]> git.saurik.com Git - bison.git/blame_incremental - ChangeLog
* src/state.h, src/state.c (state_rule_lookaheads_print): New.
[bison.git] / ChangeLog
... / ...
CommitLineData
12002-06-15 Akim Demaille <akim@epita.fr>
2
3 * src/state.h, src/state.c (state_rule_lookaheads_print): New.
4 * src/print.c (print_core): Use it.
5
6
72002-06-15 Akim Demaille <akim@epita.fr>
8
9 * src/conflicts.c (log_resolution): Accept the rule involved in
10 the sr conflicts instead of the lookahead number that points to
11 that rule.
12 (flush_reduce): Accept the current lookahead vector as argument,
13 instead of the index in LA.
14 (resolve_sr_conflict): Accept the current number of lookahead
15 bitset to consider for the STATE, instead of the index in LA.
16 (set_conflicts): Adjust.
17 * src/lalr.c, src/lalr.h, src/state.h: Comment changes.
18
19
202002-06-15 Akim Demaille <akim@epita.fr>
21
22 * src/state.h (state_t): Replace the `lookaheadsp' member, a
23 short, with `lookaheads' (bitsetv), `lookaheads_rule' (rule_t**).
24 Adjust all dependencies.
25 * src/lalr.c (initialize_lookaheads): Split into...
26 (states_lookaheads_count, states_lookaheads_initialize): these.
27 (lalr): Adjust.
28
29
302002-06-15 Akim Demaille <akim@epita.fr>
31
32 * src/gram.h, src/gram.c (grammar_rules_partial_print): New, eved
33 out of...
34 (grammar_rules_print): here.
35 * src/reduce.c (reduce_output): Use it.
36 * tests/reduce.at (Useless Rules, Reduced Automaton)
37 (Underivable Rules): Adjust.
38
39
402002-06-15 Akim Demaille <akim@epita.fr>
41
42 Copy BYacc's nice way to report the grammar.
43
44 * src/gram.h, src/gram.c (grammar_rhs_print, grammar_rules_print):
45 New.
46 Don't print the rules' location, it is confusing and useless.
47 (rule_print): Use grammar_rhs_print.
48 * src/print.c (print_grammar): Use grammar_rules_print.
49
50
512002-06-15 Akim Demaille <akim@epita.fr>
52
53 Complete and rationalize `useless thing' warnings.
54
55 * src/symtab.h, src/symtab.c (symbol_tag_get, symbol_tag_get_n)
56 (symbol_tag_print): New.
57 Use them everywhere in place of accessing directly the tag member.
58 * src/gram.h, src/gram.c (rule_print): New.
59 Use it where a rule used to be printed `by hand'.
60 * src/reduce.c (nonterminals_reduce): Report the use nonterminals.
61 (reduce_grammar_tables): Report the useless rules.
62 (reduce_print): Useless things are a warning, not an error.
63 Report it as such.
64 * tests/reduce.at (Useless Nonterminals, Useless Rules):
65 (Reduced Automaton, Underivable Rules): Adjust.
66 * tests/regression.at (Web2c Report, Web2c Report): Adjust.
67 * tests/conflicts.at (Unresolved SR Conflicts)
68 (Solved SR Conflicts): Adjust.
69
70
712002-06-15 Akim Demaille <akim@epita.fr>
72
73 Let symbols have a location.
74
75 * src/symtab.h, src/symtab.c (symbol_t): Location is a new member.
76 (getsym): Adjust.
77 Adjust all callers.
78 * src/complain.h, src/complain.c (complain_at, fatal_at, warn_at):
79 Use location_t, not int.
80 * src/symtab.c (symbol_check_defined): Take advantage of the
81 location.
82 * tests/regression.at (Invalid inputs): Adjust.
83
842002-06-15 Akim Demaille <akim@epita.fr>
85
86 * src/parse-gram.y (YYLLOC_DEFAULT, current_lhs_location): New.
87 (input): Don't try to initialize yylloc here, do it in the
88 scanner.
89 * src/scan-gram.l (YY_USER_INIT): Initialize yylloc.
90 * src/gram.h (rule_t): Change line and action_line into location
91 and action_location, of location_t type.
92 Adjust all dependencies.
93 * src/location.h, src/location.c (empty_location): New.
94 * src/reader.h, src/reader.c (grammar_start_symbol_set)
95 (grammar_symbol_append, grammar_rule_begin, grammar_rule_end)
96 (grammar_current_rule_symbol_append)
97 (grammar_current_rule_action_append): Expect a location as argument.
98 * src/reader.c (grammar_midrule_action): Adjust to attach an
99 action's location as dummy symbol location.
100 * src/symtab.h, src/symtab.c (startsymbol_location): New.
101 * tests/regression.at (Web2c Report, Rule Line Numbers): Adjust
102 the line numbers.
103
104
1052002-06-14 Akim Demaille <akim@epita.fr>
106
107 Grammar declarations may be found in the grammar section.
108
109 * src/parse-gram.y (rules_or_grammar_declaration): New.
110 (declarations): Each declaration may end with a semicolon, not
111 just...
112 (grammar_declaration): `"%union"'.
113 (grammar): Branch to rules_or_grammar_declaration.
114
1152002-06-14 Akim Demaille <akim@epita.fr>
116
117 * src/main.c (main): Invoke scanner_free.
118
1192002-06-14 Akim Demaille <akim@epita.fr>
120
121 * src/output.c (m4_invoke): Extracted from...
122 (output_skeleton): here.
123 Free tempfile.
124
1252002-06-14 Akim Demaille <akim@epita.fr>
126
127 * src/parse-gram.y (directives, directive, gram)
128 (grammar_directives, precedence_directives, precedence_directive):
129 Rename as...
130 (declarations, declaration, grammar, grammar_declaration)
131 (precedence_declaration, precedence_declarator): these.
132 (symbol_declaration): New.
133
1342002-06-14 Akim Demaille <akim@epita.fr>
135
136 * src/files.c (action_obstack): Remove, unused.
137 (output_obstack): Remove it, and all its dependencies, as it is no
138 longer needed.
139 * src/reader.c (epilogue_set): Build the epilogue in the
140 muscle_obstack.
141 * src/output.h, src/output.c (muscle_obstack): Move to...
142 * src/muscle_tab.h, src/muscle_tab.h: here.
143 (muscle_init): Initialize muscle_obstack.
144 (muscle_free): New.
145 * src/main.c (main): Call it.
146
1472002-06-14 Akim Demaille <akim@epita.fr>
148
149 * src/location.h: New, extracted from...
150 * src/reader.h: here.
151 * src/Makefile.am (noinst_HEADERS): Merge into
152 (bison_SOURCES): this.
153 Add location.h.
154 * src/parse-gram.y: Use location_t instead of Bison's.
155 * src/reader.h, src/reader.c (prologue_augment, epilogue_set):
156 Use location_t instead of ints.
157
1582002-06-14 Akim Demaille <akim@epita.fr>
159
160 * data/bison.simple, data/bison.c++: Be sure to restore the
161 current #line when returning to the skeleton contents after having
162 exposed the input file's #line.
163
1642002-06-12 Akim Demaille <akim@epita.fr>
165
166 * src/scan-gram.l (SC_BRACED_CODE): Don't use `<.*>', it is too
167 eager.
168 * tests/actions.at (Exotic Dollars): New.
169
1702002-06-12 Akim Demaille <akim@epita.fr>
171
172 * src/scan-gram.l (SC_PROLOGUE): Don't eat characters amongst
173 ['"/] too eagerly.
174 * tests/input.at (Torturing the Scanner): New.
175
1762002-06-11 Akim Demaille <akim@epita.fr>
177
178 * src/scan-gram.l (YY_OBS_INIT): Remove, replace with...
179 [SC_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE]
180 [SC_EPILOGUE]: Output the quadrigraphs only when not in a comment.
181 * src/reader.h, src/scan-gram.l (scanner_initialize): this.
182 * src/reader.c (reader): Use it.
183
1842002-06-11 Akim Demaille <akim@epita.fr>
185
186 * src/scan-gram.l (YY_OBS_FINISH): Don't set yylval.
187 Adjust all callers.
188 (scanner_last_string_free): New.
189
1902002-06-11 Akim Demaille <akim@epita.fr>
191
192 * src/scan-gram.l (YY_INIT, YY_GROW, YY_FINISH): Rename as...
193 (YY_OBS_INIT, YY_OBS_GROW, YY_OBS_FINISH): these.
194 (last_string, YY_OBS_FREE): New.
195 Use them when returning an ID.
196
1972002-06-11 Akim Demaille <akim@epita.fr>
198
199 Have Bison grammars parsed by a Bison grammar.
200
201 * src/reader.c, src/reader.h (prologue_augment): New.
202 * src/reader.c (copy_definition): Remove.
203
204 * src/reader.h, src/reader.c (gram_start_symbol_set, prologue_augment)
205 (grammar_symbol_append, grammar_rule_begin, grammar_midrule_action)
206 (grammar_current_rule_prec_set, grammar_current_rule_check)
207 (grammar_current_rule_symbol_append)
208 (grammar_current_rule_action_append): Export.
209 * src/parse-gram.y (symbol_list_new, symbol_list_symbol_append_
210 (symbol_list_action_append): Remove.
211 Hook the routines from reader.
212 * src/scan-gram.l: In INITIAL, characters and strings are tokens.
213 * src/system.h (ATTRIBUTE_NORETURN, ATTRIBUTE_UNUSED): Now.
214
215 * src/reader.c (read_declarations): Remove, unused.
216
217 * src/parse-gram.y: Handle the epilogue.
218 * src/reader.h, src/reader.c (gram_start_symbol_set): Rename as...
219 (grammar_start_symbol_set): this.
220 * src/scan-gram.l: Be sure to ``use'' yycontrol to keep GCC quiet.
221 * src/reader.c (readgram): Remove, unused.
222 (reader): Adjust to insert eoftoken and axiom where appropriate.
223
224 * src/reader.c (copy_dollar): Replace with...
225 * src/scan-gram.h (handle_dollar): this.
226 * src/parse-gram.y: Remove `%thong'.
227
228 * src/reader.c (copy_at): Replace with...
229 * src/scan-gram.h (handle_at): this.
230
231 * src/complain.h, src/complain.c (warn_at, complain_at, fatal_at):
232 New.
233
234 * src/scan-gram.l (YY_LINES): Keep lineno synchronized for the
235 time being.
236
237 * src/reader.h, src/reader.c (grammar_rule_end): New.
238
239 * src/parse.y (current_type, current_class): New.
240 Implement `%nterm', `%token' support.
241 Merge `%term' into `%token'.
242 (string_as_id): New.
243 * src/symtab.h, src/symtab.c (symbol_make_alias): Don't pass the
244 type name.
245
246 * src/parse-gram.y: Be sure to handle properly the beginning of
247 rules.
248
249 * src/parse-gram.y: Handle %type.
250 * src/reader.c (grammar_rule_end): Call grammar_current_rule_check.
251
252 * src/parse-gram.y: More directives support.
253 * src/options.c: No longer handle source directives.
254
255 * src/parse-gram.y: Fix %output.
256
257 * src/parse-gram.y: Handle %union.
258 Use the prologue locations.
259 * src/reader.c (parse_union_decl): Remove.
260
261 * src/reader.h, src/reader.c (epilogue_set): New.
262 * src/parse-gram.y: Use it.
263
264 * data/bison.simple, data/bison.c++: b4_stype is now either not
265 defined, then default to int, or to the contents of %union,
266 without `union' itself.
267 Adjust.
268 * src/muscle_tab.c (muscle_init): Don't predefine `stype'.
269
270 * src/output.c (actions_output): Don't output braces, as they are
271 already handled by the scanner.
272
273 * src/scan-gram.l (SC_CHARACTER): Set the user_token_number of
274 characters to themselves.
275
276 * tests/reduce.at (Reduced Automaton): End the grammars with %% so
277 that the epilogue has a proper #line.
278
279 * src/parse-gram.y: Handle precedence/associativity.
280
281 * src/symtab.c (symbol_precedence_set): Requires the symbol to be
282 a terminal.
283 * src/scan-gram.l (SC_BRACED_CODE): Catch strings and characters.
284 * tests/calc.at: Do not use `%token "foo"' as it makes not sense
285 at all to define terminals that cannot be emitted.
286
287 * src/scan-gram.l: Escape M4 characters.
288
289 * src/scan-gram.l: Working properly with escapes in user
290 strings/characters.
291
292 * tests/torture.at (AT_DATA_TRIANGULAR_GRAMMAR)
293 (AT_DATA_HORIZONTAL_GRAMMAR): Respect the `%token ID NUM STRING'
294 grammar.
295 Use more modest sizes, as for the time being the parser does not
296 release memory, and therefore the process swallows a huge amount
297 of memory.
298
299 * tests/torture.at (AT_DATA_LOOKAHEADS_GRAMMAR): Adjust to the
300 stricter %token grammar.
301
302 * src/symtab.h (associativity): Add `undef_assoc'.
303 (symbol_precedence_set): Do nothing when passed an undef_assoc.
304 * src/symtab.c (symbol_check_alias_consistence): Adjust.
305
306 * tests/regression.at (Invalid %directive): Remove, as it is now
307 meaningless.
308 (Invalid inputs): Adjust to the new error messages.
309 (Token definitions): The new grammar doesn't allow too many
310 eccentricities.
311
312 * src/lex.h, src/lex.c: Remove.
313 * src/reader.c (lastprec, skip_to_char, read_signed_integer)
314 (copy_character, copy_string2, copy_string, copy_identifier)
315 (copy_comment, parse_token_decl, parse_type_decl, parse_assoc_decl)
316 (parse_muscle_decl, parse_dquoted_param, parse_skel_decl)
317 (parse_action): Remove.
318 * po/POTFILES.in: Adjust.
319
3202002-06-11 Akim Demaille <akim@epita.fr>
321
322 * src/reader.c (parse_action): Don't store directly into the
323 rule's action member: return the action as a string.
324 Don't require `rule_length' as an argument: compute it.
325 (grammar_current_rule_symbol_append)
326 (grammar_current_rule_action_append): New, eved out from
327 (readgram): here.
328 Remove `action_flag', `rulelength', unused now.
329
3302002-06-11 Akim Demaille <akim@epita.fr>
331
332 * src/reader.c (grammar_current_rule_prec_set).
333 (grammar_current_rule_check): New, eved out from...
334 (readgram): here.
335 Remove `xaction', `first_rhs': useless.
336 * tests/input.at (Type clashes): New.
337 * tests/existing.at (GNU Cim Grammar): Adjust.
338
3392002-06-11 Akim Demaille <akim@epita.fr>
340
341 * src/reader.c (grammar_midrule_action): New, Eved out from
342 (readgram): here.
343
3442002-06-11 Akim Demaille <akim@epita.fr>
345
346 * src/reader.c (grammar_rule_begin, previous_rule, current_rule):
347 New.
348 (readgram): Use them as replacement of inlined code, crule and
349 crule1.
350
3512002-06-11 Akim Demaille <akim@epita.fr>
352
353 * src/reader.c (grammar_end, grammar_symbol_append): New.
354 (readgram): Use them.
355 Make the use of `p' as local as possible.
356
3572002-06-10 Akim Demaille <akim@epita.fr>
358
359 GCJ's parser requires the tokens to be defined before the prologue.
360
361 * data/bison.simple: Output the token definition before the user's
362 prologue.
363 * tests/regression.at (Braces parsing, Duplicate string)
364 (Mixing %token styles): Check the output from bison.
365 (Early token definitions): New.
