]>
Commit | Line | Data |
---|---|---|
3db472b9 AD |
1 | 2001-12-27 Akim Demaille <akim@epita.fr> |
2 | ||
3 | * src/output.c (output_gram): Rely on nritems, not the | |
4 | 0-sentinel. See below. | |
5 | Use -1 as separator, not 0. | |
6 | * src/bison.simple (yyparse): Subtract 1 to the rule numbers. | |
7 | Rely on -1 as separator in yyrhs, instead of 0. | |
8 | * tests/calc.at (AT_CHECK_CALC): Now, the parsers no longer issue | |
9 | twice `Now at end of input', therefore there are two lines less to | |
10 | expect. | |
11 | ||
b365aa05 AD |
12 | 2001-12-27 Akim Demaille <akim@epita.fr> |
13 | ||
14 | * tests/regression.at (Unresolved SR Conflicts): | |
15 | (Solved SR Conflicts, Rule Line Numbers): Adjust to the changes | |
16 | below. | |
17 | ||
30171f79 AD |
18 | 2001-12-27 Akim Demaille <akim@epita.fr> |
19 | ||
20 | * src/LR0.c (new_state): Recognize the final state by the fact it | |
21 | is reached by eoftoken. | |
22 | (insert_start_shifting_state, insert_eof_shifting_state) | |
23 | (insert_accepting_state, augment_automaton): Remove, since now | |
24 | these states are automatically computed from the initial state. | |
25 | (generate_states): Adjust. | |
26 | * src/print.c: When reporting a rule number to the user, substract | |
27 | 1, so that the axiom rule is rule 0, and the first user rule is 1. | |
28 | * src/reduce.c: Likewise. | |
29 | * src/print_graph.c (print_core): For the time being, just as for | |
30 | the report, depend upon --trace-flags to dump the full set of | |
31 | items. | |
32 | * src/reader.c (readgram): Once the grammar read, insert the rule | |
33 | 0: `$axiom: START-SYMBOL $'. | |
34 | * tests/set.at: Adjust: rule 0 is now displayed, and since the | |
35 | number of the states has changed (the final state is no longer | |
36 | necessarily the last), catch up. | |
37 | ||
75142d45 AD |
38 | 2001-12-27 Akim Demaille <akim@epita.fr> |
39 | ||
40 | Try to make the use of the eoftoken valid. Given that its value | |
41 | is 0 which was also used as a sentinel in ritem, (i) make sure >= 0 | |
42 | is used instead of > 0 where appropriate, (ii), depend upon nritems | |
43 | instead of the 0-sentinel. | |
44 | ||
45 | * src/gram.h, src/gram.c (nritems): New. | |
46 | Expected to be duplication of nitems, but for the time being... | |
47 | * src/reader.c (packgram): Assert nritems and nitems are equal. | |
48 | * src/LR0.c (allocate_itemsets, new_itemsets): Adjust. | |
49 | * src/closure.c (print_closure, print_fderives): Likewise. | |
50 | * src/gram.c (ritem_print): Likewise. | |
51 | * src/print.c (print_core, print_grammar): Likewise. | |
52 | * src/print_graph.c: Likewise. | |
53 | ||
b7c49edf AD |
54 | 2001-12-27 Akim Demaille <akim@epita.fr> |
55 | ||
56 | * src/main.c (main): If there are complains after grammar | |
57 | reductions, then output the report anyway if requested, then die. | |
58 | * src/symtab.c (bucket_new): Initialize `value' to -1, not 0. | |
59 | * src/reader.c (eoftoken): New. | |
60 | (parse_token_decl): If the token being defined has value `0', it | |
61 | is the eoftoken. | |
62 | (packsymbols): No longer hack `tags' to insert `$' by hand. | |
63 | Be sure to preserve the value of the eoftoken. | |
64 | (reader): Make sure eoftoken is defined. | |
65 | Initialize nsyms to 0: now eoftoken is created just like the others. | |
66 | * src/print.c (print_grammar): Don't special case the eof token. | |
67 | * src/regression.at: Adjust: `$' has value 0, not -1, which was a | |
68 | lie anyway, albeit pleasant. | |
69 | * tests/calc.at: Exercise error messages with eoftoken. | |
70 | Change the grammar so that empty input is invalid. | |
71 | Adjust expectations. | |
72 | When yyungeting, be sure to use a valid yylloc: use last_yylloc. | |
73 | ||
ec2da99f AD |
74 | 2001-12-27 Akim Demaille <akim@epita.fr> |
75 | ||
76 | * configure.in: Check the protos of strchr ans strspn. | |
77 | Replace strchr if needed. | |
78 | * src/system.h: Provide the protos of strchr, strspn and memchr if | |
79 | missing. | |
