebuilding and only g77.c has changed, f771 was being linked
3935 * g77.c: Include -lc between -lf2c and -lm throughout.
3937 * implic.c (ffeimplic_establish_symbol): If -Wimplicit, warn if
3938 implicit type given to symbol.
3940 * lex.c (ffelex_include_): Don't gratuitously increment line
3943 * top.h, top.c (ffe_is_warn_implicit_): New global variable and
3944 related access macros.
3945 (ffe_decode_option): Handle -W options, including -Wall and
3948 * where.c (ffewhere_line_new): Don't muck with root line (was
3949 crashing on null input since lexer changes over the past week
3952 Thu Apr 13 16:48:30 1995 Craig Burley (burley@gnu.ai.mit.edu)
3954 * com.c (ffecom_init_0): Register built-in functions for cos,
3956 (ffecom_tree_fun_type_double): New variable.
3957 (ffecom_expr_intrinsic_): Update f2c input and output files
3958 to latest version of f2c (no important g77-related changes
3959 noted, just bug fixes to f2c and such).
3960 (builtin_function): New function from c-decl.c.
3962 * com-rt.def: Refer to built-in functions for cos, sin, and sqrt.
3964 Thu Apr 13 10:25:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3966 * com.c (ffecom_expr_intrinsic_): Convert 0. to appropriate
3967 type to keep DCMPLX(I) from crashing the compiler.
3968 (ffecom_expr_): Don't convert result from ffecom_tree_divide_.
3969 (ffecom_tree_divide_): Add tree_type argument, have all callers
3970 pass one, and don't convert right-hand operand to it (this is
3971 to make this new function work as much like the old in-line
3972 code used in ffecom_expr_ as possible).
3974 * lex.c: Maintain lineno and input_filename the way the gcc
3977 * std.c (ffestd_exec_end): Save and restore lineno and
3978 input_filename around the second pass, which sets them
3979 appropriately for each saved statement.
3981 Wed Apr 12 09:44:45 1995 Craig Burley (burley@gnu.ai.mit.edu)
3983 * com.c (ffecom_expr_power_integer_): New function.
3984 (ffecom_expr_): Call new function for power op with integer second
3985 argument, for generating better code. Also replace divide
3986 code with call to new ffecom_tree_divide_ function.
3987 Canonicalize calls to ffecom_truth_value(_invert).
3988 (ffecom_tree_divide_): New function.
3990 Wed Apr 5 14:15:44 1995 Craig Burley (burley@gnu.ai.mit.edu)
3992 * lex.c: Change to allocate text for tokens only when actually
3993 needed, which should speed compilation up somewhat.
3994 Change to allow INCLUDE at any point where a statement
3995 can end, i.e. in ffelex_finish_statement_ or when a SEMICOLON
3997 Remove some old, obsolete code.
3998 Clean up layout of entire file to improve formatting,
4000 (ffelex_set_expecting_hollerith): Remove include argument.
4002 Fri Mar 31 23:19:08 1995 Craig Burley (burley@gnu.ai.mit.edu)
4004 * bad.h, bad.c (ffebad_start_msg, ffebad_start_msg_lex):
4005 New functions to generate arbitrary messages.
4006 (FFEBAD_severityPEDANTIC): New severity, to correspond
4007 to toplev's pedwarn() function.
4009 * lex.c (ffelex_backslash_): New function to implement
4010 backslash processing.
4011 (ffelex_file_fixed, ffelex_file_free): Implement new
4012 backslash processing.
4014 * std.c (ffestd_R1001dump_): Don't assume CHARACTER and
4015 HOLLERITH tokens stop at '\0' characters, now that backslash
4016 processing is supported -- use their advertised lengths instead,
4017 and double up the '\002' character for libf2c.
4019 Mon Mar 27 17:10:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
4021 * com.c (ffecom_init_local_zero_): Implement -finit-local-zero.
4022 (ffecom_sym_transform_): Same.
4023 (ffecom_transform_equiv_): Same.
4025 * options-lang.h: Add -f(no-)(init-local-zero,backslash,ugly-init).
4027 * stb.c (ffestb_V020): Reject "TYPEblah(...", which might be
4028 an array assignment.
4030 * target.h, top.h, top.c: Implement -finit-local-zero.
4032 Fri Mar 24 19:56:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
4034 * Make-lang.in, Makefile.in: Remove conf-proj(.in) and
4035 proj.h(.in) rules, plus related config.log, config.cache,
4036 and config.status stuff.
4038 * com.c (ffecom_init_0): Change messages when atof(), bsearch(),
4039 or strtoul() do not work as expected in the start-up test.
4041 * conf-proj, conf-proj.in: Delete.
4043 * lex.c (ffelex_file_fixed): Allow f2c's '&' in column 1
4044 to mean continuation line.
4046 * options-lang.h: New file, #include'd by ../toplev.c.
4048 * proj.h.in: Rename back to proj.h.
4050 * proj.h (LAME_ASSERT): Remove.
4051 (LAME_STDIO): Remove.
4052 (NO_STDDEF): Remove.
4053 (NO_STDLIB): Remove.
4054 (NO_BSEARCH): Remove auto detection, rename to !FFEPROJ_BSEARCH.
4055 (NO_STRTOUL): Remove auto detection, rename to !FFEPROJ_STRTOUL.
4056 (USE_HOST_LIMITS): Remove (maybe still needed by stand-alone?).
4057 (STR, STRX): Do only ANSI C definitions.
4059 Mon Mar 13 10:46:13 1995 Craig Burley (burley@gnu.ai.mit.edu)
4061 * BUGS: Add item about g77 requiring gcc to compile it.
4063 * NEWS: New file listing user-visible changes in the release.
4065 * PROJECTS: Update to include a new item or two, and modify
4066 or delete items that are addressed in this or previous releases.
4068 * bad.c (ffebad_finish): Don't crash if missing string &c,
4069 just substitute obviously distressed string "[REPORT BUG!!]"
4070 for cases where the message/caller are fudgy.
4072 * bad.def: Clean up error messages in a major way, add new ones
4073 for use by changes in target.c.
