1 Mon Jun 15 23:39:24 1998 Craig Burley <burley@gnu.org>
3 * Make-lang.in (f/g77.info): Use -f when removing
4 pre-existing Info files, if any. (This rm command
5 can go away once makeinfo has been changed to delete
6 .info-N files beyond the last one it creates.)
8 * Make-lang.in ($(srcdir)/f/intdoc.texi): Compile
9 using $(INCLUDES) macro to get the new hconfig.h
10 and system.h headers. [Instead of change below.]
12 1998-16-19 Manfred Hollstein <manfred@s-direktnet.de>
14 * Makefile.in ($(srcdir)/f/intdoc.texi): Add $(INCLUDES).
16 Mon Jun 15 22:21:57 1998 Craig Burley <burley@gnu.org>
45 Use toplev.h where appropriate:
56 Conditionalize all dumping/reporting routines so they don't
57 get built for gcc/egcs:
69 Use hconfig.h instead of config.h where appropriate:
70 * Makefile.in (proj-h.o): Compile with -DUSE_HCONFIG.
71 * fini.c: Define USE_HCONFIG before including proj.h.
73 * Makefile.in (deps-kinda): Redirect stderr to stdout,
74 to eliminate diagnostics vis-a-vis g77spec.c.
76 * Makefile.in: Regenerate dependencies via deps-kinda.
78 * lex.c (ffelex_file_fixed, ffelex_file_free): Eliminate
79 apparently spurious warnings about uninitialized variables
80 `c', `column', and so on.
82 Sat Jun 13 03:13:18 1998 Craig Burley <burley@gnu.org>
84 * g77spec.c (lang_specific_driver): Print out egcs
85 version info first, to be compatible with what some
86 test facilities expect.
88 Wed Jun 10 13:17:32 1998 Dave Brolley <brolley@cygnus.com>
90 * top.h (ffe_decode_option): New argc/argv interface.
91 * top.c (ffe_decode_option): New argc/argv interface.
92 * parse.c (yyparse): New argc/argv interface for ffe_decode_option.
93 * com.c (lang_decode_option): New argc/argv interface.
95 Mon Jun 1 19:37:42 1998 Craig Burley <burley@gnu.org>
97 * com.c (ffecom_init_0): Fix setup of INTEGER(KIND=7)
99 * info.c (ffeinfo_type): Don't crash on null type.
100 * expr.c (ffeexpr_fulfill_call_): Don't special-case
101 %LOC(expr) or LOC(expr).
102 Delete FFEGLOBAL_argsummaryPTR.
103 * global.c, global.h: Delete FFEGLOBAL_argsummaryPTR.
105 Thu May 28 21:32:18 1998 Craig Burley <burley@gnu.org>
107 Restore circa-0.5.22 capabilities of `g77' driver:
108 * Make-lang.in (g77spec.o): Depend on f/version.h.
109 (g77version.o): New rule to compile g77 version info.
110 (g77$(exeext)): Depend on and link in g77version.o.
111 * g77spec.c: Rewrite to be more like 0.5.22 version
112 of g77.c, making filtering of command line smarter
113 so mixed Fortran and C (etc.) can be compiled, verbose
114 version info can be obtained, etc.
115 * lang-specs.h (f77-version): New "language" to support
116 "g77 -v" command under new gcc 2.8 regime.
117 * lex.c (ffelex_file_fixed): If -fnull-version, just
118 substitute a "source file" that prints out version info.
119 * top.c, top.h: Support -fnull-version.
121 * lang-specs.h: Use "%O" instead of OO macro to specify
122 object extension. Remove old stringizing cruft.
124 * Make-lang.in (g77.c, g77spec.o, g77.o, g77$(exeext),
125 g77-cross$(exeext), f771,
126 $(srcdir)/f/g77.info, $(srcdir)/f/g77.dvi,
127 $(srcdir)/f/intdoc.texi,
128 f77.install-common, f77.install-info, f77.install-man,
129 f77.uninstall, $(G77STAGESTUFF), f77.stage1, f77.stage2,
130 f77.stage3, f77.stage4, f77.distdir): Don't do anything
131 unless user specified "f77" or "F77" in $LANGUAGES either
132 during configuration or explicitly. For convenience of
133 various tests and to work around lack of the assignment
134 "LANGUAGES=$(BOOT_LANGUAGES)" in the "make stage1" command
135 of "make bootstrap" in gcc, use a touch file named "lang-f77"
136 to communicate whether this is the case.
138 * Make-lang.in (F77_FLAGS_TO_PASS): Delete this macro,
139 replace with minimal expansion of its former self in
140 each of the two instances where it was used.
142 * Makefile.in (HOST_CC): Delete this definition.
144 * com.c (index, rindex): Delete these declarations.
146 * proj.h: (isascii): Delete this.
148 * Make-lang.in (f77.install-common): Warn if `f77-install-ok'
149 flag-file exists, since it no longer triggers any activity.
151 Rename libf2c.a and f2c.h to libg2c.a and g2c.h,
152 normalize and simplify g77/libg2c build process:
153 * Make-lang.in: Remove all support for overwriting
154 /usr/bin/f77 etc., or whatever the actual names are
155 via $(prefix) and $(local_prefix). (g++ overwrites
156 /usr/bin/c++, but then it's often the only C++ compiler
157 on the system; f77 often exists on systems that are
159 (f77.realclean): Remove obsolete target.
160 (g77.c, g77$(exeext)): Minor changes to look more like g++'s
162 (f771): Now built with srcdir=gcc/f, not srcdir=gcc, to be
163 more like g++ and such.
164 (f/Makefile): Removed, as g++ doesn't need this rule.
165 (f77.install-common): No longer install f77, etc.
166 (f77.install-man): No longer install f77.1.
167 (f77.uninstall): No longer uninstall f77, f77.1, etc.
168 (f77.stage1, f77.stage2, f77.stage3, f77.stage4): Do work
169 only if "f77" appears in $(LANGUAGES).
170 (Note: gcc's Makefile.in's bootstrap target should set
171 LANGUAGES=$(BOOT_LANGUAGES) when making the stage1 target.)
172 * Makefile.in: Update vis-a-vis gcc/cp/Makefile.in.
174 (g77-only): Relocate.
175 (all.indirect, f771, *.o): Now assumes current directory
176 is this dir (gcc/f), not the parent directory.
177 (TAGS): Remove "echo 'parse.y,0' >> TAGS ;" line.
178 * config-lang.in: Delete commented-out code.
179 Fix stagestuff definition. Add more stuff to
180 diff_excludes definition. Don't create any directories.
181 Set outputs to f/Makefile, to get variable substition
182 to happen (what does that really do, anyway?!).
183 * g77spec.c: Rename libf2c to libg2c.
185 * com.h: Remove all of the gcc back-end decls,
186 since egcs should have all of them correct.
188 * com.c: Include "proj.h" before anything else,
189 as that's how things are supposed to work.
192 * bad.c: Include "flags.j" here, since some diagnostics
193 check flag_pedantic_errors.
195 * Makefile.in (f/*.o): Rebuild dependencies via
198 * output.j: New source file.
199 * Make-lang.in (F77_SRCS): Update accordingly.
200 * Makefile.in (OUTPUT_H): Ditto.
202 * com.c: Include "output.j" here.
205 Mon May 25 03:34:42 1998 Craig Burley <burley@gnu.org>
207 * com.c (ffecom_expr_): Fix D**I and Z**I cases to
208 not convert (DOUBLE PRECISION) D and (DOUBLE COMPLEX) Z
209 to INTEGER. (This is dead code here anyway.)
211 Sat May 23 06:32:52 1998 Craig Burley <burley@gnu.org>
213 * com.c (ffecom_finish_symbol_transform_): Don't transform
214 statement (nested) functions, to avoid gcc compiling them
215 and thus producing linker errors if they refer to undefined
216 external functions. But warn if they're unused and -Wunused.
217 * bad.def (FFEBAD_SFUNC_UNUSED): New diagnostic.
219 Tue May 19 14:52:41 1998 Craig Burley <burley@gnu.org>
221 * bad.def (FFEBAD_OPEN_UNSUPPORTED, FFEBAD_INQUIRE_UNSUPPORTED,
222 FFEBAD_READ_UNSUPPORTED, FFEBAD_WRITE_UNSUPPORTED,
223 FFEBAD_QUAD_UNSUPPORTED, FFEBAD_BLOCKDATA_STMT,
224 FFEBAD_TRUNCATING_CHARACTER, FFEBAD_TRUNCATING_HOLLERITH,
225 FFEBAD_TRUNCATING_NUMERIC, FFEBAD_TRUNCATING_TYPELESS,
226 FFEBAD_TYPELESS_OVERFLOW): Change these from warnings
229 Tue May 19 14:51:59 1998 Craig Burley <burley@gnu.org>
231 * Make-lang.in (f77.install-info, f77.uninstall):
232 Use install-info as appropriate.
234 Tue May 19 12:56:54 1998 Craig Burley <burley@gnu.org>
236 * com.c (ffecom_init_0): Rename xargc to f__xargc,
237 in accord with same-dated change to f/runtime.
239 Fri May 15 10:52:49 1998 Craig Burley <burley@gnu.org>
241 * com.c (ffecom_convert_narrow_, ffecom_convert_widen_):
242 Be even more persnickety in checking for internal bugs.
243 Also, if precision isn't changing, just return the expr.
245 * expr.c (ffeexpr_token_number_): Call
246 ffeexpr_make_float_const_ to make an integer.
247 (ffeexpr_make_float_const_): Handle making an integer.
249 * intrin.c (ffeintrin_init_0): Distinguish between
250 crashes on bad arg base and kind types.
252 Thu May 14 13:30:59 1998 Craig Burley <burley@gnu.org>
254 * Make-lang.in (f/expr.c): Now depends on f/stamp-str.
255 * expr.c: Use ffestrOther in place of ffeexprDotdot_.
256 * str-ot.fin: Add more keywords for expr.c.
258 * intdoc.c (dumpimp): Trivial fix.
260 * com.c (ffecom_expr_): Add ltkt variable for clarity.
262 Wed May 13 13:05:34 1998 Craig Burley <burley@gnu.org>
264 * Make-lang.in (G77STAGESTUFF): Add g77.o, g77spec.o,
266 (f77.clean): Add removal of g77.c, g77.o, g77spec.o,
268 (f77.distclean): Delete removal of g77.c.
270 Tue Apr 28 09:23:10 1998 Craig Burley <burley@gnu.org>
273 * global.c (ffeglobal_ref_progunit_): When transitioning
274 from EXT to FUNC, discard hook, since the decl, if any, is
277 Sun Apr 26 09:05:50 1998 Craig Burley <burley@gnu.org>
279 * com.c (ffecom_char_enhance_arg_): Wrap the upper bound
280 (the PARM_DECL specifying the length of the CHARACTER*(*)
281 dummy arg) in a variable_size invocation, to prevent
282 dwarf2out.c crashing when compiling code with -g.
284 Sat Apr 18 05:03:21 1998 Craig Burley <burley@gnu.org>
286 * com.c (ffecom_check_size_overflow_): Ignore overflow
287 as well if dummy argument.
289 Fri Apr 17 17:18:04 1998 Craig Burley <burley@gnu.org>
291 * version.h: Get rid of the overly large headers
292 here too, as done in version.c.
294 Tue Apr 14 14:40:40 1998 Craig Burley <burley@gnu.org>
296 * com.c (ffecom_start_progunit_): Mark function decl
297 as used, to avoid spurious warning (-Wunused) for ENTRY.
299 Tue Apr 14 14:19:34 1998 Craig Burley <burley@gnu.org>
301 * sta.c (ffesta_second_): Check for CASE DEFAULT
302 as well as CASE, or it won't be recognized.
304 Mon Mar 23 21:20:35 1998 Craig Burley <burley@gnu.org>
306 * version.c: Reduce to a one-line file, like
307 gcc's version.c, since there's really no content
310 Mon Mar 23 11:58:43 1998 Craig Burley <burley@gnu.org>
312 * bugs.texi: Various updates.
314 * com.c (ffecom_tree_canonize_ptr_): Fix up spacing a bit.
316 Mon Mar 16 21:20:35 1998 Craig Burley <burley@gnu.org>
318 * expr.c (ffeexpr_sym_impdoitem_): Don't blindly
319 reset symbol info after calling ffesymbol_error,
322 Thu Mar 12 13:50:21 1998 Craig Burley <burley@gnu.org>
324 * lang-options.h: Sort all -f*-intrinsics-* options,
325 for consistency with other g77 versions.
327 Mon Mar 16 15:38:50 1998 Craig Burley <burley@gnu.org>
329 * Version 0.5.22 released.
331 Mon Mar 16 14:36:02 1998 Craig Burley <burley@gnu.org>
333 Make -g work better for ENTRY:
334 * com.c (ffecom_start_progunit_): Master function
335 for ENTRY-laden procedure is not really invented,
336 so it can be debugged.
337 (ffecom_do_entry_): Push/set/pop lineno for each
340 Sun Mar 15 05:48:49 1998 Craig Burley <burley@gnu.org>
342 * intrin.def: Fix spelling of mixed-case form
343 of `CPU_Time' (was `Cpu_Time').
345 1998-03-09 Dave Love <d.love@dl.ac.uk>
347 * Make-lang.in: Set CONFIG_SITE to a non-existent file since
348 /dev/null loses with bash 2.0/autoconf 2.12. Put
349 F77_FLAGS_TO_PASS before CC.
351 Sun Mar 8 16:35:34 1998 Craig Burley <burley@gnu.org>
353 * intrin.def: Use tabs instead of blanks more
354 consistently (excepting DEFGEN section for now).
356 Sat Feb 28 15:24:38 1998 Craig Burley <burley@gnu.org>
358 * intrin.def: Make CPU_TIME's arg generic real to be just
361 Fri Feb 20 12:45:53 1998 Craig Burley <burley@gnu.org>
363 * expr.c (ffeexpr_token_arguments_): Make sure
364 outer exprstack isn't null.
366 1998-02-16 Dave Love <d.love@dl.ac.uk>
368 * Makefile.in (f/fini): Don't use -W -Wall with HOST_CC.
370 Sun Jan 11 02:14:47 1998 Craig Burley <burley@gnu.org>
372 Support FORMAT(I<1+2>) (constant variable-FORMAT
374 * bad.def (FFEBAD_FORMAT_VARIABLE): New diagnostic.
375 * std.c (ffestd_R1001rtexpr_): New function.
376 (ffestd_R1001dump_, ffestd_R1001dump_1005_1_,
377 ffestd_R1001dump_1005_2_, ffestd_R1001dump_1005_3_,
378 ffestd_R1001dump_1005_4_, ffestd_R1001dump_1005_5_,
379 ffestd_R1001dump_1010_2_, ffestd_R1001dump_1010_3_,
380 ffestd_R1001dump_1010_4_, ffestd_R1001dump_1010_5_):
381 Use new function instead of ffestd_R1001error_.
383 * stb.c (ffestb_R10014_, ffestb_R10016_, ffestb_R10018_,
384 ffestb_R100110_): Restructure `for' loop for style.
386 Fix 970626-2.f by not doing most back-end processing
387 when current_function_decl is an ERROR_MARK, and by
388 making that the case when its type would be an ERROR_MARK:
389 * com.c (ffecom_start_progunit_, finish_function,
390 lang_printable_name, start_function,
391 ffecom_finish_symbol_transform_): Test for ERROR_MARK.
392 * std.c (ffestd_stmt_pass_): Don't do any downstream
393 processing if ERROR_MARK.
395 * Make-lang.in (f77.install-common): Don't install, and
396 don't uninstall existing, Info files if f/g77.info
397 doesn't exit. (This is a somewhat modified version
398 of an egcs patch on 1998-01-07 12:05:51 by Bruno Haible
399 <bruno@linuix.mathematik.uni-karlsruhe.de>.)
401 Fri Jan 9 19:09:07 1998 Craig Burley <burley@gnu.org>
403 Fix -fpedantic combined with `F()' invocation,
404 also -fugly-comma combined with `IARGC()' invocation:
405 * bad.def (FFEBAD_NULL_ARGUMENT_W): New diagnostic.
406 * expr.c (ffeexpr_finished_): Don't reject null expressions
407 in the argument-expression context -- let outer context
409 (ffeexpr_token_arguments_): Warn about null expressions
410 here if -fpedantic (as appropriate).
411 Obey -fugly-comma for only external-procedure invocations.
412 * intrin.c (ffeintrin_check_): No longer ignore explicit
413 omitted trailing args.
415 Tue Dec 23 14:58:04 1997 Craig Burley <burley@gnu.org>
417 * intrin.c (ffeintrin_fulfill_generic): Don't generate
418 FFEBAD_INTRINSIC_TYPE for CHARACTER*(*) intrinsic.
420 * com.c (ffecom_gfrt_basictype):
421 (ffecom_gfrt_kindtype):
423 (FFECOM_rttypeVOIDSTAR_): New return type `void *', for
424 the SIGNAL intrinsic.
425 * com-rt.def (FFECOM_rttypeSIGNAL): Now returns `void *'.
426 * intdoc.c: Replace `p' kind specifier with `7'.
427 * intrin.c (ffeintrin_check_, ffeintrin_init_0): Replace
428 `p' kind specifier with `7'.
429 * intrin.def (FFEINTRIN_impLOC, FFEINTRIN_impSIGNAL_func,
430 FFEINTRIN_impSIGNAL_subr): Replace `p' specifier with `7'.
431 Also, SIGNAL now returns a `void *' status, not `int'.
433 Improve run-time diagnostic for "PRINT '(I1', 42":
434 * com.c (ffecom_char_args_x_): Renamed from ffecom_char_args_,
435 which is now a macro (to avoid lots of changes to other code)
436 with new arg, ffecom_char_args_with_null_ being another new
437 macro to call same function with different value for new arg.
438 This function now appends a null byte to opCONTER expression
439 if the new arg is TRUE.
440 (ffecom_arg_ptr_to_expr): Support NULL length pointer.
441 * ste.c (ffeste_io_cilist_):
442 (ffeste_io_icilist_): Pass NULL length ptr for
443 FORMAT expression, so null byte gets appended where
445 * target.c (ffetarget_character1):
446 (ffetarget_concatenate_character1):
447 (ffetarget_substr_character1):
448 (ffetarget_convert_character1_character1):
449 (ffetarget_convert_character1_hollerith):
450 (ffetarget_convert_character1_integer4):
451 (ffetarget_convert_character1_logical4):
452 (ffetarget_convert_character1_typeless):
453 (ffetarget_hollerith): Append extra phantom null byte as
454 part of FFETARGET-NULL-BYTE kludge.
456 * intrin.def (FFEINTRIN_impCPU_TIME): Point to
457 FFECOM_gfrtSECOND as primary run-time routine.
459 Mon Dec 22 12:41:07 1997 Craig Burley <burley@gnu.org>
461 * intrin.c (ffeintrin_init_0): Remove duplicate
464 Sun Dec 14 02:49:58 1997 Craig Burley <burley@gnu.org>
466 * intrin.c (ffeintrin_init_0): Fix up indentation a bit.
467 Fix bug that prevented checking of arguments other
470 * intdoc.c: Fix up indentation a bit.
472 Tue Dec 9 16:20:57 1997 Richard Henderson <rth@cygnus.com>
474 * com.c (ffecom_type_vardesc_): Vardesc.dims is a `ftnlen*'.
476 Mon Dec 1 19:12:36 1997 Craig Burley <burley@gnu.org>
478 * intrin.c (ffeintrin_check_): Fix up indentation a bit more.
480 Mon Dec 1 16:21:08 1997 Craig Burley <burley@gnu.org>
482 * com.c (ffecom_arglist_expr_): Crash if non-supplied
483 optional arg isn't passed as an address.
484 Pass null pointer explicitly, instead of via ffecom routine.
485 If incoming argstring is NULL, substitute pointer to "0".
486 Recognize '0' as ending the usual arg stuff, just like '\0'.
488 Sun Nov 30 22:22:22 1997 Craig Burley <burley@gnu.org>
490 * intdoc.c: Minor fix-ups.
492 * intrin.c (ffeintrin_check_): Fix up indentation a bit.
494 * intrin.def: Fix up spacing a bit.
496 1997-11-17 Dave Love <d.love@dl.ac.uk>
498 * com.c (ffecom_arglist_expr_): Pass null pointers for optional
499 args which aren't supplied.
501 Sun Nov 16 21:45:43 1997 H.J. Lu (hjl@gnu.ai.mit.edu)
503 * Make-lang.in (f77.install-info): Depend on f77.info.
505 1997-11-06 Dave Love <d.love@dl.ac.uk>
507 * intrin.def: Allow non-integer args for INT2 and INT8 (per
510 Tue Oct 28 02:21:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
512 * lang-options.h: Add -fgnu-intrinsics-* and
513 -fbadu77-intrinsics-* options.
515 Sun Oct 26 02:36:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
517 * com.c (lang_print_error_function): Fix to more
518 reliably notice when the diagnosed region changes.
520 Sat Oct 25 23:43:36 1997 Craig Burley <burley@gnu.ai.mit.edu>
523 * sta.c, sta.h (ffesta_outpooldisp): New function.
524 * std.c (ffestd_stmt_pass_): Don't kill NULL pool.
525 (ffestd_R842): If pool already preserved, save NULL
526 for pool, because it should be killed only once.
528 * malloc.c [MALLOC_DEBUG]: Put initializer for `name'
529 component in braces, to avoid compiler warning.
531 Fri Oct 10 13:00:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
533 * ste.c (ffeste_begin_iterdo_): Fix loop setup so iteration
534 variable is modified only after the #iterations is calculated;
535 otherwise if the iteration variable is aliased to any of the
536 operands in the start, end, or increment expressions, the
537 wrong #iterations might be calculated.
539 * com.c (ffecom_save_tree): Fix indentation.
541 1997-10-05 Dave Love <d.love@dl.ac.uk>
543 * intrin.def: Make SECOND_subr's arg generic real for people
544 porting from Cray and making everything double precision.
546 Mon Sep 29 16:18:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
548 * stu.c (ffestu_list_exec_transition_,
549 ffestu_dummies_transition_): Specify `bool' type for
550 `in_progress' variables.
552 * com.h (assemble_string): Declare this routine (instead
553 of #include'ing "output.h" from gcc) to eliminate warnings
556 Fri Sep 19 01:12:27 1997 Craig Burley <burley@gnu.ai.mit.edu>
558 * expr.c (ffeexpr_reduced_eqop2_):
559 (ffeexpr_reduced_relop2_): Minor fixes to diagnostic code.
561 * fini.c (main): Change return type to `int'.
563 Wed Sep 17 10:47:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
565 * com-rt.def (FFECOM_gfrtDSIGN, FFECOM_gfrtISIGN,
566 FFECOM_gfrtSIGN): Add second argument.
568 * expr.c (ffeexpr_cb_comma_c_): Trivial fixes.
570 Tue Sep 9 01:59:35 1997 Craig Burley <burley@gnu.ai.mit.edu>
572 * Version 0.5.21 released.
574 Tue Sep 9 00:31:01 1997 Craig Burley <burley@gnu.ai.mit.edu>
576 * intdoc.c (dumpem): Put appropriate commentary in
577 output file, so readers know it isn't source.
579 Wed Aug 27 08:08:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
581 * proj.h: Always #include "config.j" first, to pick up
583 * com.c: Change bcopy() and bzero() calls to memcpy()
584 and memset() calls, to make more of g77 ANSI C.
586 1997-08-26 Dave Love <d.love@dl.ac.uk>
588 * Make-lang.in ($(srcdir)/f/runtime/configure,
589 $(srcdir)/f/runtime/libU77/configure): Fix for when srcdir isn't
592 Tue Aug 26 05:59:21 1997 Craig Burley <burley@gnu.ai.mit.edu>
594 * ansify.c (main): Make sure readers of stdout know
595 it's derived from stdin; omit comment text; get source
596 line numbers in future stderr output to be correct.
598 Tue Aug 26 01:36:01 1997 Craig Burley <burley@gnu.ai.mit.edu>
601 * stb.c (ffestb_R5284_): Allow OPEN_PAREN after closing
602 SLASH as well as NAME.
604 Mon Aug 25 23:48:17 1997 Craig Burley <burley@gnu.ai.mit.edu>
606 Changes to allow g77 docs to be built entirely from scratch
607 using any ANSI C compiler, not requiring GNU C:
608 * Make-lang.in ($(srcdir)/f/intdoc.texi): "Pipe" new
609 location of intrinsic documentation data base, f/intdoc.in,
610 through new `ansify' program to append `\n\' to quoted
611 newlines, into f/intdoc.h0. Do appropriate cleanups. Explain.