80 | * lib/strchr.c: New. | |
81 | * src/reader.c (symbols_save): Use strchr. | |
82 | ||
8adfa272 AD |
83 | 2001-12-27 Akim Demaille <akim@epita.fr> |
84 | ||
85 | * src/print.c, src/print_graph.c (escape): New. | |
86 | Use it to quote the TAGS outputs. | |
87 | * src/print_graph.c (print_state): Now errors are in red, and | |
88 | reductions in green. | |
89 | Prefer high to wide: output the state number on a line of its own. | |
90 | ||
80dac38c AD |
91 | 2001-12-27 Akim Demaille <akim@epita.fr> |
92 | ||
93 | * src/state.h, src/state.c (reductions_new): New. | |
94 | * src/LR0.c (set_state_table): Let all the states have a | |
95 | `reductions', even if reduced to 0. | |
96 | (save_reductions): Adjust. | |
97 | * src/lalr.c (initialize_LA, initialize_lookaheads): Adjust. | |
98 | * src/print.c (print_reductions, print_actions): Adjust. | |
99 | * src/output.c (action_row): Adjust. | |
100 | ||
2cec70b9 AD |
101 | 2001-12-27 Akim Demaille <akim@epita.fr> |
102 | ||
103 | * src/state.h, src/state.c (errs_new, errs_dup): New. | |
104 | * src/LR0.c (set_state_table): Let all the states have an errs, | |
105 | even if reduced to 0. | |
106 | * src/print.c (print_errs, print_reductions): Adjust. | |
107 | * src/output.c (output_actions, action_row): Adjust. | |
108 | * src/conflicts.c (resolve_sr_conflict): Adjust. | |
109 | ||
13ca549a AD |
110 | 2001-12-27 Akim Demaille <akim@epita.fr> |
111 | ||
112 | * src/lalr.c (set_goto_map, initialize_F): Use SHIFT_SYMBOL. | |
113 | ||
5092aba5 AD |
114 | 2001-12-27 Akim Demaille <akim@epita.fr> |
115 | ||
116 | * src/conflicts.c, src/conflicts.h (print_reductions): Move to... | |
117 | * src/print.c: here. | |
118 | (lookaheadset, shiftset): New, used as additional storage by | |
119 | print_reductions. | |
120 | (print_results): Adjust. | |
121 | (print_shifts, print_gotos, print_errs): New, extracted from... | |
122 | (print_actions): here. | |
123 | * src/print_graph.c (print_actions): Remove dead code. | |
124 | ||
11e2beca AD |
125 | 2001-12-27 Akim Demaille <akim@epita.fr> |
126 | ||
127 | * src/reader.c (copy_dollar, copy_at): Better checking of `n' in | |
128 | `$n' and `@n'. | |
129 | ||
dac3c910 AD |
130 | 2001-12-27 Akim Demaille <akim@epita.fr> |
131 | ||
132 | * src/lalr.c (add_lookback_edge): Use state_t instead of ints. | |
133 | (build_relations): Adjust. | |
134 | ||
d0b0fefa AD |
135 | 2001-12-27 Akim Demaille <akim@epita.fr> |
136 | ||
137 | * src/lalr.c (set_goto_map): Remove a wrong but benign loop | |
138 | duplication. | |
139 | ||
adc8c848 AD |
140 | 2001-12-27 Akim Demaille <akim@epita.fr> |
141 | ||
142 | * src/reader.c (packgram): Catch nitems overflows. | |
143 | ||
14d293ac AD |
144 | 2001-12-27 Akim Demaille <akim@epita.fr> |
145 | ||
146 | * src/files.c, src/files.h (guard_obstack): Remove. | |
147 | * src/output.c (output): Adjust. | |
148 | * src/reader.c (parse_braces): New, factoring... | |
149 | (copy_action, copy_guard): these two which are renamed as... | |
150 | (parse_action, parse_guard): these. | |
151 | As a voluntary consequence, using braces around guards is now | |
152 | mandatory. | |
153 | ||
f499b062 AD |
154 | 2001-12-27 Akim Demaille <akim@epita.fr> |
155 | ||
156 | * src/gram.h (rule_t): `guard' and `guard_line' are new members. | |
157 | * src/reader.c (symbol_list): `guard' and `guard_line' are new | |
158 | members. | |
159 | (symbol_list_new): Adjust. | |
160 | (copy_action): action_line is the first line, not the last. | |
161 | (copy_guard): Just as for actions, store the `action' only, not | |
162 | the switch/case/break flesh. | |
163 | Don't parse the user action that might follow the guard, let... | |
164 | (readgram): do it, i.e., now, there can be an action after a | |
165 | guard. | |
166 | In other words the guard is just explicitly optional. | |
167 | (packgram): Adjust. | |
168 | * src/output.c (guards_output): New. | |
169 | (output_parser): Call it when needed. | |
170 | (output): Also free the guard and attrs obstacks. | |