4075 * com.c (ffecom_expr_): Handle opANY in opCONVERT.
4076 (ffecom_let_char_): Disregard destinations with ERROR_MARK.
4077 (ffecom_1, ffecom_1_fn, ffecom_2, ffecom_2s, ffecom_3,
4078 ffecom_3s, &c): Check all inputs for error_mark_node.
4079 (ffecom_start_progunit_): Don't transform all symbols
4080 in BLOCK DATA, since it never executes, and it is silly
4081 to, e.g., generate all the structures for NAMELIST.
4082 (ffecom_char_length_expr_): Rename to ffecom_intrinsic_len_.
4083 (ffecom_intrinsic_ichar_): New function to handle ICHAR of
4084 arbitrary expression with possible 0-length operands.
4085 (ffecom_expr_intrinsic_): Use ffecom_intrinsic_char_.
4086 For MVBITS, set tree_type to void_type_node.
4087 (ffecom_start_progunit_): Name master function for entry points
4088 after primary entry point so users can easily guess it while
4090 (ffecom_arg_ptr_to_expr): Change treatment of Hollerith,
4091 Typeless, and %DESCR.
4092 (ffecom_expr_): Change treatment of Hollerith.
4094 * data.c (ffedata_gather_): Handle opANY in opCONVERT.
4096 * expr.c (ffeexpr_token_apostrophe_): Issue FFEBAD_NULL_CHAR_CONST
4097 warning as necessary.
4098 (ffeexpr_token_name_rhs_): Set context for args to intrinsic
4099 so that assignment-like concatenation is allowed for ICHAR(),
4100 IACHAR(), and LEN() intrinsics.
4101 (ffeexpr_reduced_*_): Say "an array" instead of "an entity" in
4102 diagnostics, since it's more informative.
4103 (ffeexpr_finished_): For many contexts, check for null expression
4104 and array before trying to do a conversion, to avoid redundant
4107 * g77.1: Fix typo for preprocessed suffix (.F, not .f).
4109 * global.c (ffeglobal_init_common): Warn if initializing
4111 (ffeglobal_pad_common): Enable code to warn if initial
4113 (ffeglobal_size_common): Complain if enlarging already-
4114 initialized common, since it won't work right anyway.
4116 * intrin.c: Add IMAG() intrinsic.
4117 (ffeintrin_check_loc_): Allow opSUBSTR in LOC().
4119 * intrin.def: Add IMAG() intrinsic.
4121 * lex.c: Don't report FFEBAD_NULL_CHAR_CONST errors.
4123 * sta.c, sta.h, stb.c: Changes to clean up error messages (see
4126 * stb.c (ffestb_R100113_): Issue FFEBAD_NULL_CHAR_CONST
4127 warning as necessary.
4129 * stc.c (ffestc_shriek_do_): Don't try to reference doref_line
4130 stuff in ANY case, since it won't be valid.
4131 (ffestc_R1227): Allow RETURN in main program unit, with
4132 appropriate warnings/errors.
4133 (ffestc_subr_format_): Array of any type is a CHAREXPR (F77 C5).
4135 * ste.c (ffeste_begin_doiter_): Couple of fixes to accurately
4136 determine if loop never executes.
4138 * target.c (ffetarget_convert_*_hollerith_): Append spaces,
4139 not zeros, to follow F77 Appendix C, and to warn when
4140 truncation of non-blanks done.
4141 (ffetarget_convert_*_typeless): Rewrite to do typeless
4142 conversions properly, and warn when truncation done.
4143 (ffetarget_print_binary, ffetarget_print_octal,
4144 ffetarget_print_hex): Rewrite to use new implementation of
4146 (ffetarget_typeless_*): Rewrite to use new implementation
4147 of typeless, and to warn about overflow.
4149 * target.h (ffetargetTypeless): New implementation of
4152 * type.h, type.c (ffetype_size_typeless): Remove (incorrect)
4153 implementation of this function and its extern.
4155 Sun Mar 5 18:46:42 1995 Craig Burley (burley@gnu.ai.mit.edu)
4157 * BUGS: Clarify that constant handling would also fix lack of
4158 adequate IEEE-754/854 support to some degree, and typeless
4159 and non-decimal constants.
4161 * com.c (ffecom_type_permanent_copy_): Comment out to avoid
4163 (duplicate_decls): New function a la gcc/c-decl.c.
4164 (pushdecl): Use duplicate_decls to decide whether to return
4165 existing decl or new one, instead of always returning existing
4167 (ffecom_expr_): opPERCENT_LOC now supports CHARACTER arguments.
4168 (ffecom_init_0): Give f2c I/O code 0 for basictypeANY/kindtypeANY.
4169 (ffecom_sym_transform_): For adjustable arrays, pass low bound
4170 through variable_size in case its primaries are changed (950302-1.f).
4172 * com.h: More decls that belong in tree.h &c.
4174 * data.c (ffedata_eval_integer1_): Fix opPAREN case to not
4175 treat value of expression as an error code.
4177 * expr.c (ffeexpr_finished_): Allow opSUBSTR in contextLOC case.
4179 * proj.c: Add "const" as appropriate.
4181 Mon Feb 27 10:04:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
4183 * bad.def (FFEBAD_BAD_SUBSTR): Fix bad grammar in message.
4185 Fri Feb 24 16:21:31 1995 Craig Burley (burley@gnu.ai.mit.edu)
4187 * Version 0.5.13 released.
4189 * INSTALL: Warn that f/zzz.o will compare differently between
4190 stages, since it puts the __TIME__ macro into a string.
4192 * com.c (ffecom_sym_transform_): Transform kindFUNCTION/whereDUMMY
4193 to pointer-to-function, not function.
4194 (ffecom_expr_): Use ffecom_arg_ptr_to_expr instead of
4195 ffecom_char_args_ to handle comparison between CHARACTER
4196 types, so either operand can be a CONCATENATE.
4197 (ffecom_transform_common_): Set size of initialized common area
4198 to global (largest-known) size, even though size of init might
4201 * equiv.c (ffeequiv_offset_): Check symbol info for ANY.