612 (f77.mostlyclean): Add f/ansify and f/intdoc.h0 to cleanups.
613 * f/ansify.c: New program.
614 * f/intdoc.c: Fix so it conforms to ANSI C.
615 #include f/intdoc.h0 instead of f/intdoc.h.
617 * f/intdoc.h, f/intdoc.in: Rename the former to the latter; no
618 changes made to the content in this patch!
619 * f/intrin.h (ffeintrinFamily): Fix to conform to ANSI C.
621 Sun Aug 24 06:52:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
623 Fix up g77 compiler data base for libf2c routines:
624 * com-rt.def (FFECOM_gfrtSIGNAL): Change return type to
625 FTNINT to match actual code.
627 * com.c (ffecomRttype_): Replace FFECOM_rttypeINT_ with
628 FFECOM_rttypeFTNINT_.
629 Add and fix up comments.
630 (ffecom_make_gfrt_, ffecom_gfrt_basictype,
631 ffecom_gfrt_kindtype): Replace FFECOM_rttypeINT_ with
632 FFECOM_rttypeFTNINT_; add FFECOM_rttypeDOUBLEREAL_.
634 Wed Aug 20 17:18:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
636 * global.c (ffeglobal_ref_progunit_): It's okay to have
637 a different CHARACTER*n length for a reference if the
638 existing length is for another reference, not a definition.
640 Mon Aug 18 14:27:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
643 * global.c (ffeglobal_new_progunit_): Distinguish
644 between previously defined, versus inferred, filewide
645 when it comes to diagnostics.
648 * global.c (ffeglobal_ref_progunit_): Change BDATA into EXT
649 right at the beginning, so EXTERNAL FOO followed later
650 by SUBROUTINE FOO is not diagnosed.
653 * com-rt.def (FFECOM_gfrtALARM): Returns `integer', not
656 Sun Aug 17 03:32:44 1997 Craig Burley <burley@gnu.ai.mit.edu>
658 Fix up problems when virtual memory exhausted:
659 * malloc.c (malloc_new_): Use gcc's xmalloc(), so we
660 print a nicer message when malloc returns no memory.
661 (malloc_resize_): Ditto for xrealloc().
663 * Make-lang.in, Makefile.in: Comment out lines containing
666 Sat Aug 16 19:41:33 1997 Craig Burley <burley@gnu.ai.mit.edu>
668 * com.c (ffecom_make_gfrt_): For rttypeREAL_F2C_, return
669 double_type_node; for rttypeREAL_GNU_, return
672 1997-08-13 Dave Love <d.love@dl.ac.uk>
674 * config-lang.in (diff_excludes): Add some hints about known
675 problematic platforms.
677 1997-08-13 Dave Love <d.love@dl.ac.uk>
679 * intdoc.h: Document `alarm'.
681 Mon Aug 11 21:19:22 1997 Craig Burley <burley@gnu.ai.mit.edu>
683 * Make-lang.in ($(RUNTIMESTAGESTUFF)): Add
686 Mon Aug 11 01:52:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
688 * com.c (ffecom_build_complex_constant_): Go with the
689 new build_complex() approach used in gcc-2.8.
691 * com.c (ffecom_sym_transform_): Don't set
692 DECL_IN_SYSTEM_HEADER for a tree node that isn't
693 a VAR_DECL, which happens when var is in common!
695 * com.c (ffecom_expr_intrinsic_) (case FFEINTRIN_impALARM):
696 No need to test codegen_imp -- there's only one valid here.
698 * intrin.def (FFEINTRIN_impALARM): Specify `Status' argument
701 Fri Aug 8 05:40:23 1997 Craig Burley <burley@gnu.ai.mit.edu>
703 Substantial changes to accommodate distinctions among
704 run-time routines that support intrinsics, and between
705 routines that compute and return the same type vs. those
706 that compute one type and return another (or `void'):
707 * com-rt.def: Specify new return type REAL_F2C_ instead
708 of many DOUBLE_, COMPLEX_F2C_ instead of COMPLEX_, and
710 Clear up the *BES* routines "once and for all".
711 * com.c: New return types.
712 (ffecom_convert_narrow_, ffecom_convert_widen_):
713 New functions that are "safe" variants of convert(),
714 to catch errors that ffecom_expr_intrinsic_() now
716 (ffecom_arglist_expr_): Ensure arguments are not
717 converted to narrower types.
718 (ffecom_call_): Ensure return value is not converted
720 (ffecom_char_args_): Use new ffeintrin_gfrt_direct()
722 (ffecom_expr_intrinsic_): Simplify how run-time
723 routine is selected (via `gfrt' only now; lose the
724 redundant `ix' variable).
725 Eliminate the `library' label; any code that doesn't
726 return directly just `break's out now with `gfrt'
728 Set `gfrt' to default choice initially, either a
729 fast direct form or, if not available, a slower
730 indirect-callable form.
731 (ffecom_make_gfrt_): No longer need to do special
732 check for complex; it's built into the new return-type
734 (ffecom_ptr_to_expr): Use new ffeintrin_gfrt_indirect()
736 * intrin.c, intrin.h: `gfrt' field replaced with three fields,
737 so it is easier to provide faster direct-callable and
738 GNU-convention indirect-callable routines in the future.
739 DEFIMP macro adjusted accordingly, along with all its uses.
740 (ffeintrin_gfrt_direct): New function.
741 (ffeintrin_gfrt_indirect): Ditto.
742 (ffeintrin_is_actualarg): If `-fno-f2c' is in effect,
743 require a GNU-callable version of intrinsic instead of
744 an f2c-callable version, so indirect calling is still checked.
745 * intrin.def: Replace one GFRT field with the three new fields,
746 as appropriate for each DEFIMP intrinsic.
748 * com.c (ffecom_stabilize_aggregate_,
749 ffecom_convert_to_complex_): Make these `static'.
751 Thu Aug 7 11:24:34 1997 Craig Burley <burley@gnu.ai.mit.edu>
753 Provide means for front end to determine actual
754 "standard" return type for an intrinsic if it is
755 passed as an actual argument:
756 * com.h, com.c (ffecom_gfrt_basictype,
757 ffecom_gfrt_kindtype): New functions.
758 (ffecom_gfrt_kind_type_): Replaced with new function.
760 (ffecom_make_gfrt_): No longer need do anything
763 * intrin.c (ffeintrin_basictype, ffeintrin_kindtype):
764 Now returns correct type info for specific intrinsic
765 (based on type of run-time-library implementation).
767 Wed Aug 6 23:08:46 1997 Craig Burley <burley@gnu.ai.mit.edu>
769 * global.c (ffeglobal_ref_progunit_): Don't reset
770 number of arguments just due to new type info,
771 so useful warnings can be issued.
773 1997-08-06 Dave Love <d.love@dl.ac.uk>
775 * intrin.def: Fix IDATE_vxt argument order.
776 * intdoc.h: Likewise.
778 Thu Jul 31 22:22:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
780 * global.c (ffeglobal_proc_ref_arg): If REF/DESCR
781 disagreement, DESCR is CHARACTER, and types disagree,
782 pretend the argsummary agrees so the message ends up
783 being about type disagreement.
784 (ffeglobal_proc_def_arg): Ditto.
786 * expr.c (ffeexpr_token_first_rhs_3_): Set info for LABTOK
787 to NONE of everything, to avoid misdiagnosing filewide
788 usage of alternate returns.
790 Sun Jul 20 23:07:47 1997 Craig Burley <burley@gnu.ai.mit.edu>
792 * com.c (ffecom_sym_transform_): If type gets set
793 to error_mark_node, just return that for transformed symbol.
794 (ffecom_member_phase2_): If type gets set to error_mark_node,
796 (ffecom_check_size_overflow_): Add `dummy' argument to
797 flag that type is for a dummy, update all callers.
799 Sun Jul 13 17:40:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
802 * where.c (ffewhere_set_from_track): If start point
803 is too large, just use initial start point. 0.6 should
804 fix all this properly.
807 * com.c (ffecom_sym_transform_): Preserve error_mark_node for type.
808 (ffecom_type_localvar_): Ditto.
809 (ffecom_sym_transform_): If type is error_mark_node,
810 don't error-check decl size, because back end responds by
811 setting that to an integer 0 instead of error_mark_node.
812 (ffecom_transform_common_): Same as earlier fix to _transform_
813 in that size is checked by dividing BITS_PER_UNIT instead of
815 (ffecom_transform_equiv_): Ditto.
818 * stb.c (ffestb_R10014_): Fix flaky fall-through in error
819 test for FFELEX_typeCONCAT by just replicating the code,
820 and do FFELEX_typeCOLONCOLON while at it.
822 1997-07-07 Dave Love <d.love@dl.ac.uk>
824 * intdoc.h: Add various missing pieces; correct GMTIME, LTIME
827 * intrin.def, com-rt.def: Add alarm.
829 * com.c (ffecom_expr_intrinsic_): Add case for alarm.
831 Thu Jun 26 04:19:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
834 * com.c (ffecom_sym_transform_): For sanity-check compare
835 of gbe size of local variable to g77 expectation,
836 use varasm.c/assemble_variable technique of dividing
837 BITS_PER_UNIT out of gbe info instead of multiplying
838 g77 info up, to avoid crash when size in bytes is very
839 large, and overflows an `int' or similar when multiplied.
842 * com.c (ffecom_finish_symbol_transform_): Don't bother
843 transforming a dummy argument, to avoid a crash.
844 * ste.c (ffeste_R1227): Don't return a value if the
845 result decl, or its type, is error_mark_node.
848 * lex.c (ffelex_splice_tokens): `-fdollar-ok' is
849 irrelevant to whether a DOLLAR token should be made
850 from an initial character of `$'.
853 * stb.c (ffestb_do3_): DO iteration variable is an
854 lhs, not rhs, expression.
856 Fix 970626-7.f and 970626-8.f:
857 * expr.c (ffeexpr_cb_comma_i_1_): Set IMPDO expression
858 to have clean info, because undefined rank, for example,
859 caused crash on mangled source on UltraSPARC but not
860 on Alpha for a series of weird reasons.
861 (ffeexpr_cb_close_paren_): If not CLOSE_PAREN, push
862 opANY expression onto stack instead of attempting
863 to mimic what program might have wanted.
864 (ffeexpr_cb_close_paren_): Don't wrap opPAREN around
865 opIMPDO, just warn that it's gratuitous.
866 * bad.def (FFEBAD_IMPDO_PAREN): New warning.
869 * expr.c (ffeexpr_declare_parenthesized_): Must shut down
870 parsing in kindANY case, otherwise the parsing engine might
871 decide there's an ambiguity.
872 (ffeexpr_token_name_rhs_): Eliminate parentypeSUBROUTINE_
873 case, so we crash right away if it comes through.
874 * st.c, st.h, sta.c, sta.h (ffest_shutdown, ffesta_shutdown):
877 Tue Jun 24 19:47:29 1997 Craig Burley <burley@gnu.ai.mit.edu>
879 * com.c (ffecom_check_size_overflow_): New function
880 catches some cases of the size of a type getting
881 too large. varasm.c must catch the rest.
882 (ffecom_sym_transform_): Use new function.
883 (ffecom_type_localvar_): Ditto.
885 Mon Jun 23 01:09:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
887 * global.c (ffeglobal_proc_def_arg): Fix comparison
889 (ffeglobal_proc_ref_arg): Ditto.
891 * lang-options.h, top.c: Rename `-fdebug' to `-fxyzzy',
892 since it's an unsupported internals option and some
893 poor user might guess that it does something.
895 * bad.def: Make a warning for each filewide diagnostic.
896 Put all filewides together.
897 * com.c (ffecom_sym_transform_): Don't substitute
898 known global tree for global entities when `-fno-globals'.
899 * global.c (ffeglobal_new_progunit_): Don't produce
900 fatal diagnostics about globals when `-fno-globals'.
901 Instead, produce equivalent warning when `-Wglobals'.
902 (ffeglobal_proc_ref_arg): Ditto.
903 (ffeglobal_proc_ref_nargs): Ditto.
904 (ffeglobal_ref_progunit_): Ditto.
905 * lang-options.h, top.c, top.h: New `-fno-globals' option.
907 Sat Jun 21 12:32:54 1997 Craig Burley <burley@gnu.ai.mit.edu>
909 * expr.c (ffeexpr_fulfill_call_): Set array variable
910 to avoid warning about uninitialized variable.
912 * Make-lang.in: Get rid of any setting of HOST_* macros,
913 since these will break gcc's build!
914 * makefile: New file to make building derived files
917 Thu Jun 19 18:19:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
919 * g77.c (main): Install Emilio Lopes' patch to support
920 Ratfor, and to fix the printing of the version string
921 to go to stderr, not stdout.
922 * lang-specs.h: Install Emilio Lopes' patch to support
923 Ratfor, and patch the result to support picking up
924 `*f771' from the `specs' file.
926 Thu Jun 12 14:36:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
928 * storag.c (ffestorag_update_init, ffestorag_update_save):
929 Also update parent, in case equivalence processing
930 has already eliminated pointers to it via the
931 local equivalence info.
933 Tue Jun 10 14:08:26 1997 Craig Burley <burley@gnu.ai.mit.edu>
935 * intdoc.c: Add cross-reference to end of description
936 of any generic intrinsic pointing to other intrinsics
939 Warn about explicit type declaration for intrinsic
940 that disagrees with invocation:
941 * expr.c (ffeexpr_paren_rhs_let_): Preserve type info
942 for intrinsic functions.
943 (ffeexpr_token_funsubstr_): Ditto.
944 * intrin.c (ffeintrin_fulfill_generic): Warn if type
945 info of fulfilled intrinsic invocation disagrees with
946 explicit type info given symbol.
947 (ffeintrin_fulfill_specific): Ditto.
948 * stc.c (ffestc_R1208_item): Preserve type info
950 (ffestc_R501_item): Ditto.
952 Mon Jun 9 17:45:44 1997 Craig Burley <burley@gnu.ai.mit.edu>
954 * com.c (ffecom_expr_intrinsic_): Fix several of the
955 libU77/libF77-unix handlers to properly convert their
958 * com-rt.def (FFECOM_gfrtFSTAT): Append missing "i" to
961 Fri Jun 6 14:37:30 1997 Craig Burley <burley@gnu.ai.mit.edu>
963 * com.c (ffecom_expr_intrinsic_): Have a case statement
964 for every intrinsic implementation, so missing ones
965 are caught via gcc warnings.
966 Don't call ffeintrin_codegen_imp anymore.
967 * intrin.c (ffeintrin_fulfill_generic): Remove cg_imp
969 (ffeintrin_codegen_imp): Delete this function.
970 * intrin.def, intrin.h: Remove DEFIMQ stuff from here
973 Thu Jun 5 13:03:07 1997 Craig Burley <burley@gnu.ai.mit.edu>
975 * top.c (ffe_decode_option): New -fbadu77-intrinsics-*
978 * intrin.h: New BADU77 family.
979 * intrin.c (ffeintrin_state_family): Ditto.
981 Implement new scheme to track intrinsic names vs. forms:
982 * intrin.c (ffeintrin_fulfill_generic),
983 (ffeintrin_fulfill_specific), (ffeintrin_is_intrinsic),
984 intrin.def: The documented name is now either in the
985 generic info or, if no generic, in the specific info.
986 For a generic, the specific info contains merely the
987 distinguishing form (usually "function" or "subroutine"),
988 used for diagnostics about ambiguous references and
989 in the documentation.
991 * intrin.def: Clean up formatting of DEFNAME block.
992 Convert many libU77 intrinsics into generics that
993 support both subroutine and function forms.
994 Put the function forms of side-effect routines into
995 the new BADU77 family.
996 Make MCLOCK and TIME return INTEGER*4 again, and add
997 INTEGER*8 equivalents called MCLOCK8 and TIME8.
998 Fix up more status return values to be written and
999 insist on them being I1 as well.
1000 * com.c (ffecom_expr_intrinsic_): Lots of changes to
1001 support new libU77 intrinsic interfaces.
1003 Mon Jun 2 00:37:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
1005 * com.c (ffecom_init_0): Pointer type is now INTEGER(KIND=7),
1006 not INTEGER(KIND=0), since we want to reserve KIND=0 for
1009 Thu May 29 14:30:33 1997 Craig Burley <burley@gnu.ai.mit.edu>
1011 Fix bugs preventing CTIME(I*4) from working correctly:
1012 * com.c (ffecom_char_args_): For FUNCREF case, process
1013 args to intrinsic just as they would be in
1014 ffecom_expr_intrinsic_.
1015 * com-rt.def (FFECOM_gfrtCTIME, FFECOM_gfrtTTYNAM): Fix
1016 argument decls to specify `&'.
1018 Wed May 28 22:19:49 1997 Craig Burley <burley@gnu.ai.mit.edu>
1020 Fix gratuitous warnings exposed by dophot aka 970528-1:
1021 * global.c (ffeglobal_proc_def_arg, ffeglobal_proc_ref_arg):
1022 Support distinct function/subroutine arguments instead of
1025 * expr.c (ffeexpr_fulfill_call_): A SYMTER with kindNONE
1026 also is a procedure (either function or subroutine).
1028 Mon May 26 20:25:31 1997 Craig Burley <burley@gnu.ai.mit.edu>
1030 * bad.def: Have several lexer diagnostics refer to
1031 documentation for people who need more info on what Fortran
1032 source code is supposed to look like.
1034 * expr.c (ffeexpr_reduced_bool1_), bad.def: New diagnostics
1035 specific to .NOT. now mention only one operand instead
1038 * g77.c: Recognize -fsyntax-only, similar to -c etc.
1039 (lookup_option): Fix bug that prevented non-`--' options
1040 from being recognized.
1042 Sun May 25 04:29:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
1044 * intrin.def (FFEINTRIN_impCTIME): Accept `I*' expression
1045 for STime instead of requiring `I2'.
1047 Tue May 20 16:14:40 1997 Craig Burley <burley@gnu.ai.mit.edu>
1049 * symbol.c (ffesymbol_reference): All references to
1050 standard intrinsics are considered explicit, so as
1051 to avoid generating basically useless warnings.
1052 * intrin.c, intrin.h (ffeintrin_is_standard): Returns TRUE
1053 if intrinsic is standard.
1055 Sun May 18 21:14:59 1997 Craig Burley <burley@gnu.ai.mit.edu>
1057 * com-rt.def: Changed all external names of the
1058 form `"\([a-z0-9]*\)_' to `"G77_\1_0"' so as to
1059 allow any name valid as an intrinsic to be used
1060 as such and as a user-defined external procedure
1061 name or common block as well.
1063 Thu May 8 13:07:10 1997 Craig Burley <burley@gnu.ai.mit.edu>
1065 * expr.c (ffeexpr_cb_end_notloc_): For %VAL, %REF, and
1066 %DESCR, copy arg info into new node.
1068 Mon May 5 14:42:17 1997 Craig Burley <burley@gnu.ai.mit.edu>
1070 From Uwe F. Mayer <mayer@math.Vanderbilt.Edu>:
1071 * Make-lang.in (g77-cross): Fix typo in g77.c path.
1073 From Brian McIlwrath <bkm@star.rl.ac.uk>:
1074 * lang-specs.h: Have g77 pick up options from a section
1075 labeled `*f771' of the `specs' file.
1077 Sat May 3 02:46:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1079 * intrin.def (FFEINTRIN_defSIGNAL): Add optional `Status'
1080 argument that com.c already expects (per Dave Love).
1082 More changes to support better tracking of (filewide)
1083 globals, in particular, the arguments to procedures:
1084 * bad.def (FFEBAD_FILEWIDE_NARGS, FFEBAD_FILEWIDE_NARGS_W,
1085 FFEBAD_FILEWIDE_ARG, FFEBAD_FILEWIDE_ARG_W): New diagnostics.
1086 * expr.c (ffebad_fulfill_call_): Provide info on each
1087 argument to ffeglobal.
1088 * global.c, global.h (ffeglobal_proc_def_arg,
1089 ffeglobal_proc_def_nargs, ffeglobal_proc_ref_arg,
1090 ffeglobal_proc_ref_args): New functions.
1091 (ffeglobalArgSummary, ffeglobalArgInfo_): New types.
1093 Tue Apr 29 18:35:41 1997 Craig Burley <burley@gnu.ai.mit.edu>
1095 More changes to support better tracking of (filewide)
1097 * expr.c (ffeexpr_fulfill_call_): New function.
1098 (ffeexpr_token_name_lhs_): Call after building procedure
1099 reference expression. Also leave info field for ANY-ized
1101 (ffeexpr_token_arguments_): Ditto.
1103 Mon Apr 28 20:04:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
1105 Changes to support better tracking of (filewide)
1106 globals, mainly to avoid crashes due to inlining:
1107 * bad.def: Go back to quoting intrinsic names,
1108 (FFEBAD_FILEWIDE_DISAGREEMENT, FFEBAD_FILEWIDE_TIFF,
1109 FFEBAD_FILEWIDE_TYPE_MISMATCH): New diagnostics.
1110 (FFEBAD_INTRINSIC_EXPIMP, FFEBAD_INTRINSIC_GLOBAL): Reword
1112 * com.c (ffecom_do_entry_, ffecom_start_progunit_,
1113 ffecom_sym_transform_): Accommodate new FFEGLOBAL_typeEXT
1115 * expr.c (ffeexpr_sym_lhs_call_, ffeexpr_sym_lhs_extfunc_,
1116 ffeexpr_sym_rhs_actualarg_, ffeexpr_declare_parenthesized_,
1117 ffeexpr_paren_rhs_let_, ffeexpr_token_funsubstr_):
1118 Fill in real kind info instead of leaving NONE where
1120 Register references to intrinsics and globals with ffesymbol
1121 using new ffesymbol_reference function instead of
1122 ffesymbol_globalize.
1123 * global.c (ffeglobal_type_string_): New array for
1125 * global.h, global.c:
1126 Replace ->init mechanism with ->tick mechanism.
1127 Move other common-related members into a substructure of
1128 a union, so the proc substructure can be introduced
1129 to include members related to externals other than commons.
1130 Don't complain about ANY-ized globals; ANY-ize globals
1131 once they're complained about, in any case where code
1132 generation could become a problem.
1133 Handle global entries that have NONE type (seen as
1134 intrinsics), EXT type (seen as EXTERNAL), and so on.
1135 Keep track of kind and type of externals, both via
1136 definition and via reference.
1137 Diagnose disagreements about kind or type of externals
1138 (such as functions).
1139 (ffeglobal_ref_intrinsic, ffeglobal_ref_progunit_): New
1141 * stc.c (ffestc_R1207_item, ffestc_R1208_item,
1142 ffestc_R1219, ffestc_R1226):
1143 Call ffesymbol_reference, not ffesymbol_globalize.
1144 * stu.c (ffestu_sym_end_transition,
1145 ffestu_sym_exec_transition):
1146 Call ffesymbol_reference, not ffesymbol_globalize.
1147 * symbol.c (ffesymbol_globalize): Removed...
1148 (ffesymbol_reference): ...to this new function,
1149 which more generally registers references to symbols,
1150 globalizes globals, and calls on the ffeglobal module
1151 to check globals filewide.
1153 * global.h, global.c: Rename some macros and functions
1154 to more clearly distinguish common from other globals.
1155 All callers changed.
1157 * com.c (ffecom_sym_transform_): Trees describing
1158 filewide globals must be allocated on permanent obstack.
1160 * expr.c (ffeexpr_token_name_lhs_): Don't generate
1161 gratuitous diagnostics for FFEINFO_whereANY case.
1163 Thu Apr 17 03:27:18 1997 Craig Burley <burley@gnu.ai.mit.edu>
1165 * global.c: Add support for flagging intrinsic/global
1166 confusion via warnings.
1167 * bad.def (FFEBAD_INTRINSIC_EXPIMP,
1168 FFEBAD_INTRINSIC_GLOBAL): New diagnostics.
1169 * expr.c (ffeexpr_token_funsubstr_): Ditto.
1170 (ffeexpr_sym_lhs_call_): Ditto.
1171 (ffeexpr_paren_rhs_let_): Ditto.
1172 * stc.c (ffestc_R1208_item): Ditto.
1174 Wed Apr 16 22:40:56 1997 Craig Burley <burley@gnu.ai.mit.edu>
1176 * expr.c (ffeexpr_declare_parenthesized_): INCLUDE
1177 context can't be an intrinsic invocation either.
1179 Fri Mar 28 10:43:28 1997 Craig Burley <burley@gnu.ai.mit.edu>
1181 * expr.c (ffeexpr_token_arguments_): Make sure top of
1182 exprstack is operand before dereferencing operand field.