171 | * src/files.c, src/files.h (obstack_save): Remove. | |
172 | (output_files): Remove. | |
173 | As a result, if one needs the former `.act' file, using an | |
174 | appropriate skeleton which requires actions and guards is now | |
175 | required. | |
176 | * src/main.c (main): Adjust. | |
177 | * tests/semantic.at: New. | |
178 | * tests/regression.at: Use `input.y' as input file name. | |
179 | Avoid 8+3 problems by requiring input.c when the test needs the | |
180 | parser. | |
181 | ||
d945f5cd AD |
182 | 2001-12-27 Akim Demaille <akim@epita.fr> |
183 | ||
184 | * src/reader.c (symbol_list_new): Be sure to initialize all the | |
185 | fields. | |
186 | ||
d200e455 AD |
187 | 2001-12-27 Akim Demaille <akim@epita.fr> |
188 | ||
189 | All the hacks using a final pseudo state are now useless. | |
190 | ||
191 | * src/LR0.c (set_state_table): state_table holds exactly nstates. | |
192 | * src/lalr.c (nLA): New. | |
193 | (initialize_LA, compute_lookaheads, initialize_lookaheads): Use it | |
194 | instead of lookaheadsp from the pseudo state (nstate + 1). | |
195 | ||
f9507c28 AD |
196 | 2001-12-27 Akim Demaille <akim@epita.fr> |
197 | ||
198 | * src/output.c (action_row, token_actions): Use a state_t instead | |
199 | of a integer, and nlookaheads instead of the following state's | |
200 | lookaheadsp. | |
201 | ||
065fbd27 AD |
202 | 2001-12-27 Akim Demaille <akim@epita.fr> |
203 | ||
204 | * src/conflicts.c (log_resolution, flush_shift) | |
205 | (resolve_sr_conflict, set_conflicts, solve_conflicts) | |
206 | (count_sr_conflicts, count_rr_conflicts, conflicts_output) | |
207 | (conflicts_print, print_reductions): Use a state_t instead of an | |
208 | integer when referring to a state. | |
209 | As much as possible, depend upon nlookaheads, instead of the | |
210 | `lookaheadsp' member of the following state (since lookaheads of | |
211 | successive states are successive, the difference between state n + 1 | |
212 | and n served as the number of lookaheads for state n). | |
213 | * src/lalr.c (add_lookback_edge): Likewise. | |
214 | * src/print.c (print_core, print_actions, print_state) | |
215 | (print_results): Likewise. | |
216 | * src/print_graph.c (print_core, print_actions, print_state) | |
217 | (print_graph): Likewise. | |
218 | * src/conflicts.h: Adjust. | |
219 | ||
1b177bd7 AD |
220 | 2001-12-27 Akim Demaille <akim@epita.fr> |
221 | ||
222 | * src/bison.hairy: Formatting/comment changes. | |
223 | ANSIfy. | |
224 | Remove `register' indications. | |
225 | Add plenty of `static'. | |
226 | ||
7742ddeb AD |
227 | 2001-12-27 Akim Demaille <akim@epita.fr> |
228 | ||
229 | * src/output.c (prepare): Drop the muscle `ntbase' which | |
230 | duplicates ntokens. | |
231 | * src/bison.simple: Formatting/comment changes. | |
232 | Use YYNTOKENS only, which is documented, but not YYNTBASE, which | |
233 | is an undocumented synonym. | |
234 | ||
1fa14068 AD |
235 | 2001-12-22 Akim Demaille <akim@epita.fr> |
236 | ||
237 | * src/output.c (output_table_data): Change the prototype to use | |
238 | `int' for array ranges: some invocations do pass an int, not a | |
239 | short. | |
240 | Reported by Wayne Green. | |
241 | ||
b9752825 AD |
242 | 2001-12-22 Akim Demaille <akim@epita.fr> |
243 | ||
244 | Some actions of web2c.y are improperly triggered. | |
245 | Reported by Mike Castle. | |
246 | ||
247 | * src/lalr.c (traverse): s/F (i)[k] = F (j)[k]/F (j)[k] = F (i)[k]/. | |
248 | * tests/regression.at (Web2c): Rename as... | |
249 | (Web2c Report): this. | |
250 | (Web2c Actions): New. | |
251 | ||
776209d6 AD |
252 | 2001-12-22 Akim Demaille <akim@epita.fr> |
253 | ||
254 | Reductions in web2c.y are improperly reported. | |
255 | Reported by Mike Castle. | |
256 | ||
257 | * src/conflicts.c (print_reductions): Fix. | |
258 | * tests/regression.at (Web2c): New. | |
259 | ||
275fc3ad AD |
260 | 2001-12-18 Akim Demaille <akim@epita.fr> |
261 | ||
262 | Some host fail on `assert (!"foo")', which expands to | |
263 | ((!"foo") ? (void)0 : __assert("!"foo."", __FILE__, __LINE__)) | |