4203 * expr.c (ffeexpr_find_close_paren_, ffeexpr_nil_*): New functions
4204 to handle following the contour of a rejected expression, so
4205 statements like "PRINT(I,I,I)=0" don't cause the PRINT statement
4206 code to get the second passed back to it as if there was a
4207 missing close-paren before it, the comma causing the PRINT code
4208 to confirm the statement, resulting in an ambiguity vis-a-vis
4209 the let statement code.
4210 Use the new ffecom_find_close_paren_ handler when an expected
4211 close-paren is missing.
4212 (ffeexpr_isdigits_): New function, use in all places that
4213 currently use isdigit in repetitive code.
4214 (ffeexpr_collapse_symter): Collapse to ANY if init-expr is ANY,
4215 so as to avoid having symbol get "transformed" if used to
4217 (ffeexpr_token_real_, ffeexpr_token_number_real_): Don't issue
4218 diagnostic about exponent, since it'll be passed along the
4219 handler path, resulting in a diagnostic anyway.
4220 (ffeexpr_token_apos_char_): Use consistent handler path
4221 regardless of whether diagnostics inhibited.
4222 (ffeexpr_token_name_apos_name_): Skip past closing quote/apos
4223 even if not a match or other diagnostic issued.
4224 (ffeexpr_sym_impdoitem_): Exec-transition local SEEN symbol.
4226 * lex.c (ffelex_image_char_): Set ffelex_saw_tab_ if TAB
4227 seen, not if anything other than TAB seen!
4229 * stc.c (ffestc_R537_item): If source is ANY but dest isn't,
4230 set dest symbol's init expr to ANY.
4231 (ffestc_R501_attrib, ffestc_R522, ffestc_R522start): Complain
4232 about conflict between "SAVE" by itself and other uses of
4233 SAVE only in pedantic mode.
4235 * ste.c (ffeste_R1212): Fix loop over labels to always
4236 increment caseno, to avoid pushcase returning 2 for duplicate
4237 values when one of the labels is invalid.
4239 Thu Feb 23 12:42:04 1995 Craig Burley (burley@gnu.ai.mit.edu)
4241 * Version 0.5.12 released.
4243 * Make-lang.in (f77.install-common): Add "else true;" before outer
4244 "fi" per Makefile.in patch.
4246 * Makefile.in (dircheck): Add "else true;" before "fi" per
4247 patch from chs1pm@surrey.ac.uk.
4249 * com.c (ffecom_push_tempvar): If type desired is ERROR_MARK,
4250 return error_mark_node, to avoid crash that results from
4251 making a VAR_DECL with error_mark_node as its type.
4253 * ste.c (ffeste_begin_iterdo_): Convert itercount to INTEGER
4254 anytime calculation of number of iterations ends up with type
4255 other than INTEGER (e.g. DOUBLE PRECISION, REAL).
4257 Thu Feb 23 02:48:38 1995 Craig Burley (burley@gnu.ai.mit.edu)
4259 * Version 0.5.11 released.
4261 * DOC: Explain -fugly-args.
4263 * bad.def (FFEBAD_ACTUALARG): Explain -fugly-args and how to
4264 rewrite code to not require it.
4266 * com.c (ffecom_vardesc_): Handle negative type code, just in
4268 (ffecom_arg_ptr_to_expr): Let ffecom_expr handle hollerith
4269 and typeless constants (move code to ffecom_constantunion).
4270 (ffecom_constantunion): Handle hollerith and typeless constants.
4272 * expr.c (ffecom_finished_): Check -fugly-args in actual-arg
4273 context where hollerith/typeless provided.
4275 * intrin.def (FFEINTRIN_genDFLOAT): Add FFEINTRIN_specDFLOAT.
4276 (FFEINTRIN_specDFLOAT): Add as f2c intrinsic.
4278 * target.h (ffetarget_convert_real[12]_integer,
4279 ffetarget_convert_complex[12]_integer): Pass -1 for high integer
4280 value if low part is negative.
4281 (FFETARGET_defaultIS_UGLY_ARGS): New macro.
4283 * top.c (ffe_is_ugly_args_): New variable.
4284 (ffe_decode_option): Handle -fugly-args and -fno-ugly-args.
4286 * top.h (ffe_is_ugly_args_, ffe_is_ugly_args(),
4287 ffe_set_is_ugly_args()): New variable and macros.
4289 Thu Feb 23 02:48:38 1995 Pedro A M Vazquez (vazquez@iqm.unicamp.br)
4291 * g77.c (sys_errlist): Use const for __FreeBSD__ systems
4294 Wed Feb 22 13:33:43 1995 Craig Burley (burley@gnu.ai.mit.edu)
4296 * Version 0.5.10 released.
4298 * CREDITS: Add Rick Niles.
4300 * INSTALL: Note how to get around lack of makeinfo.
4302 * Make-lang.in (f/proj.h): Remove # comment.
4304 * Makefile.in (f/proj.h): Remove # comment.
4306 * com.c (ffecom_expr_): Simplify opFUNCREF/opSUBRREF conversion.
4307 (ffecom_sym_transform_): For whereGLOBAL and whereDUMMY
4308 kindFUNCTION, use ffecom_tree_fun_type[][] only for non-constant
4309 (non-statement-function) f2c functions.
4310 (ffecom_init_0): ffecom_tree_fun_type[][] and _ptr_to_*_* are
4311 really f2c-interface arrays, so use base type void for COMPLEX
4314 Tue Feb 21 19:01:18 1995 Dave Love <d.love@dl.ac.uk>
4316 * Make-lang.in (f77.install-common): Expurgate the test for and
4317 possible installation of f2c in line with elsewhere. Seems to have
4318 been missing a semicolon anyhow!
4320 Tue Feb 21 11:45:25 1995 Craig Burley (burley@gnu.ai.mit.edu)
4322 * Version 0.5.9 released.
4324 * Make-lang.in (f/proj.h): touch file to register update,
4325 because the previous commands won't necessarily modify it.
4327 * Makefile.in (f/proj.h): touch file to register update,
4328 because the previous commands won't necessarily modify it.
