]>
Commit | Line | Data |
---|---|---|
1 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
2 | ||
3 | * src/conflicts.c (flush_shift): Also adjust lookaheadset. | |
4 | (flush_reduce): New. | |
5 | (resolve_sr_conflict): Adjust. | |
6 | ||
7 | ||
8 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
9 | ||
10 | * src/output.c (output_obstack): Be static and rename as... | |
11 | (format_obstack): this, to avoid any confusion with files.c's | |
12 | output_obstack. | |
13 | * src/reader.h (muscle_obstack): Move to... | |
14 | * src/output.h: here, since it's defined in output.c. | |
15 | ||
16 | ||
17 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
18 | ||
19 | * src/output.c (action_row, save_column, default_goto) | |
20 | (sort_actions, matching_state, pack_vector): Better variable | |
21 | locality. | |
22 | ||
23 | ||
24 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
25 | ||
26 | * src/output.c: Various formatting changes. | |
27 | ||
28 | ||
29 | 2001-12-17 Akim Demaille <akim@epita.fr> | |
30 | ||
31 | * src/files.c (output_files): Free the output_obstack. | |
32 | * src/main.c (main): Call print and print_graph conditionally. | |
33 | * src/print.c (print): Work unconditionally. | |
34 | * src/print_graph.c (print_graph): Work unconditionally. | |
35 | * src/conflicts.c (log_resolution): Output only if verbose_flag. | |
36 | ||
37 | 2001-12-16 Marc Autret <autret_m@epita.fr> | |
38 | ||
39 | * src/output.c (actions_output): Fix. When we use %no-lines, | |
40 | there is one less line per action. | |
41 | ||
42 | 2001-12-16 Marc Autret <autret_m@epita.fr> | |
43 | ||
44 | * src/bison.simple: Remove a useless #line directive. | |
45 | s/#line %%line %%skeleton/#line %%line "%%parser-file-name"/'. | |
46 | * src/output.c (get_lines_number): New. | |
47 | (output_parser): Adjust, now takes care about the lines of a | |
48 | output muscles. | |
49 | Fix line numbering. | |
50 | (actions_output): Computes the number of lines taken by actions. | |
51 | (output_master_parser): Insert new skeleton which is the name of | |
52 | the output parser file name. | |
53 | ||
54 | 2001-12-15 Marc Autret <autret_m@epita.fr> | |
55 | ||
56 | * src/bison.simple [YYERROR_VERBOSE]: Restore backward compatibility. | |
57 | ||
58 | 2001-12-15 Marc Autret <autret_m@epita.fr> | |
59 | ||
60 | * src/output.c (output_gram): Keep track of the hairy one. | |
61 | ||
62 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
63 | ||
64 | Make `make distcheck' work. | |
65 | ||
66 | * lib/Makefile.am (INCLUDES): Add top_srcdir/intl, since hash uses | |
67 | system.h which uses libgettext.h. | |
68 | ||
69 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
70 | ||
71 | * src/nullable.c (set_nullable): Useless rules must be skipped, | |
72 | otherwise, since we range over their symbols, we might look at a | |
73 | nonterminal which no longer ``exists'', i.e., it is not counted in | |
74 | `nvars', hence we overflow our arrays. | |
75 | ||
76 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
77 | ||
78 | The header can also be produced directly, without any obstack! | |
79 | Yahoo! | |
80 | ||
81 | * src/files.c, src/files.h (defines_obstack): Remove. | |
82 | (compute_header_macro): Global. | |
83 | (defines_obstack_save): Remove. | |
84 | * src/reader.c (parse_union_decl): No longer output to | |
85 | defines_obstack: its content can be found in the `stype' muscle | |
86 | anyway. | |
87 | (output_token_translations): Merge into... | |
88 | (symbols_output): this. | |
89 | Rename as... | |
90 | (symbols_save): this. | |
91 | (reader): Adjust. | |
92 | * src/output.c (header_output): New. | |
93 | (output): Call it. | |
94 | ||
95 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
96 | ||
97 | * src/reader.c (parse_union_decl): Instead of handling two obstack | |
98 | simultaneously, use one to define the `stype' muscle, and use the | |
99 | value of the latter to fill defines_obstack. | |
100 | (copy_comment): Remove. | |
101 | (copy_comment2): Work for a single obstack. | |
102 | Rename as... | |
103 | (copy_comment): this. | |
104 | ||
105 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
106 | ||
107 | * src/lex.c, src/lex.h (xgetc): No longer static. | |
108 | * src/reader.c (parse_union_decl): Revamp. | |
109 | ||
110 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
111 | ||
112 | Still making progress in separating Bison into (i) input, (ii) | |
113 | process, (iii) output: now we can directly output the parser file | |