264 | Reported by Nelson Beebee. | |
265 | ||
266 | * src/output.c, src/vcg.c: Replace `assert (!"it succeeded")' with | |
267 | `#define it_succeeded 0' and `assert (it_succeeded)'. | |
268 | ||
897668ee MA |
269 | 2001-12-17 Marc Autret <autret_m@epita.fr> |
270 | ||
271 | * src/bison.simple: Don't hard code the skeleton line and filename. | |
272 | * src/output.c (output_parser): Rename 'line' as 'output_line'. | |
273 | New line counter 'skeleton_line' (skeleton-line muscle). | |
274 | ||
ab3399e0 PE |
275 | 2001-12-17 Paul Eggert <eggert@twinsun.com> |
276 | ||
277 | * NEWS, doc/bison.texinfo, doc/bison.1, doc/bison.rnh: Document that | |
278 | YYDEBUG must be defined to a nonzero value. | |
279 | ||
280 | * src/bison.simple (yytname): Do not assume that the user defines | |
281 | YYDEBUG to a properly parenthesized expression. | |
282 | ||
3877f72b AD |
283 | 2001-12-17 Akim Demaille <akim@epita.fr> |
284 | ||
285 | * src/state.h (state_t): Rename lookaheads as lookaheadsp. | |
286 | nlookaheads is a new member. | |
287 | Adjust all users. | |
288 | * src/lalr.h (nlookaheads): Remove this orphan declaration. | |
289 | * src/lalr.c (initialize_lookaheads): Set nlookaheads for each | |
290 | state. | |
776209d6 | 291 | |
331dbc1b AD |
292 | 2001-12-17 Akim Demaille <akim@epita.fr> |
293 | ||
294 | * src/files.h, src/files.c (open_files, close_files): Remove. | |
295 | * src/main.c (main): Don't open/close files, nor invoke lex_free, | |
296 | let... | |
297 | * src/reader.c (reader): Do it. | |
776209d6 | 298 | |
be750e4c AD |
299 | 2001-12-17 Akim Demaille <akim@epita.fr> |
300 | ||
301 | * src/conflicts.c (print_reductions): Formatting changes. | |
776209d6 | 302 | |
709ae8c6 AD |
303 | 2001-12-17 Akim Demaille <akim@epita.fr> |
304 | ||
305 | * src/conflicts.c (flush_shift): Also adjust lookaheadset. | |
306 | (flush_reduce): New. | |
307 | (resolve_sr_conflict): Adjust. | |
776209d6 | 308 | |
f87685c3 AD |
309 | 2001-12-17 Akim Demaille <akim@epita.fr> |
310 | ||
311 | * src/output.c (output_obstack): Be static and rename as... | |
312 | (format_obstack): this, to avoid any confusion with files.c's | |
313 | output_obstack. | |
314 | * src/reader.h (muscle_obstack): Move to... | |
315 | * src/output.h: here, since it's defined in output.c. | |
316 | ||
837491d8 AD |
317 | 2001-12-17 Akim Demaille <akim@epita.fr> |
318 | ||
319 | * src/output.c (action_row, save_column, default_goto) | |
320 | (sort_actions, matching_state, pack_vector): Better variable | |
321 | locality. | |
322 | ||
796d61fb AD |
323 | 2001-12-17 Akim Demaille <akim@epita.fr> |
324 | ||
325 | * src/output.c: Various formatting changes. | |
776209d6 | 326 | |
64d15509 AD |
327 | 2001-12-17 Akim Demaille <akim@epita.fr> |
328 | ||
329 | * src/files.c (output_files): Free the output_obstack. | |
330 | * src/main.c (main): Call print and print_graph conditionally. | |
331 | * src/print.c (print): Work unconditionally. | |
332 | * src/print_graph.c (print_graph): Work unconditionally. | |
333 | * src/conflicts.c (log_resolution): Output only if verbose_flag. | |
334 | ||
fbc8ecb7 MA |
335 | 2001-12-16 Marc Autret <autret_m@epita.fr> |
336 | ||
337 | * src/output.c (actions_output): Fix. When we use %no-lines, | |
338 | there is one less line per action. | |
339 | ||
f0440388 MA |
340 | 2001-12-16 Marc Autret <autret_m@epita.fr> |
341 | ||
342 | * src/bison.simple: Remove a useless #line directive. | |
343 | s/#line %%line %%skeleton/#line %%line "%%parser-file-name"/'. | |
344 | * src/output.c (get_lines_number): New. | |
776209d6 | 345 | (output_parser): Adjust, now takes care about the lines of a |
f0440388 MA |
346 | output muscles. |
347 | Fix line numbering. | |
348 | (actions_output): Computes the number of lines taken by actions. | |
349 | (output_master_parser): Insert new skeleton which is the name of | |
350 | the output parser file name. | |
351 | ||
a79986b8 MA |
352 | 2001-12-15 Marc Autret <autret_m@epita.fr> |
353 | ||
354 | * src/bison.simple [YYERROR_VERBOSE]: Restore backward compatibility. | |