1184 * lex.c (ffelex_prepare_eos_): Fill up truncated
1185 hollerith token, so crash on null ->text field doesn't
1188 * stb.c (ffestb_R10014_): If NAMES isn't recognized (or
1189 the recognized part is followed in the token by a
1190 non-digit), don't try and collect digits, as there
1191 might be more than FFEWHERE_indexMAX letters to skip
1192 past to do so -- and the code is diagnosed anyway.
1194 Thu Mar 27 00:02:48 1997 Craig Burley <burley@gnu.ai.mit.edu>
1196 * com.c (ffecom_sym_transform_): Force local
1197 adjustable array onto stack.
1199 * stc.c (ffestc_R547_item_object): Don't actually put
1200 the symbol in COMMON if the symbol has already been
1201 EQUIVALENCE'd to a different COMMON area.
1203 * equiv.c (ffeequiv_add): Don't actually do anything
1204 if there's a disagreement over which COMMON area is
1207 Tue Mar 25 03:35:19 1997 Craig Burley <burley@gnu.ai.mit.edu>
1209 * com.c (ffecom_transform_common_): If no explicit init
1210 of COMMON area, don't actually init it even though
1211 storage area suggests it.
1213 Mon Mar 24 12:10:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1215 * lex.c (ffelex_image_char_): Avoid overflowing the
1216 column counter itself, as well as the card image.
1218 * where.c (ffewhere_line_new): Cast ffelex_line_length()
1219 to (size_t) so 255 doesn't overflow to 0!
1221 * stc.c (ffestc_labeldef_notloop_begin_): Don't gratuitously
1222 terminate loop before processing statement, so block
1223 doesn't disappear out from under EXIT/CYCLE processing.
1224 (ffestc_labeldef_notloop_): Has old code from above
1225 function, instead of just calling it.
1227 * expr.c (ffeexpr_cb_comma_i_4_): Don't skip over
1228 arbitrary token (such as EOS).
1230 * com.c (ffecom_init_zero_): Handle RECORD_TYPE and
1231 UNION_TYPE so -fno-zeros works with -femulated-complex.
1233 1997-03-12 Dave Love <d.love@dl.ac.uk>
1235 * intrin.def: New intrinsics INT2, INT8, CPU_TIME. Fix AND, OR,
1236 XOR. [Integrated by burley, AND/OR/XOR already fixed, INT8
1237 implementation changed/fixed.]
1239 Wed Mar 12 10:40:08 1997 Craig Burley <burley@gnu.ai.mit.edu>
1241 * Make-lang.in ($(srcdir)/f/intdoc.texi): Simplify rules
1242 so building f/intdoc is not always necessary; remove
1243 f/intdoc after running it if it is built.
1245 Tue Mar 11 23:42:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
1247 * intrin.def (FFEINTRIN_impAND, FFEINTRIN_impOR,
1248 FFEINTRIN_impXOR): Use the IAND, IOR, and IEOR implementations
1249 of these, instead of crashing in ffecom_expr_intrinsic_
1250 or adding case labels there.
1252 Mon Mar 10 22:51:23 1997 Craig Burley <burley@gnu.ai.mit.edu>
1254 * intdoc.c: Fix so any C compiler can compile this.
1256 Fri Feb 28 13:16:50 1997 Craig Burley <burley@gnu.ai.mit.edu>
1258 * Version 0.5.20 released.
1260 Fri Feb 28 01:45:25 1997 Craig Burley <burley@gnu.ai.mit.edu>
1262 * Make-lang.in (RUNTIMESTAGESTUFF, LIBU77STAGESTUFF):
1263 Move some files incorrectly in the former to the latter,
1264 and add another file or two to the latter.
1266 New meanings for (KIND=n), and new denotations in the
1267 little language describing intrinsics:
1268 * com.c (ffecom_init_0): Assign new meanings.
1269 * intdoc.c: Document new meanings.
1270 Support the new denotations.
1271 * intrin.c: Employ new meanings, mapping them to internal
1272 values (which are the same as they ever were for now).
1273 Support the new denotations.
1274 * intrin.def: Switch DEFIMP table to the new denotations.
1276 * intrin.c (ffeintrin_check_): Fix bug that was leaving
1277 LOC() and %LOC() returning INTEGER*4 on systems where
1278 it should return INTEGER*8.
1280 * type.c: Canonicalize function definitions, for etags
1283 Wed Feb 26 20:43:03 1997 Craig Burley <burley@gnu.ai.mit.edu>
1285 * com.c (ffecom_init_0): Choose INTEGER(KIND=n) types,
1286 where n is 2, 3, and 4, according to the new docs
1287 instead of according to the old C correspondences
1288 (which seem less useful at this point).
1290 * equiv.c (ffeequiv_destroy_): New function.
1291 (ffeequiv_layout_local_): Use this new function
1292 whenever the laying out of a local equivalence chain
1293 is aborted for any reason.
1294 Otherwise ensure that symbols no longer reference
1295 the stale ffeequiv entries that result when they
1296 are killed off in this procedure.
1297 Also, the rooted symbol is one that has storage,
1298 it really is irrelevant whether it has an equiv entry
1299 at this point (though the code to remove the equiv
1300 entry was put in at the end, just in case).
1301 (ffeequiv_kill): When doing internal checks, make
1302 sure the victim isn't named by any symbols it points
1303 to. Not as complete a check as looking through the
1304 entire symbol table (which does matter, since some
1305 code in equiv.c used to remove symbols from the lists
1306 for an ffeequiv victim but not remove that victim as the
1307 symbol's equiv info), but this check did find some
1308 real bugs in the code (that were fixed).
1310 Mon Feb 24 16:42:13 1997 Craig Burley <burley@gnu.ai.mit.edu>
1312 * com.c (ffecom_expr_intrinsic_): Fix a couple of
1313 warnings about uninitialized variables.
1314 * intrin.c (ffeintrin_check_): Ditto, but there were
1315 a couple of _real_ uninitialized-variable _bugs_ here!
1316 (ffeintrin_fulfill_specific): Ditto, no real bug here.
1318 Sun Feb 23 15:01:20 1997 Craig Burley <burley@gnu.ai.mit.edu>
1320 Clean up diagnostics (especially about intrinsics):
1321 * bad.def (FFEBAD_UNIMPL_STMT): Remove.
1322 (FFEBAD_INTRINSIC_*, FFEBAD_NEED_INTRINSIC): Clean these
1323 up so they're friendlier.
1324 (FFEBAD_INTRINSIC_CMPAMBIG): New.
1325 * intrin.c (ffeintrin_fulfill_generic,
1326 ffeintrin_fulfill_specific, ffeintrin_is_intrinsic):
1328 generic or specific name text (which is for doc purposes
1329 anyway) over implementation name text (which is for
1331 * intrin.def: Use more descriptive name texts for generics
1332 and specifics in cases where the names themselves are not
1333 enough (e.g. IDATE, which has two forms).
1335 Fix some intrinsic mappings:
1336 * intrin.def (FFEINTRIN_specIDINT, FFEINTRIN_specAND,
1337 FFEINTRIN_specDFLOAT, FFEINTRIN_specDREAL, FFEINTRIN_specOR,
1338 FFEINTRIN_specXOR): Now have their own implementations,
1339 instead of borrowing from others.
1340 (FFEINTRIN_specAJMAX0, FFEINTRIN_specAJMIN0, FFEINTRIN_specBJTEST,
1341 FFEINTRIN_specDFLOTJ, FFEINTRIN_specFLOATJ, FFEINTRIN_specJIABS,
1342 FFEINTRIN_specJIAND, FFEINTRIN_specJIBCLR, FFEINTRIN_specJIBITS,
1343 FFEINTRIN_specJIBSET, FFEINTRIN_specJIDIM, FFEINTRIN_specJIDINT,
1344 FFEINTRIN_specJIDNNT, FFEINTRIN_specJIEOR, FFEINTRIN_specJIFIX,
1345 FFEINTRIN_specJINT, FFEINTRIN_specJIOR, FFEINTRIN_specJISHFT,
1346 FFEINTRIN_specJISHFTC, FFEINTRIN_specJISIN, FFEINTRIN_specJMAX0,
1347 FFEINTRIN_specJMAX1, FFEINTRIN_specJMIN0, FFEINTRIN_specJMIN1,
1348 FFEINTRIN_specJMOD, FFEINTRIN_specJNINT, FFEINTRIN_specJNOT,):
1349 Turn these implementations off, since it's not clear
1350 just what types they expect in the context of portable Fortran.
1351 (DFLOAT): Now in FVZ family, since f2c supports them
1353 Support intrinsic inquiry functions (BIT_SIZE, LEN):
1354 * intrin.c: Allow `i' in <arg_extra>.
1355 * intrin.def (FFEINTRIN_impBIT_SIZE, FFEINTRIN_impLEN):
1358 Sat Feb 22 13:34:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
1360 Only warn, don't error, for reference to unimplemented
1362 * bad.def (FFEBAD_INTRINSIC_UNIMPLW): Warning version
1364 * intrin.c (ffeintrin_is_intrinsic): Use new warning
1365 version of _UNIMPL (FFEBAD_INTRINSIC_UNIMPLW).
1367 Complain about REAL(Z) and AIMAG(Z) (Z is DOUBLE COMPLEX):
1368 * bad.def (FFEBAD_INTRINSIC_CMPAMBIG): New diagnostic.
1369 * expr.c: Needed #include "intrin.h" anyway.
1370 (ffeexpr_token_intrincheck_): New function handles delayed
1371 diagnostic for "REAL(REAL(expr)" if next token isn't ")".
1372 (ffeexpr_token_arguments_): Do most of the actual checking here.
1373 * intrin.h, intrin.c (ffeintrin_fulfill_specific): New
1374 argument, check_intrin, to tell caller that intrin is REAL(Z)
1375 or AIMAG(Z). All callers updated, mostly to pass NULL in
1377 (ffeintrin_check_): Also has new arg check_intrin for same
1378 purpose. All callers updated the same way.
1379 * intrin.def (FFEINTRIN_impAIMAG): Change return type
1380 from "R0" to "RC", to accommodate f2c (and perhaps other
1381 non-F90 F77 compilers).
1382 * top.h, top.c: New option -fugly-complex.
1384 New GNU intrinsics REALPART, IMAGPART, and COMPLEX:
1385 * com.c (ffecom_expr_intrinsic_): Implement impCOMPLEX
1386 and impREALPART here. (specIMAGPART => specAIMAG.)
1387 * intrin.def: Add the intrinsics here.
1389 Rename implementations of VXTIDATE and VXTTIME to IDATEVXT
1390 and TIMEVXT, so they sort more consistently:
1391 * com.c (ffecom_expr_intrinsic_):
1394 Delete intrinsic group `dcp', add `gnu', etc.:
1395 * intrin.c (ffeintrin_state_family): FFEINTRIN_familyGNU
1396 replaces FFEINTRIN_familyDCP, and gets state from `gnu'
1398 Get rid of FFEINTRIN_familyF2Z, nobody needs it.
1399 Move FFEINTRIN_specDCMPLX from DCP family to FVZ family,
1401 Move FFEINTRIN_specDFLOAT from F2C family to FVZ family.
1402 (FFEINTRIN_specZABS, FFEINTRIN_specZCOS, FFEINTRIN_specZEXP,
1403 FFEINTRIN_specZLOG, FFEINTRIN_specZSIN, FFEINTRIN_specZSQRT):
1404 Move these from F2Z family to F2C family.
1405 * intrin.h (FFEINTRIN_familyF2Z, FFEINTRIN_familyDCP): Remove.
1406 (FFEINTRIN_familyGNU): Add.
1407 * top.h, top.c: Replace `dcp' with `gnu'.
1409 * com.c (ffecom_expr_intrinsic_): Clean up by collecting
1410 simple conversions into one nice, conceptual place.
1411 Fix up some intrinsic subroutines (MVBITS, KILL, UMASK) to
1412 properly push and pop call temps, to avoid wasting temp
1415 * g77.c (doit): Toon says variables should be defined
1416 before being referenced. Spoilsport.
1418 * intrin.c (ffeintrin_check_): Now Dave's worried about
1419 warnings about uninitialized variables. Okay, so for
1420 basic return values 'g' and 's', they _were_
1421 uninitialized -- is determinism really _that_ useful?
1423 * intrin.def (FFEINTRIN_impFGETC): Fix STATUS argument
1424 so that it is INTENT(OUT) instead of INTENT(IN).
1426 1997-02-21 Dave Love <d.love@dl.ac.uk>
1428 * intrin.def, com.c: Support Sun-type `short' and `long'
1429 intrinsics. Perhaps should also do Microcruft-style `int2'.
1431 Thu Feb 20 15:16:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
1433 * com.c (ffecom_expr_intrinsic_): Clean up indentation.
1434 Support SECONDSUBR intrinsic implementation.
1435 Rename SECOND to SECONDFUNC for direct support via library.
1437 * g77.c: Fix to return proper status value to shell,
1438 by obtaining it from processes it spawns.
1440 * intdoc.c: Fix minor typo.
1442 * intrin.def: Turn SECOND into generic that maps into
1443 function and subroutine forms.
1445 * intrin.def: Make FLOAT and SNGL into specific intrinsics.
1447 * intrin.def, intrin.h: Change the way DEFGEN and DEFSPEC
1448 macros work, to save on verbage.
1450 Mon Feb 17 02:08:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
1452 New subsystem to automatically generate documentation
1454 * Make-lang.in ($(srcdir)/f/g77.info,
1455 $(srcdir)/f/g77.dvi): Move g77 doc rules around.
1456 Add to g77 doc rules the new subsystem.
1457 (f77.mostlyclean, f77.maintainer-clean): Also clean up
1458 after new doc subsystem.
1459 * intdoc.c, intdoc.h: New doc subsystem code.
1460 * intrin.h [FFEINTRIN_DOC]: When 1, don't pull in
1461 stuff not needed by doc subsystem.
1463 Improve on intrinsics mechanism to both be more
1464 self-documenting and to catch more user errors:
1465 * intrin.c (ffeintrin_check_): Recognize new arg-len
1466 and arg-rank information, and check it.
1467 Move goto and signal indicators to the basic type.
1468 Permit reference to arbitrary argument number, not
1469 just first argument (for BESJN and BESYN).
1470 (ffeintrin_init_0): Check and accept new notations.
1471 * intrin.c, intrin.def: Value in COL now identifies
1472 arguments starting with number 0 being the first.
1474 Some minor intrinsics cleanups (resulting from doc work):
1475 * com.c (ffecom_expr_intrinsic_): Implement FLUSH
1476 directly once again, handle its optional argument,
1477 so it need not be a generic (awkward to handle in docs).
1478 * intrin.def (BESJ0, BESJ1, BESJN, BESY0, BESY1, BESYN,
1479 CHDIR, CHMOD, CTIME, DBESJ0, DBESJ1, DBESJN, DBESY0,
1480 DBESY1, DBESYN, DDIM, ETIME, FGETC, FNUM, FPUTC, FSTAT,
1481 GERROR, GETCWD, GETGID, GETLOG, GETPID, GETUID, GMTIME,
1482 HOSTNM, IDATE, IERRNO, IIDINT, IRAND, ISATTY, ITIME, JIDINT,
1483 LNBLNK, LSTAT, LTIME, MCLOCK, PERROR, SRAND, SYMLNK, TTYNAM,
1484 UMASK): Change capitalization of initcaps (official) name
1485 to be consistent with Burley's somewhat arbitrary rules.
1486 (BESJN, BESYN): These have return arguments of same type
1487 as their _second_ argument.
1488 (FLUSH): Now a specific, not generic, intrinsic, with one
1490 (FLUSH1): Eliminated.
1491 Add arg-len and arg-rank info to several intrinsics.
1492 (ITIME): Change argument type from REAL to INTEGER.
1494 Tue Feb 11 14:04:42 1997 Craig Burley <burley@gnu.ai.mit.edu>
1496 * Make-lang.in (f771): Invocation of Makefile now done
1497 with $(srcdir)=gcc to go along with $(VPATH)=gcc.
1498 ($(srcdir)/f/runtime/configure,
1499 $(srcdir)/f/runtime/libU77/configure): Break these out
1500 so spurious triggers of this rule don't happen (as when
1501 configure.in is more recent than libU77/configure).
1502 (f77.rebuilt): Distinguish source versus build files,
1503 so this target can be invoked from build directory and
1505 * Makefile.in: This now expects $(srcdir) to be the gcc
1506 source directory, not gcc/f, to agree with $(VPATH).
1507 Accordingly, $(INCLUDES) has been fixed, various cruft
1508 removed, the removal of f771 has been fixed to remove
1509 the _real_ f771 (not the one in gcc's parent directory),
1512 * lex.c: Part of ffelex_finish_statement_() now done
1513 by new function ffelex_prepare_eos_(), so that, in one
1514 popular case, the EOS can be prepared while the pointer
1515 is at the end of the non-continued line instead of the
1516 end of the line that marks no continuation. This improves
1517 the appearance of diagnostics substantially.
1519 Mon Feb 10 12:44:06 1997 Craig Burley <burley@gnu.ai.mit.edu>
1521 * Make-lang.in: runtime Makefile's, and include/f2c.h,
1522 also depend on f/runtime/configure and f/runtime/libU77/configure.
1524 Fix various libU77 routines:
1525 * com-rt.def (FFECOM_gfrtCTIME, FFECOM_gfrtMCLOCK,
1526 FFECOM_gfrtTIME): These now use INTEGER*8 for time values,
1527 for compatibility with systems like Alpha.
1528 (FFECOM_gfrtSYSTEM_CLOCK, FFECOM_gfrtTTYNAM): Delete incorrect
1529 trailing underscore in routine names.
1530 * intrin.c, intrin.def: Support INTEGER*8 return values and
1531 arguments ('4'). Change FFEINTRIN_impCTIME, FFEINTRIN_impMCLOCK,
1532 and FFEINTRIN_impTIME accordingly.
1533 (ffeintrin_is_intrinsic): Don't give caller a clue about
1534 form of intrinsic -- shouldn't be needed at this point.
1536 Cope with generic intrinsics that are subroutines and functions:
1537 * com.c (ffecom_finish_symbol_transform_, ffecom_expr_transform_):
1538 Don't transform an intrinsic that is not known to be a subroutine
1539 or a function. (Maybe someday have to avoid transforming
1540 any intrinsic with an undecided or unknown implementation.)
1541 * expr.c (ffeexpr_declare_unadorned_,
1542 ffeexpr_declare_parenthesized_): Ok to invoke generic
1543 intrinsic that has at least one subroutine form as a
1545 Ok to pass intrinsic as actual arg if it has a known specific
1546 intrinsic form that is valid as actual arg.
1547 (ffeexpr_declare_parenthesized_): An unknown kind of
1548 intrinsic has a paren_type chosen based on context.
1549 (ffeexpr_token_arguments_): Build funcref/subrref based
1550 on context, not on kind of procedure being called.
1551 * intrin.h, intrin.c (ffeintrin_is_intrinsic): Undo changes of
1552 Tue Feb 4 23:12:04 1997 by me, change all callers to leave
1553 intrinsics as FFEINFO_kindNONE at this point. (Some callers
1554 also had unused variables deleted as a result.)
1556 Enable all intrinsic groups (especially f90 and vxt):
1557 * target.h (FFETARGET_defaultSTATE_DCP, FFETARGET_defaultSTATE_F2C,
1558 FFETARGET_defaultSTATE_F90, FFETARGET_defaultSTATE_MIL,
1559 FFETARGET_defaultSTATE_UNIX, FFETARGET_defaultSTATE_VXT):
1560 Delete these macros, let top.c set them directly.
1561 * top.c (ffeintrinsic_state_dcp_, ffe_intrinsic_state_f2c_,
1562 ffe_intrinsic_state_f90_, ffe_intrinsic_state_mil_,
1563 ffe_intrinsic_state_unix_, ffe_intrinsic_state_vxt_):
1564 Enable all these directly.
1566 Sat Feb 8 03:21:50 1997 Craig Burley <burley@gnu.ai.mit.edu>
1568 * g77.c: Incorporate recent changes to ../gcc.c.
1569 For version magic (e.g. `g77 -v'), instead of compiling
1570 /dev/null, write, compile, run, and then delete a small
1571 program that prints the version numbers of the three
1572 components of libf2c (libF77, libI77, and libU77),
1573 so we get this info with bug reports.
1574 Also, this change reduces the chances of accidentally
1575 linking to an old (complex-alias-problem) libf2c.
1576 Fix `-L' so the argument is expected in `-Larg'.
1578 * com.h (FFECOM_f2cLONGINT): For INTEGER*8 support in f2c.h,
1579 dynamically determine proper type here, instead of
1580 assuming `long long int' is correct.
1582 Tue Feb 4 23:12:04 1997 Craig Burley <burley@gnu.ai.mit.edu>
1584 Add libU77 library from Dave Love <d.love@dl.ac.uk>:
1585 * Make-lang.in (f77-runtime): Depend on new Makefile.
1586 (f/runtime/libU77/Makefile): New rule.
1587 Also configure libU77.
1588 ($(srcdir)/f/runtime/configure: Use Makefile.in,
1589 so configuration doesn't have to have happened.
1590 (f77.mostlyclean, f77.clean, f77.distclean,
1591 f77.maintainer-clean): Some fixups here, but more work
1593 (RUNTIMESTAGESTUFF): Add libU77's config.status.
1594 (LIBU77STAGESTUFF, f77.stage1, f77.stage2, f77.stage3,
1595 f77.stage4): New macro, appropriate uses added.
1596 * com-rt.def: Add libU77 procedures.
1597 * com.c (ffecom_f2c_ptr_to_integer_type_node,
1598 ffecom_f2c_ptr_to_real_type_node): New type nodes.
1599 (FFECOM_rttypeCHARACTER_): New type of run-time function.
1600 (ffecom_char_args_): Handle CHARACTER*n intrinsics
1601 where n != 1 here, instead of in ffecom_expr_intrinsic_.
1602 (ffecom_expr_intrinsic_): New code to handle new
1604 In particular, change how FFEINTRIN_impFLUSH is handled.
1605 (ffecom_make_gfrt_): Handle new type of run-time function.
1606 (ffecom_init_0): Initialize new type nodes.
1607 * config-lang.in: New libU77 directory.
1608 * intrin.h, intrin.c (ffeintrin_is_intrinsic): Handle
1609 potential generic for subroutine _and_ function
1610 specifics via two new arguments. All callers changed.
1611 Properly ignore deleted/disabled intrinsics in resolving
1613 (ffeintrin_check_, ffeintrin_init_0): Handle CHARACTER intrinsics of (*)
1615 * intrin.def: Permission granted by FSF to place this in
1616 public domain, which will allow it to serve as source
1617 for both g77 program and its documentation.
1618 Add libU77 intrinsics.
1619 (FLUSH): Now a generic, not specific, intrinsic.
1620 (DEFIMP): Now support return modifier for CHARACTER intrinsics.
1622 * com-rt.def (FFECOM_gfrtDIM, FFECOM_gfrtERF,
1623 FFECOM_gfrtERFC, FFECOM_gfrtEXP, FFECOM_gfrtSIGN,
1624 FFECOM_gfrtSIN, FFECOM_gfrtSINH, FFECOM_gfrtTAN,
1625 FFECOM_gfrtTANH, FFECOM_gfrtPOW_RI): Change "&r" to "&f".
1627 Sat Feb 1 12:15:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
1629 * Version 0.5.19.1 released.
1631 * com.c (ffecom_expr_, ffecom_expr_intrinsic_,
1632 ffecom_tree_divide_): FFECOM_gfrtPOW_ZI,
1633 FFECOM_gfrtCONJG, FFECOM_gfrtDCONJG,
1634 FFECOM_gfrtCCOS, FFECOM_gfrtCDCOS,
1635 FFECOM_gfrtCLOG, FFECOM_gfrtCDLOG,
1636 FFECOM_gfrtCSIN, FFECOM_gfrtCDSIN,
1637 FFECOM_gfrtCSQRT, FFECOM_gfrtCDSQRT,
1638 FFECOM_gfrtDIV_CC, FFECOM_gfrtDIV_ZZ: These all require
1639 result to _not_ overlap one or more inputs.
1641 Sat Feb 1 00:25:55 1997 Craig Burley <burley@gnu.ai.mit.edu>
1643 * com.c (ffecom_init_0): Do internal checks only if
1644 -fset-g77-defaults not specified.
1646 Fix %LOC(), LOC() to return sufficiently wide type:
1647 * com.h, com.c (ffecom_pointer_kind_, ffecom_label_kind_,
1648 ffecom_pointer_kind(), ffecom_label_kind()): New globals
1649 and accessor macros hold kind for integer pointers on target
1651 (ffecom_init_0): Determine narrowest INTEGER type that
1652 can hold a pointer (usually INTEGER*4 or INTEGER*8),
1653 store it in ffecom_pointer_kind_, etc.
