]>
Commit | Line | Data |
---|---|---|
1 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
2 | ||
3 | * src/output.c (header_output): Don't forget to export YYLTYPE and | |
4 | yylloc. | |
5 | * tests/headers.at (export YYLTYPE): New, make sure it does. | |
6 | * tests/regression.at (%union and --defines, Invalid CPP headers): | |
7 | Move to... | |
8 | * tests/headers.at: here. | |
9 | ||
10 | ||
11 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
12 | ||
13 | * src/gram.h (rule_s): Member `assoc' is of type `associativity'. | |
14 | ||
15 | ||
16 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
17 | ||
18 | * tests/actions.at (Mid-rule actions): Output on a single line | |
19 | instead of several. | |
20 | ||
21 | ||
22 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
23 | ||
24 | * doc/bison.texinfo: Formatting changes. | |
25 | ||
26 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
27 | ||
28 | Don't store the token defs in a muscle, just be ready to output it | |
29 | on command. Now possible via `symbols'. Fixes a memory leak. | |
30 | ||
31 | * src/output.c (token_definitions_output): New. | |
32 | (output_parser, header_output): Use it. | |
33 | * src/reader.c (symbols_save): Remove. | |
34 | ||
35 | ||
36 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
37 | ||
38 | * src/bison.simple: Do not provide a default for YYSTYPE and | |
39 | YYLTYPE before the user's prologue. Otherwise it's hardly... a | |
40 | default. | |
41 | ||
42 | ||
43 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
44 | ||
45 | Mid-rule actions are simply... ignored! | |
46 | ||
47 | * src/reader.c (readgram): Be sure to attach mid-rule actions to | |
48 | the empty-rule associated to the dummy symbol, not to the host | |
49 | rule. | |
50 | * tests/actions.at (Mid-rule actions): New. | |
51 | ||
52 | ||
53 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
54 | ||
55 | Memory leak. | |
56 | ||
57 | * src/reader.c (reader): Free grammar. | |
58 | ||
59 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
60 | ||
61 | Memory leak. | |
62 | ||
63 | * src/LR0.c (new_itemsets): Don't allocate `shift_symbol' here, | |
64 | since it allocates it for each state, although only one is needed. | |
65 | (allocate_storage): Do it here. | |
66 | ||
67 | ||
68 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
69 | ||
70 | * src/options.h, src/options.c (create_long_option_table): Rename | |
71 | as... | |
72 | (long_option_table_new): this, with a clearer prototype. | |
73 | (percent_table): Remove, unused, | |
74 | * src/getargs.c (getargs): Adjust. | |
75 | ||
76 | ||
77 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
78 | ||
79 | * src/LR0.c, src/conflicts.c, src/lalr.c, src/lalr.h, src/output.c | |
80 | * src/print.c, src/print_graph.c, src/state.h: Rename state_table | |
81 | as states. | |
82 | ||
83 | ||
84 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
85 | ||
86 | * src/lalr.c (build_relations): Rename `states' as `states1'. | |
87 | Sorry, I don't understand exactly what it is, no better name... | |
88 | ||
89 | ||
90 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
91 | ||
92 | * src/closure.c, src/conflicts.c, src/derives.c, src/gram.c | |
93 | * src/gram.h, src/lalr.c, src/nullable.c, src/output.c, src/print.c | |
94 | * src/print_graph.c, src/reader.c, src/reduce.c: Rename rule_table | |
95 | as rules. | |
96 | ||
97 | ||
98 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
99 | ||
100 | * src/gram.c (rprec, rprecsym, rassoc): Remove, unused since long | |
101 | ago. | |
102 | ||
103 | ||
104 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
105 | ||
106 | * src/reader.c, src/reader.h (user_toknums): Remove. | |
107 | Adjust all users to use symbols[i]->user_token_number. | |
108 | ||
109 | ||
110 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
111 | ||
112 | * src/gram.c, src/gram.h (sprec, sassoc): Remove. | |
113 | Adjust all users to use symbols[i]->prec or ->assoc. | |
114 | ||
115 | ||
116 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
117 | ||
118 | * src/reader.c, src/reader.h (tags): Remove. | |
119 | Adjust all users to use symbols[i]->tag. | |
120 | ||
121 | ||
122 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
123 | ||
124 | * src/gram.h, src/gram.c (symbols): New, similar to state_table | |
125 | and rule_table. | |
126 | * src/reader.c (packsymbols): Fill this table. | |
127 | Drop sprec. | |
128 | * src/conflicts.c (resolve_sr_conflict): Adjust. | |
129 | * src/reduce.c (reduce_grammar): Adjust: just sort symbols, a | |
130 | single table. | |
131 | Use symbols[i]->tag instead of tags[i]. | |
132 | ||
133 | ||
134 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
135 | ||
136 | * tests/calc.at (_AT_DATA_CALC_Y): Also use %union. | |
137 | In addition, put a comment in there, to replace... | |
138 | * tests/regression.at (%union and C comments): Remove. | |
139 | ||
140 | ||
141 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
142 | ||
143 | * tests/regression.at (Web2c Actions): Blindly move the actual | |