4330 * Makefile.in (f/str-*.h, f/str-*.j): Explicitly specify
4331 output file names, so these targets go in build, not source,
4334 * bits.c, bits.h: Switch to valid ANSI C replacement for
4337 * com.c (ffecom_expr_): Add assignp arg to support ASSIGN better.
4338 If assignp is TRUE, use different tree for FFEBLD_opSYMTER case.
4339 (ffecom_sym_transform_assign_): New function.
4340 (ffecom_expr_assign): New function.
4341 (ffecom_expr_assign_w): New function.
4343 * com.c (ffecom_f2c_make_type_): Do make_signed_type instead
4344 of make_unsigned_type throughout.
4346 * com.c (ffecom_finish_symbol_transform_): Expand scope of
4347 commented-out code to probably produce faster compiler code.
4349 * com.c (ffecom_gen_sfuncdef_): Push/pop calltemps so
4350 COMPLEX works right.
4351 Remove obsolete comment.
4353 * com.c (ffecom_start_progunit_): If non-multi alt-entry
4354 COMPLEX function, primary (static) entry point returns result
4355 directory, not via extra arg -- to agree with ffecom_return_expr
4357 Pretransform all symbols so statement functions are defined
4358 before any code emitted.
4360 * com.c (ffecom_finish_progunit): Don't posttransform all
4361 symbols here -- pretransform them instead.
4363 * com.c (ffecom_init_0): Don't warn about possible ASSIGN
4364 crash, as this shouldn't happen now.
4366 * com.c (ffecom_push_tempvar): Fix to handle temp vars
4367 pushed while context is a statement (nested) function, and
4368 add appropriate commentary.
4370 * com.c (ffecom_return_expr): Check TREE_USED to determine
4371 where return value is unset.
4373 * com.h (struct _ffecom_symbol_): Add note about length_tree
4374 now being used to keep tree for ASSIGN version of symbol.
4376 * com.h (ffecom_expr_assign, ffecom_expr_assign_rw): New decls.
4377 (error): Add this prototype for back-end function.
4379 * fini.c (main): Grab input, output, and include names
4380 directly off the command line instead of making the latter
4381 two out of the first.
4383 * lex.c: Improve tab handling for both fixed and free source
4384 forms, and ignore carriage-returns on input, while generally
4385 improving the code. ffelex_handle_tab_ has been renamed and
4386 reinvented as ffelex_image_char_, among other things.
4388 * malloc.c, malloc.h: Switch to valid ANSI C replacement for
4389 ARRAY_ZERO, and kill the full number of bytes in pools and
4392 * proj.h.in (ARRAY_ZERO, ARRAY_ZERO_SIZE): Remove.
4394 * ste.c (ffeste_io_cilist_, ffeste_io_icilist_, ffeste_R838,
4395 ffeste_R839): Issue diagnostic if a too-narrow variable used in an
4396 ASSIGN context despite changes to this code and code in com.c.
4398 * where.c, where.h: Switch to valid ANSI C replacement for
4401 Fri Feb 17 03:35:19 1995 Craig Burley (burley@gnu.ai.mit.edu)
4403 * Version 0.5.8 released.
4405 * INSTALL: In quick-build case, list g77 target first so g77
4406 gets installed. Also, explain that gcc gets built and installed
4407 as well, even though this isn't really what we want (and maybe
4408 we'll find a way around this someday).
4410 Fri Feb 17 02:35:41 1995 Craig Burley (burley@gnu.ai.mit.edu)
4412 * Version 0.5.7 released.
4414 * Makefile.in (CONFIG_H, HCONFIG_H, TCONFIG_H, TM_H): Remove
4415 ../ prefix in front of .h files, since they're in the cd.
4417 Fri Feb 17 01:50:48 1995 Craig Burley (burley@gnu.ai.mit.edu)
4419 * Version 0.5.6 released.
4421 Thu Feb 16 20:26:54 1995 Craig Burley (burley@gnu.ai.mit.edu)
4423 * ../README.g77: Remove description of g77 as "not-yet-published".
4425 * CREDITS: More changes.
4427 * Make-lang.in (G77STAGESTUFF): Remove cktyps stuff.
4429 * Makefile.in (CONFIG_H, HCONFIG_H, TCONFIG_H, TM_H): Don't
4430 prefix gcc dir with $(srcdir) since these don't live there,
4431 they are created in the build dir by gcc's configure. Add
4432 a note explaining what these macros are about.
4433 Update dependencies via deps-kinda.
4435 * README.NEXTSTEP: Credit Toon, and per his request, add his
4438 * com.h (FFECOM_DETERMINE_TYPES): #include "config.j".
4440 * config.j, convert.j, flags.j, hconfig.j, rtl.j, tconfig.j,
4441 tm.j, tree.j: Don't #include if already done.
4443 * convert.j: #include "tree.j" first, as convert.h clearly depends
4444 on trees being defined.
4446 * rtl.j: #include "config.j" first, since there's some stuff
4447 in rtl.h that assumes it has been #included.
4449 * tree.j: #include "config.j" first, or real.h makes inconsistent
4450 decision about return type of ereal_atof, leading to bugs, and
4451 because tree.h/real.h assume config.h already included.
4453 Wed Feb 15 14:40:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
4455 * Version 0.5.5 released.
4457 * Copyright notices updated to be FSF-style.
4459 * INSTALL: Some more clarification regarding building just f77.
4461 * Make-lang.in (F77_SRCS): Update wrt changing some .h to .j.
4462 (install-libf77): Fix typo in new parenthetical note.
4464 * Makefile.in (f/*.o): Update.
4465 (CONFIG_H, CONVERT_H, FLAGS_H, GLIMITS_H, HCONFIG_H, RTL_H,
4466 TCONFIG_H, TM_H, TREE_H): Update/new symbols.
4467 (deps-kinda): More fixes wrt changing some .h to .j.
4468 Document and explain this rule a bit better.
4469 Accommodate changes in output of gcc -MM.
4471 * *.h, *.c: Change #include's so proj.h not assumed to #include
4472 malloc.h or config.h (now config.j), and so new .j files are
4473 used instead of old .h ones.