1654 * expr.c (ffeexpr_cb_end_loc_): Use right type for %LOC().
1655 * intrin.c (ffeintrin_check_, ffeintrin_init_0): Support
1656 new 'p' kind for type of intrinsic.
1657 * intrin.def (FFEINTRIN_impLOC): Returns "Ip" instead of "I1",
1658 so LOC() type is correct for target machine.
1660 Support -fugly-assign:
1661 * lang-options.h, top.h, top.c (ffe_decode_option):
1662 Accept -fugly-assign and -fno-ugly-assign.
1663 * com.c (ffecom_expr_): Handle -fugly-assign.
1664 * expr.c (ffeexpr_finished_): Check right type for ASSIGN
1667 Fri Jan 31 14:30:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
1669 Remove last vestiges of -fvxt-not-f90:
1670 * stb.c (ffestb_R10012_, ffestb_R10014_, ffestb_V0201_):
1673 Fri Jan 31 02:13:54 1997 Craig Burley <burley@gnu.ai.mit.edu>
1675 * top.c (ffe_decode_option): Warn if -fugly is specified,
1678 * symbol.h: No need to #include "bad.h".
1680 Reorganize features from -fvxt-not-f90 to -fvxt:
1681 * lang-options.h, top.h, top.c:
1682 Accept -fvxt and -fno-vxt, but not -fvxt-not-f90 or -ff90-not-vxt.
1683 Warn if the latter two are used.
1684 * expr.c (ffeexpr_nil_rhs_): Double-quote means octal constant.
1685 (ffeexpr_token_rhs_): Double-quote means octal constant.
1686 * target.h (FFETARGET_defaultIS_VXT_NOT_90): Delete macro
1687 definition, no longer needed.
1689 Make some -ff90 features the default:
1690 * data.c (ffedata_value): DATA implies SAVE.
1691 * src.h (ffesrc_is_name_noninit): Underscores always okay.
1693 Fix up some more #error directives by quoting their text:
1694 * bld.c (ffebld_constant_is_zero):
1697 Sat Jan 18 18:22:09 1997 Craig Burley <burley@gnu.ai.mit.edu>
1699 * g77.c (lookup_option, main): Recognize `-Xlinker',
1700 `-Wl,', `-l', `-L', `--library-directory', `-o',
1702 (lookup_option): Don't depend on SWITCH_TAKES_ARG
1703 being correct, it might or might not have `-x' in
1704 it depending on host.
1705 Return NULL argument if it would be an empty string.
1706 (main): If no input files (by gcc.c's definition)
1707 but `-o' or `--output' specified, produce diagnostic
1708 to avoid overwriting output via gcc.
1709 Recognize C++ `+e' options.
1710 Treat -L as another non-magical option (like -B).
1711 Don't append_arg `-x' twice.
1713 Fri Jan 10 23:36:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
1715 * top.c [BUILT_FOR_270] (ffe_decode_option): Make
1716 -fargument-noalias-global the default.
1718 Fri Jan 10 07:42:27 1997 Craig Burley <burley@gnu.ai.mit.edu>
1720 Enable inlining of previously-compiled program units:
1721 * com.c (ffecom_do_entry_, ffecom_start_progunit_):
1722 Register new public function in ffeglobal database.
1723 (ffecom_sym_transform_): Any GLOBAL or potentially GLOBAL
1724 symbol should be looked up in ffeglobal database and
1725 that tree node used, if found. That way, gcc knows
1726 the references are to those earlier definitions, so it
1727 can emit shorter branches/calls, inline, etc.
1728 (ffecom_transform_common_): Minor change for clarity.
1729 * expr.c (ffeexpr_sym-lhs_call_, ffeexpr_sym_lhs_extfunc_,
1730 ffeexpr_sym_rhs_actualarg_, ffeexpr_paren_rhs_let_,
1731 ffeexpr_token_funsubstr_): Globalize symbol as needed.
1732 * global.c (ffeglobal_promoted): New function to look up
1733 existing local symbol in ffeglobal database.
1734 * global.h: Declare new function.
1735 * name.h (ffename_token): New macro, plus alphabetize.
1736 * stc.c (ffestc_R1207_item): Globalize EXTERNAL symbol.
1737 * stu.c (ffestu_sym_end_transition, ffestu_sym_exec_transition):
1738 Globalize symbol as needed.
1739 * symbol.h, symbol.c (ffesymbol_globalize): New function.
1741 Thu Jan 9 14:20:00 1997 Craig Burley <burley@gnu.ai.mit.edu>
1743 * ste.c (ffeste_R809): Produce a diagnostic for SELECT CASE
1744 on CHARACTER type, instead of crashing.
1746 Thu Jan 9 00:52:45 1997 Craig Burley <burley@gnu.ai.mit.edu>
1748 * stc.c (ffestc_order_entry_, ffestc_order_format_,
1749 ffestc_R1226): Allow ENTRY and FORMAT before IMPLICIT
1750 NONE, by having them transition only to state 1 instead
1751 of state 2 (which is disallowed by IMPLICIT NONE).
1753 Mon Jan 6 22:44:53 1997 Craig Burley <burley@gnu.ai.mit.edu>
1755 Fix AXP bug found by Rick Niles (961201-1.f):
1756 * com.c (ffecom_init_0): Undo my 1996-05-14 change, as
1757 it is incorrect and prevented easily finding this bug.
1758 * target.h [__alpha__] (ffetargetReal1, ffetargetReal2):
1759 Use int instead of long.
1760 (ffetarget_cvt_r1_to_rv_, ffetarget_cvt_rv_to_r1_,
1761 ffetarget_cvt_r2_to_rv_, ffetarget_cvt_rv_to_r2_):
1762 New functions that intercede for callers of
1763 REAL_VALUE_(TO|UNTO)_TARGET_(SINGLE|DOUBLE).
1764 All callers changed, and damaging casts to (long *) removed.
1766 Sun Jan 5 03:26:11 1997 Craig Burley <burley@gnu.ai.mit.edu>
1768 * Make-lang.in (g77, g77-cross): Depend on both g77.c and
1769 zzz.c, in $(srcdir)/f/.
1771 Better design for -fugly-assumed:
1772 * stc.c (ffestc_R501_item, ffestc_R524_item,
1773 ffestc_R547_item_object): Pass new is_ugly_assumed flag.
1774 * stt.c, stt.h (ffestt_dimlist_as_expr,
1775 ffestt_dimlist_type): New is_ugly_assumed flag now
1776 controls whether "1" is treated as "*".
1777 Don't treat "2-1" or other collapsed constants as "*".
1779 Sat Jan 4 15:26:22 1997 Craig Burley <burley@gnu.ai.mit.edu>
1781 * stb.c (ffestb_R10012_): Don't confirm on FORMAT(A,)
1782 or even FORMAT(A,,B), as R1229 only warns about the
1783 former currently, and this seems reasonable.
1785 Improvements to diagnostics:
1786 * sta.c (ffesta_second_): Don't add any ffestb parsers
1787 unless they're specifically called for.
1788 Set up ffesta_tokens[0] before calling ffestc_exec_transition,
1789 else stale info might get used.
1790 (ffesta_save_): Do a better job picking which parser to run
1791 after running all parsers with no confirmed possibles.
1792 (FFESTA_maxPOSSIBLES_): Decrease from 100 now that so few
1793 possibles are ever on the list at a given time.
1794 (struct _ffesta_possible): Add named attribute.
1795 (ffesta_add_possible_exec_, ffesta_add_possible_nonexec_):
1796 Make these into macros that call a single function that now
1797 sets the named attribute.
1798 (ffesta_add_possible_unnamed_exec_,
1799 ffeseta_add_possible_unnamed_nonexec_): New macros.
1800 (ffesta_second_): Designate unnamed possibles as
1802 * stb.c (ffestb_R1229, ffestb_R12291_): Use more general
1803 diagnostic, so things like "POINTER (FOO, BAR)" are
1804 diagnosed as unrecognized statements, not invalid statement
1806 * stb.h, stb.c (ffestb_unimplemented): Remove function.
1808 1996-12-30 Dave Love <d.love@dl.ac.uk>
1810 * com.c: #include libU77/config.h
1811 (ffecom_f2c_ptr_to_integer_type_node,
1812 ffecom_f2c_ptr_to_integer_type_node): New variables.
1813 (ffecom_init_0): Use them.
1814 (ffecom_expr_intrinsic_): Many news cases for libU77 intrinsics.
1816 * com-rt.def: New definitions for libU77.
1817 * intrin.def: Likewise. Also correct ftell arg spec.
1819 * Makefile.in (f/runtime/libU77/config.h): New target for com.c
1821 * Make-lang.in (f771): Depend on f/runtime/Makefile for the above.
1823 Sat Dec 28 12:28:29 1996 Craig Burley <burley@gnu.ai.mit.edu>
1825 * stt.c (ffestt_dimlist_type): Treat ([...,]1) in dimlist
1826 as ([...,]*) if -fugly-assumed, so assumed-size array
1827 detected early enough.
1829 Thu Dec 19 14:01:57 1996 Craig Burley <burley@gnu.ai.mit.edu>
1831 * target.h (FFETARGET_REAL_VALUE_FROM_INT_): Conditionalize
1832 definition on BUILT_FOR_280, not BUILT_WITH_280, since
1833 the name of the macro was (properly) changed since 0.5.19.
1835 Fix warnings/errors resulting from ffetargetOffset becoming
1836 `long long int' instead of `unsigned long' as of 0.5.19,
1837 while ffebitCount remains `unsigned long':
1838 * bld.c (ffebld_constantarray_dump): Avoid warnings by
1839 using loop var of appropriate type, and using casts.
1840 * com.c (ffecom_expr_): Use right type for loop var.
1841 (ffecom_sym_transform_, ffecom_transform_equiv_):
1842 Cast to right type in assertions.
1843 * data.c (ffedata_gather_, ffedata_value_): Cast to right
1844 type in assertions and comparisons.
1846 Wed Dec 18 12:07:11 1996 Craig Burley <burley@gnu.ai.mit.edu>
1848 Patch from Alexandre Oliva <oliva@dcc.unicamp.br>:
1849 * Makefile.in (all.indirect): Don't pass -bbigtoc option
1852 Cope with new versions of gcc:
1853 * com.h (BUILT_FOR_280): New macro.
1854 * com.c (ffecom_ptr_to_expr): Conditionalize test of
1856 (ffecom_build_complex_constant_): Conditionalize calling
1857 sequence for build_complex.
1859 Sat Dec 7 07:15:17 1996 Craig Burley <burley@gnu.ai.mit.edu>
1861 * Version 0.5.19 released.
1863 Fri Dec 6 12:23:55 1996 Craig Burley <burley@gnu.ai.mit.edu>
1865 * g77.c: Default to assuming "f77" is in $LANGUAGES, since
1866 the LANGUAGE_F77 macro isn't defined by anyone anymore (but
1867 might as well leave the no-f77 code in just in case).
1868 * Make-lang.in (g77, g77-cross): Don't define LANGUAGE_F77
1871 1996-12-06 Dave Love <d.love@dl.ac.uk>
1873 * Make-lang.in (g77, g77-cross): Revert to building `g77' or not
1874 conditional on `f77' in LANGUAGES.
1876 Wed Dec 4 13:08:44 1996 Craig Burley <burley@gnu.ai.mit.edu>
1878 * Make-lang.in (g77, g77-cross): No libs or lib dependencies
1879 in case where "f77" is not in $LANGUAGES.
1881 * lex.c (ffelex_image_char_, ffelex_file_fixed,
1882 ffelex_file_free): Fixes to properly handle lines with
1883 null character, and too-long lines as well.
1885 * lex.c: Call ffebad_start_msg_lex instead of
1886 ffebad_start_msg throughout.
1888 Sun Dec 1 21:19:55 1996 Craig Burley <burley@gnu.ai.mit.edu>
1890 Fix-up for 1996-11-25 changes:
1891 * com.c (ffecom_member_phase2_): Subtract out 0 offset for
1892 elegance and consistency with EQUIVALENCE aggregates.
1893 (ffecom_sym_transform_): Ditto for LOCAL/COMMON, and
1894 ensure we get the same parent storage area.
1895 * data.c (ffedata_gather_, ffedata_value_): Subtract out
1898 Wed Nov 27 13:55:57 1996 Craig Burley <burley@gnu.ai.mit.edu>
1900 * proj.h: Quote the text of the #error message, to avoid
1901 strange-looking diagnostics from non-gcc ANSI compilers.
1903 * top.c: Make -fno-debug-kludge the default.
1905 Mon Nov 25 20:13:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
1907 Provide more info on EQUIVALENCE mismatches:
1908 * bad.def (FFEBAD_EQUIV_MISMATCH): More detailed message.
1909 * equiv.c (ffeequiv_layout_local_, ffeequiv_layout_cblock):
1910 More details for FFEBAD_EQUIV_MISMATCH.
1912 Fix problem with EQUIVALENCE handling:
1913 * equiv.c (ffeequiv_layout_local_): Redesign algorithm --
1914 old one was broken, resulting in rejection of good code.
1915 (ffeequiv_offset_): Add argument, change callers.
1916 Clean up the code, fix up the (probably unused) negative-value
1918 * com.c (ffecom_sym_transform_): For local EQUIVALENCE
1919 member, subtract out aggregate offset (which is <= 0).
1921 Thu Nov 21 12:44:56 1996 Craig Burley <burley@gnu.ai.mit.edu>
1923 Change type of ffetargetOffset from `unsigned long' to `long long':
1924 * bld.c (ffebld_constantarray_dump): Change printf formats.
1925 * storag.c (ffestorag_dump): Ditto.
1926 * symbol.c (ffesymbol_report): Ditto.
1927 * target.h (ffetargetOffset_f): Ditto and change type itself.
1929 Handle situation where list of languages does not include f77:
1930 * Make-lang.in: Define LANGUAGE_F77 to 1 only if `f77' is in
1931 the $LANGUAGES macro for the build.
1932 * g77.c: Compile to a (nearly) no-op program if LANGUAGE_F77
1933 is not defined to 1.
1935 Fixes to delay confirmation of READ, WRITE, and GOTO statements
1936 so the corresponding assignments to same-named CHAR*(*) arrays
1938 * stb.c (ffestb_R90915_, ffestb_91014_): New functions.
1939 (ffestb_goto3_, ffestb_goto5_): Move confirmation from 3 to 5
1940 for the OPEN_PAREN case.
1941 (ffestb_R9091_, ffestb_R9094_, ffestb_R90913_, ffestb_R90914_,
1942 ffestb_R91012_, ffestb_R91013_): Use new functions, and confirm
1943 except for the OPEN_PAREN case.
1945 Fixes to not confirm declarations with an open paren where
1946 an equal sign or other assignment-like token might be, so the
1947 corresponding assignments to same-named CHAR*(*) arrays work:
1948 (ffestb_decl_entsp_5_): Move assertion so we crash on that first,
1949 if it turns out to be wrong, before the less-debuggable crash
1950 on mistaken confirmation.
1951 (ffestb_decl_entsp_6_, ffestb_decl_entsp_7_, ffestb_decl_entsp_8_):
1952 Include OPEN_PAREN in list of assignment-only tokens.
1954 Fix more diagnosed-crash bugs:
1955 * stu.c (ffestu_sym_end_transition): ANY-ize an adjustable array
1956 with bad dimension expressions even if still stateUNCERTAIN.
1957 (ffestu_symter_end_transition_, ffestu_symter_exec_transition_):
1958 Return TRUE for opANY as well.
1959 For code elegance, move opSYMTER case into first switch.
1961 1996-11-17 Dave Love <d.love@dl.ac.uk>
1963 * lex.c: Fix last change.
1965 1996-11-14 Dave Love <d.love@dl.ac.uk>
1967 * Make-lang.in, config-lang.in: Remove the (broken) libU77 stuff,
1970 Thu Nov 14 15:40:59 1996 Craig Burley <burley@gnu.ai.mit.edu>
1972 * bad.def (FFEBAD_UNIMPL_STMT): Explain that invalid
1973 intrinsic references can trigger this message, too.
1975 1996-11-12 Dave Love <d.love@dl.ac.uk>
1977 * lex.c: Declare dwarfout routines.
1979 * config-lang.in: Sink grep o/p.
1981 Mon Nov 11 14:21:13 1996 Craig Burley <burley@gnu.ai.mit.edu>
1983 * g77.c (main): Might as well print version number
1984 for --verbose as well.
1986 Thu Nov 7 18:41:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
1988 * expr.c, lang-options.h, target.h, top.c, top.h: Split out
1989 remaining -fugly stuff into -fugly-logint and -fugly-comma,
1990 leaving -fugly as simply a `macro' that expands into other
1991 options, and eliminate defaults for some of the ugly stuff
1994 * Make-lang.in (gcc-cross): Compile zzz.c, not version.o (!),
1995 in to get version info for this target.
1997 * config-lang.in: Test for GBE patch application based
1998 on whether 2.6.x or 2.7.x GBE is detected.
2000 Wed Nov 6 14:19:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2002 * Make-lang.in (g77): Compile zzz.c in to get version info.
2003 * g77.c: Add support for --help and --version.
2005 * g77.c (lookup_option): Short-circuit long-winded tests
2006 when second char is not hyphen, just to save a spot of time.
2008 Sat Nov 2 13:50:31 1996 Craig Burley <burley@gnu.ai.mit.edu>
2010 * intrin.def: Add FTELL and FSEEK intrinsics, plus new
2011 `g' codes for alternate-return (GOTO) arguments.
2012 * intrin.c (ffeintrin_check_): Support `g' codes.
2013 * com-rt.def: Add ftell_() and fseek_() to database.
2014 * com.c (ffecom_expr_intrinsic_): Ditto. Also, let each
2015 subroutine intrinsic decide for itself what to do with
2016 tree_type, the default being NULL_TREE once again (so
2017 ffecom_call_ doesn't think it's supposed to cast the
2018 function call to the type in the fall-through case).
2020 * ste.c (ffeste_R909_finish): Don't special-case list-directed
2021 I/O, now that libf2c can return non-zero status codes.
2022 (ffeste_R910_finish): Ditto.
2023 (ffeste_io_call_): Simplify logic.
2025 (ffeste_subr_beru_):
2028 (ffeste_R909_start):
2030 (ffeste_R909_finish):
2031 (ffeste_R910_start):
2033 (ffeste_R910_finish):
2034 (ffeste_R911_start):
2035 (ffeste_R923A): Ditto all the above.
2037 Thu Oct 31 20:56:28 1996 Craig Burley <burley@gnu.ai.mit.edu>
2039 * config-lang.in, Make-lang.in: Rename flag file
2040 build-u77 to build-libu77, for consistency with
2041 install-libf2c and such.
2043 * config-lang.in: Don't complain about failure to patch
2044 if pre-2.7.0 gcc is involved (since our patch for that
2045 doesn't add support for tooning).
2047 Sat Oct 26 05:56:51 1996 Craig Burley <burley@gnu.ai.mit.edu>
2049 * bad.def (FFEBAD_TYPELESS_TOO_LARGE): Remove this
2050 unused and redundant diagnostic.
2052 Sat Oct 26 00:45:42 1996 Craig Burley <burley@gnu.ai.mit.edu>
2054 * target.c (ffetarget_integerhex): Fix dumb bug.
2056 1996-10-20 Dave Love <d.love@dl.ac.uk>
2058 * gbe/2.7.2.1.diff: New file.
2060 * Makefile.in (F771_LDFLAGS): Add -bbigtoc for AIX4.1 up, suggested by
2061 endo@material.tohoku.ac.jp [among others!].
2063 Sat Oct 19 03:11:14 1996 Craig Burley <burley@gnu.ai.mit.edu>
2065 * bad.def, bld.c, bld.h, expr.c, lang-options.h, target.c,
2066 target.h, top.c, top.h (ffebld_constant_new_integerbinary,
2067 ffebld_constant_new_integerhex, ffebld_constant_new_integeroctal,
2068 ffeexpr_token_name_apos_name_, ffetarget_integerbinary,
2069 ffetarget_integerhex, ffetarget_integeroctal): Support
2070 new -fno-typeless-boz option with new functions, mods to
2071 existing octal-handling functions, new macros, new error
2072 messages, and so on.
2074 * com.c, lang-options.h, top.c, top.h (ffecom_notify_primary_entry):
2075 Print program unit name on stderr if -fno-silent (new option).
2077 * lang-options.h, top.c, top.h, stt.c (ffestt_dimlist_as_expr):
2078 Treat ([...,]1) in dimlist as ([...,]*) if -fugly-assumed
2081 * lang-options.h: Comment out options duplicated in gcc/toplev.c,
2082 because, somehow, having them commented in and building on my
2083 DEC Alpha results in a cc1 that always segfaults, and gdb that
2084 also segfaults whenever it debugs it up to init_lex() calling
2087 Thu Oct 17 00:39:27 1996 Craig Burley <burley@gnu.ai.mit.edu>
2089 * stb.c (ffestb_R10013_): Don't change meaning of .sign until
2090 after previous meaning/value used to set sign of value
2093 Sun Oct 13 22:15:23 1996 Craig Burley <burley@gnu.ai.mit.edu>
2095 * top.c (ffe_decode_option): Don't set back-end flags
2096 that are nonexistent prior to gcc 2.7.0.
2098 Sun Oct 13 12:48:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2100 * com.c (convert): Don't convert emulated complex expr to
2101 real (via REALPART_EXPR) if the target type is (emulated)
2104 Wed Oct 2 21:57:12 1996 Craig Burley <burley@gnu.ai.mit.edu>
2106 * com.c (ffecom_debug_kludge_): Set DECL_IN_SYSTEM_HEADER so
2107 -Wunused doesn't complain about these manufactured decls.
2108 (ffecom_expr_): Ditto, for original (non-ASSIGN'ed) variable.
2109 (ffecom_transform_equiv_): Clear DECL_IGNORED_P for aggregate
2110 area so it shows up as a debug-accessible symbol.
2111 (pushdecl): Default for "invented" identifiers (a g77-specific
2112 concept for now) is that they are artificial, in system header,
2113 ignored for debugging purposes, used, and (for types) suppressed.
2114 This ought to be overkill.
2116 Fri Sep 27 23:13:07 1996 Craig Burley <burley@gnu.ai.mit.edu>
2118 * ste.c (ffeste_begin_iterdo_, ffeste_end_iterdo_): Support
2119 one-trip DO loops (F66-style).
2120 * lang-options.h, top.c, top.h (-fonetrip): New option.
2122 Thu Sep 26 00:18:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
2124 * com.c (ffecom_debug_kludge_): New function.
2125 (ffecom_sym_transform_): Use new function for COMMON and EQUIVALENCE
2128 * lang-options.h, top.c, top.h (-fno-debug-kludge):
2131 1996-09-24 Dave Love <d.love@dl.ac.uk>
2133 * Make-lang.in (include/f2c.h):
2134 Remove dependencies on xmake_file and tmake_file.
2135 They expand inconsistently in 2.8 c.f. 2.7; $(GCC_PARTS) depends on
2138 1996-09-22 Dave Love <d.love@dl.ac.uk>
2140 * config-lang.in: Add --enable-libu77 option handling.
2143 Conditionally add --enable-libu77 when running runtime configure.
2144 Define LIBU77STAGESTUFF and use it in relevant rules.
2146 1996-08-21 Dave Love <d.love@dl.ac.uk>
2148 * Make-lang.in (f77-runtime):
2149 `stmp-hdrs' should have been `stmp-headers'.
2151 1996-08-20 Dave Love <d.love@dl.ac.uk>
2153 * Make-lang.in (f77-runtime):
2154 Depend on stmp-hdrs, not stmp-int-hdrs, since libF77
2157 Sat Jun 22 18:17:11 1996 Craig Burley <burley@gnu.ai.mit.edu>
2159 * com.c (ffecom_tree_divide_): Fix RECORD_TYPE case to
2160 look at type of first field, properly, to determine
2161 whether to call c_div or z_div.
2163 Tue Jun 4 04:27:18 1996 Craig Burley <burley@gnu.ai.mit.edu>
2165 * com.c (ffecom_build_complex_constant_): Explicitly specify
2167 (ffecom_expr_): Fix thinko.
2168 (ffecom_2): For COMPLEX_EXPR, explicitly specify TREE_PURPOSE.
2170 Mon May 27 16:23:43 1996 Craig Burley <burley@gnu.ai.mit.edu>
2172 Changes to optionally avoid gcc's back-end complex support:
2173 * com.c (ffecom_stabilize_aggregate_): New function.
2174 (ffecom_convert_to_complex_): New function.
2175 (ffecom_make_complex_type_): New function.
2176 (ffecom_build_complex_constant_): New function.