366
3672002-06-10 Akim Demaille <akim@epita.fr>
368
369 * src/symtab.c (symbol_user_token_number_set): Don't complain when
370 assigning twice the same user number to a token, so that we can
371 use it in...
372 * src/lex.c (lex): here.
373 Also use `symbol_class_set' instead of hand written code.
374 * src/reader.c (parse_assoc_decl): Likewise.
375
3762002-06-10 Akim Demaille <akim@epita.fr>
377
378 * src/symtab.c, src/symtab.c (symbol_class_set)
379 (symbol_user_token_number_set): New.
380 * src/reader.c (parse_token_decl): Use them.
381 Use a switch instead of ifs.
382 Use a single argument.
383
3842002-06-10 Akim Demaille <akim@epita.fr>
385
386 Remove `%thong' support as it is undocumented, unused, duplicates
387 `%token's job, and creates useless e-mail traffic with people who
388 want to know what it is, why it is undocumented, unused, and
389 duplicates `%token's job.
390
391 * src/reader.c (parse_thong_decl): Remove.
392 * src/options.c (option_table): Remove "thong".
393 * src/lex.h (tok_thong): Remove.
394
3952002-06-10 Akim Demaille <akim@epita.fr>
396
397 * src/symtab.c, src/symtab.c (symbol_type_set)
398 (symbol_precedence_set): New.
399 * src/reader.c (parse_type_decl, parse_assoc_decl): Use them.
400 (value_components_used): Remove, unused.
401
4022002-06-09 Akim Demaille <akim@epita.fr>
403
404 Move symbols handling code out of the reader.
405
406 * src/reader.h, src/reader.c (errtoken, undeftoken, eoftoken)
407 (axiom): Move to...
408 * src/symtab.h, src/symtab.c: here.
409
410 * src/gram.c (start_symbol): Remove: use startsymbol->number.
411 * src/reader.c (startval): Rename as...
412 * src/symtab.h, src/symtab.c (startsymbol): this.
413 * src/reader.c: Adjust.
414
415 * src/reader.c (symbol_check_defined, symbol_make_alias)
416 (symbol_check_alias_consistence, symbol_pack, symbol_translation)
417 (token_translations_init)
418 Move to...
419 * src/symtab.c: here.
420 * src/reader.c (packsymbols): Move to...
421 * src/symtab.h, src/symtab.c (symbols_pack): here.
422 * src/symtab.h, src/symtab.c (symbol_make_alias): Takes SYMVAL as
423 argument.
424
4252002-06-03 Akim Demaille <akim@epita.fr>
426
427 * src/muscle_tab.c (muscle_insert, muscle_find): Declarations,
428 then statements.
429
4302002-06-03 Akim Demaille <akim@epita.fr>
431
432 * src/muscle_tab.c (muscle_find, muscle_insert): Don't initialize
433 structs with non literals.
434 * src/scan-skel.l: never-interactive.
435 * src/conflicts.c (enum conflict_resolution_e): No trailing
436 comma.
437 * src/getargs.c (usage): Split long literal strings.
438 Reported by Hans Aberg.
439
4402002-05-28 Akim Demaille <akim@epita.fr>
441
442 * data/bison.c++: Use C++ ostreams.
443 (cdebug_): New member.
444
4452002-05-28 Akim Demaille <akim@epita.fr>
446
447 * src/output.c (output_skeleton): Be sure to allocate enough room
448 for `/' _and_ for `\0' in full_skeleton.
449
4502002-05-28 Akim Demaille <akim@epita.fr>
451
452 * data/bison.c++: Catch up with bison.simple:
453 2002-05-24 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
454 and Paul Eggert <eggert@twinsun.com>: `error' handing.
455 2002-05-26 Akim Demaille <akim@epita.fr>: stos_, token_number_,
456 and popping traces.
457
4582002-05-27 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
459
460 * src/output.c (output_skeleton): Put an explicit path in front of
461 the skeleton file name, rather than relying on the -I directory,
462 to partially alleviate effects of having a skeleton file lying around
463 in the current directory.
464
4652002-05-27 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
466
467 * src/conflicts.c (log_resolution): Correct typo:
468 obstack_printf should be obstack_fgrow1.
469
4702002-05-26 Akim Demaille <akim@epita.fr>
471
472 * src/state.h (state_t): `solved_conflicts' is a new member.
473 * src/LR0.c (new_state): Set it to 0.
474 * src/conflicts.h, src/conflicts.c (print_conflicts)
475 (free_conflicts, solve_conflicts): Rename as...
476 (conflicts_print, conflicts_free, conflicts_solve): these.
477 Adjust callers.
478 * src/conflicts.c (enum conflict_resolution_e)
479 (solved_conflicts_obstack): New, used by...
480 (log_resolution): this.
481 Adjust to attach the conflict resolution to each state.
482 Complete the description with the precedence/associativity
483 information.
484 (resolve_sr_conflict): Adjust.
485 * src/print.c (print_state): Output its solved_conflicts.
486 * tests/conflicts.at (Unresolved SR Conflicts)
487 (Solved SR Conflicts): Exercise --report=all.
488
4892002-05-26 Akim Demaille <akim@epita.fr>
490
491 * src/LR0.c, src/derives.c, src/gram.c, src/gram.h, src/lalr.c,
492 * src/nullable.c, src/output.c, src/print.c, src/print_graph.c,
493 * src/reader.c, src/reduce.c, src/state.h, src/symtab.h
494 (token_number_t, item_number_as_token_number)
495 (token_number_as_item_number, muscle_insert_token_number_table):
496 Rename as...
497 (symbol_number_t, item_number_as_symbol_number)
498 (symbol_number_as_item_number, muscle_insert_symbol_number_table):
499 these, since it is more appropriate.
500
5012002-05-26 Akim Demaille <akim@epita.fr>
502
503 * tests/calc.at (AT_CHECK_CALC): Adjust: there are now additional
504 `Error:' lines.
505 * data/bison.simple (yystos) [YYDEBUG]: New.
506 (yyparse) [YYDEBUG]: Display the symbols which are popped during
507 error recovery.
508 * tests/regression.at (Web2c Actions): Adjust: yystos is output now.
509
5102002-05-25 Akim Demaille <akim@epita.fr>
511
512 * doc/bison.texinfo (Debugging): Split into...
513 (Tracing): this new section, its former contents, and...
514 (Understanding): this new section.
515 * src/getargs.h, src/getargs.c (verbose_flag): Remove, replaced
516 by...
517 (report_flag): this.
518 Adjust all dependencies.
519 (report_args, report_types, report_argmatch): New.
520 (usage, getargs): Report/support -r, --report.
521 * src/options.h
522 (struct option_table_struct): Rename as..,
523 (struct option_table_s): this.
524 Rename the `set_flag' member to `flag' to match with getopt_long's
525 struct.
526 * src/options.c (option_table): Split verbose into an entry for
527 %verbose, and another for --verbose.
528 Support --report/-r, so remove -r from the obsolete --raw.
529 * src/print.c: Attach full item sets and lookaheads reports to
530 report_flag instead of trace_flag.
531 * lib/argmatch.h, lib/argmatch.c: New, from Fileutils 4.1.
532
5332002-05-24 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
534 and Paul Eggert <eggert@twinsun.com>
535
536 * data/bison.simple (yyparse): Correct error handling to conform to
537 POSIX and yacc. Specifically, after syntax error is discovered,
538 do not reduce further before shifting the error token.
539 Clean up the code a bit by removing the labels yyerrdefault,
540 yyerrhandle, yyerrpop.
541 * NEWS: Document the above.
542
5432002-05-20 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
544
545 * data/bison.simple (yyr1): Don't use yy_token_number_type as element
546 type; it isn't always big enough, since it doesn't necessarily
547 include non-terminals.
548 (yytranslate): Expand definition of yy_token_number_type, so that
549 the latter can be removed.
550 (yy_token_number_type): Remove, only one use.
551 * data/bison.c++ (r1_): Parallel change to yyr1 in bison.simple---
552 don't use TokenNumberType as element type.
553
554 * tests/regression.at: Modify expected output to agree with change
555 to yyr1 and yytranslate.
556
5572002-05-13 Florian Krohm <florian@edamail.fishkill.ibm.com>
558
559 * src/reader.c (parse_action): Use copy_character instead of
560 obstack_1grow.
561
5622002-05-13 Akim Demaille <akim@epita.fr>
563
564 * tests/regression.at (Token definitions): Prototype yylex and
565 yyerror.
566
5672002-05-12 Paul Hilfinger <Hilfinger@CS.Berkeley.EDU>
568
569 * src/scan-skel.l: Correct off-by-one error in handling of __oline__.
570 * data/bison.simple (b4_sint_type, b4_uint_type): Correct to reflect
571 32-bit arithmetic.
572 * data/bison.c++ (b4_sint_type, b4_uint_type): Ditto.
573
5742002-05-07 Akim Demaille <akim@epita.fr>
575
576 * tests/synclines.at: Be sure to prototype yylex and yyerror to
577 avoid GCC warnings.
578
5792002-05-07 Akim Demaille <akim@epita.fr>
580
581 Kill GCC warnings.
582
583 * src/reduce.c (nonterminals_reduce): Don't loop over RITEM: loop
584 over the RHS of each rule.
585 * src/gram.h, src/gram.c (nritems): Is `unsigned int', not int.
586 * src/state.h (state_t): Member `nitems' is unsigned short.
587 * src/LR0.c (get_state): Adjust.
588 * src/reader.c (packgram): Likewise.
589 * src/output.c (GENERATE_MUSCLE_INSERT_TABLE): `max' is of type
590 `Type'.
591 (muscle_insert_int_table): Remove, unused.
592 (prepare_rules): Remove `max'.
593
5942002-05-06 Akim Demaille <akim@epita.fr>
595
596 * src/closure.c (print_firsts): Display of the symbol tags.
597 (bitmatrix_print): Move to...
598 * lib/bitsetv-print.h, lib/bitsetv-print.c (bitsetv_matrix_dump):
599 here.
600 * tests/sets.at (Nullable, Broken Closure, Firsts): Adjust.
601
6022002-05-06 Akim Demaille <akim@epita.fr>
603
604 * src/muscle_tab.c (muscle_m4_output): Must return TRUE for
605 hash_do_for_each.
606
6072002-05-06 Akim Demaille <akim@epita.fr>
608
609 * src/LR0.c (new_state, get_state): Instead of using the global
610 `kernel_size' and `kernel_base', have two new arguments:
611 `core_size' and `core'.
612 Adjust callers.
613
6142002-05-06 Akim Demaille <akim@epita.fr>
615
616 * src/reader.c (packgram): No longer end `ritem' with a 0
617 sentinel: it is not used.
618
6192002-05-05 Akim Demaille <akim@epita.fr>
620
621 New experimental feature: display the lookaheads in the report and
622 graph.
623
624 * src/print (print_core): When --trace-flag, display the rules
625 lookaheads.
626 * src/print_graph.c (print_core): Likewise.
627 Swap the arguments.
628 Adjust caller.
629
6302002-05-05 Akim Demaille <akim@epita.fr>
631
632 * tests/torture.at (Many lookaheads): New test.
633
6342002-05-05 Akim Demaille <akim@epita.fr>
635
636 * src/output.c (GENERATE_OUTPUT_TABLE): Replace with...
637 (GENERATE_MUSCLE_INSERT_TABLE): this.
638 (output_int_table, output_unsigned_int_table, output_short_table)
639 (output_token_number_table, output_item_number_table): Replace with...
640 (muscle_insert_int_table, muscle_insert_unsigned_int_table)
641 (muscle_insert_short_table, muscle_insert_token_number_table)
642 (muscle_insert_item_number_table): these.
643 Adjust all callers.
644 (prepare_tokens): Don't free `translations', since...
645 * src/reader.h, src/reader.c (grammar_free): do it.
646 Move to...
647 * src/gram.h, src/gram.c (grammar_free): here.
648 * data/bison.simple, data/bison.c++: b4_token_number_max is now
649 b4_translate_max.
650
6512002-05-05 Akim Demaille <akim@epita.fr>
652
653 * src/output.c (output_unsigned_int_table): New.
654 (prepare_rules): `i' is unsigned.
655 `prhs', `rline', `r2' are unsigned int.
656 Rename muscle `rhs_number_max' as `rhs_max'.
657 Output muscles `prhs_max', `rline_max', and `r2_max'.
658 Free rline and r1.
659 * data/bison.simple, data/bison.c++: Adjust to use these muscles
660 to compute types instead of constant types.
661 * tests/regression.at (Web2c Actions): Adjust.
662
6632002-05-04 Akim Demaille <akim@epita.fr>
664
665 * src/symtab.h (SALIAS, SUNDEF): Rename as...
666 (USER_NUMBER_ALIAS, USER_NUMBER_UNDEFINED): these.
667 Adjust dependencies.
668 * src/output.c (token_definitions_output): Be sure not to output a
669 `#define 'a'' when fed with `%token 'a' "a"'.
670 * tests/regression.at (Token definitions): New.
671
6722002-05-03 Paul Eggert <eggert@twinsun.com>
673
674 * data/bison.simple (b4_token_defines): Also define YYTOKENTYPE
675 for K&R C.
676
6772002-05-03 gettextize <bug-gnu-gettext@gnu.org>
678
679 * Makefile.am (SUBDIRS): Remove intl.
680 (EXTRA_DIST): Add config/config.rpath.
681
6822002-05-03 Akim Demaille <akim@epita.fr>
683
684 * data/bison.simple (m4_if): Don't output empty enums.
685 And actually, output valid enum definitions :(.
686
6872002-05-03 Akim Demaille <akim@epita.fr>
688
689 * configure.bat: Remove, completely obsolete.
690 * Makefile.am (EXTRA_DIST): Adjust.
691 Don't distribute config.rpath...
692 * config/Makefile.am (EXTRA_DIST): Do it.
693
6942002-05-03 Akim Demaille <akim@epita.fr>
695
696 * configure.in (GETTEXT_VERSION): New.
697 Suggested by Bruno Haible for the forthcoming Gettext 0.10.3.
698
6992002-05-03 Akim Demaille <akim@epita.fr>
700
701 * data/bison.simple (b4_token_enum): New.
702 (b4_token_defines): Use it to output tokens both as #define and
703 enums.
704 Suggested by Paul Eggert.
705 * src/output.c (token_definitions_output): Don't output spurious
706 white spaces.
707
7082002-05-03 Akim Demaille <akim@epita.fr>
709
710 * data/m4sugar/m4sugar.m4: Update from CVS Autoconf.
711
7122002-05-02 Robert Anisko <robert@lrde.epita.fr>
713
714 * data/bison.c++: Adapt expansion of $s and @s to the C++ parser.
715 Update the stack class, give a try to deque as the default container.
716
7172002-05-02 Akim Demaille <akim@epita.fr>
718
719 * data/bison.simple (yyparse): Do not implement @$ = @1.
720 (YYLLOC_DEFAULT): Adjust to do it.
721 * doc/bison.texinfo (Location Default Action): Fix.
722
7232002-05-02 Akim Demaille <akim@epita.fr>
724
725 * src/reader.c (parse_braces): Merge into...
726 (parse_action): this.
727
7282002-05-02 Akim Demaille <akim@epita.fr>
729
730 * configure.in (ALL_LINGUAS): Remove.
731 * po/LINGUAS, hr.po: New.
732
7332002-05-02 Akim Demaille <akim@epita.fr>
734
735 Remove the so called hairy (semantic) parsers.
736
737 * src/system.h (EXT_GUARD_C, EXT_STYPE_H): Remove.
738 * src/gram.h, src/gram.c (semantic_parser): Remove.
739 (rule_t): Remove the guard and guard_line members.
740 * src/lex.h (token_t): remove tok_guard.