144 | output as expected output. The contents *seem* right to me, but I | |
145 | can't pretend reading perfectly parser tables... Nonetheless, all | |
146 | the other tests pass correctly, the table look OK, even though the | |
147 | presence of `$axiom' is to be noted: AFAICS it is useless (but | |
148 | harmless). | |
149 | ||
150 | ||
151 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
152 | ||
153 | * src/reader.c (readgram): Don't add the rule 0 if there were no | |
154 | rules read. In other words, add it _after_ having performed | |
155 | grammar sanity checks. | |
156 | Fixes the `tests/regression.at (Invalid input: 1)' Failure. | |
157 | ||
158 | ||
159 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
160 | ||
161 | * tests/regression.at (Web2c Report): Catch up: the rule 0 is now | |
162 | visible, and some states have now a different number. | |
163 | ||
164 | ||
165 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
166 | ||
167 | * src/reader.c (readgram): Bind the initial rule's lineno to that | |
168 | of the first rule. | |
169 | * tests/regression.at (Rule Line Numbers, Unresolved SR Conflicts): | |
170 | (Solved SR Conflicts): Adjust rule 0's line number. | |
171 | ||
172 | ||
173 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
174 | ||
175 | Fix the `GAWK Grammar' failure. | |
176 | ||
177 | * src/LR0.c (final_state): Initialize to -1 so that we do compute | |
178 | the reductions of the first state which was mistakenly confused | |
179 | with the final state because precisely final_state was initialized | |
180 | to 0. | |
181 | * tests/sets.at (Nullable): Adjust: state 0 does have lookaheads, | |
182 | now noticed by Bison. | |
183 | * tests/regression.at (Rule Line Numbers): Adjust: state 0 does | |
184 | have a reduction on $default. | |
185 | ||
186 | ||
187 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
188 | ||
189 | * src/gram.c (ritem_print): Be sure to subtract 1 when displaying | |
190 | rule line numbers. | |
191 | * src/closure.c (print_closure): Likewise. | |
192 | * src/derives.c (print_derives): Likewise. | |
193 | * tests/sets.at (Nullable): Adjust: the rule numbers are correct | |
194 | now. | |
195 | ||
196 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
197 | ||
198 | * src/lalr.c (lookaheads_print): New. | |
199 | (lalr): Call it when --trace-flag. | |
200 | * tests/sets.at (Nullable): Adjust: when tracing, the lookaheads | |
201 | are dumped. | |
202 | ||
203 | ||
204 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
205 | ||
206 | * src/derives.c (print_derives): Be sure to use `>= 0', not `> 0', | |
207 | when walking through ritem, even via rule->rhs. | |
208 | * src/reduce.c (dump_grammar, useful_production, reduce_output) | |
209 | (useful_production, useless_nonterminals): Likewise. | |
210 | (reduce_grammar_tables): Likewise, plus update nritems. | |
211 | * src/nullable.c (set_nullable): Likewise. | |
212 | * src/lalr.c (build_relations): Likewise. | |
213 | * tests/sets.at (Nullable): Adjust. | |
214 | Fortunately, now, the $axiom is no longer nullable. | |
215 | ||
216 | ||
217 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
218 | ||
219 | * src/LR0.c (generate_states): Use nritems, not nitems, nor using | |
220 | the 0-sentinel. | |
221 | * src/gram.c (ritem_longest_rhs): Likewise. | |
222 | * src/reduce.c (nonterminals_reduce): Likewise. | |
223 | * src/print_graph.c (print_graph): Likewise. | |
224 | * src/output.c (output_rule_data): Likewise. | |
225 | * src/nullable.c (set_nullable): Likewise. | |
226 | ||
227 | 2001-12-29 Akim Demaille <akim@epita.fr> | |
228 | ||
229 | * src/output.c: Comment changes. | |
230 | ||
231 | 2001-12-27 Paul Eggert <eggert@twinsun.com> | |
232 | ||
233 | * src/bison.simple (YYSTACK_ALLOC, YYSIZE_T): Remove special | |
234 | cases for non-GNU systems like AIX, HP-UX, SGI, Sun, and | |
235 | Sparc, as they were causing more porting problems than the | |
236 | (minor) performance improvement was worth. | |
237 | ||
238 | Also, catch up with 1.31's YYSTD. | |
239 | ||
240 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
241 | ||
242 | * src/output.c (output_gram): Rely on nritems, not the | |
243 | 0-sentinel. See below. | |
244 | Use -1 as separator, not 0. | |
245 | * src/bison.simple (yyparse): Subtract 1 to the rule numbers. | |
246 | Rely on -1 as separator in yyrhs, instead of 0. | |
247 | * tests/calc.at (AT_CHECK_CALC): Now, the parsers no longer issue | |
248 | twice `Now at end of input', therefore there are two lines less to | |
249 | expect. | |
250 | ||
251 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
252 | ||
253 | * tests/regression.at (Unresolved SR Conflicts): | |
254 | (Solved SR Conflicts, Rule Line Numbers): Adjust to the changes | |
255 | below. | |
256 | ||
257 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
258 | ||
259 | * src/LR0.c (new_state): Recognize the final state by the fact it | |
260 | is reached by eoftoken. | |
261 | (insert_start_shifting_state, insert_eof_shifting_state) | |
262 | (insert_accepting_state, augment_automaton): Remove, since now | |