2177 (ffecom_expr_): For opCONVERT of non-COMPLEX to COMPLEX,
2178 don't bother explicitly converting to the subtype first,
2179 because gcc does that anyway, and more code would have
2180 to be added to find the subtype for the emulated-complex
2182 (ffecom_f2c_make_type_): Use ffecom_make_complex_type_
2183 instead of make_node etc. to make a complex type.
2184 (ffecom_1, ffecom_2): Translate operations on COMPLEX operands
2185 to appropriate operations when emulating complex.
2186 (ffecom_constantunion): Use ffecom_build_complex_constant_
2187 instead of build_complex to build a complex constant.
2188 (ffecom_init_0): Change point at which types are laid out
2189 for improved consistency.
2190 Use ffecom_make_complex_type_ instead of make_node etc.
2191 to make a complex type.
2192 Always calculate storage sizes from TYPE_SIZE, never TYPE_PRECISION.
2193 (convert): Use e, not expr, since we've copied into that anyway.
2194 For RECORD_TYPE cases, do emulated-complex conversions.
2195 (ffecom_f2c_set_lio_code_): Always calculate storage sizes
2196 from TYPE_SIZE, never TYPE_PRECISION.
2197 (ffecom_tree_divide_): Allow RECORD_TYPE to also be handled
2198 by run-time library.
2199 (ffecom_expr_intrinsic_): Handle possible RECORD_TYPE as argument
2202 * top.h, top.c, lang-options.h: Support new -f(no-)emulate-complex option.
2204 * com.c (ffecom_sym_transform_): Clarify and fix typos in comments.
2206 Mon May 20 02:06:27 1996 Craig Burley <burley@gnu.ai.mit.edu>
2208 * target.h: Use new REAL_VALUE_UNTO_TARGET_* macros instead
2209 of REAL_VALUE_FROM_TARGET_DOUBLE and _SINGLE.
2210 Explicitly use long instead of HOST_WIDE_INT for emulation
2211 of ffetargetReal1 and ffetargetReal2.
2213 1996-05-20 Dave Love <d.love@dl.ac.uk>
2216 Test for patch being applied with flag_move_all_movables in toplev.c.
2218 * install.texi (Patching GNU Fortran):
2219 Mention overriding X_CFLAGS rather than
2220 editing proj.h on SunOS4.
2222 * Make-lang.in (F77_FLAGS_TO_PASS):
2223 Add X_CFLAGS (convenient for SunOS4 kluge, in
2225 (f77.{,mostly,dist}clean): Reorder things, in particular not to delete
2226 Makefiles too early.
2228 * g77.c (DEFAULT_SWITCH_TAKES_ARG): Define a la gcc.c in the
2229 current GCC snapshot.
2231 Tue May 14 00:24:07 1996 Craig Burley <burley@gnu.ai.mit.edu>
2233 Changes for DEC Alpha AXP support:
2234 * com.c (ffecom_init_0): REAL_ARITHMETIC means internal
2235 REAL/DOUBLE PRECISION might well have a different size
2236 than the compiled type, so don't crash if this is the
2238 * target.h: Use `int' for ffetargetInteger1,
2239 ffetargetLogical1, and magical tests. Set _f format
2240 strings accordingly.
2242 Tue Apr 16 14:08:28 1996 Craig Burley <burley@gnu.ai.mit.edu>
2244 * top.c (ffe_decode_option): -Wall no longer implies
2247 Sat Apr 13 14:50:06 1996 Craig Burley <burley@gnu.ai.mit.edu>
2249 * com.c (ffecom_char_args_): If item is error_mark_node,
2250 set *length that way, too.
2252 * com.c (ffecom_expr_power_integer_): If either operand
2253 is error_mark_node, return that.
2255 * com.c (ffecom_intrinsic_len_): If item is error_mark_node,
2256 return that for length.
2258 * expr.c (ffeexpr_declare_unadorned_,
2259 ffeexpr_declare_parenthesized_): Instead of crashing
2260 on unexpected contexts, produce a diagnostic.
2262 * intrin.c (ffeintrin_check_), intrin.def (impSIGNAL):
2263 Allow procedure as second arg to SIGNAL intrinsic.
2265 * stu.c (ffestu_symter_end_transition_): New function.
2266 (ffestu_symter_exec_transition_): Return bool arg.
2267 Always transition symbol (don't inhibit when !whereNONE).
2268 (ffestu_sym_end_transition): If DUMMY/LOCAL arg has any
2269 opANY exprs in its dimlist, diagnose it so it doesn't
2270 make it through to later stages that try to deal with
2272 (ffestu_sym_exec_transition): If sym has any opANY exprs
2273 in its dimlist, diagnose it so it becomes opANY itself.
2275 * symbol.c (ffesymbol_error): If token arg is NULL,
2276 just ANY-ize the symbol -- don't produce diagnostic.
2278 Mon Apr 1 10:14:02 1996 Craig Burley <burley@gnu.ai.mit.edu>
2280 * Version 0.5.18 released.
2282 Mon Mar 25 20:52:24 1996 Craig Burley <burley@gnu.ai.mit.edu>
2284 * com.c (ffecom_expr_power_integer_): Don't generate code
2285 that compares COMPLEX (or, as it happens, REAL) via "LT_EXPR",
2286 since the back end crashes on that. (This code would never
2287 be executed anyway, but the test that avoids it has now been
2288 translated to control whether the code gets generated at all.)
2291 * com.c (ffecom_type_localvar_): Handle variable-sized
2292 dimension bounds expressions here, so they get calculated
2293 and saved on procedure entry. Fixes 960323-4.f.
2295 * com.c (ffecom_notify_init_symbol): Symbol has no init
2296 info at all if only zeros have been used to initialize it.
2299 * expr.c, expr.h (ffeexpr_type_combine): Renamed from
2300 ffeexpr_type_combine_ and now a public procedure; last arg now
2301 a token, instead of an internal structure used to extract a token.
2302 Now allows the outputs to be aliased with the inputs.
2303 Now allows a NULL token to mean "don't report error".
2304 (ffeexpr_reduced_bool2_, ffeexpr_reduced_eqop2_,
2305 ffeexpr_reduced_math2_, ffeexpr_reduced_power_,
2306 ffeexpr_reduced_relop2_): Handle new calling sequence for
2307 ffeexpr_type_combine.
2308 * (ffeexpr_convert): Don't put an opCONVERT node
2309 in just because the size is unknown; all downstream code
2310 should be able to deal without it being there anyway, and
2311 getting rid of it allows new intrinsic code to more easily
2312 combine types and such without generating bad code.
2313 * info.c, info.h (ffeinfo_kindtype_max): Rewrite to do
2314 proper comparison of size of types, not just comparison
2315 of their internal kind numbers (so I2.eq.I1 doesn't promote
2316 I1 to I2, rather the other way around).
2317 * intrin.c (ffeintrin_check_): Combine types of arguments
2318 in COL a la expression handling, for greater flexibility
2319 and permissiveness (though, someday, -fpedantic should
2320 report use of this kind of thing).
2321 Make sure Hollerith/typeless where CHARACTER expected is
2322 rejected. This all fixes 960323-2.f.
2324 * ste.c (ffeste_begin_iterdo_): Fix some more type conversions
2325 so INTEGER*2-laden DO loops don't crash at compile time on
2326 certain machines. Believed to fix 960323-1.f.
2328 * stu.c (ffestu_sym_end_transition): Certainly reject
2329 whereDUMMY not in any dummy list, whether stateUNCERTAIN
2330 or stateUNDERSTOOD. Fixes 960323-0.f.
2332 Tue Mar 19 13:12:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
2334 * data.c (ffedata_value): Fix crash on opANY, and simplify
2335 the code at the same time.
2337 * Make-lang.in (f77-runtime): Also depends on lib[FI]77/Makefile...
2338 (include/f2c.h...): ...which in turn depend on */Makefile.in.
2339 (f77.rebuilt): Rebuild runtime stuff too.
2341 * intrin.c (ffeintrin_check_): Accommodate TYPELESS/HOLLERITH
2342 types, convert args as necessary, etc.
2344 * expr.c (ffeexpr_convert): Fix test for TYPELESS/HOLLERITH
2345 to obey the docs; crash if no source token when error.
2346 (ffeexpr_collapse_convert): Crash if no token when error.
2348 Mon Mar 18 15:51:30 1996 Craig Burley <burley@gnu.ai.mit.edu>
2350 * com.c (ffecom_init_zero_): Renamed from
2351 ffecom_init_local_zero_; now handles top-level
2352 (COMMON) initializations too.
2354 * bld.c (ffebld_constant_is_zero):
2355 * com.c (ffecom_symbol_transform_, ffecom_sym_transform_assign_,
2356 ffecom_transform_common_, ffecom_transform_equiv_):
2369 * top.h: All of this is mostly housekeeping-type changes
2370 to support -f(no-)zeros, i.e. not always stuff zero
2371 values into the initializer fields of symbol/storage objects,
2372 but still track that they have been given initial values.
2374 * bad.def: Fix wording for DATA-related diagnostics.
2376 * com.c (ffecom_sym_transform_assign_): Don't check
2377 any EQUIVALENCE stuff for local ASSIGN, the check was
2378 bad (crashing), and it's not necessary, anyway.
2380 * com.c (ffecom_expr_intrinsic_): For MAX and MIN,
2381 ignore null arguments as far arg[123], and fix handling
2382 of ANY arguments. (New intrinsic support now allows
2383 spurious trailing null arguments.)
2385 * com.c (ffecom_init_0): Add HOLLERITH (unsigned)
2386 equivalents for INTEGER*2, *4, and *8, so shift intrinsics
2387 and other things that need unsigned versions of signed
2390 Sat Mar 16 12:11:40 1996 Craig Burley <burley@gnu.ai.mit.edu>
2392 * storag.c (ffestorag_exec_layout): Treat adjustable
2393 local array like dummy -- don't create storage object.
2394 * com.c (ffecom_sym_transform_): Allow for NULL storage
2395 object in LOCAL case (adjustable array).
2397 Fri Mar 15 13:09:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
2399 * com.c (ffecom_sym_transform_): Allow local symbols
2400 with nonconstant sizes (adjustable local arrays).
2401 (ffecom_type_localvar_): Allow dimensions with nonconstant
2402 component (adjustable local arrays).
2403 * expr.c: Various minor changes to handle adjustable
2404 local arrays (a new case of stateUNCERTAIN).
2405 * stu.c (ffestu_sym_end_transition,
2406 ffestu_sym_exec_transition): Ditto.
2407 * symbol.def: Update docs to reflect these changes.
2409 * com.c (ffecom_expr_): Reduce space/time needed for
2410 opACCTER case by handling it here instead of converting
2411 it to opARRTER earlier on.
2412 (ffecom_notify_init_storage): Don't convert ACCTER to ARRTER.
2413 (ffecom_notify_init_symbol): Ditto.
2415 * com.c (ffecom_init_0): Crash and burn if any of the types'
2416 sizes, according to the GBE, disagrees with the sizes of
2417 the FFE's internal implementation. This might catch
2418 Alpha/SGI bugs earlier.
2420 Fri Mar 15 01:09:41 1996 Craig Burley <burley@gnu.ai.mit.edu>
2422 * com-rt.def, com.c, com.h: Changes for rewrite of intrinsic
2424 * com.c (ffecom_arglist_expr_): New function.
2425 (ffecom_widest_expr_type_): New function.
2426 (ffecom_expr_intrinsic_): Reorganize, some rewriting.
2427 (ffecom_f2c_make_type_): Layout complex types.
2428 (ffecom_gfrt_args_): New function.
2429 (ffecom_list_expr): Trivial change for consistency.
2431 * expr.c (ffeexpr_token_name_rhs_): Go back to getting
2432 type from specific, not implementation, info.
2433 (ffeexpr_token_funsubstr_): Set intrinsic implementation too!
2434 * intrin.c: Major rewrite of most portions.
2435 * intrin.def: Major rearchitecting of tables.
2436 * intrin.h (ffeintrin_basictype, ffeintrin_kindtype):
2437 Now (once again) take ffeintrinSpec as arg, not ffeintrinImp;
2438 for now, these return NONE, since they're not really needed
2439 and adding the necessary info to the tables is not trivial.
2440 (ffeintrin_codegen_imp): New function.
2441 * stc.c (ffestc_R1208_item): Change way ffeintrin funcs called,
2442 back to original per above; but comment out the code anyway.
2444 * intrin.c (ffe_init_0): Do internal checks only if
2445 -fset-g77-defaults not specified.
2447 * lang-options.h: Add -fset-g77-defaults option.
2448 * lang-specs.h: Always pass -fset-g77-defaults.
2449 * top.c, top.h: New option.
2451 Sat Mar 9 17:49:50 1996 Craig Burley <burley@gnu.ai.mit.edu>
2453 * Make-lang.in (stmp-int-hdrs): Use --no-validate when
2454 generating the f77.rebuilt files (BUGS, INSTALL, NEWS)
2455 so cross-references can work properly in g77.info
2456 without a lot of hassle. Users can probably deal with
2457 the way they end up looking in the f77.rebuilt files.
2459 * bld.c (ffebld_constant_new_integer4_val): INTEGER*8
2460 support -- new function.
2461 (ffebld_constant_new_logical4_val): New function.
2462 * com.c (ffecom_f2c_longint_type_node): New type.
2463 (FFECOM_rttypeLONGINT_): New return type code.
2464 (ffecom_expr_): Add code to invoke pow_qq instead
2465 of pow_ii for INTEGER4 (INTEGER*8) case.
2466 If ffecom_expr_power_integer_ returns NULL_TREE, just do
2468 (ffecom_make_gfrt_): Handle new type.
2469 (ffecom_expr_power_integer_): Let caller do the work if in
2470 dummy-transforming case, since
2471 caller now knows about INTEGER*8 and such, by returning
2473 * expr.c (ffeexpr_reduced_power_): Complain about non-INTEGER
2474 raised to INTEGER4 (INTEGER*8) power.
2476 * target.c (ffetarget_power_integerdefault_integerdefault):
2478 * com.c (ffecom_expr_power_integer_): Fix (-1)**(-8) and similar
2479 to ABS() the integral result if the exponent is negative
2482 * ste.c (ffeste_begin_iterdo_): Clean up a type ref.
2483 Always convert iteration count to _default_ INTEGER.
2485 * sta.c (ffesta_second_): Add BYTE and WORD type/stmts;
2486 changes by Scott Snyder <snyder@d0sgif.fnal.gov>.
2487 * stb.c (ffestb_decl_recursive): Ditto.
2488 (ffestb_decl_recursive): Ditto.
2489 (ffestb_decl_entsp_2_): Ditto.
2490 (ffestb_decl_entsp_3_): Ditto.
2491 (ffestb_decl_funcname_2_): Ditto.
2492 (ffestb_decl_R539): Ditto.
2493 (ffestb_decl_R5395_): Ditto.
2494 * stc.c (ffestc_establish_declstmt_): Ditto.
2495 * std.c (ffestd_R539item): Ditto.
2496 (ffestd_R1219): Ditto.
2498 * str-1t.fin: Ditto.
2499 * str-2t.fin: Ditto.
2501 * expr.c (ffeexpr_finished_): For DO loops, allow
2502 any INTEGER type; convert LOGICAL (assuming -fugly)
2503 to corresponding INTEGER type instead of always default
2504 INTEGER; let later phases do conversion of DO start,
2505 end, incr vars for implied-DO; change checks for non-integral
2506 DO vars to be -Wsurprising warnings.
2507 * ste.c (ffeste_io_impdo_): Convert start, end, and incr
2508 to type of DO variable.
2510 * com.c (ffecom_init_0): Add new types for [IL][234],
2511 much of which was done by Scott Snyder <snyder@d0sgif.fnal.gov>.
2515 Wed Mar 6 14:08:45 1996 Craig Burley <burley@gnu.ai.mit.edu>
2517 * top.c (ffe_init_gbe_): Make -frerun-loop-opt the default.
2519 Mon Mar 4 12:27:00 1996 Craig Burley <burley@gnu.ai.mit.edu>
2521 * expr.c (ffeexpr_exprstack_push_unary_): Really warn only
2522 about two successive _arithmetic_ operators.
2524 * stc.c (ffestc_R522item_object): Allow SAVE of (understood)
2527 * top.c (ffe_decode_option): New -f(no-)second-underscore options.
2528 * top.h: New options.
2529 * com.c (ffecom_get_external_identifier_, ffecom_get_identifier_):
2532 * Make-lang.in (f77.maintainer-clean): Clean f/BUGS, f/INSTALL,
2534 ($(srcdir)/f/BUGS, $(srcdir)/f/INSTALL, $(srcdir)/f/NEWS):
2536 ($(srcdir)/f/g77.info, $(srcdir)/f/g77.dvi): Depend on
2537 f/bugs.texi and f/news.texi.
2538 (f77.install-man): Install f77 man pages (if enabled).
2539 (f77.uninstall): Uninstall info docs, f77 man pages (if enabled).
2541 * top.c (ffe_init_gbe_): New function.
2542 (ffe_decode_option, ffe_file): Call ffe_init_gbe_ to
2543 set defaults for gcc options.
2545 Sat Jan 20 13:57:19 1996 Craig Burley <burley@gnu.ai.mit.edu>
2547 * com.c (ffecom_get_identifier_): Eliminate needless
2548 comparison of results of strchr.
2550 Tue Dec 26 11:41:56 1995 Craig Burley <burley@gnu.ai.mit.edu>
2552 * Make-lang.in: Add rules for new files g77.texi, g77.info,
2554 Reorganize the *clean rules to more closely parallel gcc's.
2556 * config-lang.in: Exclude g77.info from diffs.
2558 Sun Dec 10 02:29:13 1995 Craig Burley <burley@gnu.ai.mit.edu>
2560 * expr.c (ffeexpr_declare_unadorned_,
2561 ffeexpr_declare_parenthesized_): Break out handling of
2562 contextDATAIMPDO[INDEX,CTRL] so it's independent of symbol state.
2563 Don't exec-transition these here (let ffeexpr_sym_impdoitem_
2564 handle that when appropriate). Don't "declare" them twice.
2566 Tue Dec 5 06:48:26 1995 Craig Burley <burley@gnu.ai.mit.edu>
2568 * stc.c (ffestc_promote_sfdummy_): Allow whereNONE parent
2569 symbol, since it is not necessarily known whether it will
2570 become LOCAL or DUMMY.
2572 Mon Dec 4 03:46:55 1995 Craig Burley <burley@gnu.ai.mit.edu>
2574 * lex.c (ffelex_display_token, ffelex_type_string_): Resurrect
2575 these from their old versions and update them for possible invocation
2577 * lex.h (ffelex_display_token): Declare this in case anyone
2578 else wants to call it.
2580 * lex.c (ffelex_total_tokens_): Have this reflect actual allocated
2581 tokens, no longer include outstanding "uses" of tokens.
2583 * malloc.c, malloc.h (MALLOC_DEBUG): New macro to control
2584 checking of whether callers follow rules, now defaults to 0
2585 for "no checking" to improve compile times.
2587 * malloc.c (malloc_pool_kill): Fix bug that could prevent
2588 subpool from actually being killed (wasn't setting its use
2591 * proj.h, *.c (dmpout): Replace all occurrences of `stdout'
2592 and some of `stderr' with `dmpout', so where to dump debugging
2593 output can be easily controlled during build; add default
2594 for `dmpout' of `stderr' to proj.h.
2596 Sun Dec 3 00:56:29 1995 Craig Burley <burley@gnu.ai.mit.edu>
2598 * com.c (ffecom_return_expr): Eliminate attempt at warning
2599 about unset return values, since the back end does this better,
2600 with better wording, and is not triggered by clearly working
2601 (but spaghetti) code as easily as this test.
2603 Sat Dec 2 08:28:56 1995 Craig Burley <burley@gnu.ai.mit.edu>
2605 * target.c (ffetarget_power_*_integerdefault): Raising 0 to
2606 integer constant power should not be an error condition;
2607 if so, other code should catch 0 to any power, etc.
2609 * bad.def (FFEBAD_BAD_POWER): 0**integer now a warning instead
2612 Fri Dec 1 00:12:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
2614 * bad.def: Clarify diagnostic regarding complex constant elements.
2615 * expr.c (ffeexpr_cb_comma_c_): Capitalize real/imaginary
2616 for clarified diagnostic.
2618 * com.c (ffecom_close_include_): Close the file!
2620 * lex.c (ffelex_file_fixed): Update line info if the line
2621 has any content, not just if it finishes a previous line
2623 (ffelex_file_free): Clarify switch statement code.
2625 Sat Nov 18 19:37:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
2627 * Version 0.5.17 released.
2629 Fri Nov 17 14:27:24 1995 Craig Burley (burley@gnu.ai.mit.edu)
2631 * Make-lang.in: Fix typo in comment.
2633 * Makefile.in (f/fini.o, f/proj-h.o): Don't use `$<' since
2634 not all makes support it (e.g. NeXT make), use explicit
2635 source name instead (with $(srcdir) and munging).
2636 (ASSERT_H): assert.h lives in source dir, not build dir.
2638 Thu Nov 16 12:47:50 1995 Craig Burley (burley@gnu.ai.mit.edu)
2640 * com.c (ffecom_init_0): Fix dumb bug in code to produce
2641 warning message about non-32-bit-systems.
2643 * stc.c (ffestc_R501_item): Parenthesize test to make
2644 warning go away (and perhaps fix bug).
2646 Thu Nov 16 03:43:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
2648 * g77.c: Upgrade to 2.7.0's gcc.c.
2649 Fix -v to pass a temp name instead of "/dev/null" for "-o".
2651 Fri Nov 10 19:16:05 1995 Craig Burley (burley@gnu.ai.mit.edu)
2653 * ste.c (ffeste_begin_iterdo_): Add Toon's change to
2654 make loops faster on some machines (implement termination
2655 condition as "--i >= 0" instead of "i-- > 0").
2657 Thu Nov 2 03:58:17 1995 Craig Burley (burley@gnu.ai.mit.edu)
2659 * Make-lang.in: Remove unnecessary $(exeext) a la cp/Make-lang.in.
2661 * com.c (ffecom_expr_): Restore old strategy for assignp variant
2662 of opSYMTER case...always return the ASSIGN version of var.
2663 That way, `-O -Wuninitialized' will catch "I=3;GOTO I;END"
2664 (though the diagnostic will refer to `__g77_ASSIGN_i').
2666 * com.c (ffecom_expr_power_integer_): For constant rhs case,
2667 wrap every new eval of lhs in save_expr() so it is clear to
2668 back end that MULT_EXPR(lhs,lhs) has identical operands,
2669 otherwise for an rhs like 32767 it generates around 65K pseudo
2670 registers, with which stupid_life_analysis cannot cope
2671 (due to reg_renumber in regs.h being `short *' instead of
2674 * com.c (ffecom_expr_): Speed up implementation of LOGICAL
2675 versions of opNOT, opAND, opOR, opXOR/opNEQV, and opEQV by
2676 assuming the values actually are kosher LOGICAL bit patterns.
2677 Also simplify code that implements some of the INTEGER versions
2680 * com.c (skip_redundant_dir_prefix, read_name_map,
2681 ffecom_open_include_, signed_type, unsigned_type): Fold in
2682 changes to cccp.c made from 2.7.0 through ss-950826.
2684 * equiv.c (ffeequiv_layout_local_): Kill the equiv list
2687 * expr.c (ffeexpr_reduced_eqop2_): Issue specific diagnostic
2688 regarding usage of .EQV./.NEQV. in preference to .EQ./.NE..
2690 * intrin.c: Add ERF and ERFC as generic intrinsics.
2693 * sta.c (ffesta_save_, ffesta_second_): Whoever calls
2694 ffestd_exec_begin must also set ffesta_seen_first_exec = TRUE,
2695 and anytime stc sees an exec transition, it must do both.
2696 stc.c (ffestc_eof): Same.
2698 * stc.c (ffestc_promote_sfdummy_): If failed implicit typing
2699 or CHARACTER*(*) arg, after calling ffesymbol_error, don't
2700 reset info to ENTITY/DUMMY, because ffecom_sym_transform_
2701 doesn't expect such a thing with ANY/ANY type.
2703 * target.h (*logical*): Change some of these so they parallel
2704 changes in com.c, e.g. for _eqv_, use (l)==(r) instead of
2705 !!(l)==!!(r), to get a more faithful result.
2707 Fri Oct 27 07:06:59 1995 Craig Burley (burley@gnu.ai.mit.edu)
2709 * com.c (ffecom_sym_transform_): Simplify code for local
2712 * expr.c (ffeexpr_exprstack_push_unary_): Warn about two
2713 successive operators.
2714 (ffeexpr_exprstack_push_binary_): Warn about "surprising"
2715 operator precedence, as in "-2**2".