114 | without using table_obstack at all. | |
115 | ||
116 | * src/files.c, src/files.h (table_obstack): Bye bye. | |
117 | (parser_file_name): New. | |
118 | * src/files.c (compute_output_file_names): Compute it. | |
119 | * src/output.c (actions_output, output_parser) | |
120 | (output_master_parser): To a file instead of an obstack. | |
121 | ||
122 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
123 | ||
124 | Attach actions to rules, instead of pre-outputting them to | |
125 | actions_obstack. | |
126 | ||
127 | * src/gram.h (rule_t): action and action_line are new members. | |
128 | * src/reader.c (symbol_list): Likewise. | |
129 | (copy_action): Save the actions within the rule. | |
130 | (packgram): Save them in rule_table. | |
131 | * src/output.c (actions_output): New. | |
132 | (output_parser): Use it on `%%actions'. | |
133 | (output_rule_data): Don't free rule_table. | |
134 | (output): Do it. | |
135 | (prepare): Don't save the `action' muscle. | |
136 | * src/bison.simple: s/%%action/%%actions/. | |
137 | ||
138 | 2001-12-15 Akim Demaille <akim@epita.fr> | |
139 | ||
140 | * src/reader.c (copy_action): When --yacc, don't append a `;' | |
141 | to the user action: let it fail if lacking. | |
142 | Suggested by Aharon Robbins and Tom Tromey. | |
143 | ||
144 | 2001-12-14 Akim Demaille <akim@epita.fr> | |
145 | ||
146 | * src/lex.c (literalchar): Simply return the char you decoded, non | |
147 | longer mess around with obstacks and int pointers. | |
148 | Adjust all callers. | |
149 | ||
150 | 2001-12-14 Akim Demaille <akim@epita.fr> | |
151 | ||
152 | * src/lex.c (literalchar): Don't escape the special characters, | |
153 | just decode them, and keep them as char (before, eol was output as | |
154 | the 2 char string `\n' etc.). | |
155 | * src/output.c (output_rule_data): Use quotearg to output the | |
156 | token strings. | |
157 | ||
158 | 2001-12-13 Paul Eggert <eggert@twinsun.com> | |
159 | ||
160 | * src/bison.simple (YYSIZE_T, YYSTACK_ALLOC, YYSTACK_FREE): | |
161 | Do not infringe on the global user namespace when using C++. | |
162 | (YYFPRINTF, YYSTDERR): New macros, needed for the above. | |
163 | All uses of `fprintf' and `stderr' changed. | |
164 | ||
165 | * doc/bison.texinfo: Document YYFPRINTF, YYSTDERR. | |
166 | ||
167 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
168 | ||
169 | The computation of nullable is broken: it doesn't handle empty | |
170 | RHS's properly. | |
171 | ||
172 | * tests/torture.at (GNU AWK Grammar): New. | |
173 | * tests/sets.at (Nullable): New. | |
174 | * src/nullable.c (set_nullable): Instead of blindly looping over | |
175 | `ritems', loop over the rules, and then over their rhs's. | |
176 | ||
177 | Work around Autotest bugs. | |
178 | ||
179 | * src/warshall.c (bitmatrix_print): Don't use `+--+' as table | |
180 | frame, because Autotest understand lines starting with a `+' as | |
181 | traces from the shell. Then, they are not processed properly. | |
182 | Admittedly an Autotest bug, but we don't have time to wait for | |
183 | Autotest to catch up. | |
184 | * tests/regression.at (Broken Closure): Adjust to the new table | |
185 | frames. | |
186 | Move to... | |
187 | * tests/sets.at: here. | |
188 | ||
189 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
190 | ||
191 | * src/closure.c (closure): Use nrules instead of playing tricks | |
192 | with BITS_PER_WORD. | |
193 | ||
194 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
195 | ||
196 | * src/print.c (print_actions): Output the handling of `$' as the | |
197 | traces do: shifting the token EOF. Before EOF was treated as a | |
198 | nonterminal. | |
199 | * tests/regression.at: Adjust some tests. | |
200 | * src/print_graph.c (print_core): Complete the set of items via | |
201 | closure. The next-to-final and final states are still unsatisfying, | |
202 | but that's to be addressed elsewhere. | |
203 | No longer output the rule numbers, but do output the state number. | |
204 | A single loop for the shifts + gotos is enough, but picked a | |
205 | distinct color for each. | |
206 | (print_graph): Initialize and finalize closure. | |
207 | ||
208 | 2001-12-13 Akim Demaille <akim@epita.fr> | |
209 | ||
210 | * src/reader.c (readgram): Remove dead code, an strip useless | |
211 | braces. | |
212 | (get_type): Remove, unused. | |
213 | ||
214 | 2001-12-12 Akim Demaille <akim@epita.fr> | |
215 | ||
216 | * src/complain.h, src/complain.c: Remove error_one_per_line, rely | |