263 | these states are automatically computed from the initial state. | |
264 | (generate_states): Adjust. | |
265 | * src/print.c: When reporting a rule number to the user, substract | |
266 | 1, so that the axiom rule is rule 0, and the first user rule is 1. | |
267 | * src/reduce.c: Likewise. | |
268 | * src/print_graph.c (print_core): For the time being, just as for | |
269 | the report, depend upon --trace-flags to dump the full set of | |
270 | items. | |
271 | * src/reader.c (readgram): Once the grammar read, insert the rule | |
272 | 0: `$axiom: START-SYMBOL $'. | |
273 | * tests/set.at: Adjust: rule 0 is now displayed, and since the | |
274 | number of the states has changed (the final state is no longer | |
275 | necessarily the last), catch up. | |
276 | ||
277 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
278 | ||
279 | Try to make the use of the eoftoken valid. Given that its value | |
280 | is 0 which was also used as a sentinel in ritem, (i) make sure >= 0 | |
281 | is used instead of > 0 where appropriate, (ii), depend upon nritems | |
282 | instead of the 0-sentinel. | |
283 | ||
284 | * src/gram.h, src/gram.c (nritems): New. | |
285 | Expected to be duplication of nitems, but for the time being... | |
286 | * src/reader.c (packgram): Assert nritems and nitems are equal. | |
287 | * src/LR0.c (allocate_itemsets, new_itemsets): Adjust. | |
288 | * src/closure.c (print_closure, print_fderives): Likewise. | |
289 | * src/gram.c (ritem_print): Likewise. | |
290 | * src/print.c (print_core, print_grammar): Likewise. | |
291 | * src/print_graph.c: Likewise. | |
292 | ||
293 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
294 | ||
295 | * src/main.c (main): If there are complains after grammar | |
296 | reductions, then output the report anyway if requested, then die. | |
297 | * src/symtab.c (bucket_new): Initialize `value' to -1, not 0. | |
298 | * src/reader.c (eoftoken): New. | |
299 | (parse_token_decl): If the token being defined has value `0', it | |
300 | is the eoftoken. | |
301 | (packsymbols): No longer hack `tags' to insert `$' by hand. | |
302 | Be sure to preserve the value of the eoftoken. | |
303 | (reader): Make sure eoftoken is defined. | |
304 | Initialize nsyms to 0: now eoftoken is created just like the others. | |
305 | * src/print.c (print_grammar): Don't special case the eof token. | |
306 | * src/regression.at: Adjust: `$' has value 0, not -1, which was a | |
307 | lie anyway, albeit pleasant. | |
308 | * tests/calc.at: Exercise error messages with eoftoken. | |
309 | Change the grammar so that empty input is invalid. | |
310 | Adjust expectations. | |
311 | When yyungeting, be sure to use a valid yylloc: use last_yylloc. | |
312 | ||
313 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
314 | ||
315 | * configure.in: Check the protos of strchr ans strspn. | |
316 | Replace strchr if needed. | |
317 | * src/system.h: Provide the protos of strchr, strspn and memchr if | |
318 | missing. | |
319 | * lib/strchr.c: New. | |
320 | * src/reader.c (symbols_save): Use strchr. | |
321 | ||
322 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
323 | ||
324 | * src/print.c, src/print_graph.c (escape): New. | |
325 | Use it to quote the TAGS outputs. | |
326 | * src/print_graph.c (print_state): Now errors are in red, and | |
327 | reductions in green. | |
328 | Prefer high to wide: output the state number on a line of its own. | |
329 | ||
330 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
331 | ||
332 | * src/state.h, src/state.c (reductions_new): New. | |
333 | * src/LR0.c (set_state_table): Let all the states have a | |
334 | `reductions', even if reduced to 0. | |
335 | (save_reductions): Adjust. | |
336 | * src/lalr.c (initialize_LA, initialize_lookaheads): Adjust. | |
337 | * src/print.c (print_reductions, print_actions): Adjust. | |
338 | * src/output.c (action_row): Adjust. | |
339 | ||
340 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
341 | ||
342 | * src/state.h, src/state.c (errs_new, errs_dup): New. | |
343 | * src/LR0.c (set_state_table): Let all the states have an errs, | |
344 | even if reduced to 0. | |
345 | * src/print.c (print_errs, print_reductions): Adjust. | |
346 | * src/output.c (output_actions, action_row): Adjust. | |
347 | * src/conflicts.c (resolve_sr_conflict): Adjust. | |
348 | ||
349 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
350 | ||
351 | * src/lalr.c (set_goto_map, initialize_F): Use SHIFT_SYMBOL. | |
352 | ||
353 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
354 | ||
355 | * src/conflicts.c, src/conflicts.h (print_reductions): Move to... | |
356 | * src/print.c: here. | |
357 | (lookaheadset, shiftset): New, used as additional storage by | |
358 | print_reductions. | |
359 | (print_results): Adjust. | |
360 | (print_shifts, print_gotos, print_errs): New, extracted from... | |
361 | (print_actions): here. | |
362 | * src/print_graph.c (print_actions): Remove dead code. | |
363 | ||
364 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
365 | ||