4475 * com.c (ffecom_init_0): Use FLOAT_TYPE_SIZE for f2c's
4476 TYLONG/TYLOGICAL type codes, to get g77 working on Alpha.
4478 * com.h: Make all f2c-related integral types "int", not "long
4481 * config.j, convert.j, flags.j, glimits.j, hconfig.j, rtl.j,
4482 tconfig.j, tm.j, tree.j: New files wrapping around gbe
4485 * config.h, convert.h, flags.h, glimits.h, hconfig.h, rtl.h,
4486 tconfig.h, tm.h, tree.h: Deleted so new .j files
4487 can #include the gbe files directly, instead of using "../",
4488 and thus do better with various kinds of builds.
4490 * proj.h: Delete unused NO_STDDEF and related stuff.
4492 Tue Feb 14 08:28:08 1995 Craig Burley (burley@gnu.ai.mit.edu)
4494 * BUGS: Remove item #12, cross-compiling & autoconf scripts
4495 reportedly expected to work properly (according to d.love).
4497 * INSTALL: Add explanation of d.love's patch to config-lang.in.
4498 Add explanation of how to install just g77 when gcc already installed.
4499 Add note about usability of "-Wall". Add note about bug-
4502 * Make-lang.in ($(srcdir)/f/conf-proj): Add comment about why
4504 (install-libf77): Echo parenthetical note to user about how to do
4505 just the (aborted) libf2c installation.
4506 (deps-kinda): Update to work with new configuration/build stuff.
4508 * bad.c (ffebad_finish): Put capitalized "warning:" &c message
4509 as prefix on any diagnostic without pointers into source.
4511 * bad.def (FFEBAD_TOO_BIG_INIT): Add this warning message.
4513 * config-lang.in: Add Dave Love's patch to catch case where
4514 back-end patches not applied and abort configuration.
4516 * data.c (ffedata_gather_, ffedata_value_): Warn when about
4517 to initialize a large aggregate area, due to design flaw resulting
4518 in too much time/space used to handle such cases.
4519 Use COMMON area name, and first notice of symbol, for multiple-
4520 initialization diagnostic, instead of member symbol and unknown
4522 (FFEDATA_sizeTOO_BIG_INIT_): New macro per above.
4524 Mon Feb 13 13:54:26 1995 Dave Love <d.love@dl.ac.uk>
4526 * Make-lang.in (F77_SRCS): Use $(srcdir)/f/proj.h.in, not
4527 $(srcdir)/f/proj.h for build outside srcdir.
4529 Sun Feb 12 13:37:11 1995 Craig Burley (burley@gnu.ai.mit.edu)
4531 * ../README.g77: Clarify procedures for unpacking, add asterisks
4532 to mark important things the user must do.
4534 * Fix dates in/add dates to ../README.g77, BUGS, CREDITS, DOC,
4535 INSTALL, PROJECTS, README.
4537 Sun Feb 12 00:26:10 1995 Craig Burley (burley@gnu.ai.mit.edu)
4539 * Version 0.5.4 released.
4541 * Make-lang.in (f/proj.h): Reproduce this rule here from
4543 ($(srcdir)/f/conf-proj): Put autoconf's stdout in temp file
4544 conf-proj.out, then mv to conf-proj only if successful, so
4545 conf-proj not touched if autoconf not installed.
4547 * Makefile.in ($(srcdir)/conf-proj): See Make-lang.in's similar
4550 Sat Feb 11 20:56:02 1995 Craig Burley (burley@gnu.ai.mit.edu)
4552 * BUGS: Clarify some bugs.
4554 * DOC: Many improvements and fixes.
4556 * README: Move bulk of text, edited, to ../README.g77, and
4557 replace with pointer to that file.
4559 * com.c (ffecom_init_0): Comment out warning about sizeof(ftnlen)
4560 as per ste.c change. Add text about ASSIGN to help user understand
4561 what is being warned about.
4563 * conf-proj.in: Fix typos in comments.
4565 * proj.h.in: Add ARRAY_ZERO_SIZE to parallel malloc.h's version,
4566 in case it proves to be needed.
4568 * ste.c: Comment out assertions requiring sizeof(ftnlen) >=
4569 sizeof(char *), in the hopes that overflow will never happen.
4570 (ffeste_R838): Change assertion to fatal() with at least
4571 partially helpful message.
4573 Sat Feb 11 12:38:00 1995 Craig Burley (burley@gnu.ai.mit.edu)
4575 * com.c (ffecom_vardesc_): Crash if typecode is -1.
4577 * ste.c (ffeste_io_dolio_): Crash if typecode is -1.
4579 Sat Feb 11 09:51:57 1995 Craig Burley (burley@gnu.ai.mit.edu)
4581 * ste.c: In I/O code tests for item arrayness, sort of revert
4582 to much earlier code that tests original exp, but also check
4583 in newer way just in case. Newer way alone treated FOO(1:40)
4584 as an array, not sure why older way alone didn't work, but I
4585 think maybe it was when diagnosed code was involved, and
4586 since there are now checks for error_mark_node, maybe the old
4587 way alone would work. But better to be safe; both original
4588 ffebld exp _and_ the transformed tree must indicate an array
4589 for the size-determination code to be used, else just 1/2 elements
4590 assumed. And this text is for EMACS: (foo at bar).
4592 Fri Feb 10 11:05:50 1995 Craig Burley (burley@gnu.ai.mit.edu)
4594 * ste.c: In many cases, surround statement-expansion code
4595 with ffecom_push_calltemps () and ffecom_pop_calltemps ()
4596 so COMPLEX-returning functions can have temporaries pushed
4597 in "auto-pop" mode and have them auto-popped at the end of
4600 Wed Feb 8 14:35:10 1995 Dave Love <d.love@dl.ac.uk>
4602 * runtime/f2c.h.in (ftnlen, ftnint): Make same size as integer.
4604 * runtime/libI77/err.c (f_init): Thinko in MISSING_FILE_ELEMS
4606 * runtime/libI77/wrtfmt.c (mv_cur): Likewise.
4607 * runtime/libI77/wsfe.c (x_putc): Likewise.