217 | on that of lib/error.c. | |
218 | ||
219 | 2001-12-12 Akim Demaille <akim@epita.fr> | |
220 | ||
221 | Some hosts don't like `/' in includes. | |
222 | ||
223 | * src/system.h: Include libgettext.h without qualifying the path. | |
224 | * src/Makefile.am (INCLUDES): Add $(top_srcdir)/intl, remove | |
225 | $(top_srcdir). | |
226 | ||
227 | 2001-12-11 Marc Autret <autret_m@epita.fr> | |
228 | ||
229 | * src/output.c (output_parser): Remove useless muscle. | |
230 | ||
231 | 2001-12-11 Marc Autret <autret_m@epita.fr> | |
232 | ||
233 | * src/bison.simple: Remove #line just before %%epilogue. It | |
234 | is now handled in ... | |
235 | * src/reader.c (read_additionnal_code): Add the output of a | |
236 | #line for the epilogue. | |
237 | ||
238 | 2001-12-10 Marc Autret <autret_m@epita.fr> | |
239 | ||
240 | * src/reader.c (copy_definition): Re-use CPP-outed code which | |
241 | replace precedent remove. | |
242 | * src/bison.simple: Remove #line before %%prologue because | |
243 | %%input-line is wrong at this time. | |
244 | ||
245 | 2001-12-10 Marc Autret <autret_m@epita.fr> | |
246 | ||
247 | * src/reader.c (symbols_output): Clean up. | |
248 | * src/output.c (output_gram, output): Clean up. | |
249 | ||
250 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
251 | ||
252 | * src/lalr.c (initialize_lookaheads): New. Extracted from... | |
253 | * src/LR0.c (set_state_table): here. | |
254 | * src/lalr.c (lalr): Call it. | |
255 | ||
256 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
257 | ||
258 | * src/state.h (shifts): Remove the `number' member: shifts are | |
259 | attached to state, hence no longer need to be labelled with a | |
260 | state number. | |
261 | ||
262 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
263 | ||
264 | Now that states have a complete set of members, the linked list of | |
265 | shifts is useless: just fill directly the state's shifts member. | |
266 | ||
267 | * src/state.h (shifts): Remove the `next' member. | |
268 | * src/LR0.c (first_state, last_state): Remove. | |
269 | Adjust the callers. | |
270 | (augment_automaton): Don't look for the shifts that must be added | |
271 | a shift on EOF: it is those of the state we looked for! But now, | |
272 | since shifts are attached, it is no longer needed to looking | |
273 | merely by its id: its number. | |
274 | ||
275 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
276 | ||
277 | * src/LR0.c (augment_automaton): Better variable locality. | |
278 | Remove an impossible branch: if there is a state corresponding to | |
279 | the start symbol being shifted, then there is shift for the start | |
280 | symbol from the initial state. | |
281 | ||
282 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
283 | ||
284 | * src/LR0.c (augment_automaton): Call `insert_eof_shifting_state' | |
285 | only when appropriate: when insert_start_shifting_state' is not | |
286 | invoked. | |
287 | * tests/regression.at (Rule Line Numbers): Adjust. | |
288 | ||
289 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
290 | ||
291 | * src/LR0.c (augment_automaton): Now that all states have shifts, | |
292 | merge the two cases addition shifts to the initial state. | |
293 | ||
294 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
295 | ||
296 | * src/lalr.c (set_state_table): Move to... | |
297 | * src/LR0.c: here. | |
298 | * src/lalr.c (lalr): Don't call it... | |
299 | * src/LR0.c (generate_states): do it. | |
300 | * src/LR0.h (first_state): Remove, only the table is used. | |
301 | ||
302 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
303 | ||
304 | * src/LR0.h (first_shift, first_reduction): Remove. | |
305 | * src/lalr.c: Don't use first_shift: find shifts through the | |
306 | states. | |
307 | ||
308 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
309 | ||
310 | * src/LR0.c: Attach shifts to states as soon as they are | |
311 | computed. | |
312 | * src/lalr.c (set_state_table): Instead of assigning shifts to | |
313 | state, just assert that the mapping was properly done. | |
314 | ||
315 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
316 | ||
317 | * src/LR0.c (insert_start_shift): Rename as... | |
318 | (insert_start_shifting_state): this. | |
319 | (insert_eof_shifting_state, insert_accepting_state): New. | |
320 | (augment_automaton): Adjust. | |
321 | Better locality of the variables. | |
322 | When looking if the start_symbol is shifted from the initial | |