366 | * src/reader.c (copy_dollar, copy_at): Better checking of `n' in | |
367 | `$n' and `@n'. | |
368 | ||
369 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
370 | ||
371 | * src/lalr.c (add_lookback_edge): Use state_t instead of ints. | |
372 | (build_relations): Adjust. | |
373 | ||
374 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
375 | ||
376 | * src/lalr.c (set_goto_map): Remove a wrong but benign loop | |
377 | duplication. | |
378 | ||
379 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
380 | ||
381 | * src/reader.c (packgram): Catch nitems overflows. | |
382 | ||
383 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
384 | ||
385 | * src/files.c, src/files.h (guard_obstack): Remove. | |
386 | * src/output.c (output): Adjust. | |
387 | * src/reader.c (parse_braces): New, factoring... | |
388 | (copy_action, copy_guard): these two which are renamed as... | |
389 | (parse_action, parse_guard): these. | |
390 | As a voluntary consequence, using braces around guards is now | |
391 | mandatory. | |
392 | ||
393 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
394 | ||
395 | * src/gram.h (rule_t): `guard' and `guard_line' are new members. | |
396 | * src/reader.c (symbol_list): `guard' and `guard_line' are new | |
397 | members. | |
398 | (symbol_list_new): Adjust. | |
399 | (copy_action): action_line is the first line, not the last. | |
400 | (copy_guard): Just as for actions, store the `action' only, not | |
401 | the switch/case/break flesh. | |
402 | Don't parse the user action that might follow the guard, let... | |
403 | (readgram): do it, i.e., now, there can be an action after a | |
404 | guard. | |
405 | In other words the guard is just explicitly optional. | |
406 | (packgram): Adjust. | |
407 | * src/output.c (guards_output): New. | |
408 | (output_parser): Call it when needed. | |
409 | (output): Also free the guard and attrs obstacks. | |
410 | * src/files.c, src/files.h (obstack_save): Remove. | |
411 | (output_files): Remove. | |
412 | As a result, if one needs the former `.act' file, using an | |
413 | appropriate skeleton which requires actions and guards is now | |
414 | required. | |
415 | * src/main.c (main): Adjust. | |
416 | * tests/semantic.at: New. | |
417 | * tests/regression.at: Use `input.y' as input file name. | |
418 | Avoid 8+3 problems by requiring input.c when the test needs the | |
419 | parser. | |
420 | ||
421 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
422 | ||
423 | * src/reader.c (symbol_list_new): Be sure to initialize all the | |
424 | fields. | |
425 | ||
426 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
427 | ||
428 | All the hacks using a final pseudo state are now useless. | |
429 | ||
430 | * src/LR0.c (set_state_table): state_table holds exactly nstates. | |
431 | * src/lalr.c (nLA): New. | |
432 | (initialize_LA, compute_lookaheads, initialize_lookaheads): Use it | |
433 | instead of lookaheadsp from the pseudo state (nstate + 1). | |
434 | ||
435 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
436 | ||
437 | * src/output.c (action_row, token_actions): Use a state_t instead | |
438 | of a integer, and nlookaheads instead of the following state's | |
439 | lookaheadsp. | |
440 | ||
441 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
442 | ||
443 | * src/conflicts.c (log_resolution, flush_shift) | |
444 | (resolve_sr_conflict, set_conflicts, solve_conflicts) | |
445 | (count_sr_conflicts, count_rr_conflicts, conflicts_output) | |
446 | (conflicts_print, print_reductions): Use a state_t instead of an | |
447 | integer when referring to a state. | |
448 | As much as possible, depend upon nlookaheads, instead of the | |
449 | `lookaheadsp' member of the following state (since lookaheads of | |
450 | successive states are successive, the difference between state n + 1 | |
451 | and n served as the number of lookaheads for state n). | |
452 | * src/lalr.c (add_lookback_edge): Likewise. | |
453 | * src/print.c (print_core, print_actions, print_state) | |
454 | (print_results): Likewise. | |
455 | * src/print_graph.c (print_core, print_actions, print_state) | |
456 | (print_graph): Likewise. | |
457 | * src/conflicts.h: Adjust. | |
458 | ||
459 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
460 | ||
461 | * src/bison.hairy: Formatting/comment changes. | |
462 | ANSIfy. | |
463 | Remove `register' indications. | |
464 | Add plenty of `static'. | |
465 | ||
466 | 2001-12-27 Akim Demaille <akim@epita.fr> | |
467 | ||
468 | * src/output.c (prepare): Drop the muscle `ntbase' which | |
469 | duplicates ntokens. | |
470 | * src/bison.simple: Formatting/comment changes. | |
471 | Use YYNTOKENS only, which is documented, but not YYNTBASE, which | |
472 | is an undocumented synonym. | |
473 | ||
474 | 2001-12-22 Akim Demaille <akim@epita.fr> | |
475 | ||
476 | * src/output.c (output_table_data): Change the prototype to use | |
477 | `int' for array ranges: some invocations do pass an int, not a | |
478 | short. | |
479 | Reported by Wayne Green. | |
480 | ||
481 | 2001-12-22 Akim Demaille <akim@epita.fr> | |
482 | ||