741 * src/options.c (option_table): Remove %guard and %semantic_parser
742 support.
743 * src/output.c, src/output.h (guards_output): Remove.
744 (prepare): Adjust.
745 (token_definitions_output): Don't output the `T'
746 tokens (???).
747 (output_skeleton): Don't output the guards.
748 * src/files.c, src/files.c (attrsfile): Remove.
749 * src/reader.c (symbol_list): Remove the guard and guard_line
750 members.
751 Adjust dependencies.
752 (parse_guard): Remove.
753 * data/bison.hairy: Remove.
754 * doc/bison.texinfo (Environment Variables): Remove occurrences of
755 BISON_HAIRY.
756
7572002-05-02 Akim Demaille <akim@epita.fr>
758
759 * src/reader.c (copy_at, copy_dollarm parse_braces, parse_action)
760 (parse_guard): Rename the formal argument `stack_offset' as
761 `rule_length', which is more readable.
762 Adjust callers.
763 (copy_at, copy_dollar): Instead of outputting the hard coded
764 values of $$, $n and so forth, output invocation to b4_lhs_value,
765 b4_lhs_location, b4_rhs_value, and b4_rhs_location.
766 Note: this patch partially drops `semantic-parser' support: it
767 always does `rule_length - n', where semantic parsers ought to
768 always use `-n'.
769 * data/bison.simple, data/bison.c++ (b4_lhs_value)
770 (b4_lhs_location, b4_rhs_value, and b4_rhs_location: New.
771
7722002-05-02 Akim Demaille <akim@epita.fr>
773
774 * configure.in (AC_INIT): Bump to 1.49b.
775 (AM_INIT_AUTOMAKE): Short invocation.
776
7772002-05-02 Akim Demaille <akim@epita.fr>
778
779 Version 1.49a.
780
7812002-05-01 Akim Demaille <akim@epita.fr>
782
783 * src/skeleton.h: Remove.
784
7852002-05-01 Akim Demaille <akim@epita.fr>
786
787 * src/skeleton.h: Fix the #endif.
788 Reported by Magnus Fromreide.
789
7902002-04-26 Paul Eggert <eggert@twinsun.com>
791
792 * data/bison.simple (YYSTYPE_IS_TRIVIAL, YYLTYPE_IS_TRIVIAL):
793 Define if we define YYSTYPE and YYLTYPE, respectively.
794 (YYCOPY): Fix [] quoting problem in the non-GCC case.
795
7962002-04-25 Robert Anisko <robert@lrde.epita.fr>
797
798 * src/scan-skel.l: Postprocess quadrigraphs.
799
800 * src/reader.c (copy_character): New function, used to output
801 single characters while replacing `[' and `]' with quadrigraphs, to
802 avoid troubles with M4 quotes.
803 (copy_comment): Output characters with copy_character.
804 (read_additionnal_code): Likewise.
805 (copy_string2): Likewise.
806 (copy_definition): Likewise.
807
808 * tests/calc.at: Exercise M4 quoting.
809
8102002-04-25 Akim Demaille <akim@epita.fr>
811
812 * tests/sets.at (AT_EXTRACT_SETS): Sed portability issue: no space
813 between `!' and the command.
814 Reported by Paul Eggert.
815
8162002-04-24 Robert Anisko <robert@lrde.epita.fr>
817
818 * tests/calc.at: Exercise prologue splitting.
819
820 * data/bison.simple, data/bison.c++: Use `b4_pre_prologue' and
821 `b4_post_prologue' instead of `b4_prologue'.
822
823 * src/output.c (prepare): Add the `pre_prologue' and `post_prologue'
824 muscles.
825 (output): Free pre_prologue_obstack and post_prologue_obstack.
826 * src/files.h, src/files.c (attrs_obstack): Remove.
827 (pre_prologue_obstack, post_prologue_obstack): New.
828 * src/reader.c (copy_definition): Add a parameter to specify the
829 obstack to fill, instead of using attrs_obstack unconditionally.
830 (read_declarations): Pass pre_prologue_obstack to copy_definition if
831 `%union' has not yet been seen, pass post_prologue_obstack otherwise.
832
8332002-04-23 Paul Eggert <eggert@twinsun.com>
834
835 * data/bison.simple: Remove unnecessary commentary and white
836 space differences from 1_29-branch.
837 Depend on YYERROR_VERBOSE, not defined (YYERROR_VERBOSE).
838
839 (union yyalloc, YYSTACK_GAP_MAX, YYSTACK_BYTES, YYCOPY,
840 YYSTACK_RELOCATE): Do not define if yyoverflow is defined, or
841 if this is a C++ parser and YYSTYPE or YYLTYPE has nontrivial
842 constructors or destructors.
843
844 (yyparse) [! defined YYSTACK_RELOCATE]: Do not relocate the stack.
845
8462002-04-23 Akim Demaille <akim@epita.fr>
847
848 * tests/sets.at (AT_EXTRACT_SETS): Don't use 8 char long sed labels.
849 * tests/synclines.at (AT_TEST_SYNCLINE): Be robust to GCC's
850 location with columns.
851 * tests/conflicts.at (%nonassoc and eof): Don't use `error.h'.
852 All reported by Paul Eggert.
853
8542002-04-22 Akim Demaille <akim@epita.fr>
855
856 * src/reduce.c (dump_grammar): Move to...
857 * src/gram.h, src/gram.c (grammar_dump): here.
858 Be sure to separate long item numbers.
859 Don't read the members of a rule's prec if its nil.
860
8612002-04-22 Akim Demaille <akim@epita.fr>
862
863 * src/output.c (table_size, table_grow): New.
864 (MAXTABLE): Remove, replace uses with table_size.
865 (pack_vector): Instead of dying when the table is too big, grow it.
866
8672002-04-22 Akim Demaille <akim@epita.fr>
868
869 * data/bison.simple (yyr1): Its type is that of a token number.
870 * data/bison.c++ (r1_): Likewise.
871 * tests/regression.at (Web2c Actions): Adjust.
872
8732002-04-22 Akim Demaille <akim@epita.fr>
874
875 * src/reader.c (token_translations_init): 256 is now the default
876 value for the error token, i.e., it will be assigned another
877 number if the user assigned 256 to one of her tokens.
878 (reader): Don't force 256 to error.
879 * doc/bison.texinfo (Symbols): Adjust.
880 * tests/torture.at (AT_DATA_HORIZONTAL_GRAMMAR)
881 (AT_DATA_TRIANGULAR_GRAMMAR): Number the tokens as 1, 2, 3
882 etc. instead of 10, 20, 30 (which was used to `jump' over error
883 (256) and undefined (2)).
884
8852002-04-22 Akim Demaille <akim@epita.fr>
886
887 Propagate more token_number_t.
888
889 * src/gram.h (token_number_as_item_number)
890 (item_number_as_token_number): New.
891 * src/output.c (GENERATE_OUTPUT_TABLE): New.
892 Use it to create output_item_number_table and
893 output_token_number_table.
894 * src/LR0.c, src/derives.c, src/gram.c, src/gram.h, src/lalr.c,
895 * src/lex.c, src/nullable.c, src/output.c, src/print.c,
896 * src/print_graph.c, src/reader.c, src/reduce.c, src/state.h,
897 * src/symtab.c, src/symtab.h: Use token_number_t instead of shorts.
898
8992002-04-22 Akim Demaille <akim@epita.fr>
900
901 * src/output.h, src/output.c (get_lines_number): Remove.
902
9032002-04-19 Akim Demaille <akim@epita.fr>
904
905 * doc/bison.texinfo (Actions): Make clear that `|' is not the same
906 as Lex/Flex'.
907 (Debugging): More details about enabling the debugging features.
908 (Table of Symbols): Describe $$, $n, @$, and @n.
909 Suggested by Tim Josling.
910
9112002-04-19 Akim Demaille <akim@epita.fr>
912
913 * doc/bison.texinfo: Remove the uses of the obsolete @refill.
914
9152002-04-10 Akim Demaille <akim@epita.fr>
916
917 * src/system.h: Rely on HAVE_LIMITS_H.
918 Suggested by Paul Eggert.
919
9202002-04-09 Akim Demaille <akim@epita.fr>
921
922 * tests/calc.at (_AT_CHECK_CALC_ERROR): Receive as argument the
923 full stderr, and strip it according to the bison options, instead
924 of composing the error message from different bits.
925 This makes it easier to check for several error messages.
926 Adjust all the invocations.
927 Add an invocation exercising the error token.
928 Add an invocation demonstrating a stupid error message.
929 (_AT_DATA_CALC_Y): Follow the GCS: initial column is 1, not 0.
930 Adjust the tests.
931 Error message are for stderr, not stdout.
932
9332002-04-09 Akim Demaille <akim@epita.fr>
934
935 * src/gram.h, src/gram.c (error_token_number): Remove, use
936 errtoken->number.
937 * src/reader.c (reader): Don't specify the user token number (2)
938 for $undefined, as it uselessly prevents using it.
939 * src/gram.h (token_number_t): Move to...
940 * src/symtab.h: here.
941 (state_t.number): Is a token_number_t.
942 * src/print.c, src/reader.c: Use undeftoken->number instead of
943 hard coded 2.
944 (Even though this 2 is not the same as above: the number of the
945 undeftoken remains being 2, it is its user token number which
946 might not be 2).
947 * src/output.c (prepare_tokens): Rename the `maxtok' muscle with
948 `user_token_number_max'.
949 Output `undef_token_number'.
950 * data/bison.simple, data/bison.c++: Use them.
951 Be sure to map invalid yylex return values to
952 `undef_token_number'. This saves us from gratuitous SEGV.
953
954 * tests/conflicts.at (Solved SR Conflicts)
955 (Unresolved SR Conflicts): Adjust.
956 * tests/regression.at (Web2c Actions): Adjust.
957
9582002-04-08 Akim Demaille <akim@epita.fr>
959
960 * data/bison.c++: s/b4_item_number_max/b4_rhs_number_max/.
961 Adding #line.
962 Remove the duplicate `typedefs'.
963 (RhsNumberType): Fix the declaration and various other typos.
964 Use __ofile__.
965 * data/bison.simple: Use __ofile__.
966 * src/scan-skel.l: Handle __ofile__.
967
9682002-04-08 Akim Demaille <akim@epita.fr>
969
970 * src/gram.h (item_number_t): New, the type of item numbers in
971 RITEM. Note that it must be able to code symbol numbers as
972 positive number, and the negation of rule numbers as negative
973 numbers.
974 Adjust all dependencies (pretty many).
975 * src/reduce.c (rule): Remove this `short *' pointer: use
976 item_number_t.
977 * src/system.h (MINSHORT, MAXSHORT): Remove.
978 Include `limits.h'.
979 Adjust dependencies to using SHRT_MAX and SHRT_MIN.
980 (shortcpy): Remove.
981 (MAXTABLE): Move to...
982 * src/output.c (MAXTABLE): here.
983 (prepare_rules): Use output_int_table to output rhs.
984 * data/bison.simple, data/bison.c++: Adjust.
985 * tests/torture.at (Big triangle): Move the limit from 254 to
986 500.
987 * tests/regression.at (Web2c Actions): Ajust.
988
989 Trying with bigger grammars shows various phenomena: at 3000 (28Mb
990 of grammar file) bison is killed by my system, at 2000 (12Mb) bison
991 passes, but produces negative #line number, once fixed, GCC is
992 killed while compiling 14Mb, at 1500 (6.7 Mb of grammar, 8.2Mb of
993 C), it passes.
994 * src/state.h (state_h): Code input lines on ints, not shorts.
995
9962002-04-08 Akim Demaille <akim@epita.fr>
997
998 * src/reduce.c (reduce_grammar): First reduce the nonterminals,
999 and then the grammar.
1000
10012002-04-08 Akim Demaille <akim@epita.fr>
1002
1003 * src/system.h: No longer using strndup.
1004
10052002-04-07 Akim Demaille <akim@epita.fr>
1006
1007 * src/muscle_tab.h (MUSCLE_INSERT_LONG_INT): New.
1008 * src/output.c (output_table_data): Return the longest number.
1009 (prepare_tokens): Output `token_number_max').
1010 * data/bison.simple, data/bison.c++ (b4_sint_type, b4_uint_type):
1011 New.
1012 Use them to define yy_token_number_type/TokenNumberType.
1013 Use this type for yytranslate.
1014 * tests/torture.at (Big triangle): Push the limit from 124 to
1015 253.
1016 * tests/regression.at (Web2c Actions): Adjust.
1017
10182002-04-07 Akim Demaille <akim@epita.fr>
1019
1020 * tests/torture.at (Big triangle): New.
1021 (GNU AWK Grammar, GNU Cim Grammar): Move to...
1022 * tests/existing.at: here.
1023
10242002-04-07 Akim Demaille <akim@epita.fr>
1025
1026 * src/gram.h, src/gram.c (nitems): Remove, it is an alias of
1027 nritems.
1028 Adjust dependencies.
1029
10302002-04-07 Akim Demaille <akim@epita.fr>
1031
1032 * src/reader.c: Normalize increments to prefix form.
1033
10342002-04-07 Akim Demaille <akim@epita.fr>
1035
1036 * src/reader.c, symtab.c: Remove debugging code.
1037
10382002-04-07 Akim Demaille <akim@epita.fr>
1039
1040 Rename all the `bucket's as `symbol_t'.
1041
1042 * src/gram.c, src/gram.h, src/lex.c, src/lex.h, src/output.c,
1043 * src/reader.c, src/reader.h, src/reduce.c, src/state.h,
1044 * src/symtab.c, src/symtab.h (bucket): Rename as...
1045 (symbol_t): this.
1046 (symbol_list_new, bucket_check_defined, bucket_make_alias)
1047 (bucket_check_alias_consistence, bucket_pack, bucket_translation)
1048 (bucket_new, bucket_free, hash_compare_bucket, hash_bucket)
1049 (buckets_new, buckets_free, buckets_do): Rename as...
1050 (symbol_list_new, symbol_check_defined, symbol_make_alias)
1051 (symbol_check_alias_consistence, symbol_pack, symbol_translation)
1052 (symbol_new, symbol_free, hash_compare_symbol_t, hash_symbol_t)
1053 (symbols_new, symbols_free, symbols_do): these.
1054
10552002-04-07 Akim Demaille <akim@epita.fr>
1056
1057 Use lib/hash for the symbol table.
1058
1059 * src/gram.c (ntokens): Initialize to 1, to reserve a slot for
1060 EOF.
1061 * src/lex.c (lex): Set the `number' member of new terminals.
1062 * src/reader.c (bucket_check_defined, bucket_make_alias)
1063 (bucket_check_alias_consistence, bucket_translation): New.
1064 (reader, grammar_free, readgram, token_translations_init)
1065 (packsymbols): Adjust.
1066 (reader): Number the predefined tokens.
1067 * src/reduce.c (inaccessable_symbols): Just use hard coded numbers
1068 for predefined tokens.
1069 * src/symtab.h (bucket): Remove all the hash table related
1070 members.
1071 * src/symtab.c (symtab): Replace by...
1072 (bucket_table): this.
1073 (bucket_new, bucket_free, hash_compare_bucket, hash_bucket)
1074 (buckets_new, buckets_do): New.
1075
10762002-04-07 Akim Demaille <akim@epita.fr>
1077
1078 * src/gram.c (nitems, nrules, nsyms, ntokens, nvars, nritems)
1079 (start_symbol, max_user_token_number, semantic_parser)
1080 (error_token_number): Initialize.
1081 * src/reader.c (grammar, start_flag, startval, typed, lastprec):
1082 Initialize.
1083 (reader): Don't.