323 | state, using `while (... symbol != start_symbol ...)' sounds | |
324 | better than `while (... symbol < start_symbol ...)': If fail | |
325 | to see how the order between symbols could be relevant! | |
326 | ||
327 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
328 | ||
329 | * src/getargs.h: Don't declare `spec_name_prefix' and | |
330 | `spec_file_prefix', declared by src/files.h. | |
331 | * src/files.c, src/files.h: Default for spec_name_prefix is "yy". | |
332 | * src/muscle_tab.c (muscle_init): Default prefix to NULL. | |
333 | * src/output.c (prepare): Adjust. | |
334 | * src/reader.c (symbols_output): Likewise. | |
335 | * src/vmsgetargs.c: Vaguely adjust, but who cares? | |
336 | ||
337 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
338 | ||
339 | * src/muscle_tab.c (muscle_init): NULL is a better default than | |
340 | `"0"'. | |
341 | ||
342 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
343 | ||
344 | * src/reader.c (reader): Calling symbols_output once is enough. | |
345 | ||
346 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
347 | ||
348 | Now that states have a complete set of members, the linked list of | |
349 | reductions is useless: just fill directly the state's reductions | |
350 | member. | |
351 | ||
352 | * src/state.h (struct reductions): Remove member `number' and | |
353 | `next'. | |
354 | * src/LR0.c (first_reduction, last_reduction): Remove. | |
355 | (save_reductions): Don't link the new reductions, store them in | |
356 | this_state. | |
357 | * src/lalr.c (set_state_table): No need to attach reductions to | |
358 | states, it's already done. | |
359 | * src/output.c (output_actions): No longer free the shifts, then | |
360 | the reductions, then the states: free all the states and their | |
361 | members. | |
362 | ||
363 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
364 | ||
365 | * src/options.c (OPTN, DRTV, BOTH): New. | |
366 | (option_table): Use them. | |
367 | ||
368 | * src/muscle_tab.c: Don't include xalloc.h and string.h: that's | |
369 | the job of system.h. | |
370 | * src/options.c: Don't include stdio.h and xalloc.h for the same | |
371 | reasons. | |
372 | ||
373 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
374 | ||
375 | * src/output.c (output, prepare): Make sure the values of the | |
376 | muscles `action' and `prologue' are 0-terminated. | |
377 | ||
378 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
379 | ||
380 | Clean up GCC warnings. | |
381 | ||
382 | * src/reader.c (copy_action): `buf' is not used. | |
383 | (parse_skel_decl): Be static. | |
384 | * src/muscle_tab.c (mhash1, mhash2, muscle_insert): Preserve `const'. | |
385 | * src/options.h (create_long_option_table): Have a real prototype. | |
386 | * lib/hash.c, lib/hash.h (hash_insert, hash_insert_at, hash_delete) | |
387 | (hash_delete_at): Return const void *. | |
388 | Adjust casts to preserve the const. | |
389 | ||
390 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
391 | ||
392 | * configure.in: Require 2.52g. | |
393 | M4 is not needed, but AUTOM4TE is. | |
394 | * m4/m4.m4: Remove. | |
395 | * tests/Makefile.am: Adjust. | |
396 | ||
397 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
398 | ||
399 | One structure for states is enough, even though theoretically | |
400 | there are LR(0) states and LALR(1) states. | |
401 | ||
402 | * src/lalr.h (state_t): Remove. | |
403 | (state_table): Be state_t **, not state_t *. | |
404 | * src/state.h (core, CORE_ALLOC): Rename as... | |
405 | (state_t, STATE_ALLOC): this. | |
406 | Add the LALR(1) members: shifts, reductions, errs. | |
407 | * src/LR0.c (state_table): Rename as... | |
408 | (state_hash): this, to avoid name clashes with the global | |
409 | `state_table'. | |
410 | * src/print_graph.c, src/LR0.c, src/LR0.h, src/conflicts.c | |
411 | * src/lalr.c, src/lalr.h, src/output.c, src/print.c: Adjust. | |
412 | ||
413 | 2001-12-10 Akim Demaille <akim@epita.fr> | |
414 | ||
415 | Bison dumps core on bash.y. | |
416 | Reported by Pascal Bart. | |
417 | ||
418 | * src/warshall.c (bitmatrix_print): New. | |
419 | (TC): Use it. | |
420 | When performing a transitive closure R(i, j) && R(j, k) => R(i, k), | |
421 | j must be the outer loop. | |
422 | * tests/regression.at (Broken Closure): New. | |
423 | ||
424 | 2001-12-05 Akim Demaille <akim@epita.fr> | |
425 | ||
426 | * tests/atlocal.in (CPPFLAGS): Do not leave a space between -I and | |
427 | its argument. | |
428 |