483 | Some actions of web2c.y are improperly triggered. | |
484 | Reported by Mike Castle. | |
485 | ||
486 | * src/lalr.c (traverse): s/F (i)[k] = F (j)[k]/F (j)[k] = F (i)[k]/. | |
487 | * tests/regression.at (Web2c): Rename as... | |
488 | (Web2c Report): this. | |
489 | (Web2c Actions): New. | |
490 | ||
491 | 2001-12-22 Akim Demaille <akim@epita.fr> | |
492 | ||
493 | Reductions in web2c.y are improperly reported. | |
494 | Reported by Mike Castle. | |
495 | ||
496 | * src/conflicts.c (print_reductions): Fix. | |
497 | * tests/regression.at (Web2c): New. | |
498 | ||
499 | 2001-12-18 Akim Demaille <akim@epita.fr> | |
500 | ||
501 | Some host fail on `assert (!"foo")', which expands to | |
502 | ((!"foo") ? (void)0 : __assert("!"foo."", __FILE__, __LINE__)) | |
503 | Reported by Nelson Beebee. | |
504 | ||
505 | * src/output.c, src/vcg.c: Replace `assert (!"it succeeded")' with | |
506 | `#define it_succeeded 0' and `assert (it_succeeded)'. | |
507 | ||
508 | 2001-12-17 Marc Autret <autret_m@epita.fr> | |
509 | ||
510 | * src/bison.simple: Don't hard code the skeleton line and filename. | |
511 | * src/output.c (output_parser): Rename 'line' as 'output_line'. | |
512 | New line counter 'skeleton_line' (skeleton-line muscle). | |
513 | ||
514 | 2001-12-17 Paul Eggert <eggert@twinsun.com> | |
515 | ||
516 | * NEWS, doc/bison.texinfo, doc/bison.1, doc/bison.rnh: Document that | |
517 | YYDEBUG must be defined to a nonzero value. | |
518 | ||
519 | * src/bison.simple (yytname): Do not assume that the user defines | |
520 | YYDEBUG to a properly parenthesized expression. | |
521 | ||
522 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
523 | ||
524 | * src/state.h (state_t): Rename lookaheads as lookaheadsp. | |
525 | nlookaheads is a new member. | |
526 | Adjust all users. | |
527 | * src/lalr.h (nlookaheads): Remove this orphan declaration. | |
528 | * src/lalr.c (initialize_lookaheads): Set nlookaheads for each | |
529 | state. | |
530 | ||
531 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
532 | ||
533 | * src/files.h, src/files.c (open_files, close_files): Remove. | |
534 | * src/main.c (main): Don't open/close files, nor invoke lex_free, | |
535 | let... | |
536 | * src/reader.c (reader): Do it. | |
537 | ||
538 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
539 | ||
540 | * src/conflicts.c (print_reductions): Formatting changes. | |
541 | ||
542 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
543 | ||
544 | * src/conflicts.c (flush_shift): Also adjust lookaheadset. | |
545 | (flush_reduce): New. | |
546 | (resolve_sr_conflict): Adjust. | |
547 | ||
548 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
549 | ||
550 | * src/output.c (output_obstack): Be static and rename as... | |
551 | (format_obstack): this, to avoid any confusion with files.c's | |
552 | output_obstack. | |
553 | * src/reader.h (muscle_obstack): Move to... | |
554 | * src/output.h: here, since it's defined in output.c. | |
555 | ||
556 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
557 | ||
558 | * src/output.c (action_row, save_column, default_goto) | |
559 | (sort_actions, matching_state, pack_vector): Better variable | |
560 | locality. | |
561 | ||
562 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
563 | ||
564 | * src/output.c: Various formatting changes. | |
565 | ||
566 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
567 | ||
568 | * src/files.c (output_files): Free the output_obstack. | |
569 | * src/main.c (main): Call print and print_graph conditionally. | |
570 | * src/print.c (print): Work unconditionally. | |
571 | * src/print_graph.c (print_graph): Work unconditionally. | |
572 | * src/conflicts.c (log_resolution): Output only if verbose_flag. | |
573 | ||
574 | 2001-12-16 Marc Autret <autret_m@epita.fr> | |
575 | ||
576 | * src/output.c (actions_output): Fix. When we use %no-lines, | |
577 | there is one less line per action. | |
578 | ||
579 | 2001-12-16 Marc Autret <autret_m@epita.fr> | |
580 | ||
581 | * src/bison.simple: Remove a useless #line directive. | |
582 | s/#line %%line %%skeleton/#line %%line "%%parser-file-name"/'. | |
583 | * src/output.c (get_lines_number): New. | |
584 | (output_parser): Adjust, now takes care about the lines of a | |
585 | output muscles. | |
586 | Fix line numbering. | |
587 | (actions_output): Computes the number of lines taken by actions. | |
588 | (output_master_parser): Insert new skeleton which is the name of | |
589 | the output parser file name. | |
590 | ||
591 | 2001-12-15 Marc Autret <autret_m@epita.fr> | |
592 | ||
593 | * src/bison.simple [YYERROR_VERBOSE]: Restore backward compatibility. | |
594 | ||
595 | 2001-12-15 Marc Autret <autret_m@epita.fr> | |
596 | ||
597 | * src/output.c (output_gram): Keep track of the hairy one. | |
598 | ||
599 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
600 | ||
601 | Make `make distcheck' work. | |
602 | ||
603 | * lib/Makefile.am (INCLUDES): Add top_srcdir/intl, since hash uses | |
604 | system.h which uses libgettext.h. | |
605 | ||
606 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
607 | ||