355 | ||
4ec8e00f MA |
356 | 2001-12-15 Marc Autret <autret_m@epita.fr> |
357 | ||
358 | * src/output.c (output_gram): Keep track of the hairy one. | |
359 | ||
1a4648ff AD |
360 | 2001-12-15 Akim Demaille <akim@epita.fr> |
361 | ||
362 | Make `make distcheck' work. | |
363 | ||
364 | * lib/Makefile.am (INCLUDES): Add top_srcdir/intl, since hash uses | |
365 | system.h which uses libgettext.h. | |
366 | ||
9c2c67e6 AD |
367 | 2001-12-15 Akim Demaille <akim@epita.fr> |
368 | ||
369 | * src/nullable.c (set_nullable): Useless rules must be skipped, | |
370 | otherwise, since we range over their symbols, we might look at a | |
371 | nonterminal which no longer ``exists'', i.e., it is not counted in | |
372 | `nvars', hence we overflow our arrays. | |
373 | ||
93ede233 AD |
374 | 2001-12-15 Akim Demaille <akim@epita.fr> |
375 | ||
376 | The header can also be produced directly, without any obstack! | |
377 | Yahoo! | |
378 | ||
379 | * src/files.c, src/files.h (defines_obstack): Remove. | |
380 | (compute_header_macro): Global. | |
381 | (defines_obstack_save): Remove. | |
382 | * src/reader.c (parse_union_decl): No longer output to | |
383 | defines_obstack: its content can be found in the `stype' muscle | |
384 | anyway. | |
385 | (output_token_translations): Merge into... | |
386 | (symbols_output): this. | |
387 | Rename as... | |
388 | (symbols_save): this. | |
389 | (reader): Adjust. | |
390 | * src/output.c (header_output): New. | |
391 | (output): Call it. | |
392 | ||
2666f928 AD |
393 | 2001-12-15 Akim Demaille <akim@epita.fr> |
394 | ||
395 | * src/reader.c (parse_union_decl): Instead of handling two obstack | |
396 | simultaneously, use one to define the `stype' muscle, and use the | |
397 | value of the latter to fill defines_obstack. | |
398 | (copy_comment): Remove. | |
399 | (copy_comment2): Work for a single obstack. | |
400 | Rename as... | |
401 | (copy_comment): this. | |
402 | ||
428046f8 AD |
403 | 2001-12-15 Akim Demaille <akim@epita.fr> |
404 | ||
405 | * src/lex.c, src/lex.h (xgetc): No longer static. | |
406 | * src/reader.c (parse_union_decl): Revamp. | |
407 | ||
ea52d706 AD |
408 | 2001-12-15 Akim Demaille <akim@epita.fr> |
409 | ||
410 | Still making progress in separating Bison into (i) input, (ii) | |
411 | process, (iii) output: now we can directly output the parser file | |
412 | without using table_obstack at all. | |
413 | ||
414 | * src/files.c, src/files.h (table_obstack): Bye bye. | |
415 | (parser_file_name): New. | |
416 | * src/files.c (compute_output_file_names): Compute it. | |
417 | * src/output.c (actions_output, output_parser) | |
418 | (output_master_parser): To a file instead of an obstack. | |
419 | ||
3f96f4dc AD |
420 | 2001-12-15 Akim Demaille <akim@epita.fr> |
421 | ||
422 | Attach actions to rules, instead of pre-outputting them to | |
423 | actions_obstack. | |
424 | ||
425 | * src/gram.h (rule_t): action and action_line are new members. | |
426 | * src/reader.c (symbol_list): Likewise. | |
427 | (copy_action): Save the actions within the rule. | |
428 | (packgram): Save them in rule_table. | |
429 | * src/output.c (actions_output): New. | |
430 | (output_parser): Use it on `%%actions'. | |
431 | (output_rule_data): Don't free rule_table. | |
432 | (output): Do it. | |
433 | (prepare): Don't save the `action' muscle. | |
434 | * src/bison.simple: s/%%action/%%actions/. | |
435 | ||
51576fb3 AD |
436 | 2001-12-15 Akim Demaille <akim@epita.fr> |
437 | ||
438 | * src/reader.c (copy_action): When --yacc, don't append a `;' | |
439 | to the user action: let it fail if lacking. | |
dee049eb | 440 | Suggested by Arnold Robbins and Tom Tromey. |
51576fb3 | 441 | |
2648a72d AD |
442 | 2001-12-14 Akim Demaille <akim@epita.fr> |
443 | ||
444 | * src/lex.c (literalchar): Simply return the char you decoded, non | |
445 | longer mess around with obstacks and int pointers. | |
446 | Adjust all callers. | |
447 | ||
92790e5b AD |
448 | 2001-12-14 Akim Demaille <akim@epita.fr> |
449 | ||