4609 * runtime/libF77/signal_.c (signal_): Return 0 (this is a
4612 * Makefile.in (f/proj.h): Depend on com.h.
4613 * Make-lang.in (include/f2c.h): Likewise (and proj.h).
4614 (install-libf77): Also install f2c.h.
4616 * runtime/libI77/Makefile.in (*.o): Add f2c.h dependency.
4617 * runtime/libF77/Makefile.in: Likewise.
4619 Wed Feb 8 13:56:47 1995 Craig Burley (burley@gnu.ai.mit.edu)
4621 * stc.c (ffestc_R501_item): Don't reset kind/where to NONE when
4622 setting basictype/kindtype info for symbol, or especially
4623 its function/result twin, because kind/where might not be NONE.
4625 Tue Feb 7 14:47:26 1995 Dave Love <d.love@dl.ac.uk>
4627 * Make-lang.in (include/f2c.h:): Set shell variable src more
4628 robustly (independent of whether srcdir is relative or absolute).
4629 * Makefile.in (f/proj.h:): Likewise.
4631 * conf-proj.in: Check need for LAME_ASSERT. Fix indentation in
4632 check for LAME_STDIO (cosmetic only with ANSI C).
4634 * com.h: Extra ...SIZE stuff taken from com.c.
4636 * com.c (FFECOM_DETERMINE_TYPES): Define before including com.h.
4637 (BITS_PER_WORD etc.) Remove and use conditional definitions to com.h.
4639 * runtime/configure.in: #define FFECOM_DETERMINE_TYPES for com.h in
4640 f2c type determination.
4642 * tm.h: Remove (at least pro tem) because of relative path and use
4645 * Make-lang.in (include/f2c.h:): Set shell variable src more
4646 robustly (independent of whether srcdir is relative or absolute).
4647 * Makefile.in (f/proj.h:): Likewise.
4649 Mon Feb 6 19:58:32 1995 Dave Love <d.love@dl.ac.uk>
4651 * g77.c (append_arg): Use K&R declaration for, e.g. SunOS4 build.
4653 Fri Feb 3 20:33:14 1995 Craig Burley (burley@gnu.ai.mit.edu)
4655 * g77.c (main): Treat -l like filename in terms of -x handling.
4656 Rewrite arglist mechanism for ease of maintenance.
4657 Make sure every -lf2c is followed by -lm and vice versa.
4659 * Make-lang.in: Put complete list of sources in F77_SRCS def
4660 so changing a .h file, for example, causes rebuild.
4662 * Makefile.in: Change test for nextstep to m68k-next-nextstep* so
4663 all versions of nextstep on m68k get the necessary flag.
4665 Fri Feb 3 19:10:32 1995 Dave Love <d.love@dl.ac.uk>
4667 * INSTALL: Note about possible conflict with existing libf2c.a and
4670 * Make-lang.in (f77.distclean): Tidy and move deletion of
4671 f/config.cache to mostlyclean.
4672 (install-libf77): Test for $(libdir)/libf2c.* and barf if found
4673 unless F2CLIBOK defined.
4675 * runtime/Makefile.in (all): Change path to include directory (and
4677 (INCLUDES): Remove (unused/misleading).
4678 (distclean): Include f2c.h.
4679 (clean): Include config.cache.
4681 * runtime/libF77/Makefile.in (.SUFFIXES): Correct typo.
4682 (ALL_CFLAGS) Fix up include search path to find f2c.h in top level
4684 (all): Depend on f2c.h.
4685 * runtime/libI77/Makefile.in (.SUFFIXES): Likewise.
4687 Thu Feb 2 17:17:06 1995 Dave Love <d.love@dl.ac.uk>
4689 * INSTALL: Note about --srcdir and GNU make.
4691 * runtime/f2c.h.in (Pad_UDread, ALWAYS_FLUSH): Reomve the #defines
4694 * runtime/configure.in (Pad_UDread, ALWAYS_FLUSH): Define these
4695 here, not in f2c.h as they'r eonly relevant for building.
4696 * runtime/configure: Regenerated.
4698 * config-lang.in: Warn about using GNU make outside source tree
4699 since I can't get Irix5 or SunOS4 makes to work in this case.
4701 * Makefile.in (VPATH): Don't set it here.
4702 (srcdir): Make it the normal `.' (overridden) at top level.
4703 (all.indirect): New dependency `dircheck'.
4705 (dircheck): New target for foolproofing.
4706 (f/proj.h:): Change finding source.
4707 (CONFIG_H): Don't use this as the relative path in the include loses
4708 f builddir != srcdir.
4710 * config.h: Remove per CONFIG_H change above.
4712 * Make-lang.in (F77_FLAGS_TO_PASS): Remove GCC_FOR_TARGET.
4713 (f771:): Pass VPATH, srcdir to sub-make.
4714 (f/Makefile:): New target.
4715 (stmp-int-hdrs): new variable for cheating build.
4716 (f77-runtime:): Alter GCC_FOR_TARGET treatment.
4717 (include/f2c.h f/runtime/Makefile:) Likewise.
4718 (f77-runtime-unsafe:): New (cheating) target.
4720 Thu Feb 2 12:09:51 1995 Craig Burley (burley@gnu.ai.mit.edu)
4722 * BUGS: Update regarding losing EQUIVALENCE members in -g, and
4723 regarding RS/6000 problems in the back end.
4725 * CREDITS: Make some changes as requested.
4727 * com.c (ffecom_member_trunk_): Remove unused static variable.
4728 (ffecom_finish_symbol_transform_): Improve comments.
4729 (ffecom_let_char_): Fix size of temp address-type var.
4730 (ffecom_member_phase2_): Try fixing problem fixed by change
4731 to ffecom_transform_equiv_ (f_m_p2_ function currently not used).
4732 (ffecom_transform_equiv_): Remove def of unused static variable.
4733 Comment-out use of ffecom_member_phase2_, until problems with
4735 (ffecom_push_tempvar): Fix assertion to not crash okay code.
4737 * com.h: Remove old, commented-out code.