608 | * src/nullable.c (set_nullable): Useless rules must be skipped, | |
609 | otherwise, since we range over their symbols, we might look at a | |
610 | nonterminal which no longer ``exists'', i.e., it is not counted in | |
611 | `nvars', hence we overflow our arrays. | |
612 | ||
613 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
614 | ||
615 | The header can also be produced directly, without any obstack! | |
616 | Yahoo! | |
617 | ||
618 | * src/files.c, src/files.h (defines_obstack): Remove. | |
619 | (compute_header_macro): Global. | |
620 | (defines_obstack_save): Remove. | |
621 | * src/reader.c (parse_union_decl): No longer output to | |
622 | defines_obstack: its content can be found in the `stype' muscle | |
623 | anyway. | |
624 | (output_token_translations): Merge into... | |
625 | (symbols_output): this. | |
626 | Rename as... | |
627 | (symbols_save): this. | |
628 | (reader): Adjust. | |
629 | * src/output.c (header_output): New. | |
630 | (output): Call it. | |
631 | ||
632 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
633 | ||
634 | * src/reader.c (parse_union_decl): Instead of handling two obstack | |
635 | simultaneously, use one to define the `stype' muscle, and use the | |
636 | value of the latter to fill defines_obstack. | |
637 | (copy_comment): Remove. | |
638 | (copy_comment2): Work for a single obstack. | |
639 | Rename as... | |
640 | (copy_comment): this. | |
641 | ||
642 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
643 | ||
644 | * src/lex.c, src/lex.h (xgetc): No longer static. | |
645 | * src/reader.c (parse_union_decl): Revamp. | |
646 | ||
647 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
648 | ||
649 | Still making progress in separating Bison into (i) input, (ii) | |
650 | process, (iii) output: now we can directly output the parser file | |
651 | without using table_obstack at all. | |
652 | ||
653 | * src/files.c, src/files.h (table_obstack): Bye bye. | |
654 | (parser_file_name): New. | |
655 | * src/files.c (compute_output_file_names): Compute it. | |
656 | * src/output.c (actions_output, output_parser) | |
657 | (output_master_parser): To a file instead of an obstack. | |
658 | ||
659 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
660 | ||
661 | Attach actions to rules, instead of pre-outputting them to | |
662 | actions_obstack. | |
663 | ||
664 | * src/gram.h (rule_t): action and action_line are new members. | |
665 | * src/reader.c (symbol_list): Likewise. | |
666 | (copy_action): Save the actions within the rule. | |
667 | (packgram): Save them in rule_table. | |
668 | * src/output.c (actions_output): New. | |
669 | (output_parser): Use it on `%%actions'. | |
670 | (output_rule_data): Don't free rule_table. | |
671 | (output): Do it. | |
672 | (prepare): Don't save the `action' muscle. | |
673 | * src/bison.simple: s/%%action/%%actions/. | |
674 | ||
675 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
676 | ||
677 | * src/reader.c (copy_action): When --yacc, don't append a `;' | |
678 | to the user action: let it fail if lacking. | |
679 | Suggested by Arnold Robbins and Tom Tromey. | |
680 | ||
681 | 2001-12-14 Akim Demaille <akim@epita.fr> | |
682 | ||
683 | * src/lex.c (literalchar): Simply return the char you decoded, non | |
684 | longer mess around with obstacks and int pointers. | |
685 | Adjust all callers. | |
686 | ||
687 | 2001-12-14 Akim Demaille <akim@epita.fr> | |
688 | ||
689 | * src/lex.c (literalchar): Don't escape the special characters, | |
690 | just decode them, and keep them as char (before, eol was output as | |
691 | the 2 char string `\n' etc.). | |
692 | * src/output.c (output_rule_data): Use quotearg to output the | |
693 | token strings. | |
694 | ||
695 | 2001-12-13 Paul Eggert <eggert@twinsun.com> | |
696 | ||
697 | * src/bison.simple (YYSIZE_T, YYSTACK_ALLOC, YYSTACK_FREE): | |
698 | Do not infringe on the global user namespace when using C++. | |
699 | (YYFPRINTF, YYSTDERR): New macros, needed for the above. | |
700 | All uses of `fprintf' and `stderr' changed. | |
701 | ||
702 | * doc/bison.texinfo: Document YYFPRINTF, YYSTDERR. | |
703 | ||
704 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
705 | ||
706 | The computation of nullable is broken: it doesn't handle empty | |
707 | RHS's properly. | |
708 | ||
709 | * tests/torture.at (GNU AWK Grammar): New. | |
710 | * tests/sets.at (Nullable): New. | |
711 | * src/nullable.c (set_nullable): Instead of blindly looping over | |
712 | `ritems', loop over the rules, and then over their rhs's. | |
713 | ||
714 | Work around Autotest bugs. | |
715 | ||
716 | * src/warshall.c (bitmatrix_print): Don't use `+--+' as table | |
717 | frame, because Autotest understand lines starting with a `+' as | |
718 | traces from the shell. Then, they are not processed properly. | |
719 | Admittedly an Autotest bug, but we don't have time to wait for | |
720 | Autotest to catch up. | |
721 | * tests/regression.at (Broken Closure): Adjust to the new table | |
722 | frames. | |
723 | Move to... | |
724 | * tests/sets.at: here. | |
725 | ||