450 | * src/lex.c (literalchar): Don't escape the special characters, | |
451 | just decode them, and keep them as char (before, eol was output as | |
452 | the 2 char string `\n' etc.). | |
453 | * src/output.c (output_rule_data): Use quotearg to output the | |
454 | token strings. | |
455 | ||
927c1557 PE |
456 | 2001-12-13 Paul Eggert <eggert@twinsun.com> |
457 | ||
458 | * src/bison.simple (YYSIZE_T, YYSTACK_ALLOC, YYSTACK_FREE): | |
459 | Do not infringe on the global user namespace when using C++. | |
460 | (YYFPRINTF, YYSTDERR): New macros, needed for the above. | |
461 | All uses of `fprintf' and `stderr' changed. | |
462 | ||
463 | * doc/bison.texinfo: Document YYFPRINTF, YYSTDERR. | |
464 | ||
ed8e1f68 AD |
465 | 2001-12-13 Akim Demaille <akim@epita.fr> |
466 | ||
467 | The computation of nullable is broken: it doesn't handle empty | |
468 | RHS's properly. | |
469 | ||
470 | * tests/torture.at (GNU AWK Grammar): New. | |
471 | * tests/sets.at (Nullable): New. | |
472 | * src/nullable.c (set_nullable): Instead of blindly looping over | |
473 | `ritems', loop over the rules, and then over their rhs's. | |
474 | ||
475 | Work around Autotest bugs. | |
476 | ||
477 | * src/warshall.c (bitmatrix_print): Don't use `+--+' as table | |
478 | frame, because Autotest understand lines starting with a `+' as | |
479 | traces from the shell. Then, they are not processed properly. | |
480 | Admittedly an Autotest bug, but we don't have time to wait for | |
481 | Autotest to catch up. | |
482 | * tests/regression.at (Broken Closure): Adjust to the new table | |
483 | frames. | |
484 | Move to... | |
485 | * tests/sets.at: here. | |
486 | ||
cb581495 AD |
487 | 2001-12-13 Akim Demaille <akim@epita.fr> |
488 | ||
489 | * src/closure.c (closure): Use nrules instead of playing tricks | |
490 | with BITS_PER_WORD. | |
491 | ||
2e729273 AD |
492 | 2001-12-13 Akim Demaille <akim@epita.fr> |
493 | ||
494 | * src/print.c (print_actions): Output the handling of `$' as the | |
495 | traces do: shifting the token EOF. Before EOF was treated as a | |
496 | nonterminal. | |
497 | * tests/regression.at: Adjust some tests. | |
498 | * src/print_graph.c (print_core): Complete the set of items via | |
499 | closure. The next-to-final and final states are still unsatisfying, | |
500 | but that's to be addressed elsewhere. | |
501 | No longer output the rule numbers, but do output the state number. | |
502 | A single loop for the shifts + gotos is enough, but picked a | |
503 | distinct color for each. | |
504 | (print_graph): Initialize and finalize closure. | |
505 | ||
107f7dfb AD |
506 | 2001-12-13 Akim Demaille <akim@epita.fr> |
507 | ||
508 | * src/reader.c (readgram): Remove dead code, an strip useless | |
509 | braces. | |
510 | (get_type): Remove, unused. | |
511 | ||
9b53a24f AD |
512 | 2001-12-12 Akim Demaille <akim@epita.fr> |
513 | ||
514 | * src/complain.h, src/complain.c: Remove error_one_per_line, rely | |
515 | on that of lib/error.c. | |
516 | ||
dbfb6dcd AD |
517 | 2001-12-12 Akim Demaille <akim@epita.fr> |
518 | ||
519 | Some hosts don't like `/' in includes. | |
520 | ||
521 | * src/system.h: Include libgettext.h without qualifying the path. | |
522 | * src/Makefile.am (INCLUDES): Add $(top_srcdir)/intl, remove | |
523 | $(top_srcdir). | |
524 | ||
c25fb648 MA |
525 | 2001-12-11 Marc Autret <autret_m@epita.fr> |
526 | ||
527 | * src/output.c (output_parser): Remove useless muscle. | |
528 | ||
710ddc4f MA |
529 | 2001-12-11 Marc Autret <autret_m@epita.fr> |
530 | ||
531 | * src/bison.simple: Remove #line just before %%epilogue. It | |
532 | is now handled in ... | |
533 | * src/reader.c (read_additionnal_code): Add the output of a | |
534 | #line for the epilogue. | |
535 | ||
e83d80b8 MA |
536 | 2001-12-10 Marc Autret <autret_m@epita.fr> |
537 | ||
927c1557 | 538 | * src/reader.c (copy_definition): Re-use CPP-outed code which |
e83d80b8 MA |
539 | replace precedent remove. |
540 | * src/bison.simple: Remove #line before %%prologue because | |
541 | %%input-line is wrong at this time. | |