4738 Add prototype for warning() in back end.
4740 * ste.c (ffeste_io_dofio_, ffeste_io_dolio_, ffeste_io_douio_,
4741 ffeste_io_icilist_): Check correct type of variable for arrayness.
4743 Sun Jan 29 14:41:42 1995 Dave Love <d.love@dl.ac.uk>
4745 * BUGS: Remove references to my configure bugs; add another.
4747 * runtime/Makefile.in (AR_FLAGS): Provide default value.
4749 * runtime/f2c.h.in (integer, logical): Take typedefs from
4750 F2C_INTEGER configuration parameter again.
4751 (NON_UNIX_STDIO): don't define it.
4753 * runtime/configure.in: Bring type checks for f2c.h in line with
4755 (MISSING_FILE_ELEMS): New variable to determine whether the relevant
4756 elements of the FILE struct exist, independent of NON_UNIX_STDIO.
4757 * runtime/libI77/{err,wrtfmt,wsfe}.c (MISSING_FILE_ELEMS): Use new
4760 * config-lang.in: Comment out more of f2c rules (c.f. Make-lang.in).
4761 (This stuff is relevant iff you gave configure --enable-f2c.)
4762 Create f/runtime directory tree iff not building in source
4765 * Makefile.in (srcdir): Append slash so we get the right value when
4766 not building in the source directory. This is a consequence of not
4767 building the `f' sources in `f'.
4768 (VPATH): Override configure's value for reasons above.
4769 (f/proj.h f/conf-proj): New rules to build proj.h by
4772 * proj.h: Rename to proj.h.in for autoconfiguration.
4773 * proj.h.in: New as above.
4774 * conf-proj conf-proj.in: New files for autoconfiguration.
4776 * Make-lang.in (include/f2c.h f/runtime/Makefile:): Change the order
4777 of setting the sh variables so that the right GCC_FOR_TARGET is
4779 (f77.*clean:) Add products of new configuration files and make sure
4780 all the *clean targets do something (unlike the ones in
4783 * com.h (FFECOM_f2cINTEGER, FFECOM_f2cLOGICAL): Define as long or
4784 int appropriately to ensure sizeof(real) == sizeof(integer).
4786 * PROJECTS: Library section.
4788 * runtime/libI77/endfile.c: Don't #include sys/types.h conditional
4789 on NON_UNIX_STDIO since rawio.h needs size_t.
4790 * runtime/libI77/uio.c: #include <sys/types.h> for size_t if not
4793 Wed Jan 25 03:31:51 1995 Craig Burley (burley@gnu.ai.mit.edu)
4795 * Version 0.5.3 released.
4799 * Make-lang.in: Comment out rules for building f2c itself (f/f2c/).
4803 * com.c (ffecom_init_0): Warn if ftnlen or INTEGER not big enough
4806 * gbe/2.6.2.diff: Update.
4808 Mon Jan 23 17:10:49 1995 Craig Burley (burley@gnu.ai.mit.edu)
4816 Make-lang.in: Remove cktyps stuff.
4817 Makefile.in: Remove cktyps stuff.
4819 * DOC: Add info on changes for 0.5.3.
4821 * bad.c: Put "warning:" &c on diagnostic messages.
4822 Don't output informational messages if warnings disabled.
4824 Thu Jan 19 12:38:13 1995 Craig Burley (burley@gnu.ai.mit.edu)
4826 * g77.c: Avoid putting out useless "-xnone -xf77" pairs so
4827 larger command lines can be accommodated.
4828 Recognize both `-xlang' and `-x lang'.
4829 Recognize `-xnone' and `-x none' to mean what it does, instead
4830 of treating "none" as any other language.
4831 Some minor, slight improvements in the way args are handled
4832 (hopefully for clearer, more maintainable code), including
4833 consistency checks on arg count just in case.
4835 Wed Jan 18 16:41:57 1995 Craig Burley (burley@gnu.ai.mit.edu)
4837 * DOC: Explain -fautomatic better.
4839 * INSTALL: Describe libf2c.a better.
4841 * Make-lang.in, Makefile.in: Build f771 &c with gcc/ as cd instead
4842 of gcc/f/ so debugging info is better (source file tracking).
4843 Add new source file type.c.
4845 * Makefile.in: For nextstep3, link f771 with -segaddr __DATA
4846 6000000. Fix typo. Change deps-kinda target to handle building
4847 from gcc/. Update dependencies.
4849 * bld-op.def, bld.h, bld.c: Remove opBACKEND and all related
4851 Remove consistency tests that cause compiler warnings.
4853 * cktyps.c: Remove all typing checking.
4855 * com-rt.def: Change all rttypeFLOAT_ intrinsics to rttypeDOUBLE_,
4856 to precisely match how they're declared in libf2c.
4858 * com.h, com.c: Revise to more elegantly track related stuff
4859 in the version of f2c.h used to build libf2c.
4861 * com.c: Increase FFECOM_sizeMAXSTACKITEM, and if 0 or undefined
4862 when checked to determine where to put entity, treat as infinite.
4863 Rewrite temporary mechanism to be based on trees instead of
4864 ffeinfo stuff, and make it much simpler. Change interface
4866 Fixes to better track types of things, make appropriate
4867 conversions, etc. E.g. when making an arg for a libf2c
4868 function, make sure it's of the right type (such as ftnlen).
4869 Delete opBACKEND transformation code.
4870 (ffecom_init_0): Smoother initialization of types, especially
4871 paying attention to using consistent rules for making INTEGER,
4872 REAL, DOUBLE PRECISION, etc., and for deciding their "*N"
4873 and kind values that will work across all g77 platforms.
4874 No longer require per-target configuration info in target.h
4875 or config/*/*; use new type module to store size, alignment.
4876 (ffecom_member_phase2): Declare COMMON/EQUIVALENCE group members
4877 so debugger sees them.
4878 (ffecom_finish_progunit): Transform all symbols in program unit,
4879 so -g will show they all exist.
4881 * expr.c (ffeexpr_collapse_substr): Handle strange substring
4884 * info.h, info.c: Provide connection to new type module.
4885 Remove tests that yield compiler warnings.
4887 * intrin.c (ffeintrin_is_intrinsic): Properly handle deleted
4890 * lex.c (ffelex_file_fixed): Remove redundant/buggy code.