1084 (errtoken, eoftoken, undeftoken, axiom): Extern.
1085
10862002-04-07 Akim Demaille <akim@epita.fr>
1087
1088 * src/gram.h (rule_s): prec and precsym are now pointers
1089 to the bucket giving the priority/associativity.
1090 Member `associativity' removed: useless.
1091 * src/reduce.c, src/conflicts.c: Adjust.
1092
10932002-04-07 Akim Demaille <akim@epita.fr>
1094
1095 * src/lalr.c, src/LR0.c, src/closure.c, src/gram.c, src/reduce.c:
1096 Properly escape the symbols' TAG when outputting them.
1097
10982002-04-07 Akim Demaille <akim@epita.fr>
1099
1100 * src/lalr.h (LA): Is a bitsetv, not bitset*.
1101
11022002-04-07 Akim Demaille <akim@epita.fr>
1103
1104 * src/lalr.h, src/lalr.c (LAruleno): Replace with...
1105 (LArule): this, which is an array to rule_t*.
1106 * src/print.c, src/conflicts.c: Adjust.
1107
11082002-04-07 Akim Demaille <akim@epita.fr>
1109
1110 * src/gram.h (rule_t): Rename `number' as `user_number'.
1111 `number' is a new member.
1112 Adjust dependencies.
1113 * src/reduce.c (reduce_grammar_tables): Renumber rule_t.number.
1114
11152002-04-07 Akim Demaille <akim@epita.fr>
1116
1117 As a result of the previous patch, it is no longer needed
1118 to reorder ritem itself.
1119
1120 * src/reduce.c (reduce_grammar_tables): Don't sort RITEM.
1121
11222002-04-07 Akim Demaille <akim@epita.fr>
1123
1124 Be sure never to walk through RITEMS, but use only data related to
1125 the rules themselves. RITEMS should be banished.
1126
1127 * src/output.c (output_token_translations): Rename as...
1128 (prepare_tokens): this.
1129 In addition to `translate', prepare the muscles `tname' and
1130 `toknum', which were handled by...
1131 (output_rule_data): this.
1132 Remove, and move the remainder of its outputs into...
1133 (prepare_rules): this new routines, which also merges content from
1134 (output_gram): this.
1135 (prepare_rules): Be sure never to walk through RITEMS.
1136 (output_stos): Rename as...
1137 (prepare_stos): this.
1138 (output): Always invoke prepare_states, after all, just don't use it
1139 in the output if you don't need it.
1140
11412002-04-07 Akim Demaille <akim@epita.fr>
1142
1143 * src/LR0.c (new_state): Display `nstates' as the name of the
1144 newly created state.
1145 Adjust to initialize first_state and last_state if needed.
1146 Be sure to distinguish the initial from the final state.
1147 (new_states): Create the itemset of the initial state, and use
1148 new_state.
1149 * src/closure.c (closure): Now that the initial state has its
1150 items properly set, there is no need for a special case when
1151 creating `ruleset'.
1152
1153 As a result, now the rule 0, reducing to $axiom, is visible in the
1154 outputs. Adjust the test suite.
1155
1156 * tests/conflicts.at (Solved SR Conflicts)
1157 (Unresolved SR Conflicts): Adjust.
1158 * tests/regression.at (Web2c Report, Rule Line Numbers): Idem.
1159 * tests/conflicts.at (S/R in initial): New.
1160
11612002-04-07 Akim Demaille <akim@epita.fr>
1162
1163 * src/LR0.c (allocate_itemsets): Don't loop over ritem: loop over
1164 the RHS of the rules.
1165 * src/output.c (output_gram): Likewise.
1166
11672002-04-07 Akim Demaille <akim@epita.fr>
1168
1169 * src/gram.h (rule_t): `lhs' is now a pointer to the symbol's
1170 bucket.
1171 Adjust all dependencies.
1172 * src/reduce.c (nonterminals_reduce): Don't forget to renumber the
1173 `number' of the buckets too.
1174 * src/gram.h: Include `symtab.h'.
1175 (associativity): Move to...
1176 * src/symtab.h: here.
1177 No longer include `gram.h'.
1178
11792002-04-07 Akim Demaille <akim@epita.fr>
1180
1181 * src/gram.h, src/gram.c (rules_rhs_length): New.
1182 (ritem_longest_rhs): Use it.
1183 * src/gram.h (rule_t): `number' is a new member.
1184 * src/reader.c (packgram): Set it.
1185 * src/reduce.c (reduce_grammar_tables): Move the useless rules at
1186 the end of `rules', and count them out of `nrules'.
1187 (reduce_output, dump_grammar): Adjust.
1188 * src/print.c (print_grammar): It is no longer needed to check for
1189 the usefulness of a rule, as useless rules are beyond `nrules + 1'.
1190 * tests/reduce.at (Reduced Automaton): New test.
1191
11922002-04-07 Akim Demaille <akim@epita.fr>
1193
1194 * src/reduce.c (inaccessable_symbols): Fix a buglet: because of a
1195 lacking `+ 1' to nrules, Bison reported as useless a token if it
1196 was used solely to set the precedence of the last rule...
1197
11982002-04-07 Akim Demaille <akim@epita.fr>
1199
1200 * data/bison.c++, data/bison.simple: Don't output the current file
1201 name in #line, to avoid useless diffs between two identical
1202 outputs under different names.
1203
12042002-04-07 Akim Demaille <akim@epita.fr>
1205
1206 * src/closure.c, src/print.c, src/reader.c, src/reduce.c:
1207 Normalize loops to using `< nrules + 1', not `<= nrules'.
1208
12092002-04-07 Akim Demaille <akim@epita.fr>
1210
1211 * TODO: Update.
1212
12132002-04-07 Akim Demaille <akim@epita.fr>
1214
1215 * src/output.c, src/reader.c, src/symtab.c, src/symtab.h: Rename
1216 bucket.value as bucket.number.
1217
12182002-04-07 Akim Demaille <akim@epita.fr>
1219
1220 * src/closure.c, src/derives.c, src/gram.h, src/lalr.c,
1221 * src/nullable.c, src/output.c, src/print.c, src/print_graph.c,
1222 * src/reader.c, src/reduce.c: Let rule_t.rhs point directly to the
1223 RHS, instead of being an index in RITEMS.
1224
12252002-04-04 Paul Eggert <eggert@twinsun.com>
1226
1227 * doc/bison.texinfo: Update copyright date.
1228 (Rpcalc Lexer, Symbols, Token Decl): Don't assume ASCII.
1229 (Symbols): Warn about running Bison in one character set,
1230 but compiling and/or running in an incompatible one.
1231 Warn about character code 256, too.
1232
12332002-04-03 Paul Eggert <eggert@twinsun.com>
1234
1235 * src/bison.data (YYSTACK_ALLOC): Depend on whether
1236 YYERROR_VERBOSE is nonzero, not whether it is defined.
1237
1238 Merge changes from bison-1_29-branch.
1239
12402002-03-20 Paul Eggert <eggert@twinsun.com>
1241
1242 Merge fixes from Debian bison_1.34-1.diff.
1243
1244 * configure.in (AC_PREREQ): 2.53.
1245
12462002-03-20 Akim Demaille <akim@epita.fr>
1247
1248 * src/conflicts.c (log_resolution): Argument `resolution' is const.
1249
12502002-03-19 Paul Eggert <eggert@twinsun.com>
1251
1252 * src/bison.simple (YYCOPY): New macro.
1253 (YYSTACK_RELOCATE): Use it.
1254 Remove Type arg; no longer needed. All callers changed.
1255 (yymemcpy): Remove; no longer needed.
1256
1257 * Makefile.am (AUTOMAKE_OPTIONS): 1.6.
1258 * doc/Makefile.am (AUTOMAKE_OPTIONS): Remove.
1259
12602002-03-19 Akim Demaille <akim@epita.fr>
1261
1262 Test and fix the #line outputs.
1263
1264 * tests/atlocal.at (GCC): New.
1265 * tests/synclines.at (AT_TEST_SYNCLINE): New macro.
1266 (Prologue synch line, ,%union synch line, Postprologue synch line)
1267 (Action synch line, Epilogue synch line): New tests.
1268 * src/reader.c (parse_union_decl): Define the muscle stype_line.
1269 * data/bison.simple, data/bison.c++: Use it.
1270
12712002-03-19 Akim Demaille <akim@epita.fr>
1272
1273 * tests/regression.at (%nonassoc and eof, Unresolved SR Conflicts)
1274 (Solved SR Conflicts, %expect not enough, %expect right)
1275 (%expect too much): Move to...
1276 * tests/conflicts.at: this new file.
1277
12782002-03-19 Akim Demaille <akim@epita.fr>
1279
1280 * data/m4sugar/m4sugar.m4: Update from CVS Autoconf.
1281 * data/bison.simple, data/bison.c++: Handle the `#define' part, so
1282 that we can move to enums for instance.
1283 * src/output.c (token_definitions_output): Output a list of
1284 `token-name, token-number' instead of the #define.
1285 (output_skeleton): Name this list `b4_tokens', not `b4_tokendefs'.
1286
12872002-03-14 Akim Demaille <akim@epita.fr>
1288
1289 Use Gettext 0.11.1.
1290
12912002-03-09 Robert Anisko <robert@lrde.epita.fr>
1292
1293 * data/bison.c++: Make the user able to add members to the generated
1294 parser by subclassing.
1295
12962002-03-05 Robert Anisko <robert@lrde.epita.fr>
1297
1298 * src/reader.c (read_additionnal_code): `c' should be an integer, not
1299 a character.
1300 Reported by Nicolas Tisserand and Nicolas Burrus.
1301
13022002-03-04 Robert Anisko <robert@lrde.epita.fr>
1303
1304 * src/reader.c: Warn about lacking semi-colons, do not complain.
1305
13062002-03-04 Robert Anisko <robert@lrde.epita.fr>
1307
1308 * data/bison.c++: Remove a debug line.
1309
13102002-03-04 Robert Anisko <robert@lrde.epita.fr>
1311
1312 * data/bison.c++: Unmerge value as yylval and value as yyval. Unmerge
1313 location as yylloc and location as yyloc. Use YYLLOC_DEFAULT, and
1314 provide a default implementation.
1315
13162002-03-04 Akim Demaille <akim@epita.fr>
1317
1318 * tests/input.at (Invalid $n, Invalid @n): Add the ending `;'.
1319 * tests/output.at (AT_CHECK_OUTPUT): Likewise.
1320 * tests/headers.at (AT_TEST_CPP_GUARD_H): Ditto.
1321 * tests/semantic.at (Parsing Guards): Similarly.
1322 * src/reader.at (readgram): Complain if the last rule is not ended
1323 with a semi-colon.
1324
13252002-03-04 Akim Demaille <akim@epita.fr>
1326
1327 * src/warshall.h, src/warshall.c (bitmatrix_print): Move to...
1328 * src/closure.c: here.
1329 (set_firsts): Use bitsetv_reflexive_transitive_closure instead of
1330 RTC.
1331 * src/warshall.h, src/warshall.c: Remove.
1332 * tests/sets.at (Broken Closure): Adjust.
1333
13342002-03-04 Akim Demaille <akim@epita.fr>
1335
1336 * src/output.c (output_skeleton): tempdir is const.
1337 bytes_read is unused.
1338
13392002-03-04 Akim Demaille <akim@epita.fr>
1340
1341 * lib/bbitset.h, lib/bitset.c, lib/bitset.h, lib/bitsetv.c,
1342 * lib/bitsetv.h, lib/ebitset.c, lib/lbitset.c, lib/sbitset.c:
1343 Update.
1344 From Michael Hayes.
1345
13462002-03-04 Akim Demaille <akim@epita.fr>
1347
1348 * src/closure.c (closure): `r' is unused.
1349
13502002-03-04 Akim Demaille <akim@epita.fr>
1351
1352 * tests/sets.at (Broken Closure): Add the ending `;'.
1353 * src/reader.at (readgram): Complain if a rule is not ended with a
1354 semi-colon.
1355
13562002-03-04 Akim Demaille <akim@epita.fr>
1357
1358 * src/conflicts.c (set_conflicts): Use bitset_disjoint_p.
1359 (count_sr_conflicts): Use bitset_count.
1360 * src/reduce.c (inaccessable_symbols): Ditto.
1361 (bits_size): Remove.
1362 * src/warshall.h, src/warshall.c: Convert to bitsetv.
1363
13642002-03-04 Akim Demaille <akim@epita.fr>
1365
1366 * src/closure.c, src/conflicts.c, src/lalr.c, src/print.c,
1367 * src/reduce.c: Remove the `bitset_zero's following the
1368 `bitset_create's, as now it is performed by the latter.
1369
13702002-03-04 Akim Demaille <akim@epita.fr>
1371
1372 * lib/bitset.c, lib/bitset.h, lib/bitsetv.c, lib/bitsetv.h,
1373 * lib/ebitset.c, lib/ebitset.h, lib/lbitset.c, lib/lbitset.h,
1374 * lib/sbitset.c, lib/sbitset.h, lib/bbitset.h: Update from the
1375 latest sources from Michael.
1376
13772002-03-04 Akim Demaille <akim@epita.fr>
1378
1379 * src/output.c (output): Don't free the grammar.
1380 * src/reader.c (grammar_free): New.
1381 * src/main.c (main): Call it and don't free symtab here.
1382
13832002-03-04 Akim Demaille <akim@epita.fr>
1384
1385 * src/lex.c (parse_percent_token): Be sure to 0-end token_buffer
1386 before returning.
1387 Reported by Benoit Perrot.
1388
13892002-03-04 Akim Demaille <akim@epita.fr>
1390
1391 Use bitset operations when possible, not loops over bits.
1392
1393 * src/conflicts.c (set_conflicts, count_sr_conflicts): Use
1394 bitset_or.
1395 * src/print.c (print_reductions): Use bitset_and, bitset_andn.
1396 * src/reduce.c (useless_nonterminals): Formatting changes.
1397 * src/warshall.c (TC): Use bitset_or.
1398
13992002-03-04 Akim Demaille <akim@epita.fr>
1400
1401 * src/lalr.h, src/lalr.c (tokensetsize): Remove, unused.
1402 * src/system.h (BITS_PER_WORD, WORDSIZE, SETBIT, RESETBIT, BITISSET):
1403 Ditto.
1404
14052002-03-04 Akim Demaille <akim@epita.fr>
1406
1407 * src/lalr.c (F): Now a bitset*.
1408 Adjust all dependencies.
1409
14102002-03-04 Akim Demaille <akim@epita.fr>
1411
1412 * src/conflicts.c (shiftset, lookaheadset): Now bitset.
1413 Adjust all dependencies.
1414
14152002-03-04 Akim Demaille <akim@epita.fr>
1416
1417 * src/L0.c, src/LR0.h (nstates): Be size_t.
1418 Adjust comparisons (signed vs unsigned).
1419 * src/conflics.c, src/lalr.c, src/lalr.h, src/output.c (LA): Now a
1420 bitset*.
1421 Adjust all dependencies.
1422
14232002-03-04 Akim Demaille <akim@epita.fr>
1424
1425 * src/closure.c (firsts): Now, also a bitset.
1426 Adjust all dependencies.
1427 (varsetsize): Remove, now unused.
1428 * src/warshall.h, src/warshall.c: Now work on arrays of bitsets.
1429
14302002-03-04 Akim Demaille <akim@epita.fr>
1431
1432 * src/print.c: Convert to use bitset.h, not hand coded iterations
1433 over ints.
1434
14352002-03-04 Akim Demaille <akim@epita.fr>
1436
1437 * src/reduce.c: Convert to use bitset.h, not hand coded BSet.