542 | ||
971d5158 MA |
543 | 2001-12-10 Marc Autret <autret_m@epita.fr> |
544 | ||
545 | * src/reader.c (symbols_output): Clean up. | |
927c1557 | 546 | * src/output.c (output_gram, output): Clean up. |
971d5158 | 547 | |
5edafffd AD |
548 | 2001-12-10 Akim Demaille <akim@epita.fr> |
549 | ||
550 | * src/lalr.c (initialize_lookaheads): New. Extracted from... | |
551 | * src/LR0.c (set_state_table): here. | |
552 | * src/lalr.c (lalr): Call it. | |
553 | ||
0279f8e9 AD |
554 | 2001-12-10 Akim Demaille <akim@epita.fr> |
555 | ||
556 | * src/state.h (shifts): Remove the `number' member: shifts are | |
557 | attached to state, hence no longer need to be labelled with a | |
558 | state number. | |
559 | ||
190c4f5f AD |
560 | 2001-12-10 Akim Demaille <akim@epita.fr> |
561 | ||
562 | Now that states have a complete set of members, the linked list of | |
563 | shifts is useless: just fill directly the state's shifts member. | |
564 | ||
565 | * src/state.h (shifts): Remove the `next' member. | |
566 | * src/LR0.c (first_state, last_state): Remove. | |
567 | Adjust the callers. | |
568 | (augment_automaton): Don't look for the shifts that must be added | |
569 | a shift on EOF: it is those of the state we looked for! But now, | |
570 | since shifts are attached, it is no longer needed to looking | |
571 | merely by its id: its number. | |
572 | ||
2a73b93d AD |
573 | 2001-12-10 Akim Demaille <akim@epita.fr> |
574 | ||
575 | * src/LR0.c (augment_automaton): Better variable locality. | |
576 | Remove an impossible branch: if there is a state corresponding to | |
577 | the start symbol being shifted, then there is shift for the start | |
578 | symbol from the initial state. | |
579 | ||
74392f6a AD |
580 | 2001-12-10 Akim Demaille <akim@epita.fr> |
581 | ||
582 | * src/LR0.c (augment_automaton): Call `insert_eof_shifting_state' | |
583 | only when appropriate: when insert_start_shifting_state' is not | |
584 | invoked. | |
585 | * tests/regression.at (Rule Line Numbers): Adjust. | |
586 | ||
37c82725 AD |
587 | 2001-12-10 Akim Demaille <akim@epita.fr> |
588 | ||
589 | * src/LR0.c (augment_automaton): Now that all states have shifts, | |
590 | merge the two cases addition shifts to the initial state. | |
591 | ||
6a164e0c AD |
592 | 2001-12-10 Akim Demaille <akim@epita.fr> |
593 | ||
594 | * src/lalr.c (set_state_table): Move to... | |
595 | * src/LR0.c: here. | |
596 | * src/lalr.c (lalr): Don't call it... | |
597 | * src/LR0.c (generate_states): do it. | |
598 | * src/LR0.h (first_state): Remove, only the table is used. | |
599 | ||
7215de24 AD |
600 | 2001-12-10 Akim Demaille <akim@epita.fr> |
601 | ||
602 | * src/LR0.h (first_shift, first_reduction): Remove. | |
603 | * src/lalr.c: Don't use first_shift: find shifts through the | |
604 | states. | |
605 | ||
80e25d4d AD |
606 | 2001-12-10 Akim Demaille <akim@epita.fr> |
607 | ||
608 | * src/LR0.c: Attach shifts to states as soon as they are | |
609 | computed. | |
610 | * src/lalr.c (set_state_table): Instead of assigning shifts to | |
611 | state, just assert that the mapping was properly done. | |
612 | ||
0ab3728b AD |
613 | 2001-12-10 Akim Demaille <akim@epita.fr> |
614 | ||
615 | * src/LR0.c (insert_start_shift): Rename as... | |
616 | (insert_start_shifting_state): this. | |
617 | (insert_eof_shifting_state, insert_accepting_state): New. | |
618 | (augment_automaton): Adjust. | |
619 | Better locality of the variables. | |
620 | When looking if the start_symbol is shifted from the initial | |
621 | state, using `while (... symbol != start_symbol ...)' sounds | |
622 | better than `while (... symbol < start_symbol ...)': If fail | |
623 | to see how the order between symbols could be relevant! | |
624 | ||
78af9bbc AD |
625 | 2001-12-10 Akim Demaille <akim@epita.fr> |
626 | ||
627 | * src/getargs.h: Don't declare `spec_name_prefix' and | |
628 | `spec_file_prefix', declared by src/files.h. | |
629 | * src/files.c, src/files.h: Default for spec_name_prefix is "yy". | |
630 | * src/muscle_tab.c (muscle_init): Default prefix to NULL. | |
