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