1438
14392002-03-04 Akim Demaille <akim@epita.fr>
1440
1441 * src/closure.c (ruleset): Be a bitset.
1442 (rulesetsize): Remove.
1443
14442002-03-04 Akim Demaille <akim@epita.fr>
1445
1446 * lib/bitset-int.h, lib/bitset.c, lib/bitset.h, lib/bitsetv.c,
1447 * lib/bitsetv.h, lib/ebitset.c, lib/ebitset.h, lib/lbitset.c,
1448 * lib/lbitset.h, lib/sbitset.c, lib/sbitset.h: New.
1449 * src/closure.c (fderives): Be an array of bitsets.
1450
14512002-02-28 Robert Anisko <robert@lrde.epita.fr>
1452
1453 * data/bison.c++: Merge the two generated headers. Insert a copyright
1454 notice in each output file.
1455
14562002-02-28 Akim Demaille <akim@epita.fr>
1457
1458 * data/bison.c++: Copy the prologue of bison.simple to fetch
1459 useful M4 definitions, such as b4_header_guard.
1460
14612002-02-25 Akim Demaille <akim@epita.fr>
1462
1463 * src/getargs.c (version): Give the name of the authors, and use a
1464 translator friendly scheme for the bgr
1465 copyright notice.
1466
14672002-02-25 Akim Demaille <akim@epita.fr>
1468
1469 * src/output.c (header_output): Remove, now handled completely via
1470 M4.
1471
14722002-02-25 Akim Demaille <akim@epita.fr>
1473
1474 * m4/m4.m4: New, from CVS Autoconf.
1475 * configure.in: Invoke it.
1476 * src/output.c (output_skeleton): Use its result instead of the
1477 hard coded name.
1478
14792002-02-25 Akim Demaille <akim@epita.fr>
1480
1481 * lib/tempname.c, lib/mkstemp.c, m4/mkstemp.m4: New, stolen from
1482 Fileutils 4.1.5.
1483 * configure.in: Invoke UTILS_FUNC_MKSTEMP.
1484 * src/output.c (output_skeleton): Use mkstemp to create a real
1485 temporary file.
1486 Move the filling of `skeleton' and its muscle to...
1487 (prepare): here.
1488 (output): Move the definition of the prologue muscle to...
1489 (prepare): here.
1490 * src/system.h (DEFAULT_TMPDIR): New.
1491
14922002-02-14 Paul Eggert <eggert@twinsun.com>
1493
1494 Remove the support for C++ namespace cleanliness; it was
1495 causing more problems than it was curing, since it didn't work
1496 properly on some nonstandard C++ compilers. This can wait
1497 for a proper C++ parser.
1498
1499 * NEWS: Document this.
1500 * doc/bison.texinfo (Bison Parser, Debugging): Remove special mention
1501 of C++, as it's treated like C now.
1502 * src/bison.simple (YYSTD): Remove.
1503 (YYSIZE_T, YYFPRINTF, YYPARSE_PARAM_ARG, YYPARSE_PARAM_DECL):
1504 Treat C++ just like Standard C instead of trying to support
1505 namespace cleanliness.
1506
15072002-02-14 Akim Demaille <akim@epita.fr>
1508
1509 * tests/regression.at (else): Adjust to Andreas' change.
1510
15112002-02-14 Akim Demaille <akim@epita.fr>
1512
1513 * lib/Makefile.am (EXTRA_DIST): Ship strnlen.c.
1514
15152002-02-13 Andreas Schwab <schwab@suse.de>
1516
1517 * src/output.c (output_rule_data): Don't output NULL, it might
1518 not be defined yet.
1519
15202002-02-11 Robert Anisko <robert@lrde.epita.fr>
1521
1522 * data/bison.c++ (YYDEBUG, YYERROR_VERBOSE): After the prologue.
1523 (Copyright notice): Update.
1524
15252002-02-11 Akim Demaille <akim@epita.fr>
1526
1527 * tests/regression.at (%nonassoc and eof): Don't include
1528 nonportable headers.
1529
15302002-02-08 Robert Anisko <robert@lrde.epita.fr>
1531
1532 * data/bison.c++: Correct error recovery. Make the user able to
1533 initialize the starting location.
1534
15352002-02-07 Akim Demaille <akim@epita.fr>
1536
1537 * tests/input.at: New.
1538
15392002-02-07 Robert Anisko <robert@lrde.epita.fr>
1540
1541 * data/bison.c++: Replace some direct m4 expansions by constants. Be
1542 more consistent when naming methods and variables. Put preprocessor
1543 directives around tables only needed for debugging.
1544
15452002-02-07 Robert Anisko <robert@lrde.epita.fr>
1546
1547 * data/bison.c++ (yy::b4_name::print_): New method, replaces yyprint in
1548 C++ parsers.
1549 (yy::b4_name::parse): Use print_.
1550
15512002-02-07 Robert Anisko <robert@lrde.epita.fr>
1552
1553 * data/bison.c++ (yy::b4_name::parse): Error recovery is back.
1554
15552002-02-07 Robert Anisko <robert@lrde.epita.fr>
1556
1557 * data/bison.c++ (yy::b4_name::error_): New method, replaces yyerror in
1558 C++ parsers.
1559 (yy::b4_name::parse): Build verbose error messages, and use error_.
1560
15612002-02-06 Robert Anisko <robert@lrde.epita.fr>
1562
1563 * data/bison.c++: Fix m4 quoting in comments.
1564
15652002-02-06 Robert Anisko <robert@lrde.epita.fr>
1566
1567 * data/bison.c++: Adjust the parser code. Fix some muscles that were
1568 not expanded by m4.
1569
15702002-02-05 Akim Demaille <akim@epita.fr>
1571
1572 * data/bison.c++: Adjust to the M4 back end.
1573 More is certainly needed.
1574
15752002-02-05 Akim Demaille <akim@epita.fr>
1576
1577 Give a try to M4 as a back end.
1578
1579 * lib/readpipe.c: New, from wdiff.
1580 * src/Makefile.am (DEFS): Define PKGDATADIR, not BISON_SIMPLE and
1581 BISON_HAIRY.
1582 * src/system.h (BISON_HAIRY, BISON_SIMPLE): Remove the DOS and VMS
1583 specific values. Now it is m4 that performs the lookup.
1584 * src/parse-skel.y: Remove.
1585 * src/muscle_tab.c, src/muscle_tab.h (muscles_m4_output): New.
1586 * src/output.c (actions_output, guards_output)
1587 (token_definitions_output): No longer keeps track of the output
1588 line number, hence remove the second argument.
1589 (guards_output): Check against the guard member of a rule, not the
1590 action member.
1591 Adjust callers.
1592 (output_skeleton): Don't look for the skeleton location, let m4 do
1593 that.
1594 Create `/tmp/muscles.m4'. This is temporary, a proper temporary
1595 file will be used.
1596 Invoke `m4' on m4sugar.m4, muscles.m4, and the skeleton.
1597 (prepare): Given that for the time being changesyntax is not
1598 usable in M4, rename the muscles using `-' to `_'.
1599 Define `defines_flag', `output_parser_name' and `output_header_name'.
1600 * src/output.h (actions_output, guards_output)
1601 (token_definitions_output): Adjust prototypes.
1602 * src/scan-skel.l: Instead of scanning the skeletons, it now
1603 processes the output of m4: `__oline__' and `#output'.
1604 * data/bison.simple: Adjust to be used by M4(sugar).
1605 * tests/Makefile.am: Use check_SCRIPTS to make sure `bison' is up
1606 to date.
1607 * tests/bison.in: Use the secrete envvar `BISON_PKGDATADIR'
1608 instead of the dead `BISON_SIMPLE' and `BISON_HAIRY'.
1609 * data/m4sugar/m4sugar.m4, data/m4sugar/version.m4: New,
1610 shamelessly stolen from CVS Autoconf.
1611
16122002-02-05 Akim Demaille <akim@epita.fr>
1613
1614 * lib/hash.c, lib/hash.h: Replace with Fileutils 4.1's version.
1615 * configure.in: Check for the declarations of free and malloc.
1616 * src/muscle_tab.c: Adjust.
1617
16182002-02-05 Akim Demaille <akim@epita.fr>
1619
1620 * src/muscle_tab.c (muscle_init): Don't default to NULL muscle
1621 which have no values.
1622
16232002-02-05 Akim Demaille <akim@epita.fr>
1624
1625 * src/bison.simple, src/bison.hairy, src/bison.c++: Move to...
1626 * data/: here.
1627
16282002-01-29 Paul Eggert <eggert@twinsun.com>
1629
1630 * src/bison.simple (YYSIZE_T): Do not define merely because
1631 YYSTACK_USE_ALLOCA is nonzero or alloca or _ALLOCA_H are defined.
1632 On some platforms, <alloca.h> does not declare YYSTD (size_t).
1633
16342002-01-27 Akim Demaille <akim@epita.fr>
1635
1636 Fix `%nonassoc and eof'.
1637
1638 * src/state.c (errs_dup): Aaaah! The failure was due to bytes
1639 which were not properly copied! Replace
1640 memcpy (res->errs, src->errs, src->nerrs);
1641 with
1642 memcpy (res->errs, src->errs, src->nerrs * sizeof (src->errs[0]));
1643 !!!
1644 * tests/regression.at (%nonassoc and eof): Adjust to newest
1645 Autotest: `.' is not in the PATH.
1646
16472002-01-27 Akim Demaille <akim@epita.fr>
1648
1649 * tests/sets.at (AT_EXTRACT_SETS): New.
1650 (Nullable): Use it.
1651 (Firsts): New.
1652
16532002-01-26 Akim Demaille <akim@epita.fr>
1654
1655 * tests/actions.at, tests/calc.at, tests/headers.at,
1656 * tests/torture.at: Adjust to the newest Autotest which no longer
1657 forces `.' in the PATH.
1658
16592002-01-25 Akim Demaille <akim@epita.fr>
1660
1661 * tests/regression.at (%nonassoc and eof): New.
1662 Suggested by Robert Anisko.
1663
16642002-01-24 Akim Demaille <akim@epita.fr>
1665
1666 Bison dumps core when trying to complain about broken input files.
1667 Reported by Cris van Pelt.
1668
1669 * src/lex.c (parse_percent_token): Be sure to set token_buffer.
1670 * tests/regression.at (Invalid input: 1, Invalid input: 2): Merge
1671 into...
1672 (Invalid inputs): Strengthen: exercise parse_percent_token.
1673
16742002-01-24 Robert Anisko <robert.anisko@epita.fr>
1675
1676 * src/Makefile.am: Add bison.c++.
1677 * src/bison.c++: New skeleton.
1678
16792002-01-21 Paolo Bonzini <bonzini@gnu.org>
1680
1681 * po/it.po: New.
1682
16832002-01-21 Kees Zeelenberg <kzlg@users.sourceforge.net>
1684
1685 * src/files.c (skeleton_find) [MSDOS]: Fix cp definition.
1686
16872002-01-20 Marc Autret <marc@gnu.org>
1688
1689 * src/files.c (compute_output_file_names): Fix
1690
16912002-01-20 Marc Autret <marc@gnu.org>
1692
1693 * tests/output.at: New test.
1694 * src/files.c (compute_base_names): Don't map extensions when
1695 the YACC flag is set, use defaults.
1696 Reported by Evgeny Stambulchik.
1697
16982002-01-20 Marc Autret <marc@gnu.org>
1699
1700 * src/system.h: Need to define __attribute__ away for non-GCC
1701 compilers as well (i.e. the vendor C compiler).
1702 Suggested by Albert Chin-A-Young.
1703
17042002-01-11 Tim Van Holder <tim.van.holder@pandora.be>
1705
1706 * lib/hash.h, lib/hash.c: Renamed __P to PARAMS and used the
1707 canonical definition.
1708 * src/system.h: Use the canonical definition for PARAMS (avoids
1709 a conflict with the macro from lib/hash.h).
1710
17112002-01-11 Akim Demaille <akim@epita.fr>
1712
1713 * configure.in: Use AC_FUNC_STRNLEN.
1714 Fixes the failures observed on AIX 4.3 by H.Merijn Brand.
1715
17162002-01-09 Akim Demaille <akim@epita.fr>
1717
1718 * src/files.c, src/files.h (output_infix): New.
1719 (tab_extension): Remove.
1720 (compute_base_names): Compute the former, drop the latter.
1721 * src/output.c (prepare): Insert the muscles `output-infix', and
1722 `output-suffix'.
1723 * src/parse-skel.y (string, string.1): New.
1724 (section.header): Use it.
1725 (section.yacc): Remove.
1726 (prefix): Remove too.
1727 * src/scan-skel.l: Adjust.
1728 * src/bison.simple, src/bison.hairy: Adjust.
1729
17302002-01-09 Akim Demaille <akim@epita.fr>
1731
1732 * configure.in (WERROR_CFLAGS): Compute it.
1733 * src/Makefile.am (CFLAGS): Pass it.
1734 * tests/atlocal.in (CFLAGS): Idem.
1735 * src/files.c: Fix a few warnings.
1736 (get_extension_index): Remove, unused.
1737
17382002-01-08 Akim Demaille <akim@epita.fr>
1739
1740 * src/getargs.c (AS_FILE_NAME): New.
1741 (getargs): Use it to convert DOSish file names.
1742 * src/files.c (base_name): Rename as full_base_name to avoid
1743 clashes with `base_name ()'.
1744 (filename_split): New.
1745 (compute_base_names): N-th rewrite, using filename_split.
1746
17472002-01-08 Akim Demaille <akim@epita.fr>
1748
1749 * lib/basename.c, lib/dirname.h, lib/dirname.c, lib/memrchr.c:
1750 New, stolen from the Fileutils 4.1.
1751 * lib/Makefile.am (libbison_a_SOURCES): Adjust.
1752 * configure.in: Check for the presence of memrchr, and of its
1753 prototype.
1754
17552002-01-07 Tim Van Holder <tim.van.holder@pandora.be>
1756
1757 * lib/hash.h (__P): Added definition for this macro.
1758 * src/Makefile.am: Add parse-skel.c and scan-skel.c to
1759 BUILT_SOURCES, to ensure they are generated first.
1760 * src/parse-skel.y: Use YYERROR_VERBOSE instead of
1761 %error-verbose to allow bootstrapping with bison 1.30x.
1762
17632002-01-06 Akim Demaille <akim@epita.fr>
1764
1765 * src/reader.c (parse_braces): Don't fetch the next char, the
1766 convention is to fetch on entry.
1767 * tests/torture.at (GNU Cim Grammar): Reintroduce their weird
1768 'switch' without a following semicolon.
1769 * tests/regression.at (braces parsing): New.
1770
17712002-01-06 Akim Demaille <akim@epita.fr>
1772
1773 Bison is dead wrong in its RR conflict reports.
1774
1775 * tests/torture.at (GNU Cim Grammar): New.
1776 * src/conflicts.c (count_rr_conflicts): Fix.
1777
17782002-01-06 Akim Demaille <akim@epita.fr>
1779
1780 Creating package.m4 from configure.ac causes too many problems.
1781
1782 * tests/Makefile.am (package.m4): Create it by hand,
1783 AC_CONFIG_TESTDIR no longer does in the most recent CVS Autoconf.
1784
17852002-01-06 Akim Demaille <akim@epita.fr>
1786
1787 * src/Makefile.am (bison_SOURCES): Add parse-skel.h and
1788 skeleton.h.
1789
17902002-01-04 Paul Eggert <eggert@twinsun.com>
1791
1792 * doc/bison.texinfo (Debugging):
1793 Remove YYSTDERR; it's no longer defined or used.
1794 Also, s/cstdio.h/cstdio/.
1795
17962002-01-03 Akim Demaille <akim@epita.fr>
1797
1798 * tests/bison.in, tests/atlocal.in: Adjust to CVS Autoconf.