4892 * stc.c (ffestc_kindtype_kind_, ffestc_kindtype_star_): Replace
4893 boring switch stmt with simple call to new type module. This
4894 sort of thing is a reason to get up in the morning.
4896 * ste.c: Update to handle new interface for
4897 ffecom_push/pop_tempvar.
4898 Fixes to better track types of things.
4899 Fixes to not crash for certain diagnosed constructs.
4900 (ffeste_begin_iterdo_): Check only constants for overflow to avoid
4901 spurious diagnostics.
4902 Don't convert larger integer (say, INTEGER*8) to canonical integer
4903 for iteration count.
4905 * stw.h: Track DO iteration count temporary variable.
4907 * symbol.c: Remove consistency tests that cause compiler warnings.
4909 * target.c (ffetarget_aggregate_info): Replace big switch with
4910 little call to new type module.
4911 (ffetarget_layout): Remove consistency tests that cause
4913 (ffetarget_convert_character1_typeless): Pick up length of
4914 typeless type from new type module.
4916 * target.h: Crash build if target float bit pattern cannot be
4917 precisely determined.
4918 Remove all the type cruft now determined by ffecom_init_0
4919 at invocation time and maintained in new type module.
4920 Put casts on second arg of all REAL_VALUE_TO_TARGET_DOUBLE
4921 uses so compiler warnings avoided (requires target float bit
4922 pattern to be precisely determined, hence code to crash build).
4924 * top.c: Add inits/terminates for new type module.
4926 * type.h, type.c: New module.
4928 * gbe/2.6.2.diff: Remove all patches to files in gcc/config/
4929 directory and its subdirectories.
4931 Mon Jan 9 19:23:25 1995 Dave Love <d.love@dl.ac.uk>
4933 * com.h (FFECOM_F2C_INTEGER_TYPE_NODE): Define and use instead of
4934 long_integer_type_node where appropriate.
4936 Tue Jan 3 14:56:18 1995 Dave Love <d.love@dl.ac.uk>
4938 * com.h: Make ffecom_f2c_logical_type_node long, consistent with
4941 Fri Dec 2 20:07:37 1994 Dave Love <d.love@dl.ac.uk>
4943 * config-lang.in (stagestuff): Add f2c conditionally.
4944 * Make-lang.in: Add f2c and related targets.
4945 * f2c: Add the directory.
4947 Fri Nov 25 22:17:26 1994 Dave Love <d.love@dl.ac.uk>
4949 * Makefile.in (FLAGS_TO_PASS): pass $(CROSS)
4950 * Make-lang.in: more changes to runtime targets
4952 Thu Nov 24 18:03:21 1994 Dave Love <d.love@dl.ac.uk>
4954 * Makefile.in (FLAGS_TO_PASS): define for sub-makes
4956 * g77.c (main): change f77-cpp-output to f77-cpp-input (.F files)
4958 Wed Nov 23 15:22:53 1994 Dave Love <d.love@dl.ac.uk>
4960 * bad.c (ffebad_finish): kluge to fool emacs19 into finding errors:
4961 add trailing space to <file>:<line>:
4963 Tue Nov 22 11:30:50 1994 Dave Love <d.love@dl.ac.uk>
4965 * runtime/libF77/signal_.c (RETSIGTYPE): added
4967 Mon Nov 21 13:04:13 1994 Dave Love <d.love@dl.ac.uk>
4969 * Makefile.in (compiler): add runtime
4971 * config-lang.in (stagestuff): add libf2c.a to stagestuff
4974 G77STAGESTUFF <- MORESTAGESTUFF
4975 f77-runtime: new target, plus supporting ones
4977 * runtime: add the directory, containing libI77, libF77 and autoconf
4982 * g77.1: minor fixes
4984 Thu Nov 17 15:18:05 1994 Craig Burley (burley@gnu.ai.mit.edu)
4986 * Version 0.5.2 released.
4988 * bad.def: Modify wording of FFEBAD_UNIMPL_STMT to indicate
4989 that it covers a wide array of possible problems (that, someday,
4990 should be handled via separate diagnostics).
4992 * lex.c: Allow $ in identifiers if -fdollar-ok.
4993 * top.c: Support -fdollar-ok.
4994 * top.h: Support -fdollar-ok.
4995 * target.h: Support -fdollar-ok.
4996 * DOC: Describe -fdollar-ok.
4998 * std.c (ffestd_R1229_finish): Fix bug so stand-alone build works.
4999 * ste.c (ffeste_R819A): Fix bug so stand-alone build works.
5001 * Make: Improvements for stand-alone build.
5003 * Makefile.in: Fix copyright text at top of file.
5005 * LINK, SRCS, UNLINK: Removed. Not particularly useful now that
5006 g77 sources live in their own subdirectory.
5008 * g77.c (main): Cast arg to bzero to avoid warning. (This is
5009 identical to Kenner's fix to cp/g++.c.)
5011 * gbe/: New subdirectory, to contain .diff files for various
5012 versions of the GNU CC back end.
5014 * gbe/README: New file.
5015 * gbe/2.6.2.diff: New file.
5017 Tue Nov 8 10:23:10 1994 Dave Love <d.love@dl.ac.uk>
5019 * Make-lang.in: don't install as f77 as well as g77 to avoid
5020 confusion with system's compiler (especially while testing)
5022 * g77.c (main): use -lf2c and -lm; fix sense of test for .f/.F files
5024 Fri Oct 28 09:45:00 1994 Craig Burley (burley@gnu.ai.mit.edu)
5026 * Version 0.5.1 released.
5028 * gcc.c: Invoke f771 instead of f-771.
5030 Fri Oct 28 02:00:44 1994 Craig Burley (burley@gnu.ai.mit.edu)
5032 * Version 0.5.0 released.
5034 Fri Oct 14 15:03:35 1994 Craig Burley (burley@gnu.ai.mit.edu)
5036 * Makefile.in: Handle the Fortran-77 front-end in a subdirectory.
5037 * f-*: Move Fortran-77 front-end to f/*.