726 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
727 | ||
728 | * src/closure.c (closure): Use nrules instead of playing tricks | |
729 | with BITS_PER_WORD. | |
730 | ||
731 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
732 | ||
733 | * src/print.c (print_actions): Output the handling of `$' as the | |
734 | traces do: shifting the token EOF. Before EOF was treated as a | |
735 | nonterminal. | |
736 | * tests/regression.at: Adjust some tests. | |
737 | * src/print_graph.c (print_core): Complete the set of items via | |
738 | closure. The next-to-final and final states are still unsatisfying, | |
739 | but that's to be addressed elsewhere. | |
740 | No longer output the rule numbers, but do output the state number. | |
741 | A single loop for the shifts + gotos is enough, but picked a | |
742 | distinct color for each. | |
743 | (print_graph): Initialize and finalize closure. | |
744 | ||
745 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
746 | ||
747 | * src/reader.c (readgram): Remove dead code, an strip useless | |
748 | braces. | |
749 | (get_type): Remove, unused. | |
750 | ||
751 | 2001-12-12 Akim Demaille <akim@epita.fr> | |
752 | ||
753 | * src/complain.h, src/complain.c: Remove error_one_per_line, rely | |
754 | on that of lib/error.c. | |
755 | ||
756 | 2001-12-12 Akim Demaille <akim@epita.fr> | |
757 | ||
758 | Some hosts don't like `/' in includes. | |
759 | ||
760 | * src/system.h: Include libgettext.h without qualifying the path. | |
761 | * src/Makefile.am (INCLUDES): Add $(top_srcdir)/intl, remove | |
762 | $(top_srcdir). | |
763 | ||
764 | 2001-12-11 Marc Autret <autret_m@epita.fr> | |
765 | ||
766 | * src/output.c (output_parser): Remove useless muscle. | |
767 | ||
768 | 2001-12-11 Marc Autret <autret_m@epita.fr> | |
769 | ||
770 | * src/bison.simple: Remove #line just before %%epilogue. It | |
771 | is now handled in ... | |
772 | * src/reader.c (read_additionnal_code): Add the output of a | |
773 | #line for the epilogue. | |
774 | ||
775 | 2001-12-10 Marc Autret <autret_m@epita.fr> | |
776 | ||
777 | * src/reader.c (copy_definition): Re-use CPP-outed code which | |
778 | replace precedent remove. | |
779 | * src/bison.simple: Remove #line before %%prologue because | |
780 | %%input-line is wrong at this time. | |
781 | ||
782 | 2001-12-10 Marc Autret <autret_m@epita.fr> | |
783 | ||
784 | * src/reader.c (symbols_output): Clean up. | |
785 | * src/output.c (output_gram, output): Clean up. | |
786 | ||
787 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
788 | ||
789 | * src/lalr.c (initialize_lookaheads): New. Extracted from... | |
790 | * src/LR0.c (set_state_table): here. | |
791 | * src/lalr.c (lalr): Call it. | |
792 | ||
793 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
794 | ||
795 | * src/state.h (shifts): Remove the `number' member: shifts are | |
796 | attached to state, hence no longer need to be labelled with a | |
797 | state number. | |
798 | ||
799 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
800 | ||
801 | Now that states have a complete set of members, the linked list of | |
802 | shifts is useless: just fill directly the state's shifts member. | |
803 | ||
804 | * src/state.h (shifts): Remove the `next' member. | |
805 | * src/LR0.c (first_state, last_state): Remove. | |
806 | Adjust the callers. | |
807 | (augment_automaton): Don't look for the shifts that must be added | |
808 | a shift on EOF: it is those of the state we looked for! But now, | |
809 | since shifts are attached, it is no longer needed to looking | |
810 | merely by its id: its number. | |
811 | ||
812 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
813 | ||
814 | * src/LR0.c (augment_automaton): Better variable locality. | |
815 | Remove an impossible branch: if there is a state corresponding to | |
816 | the start symbol being shifted, then there is shift for the start | |
817 | symbol from the initial state. | |
818 | ||
819 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
820 | ||
821 | * src/LR0.c (augment_automaton): Call `insert_eof_shifting_state' | |
822 | only when appropriate: when insert_start_shifting_state' is not | |
823 | invoked. | |
824 | * tests/regression.at (Rule Line Numbers): Adjust. | |
825 | ||
826 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
827 | ||
828 | * src/LR0.c (augment_automaton): Now that all states have shifts, | |
829 | merge the two cases addition shifts to the initial state. | |
830 | ||
831 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
832 | ||
833 | * src/lalr.c (set_state_table): Move to... | |
834 | * src/LR0.c: here. | |
835 | * src/lalr.c (lalr): Don't call it... | |
836 | * src/LR0.c (generate_states): do it. | |
837 | * src/LR0.h (first_state): Remove, only the table is used. | |
838 | ||
839 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
840 | ||
841 | * src/LR0.h (first_shift, first_reduction): Remove. | |
842 | * src/lalr.c: Don't use first_shift: find shifts through the | |
843 | states. | |
844 | ||
845 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
846 | ||