1799
18002002-01-03 Akim Demaille <akim@epita.fr>
1801
1802 * src/parse-skel.y (process_skeleton): Don't bind the parser's
1803 tracing code to --trace, wait for a better --trace option, with
1804 args.
1805
18062002-01-03 Akim Demaille <akim@epita.fr>
1807
1808 * src/bison.simple (YYSTDERR): Remove, replace `stderr'.
1809 The ISO C++ standard is extremely clear about it: stderr is
1810 considered a macro, not a regular symbol (see table 94 `Header
1811 <cstdio> synopsis', [lib.c.files] 27.8.2 C Library files).
1812 Therefore std:: does not apply to it. It still does with fprintf.
1813 Also, s/cstdio.h/cstdio/.
1814
18152002-01-03 Akim Demaille <akim@epita.fr>
1816
1817 * lib/quotearg.c: Use `#include "..."' instead of `#include <...>'
1818 for non system headers.
1819
18202002-01-02 Akim Demaille <akim@epita.fr>
1821
1822 Equip the skeleton chain with location tracking, runtime trace,
1823 pure parser and scanner.
1824
1825 * src/parse-skel.y: Request a pure parser, locations, and prefix
1826 renaming.
1827 (%union): Having several members with the same type does not help
1828 type mismatches, simplify.
1829 (YYPRINT, yyprint): New.
1830 (yyerror): ``Rename'' (there is a #define yyerror skel_error) as...
1831 (skel_error): this.
1832 Handle locations.
1833 * src/scan-skel.l: Adjust to these changes.
1834 * src/skeleton.h (LOCATION_RESET, LOCATION_LINES, LOCATION_STEP)
1835 (LOCATION_PRINT, skel_control_t): New.
1836
18372001-12-30 Akim Demaille <akim@epita.fr>
1838
1839 * src/parse-skel.y: Get rid of the shift/reduce conflict:
1840 replace `gb' with BLANKS.
1841 * src/scan-skel.l: Adjust.
1842
18432001-12-30 Akim Demaille <akim@epita.fr>
1844
1845 * src/system.h: We don't need nor want bcopy.
1846 Throw away MS-DOS crap: we don't need getpid.
1847 * configure.in: We don't need strndup. It was even causing
1848 problems: because Flex includes the headers *before* us,
1849 _GNU_SOURCE is not defined by config.h, and therefore strndup was
1850 not visible.
1851 * lib/xstrndup.c: New.
1852 * src/scan-skel.l: Use it.
1853 Be sure to initialize yylval.muscle member when scanning a MUSCLE.
1854 * src/parse-skel.y: Use %directives instead of #defines.
1855
18562001-12-30 Akim Demaille <akim@epita.fr>
1857
1858 * src/skeleton.h: New.
1859 * src/output.c (output_parser, output_master_parser): Remove, dead
1860 code.
1861 * src/output.h (get_lines_number, actions_output, guards_output)
1862 (token_definitions_output): Prototype them.
1863 * src/parse-skel.y: Add the license notice.
1864 Include output.h and skeleton.h.
1865 (process_skeleton): Returns void, and takes a single parameter.
1866 * src/scan-skel.l: Add the license notice.
1867 Include skeleton.h.
1868 Don't use %option yylineno: it seems that then Flex imagines
1869 REJECT has been used, and therefore it won't reallocate its
1870 buffers (which makes no other sense to me than a bug). It results
1871 in warnings for `unused: yy_flex_realloc'.
1872
18732001-12-30 Robert Anisko <robert.anisko@epita.fr>
1874
1875 * src/muscle_tab.h (MUSCLE_INSERT_INT, MUSCLE_INSERT_STRING)
1876 (MUSCLE_INSERT_PREFIX): ...to there.
1877 * src/output.c (MUSCLE_INSERT_INT, MUSCLE_INSERT_STRING)
1878 (MUSCLE_INSERT_PREFIX): Move from here...
1879
1880 * src/bison.hairy: Add a section directive. Put braces around muscle
1881 names. This parser skeleton is still broken, but Bison should not
1882 choke on a bad muscle 'syntax'.
1883 * src/bison.simple: Add a section directive. Put braces around muscle
1884 names.
1885
1886 * src/files.h (strsuffix, stringappend): Add declarations.
1887 (tab_extension): Add declaration.
1888 (short_base_name): Add declaration.
1889
1890 * src/files.c (strsuffix, stringappend): No longer static. These
1891 functions are used in the skeleton parser.
1892 (tab_extension): New.
1893 (compute_base_names): Use the computations done in this function
1894 to guess if the generated parsers should have '.tab' in their
1895 names.
1896 (short_base_name): No longer static.
1897
1898 * src/output.c (output_skeleton): New.
1899 (output): Disable call to output_master_parser, and give a try to
1900 a new skeleton handling system.
1901 (guards_output, actions_output): No longer static.
1902 (token_definitions_output, get_lines_number): No longer static.
1903
1904 * configure.in: Use AM_PROG_LEX and AC_PROG_YACC.
1905
1906 * src/Makefile.am (bison_SOURCES): Add scan-skel.l and
1907 parse-skel.y.
1908
1909 * src/parse-skel.y: New file.
1910 * src/scan-skel.l: New file.
1911
19122001-12-29 Akim Demaille <akim@epita.fr>
1913
1914 %name-prefix is broken.
1915
1916 * src/files.c (spec_name_prefix): Initialize to NULL, not to "yy".
1917 Adjust all dependencies.
1918 * tests/headers.at (export YYLTYPE): Strengthen this test: use
1919 %name-prefix.
1920
1921 Renaming yylval but not yylloc is not consistent. Now we do.
1922
1923 * src/bison.simple: Prefix yylloc if used.
1924 * doc/bison.texinfo (Decl Summary): Document that.
1925
19262001-12-29 Akim Demaille <akim@epita.fr>
1927
1928 * doc/bison.texinfo: Promote `%long-directive' over
1929 `%long_directive'.
1930 Remove all references to fixed-output-files, yacc is enough.
1931
19322001-12-29 Akim Demaille <akim@epita.fr>
1933
1934 * src/bison.simple: Define YYDEBUG and YYERROR_VERBOSE *after* the
1935 user prologue. These are defaults.
1936 * tests/actions.at (Mid-rule actions): Make sure the user can
1937 define YYDEBUG and YYERROR_VERBOSE.
1938
19392001-12-29 Akim Demaille <akim@epita.fr>
1940
1941 * src/output.c (header_output): Don't forget to export YYLTYPE and
1942 yylloc.
1943 * tests/headers.at (export YYLTYPE): New, make sure it does.
1944 * tests/regression.at (%union and --defines, Invalid CPP headers):
1945 Move to...
1946 * tests/headers.at: here.
1947
19482001-12-29 Akim Demaille <akim@epita.fr>
1949
1950 * src/gram.h (rule_s): Member `assoc' is of type `associativity'.
1951
19522001-12-29 Akim Demaille <akim@epita.fr>
1953
1954 * tests/actions.at (Mid-rule actions): Output on a single line
1955 instead of several.
1956
19572001-12-29 Akim Demaille <akim@epita.fr>
1958
1959 * doc/bison.texinfo: Formatting changes.
1960
19612001-12-29 Akim Demaille <akim@epita.fr>
1962
1963 Don't store the token defs in a muscle, just be ready to output it
1964 on command. Now possible via `symbols'. Fixes a memory leak.
1965
1966 * src/output.c (token_definitions_output): New.
1967 (output_parser, header_output): Use it.
1968 * src/reader.c (symbols_save): Remove.
1969
19702001-12-29 Akim Demaille <akim@epita.fr>
1971
1972 * src/bison.simple: Do not provide a default for YYSTYPE and
1973 YYLTYPE before the user's prologue. Otherwise it's hardly... a
1974 default.
1975
19762001-12-29 Akim Demaille <akim@epita.fr>
1977
1978 Mid-rule actions are simply... ignored!
1979
1980 * src/reader.c (readgram): Be sure to attach mid-rule actions to
1981 the empty-rule associated to the dummy symbol, not to the host
1982 rule.
1983 * tests/actions.at (Mid-rule actions): New.
1984
19852001-12-29 Akim Demaille <akim@epita.fr>
1986
1987 Memory leak.
1988
1989 * src/reader.c (reader): Free grammar.
1990
19912001-12-29 Akim Demaille <akim@epita.fr>
1992
1993 Memory leak.
1994
1995 * src/LR0.c (new_itemsets): Don't allocate `shift_symbol' here,
1996 since it allocates it for each state, although only one is needed.
1997 (allocate_storage): Do it here.
1998
19992001-12-29 Akim Demaille <akim@epita.fr>
2000
2001 * src/options.h, src/options.c (create_long_option_table): Rename
2002 as...
2003 (long_option_table_new): this, with a clearer prototype.
2004 (percent_table): Remove, unused,
2005 * src/getargs.c (getargs): Adjust.
2006
20072001-12-29 Akim Demaille <akim@epita.fr>
2008
2009 * src/LR0.c, src/conflicts.c, src/lalr.c, src/lalr.h, src/output.c
2010 * src/print.c, src/print_graph.c, src/state.h: Rename state_table
2011 as states.
2012
20132001-12-29 Akim Demaille <akim@epita.fr>
2014
2015 * src/lalr.c (build_relations): Rename `states' as `states1'.
2016 Sorry, I don't understand exactly what it is, no better name...
2017
20182001-12-29 Akim Demaille <akim@epita.fr>
2019
2020 * src/closure.c, src/conflicts.c, src/derives.c, src/gram.c
2021 * src/gram.h, src/lalr.c, src/nullable.c, src/output.c, src/print.c
2022 * src/print_graph.c, src/reader.c, src/reduce.c: Rename rule_table
2023 as rules.
2024
20252001-12-29 Akim Demaille <akim@epita.fr>
2026
2027 * src/gram.c (rprec, rprecsym, rassoc): Remove, unused since long
2028 ago.
2029
20302001-12-29 Akim Demaille <akim@epita.fr>
2031
2032 * src/reader.c, src/reader.h (user_toknums): Remove.
2033 Adjust all users to use symbols[i]->user_token_number.
2034
20352001-12-29 Akim Demaille <akim@epita.fr>
2036
2037 * src/gram.c, src/gram.h (sprec, sassoc): Remove.
2038 Adjust all users to use symbols[i]->prec or ->assoc.
2039
20402001-12-29 Akim Demaille <akim@epita.fr>
2041
2042 * src/reader.c, src/reader.h (tags): Remove.
2043 Adjust all users to use symbols[i]->tag.
2044
20452001-12-29 Akim Demaille <akim@epita.fr>
2046
2047 * src/gram.h, src/gram.c (symbols): New, similar to state_table
2048 and rule_table.
2049 * src/reader.c (packsymbols): Fill this table.
2050 Drop sprec.
2051 * src/conflicts.c (resolve_sr_conflict): Adjust.
2052 * src/reduce.c (reduce_grammar): Adjust: just sort symbols, a
2053 single table.
2054 Use symbols[i]->tag instead of tags[i].
2055
20562001-12-29 Akim Demaille <akim@epita.fr>
2057
2058 * tests/calc.at (_AT_DATA_CALC_Y): Also use %union.
2059 In addition, put a comment in there, to replace...
2060 * tests/regression.at (%union and C comments): Remove.
2061
20622001-12-29 Akim Demaille <akim@epita.fr>
2063
2064 * tests/regression.at (Web2c Actions): Blindly move the actual
2065 output as expected output. The contents *seem* right to me, but I
2066 can't pretend reading perfectly parser tables... Nonetheless, all
2067 the other tests pass correctly, the table look OK, even though the
2068 presence of `$axiom' is to be noted: AFAICS it is useless (but
2069 harmless).
2070
20712001-12-29 Akim Demaille <akim@epita.fr>
2072
2073 * src/reader.c (readgram): Don't add the rule 0 if there were no
2074 rules read. In other words, add it _after_ having performed
2075 grammar sanity checks.
2076 Fixes the `tests/regression.at (Invalid input: 1)' Failure.
2077
20782001-12-29 Akim Demaille <akim@epita.fr>
2079
2080 * tests/regression.at (Web2c Report): Catch up: the rule 0 is now
2081 visible, and some states have now a different number.
2082
20832001-12-29 Akim Demaille <akim@epita.fr>
2084
2085 * src/reader.c (readgram): Bind the initial rule's lineno to that
2086 of the first rule.
2087 * tests/regression.at (Rule Line Numbers, Unresolved SR Conflicts):
2088 (Solved SR Conflicts): Adjust rule 0's line number.
2089
20902001-12-29 Akim Demaille <akim@epita.fr>
2091
2092 Fix the `GAWK Grammar' failure.
2093
2094 * src/LR0.c (final_state): Initialize to -1 so that we do compute
2095 the reductions of the first state which was mistakenly confused
2096 with the final state because precisely final_state was initialized
2097 to 0.
2098 * tests/sets.at (Nullable): Adjust: state 0 does have lookaheads,
2099 now noticed by Bison.
2100 * tests/regression.at (Rule Line Numbers): Adjust: state 0 does
2101 have a reduction on $default.
2102
21032001-12-29 Akim Demaille <akim@epita.fr>
2104
2105 * src/gram.c (ritem_print): Be sure to subtract 1 when displaying
2106 rule line numbers.
2107 * src/closure.c (print_closure): Likewise.
2108 * src/derives.c (print_derives): Likewise.
2109 * tests/sets.at (Nullable): Adjust: the rule numbers are correct
2110 now.
2111
21122001-12-29 Akim Demaille <akim@epita.fr>
2113
2114 * src/lalr.c (lookaheads_print): New.
2115 (lalr): Call it when --trace-flag.
2116 * tests/sets.at (Nullable): Adjust: when tracing, the lookaheads
2117 are dumped.
2118
21192001-12-29 Akim Demaille <akim@epita.fr>
2120
2121 * src/derives.c (print_derives): Be sure to use `>= 0', not `> 0',
2122 when walking through ritem, even via rule->rhs.
2123 * src/reduce.c (dump_grammar, useful_production, reduce_output)
2124 (useful_production, useless_nonterminals): Likewise.
2125 (reduce_grammar_tables): Likewise, plus update nritems.
2126 * src/nullable.c (set_nullable): Likewise.
2127 * src/lalr.c (build_relations): Likewise.
2128 * tests/sets.at (Nullable): Adjust.
2129 Fortunately, now, the $axiom is no longer nullable.
2130
21312001-12-29 Akim Demaille <akim@epita.fr>
2132
2133 * src/LR0.c (generate_states): Use nritems, not nitems, nor using
2134 the 0-sentinel.
2135 * src/gram.c (ritem_longest_rhs): Likewise.
2136 * src/reduce.c (nonterminals_reduce): Likewise.
2137 * src/print_graph.c (print_graph): Likewise.
2138 * src/output.c (output_rule_data): Likewise.
2139 * src/nullable.c (set_nullable): Likewise.
2140
21412001-12-29 Akim Demaille <akim@epita.fr>
2142
2143 * src/output.c: Comment changes.
2144
21452001-12-27 Paul Eggert <eggert@twinsun.com>
2146
2147 * src/bison.simple (YYSTACK_ALLOC, YYSIZE_T): Remove special
2148 cases for non-GNU systems like AIX, HP-UX, SGI, Sun, and
2149 Sparc, as they were causing more porting problems than the
2150 (minor) performance improvement was worth.