2717 * lang-options.h: Add -W(no-)surprising options.
2719 * parse.c (yyparse): Don't reset -fpedantic if not -pedantic.
2721 * top.c (ffe_decode_option): Support new -Wsurprising option.
2724 Mon Oct 23 09:14:15 1995 Craig Burley (burley@gnu.ai.mit.edu)
2726 * com.c (ffecom_finish_symbol_transform_): Don't transform
2727 NONE/NONE (CHARACTER*(*)) vars, as these don't mean anything
2728 in debugging terms, and can't be turned into anything
2729 in the back end (so ffecom_sym_transform_ crashes on them).
2731 * com.c (ffecom_expr_): Change strategy for assignp variant
2732 of opSYMTER case...always return the original var unless
2733 it is not wide enough.
2735 * ste.c (ffeste_io_cilist_): Clarify diagnostic for ASSIGN
2736 involving too-narrow variable. This shouldn't happen, though.
2737 (ffeste_io_icilist_): Ditto.
2738 (ffeste_R838): Ditto.
2739 (ffeste_R839): Ditto.
2741 Thu Oct 19 03:21:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
2743 * com.c (ffecom_sym_transform_assign_): Set TREE_STATIC
2744 using the same decision-making process as used for their twin
2745 variables, so ASSIGN can last across RETURN/CALL as appropriate.
2747 Fri Sep 22 20:21:18 1995 Craig Burley (burley@gnu.ai.mit.edu)
2749 * Makefile.in: fini is a host program, so it needs a host-compiled
2750 version of proj.o, named proj-h.o. f/fini, f/fini.o, and
2751 f/proj-h.o targets updated accordingly.
2753 * com.c (__eprintf): New function.
2755 Wed Sep 20 02:26:36 1995 Craig Burley (burley@gnu.ai.mit.edu)
2757 * lang-options.h: Add omitted -funix-intrinsics-* options.
2759 * malloc.c (malloc_find_inpool_): Check for infinite
2760 loop, crash if detected (user reports encountering
2761 them in some large programs, this might help track
2764 Thu Sep 7 13:00:32 1995 Craig Burley (burley@gnu.ai.mit.edu)
2766 * com.c (lang_print_error_function): Don't dereference null
2767 pointer when outside any program unit.
2768 (ffecom_let_char_, ffecom_arg_ptr_to_expr): If catlist
2769 item or length ever error_mark_node, don't continue processing,
2770 since back-end functions like build_pointer_type crash on
2771 error_mark_node's (due to pushing bad obstacks, etc.).
2773 Wed Aug 30 15:58:35 1995 Craig Burley (burley@gnu.ai.mit.edu)
2775 * Version 0.5.16 released.
2777 Mon Aug 28 12:24:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
2779 * bad.c (ffebad_finish): Fix botched message when no places
2780 are printed (due to unknown line info, etc.).
2782 * std.c (ffestd_subr_labels_): Do a better job finding
2783 line info in the case of typeANY and diagnostics.
2785 Fri Aug 25 15:19:29 1995 Craig Burley (burley@gnu.ai.mit.edu)
2787 * com.c (DECL_ARTIFICIAL): Surround all references to this
2788 macro with #if !BUILT_FOR_270 and #endif.
2789 (init_lex): Surround print_error_function decl with
2790 #if !BUILT_FOR_270 and #endif.
2791 (lang_init): Call new ffelex_hash_kludge function to solve
2792 problem with preprocessed files that have INCLUDE statements.
2794 * lex.c (ffelex_getc_): New function.
2795 (ffelex_cfelex_): Use ffelex_getc_ instead of getc in any
2796 paths of code that can be affected by ffelex_hash_kludge.
2797 Don't make an EOF token for unrecognized token; set token
2798 to NULL instead, to avoid problems when not initialized.
2799 (ffelex_hash_): Use ffelex_getc_ instead of getc in any
2800 paths of code that can be affected by ffelex_hash_kludge.
2801 Test token returned by ffelex_cfelex_ for NULL, meaning
2803 Get rid of useless used_up variable.
2804 Don't do ffewhere stuff or kill any tokens if in
2806 (ffelex_file_fixed, ffelex_file_free): Use ffelex_getc_
2807 instead of getc in any paths of code that can be affected
2808 by ffelex_hash_kludge.
2809 (ffelex_hash_kludge): New function.
2811 * lex.h (ffelex_hash_kludge): New function.
2813 Wed Aug 23 15:17:40 1995 Craig Burley (burley@gnu.ai.mit.edu)
2815 * com.c: Implement -f(no-)underscoring options by always
2816 compiling in code to do it, and having that code inhibit
2817 itself when -fno-underscoring is in effect. This option
2818 overrides -f(no-)f2c for this purpose; -f(no-)f2c returns
2819 to it's <=0.5.15 behavior of affecting only how code
2820 is generated, not how/whether names are mangled.
2822 * target.h: Redo specification of appending underscores so
2823 the macros are named "_default" instead of "_is" and the
2824 two-underscore macro defaults to 1.
2826 * top.c, top.h (underscoring): Add appropriate stuff
2827 for the -f(no-)underscoring options.
2829 Tue Aug 22 10:25:01 1995 Craig Burley (burley@gnu.ai.mit.edu)
2831 * bad.c (ffebad_finish): Call report_error_function (in toplev.c)
2832 to better identify location of problem.
2833 Say "(continued):" instead of "(continued:)" for consistency.
2835 * com.c (ffecom_gen_sfuncdef_): Set and reset new
2836 ffecom_nested_entry_ variable to hold ffesymbol being compiled.
2837 (lang_print_error_function): New function from toplev.c.
2838 Use ffecom_nested_entry_ to help determine which name
2839 and kind-string to print.
2840 (ffecom_expr_intrinsic_): Handle EXIT and FLUSH invocations
2841 with different calling sequences than library functions.
2842 Have SIGNAL and SYSTEM push and pop calltemps, and convert
2843 their return values to the destination type (just in case).
2844 (FFECOM_rttypeINT_): New return type for `int', in case
2845 gcc/f/runtime/libF77/system_.c(system_) is really supposed
2846 to return `int' instead of `ftnint'.
2848 * com.h (report_error_function): Declare this.
2850 * equiv.c (ffeequiv_layout_local_): Don't forget to consider
2851 root variable itself as possible "first rooted variable",
2852 else might never set symbol and then crash later.
2854 * intrin.c (ffeintrin_check_exit_): Change to allow no args
2855 and rename to ffeintrin_check_int_1_o_ for `optional'.
2856 #define ffeintrin_check_exit_ and _flush_ to this new
2857 function, so intrin.def can refer to the appropriate names.
2859 * intrin.def (FFEINTRIN_impFLUSH): Validate using
2860 ffeintrin_check_flush_ so passing an INTEGER arg is allowed.
2862 * lex.c (ffelex_file_push_, ffelex_file_pop_): New functions
2863 to manage input_file_stack in gbe.
2864 (ffelex_hash_): Call new functions (instead of doing code).
2865 (ffelex_include_): Call new functions to update stack for
2866 INCLUDE (_hash_ handles cpp output of #include).
2868 Mon Aug 21 08:09:04 1995 Craig Burley (burley@gnu.ai.mit.edu)
2870 * Makefile.in: Put `-W' in front of every `-Wall', since
2871 2.7.0 requires that to engage `-Wunused' for parameters.
2873 * com.c: Mark all parameters as artificial, so
2874 `-W -Wunused' doesn't complain about unused ones (since
2875 there's no way right not to individually specify attributes
2878 * proj.h: Don't #define UNUSED if already defined, regardless
2881 Sun Aug 20 16:03:56 1995 Craig Burley (burley@gnu.ai.mit.edu)
2883 * gbe/2.7.0.diff: Regenerate.
2885 * lang-options.h, lang-specs.h: If not __STDC__ (ANSI C),
2886 avoid doing anything, especially the stringizing in -specs.h.
2888 Thu Aug 17 03:36:12 1995 Craig Burley (burley@gnu.ai.mit.edu)
2890 * lang-specs.h: Remove useless optional settings of -traditional,
2891 since -traditional is always set anyway.
2893 Wed Aug 16 16:56:46 1995 Craig Burley (burley@gnu.ai.mit.edu)
2895 * Make-lang.in (F2C_INSTALL_FLAG, F2CLIBOK): More
2896 control over whether to install f2c-related stuff.
2897 (install-f2c-*): New targets to install f2c-related
2898 stuff in system, not just gcc, directories.
2900 * com.c: Change calls to ffecom_get_invented_identifier
2901 to use generally more predictable names.
2902 Change calls to build_range_type to ensure consistency
2903 of types of operands.
2904 (ffecom_get_external_identifier_): Change to accept
2905 symbol info, not just text, so it can use f2c flag for
2906 symbol to decide whether to append underscore(s).
2907 (ffecom_get_identifier_): Don't change names if f2c flag
2908 off for compilation.
2909 (ffecom_type_permanent_copy_): Use same type for new max as
2911 (ffecom_notify_init_storage): Offline fixups for stand-alone.
2913 * data.c (ffedata_gather): Explicitly test for common block,
2914 since it's no longer always the case that a local EQUIVALENCE
2915 group has no symbol ptr (it now can, if a user-predictable
2916 "rooted" symbol has been identified).
2918 * equiv.c: Add some debugging stuff.
2919 (ffeequiv_layout_local_): Set symbol ptr with user-predictable
2920 "rooted" symbol, for giving the invented aggregate a
2923 * g77.c (append_arg): Allow for 20 extra args instead of 10.
2924 (main): For version-only case, add `-fnull-version' and, unless
2925 explicitly omitted, `-lf2c -lm'.
2927 * lang-options.h: New "-fnull-version" option.
2929 * lang-specs.h: Support ".fpp" suffix for preprocessed source
2930 (useful for OS/2, MS-DOS, other case-insensitive systems).
2932 * stc.c (ffestc_R544_equiv_): Swap way lists are merged so this
2933 is consistent with the order in which lists are built, making
2934 user predictability of invented aggregate name much higher.
2936 * storag.c, storag.h (FFESTORAG_typeDUMMY): Delete this enum.
2938 * top.c: Accept, but otherwise ignore, `-fnull-version'.
2940 Tue Aug 15 07:01:07 1995 Craig Burley (burley@gnu.ai.mit.edu)
2942 * DOC, INSTALL, PROJECTS: Extensive improvements to documentation.
2944 Sun Aug 13 01:55:18 1995 Craig Burley (burley@gnu.ai.mit.edu)
2946 * INSTALL (f77-install-ok): Document the use of this file.
2948 * Make-lang.in (F77_INSTALL_FLAG): New flag to control
2949 whether to install an `f77' command (based on whether
2950 a file named `f77-install-ok' exists in the source or
2951 build directory) to replace the broken attempt to use
2952 comment lines to avoid installing `f77' (broken in the
2953 sense that it prevented installation of `g77').
2955 Mon Aug 7 06:14:26 1995 Craig Burley (burley@gnu.ai.mit.edu)
2957 * DOC: Add new sections for g77 & gcc compiler options,
2958 source code form, and types, sizes and precisions.
2959 Remove lots of old "delta-version" info, or at least
2962 * INSTALL: Add info here that used to be in DOC.
2965 * g77.c (lookup_option, main): Check for --print-* options,
2966 so we avoid adding version-determining stuff.
2968 Wed Jul 26 15:51:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
2970 * Make-lang.in, Makefile.in (input.j, INPUT_H): New file.
2971 Update dependencies accordingly.
2973 * bad.c (ffebad_here): Okay to use unknown line/col.
2975 * compilers.h (@f77-cpp-input): Remove -P option now that
2976 # directives are handled by f771. Update other options
2977 to be more consistent with @c in gcc/gcc.c. Don't run f771
2978 if -E specified, etc., a la @c.
2979 (@f77): Don't run f771 if -E specified, etc., a la @c.
2981 * config-lang.in: Avoid use of word "guaranteed".
2983 * input.j: New file to wrap around gcc/input.h.
2985 * lex.j: Add support for parsing # directives output by cpp.
2986 (ffelex_cfebackslash_): New function.
2987 (ffelex_cfelex_): New function.
2988 (ffelex_get_directive_line_): New function.
2989 (ffelex_hash_): New function.
2990 (ffelex_include_): Change to not use ffewhere_file_(begin|end).
2991 Also fix bug in pointing to next line (for diagnostics, &c)
2992 following successful INCLUDE.
2993 (ffelex_next_line_): New function that does chunk of code
2994 seen in several places elsewhere in the lexers.
2995 (ffelex_file_fixed): Delay finishing statement until source
2996 line is registered with ffewhere, so INCLUDE processing
2997 picks up the info correctly.
2998 Okay to kill or use unknown line/col objects now.
2999 Handle HASH (#) lines.
3000 Reorder tests for insubstantial lines to put most frequent
3001 occurrences at top, for possible minor speedup.
3002 Some general consolidation of code.
3003 (ffelex_file_free): Handle HASH (#) lines.
3004 Okay to kill or use unknown line/col objects now.
3005 Some general consolidation of code.
3006 (ffelex_init_1): Detect HASH (#) lines.
3007 (ffelex_set_expecting_hollerith): Okay to kill or use unknown
3008 line/col objects now.
3010 * lex.h (FFELEX_typeHASH): New enum.
3012 * options-lang.h (-fident, -fno-ident): New options.
3014 * stw.c (ffestw_update): Okay to kill unknown line/col objects
3017 * target.h (FFETARGET_okREALQUAD, FFETARGET_okCOMPLEXDOUBLE,
3018 FFETARGET_okCOMPLEXQUAD): #define these appropriately.
3020 * top.c: Include flag.j wrapper, not flags.h directly.
3021 (ffe_is_ident_): New flag.
3022 (ffe_decode_option): Handle -fident and -fno-ident.
3023 (ffe_file): Replace obsolete ffewhere_file_(begin|end) with
3026 * top.h (ffe_is_ident_, ffe_is_ident, ffe_set_is_ident):
3027 New flag and access functions.
3029 * where.c, where.h: Remove all tracking of parent file.
3030 (ffewhere_file_begin, ffewhere_file_end): Delete these.
3031 (ffewhere_line_use): Make it work with unknown line object.
3033 Mon Jul 17 03:04:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3035 * com.c (ffecom_sym_transform_): Set DECL_IN_SYSTEM_HEADER
3036 flag for any local vars used as stmtfunc dummies or DATA
3037 implied-DO iter vars, so no -Wunused warnings are produced
3038 for them (a la f2c).
3039 (ffecom_init_0): Do "extern int xargc;" for IARGC() intrinsic.
3040 Warn if target machine not 32 bits, since g77 isn't yet
3041 working on them at all well.
3043 * expr.c (ffeexpr_sym_lhs_call_, ffeexpr_sym_lhs_data_,
3044 ffeexpr_sym_lhs_extfunc_, ffeexpr_sym_rhs_actualarg_,
3045 ffeexpr_sym_rhs_let_, ffeexpr_paren_rhs_let_): Don't
3046 gratuitously set attr bits that don't apply just
3047 to avoid null set meaning error; instead, use explicit
3048 error flag, and allow null attr set, to
3049 fix certain bugs discovered by looking at this code.
3051 * g77.c: Major changes to improve support for gcc long options,
3052 to make `g77 -v' report more useful info, and so on.
3054 Mon Jul 3 14:49:16 1995 Craig Burley (burley@gnu.ai.mit.edu)
3056 * DOC, com.c, intrin.h, intrin.c, intrin.def, target.h, top.c,
3057 top.h: Add new `unix' group of intrinsics, which includes the
3058 newly added ERF, ERFC, EXIT, plus even newer ABORT, DERF, DERFC,
3059 FLUSH, GETARG, GETENV, SIGNAL, and SYSTEM.
3061 Tue Jun 27 23:01:05 1995 Craig Burley (burley@gnu.ai.mit.edu)
3063 * bld.c, bld.h (ffebld_constant_pool,
3064 ffebld_constant_character_pool): Use a single macro (the
3065 former) to access the pool for allocating constants, instead
3066 of latter in public and FFEBLD_CONSTANT_POOL_ internally
3067 in bld.c (which was the only one that was correct before
3068 these changes). Add verification of integrity of certain
3069 heap-allocated areas.
3071 * com.c (ffecom_overlap_, ffecom_args_overlap_,
3072 ffecom_tree_canonize_ptr_, ffecom_tree_canonize_ref_): New
3073 functions to optimize calling COMPLEX and, someday, CHARACTER
3074 functions requiring additional argument to be passed.
3075 (ffecom_call_, ffecom_call_binop_, ffecom_expr_,
3076 ffecom_expr_intrinsic_): Change calling
3077 sequences to include more info on possible destination.
3078 (ffecom_expr_intrinsic_): Add ERF(), ERFC(), and EXIT()
3080 (ffecom_sym_transform_): For assumed-size arrays, set high
3081 bound to highest possible value instead of low bound, to
3082 improve validity of overlap checking.
3083 (duplicate_decls): If olddecl and newdecl are the same,
3084 don't do any munging, just return affirmative.
3086 * expr.c: Change ffecom_constant_character_pool() to
3087 ffecom_constant_pool().
3089 * info.c (ffeinfo_new): Compile this version if not being
3092 * info.h (ffeinfo_new): Don't define macro if not being
3095 * intrin.c, intrin.def: Add ERF(), ERFC(), and EXIT() intrinsics.
3096 (ffeintrin_check_exit_): New for EXIT() subroutine intrinsic.
3098 * malloc.c, malloc.h (malloc_verify_*): New functions to verify
3099 integrity of heap-storage areas.
3101 * stc.c (ffestc_R834, ffestc_R835): Handle possibility that
3102 an enclosing DO won't have a construct name even when the
3103 CYCLE/EXIT does (i.e. without dereferencing NULL).
3105 * target.c, target.h (ffetarget_verify_character1): New function
3106 to verify integrity of heap storage used to hold character constant.
3108 Thu Jun 22 15:36:39 1995 Howard Gordon (flash@super.org)
3110 * stp.h (ffestpVxtcodeIx): Fix typo in typedef for this.
3112 Mon May 29 15:22:31 1995 Craig Burley (burley@gnu.ai.mit.edu)
3114 * *: Make all sorts of changes to accommodate upcoming gcc-2.7.0.
3115 I didn't keep track of them, nor just when I made them, nor
3116 when I (much later, probably in early August 1995) modified
3117 them so they could properly handle both 2.7.0 and 2.6.x.
3119 * com.c (ffecom_expr_power_integer_): Don't expand_start_stmt_expr
3120 if transforming dummy args, because the back end cannot handle
3121 that (it's rejected by the gcc front end), just generate
3122 call to run-time library.
3123 Back out changes in 0.5.15 because more temporaries might be
3124 needed anyway (for COMPLEX**INTEGER).
3125 (ffecom_push_tempvar): Remove inhibitor.
3126 Around start_decl and finish_decl (in particular, arround
3127 expand_decl, which is called by them), push NULL_TREE into
3128 sequence_rtl_expr, an external published by gcc/function.c.
3129 This makes sure the temporary is truly in the function's
3130 context, not the inner context of a statement-valued expression.
3131 (I think the back end is inconsistent here, but am not
3132 interested in convincing the gbe maintainers about this now.)
3133 (pushdecl): Make sure that when pushing PARM_DECLs, nothing
3134 other than them are pushed, as happened for 0.5.15 and which,
3135 if done for other reasons not fixed here, might well indicate
3136 some other problem -- so crash if it happens.
3138 * equiv.c (ffeequiv_layout_local_): If the local equiv group
3139 has a non-nil COMMON field, it should mean that an error has
3140 occurred and been reported, so just trash the local equiv
3141 group and do nothing.
3143 * stc.c (ffestc_promote_sfdummy_): Set sfdummy arg state to
3144 UNDERSTOOD so above checking for duplicate args actually
3145 works, and so we don't crash later in pushdecl.
3147 * ste.c (ffeste_R1001): Set initial value only for VAR_DECLs,
3148 not for, e.g., LABEL_DECLs, which the FORMAT label can be
3149 if it was previously treated as an executable label.
3151 Sat May 20 01:53:53 1995 Craig Burley (burley@gnu.ai.mit.edu)
3153 * com.c (ffecom_sym_transform_): For adjustable arrays,
3154 pass high bound through variable_size in case its primaries
3155 are changed (dumb0.f, and this might also improve
3156 performance so it approaches f2c|gcc).
3158 Fri May 19 11:00:36 1995 Craig Burley (burley@gnu.ai.mit.edu)
3160 * Version 0.5.15 released.
3162 * com.c (ffecom_expr_power_integer_): Push temp vars
3163 before expanding a statement expression, since that seems
3164 to cause temp vars to be "forgotten" after the end of the
3165 expansion in the back end. Disallow more temp-var
3166 pushing during such an expansion, just in case.
3167 (ffecom_push_tempvar): Crash if a new variable needs to be
3168 pushed but cannot be at this point (should never happen).
3170 Wed May 17 12:26:16 1995 Craig Burley (burley@gnu.ai.mit.edu)
3172 * expr.c (ffeexpr_collapse_convert): Add code to convert
3173 LOGICAL to CHARACTER. Reject conversion of REAL or COMPLEX
3174 to CHARACTER entirely, as it cannot be supported with all
3177 * target.h, target.c (ffetarget_convert_character1_logical1):
3180 Sun May 14 00:00:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3182 * com.c (ffecom_do_entry_, ffecom_gen_sfuncdef_,
3183 ffecom_start_progunit_, ffecom_sym_transform_,
3184 ffecom_init_0, start_function): Changes to have REAL
3185 external functions return same type as DOUBLE PRECISION
3186 external functions when -ff2c is in force; while at it,
3187 some code cleanups done.
3189 * stc.c (ffestc_R547_item_object): Disallow array declarator
3190 if one already exists for symbol.
3192 * ste.c (ffeste_R1227): Convert result variable to type
3193 of function result as seen by back end (e.g. for when REAL
3194 external function actually returns result as double).
3196 * target.h (FFETARGET_defaultFIXED_LINE_LENGTH): New
3197 macro for default for -ffixed-line-length-N option.
3199 * top.c (ffe_fixed_line_length_): Initialize this to new
3200 target.h macro instead of constant 72.
3202 Tue May 9 01:20:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
3204 * lex.c (ffelex_send_token_): If sending CHARACTER token with
3205 null text field, put a single '\0' in it and set length/size
3206 fields to 0 (to fix 950508-0.f).
3207 (ffelex_image_char_): When setting ffelex_bad_line_ to TRUE,
3208 always "close" card image by appending a null char and setting
3209 ffelex_card_length_. As part of this, append useful text
3210 to identify the two kinds of problems that involve this.
3211 (ffelex_file_fixed): Reset ffelex_bad_line_ to FALSE after
3212 seeing a line with invalid first character (fixes 950508-1.f).
3213 If final nontab column is zero, assume tab seen in line.
3214 (ffelex_card_image_): Always make this array 8 characters
3215 longer than reflected by ffelex_card_size_.
3216 (ffelex_init_1): Get final nontab column info from top instead
3219 * options-lang.h: Add -ffixed-line-length- prefix.
3221 * top.h: Add ffe_fixed_line_length() and _set_ version, plus
3222 corresponding extern.
3224 * top.c: Handle -ffixed-line-length- option prefix.
3226 Fri Apr 28 05:40:25 1995 Craig Burley (burley@gnu.ai.mit.edu)
3228 * Version 0.5.14 released.
3230 * Make-lang.in: Add assert.j.
3232 * Makefile.in: Add assert.j.
3234 * assert.j: New file.
3236 Thu Apr 27 16:24:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
3238 * bad.h (ffebad_severity): New function.
3240 * bad.c (ffebad_severity): New function.
3242 * bad.def (FFEBAD_OPEN_INCLUDE): Change severity from SEVERE
3243 to FATAL, since processing continues, and that seems fine.
3245 * com.c: Add facility to handle -I.
3246 (ffecom_file, ffecom_close_include, ffecom_open_include,
3247 ffecom_decode_include_option): New global functions for -I.
3248 (ffecom_file_, ffecom_initialize_char_syntax_,
3249 ffecom_close_include_, ffecom_decode_include_option_,
3250 ffecom_open_include_, append_include_chain, open_include_file,
3251 print_containing_files, read_filename_string, file_name_map,
3252 savestring): New internal functions for -I.
3254 * compilers.h: Pass -I flag(s) to f771 (via "%{I*}").
3256 * lex.c (ffelex_include_): Call ffecom_close_include
3257 to close include file, for its tracking needs for -I,
3258 instead of using fclose.
3260 * options-lang.h: Add -I prefix.
3262 * parse.c (yyparse): Call ffecom_file for main input file,
3263 so -I handling works (diagnostics).