631 | * src/output.c (prepare): Adjust. | |
632 | * src/reader.c (symbols_output): Likewise. | |
633 | * src/vmsgetargs.c: Vaguely adjust, but who cares? | |
634 | ||
bdef2a41 AD |
635 | 2001-12-10 Akim Demaille <akim@epita.fr> |
636 | ||
637 | * src/muscle_tab.c (muscle_init): NULL is a better default than | |
638 | `"0"'. | |
639 | ||
3735969c AD |
640 | 2001-12-10 Akim Demaille <akim@epita.fr> |
641 | ||
642 | * src/reader.c (reader): Calling symbols_output once is enough. | |
643 | ||
49701457 AD |
644 | 2001-12-10 Akim Demaille <akim@epita.fr> |
645 | ||
646 | Now that states have a complete set of members, the linked list of | |
647 | reductions is useless: just fill directly the state's reductions | |
648 | member. | |
649 | ||
650 | * src/state.h (struct reductions): Remove member `number' and | |
651 | `next'. | |
652 | * src/LR0.c (first_reduction, last_reduction): Remove. | |
653 | (save_reductions): Don't link the new reductions, store them in | |
654 | this_state. | |
655 | * src/lalr.c (set_state_table): No need to attach reductions to | |
656 | states, it's already done. | |
657 | * src/output.c (output_actions): No longer free the shifts, then | |
658 | the reductions, then the states: free all the states and their | |
659 | members. | |
660 | ||
0edad749 AD |
661 | 2001-12-10 Akim Demaille <akim@epita.fr> |
662 | ||
663 | * src/options.c (OPTN, DRTV, BOTH): New. | |
664 | (option_table): Use them. | |
665 | ||
0edad749 AD |
666 | * src/muscle_tab.c: Don't include xalloc.h and string.h: that's |
667 | the job of system.h. | |
668 | * src/options.c: Don't include stdio.h and xalloc.h for the same | |
669 | reasons. | |
670 | ||
5449dd0f AD |
671 | 2001-12-10 Akim Demaille <akim@epita.fr> |
672 | ||
673 | * src/output.c (output, prepare): Make sure the values of the | |
674 | muscles `action' and `prologue' are 0-terminated. | |
675 | ||
a870c567 AD |
676 | 2001-12-10 Akim Demaille <akim@epita.fr> |
677 | ||
678 | Clean up GCC warnings. | |
679 | ||
680 | * src/reader.c (copy_action): `buf' is not used. | |
681 | (parse_skel_decl): Be static. | |
682 | * src/muscle_tab.c (mhash1, mhash2, muscle_insert): Preserve `const'. | |
683 | * src/options.h (create_long_option_table): Have a real prototype. | |
684 | * lib/hash.c, lib/hash.h (hash_insert, hash_insert_at, hash_delete) | |
685 | (hash_delete_at): Return const void *. | |
686 | Adjust casts to preserve the const. | |
687 | ||
80df8768 AD |
688 | 2001-12-10 Akim Demaille <akim@epita.fr> |
689 | ||
690 | * configure.in: Require 2.52g. | |
691 | M4 is not needed, but AUTOM4TE is. | |
692 | * m4/m4.m4: Remove. | |
693 | * tests/Makefile.am: Adjust. | |
694 | ||
f693ad14 AD |
695 | 2001-12-10 Akim Demaille <akim@epita.fr> |
696 | ||
697 | One structure for states is enough, even though theoretically | |
698 | there are LR(0) states and LALR(1) states. | |
699 | ||
700 | * src/lalr.h (state_t): Remove. | |
701 | (state_table): Be state_t **, not state_t *. | |
702 | * src/state.h (core, CORE_ALLOC): Rename as... | |
703 | (state_t, STATE_ALLOC): this. | |
704 | Add the LALR(1) members: shifts, reductions, errs. | |
705 | * src/LR0.c (state_table): Rename as... | |
706 | (state_hash): this, to avoid name clashes with the global | |
707 | `state_table'. | |
708 | * src/print_graph.c, src/LR0.c, src/LR0.h, src/conflicts.c | |
709 | * src/lalr.c, src/lalr.h, src/output.c, src/print.c: Adjust. | |
710 | ||
74ffbcb6 AD |
711 | 2001-12-10 Akim Demaille <akim@epita.fr> |
712 | ||
713 | Bison dumps core on bash.y. | |
714 | Reported by Pascal Bart. | |
715 | ||
716 | * src/warshall.c (bitmatrix_print): New. | |
717 | (TC): Use it. | |
718 | When performing a transitive closure R(i, j) && R(j, k) => R(i, k), | |
719 | j must be the outer loop. | |
720 | * tests/regression.at (Broken Closure): New. | |
721 | ||
07708e19 AD |
722 | 2001-12-05 Akim Demaille <akim@epita.fr> |
723 | ||
724 | * tests/atlocal.in (CPPFLAGS): Do not leave a space between -I and | |
725 | its argument. | |
726 |