2151
2152 Also, catch up with 1.31's YYSTD.
2153
21542001-12-27 Akim Demaille <akim@epita.fr>
2155
2156 * src/output.c (output_gram): Rely on nritems, not the
2157 0-sentinel. See below.
2158 Use -1 as separator, not 0.
2159 * src/bison.simple (yyparse): Subtract 1 to the rule numbers.
2160 Rely on -1 as separator in yyrhs, instead of 0.
2161 * tests/calc.at (AT_CHECK_CALC): Now, the parsers no longer issue
2162 twice `Now at end of input', therefore there are two lines less to
2163 expect.
2164
21652001-12-27 Akim Demaille <akim@epita.fr>
2166
2167 * tests/regression.at (Unresolved SR Conflicts):
2168 (Solved SR Conflicts, Rule Line Numbers): Adjust to the changes
2169 below.
2170
21712001-12-27 Akim Demaille <akim@epita.fr>
2172
2173 * src/LR0.c (new_state): Recognize the final state by the fact it
2174 is reached by eoftoken.
2175 (insert_start_shifting_state, insert_eof_shifting_state)
2176 (insert_accepting_state, augment_automaton): Remove, since now
2177 these states are automatically computed from the initial state.
2178 (generate_states): Adjust.
2179 * src/print.c: When reporting a rule number to the user, substract
2180 1, so that the axiom rule is rule 0, and the first user rule is 1.
2181 * src/reduce.c: Likewise.
2182 * src/print_graph.c (print_core): For the time being, just as for
2183 the report, depend upon --trace-flags to dump the full set of
2184 items.
2185 * src/reader.c (readgram): Once the grammar read, insert the rule
2186 0: `$axiom: START-SYMBOL $'.
2187 * tests/set.at: Adjust: rule 0 is now displayed, and since the
2188 number of the states has changed (the final state is no longer
2189 necessarily the last), catch up.
2190
21912001-12-27 Akim Demaille <akim@epita.fr>
2192
2193 Try to make the use of the eoftoken valid. Given that its value
2194 is 0 which was also used as a sentinel in ritem, (i) make sure >= 0
2195 is used instead of > 0 where appropriate, (ii), depend upon nritems
2196 instead of the 0-sentinel.
2197
2198 * src/gram.h, src/gram.c (nritems): New.
2199 Expected to be duplication of nitems, but for the time being...
2200 * src/reader.c (packgram): Assert nritems and nitems are equal.
2201 * src/LR0.c (allocate_itemsets, new_itemsets): Adjust.
2202 * src/closure.c (print_closure, print_fderives): Likewise.
2203 * src/gram.c (ritem_print): Likewise.
2204 * src/print.c (print_core, print_grammar): Likewise.
2205 * src/print_graph.c: Likewise.
2206
22072001-12-27 Akim Demaille <akim@epita.fr>
2208
2209 * src/main.c (main): If there are complains after grammar
2210 reductions, then output the report anyway if requested, then die.
2211 * src/symtab.c (bucket_new): Initialize `value' to -1, not 0.
2212 * src/reader.c (eoftoken): New.
2213 (parse_token_decl): If the token being defined has value `0', it
2214 is the eoftoken.
2215 (packsymbols): No longer hack `tags' to insert `$' by hand.
2216 Be sure to preserve the value of the eoftoken.
2217 (reader): Make sure eoftoken is defined.
2218 Initialize nsyms to 0: now eoftoken is created just like the others.
2219 * src/print.c (print_grammar): Don't special case the eof token.
2220 * src/regression.at: Adjust: `$' has value 0, not -1, which was a
2221 lie anyway, albeit pleasant.
2222 * tests/calc.at: Exercise error messages with eoftoken.
2223 Change the grammar so that empty input is invalid.
2224 Adjust expectations.
2225 When yyungeting, be sure to use a valid yylloc: use last_yylloc.
2226
22272001-12-27 Akim Demaille <akim@epita.fr>
2228
2229 * configure.in: Check the protos of strchr ans strspn.
2230 Replace strchr if needed.
2231 * src/system.h: Provide the protos of strchr, strspn and memchr if
2232 missing.
2233 * lib/strchr.c: New.
2234 * src/reader.c (symbols_save): Use strchr.
2235
22362001-12-27 Akim Demaille <akim@epita.fr>
2237
2238 * src/print.c, src/print_graph.c (escape): New.
2239 Use it to quote the TAGS outputs.
2240 * src/print_graph.c (print_state): Now errors are in red, and
2241 reductions in green.
2242 Prefer high to wide: output the state number on a line of its own.
2243
22442001-12-27 Akim Demaille <akim@epita.fr>
2245
2246 * src/state.h, src/state.c (reductions_new): New.
2247 * src/LR0.c (set_state_table): Let all the states have a
2248 `reductions', even if reduced to 0.
2249 (save_reductions): Adjust.
2250 * src/lalr.c (initialize_LA, initialize_lookaheads): Adjust.
2251 * src/print.c (print_reductions, print_actions): Adjust.
2252 * src/output.c (action_row): Adjust.
2253
22542001-12-27 Akim Demaille <akim@epita.fr>
2255
2256 * src/state.h, src/state.c (errs_new, errs_dup): New.
2257 * src/LR0.c (set_state_table): Let all the states have an errs,
2258 even if reduced to 0.
2259 * src/print.c (print_errs, print_reductions): Adjust.
2260 * src/output.c (output_actions, action_row): Adjust.
2261 * src/conflicts.c (resolve_sr_conflict): Adjust.
2262
22632001-12-27 Akim Demaille <akim@epita.fr>
2264
2265 * src/lalr.c (set_goto_map, initialize_F): Use SHIFT_SYMBOL.
2266
22672001-12-27 Akim Demaille <akim@epita.fr>
2268
2269 * src/conflicts.c, src/conflicts.h (print_reductions): Move to...
2270 * src/print.c: here.
2271 (lookaheadset, shiftset): New, used as additional storage by
2272 print_reductions.
2273 (print_results): Adjust.
2274 (print_shifts, print_gotos, print_errs): New, extracted from...
2275 (print_actions): here.
2276 * src/print_graph.c (print_actions): Remove dead code.
2277
22782001-12-27 Akim Demaille <akim@epita.fr>
2279
2280 * src/reader.c (copy_dollar, copy_at): Better checking of `n' in
2281 `$n' and `@n'.
2282
22832001-12-27 Akim Demaille <akim@epita.fr>
2284
2285 * src/lalr.c (add_lookback_edge): Use state_t instead of ints.
2286 (build_relations): Adjust.
2287
22882001-12-27 Akim Demaille <akim@epita.fr>
2289
2290 * src/lalr.c (set_goto_map): Remove a wrong but benign loop
2291 duplication.
2292
22932001-12-27 Akim Demaille <akim@epita.fr>
2294
2295 * src/reader.c (packgram): Catch nitems overflows.
2296
22972001-12-27 Akim Demaille <akim@epita.fr>
2298
2299 * src/files.c, src/files.h (guard_obstack): Remove.
2300 * src/output.c (output): Adjust.
2301 * src/reader.c (parse_braces): New, factoring...
2302 (copy_action, copy_guard): these two which are renamed as...
2303 (parse_action, parse_guard): these.
2304 As a voluntary consequence, using braces around guards is now
2305 mandatory.
2306
23072001-12-27 Akim Demaille <akim@epita.fr>
2308
2309 * src/gram.h (rule_t): `guard' and `guard_line' are new members.
2310 * src/reader.c (symbol_list): `guard' and `guard_line' are new
2311 members.
2312 (symbol_list_new): Adjust.
2313 (copy_action): action_line is the first line, not the last.
2314 (copy_guard): Just as for actions, store the `action' only, not
2315 the switch/case/break flesh.
2316 Don't parse the user action that might follow the guard, let...
2317 (readgram): do it, i.e., now, there can be an action after a
2318 guard.
2319 In other words the guard is just explicitly optional.
2320 (packgram): Adjust.
2321 * src/output.c (guards_output): New.
2322 (output_parser): Call it when needed.
2323 (output): Also free the guard and attrs obstacks.
2324 * src/files.c, src/files.h (obstack_save): Remove.
2325 (output_files): Remove.
2326 As a result, if one needs the former `.act' file, using an
2327 appropriate skeleton which requires actions and guards is now
2328 required.
2329 * src/main.c (main): Adjust.
2330 * tests/semantic.at: New.
2331 * tests/regression.at: Use `input.y' as input file name.
2332 Avoid 8+3 problems by requiring input.c when the test needs the
2333 parser.
2334
23352001-12-27 Akim Demaille <akim@epita.fr>
2336
2337 * src/reader.c (symbol_list_new): Be sure to initialize all the
2338 fields.
2339
23402001-12-27 Akim Demaille <akim@epita.fr>
2341
2342 All the hacks using a final pseudo state are now useless.
2343
2344 * src/LR0.c (set_state_table): state_table holds exactly nstates.
2345 * src/lalr.c (nLA): New.
2346 (initialize_LA, compute_lookaheads, initialize_lookaheads): Use it
2347 instead of lookaheadsp from the pseudo state (nstate + 1).
2348
23492001-12-27 Akim Demaille <akim@epita.fr>
2350
2351 * src/output.c (action_row, token_actions): Use a state_t instead
2352 of a integer, and nlookaheads instead of the following state's
2353 lookaheadsp.
2354
23552001-12-27 Akim Demaille <akim@epita.fr>
2356
2357 * src/conflicts.c (log_resolution, flush_shift)
2358 (resolve_sr_conflict, set_conflicts, solve_conflicts)
2359 (count_sr_conflicts, count_rr_conflicts, conflicts_output)
2360 (conflicts_print, print_reductions): Use a state_t instead of an
2361 integer when referring to a state.
2362 As much as possible, depend upon nlookaheads, instead of the
2363 `lookaheadsp' member of the following state (since lookaheads of
2364 successive states are successive, the difference between state n + 1
2365 and n served as the number of lookaheads for state n).
2366 * src/lalr.c (add_lookback_edge): Likewise.
2367 * src/print.c (print_core, print_actions, print_state)
2368 (print_results): Likewise.
2369 * src/print_graph.c (print_core, print_actions, print_state)
2370 (print_graph): Likewise.
2371 * src/conflicts.h: Adjust.
2372
23732001-12-27 Akim Demaille <akim@epita.fr>
2374
2375 * src/bison.hairy: Formatting/comment changes.
2376 ANSIfy.
2377 Remove `register' indications.
2378 Add plenty of `static'.
2379
23802001-12-27 Akim Demaille <akim@epita.fr>
2381
2382 * src/output.c (prepare): Drop the muscle `ntbase' which
2383 duplicates ntokens.
2384 * src/bison.simple: Formatting/comment changes.
2385 Use YYNTOKENS only, which is documented, but not YYNTBASE, which
2386 is an undocumented synonym.
2387
23882001-12-22 Akim Demaille <akim@epita.fr>
2389
2390 * src/output.c (output_table_data): Change the prototype to use
2391 `int' for array ranges: some invocations do pass an int, not a
2392 short.
2393 Reported by Wayne Green.
2394
23952001-12-22 Akim Demaille <akim@epita.fr>
2396
2397 Some actions of web2c.y are improperly triggered.
2398 Reported by Mike Castle.
2399
2400 * src/lalr.c (traverse): s/F (i)[k] = F (j)[k]/F (j)[k] = F (i)[k]/.
2401 * tests/regression.at (Web2c): Rename as...
2402 (Web2c Report): this.
2403 (Web2c Actions): New.
2404
24052001-12-22 Akim Demaille <akim@epita.fr>
2406
2407 Reductions in web2c.y are improperly reported.
2408 Reported by Mike Castle.
2409
2410 * src/conflicts.c (print_reductions): Fix.
2411 * tests/regression.at (Web2c): New.
2412
24132001-12-18 Akim Demaille <akim@epita.fr>
2414
2415 Some host fail on `assert (!"foo")', which expands to
2416 ((!"foo") ? (void)0 : __assert("!"foo."", __FILE__, __LINE__))
2417 Reported by Nelson Beebee.
2418
2419 * src/output.c, src/vcg.c: Replace `assert (!"it succeeded")' with
2420 `#define it_succeeded 0' and `assert (it_succeeded)'.
2421
24222001-12-17 Marc Autret <autret_m@epita.fr>
2423
2424 * src/bison.simple: Don't hard code the skeleton line and filename.
2425 * src/output.c (output_parser): Rename 'line' as 'output_line'.
2426 New line counter 'skeleton_line' (skeleton-line muscle).
2427
24282001-12-17 Paul Eggert <eggert@twinsun.com>
2429
2430 * NEWS, doc/bison.texinfo, doc/bison.1, doc/bison.rnh: Document that
2431 YYDEBUG must be defined to a nonzero value.
2432
2433 * src/bison.simple (yytname): Do not assume that the user defines
2434 YYDEBUG to a properly parenthesized expression.
2435
24362001-12-17 Akim Demaille <akim@epita.fr>
2437
2438 * src/state.h (state_t): Rename lookaheads as lookaheadsp.
2439 nlookaheads is a new member.
2440 Adjust all users.
2441 * src/lalr.h (nlookaheads): Remove this orphan declaration.
2442 * src/lalr.c (initialize_lookaheads): Set nlookaheads for each
2443 state.
2444
24452001-12-17 Akim Demaille <akim@epita.fr>
2446
2447 * src/files.h, src/files.c (open_files, close_files): Remove.
2448 * src/main.c (main): Don't open/close files, nor invoke lex_free,
2449 let...
2450 * src/reader.c (reader): Do it.
2451
24522001-12-17 Akim Demaille <akim@epita.fr>
2453
2454 * src/conflicts.c (print_reductions): Formatting changes.
2455
24562001-12-17 Akim Demaille <akim@epita.fr>
2457
2458 * src/conflicts.c (flush_shift): Also adjust lookaheadset.
2459 (flush_reduce): New.
2460 (resolve_sr_conflict): Adjust.
2461
24622001-12-17 Akim Demaille <akim@epita.fr>
2463
2464 * src/output.c (output_obstack): Be static and rename as...
2465 (format_obstack): this, to avoid any confusion with files.c's
2466 output_obstack.
2467 * src/reader.h (muscle_obstack): Move to...
2468 * src/output.h: here, since it's defined in output.c.
2469
24702001-12-17 Akim Demaille <akim@epita.fr>
2471
2472 * src/output.c (action_row, save_column, default_goto)
2473 (sort_actions, matching_state, pack_vector): Better variable
2474 locality.
2475
24762001-12-17 Akim Demaille <akim@epita.fr>
2477
2478 * src/output.c: Various formatting changes.
2479
24802001-12-17 Akim Demaille <akim@epita.fr>
2481
2482 * src/files.c (output_files): Free the output_obstack.
2483 * src/main.c (main): Call print and print_graph conditionally.
2484 * src/print.c (print): Work unconditionally.
2485 * src/print_graph.c (print_graph): Work unconditionally.
2486 * src/conflicts.c (log_resolution): Output only if verbose_flag.
2487
24882001-12-16 Marc Autret <autret_m@epita.fr>
2489
2490 * src/output.c (actions_output): Fix. When we use %no-lines,
2491 there is one less line per action.
2492
24932001-12-16 Marc Autret <autret_m@epita.fr>
2494
2495 * src/bison.simple: Remove a useless #line directive.
2496 s/#line %%line %%skeleton/#line %%line "%%parser-file-name"/'.
2497 * src/output.c (get_lines_number): New.
2498 (output_parser): Adjust, now takes care about the lines of a
2499 output muscles.
2500 Fix line numbering.
2501 (actions_output): Computes the number of lines taken by actions.
2502 (output_master_parser): Insert new skeleton which is the name of
2503 the output parser file name.