3265 * std.c (ffestd_S3P4): Have ffecom_open_include handle
3266 opening and diagnosing errors with INCLUDE files.
3268 * ste.c (ffeste_begin_iterdo_): Use correct algorithm for
3269 calculating # of iterations -- mathematically similar but
3270 computationally different algorithm was not handling cases
3271 like "DO I=6,5,2" correctly, because (5-6)/2+1 => 1, not 0.
3273 * top.c (ffe_decode_option): Allow -I, restructure a bit
3274 for clarity and, maybe, speed.
3276 Mon Apr 17 13:31:11 1995 Craig Burley (burley@gnu.ai.mit.edu)
3278 * g77.c: Remove -lc, turns out not all systems has it, but
3279 leave other changes in for clarity of code.
3281 Sun Apr 16 21:50:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
3283 * com.c (ffecom_expr_): Implement ARRAY_EXPR as INDIRECT_REF
3284 of appropriate PLUS_EXPRs of ptr_to_expr of array, to see
3285 if this generates better code. (Conditional on
3286 FFECOM_FASTER_ARRAY_REFS.)
3288 Sun Apr 16 00:22:48 1995 Craig Burley (burley@gnu.ai.mit.edu)
3290 * Make-lang.in (F77_SRCS): Remove g77.c, since it doesn't
3291 contribute to building f771.
3293 * Makefile.in (dircheck): Remove/replace with f/Makefile, because
3294 phony targets that are referenced in other real targets get run
3295 when those targets are specified, which is a waste of time (e.g.
3296 when rebuilding and only g77.c has changed, f771 was being linked
3299 * g77.c: Include -lc between -lf2c and -lm throughout.
3301 * implic.c (ffeimplic_establish_symbol): If -Wimplicit, warn if
3302 implicit type given to symbol.
3304 * lex.c (ffelex_include_): Don't gratuitously increment line
3307 * top.h, top.c (ffe_is_warn_implicit_): New global variable and
3308 related access macros.
3309 (ffe_decode_option): Handle -W options, including -Wall and
3312 * where.c (ffewhere_line_new): Don't muck with root line (was
3313 crashing on null input since lexer changes over the past week
3316 Thu Apr 13 16:48:30 1995 Craig Burley (burley@gnu.ai.mit.edu)
3318 * com.c (ffecom_init_0): Register built-in functions for cos,
3320 (ffecom_tree_fun_type_double): New variable.
3321 (ffecom_expr_intrinsic_): Update f2c input and output files
3322 to latest version of f2c (no important g77-related changes
3323 noted, just bug fixes to f2c and such).
3324 (builtin_function): New function from c-decl.c.
3326 * com-rt.def: Refer to built-in functions for cos, sin, and sqrt.
3328 Thu Apr 13 10:25:09 1995 Craig Burley (burley@gnu.ai.mit.edu)
3330 * com.c (ffecom_expr_intrinsic_): Convert 0. to appropriate
3331 type to keep DCMPLX(I) from crashing the compiler.
3332 (ffecom_expr_): Don't convert result from ffecom_tree_divide_.
3333 (ffecom_tree_divide_): Add tree_type argument, have all callers
3334 pass one, and don't convert right-hand operand to it (this is
3335 to make this new function work as much like the old in-line
3336 code used in ffecom_expr_ as possible).
3338 * lex.c: Maintain lineno and input_filename the way the gcc
3341 * std.c (ffestd_exec_end): Save and restore lineno and
3342 input_filename around the second pass, which sets them
3343 appropriately for each saved statement.
3345 Wed Apr 12 09:44:45 1995 Craig Burley (burley@gnu.ai.mit.edu)
3347 * com.c (ffecom_expr_power_integer_): New function.
3348 (ffecom_expr_): Call new function for power op with integer second
3349 argument, for generating better code. Also replace divide
3350 code with call to new ffecom_tree_divide_ function.
3351 Canonicalize calls to ffecom_truth_value(_invert).
3352 (ffecom_tree_divide_): New function.
3354 Wed Apr 5 14:15:44 1995 Craig Burley (burley@gnu.ai.mit.edu)
3356 * lex.c: Change to allocate text for tokens only when actually
3357 needed, which should speed compilation up somewhat.
3358 Change to allow INCLUDE at any point where a statement
3359 can end, i.e. in ffelex_finish_statement_ or when a SEMICOLON
3361 Remove some old, obsolete code.
3362 Clean up layout of entire file to improve formatting,
3364 (ffelex_set_expecting_hollerith): Remove include argument.
3366 Fri Mar 31 23:19:08 1995 Craig Burley (burley@gnu.ai.mit.edu)
3368 * bad.h, bad.c (ffebad_start_msg, ffebad_start_msg_lex):
3369 New functions to generate arbitrary messages.
3370 (FFEBAD_severityPEDANTIC): New severity, to correspond
3371 to toplev's pedwarn() function.
3373 * lex.c (ffelex_backslash_): New function to implement
3374 backslash processing.
3375 (ffelex_file_fixed, ffelex_file_free): Implement new
3376 backslash processing.
3378 * std.c (ffestd_R1001dump_): Don't assume CHARACTER and
3379 HOLLERITH tokens stop at '\0' characters, now that backslash
3380 processing is supported -- use their advertised lengths instead,
3381 and double up the '\002' character for libf2c.
3383 Mon Mar 27 17:10:33 1995 Craig Burley (burley@gnu.ai.mit.edu)
3385 * com.c (ffecom_init_local_zero_): Implement -finit-local-zero.
3386 (ffecom_sym_transform_): Same.
3387 (ffecom_transform_equiv_): Same.
3389 * options-lang.h: Add -f(no-)(init-local-zero,backslash,ugly-init).
3391 * stb.c (ffestb_V020): Reject "TYPEblah(...", which might be
3392 an array assignment.
3394 * target.h, top.h, top.c: Implement -finit-local-zero.
3396 Fri Mar 24 19:56:22 1995 Craig Burley (burley@gnu.ai.mit.edu)
3398 * Make-lang.in, Makefile.in: Remove conf-proj(.in) and
3399 proj.h(.in) rules, plus related config.log, config.cache,
3400 and config.status stuff.
3402 * com.c (ffecom_init_0): Change messages when atof(), bsearch(),
3403 or strtoul() do not work as expected in the start-up test.
3405 * conf-proj, conf-proj.in: Delete.
3407 * lex.c (ffelex_file_fixed): Allow f2c's '&' in column 1
3408 to mean continuation line.
3410 * options-lang.h: New file, #include'd by ../toplev.c.
3412 * proj.h.in: Rename back to proj.h.
3414 * proj.h (LAME_ASSERT): Remove.
3415 (LAME_STDIO): Remove.
3416 (NO_STDDEF): Remove.
3417 (NO_STDLIB): Remove.
3418 (NO_BSEARCH): Remove auto detection, rename to !FFEPROJ_BSEARCH.
3419 (NO_STRTOUL): Remove auto detection, rename to !FFEPROJ_STRTOUL.
3420 (USE_HOST_LIMITS): Remove (maybe still needed by stand-alone?).
3421 (STR, STRX): Do only ANSI C definitions.
3423 Mon Mar 13 10:46:13 1995 Craig Burley (burley@gnu.ai.mit.edu)
3425 * BUGS: Add item about g77 requiring gcc to compile it.
3427 * NEWS: New file listing user-visible changes in the release.
3429 * PROJECTS: Update to include a new item or two, and modify
3430 or delete items that are addressed in this or previous releases.
3432 * bad.c (ffebad_finish): Don't crash if missing string &c,
3433 just substitute obviously distressed string "[REPORT BUG!!]"
3434 for cases where the message/caller are fudgy.
3436 * bad.def: Clean up error messages in a major way, add new ones
3437 for use by changes in target.c.
3439 * com.c (ffecom_expr_): Handle opANY in opCONVERT.
3440 (ffecom_let_char_): Disregard destinations with ERROR_MARK.
3441 (ffecom_1, ffecom_1_fn, ffecom_2, ffecom_2s, ffecom_3,
3442 ffecom_3s, &c): Check all inputs for error_mark_node.
3443 (ffecom_start_progunit_): Don't transform all symbols
3444 in BLOCK DATA, since it never executes, and it is silly
3445 to, e.g., generate all the structures for NAMELIST.
3446 (ffecom_char_length_expr_): Rename to ffecom_intrinsic_len_.
3447 (ffecom_intrinsic_ichar_): New function to handle ICHAR of
3448 arbitrary expression with possible 0-length operands.
3449 (ffecom_expr_intrinsic_): Use ffecom_intrinsic_char_.
3450 For MVBITS, set tree_type to void_type_node.
3451 (ffecom_start_progunit_): Name master function for entry points
3452 after primary entry point so users can easily guess it while
3454 (ffecom_arg_ptr_to_expr): Change treatment of Hollerith,
3455 Typeless, and %DESCR.
3456 (ffecom_expr_): Change treatment of Hollerith.
3458 * data.c (ffedata_gather_): Handle opANY in opCONVERT.
3460 * expr.c (ffeexpr_token_apostrophe_): Issue FFEBAD_NULL_CHAR_CONST
3461 warning as necessary.
3462 (ffeexpr_token_name_rhs_): Set context for args to intrinsic
3463 so that assignment-like concatenation is allowed for ICHAR(),
3464 IACHAR(), and LEN() intrinsics.
3465 (ffeexpr_reduced_*_): Say "an array" instead of "an entity" in
3466 diagnostics, since it's more informative.
3467 (ffeexpr_finished_): For many contexts, check for null expression
3468 and array before trying to do a conversion, to avoid redundant
3471 * g77.1: Fix typo for preprocessed suffix (.F, not .f).
3473 * global.c (ffeglobal_init_common): Warn if initializing
3475 (ffeglobal_pad_common): Enable code to warn if initial
3477 (ffeglobal_size_common): Complain if enlarging already-
3478 initialized common, since it won't work right anyway.
3480 * intrin.c: Add IMAG() intrinsic.
3481 (ffeintrin_check_loc_): Allow opSUBSTR in LOC().
3483 * intrin.def: Add IMAG() intrinsic.
3485 * lex.c: Don't report FFEBAD_NULL_CHAR_CONST errors.
3487 * sta.c, sta.h, stb.c: Changes to clean up error messages (see
3490 * stb.c (ffestb_R100113_): Issue FFEBAD_NULL_CHAR_CONST
3491 warning as necessary.
3493 * stc.c (ffestc_shriek_do_): Don't try to reference doref_line
3494 stuff in ANY case, since it won't be valid.
3495 (ffestc_R1227): Allow RETURN in main program unit, with
3496 appropriate warnings/errors.
3497 (ffestc_subr_format_): Array of any type is a CHAREXPR (F77 C5).
3499 * ste.c (ffeste_begin_doiter_): Couple of fixes to accurately
3500 determine if loop never executes.
3502 * target.c (ffetarget_convert_*_hollerith_): Append spaces,
3503 not zeros, to follow F77 Appendix C, and to warn when
3504 truncation of non-blanks done.
3505 (ffetarget_convert_*_typeless): Rewrite to do typeless
3506 conversions properly, and warn when truncation done.
3507 (ffetarget_print_binary, ffetarget_print_octal,
3508 ffetarget_print_hex): Rewrite to use new implementation of
3510 (ffetarget_typeless_*): Rewrite to use new implementation
3511 of typeless, and to warn about overflow.
3513 * target.h (ffetargetTypeless): New implementation of
3516 * type.h, type.c (ffetype_size_typeless): Remove (incorrect)
3517 implementation of this function and its extern.
3519 Sun Mar 5 18:46:42 1995 Craig Burley (burley@gnu.ai.mit.edu)
3521 * BUGS: Clarify that constant handling would also fix lack of
3522 adequate IEEE-754/854 support to some degree, and typeless
3523 and non-decimal constants.
3525 * com.c (ffecom_type_permanent_copy_): Comment out to avoid
3527 (duplicate_decls): New function a la gcc/c-decl.c.
3528 (pushdecl): Use duplicate_decls to decide whether to return
3529 existing decl or new one, instead of always returning existing
3531 (ffecom_expr_): opPERCENT_LOC now supports CHARACTER arguments.
3532 (ffecom_init_0): Give f2c I/O code 0 for basictypeANY/kindtypeANY.
3533 (ffecom_sym_transform_): For adjustable arrays, pass low bound
3534 through variable_size in case its primaries are changed (950302-1.f).
3536 * com.h: More decls that belong in tree.h &c.
3538 * data.c (ffedata_eval_integer1_): Fix opPAREN case to not
3539 treat value of expression as an error code.
3541 * expr.c (ffeexpr_finished_): Allow opSUBSTR in contextLOC case.
3543 * proj.c: Add "const" as appropriate.
3545 Mon Feb 27 10:04:03 1995 Craig Burley (burley@gnu.ai.mit.edu)
3547 * bad.def (FFEBAD_BAD_SUBSTR): Fix bad grammar in message.
3549 Fri Feb 24 16:21:31 1995 Craig Burley (burley@gnu.ai.mit.edu)
3551 * Version 0.5.13 released.
3553 * INSTALL: Warn that f/zzz.o will compare differently between
3554 stages, since it puts the __TIME__ macro into a string.
3556 * com.c (ffecom_sym_transform_): Transform kindFUNCTION/whereDUMMY
3557 to pointer-to-function, not function.
3558 (ffecom_expr_): Use ffecom_arg_ptr_to_expr instead of
3559 ffecom_char_args_ to handle comparison between CHARACTER
3560 types, so either operand can be a CONCATENATE.
3561 (ffecom_transform_common_): Set size of initialized common area
3562 to global (largest-known) size, even though size of init might
3565 * equiv.c (ffeequiv_offset_): Check symbol info for ANY.
3567 * expr.c (ffeexpr_find_close_paren_, ffeexpr_nil_*): New functions
3568 to handle following the contour of a rejected expression, so
3569 statements like "PRINT(I,I,I)=0" don't cause the PRINT statement
3570 code to get the second passed back to it as if there was a
3571 missing close-paren before it, the comma causing the PRINT code
3572 to confirm the statement, resulting in an ambiguity vis-a-vis
3573 the let statement code.
3574 Use the new ffecom_find_close_paren_ handler when an expected
3575 close-paren is missing.
3576 (ffeexpr_isdigits_): New function, use in all places that
3577 currently use isdigit in repetitive code.
3578 (ffeexpr_collapse_symter): Collapse to ANY if init-expr is ANY,
3579 so as to avoid having symbol get "transformed" if used to
3581 (ffeexpr_token_real_, ffeexpr_token_number_real_): Don't issue
3582 diagnostic about exponent, since it'll be passed along the
3583 handler path, resulting in a diagnostic anyway.
3584 (ffeexpr_token_apos_char_): Use consistent handler path
3585 regardless of whether diagnostics inhibited.
3586 (ffeexpr_token_name_apos_name_): Skip past closing quote/apos
3587 even if not a match or other diagnostic issued.
3588 (ffeexpr_sym_impdoitem_): Exec-transition local SEEN symbol.
3590 * lex.c (ffelex_image_char_): Set ffelex_saw_tab_ if TAB
3591 seen, not if anything other than TAB seen!
3593 * stc.c (ffestc_R537_item): If source is ANY but dest isn't,
3594 set dest symbol's init expr to ANY.
3595 (ffestc_R501_attrib, ffestc_R522, ffestc_R522start): Complain
3596 about conflict between "SAVE" by itself and other uses of
3597 SAVE only in pedantic mode.
3599 * ste.c (ffeste_R1212): Fix loop over labels to always
3600 increment caseno, to avoid pushcase returning 2 for duplicate
3601 values when one of the labels is invalid.
3603 Thu Feb 23 12:42:04 1995 Craig Burley (burley@gnu.ai.mit.edu)
3605 * Version 0.5.12 released.
3607 * Make-lang.in (f77.install-common): Add "else true;" before outer
3608 "fi" per Makefile.in patch.
3610 * Makefile.in (dircheck): Add "else true;" before "fi" per
3611 patch from chs1pm@surrey.ac.uk.
3613 * com.c (ffecom_push_tempvar): If type desired is ERROR_MARK,
3614 return error_mark_node, to avoid crash that results from
3615 making a VAR_DECL with error_mark_node as its type.
3617 * ste.c (ffeste_begin_iterdo_): Convert itercount to INTEGER
3618 anytime calculation of number of iterations ends up with type
3619 other than INTEGER (e.g. DOUBLE PRECISION, REAL).
3621 Thu Feb 23 02:48:38 1995 Craig Burley (burley@gnu.ai.mit.edu)
3623 * Version 0.5.11 released.
3625 * DOC: Explain -fugly-args.
3627 * bad.def (FFEBAD_ACTUALARG): Explain -fugly-args and how to
3628 rewrite code to not require it.
3630 * com.c (ffecom_vardesc_): Handle negative type code, just in
3632 (ffecom_arg_ptr_to_expr): Let ffecom_expr handle hollerith
3633 and typeless constants (move code to ffecom_constantunion).
3634 (ffecom_constantunion): Handle hollerith and typeless constants.
3636 * expr.c (ffecom_finished_): Check -fugly-args in actual-arg
3637 context where hollerith/typeless provided.
3639 * intrin.def (FFEINTRIN_genDFLOAT): Add FFEINTRIN_specDFLOAT.
3640 (FFEINTRIN_specDFLOAT): Add as f2c intrinsic.
3642 * target.h (ffetarget_convert_real[12]_integer,
3643 ffetarget_convert_complex[12]_integer): Pass -1 for high integer
3644 value if low part is negative.
3645 (FFETARGET_defaultIS_UGLY_ARGS): New macro.
3647 * top.c (ffe_is_ugly_args_): New variable.
3648 (ffe_decode_option): Handle -fugly-args and -fno-ugly-args.
3650 * top.h (ffe_is_ugly_args_, ffe_is_ugly_args(),
3651 ffe_set_is_ugly_args()): New variable and macros.
3653 Thu Feb 23 02:48:38 1995 Pedro A M Vazquez (vazquez@iqm.unicamp.br)
3655 * g77.c (sys_errlist): Use const for __FreeBSD__ systems
3658 Wed Feb 22 13:33:43 1995 Craig Burley (burley@gnu.ai.mit.edu)
3660 * Version 0.5.10 released.
3662 * CREDITS: Add Rick Niles.
3664 * INSTALL: Note how to get around lack of makeinfo.
3666 * Make-lang.in (f/proj.h): Remove # comment.
3668 * Makefile.in (f/proj.h): Remove # comment.
3670 * com.c (ffecom_expr_): Simplify opFUNCREF/opSUBRREF conversion.
3671 (ffecom_sym_transform_): For whereGLOBAL and whereDUMMY
3672 kindFUNCTION, use ffecom_tree_fun_type[][] only for non-constant
3673 (non-statement-function) f2c functions.
3674 (ffecom_init_0): ffecom_tree_fun_type[][] and _ptr_to_*_* are
3675 really f2c-interface arrays, so use base type void for COMPLEX
3678 Tue Feb 21 19:01:18 1995 Dave Love <d.love@dl.ac.uk>
3680 * Make-lang.in (f77.install-common): Expurgate the test for and
3681 possible installation of f2c in line with elsewhere. Seems to have
3682 been missing a semicolon anyhow!
3684 Tue Feb 21 11:45:25 1995 Craig Burley (burley@gnu.ai.mit.edu)
3686 * Version 0.5.9 released.
3688 * Make-lang.in (f/proj.h): touch file to register update,
3689 because the previous commands won't necessarily modify it.
3691 * Makefile.in (f/proj.h): touch file to register update,
3692 because the previous commands won't necessarily modify it.
3694 * Makefile.in (f/str-*.h, f/str-*.j): Explicitly specify
3695 output file names, so these targets go in build, not source,
3698 * bits.c, bits.h: Switch to valid ANSI C replacement for
3701 * com.c (ffecom_expr_): Add assignp arg to support ASSIGN better.
3702 If assignp is TRUE, use different tree for FFEBLD_opSYMTER case.
3703 (ffecom_sym_transform_assign_): New function.
3704 (ffecom_expr_assign): New function.
3705 (ffecom_expr_assign_w): New function.
3707 * com.c (ffecom_f2c_make_type_): Do make_signed_type instead
3708 of make_unsigned_type throughout.
3710 * com.c (ffecom_finish_symbol_transform_): Expand scope of
3711 commented-out code to probably produce faster compiler code.
3713 * com.c (ffecom_gen_sfuncdef_): Push/pop calltemps so
3714 COMPLEX works right.
3715 Remove obsolete comment.
3717 * com.c (ffecom_start_progunit_): If non-multi alt-entry
3718 COMPLEX function, primary (static) entry point returns result
3719 directory, not via extra arg -- to agree with ffecom_return_expr
3721 Pretransform all symbols so statement functions are defined
3722 before any code emitted.
3724 * com.c (ffecom_finish_progunit): Don't posttransform all
3725 symbols here -- pretransform them instead.
3727 * com.c (ffecom_init_0): Don't warn about possible ASSIGN
3728 crash, as this shouldn't happen now.
3730 * com.c (ffecom_push_tempvar): Fix to handle temp vars
3731 pushed while context is a statement (nested) function, and
3732 add appropriate commentary.
3734 * com.c (ffecom_return_expr): Check TREE_USED to determine
3735 where return value is unset.
3737 * com.h (struct _ffecom_symbol_): Add note about length_tree
3738 now being used to keep tree for ASSIGN version of symbol.
3740 * com.h (ffecom_expr_assign, ffecom_expr_assign_rw): New decls.
3741 (error): Add this prototype for back-end function.
3743 * fini.c (main): Grab input, output, and include names
3744 directly off the command line instead of making the latter
3745 two out of the first.
3747 * lex.c: Improve tab handling for both fixed and free source
3748 forms, and ignore carriage-returns on input, while generally
3749 improving the code. ffelex_handle_tab_ has been renamed and
3750 reinvented as ffelex_image_char_, among other things.
3752 * malloc.c, malloc.h: Switch to valid ANSI C replacement for
3753 ARRAY_ZERO, and kill the full number of bytes in pools and
3756 * proj.h.in (ARRAY_ZERO, ARRAY_ZERO_SIZE): Remove.
3758 * ste.c (ffeste_io_cilist_, ffeste_io_icilist_, ffeste_R838,
3759 ffeste_R839): Issue diagnostic if a too-narrow variable used in an
3760 ASSIGN context despite changes to this code and code in com.c.
3762 * where.c, where.h: Switch to valid ANSI C replacement for
3765 Fri Feb 17 03:35:19 1995 Craig Burley (burley@gnu.ai.mit.edu)
3767 * Version 0.5.8 released.
3769 * INSTALL: In quick-build case, list g77 target first so g77
3770 gets installed. Also, explain that gcc gets built and installed
3771 as well, even though this isn't really what we want (and maybe
3772 we'll find a way around this someday).
3774 Fri Feb 17 02:35:41 1995 Craig Burley (burley@gnu.ai.mit.edu)
3776 * Version 0.5.7 released.
3778 * Makefile.in (CONFIG_H, HCONFIG_H, TCONFIG_H, TM_H): Remove
3779 ../ prefix in front of .h files, since they're in the cd.
3781 Fri Feb 17 01:50:48 1995 Craig Burley (burley@gnu.ai.mit.edu)
3783 * Version 0.5.6 released.
3785 Thu Feb 16 20:26:54 1995 Craig Burley (burley@gnu.ai.mit.edu)
3787 * ../README.g77: Remove description of g77 as "not-yet-published".
3789 * CREDITS: More changes.
3791 * Make-lang.in (G77STAGESTUFF): Remove cktyps stuff.
3793 * Makefile.in (CONFIG_H, HCONFIG_H, TCONFIG_H, TM_H): Don't
3794 prefix gcc dir with $(srcdir) since these don't live there,
3795 they are created in the build dir by gcc's configure. Add
3796 a note explaining what these macros are about.
3797 Update dependencies via deps-kinda.
3799 * README.NEXTSTEP: Credit Toon, and per his request, add his
3802 * com.h (FFECOM_DETERMINE_TYPES): #include "config.j".
3804 * config.j, convert.j, flags.j, hconfig.j, rtl.j, tconfig.j,
3805 tm.j, tree.j: Don't #include if already done.
3807 * convert.j: #include "tree.j" first, as convert.h clearly depends
3808 on trees being defined.
3810 * rtl.j: #include "config.j" first, since there's some stuff
3811 in rtl.h that assumes it has been #included.
3813 * tree.j: #include "config.j" first, or real.h makes inconsistent
3814 decision about return type of ereal_atof, leading to bugs, and
3815 because tree.h/real.h assume config.h already included.
3817 Wed Feb 15 14:40:20 1995 Craig Burley (burley@gnu.ai.mit.edu)
3819 * Version 0.5.5 released.