847 | * src/LR0.c: Attach shifts to states as soon as they are | |
848 | computed. | |
849 | * src/lalr.c (set_state_table): Instead of assigning shifts to | |
850 | state, just assert that the mapping was properly done. | |
851 | ||
852 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
853 | ||
854 | * src/LR0.c (insert_start_shift): Rename as... | |
855 | (insert_start_shifting_state): this. | |
856 | (insert_eof_shifting_state, insert_accepting_state): New. | |
857 | (augment_automaton): Adjust. | |
858 | Better locality of the variables. | |
859 | When looking if the start_symbol is shifted from the initial | |
860 | state, using `while (... symbol != start_symbol ...)' sounds | |
861 | better than `while (... symbol < start_symbol ...)': If fail | |
862 | to see how the order between symbols could be relevant! | |
863 | ||
864 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
865 | ||
866 | * src/getargs.h: Don't declare `spec_name_prefix' and | |
867 | `spec_file_prefix', declared by src/files.h. | |
868 | * src/files.c, src/files.h: Default for spec_name_prefix is "yy". | |
869 | * src/muscle_tab.c (muscle_init): Default prefix to NULL. | |
870 | * src/output.c (prepare): Adjust. | |
871 | * src/reader.c (symbols_output): Likewise. | |
872 | * src/vmsgetargs.c: Vaguely adjust, but who cares? | |
873 | ||
874 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
875 | ||
876 | * src/muscle_tab.c (muscle_init): NULL is a better default than | |
877 | `"0"'. | |
878 | ||
879 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
880 | ||
881 | * src/reader.c (reader): Calling symbols_output once is enough. | |
882 | ||
883 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
884 | ||
885 | Now that states have a complete set of members, the linked list of | |
886 | reductions is useless: just fill directly the state's reductions | |
887 | member. | |
888 | ||
889 | * src/state.h (struct reductions): Remove member `number' and | |
890 | `next'. | |
891 | * src/LR0.c (first_reduction, last_reduction): Remove. | |
892 | (save_reductions): Don't link the new reductions, store them in | |
893 | this_state. | |
894 | * src/lalr.c (set_state_table): No need to attach reductions to | |
895 | states, it's already done. | |
896 | * src/output.c (output_actions): No longer free the shifts, then | |
897 | the reductions, then the states: free all the states and their | |
898 | members. | |
899 | ||
900 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
901 | ||
902 | * src/options.c (OPTN, DRTV, BOTH): New. | |
903 | (option_table): Use them. | |
904 | ||
905 | * src/muscle_tab.c: Don't include xalloc.h and string.h: that's | |
906 | the job of system.h. | |
907 | * src/options.c: Don't include stdio.h and xalloc.h for the same | |
908 | reasons. | |
909 | ||
910 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
911 | ||
912 | * src/output.c (output, prepare): Make sure the values of the | |
913 | muscles `action' and `prologue' are 0-terminated. | |
914 | ||
915 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
916 | ||
917 | Clean up GCC warnings. | |
918 | ||
919 | * src/reader.c (copy_action): `buf' is not used. | |
920 | (parse_skel_decl): Be static. | |
921 | * src/muscle_tab.c (mhash1, mhash2, muscle_insert): Preserve `const'. | |
922 | * src/options.h (create_long_option_table): Have a real prototype. | |
923 | * lib/hash.c, lib/hash.h (hash_insert, hash_insert_at, hash_delete) | |
924 | (hash_delete_at): Return const void *. | |
925 | Adjust casts to preserve the const. | |
926 | ||
927 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
928 | ||
929 | * configure.in: Require 2.52g. | |
930 | M4 is not needed, but AUTOM4TE is. | |
931 | * m4/m4.m4: Remove. | |
932 | * tests/Makefile.am: Adjust. | |
933 | ||
934 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
935 | ||
936 | One structure for states is enough, even though theoretically | |
937 | there are LR(0) states and LALR(1) states. | |
938 | ||
939 | * src/lalr.h (state_t): Remove. | |
940 | (state_table): Be state_t **, not state_t *. | |
941 | * src/state.h (core, CORE_ALLOC): Rename as... | |
942 | (state_t, STATE_ALLOC): this. | |
943 | Add the LALR(1) members: shifts, reductions, errs. | |
944 | * src/LR0.c (state_table): Rename as... | |
945 | (state_hash): this, to avoid name clashes with the global | |
946 | `state_table'. | |
947 | * src/print_graph.c, src/LR0.c, src/LR0.h, src/conflicts.c | |
948 | * src/lalr.c, src/lalr.h, src/output.c, src/print.c: Adjust. | |
949 | ||
950 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
951 | ||
952 | Bison dumps core on bash.y. | |
953 | Reported by Pascal Bart. | |
954 | ||
955 | * src/warshall.c (bitmatrix_print): New. | |
956 | (TC): Use it. | |
957 | When performing a transitive closure R(i, j) && R(j, k) => R(i, k), | |
958 | j must be the outer loop. | |
959 | * tests/regression.at (Broken Closure): New. | |
960 | ||
961 | 2001-12-05 Akim Demaille <akim@epita.fr> | |
962 | ||
963 | * tests/atlocal.in (CPPFLAGS): Do not leave a space between -I and | |
964 | its argument. | |
965 |