2504
25052001-12-15 Marc Autret <autret_m@epita.fr>
2506
2507 * src/bison.simple [YYERROR_VERBOSE]: Restore backward compatibility.
2508
25092001-12-15 Marc Autret <autret_m@epita.fr>
2510
2511 * src/output.c (output_gram): Keep track of the hairy one.
2512
25132001-12-15 Akim Demaille <akim@epita.fr>
2514
2515 Make `make distcheck' work.
2516
2517 * lib/Makefile.am (INCLUDES): Add top_srcdir/intl, since hash uses
2518 system.h which uses libgettext.h.
2519
25202001-12-15 Akim Demaille <akim@epita.fr>
2521
2522 * src/nullable.c (set_nullable): Useless rules must be skipped,
2523 otherwise, since we range over their symbols, we might look at a
2524 nonterminal which no longer ``exists'', i.e., it is not counted in
2525 `nvars', hence we overflow our arrays.
2526
25272001-12-15 Akim Demaille <akim@epita.fr>
2528
2529 The header can also be produced directly, without any obstack!
2530 Yahoo!
2531
2532 * src/files.c, src/files.h (defines_obstack): Remove.
2533 (compute_header_macro): Global.
2534 (defines_obstack_save): Remove.
2535 * src/reader.c (parse_union_decl): No longer output to
2536 defines_obstack: its content can be found in the `stype' muscle
2537 anyway.
2538 (output_token_translations): Merge into...
2539 (symbols_output): this.
2540 Rename as...
2541 (symbols_save): this.
2542 (reader): Adjust.
2543 * src/output.c (header_output): New.
2544 (output): Call it.
2545
25462001-12-15 Akim Demaille <akim@epita.fr>
2547
2548 * src/reader.c (parse_union_decl): Instead of handling two obstack
2549 simultaneously, use one to define the `stype' muscle, and use the
2550 value of the latter to fill defines_obstack.
2551 (copy_comment): Remove.
2552 (copy_comment2): Work for a single obstack.
2553 Rename as...
2554 (copy_comment): this.
2555
25562001-12-15 Akim Demaille <akim@epita.fr>
2557
2558 * src/lex.c, src/lex.h (xgetc): No longer static.
2559 * src/reader.c (parse_union_decl): Revamp.
2560
25612001-12-15 Akim Demaille <akim@epita.fr>
2562
2563 Still making progress in separating Bison into (i) input, (ii)
2564 process, (iii) output: now we can directly output the parser file
2565 without using table_obstack at all.
2566
2567 * src/files.c, src/files.h (table_obstack): Bye bye.
2568 (parser_file_name): New.
2569 * src/files.c (compute_output_file_names): Compute it.
2570 * src/output.c (actions_output, output_parser)
2571 (output_master_parser): To a file instead of an obstack.
2572
25732001-12-15 Akim Demaille <akim@epita.fr>
2574
2575 Attach actions to rules, instead of pre-outputting them to
2576 actions_obstack.
2577
2578 * src/gram.h (rule_t): action and action_line are new members.
2579 * src/reader.c (symbol_list): Likewise.
2580 (copy_action): Save the actions within the rule.
2581 (packgram): Save them in rule_table.
2582 * src/output.c (actions_output): New.
2583 (output_parser): Use it on `%%actions'.
2584 (output_rule_data): Don't free rule_table.
2585 (output): Do it.
2586 (prepare): Don't save the `action' muscle.
2587 * src/bison.simple: s/%%action/%%actions/.
2588
25892001-12-15 Akim Demaille <akim@epita.fr>
2590
2591 * src/reader.c (copy_action): When --yacc, don't append a `;'
2592 to the user action: let it fail if lacking.
2593 Suggested by Arnold Robbins and Tom Tromey.
2594
25952001-12-14 Akim Demaille <akim@epita.fr>
2596
2597 * src/lex.c (literalchar): Simply return the char you decoded, non
2598 longer mess around with obstacks and int pointers.
2599 Adjust all callers.
2600
26012001-12-14 Akim Demaille <akim@epita.fr>
2602
2603 * src/lex.c (literalchar): Don't escape the special characters,
2604 just decode them, and keep them as char (before, eol was output as
2605 the 2 char string `\n' etc.).
2606 * src/output.c (output_rule_data): Use quotearg to output the
2607 token strings.
2608
26092001-12-13 Paul Eggert <eggert@twinsun.com>
2610
2611 * src/bison.simple (YYSIZE_T, YYSTACK_ALLOC, YYSTACK_FREE):
2612 Do not infringe on the global user namespace when using C++.
2613 (YYFPRINTF, YYSTDERR): New macros, needed for the above.
2614 All uses of `fprintf' and `stderr' changed.
2615
2616 * doc/bison.texinfo: Document YYFPRINTF, YYSTDERR.
2617
26182001-12-13 Akim Demaille <akim@epita.fr>
2619
2620 The computation of nullable is broken: it doesn't handle empty
2621 RHS's properly.
2622
2623 * tests/torture.at (GNU AWK Grammar): New.
2624 * tests/sets.at (Nullable): New.
2625 * src/nullable.c (set_nullable): Instead of blindly looping over
2626 `ritems', loop over the rules, and then over their rhs's.
2627
2628 Work around Autotest bugs.
2629
2630 * src/warshall.c (bitmatrix_print): Don't use `+--+' as table
2631 frame, because Autotest understand lines starting with a `+' as
2632 traces from the shell. Then, they are not processed properly.
2633 Admittedly an Autotest bug, but we don't have time to wait for
2634 Autotest to catch up.
2635 * tests/regression.at (Broken Closure): Adjust to the new table
2636 frames.
2637 Move to...
2638 * tests/sets.at: here.
2639
26402001-12-13 Akim Demaille <akim@epita.fr>
2641
2642 * src/closure.c (closure): Use nrules instead of playing tricks
2643 with BITS_PER_WORD.
2644
26452001-12-13 Akim Demaille <akim@epita.fr>
2646
2647 * src/print.c (print_actions): Output the handling of `$' as the
2648 traces do: shifting the token EOF. Before EOF was treated as a
2649 nonterminal.
2650 * tests/regression.at: Adjust some tests.
2651 * src/print_graph.c (print_core): Complete the set of items via
2652 closure. The next-to-final and final states are still unsatisfying,
2653 but that's to be addressed elsewhere.
2654 No longer output the rule numbers, but do output the state number.
2655 A single loop for the shifts + gotos is enough, but picked a
2656 distinct color for each.
2657 (print_graph): Initialize and finalize closure.
2658
26592001-12-13 Akim Demaille <akim@epita.fr>
2660
2661 * src/reader.c (readgram): Remove dead code, an strip useless
2662 braces.
2663 (get_type): Remove, unused.
2664
26652001-12-12 Akim Demaille <akim@epita.fr>
2666
2667 * src/complain.h, src/complain.c: Remove error_one_per_line, rely
2668 on that of lib/error.c.
2669
26702001-12-12 Akim Demaille <akim@epita.fr>
2671
2672 Some hosts don't like `/' in includes.
2673
2674 * src/system.h: Include libgettext.h without qualifying the path.
2675 * src/Makefile.am (INCLUDES): Add $(top_srcdir)/intl, remove
2676 $(top_srcdir).
2677
26782001-12-11 Marc Autret <autret_m@epita.fr>
2679
2680 * src/output.c (output_parser): Remove useless muscle.
2681
26822001-12-11 Marc Autret <autret_m@epita.fr>
2683
2684 * src/bison.simple: Remove #line just before %%epilogue. It
2685 is now handled in ...
2686 * src/reader.c (read_additionnal_code): Add the output of a
2687 #line for the epilogue.
2688
26892001-12-10 Marc Autret <autret_m@epita.fr>
2690
2691 * src/reader.c (copy_definition): Re-use CPP-outed code which
2692 replace precedent remove.
2693 * src/bison.simple: Remove #line before %%prologue because
2694 %%input-line is wrong at this time.
2695
26962001-12-10 Marc Autret <autret_m@epita.fr>
2697
2698 * src/reader.c (symbols_output): Clean up.
2699 * src/output.c (output_gram, output): Clean up.
2700
27012001-12-10 Akim Demaille <akim@epita.fr>
2702
2703 * src/lalr.c (initialize_lookaheads): New. Extracted from...
2704 * src/LR0.c (set_state_table): here.
2705 * src/lalr.c (lalr): Call it.
2706
27072001-12-10 Akim Demaille <akim@epita.fr>
2708
2709 * src/state.h (shifts): Remove the `number' member: shifts are
2710 attached to state, hence no longer need to be labelled with a
2711 state number.
2712
27132001-12-10 Akim Demaille <akim@epita.fr>
2714
2715 Now that states have a complete set of members, the linked list of
2716 shifts is useless: just fill directly the state's shifts member.
2717
2718 * src/state.h (shifts): Remove the `next' member.
2719 * src/LR0.c (first_state, last_state): Remove.
2720 Adjust the callers.
2721 (augment_automaton): Don't look for the shifts that must be added
2722 a shift on EOF: it is those of the state we looked for! But now,
2723 since shifts are attached, it is no longer needed to looking
2724 merely by its id: its number.
2725
27262001-12-10 Akim Demaille <akim@epita.fr>
2727
2728 * src/LR0.c (augment_automaton): Better variable locality.
2729 Remove an impossible branch: if there is a state corresponding to
2730 the start symbol being shifted, then there is shift for the start
2731 symbol from the initial state.
2732
27332001-12-10 Akim Demaille <akim@epita.fr>
2734
2735 * src/LR0.c (augment_automaton): Call `insert_eof_shifting_state'
2736 only when appropriate: when insert_start_shifting_state' is not
2737 invoked.
2738 * tests/regression.at (Rule Line Numbers): Adjust.
2739
27402001-12-10 Akim Demaille <akim@epita.fr>
2741
2742 * src/LR0.c (augment_automaton): Now that all states have shifts,
2743 merge the two cases addition shifts to the initial state.
2744
27452001-12-10 Akim Demaille <akim@epita.fr>
2746
2747 * src/lalr.c (set_state_table): Move to...
2748 * src/LR0.c: here.
2749 * src/lalr.c (lalr): Don't call it...
2750 * src/LR0.c (generate_states): do it.
2751 * src/LR0.h (first_state): Remove, only the table is used.
2752
27532001-12-10 Akim Demaille <akim@epita.fr>
2754
2755 * src/LR0.h (first_shift, first_reduction): Remove.
2756 * src/lalr.c: Don't use first_shift: find shifts through the
2757 states.
2758
27592001-12-10 Akim Demaille <akim@epita.fr>
2760
2761 * src/LR0.c: Attach shifts to states as soon as they are
2762 computed.
2763 * src/lalr.c (set_state_table): Instead of assigning shifts to
2764 state, just assert that the mapping was properly done.
2765
27662001-12-10 Akim Demaille <akim@epita.fr>
2767
2768 * src/LR0.c (insert_start_shift): Rename as...
2769 (insert_start_shifting_state): this.
2770 (insert_eof_shifting_state, insert_accepting_state): New.
2771 (augment_automaton): Adjust.
2772 Better locality of the variables.
2773 When looking if the start_symbol is shifted from the initial
2774 state, using `while (... symbol != start_symbol ...)' sounds
2775 better than `while (... symbol < start_symbol ...)': If fail
2776 to see how the order between symbols could be relevant!
2777
27782001-12-10 Akim Demaille <akim@epita.fr>
2779
2780 * src/getargs.h: Don't declare `spec_name_prefix' and
2781 `spec_file_prefix', declared by src/files.h.
2782 * src/files.c, src/files.h: Default for spec_name_prefix is "yy".
2783 * src/muscle_tab.c (muscle_init): Default prefix to NULL.
2784 * src/output.c (prepare): Adjust.
2785 * src/reader.c (symbols_output): Likewise.
2786 * src/vmsgetargs.c: Vaguely adjust, but who cares?
2787
27882001-12-10 Akim Demaille <akim@epita.fr>
2789
2790 * src/muscle_tab.c (muscle_init): NULL is a better default than
2791 `"0"'.
2792
27932001-12-10 Akim Demaille <akim@epita.fr>
2794
2795 * src/reader.c (reader): Calling symbols_output once is enough.
2796
27972001-12-10 Akim Demaille <akim@epita.fr>
2798
2799 Now that states have a complete set of members, the linked list of
2800 reductions is useless: just fill directly the state's reductions
2801 member.
2802
2803 * src/state.h (struct reductions): Remove member `number' and
2804 `next'.
2805 * src/LR0.c (first_reduction, last_reduction): Remove.
2806 (save_reductions): Don't link the new reductions, store them in
2807 this_state.
2808 * src/lalr.c (set_state_table): No need to attach reductions to
2809 states, it's already done.
2810 * src/output.c (output_actions): No longer free the shifts, then
2811 the reductions, then the states: free all the states and their
2812 members.
2813
28142001-12-10 Akim Demaille <akim@epita.fr>
2815
2816 * src/options.c (OPTN, DRTV, BOTH): New.
2817 (option_table): Use them.
2818
2819 * src/muscle_tab.c: Don't include xalloc.h and string.h: that's
2820 the job of system.h.
2821 * src/options.c: Don't include stdio.h and xalloc.h for the same
2822 reasons.
2823
28242001-12-10 Akim Demaille <akim@epita.fr>
2825
2826 * src/output.c (output, prepare): Make sure the values of the
2827 muscles `action' and `prologue' are 0-terminated.
2828
28292001-12-10 Akim Demaille <akim@epita.fr>
2830
2831 Clean up GCC warnings.
2832
2833 * src/reader.c (copy_action): `buf' is not used.
2834 (parse_skel_decl): Be static.
2835 * src/muscle_tab.c (mhash1, mhash2, muscle_insert): Preserve `const'.
2836 * src/options.h (create_long_option_table): Have a real prototype.
2837 * lib/hash.c, lib/hash.h (hash_insert, hash_insert_at, hash_delete)
2838 (hash_delete_at): Return const void *.
2839 Adjust casts to preserve the const.
2840
28412001-12-10 Akim Demaille <akim@epita.fr>
2842
2843 * configure.in: Require 2.52g.
2844 M4 is not needed, but AUTOM4TE is.
2845 * m4/m4.m4: Remove.
2846 * tests/Makefile.am: Adjust.
2847
28482001-12-10 Akim Demaille <akim@epita.fr>
2849
2850 One structure for states is enough, even though theoretically
2851 there are LR(0) states and LALR(1) states.
2852
2853 * src/lalr.h (state_t): Remove.
2854 (state_table): Be state_t **, not state_t *.
2855 * src/state.h (core, CORE_ALLOC): Rename as...
2856 (state_t, STATE_ALLOC): this.
2857 Add the LALR(1) members: shifts, reductions, errs.
2858 * src/LR0.c (state_table): Rename as...
2859 (state_hash): this, to avoid name clashes with the global
2860 `state_table'.
2861 * src/print_graph.c, src/LR0.c, src/LR0.h, src/conflicts.c
2862 * src/lalr.c, src/lalr.h, src/output.c, src/print.c: Adjust.
2863
28642001-12-10 Akim Demaille <akim@epita.fr>
2865
2866 Bison dumps core on bash.y.
2867 Reported by Pascal Bart.
2868
2869 * src/warshall.c (bitmatrix_print): New.
2870 (TC): Use it.
2871 When performing a transitive closure R(i, j) && R(j, k) => R(i, k),
2872 j must be the outer loop.
2873 * tests/regression.at (Broken Closure): New.
2874
28752001-12-05 Akim Demaille <akim@epita.fr>
2876
2877 * tests/atlocal.in (CPPFLAGS): Do not leave a space between -I and
2878 its argument.
2879