3821 * Copyright notices updated to be FSF-style.
3823 * INSTALL: Some more clarification regarding building just f77.
3825 * Make-lang.in (F77_SRCS): Update wrt changing some .h to .j.
3826 (install-libf77): Fix typo in new parenthetical note.
3828 * Makefile.in (f/*.o): Update.
3829 (CONFIG_H, CONVERT_H, FLAGS_H, GLIMITS_H, HCONFIG_H, RTL_H,
3830 TCONFIG_H, TM_H, TREE_H): Update/new symbols.
3831 (deps-kinda): More fixes wrt changing some .h to .j.
3832 Document and explain this rule a bit better.
3833 Accommodate changes in output of gcc -MM.
3835 * *.h, *.c: Change #include's so proj.h not assumed to #include
3836 malloc.h or config.h (now config.j), and so new .j files are
3837 used instead of old .h ones.
3839 * com.c (ffecom_init_0): Use FLOAT_TYPE_SIZE for f2c's
3840 TYLONG/TYLOGICAL type codes, to get g77 working on Alpha.
3842 * com.h: Make all f2c-related integral types "int", not "long
3845 * config.j, convert.j, flags.j, glimits.j, hconfig.j, rtl.j,
3846 tconfig.j, tm.j, tree.j: New files wrapping around gbe
3849 * config.h, convert.h, flags.h, glimits.h, hconfig.h, rtl.h,
3850 tconfig.h, tm.h, tree.h: Deleted so new .j files
3851 can #include the gbe files directly, instead of using "../",
3852 and thus do better with various kinds of builds.
3854 * proj.h: Delete unused NO_STDDEF and related stuff.
3856 Tue Feb 14 08:28:08 1995 Craig Burley (burley@gnu.ai.mit.edu)
3858 * BUGS: Remove item #12, cross-compiling & autoconf scripts
3859 reportedly expected to work properly (according to d.love).
3861 * INSTALL: Add explanation of d.love's patch to config-lang.in.
3862 Add explanation of how to install just g77 when gcc already installed.
3863 Add note about usability of "-Wall". Add note about bug-
3866 * Make-lang.in ($(srcdir)/f/conf-proj): Add comment about why
3868 (install-libf77): Echo parenthetical note to user about how to do
3869 just the (aborted) libf2c installation.
3870 (deps-kinda): Update to work with new configuration/build stuff.
3872 * bad.c (ffebad_finish): Put capitalized "warning:" &c message
3873 as prefix on any diagnostic without pointers into source.
3875 * bad.def (FFEBAD_TOO_BIG_INIT): Add this warning message.
3877 * config-lang.in: Add Dave Love's patch to catch case where
3878 back-end patches not applied and abort configuration.
3880 * data.c (ffedata_gather_, ffedata_value_): Warn when about
3881 to initialize a large aggregate area, due to design flaw resulting
3882 in too much time/space used to handle such cases.
3883 Use COMMON area name, and first notice of symbol, for multiple-
3884 initialization diagnostic, instead of member symbol and unknown
3886 (FFEDATA_sizeTOO_BIG_INIT_): New macro per above.
3888 Mon Feb 13 13:54:26 1995 Dave Love <d.love@dl.ac.uk>
3890 * Make-lang.in (F77_SRCS): Use $(srcdir)/f/proj.h.in, not
3891 $(srcdir)/f/proj.h for build outside srcdir.
3893 Sun Feb 12 13:37:11 1995 Craig Burley (burley@gnu.ai.mit.edu)
3895 * ../README.g77: Clarify procedures for unpacking, add asterisks
3896 to mark important things the user must do.
3898 * Fix dates in/add dates to ../README.g77, BUGS, CREDITS, DOC,
3899 INSTALL, PROJECTS, README.
3901 Sun Feb 12 00:26:10 1995 Craig Burley (burley@gnu.ai.mit.edu)
3903 * Version 0.5.4 released.
3905 * Make-lang.in (f/proj.h): Reproduce this rule here from
3907 ($(srcdir)/f/conf-proj): Put autoconf's stdout in temp file
3908 conf-proj.out, then mv to conf-proj only if successful, so
3909 conf-proj not touched if autoconf not installed.
3911 * Makefile.in ($(srcdir)/conf-proj): See Make-lang.in's similar
3914 Sat Feb 11 20:56:02 1995 Craig Burley (burley@gnu.ai.mit.edu)
3916 * BUGS: Clarify some bugs.
3918 * DOC: Many improvements and fixes.
3920 * README: Move bulk of text, edited, to ../README.g77, and
3921 replace with pointer to that file.
3923 * com.c (ffecom_init_0): Comment out warning about sizeof(ftnlen)
3924 as per ste.c change. Add text about ASSIGN to help user understand
3925 what is being warned about.
3927 * conf-proj.in: Fix typos in comments.
3929 * proj.h.in: Add ARRAY_ZERO_SIZE to parallel malloc.h's version,
3930 in case it proves to be needed.
3932 * ste.c: Comment out assertions requiring sizeof(ftnlen) >=
3933 sizeof(char *), in the hopes that overflow will never happen.
3934 (ffeste_R838): Change assertion to fatal() with at least
3935 partially helpful message.
3937 Sat Feb 11 12:38:00 1995 Craig Burley (burley@gnu.ai.mit.edu)
3939 * com.c (ffecom_vardesc_): Crash if typecode is -1.
3941 * ste.c (ffeste_io_dolio_): Crash if typecode is -1.
3943 Sat Feb 11 09:51:57 1995 Craig Burley (burley@gnu.ai.mit.edu)
3945 * ste.c: In I/O code tests for item arrayness, sort of revert
3946 to much earlier code that tests original exp, but also check
3947 in newer way just in case. Newer way alone treated FOO(1:40)
3948 as an array, not sure why older way alone didn't work, but I
3949 think maybe it was when diagnosed code was involved, and
3950 since there are now checks for error_mark_node, maybe the old
3951 way alone would work. But better to be safe; both original
3952 ffebld exp _and_ the transformed tree must indicate an array
3953 for the size-determination code to be used, else just 1/2 elements
3954 assumed. And this text is for EMACS: (foo at bar).
3956 Fri Feb 10 11:05:50 1995 Craig Burley (burley@gnu.ai.mit.edu)
3958 * ste.c: In many cases, surround statement-expansion code
3959 with ffecom_push_calltemps () and ffecom_pop_calltemps ()
3960 so COMPLEX-returning functions can have temporaries pushed
3961 in "auto-pop" mode and have them auto-popped at the end of
3964 Wed Feb 8 14:35:10 1995 Dave Love <d.love@dl.ac.uk>
3966 * runtime/f2c.h.in (ftnlen, ftnint): Make same size as integer.
3968 * runtime/libI77/err.c (f_init): Thinko in MISSING_FILE_ELEMS
3970 * runtime/libI77/wrtfmt.c (mv_cur): Likewise.
3971 * runtime/libI77/wsfe.c (x_putc): Likewise.
3973 * runtime/libF77/signal_.c (signal_): Return 0 (this is a
3976 * Makefile.in (f/proj.h): Depend on com.h.
3977 * Make-lang.in (include/f2c.h): Likewise (and proj.h).
3978 (install-libf77): Also install f2c.h.
3980 * runtime/libI77/Makefile.in (*.o): Add f2c.h dependency.
3981 * runtime/libF77/Makefile.in: Likewise.
3983 Wed Feb 8 13:56:47 1995 Craig Burley (burley@gnu.ai.mit.edu)
3985 * stc.c (ffestc_R501_item): Don't reset kind/where to NONE when
3986 setting basictype/kindtype info for symbol, or especially
3987 its function/result twin, because kind/where might not be NONE.
3989 Tue Feb 7 14:47:26 1995 Dave Love <d.love@dl.ac.uk>
3991 * Make-lang.in (include/f2c.h:): Set shell variable src more
3992 robustly (independent of whether srcdir is relative or absolute).
3993 * Makefile.in (f/proj.h:): Likewise.
3995 * conf-proj.in: Check need for LAME_ASSERT. Fix indentation in
3996 check for LAME_STDIO (cosmetic only with ANSI C).
3998 * com.h: Extra ...SIZE stuff taken from com.c.
4000 * com.c (FFECOM_DETERMINE_TYPES): Define before including com.h.
4001 (BITS_PER_WORD etc.) Remove and use conditional definitions to com.h.
4003 * runtime/configure.in: #define FFECOM_DETERMINE_TYPES for com.h in
4004 f2c type determination.
4006 * tm.h: Remove (at least pro tem) because of relative path and use
4009 * Make-lang.in (include/f2c.h:): Set shell variable src more
4010 robustly (independent of whether srcdir is relative or absolute).
4011 * Makefile.in (f/proj.h:): Likewise.
4013 Mon Feb 6 19:58:32 1995 Dave Love <d.love@dl.ac.uk>
4015 * g77.c (append_arg): Use K&R declaration for, e.g. SunOS4 build.
4017 Fri Feb 3 20:33:14 1995 Craig Burley (burley@gnu.ai.mit.edu)
4019 * g77.c (main): Treat -l like filename in terms of -x handling.
4020 Rewrite arglist mechanism for ease of maintenance.
4021 Make sure every -lf2c is followed by -lm and vice versa.
4023 * Make-lang.in: Put complete list of sources in F77_SRCS def
4024 so changing a .h file, for example, causes rebuild.
4026 * Makefile.in: Change test for nextstep to m68k-next-nextstep* so
4027 all versions of nextstep on m68k get the necessary flag.
4029 Fri Feb 3 19:10:32 1995 Dave Love <d.love@dl.ac.uk>
4031 * INSTALL: Note about possible conflict with existing libf2c.a and
4034 * Make-lang.in (f77.distclean): Tidy and move deletion of
4035 f/config.cache to mostlyclean.
4036 (install-libf77): Test for $(libdir)/libf2c.* and barf if found
4037 unless F2CLIBOK defined.
4039 * runtime/Makefile.in (all): Change path to include directory (and
4041 (INCLUDES): Remove (unused/misleading).
4042 (distclean): Include f2c.h.
4043 (clean): Include config.cache.
4045 * runtime/libF77/Makefile.in (.SUFFIXES): Correct typo.
4046 (ALL_CFLAGS) Fix up include search path to find f2c.h in top level
4048 (all): Depend on f2c.h.
4049 * runtime/libI77/Makefile.in (.SUFFIXES): Likewise.
4051 Thu Feb 2 17:17:06 1995 Dave Love <d.love@dl.ac.uk>
4053 * INSTALL: Note about --srcdir and GNU make.
4055 * runtime/f2c.h.in (Pad_UDread, ALWAYS_FLUSH): Reomve the #defines
4058 * runtime/configure.in (Pad_UDread, ALWAYS_FLUSH): Define these
4059 here, not in f2c.h as they'r eonly relevant for building.
4060 * runtime/configure: Regenerated.
4062 * config-lang.in: Warn about using GNU make outside source tree
4063 since I can't get Irix5 or SunOS4 makes to work in this case.
4065 * Makefile.in (VPATH): Don't set it here.
4066 (srcdir): Make it the normal `.' (overridden) at top level.
4067 (all.indirect): New dependency `dircheck'.
4069 (dircheck): New target for foolproofing.
4070 (f/proj.h:): Change finding source.
4071 (CONFIG_H): Don't use this as the relative path in the include loses
4072 f builddir != srcdir.
4074 * config.h: Remove per CONFIG_H change above.
4076 * Make-lang.in (F77_FLAGS_TO_PASS): Remove GCC_FOR_TARGET.
4077 (f771:): Pass VPATH, srcdir to sub-make.
4078 (f/Makefile:): New target.
4079 (stmp-int-hdrs): new variable for cheating build.
4080 (f77-runtime:): Alter GCC_FOR_TARGET treatment.
4081 (include/f2c.h f/runtime/Makefile:) Likewise.
4082 (f77-runtime-unsafe:): New (cheating) target.
4084 Thu Feb 2 12:09:51 1995 Craig Burley (burley@gnu.ai.mit.edu)
4086 * BUGS: Update regarding losing EQUIVALENCE members in -g, and
4087 regarding RS/6000 problems in the back end.
4089 * CREDITS: Make some changes as requested.
4091 * com.c (ffecom_member_trunk_): Remove unused static variable.
4092 (ffecom_finish_symbol_transform_): Improve comments.
4093 (ffecom_let_char_): Fix size of temp address-type var.
4094 (ffecom_member_phase2_): Try fixing problem fixed by change
4095 to ffecom_transform_equiv_ (f_m_p2_ function currently not used).
4096 (ffecom_transform_equiv_): Remove def of unused static variable.
4097 Comment-out use of ffecom_member_phase2_, until problems with
4099 (ffecom_push_tempvar): Fix assertion to not crash okay code.
4101 * com.h: Remove old, commented-out code.
4102 Add prototype for warning() in back end.
4104 * ste.c (ffeste_io_dofio_, ffeste_io_dolio_, ffeste_io_douio_,
4105 ffeste_io_icilist_): Check correct type of variable for arrayness.
4107 Sun Jan 29 14:41:42 1995 Dave Love <d.love@dl.ac.uk>
4109 * BUGS: Remove references to my configure bugs; add another.
4111 * runtime/Makefile.in (AR_FLAGS): Provide default value.
4113 * runtime/f2c.h.in (integer, logical): Take typedefs from
4114 F2C_INTEGER configuration parameter again.
4115 (NON_UNIX_STDIO): don't define it.
4117 * runtime/configure.in: Bring type checks for f2c.h in line with
4119 (MISSING_FILE_ELEMS): New variable to determine whether the relevant
4120 elements of the FILE struct exist, independent of NON_UNIX_STDIO.
4121 * runtime/libI77/{err,wrtfmt,wsfe}.c (MISSING_FILE_ELEMS): Use new
4124 * config-lang.in: Comment out more of f2c rules (c.f. Make-lang.in).
4125 (This stuff is relevant iff you gave configure --enable-f2c.)
4126 Create f/runtime directory tree iff not building in source
4129 * Makefile.in (srcdir): Append slash so we get the right value when
4130 not building in the source directory. This is a consequence of not
4131 building the `f' sources in `f'.
4132 (VPATH): Override configure's value for reasons above.
4133 (f/proj.h f/conf-proj): New rules to build proj.h by
4136 * proj.h: Rename to proj.h.in for autoconfiguration.
4137 * proj.h.in: New as above.
4138 * conf-proj conf-proj.in: New files for autoconfiguration.
4140 * Make-lang.in (include/f2c.h f/runtime/Makefile:): Change the order
4141 of setting the sh variables so that the right GCC_FOR_TARGET is
4143 (f77.*clean:) Add products of new configuration files and make sure
4144 all the *clean targets do something (unlike the ones in
4147 * com.h (FFECOM_f2cINTEGER, FFECOM_f2cLOGICAL): Define as long or
4148 int appropriately to ensure sizeof(real) == sizeof(integer).
4150 * PROJECTS: Library section.
4152 * runtime/libI77/endfile.c: Don't #include sys/types.h conditional
4153 on NON_UNIX_STDIO since rawio.h needs size_t.
4154 * runtime/libI77/uio.c: #include <sys/types.h> for size_t if not
4157 Wed Jan 25 03:31:51 1995 Craig Burley (burley@gnu.ai.mit.edu)
4159 * Version 0.5.3 released.
4163 * Make-lang.in: Comment out rules for building f2c itself (f/f2c/).
4167 * com.c (ffecom_init_0): Warn if ftnlen or INTEGER not big enough
4170 * gbe/2.6.2.diff: Update.
4172 Mon Jan 23 17:10:49 1995 Craig Burley (burley@gnu.ai.mit.edu)
4180 Make-lang.in: Remove cktyps stuff.
4181 Makefile.in: Remove cktyps stuff.
4183 * DOC: Add info on changes for 0.5.3.
4185 * bad.c: Put "warning:" &c on diagnostic messages.
4186 Don't output informational messages if warnings disabled.
4188 Thu Jan 19 12:38:13 1995 Craig Burley (burley@gnu.ai.mit.edu)
4190 * g77.c: Avoid putting out useless "-xnone -xf77" pairs so
4191 larger command lines can be accommodated.
4192 Recognize both `-xlang' and `-x lang'.
4193 Recognize `-xnone' and `-x none' to mean what it does, instead
4194 of treating "none" as any other language.
4195 Some minor, slight improvements in the way args are handled
4196 (hopefully for clearer, more maintainable code), including
4197 consistency checks on arg count just in case.
4199 Wed Jan 18 16:41:57 1995 Craig Burley (burley@gnu.ai.mit.edu)
4201 * DOC: Explain -fautomatic better.
4203 * INSTALL: Describe libf2c.a better.
4205 * Make-lang.in, Makefile.in: Build f771 &c with gcc/ as cd instead
4206 of gcc/f/ so debugging info is better (source file tracking).
4207 Add new source file type.c.
4209 * Makefile.in: For nextstep3, link f771 with -segaddr __DATA
4210 6000000. Fix typo. Change deps-kinda target to handle building
4211 from gcc/. Update dependencies.
4213 * bld-op.def, bld.h, bld.c: Remove opBACKEND and all related
4215 Remove consistency tests that cause compiler warnings.
4217 * cktyps.c: Remove all typing checking.
4219 * com-rt.def: Change all rttypeFLOAT_ intrinsics to rttypeDOUBLE_,
4220 to precisely match how they're declared in libf2c.
4222 * com.h, com.c: Revise to more elegantly track related stuff
4223 in the version of f2c.h used to build libf2c.
4225 * com.c: Increase FFECOM_sizeMAXSTACKITEM, and if 0 or undefined
4226 when checked to determine where to put entity, treat as infinite.
4227 Rewrite temporary mechanism to be based on trees instead of
4228 ffeinfo stuff, and make it much simpler. Change interface
4230 Fixes to better track types of things, make appropriate
4231 conversions, etc. E.g. when making an arg for a libf2c
4232 function, make sure it's of the right type (such as ftnlen).
4233 Delete opBACKEND transformation code.
4234 (ffecom_init_0): Smoother initialization of types, especially
4235 paying attention to using consistent rules for making INTEGER,
4236 REAL, DOUBLE PRECISION, etc., and for deciding their "*N"
4237 and kind values that will work across all g77 platforms.
4238 No longer require per-target configuration info in target.h
4239 or config/*/*; use new type module to store size, alignment.
4240 (ffecom_member_phase2): Declare COMMON/EQUIVALENCE group members
4241 so debugger sees them.
4242 (ffecom_finish_progunit): Transform all symbols in program unit,
4243 so -g will show they all exist.
4245 * expr.c (ffeexpr_collapse_substr): Handle strange substring
4248 * info.h, info.c: Provide connection to new type module.
4249 Remove tests that yield compiler warnings.
4251 * intrin.c (ffeintrin_is_intrinsic): Properly handle deleted
4254 * lex.c (ffelex_file_fixed): Remove redundant/buggy code.
4256 * stc.c (ffestc_kindtype_kind_, ffestc_kindtype_star_): Replace
4257 boring switch stmt with simple call to new type module. This
4258 sort of thing is a reason to get up in the morning.
4260 * ste.c: Update to handle new interface for
4261 ffecom_push/pop_tempvar.
4262 Fixes to better track types of things.
4263 Fixes to not crash for certain diagnosed constructs.
4264 (ffeste_begin_iterdo_): Check only constants for overflow to avoid
4265 spurious diagnostics.
4266 Don't convert larger integer (say, INTEGER*8) to canonical integer
4267 for iteration count.
4269 * stw.h: Track DO iteration count temporary variable.
4271 * symbol.c: Remove consistency tests that cause compiler warnings.
4273 * target.c (ffetarget_aggregate_info): Replace big switch with
4274 little call to new type module.
4275 (ffetarget_layout): Remove consistency tests that cause
4277 (ffetarget_convert_character1_typeless): Pick up length of
4278 typeless type from new type module.
4280 * target.h: Crash build if target float bit pattern cannot be
4281 precisely determined.
4282 Remove all the type cruft now determined by ffecom_init_0
4283 at invocation time and maintained in new type module.
4284 Put casts on second arg of all REAL_VALUE_TO_TARGET_DOUBLE
4285 uses so compiler warnings avoided (requires target float bit
4286 pattern to be precisely determined, hence code to crash build).
4288 * top.c: Add inits/terminates for new type module.
4290 * type.h, type.c: New module.
4292 * gbe/2.6.2.diff: Remove all patches to files in gcc/config/
4293 directory and its subdirectories.
4295 Mon Jan 9 19:23:25 1995 Dave Love <d.love@dl.ac.uk>
4297 * com.h (FFECOM_F2C_INTEGER_TYPE_NODE): Define and use instead of
4298 long_integer_type_node where appropriate.
4300 Tue Jan 3 14:56:18 1995 Dave Love <d.love@dl.ac.uk>
4302 * com.h: Make ffecom_f2c_logical_type_node long, consistent with
4305 Fri Dec 2 20:07:37 1994 Dave Love <d.love@dl.ac.uk>
4307 * config-lang.in (stagestuff): Add f2c conditionally.
4308 * Make-lang.in: Add f2c and related targets.
4309 * f2c: Add the directory.
4311 Fri Nov 25 22:17:26 1994 Dave Love <d.love@dl.ac.uk>
4313 * Makefile.in (FLAGS_TO_PASS): pass $(CROSS)
4314 * Make-lang.in: more changes to runtime targets
4316 Thu Nov 24 18:03:21 1994 Dave Love <d.love@dl.ac.uk>
4318 * Makefile.in (FLAGS_TO_PASS): define for sub-makes
4320 * g77.c (main): change f77-cpp-output to f77-cpp-input (.F files)
4322 Wed Nov 23 15:22:53 1994 Dave Love <d.love@dl.ac.uk>
4324 * bad.c (ffebad_finish): kluge to fool emacs19 into finding errors:
4325 add trailing space to <file>:<line>:
4327 Tue Nov 22 11:30:50 1994 Dave Love <d.love@dl.ac.uk>
4329 * runtime/libF77/signal_.c (RETSIGTYPE): added
4331 Mon Nov 21 13:04:13 1994 Dave Love <d.love@dl.ac.uk>
4333 * Makefile.in (compiler): add runtime
4335 * config-lang.in (stagestuff): add libf2c.a to stagestuff
4338 G77STAGESTUFF <- MORESTAGESTUFF
4339 f77-runtime: new target, plus supporting ones
4341 * runtime: add the directory, containing libI77, libF77 and autoconf
4346 * g77.1: minor fixes
4348 Thu Nov 17 15:18:05 1994 Craig Burley (burley@gnu.ai.mit.edu)
4350 * Version 0.5.2 released.
4352 * bad.def: Modify wording of FFEBAD_UNIMPL_STMT to indicate
4353 that it covers a wide array of possible problems (that, someday,
4354 should be handled via separate diagnostics).
4356 * lex.c: Allow $ in identifiers if -fdollar-ok.
4357 * top.c: Support -fdollar-ok.
4358 * top.h: Support -fdollar-ok.
4359 * target.h: Support -fdollar-ok.
4360 * DOC: Describe -fdollar-ok.
4362 * std.c (ffestd_R1229_finish): Fix bug so stand-alone build works.
4363 * ste.c (ffeste_R819A): Fix bug so stand-alone build works.
4365 * Make: Improvements for stand-alone build.
4367 * Makefile.in: Fix copyright text at top of file.
4369 * LINK, SRCS, UNLINK: Removed. Not particularly useful now that
4370 g77 sources live in their own subdirectory.
4372 * g77.c (main): Cast arg to bzero to avoid warning. (This is
4373 identical to Kenner's fix to cp/g++.c.)
4375 * gbe/: New subdirectory, to contain .diff files for various
4376 versions of the GNU CC back end.
4378 * gbe/README: New file.
4379 * gbe/2.6.2.diff: New file.
4381 Tue Nov 8 10:23:10 1994 Dave Love <d.love@dl.ac.uk>
4383 * Make-lang.in: don't install as f77 as well as g77 to avoid
4384 confusion with system's compiler (especially while testing)
4386 * g77.c (main): use -lf2c and -lm; fix sense of test for .f/.F files
4388 Fri Oct 28 09:45:00 1994 Craig Burley (burley@gnu.ai.mit.edu)
4390 * Version 0.5.1 released.
4392 * gcc.c: Invoke f771 instead of f-771.
4394 Fri Oct 28 02:00:44 1994 Craig Burley (burley@gnu.ai.mit.edu)
4396 * Version 0.5.0 released.
4398 Fri Oct 14 15:03:35 1994 Craig Burley (burley@gnu.ai.mit.edu)
4400 * Makefile.in: Handle the Fortran-77 front-end in a subdirectory.
4401 * f-*: Move Fortran-77 front-end to f/*.