]>
Commit | Line | Data |
---|---|---|
1 | /* Compute lookahead criteria for Bison. | |
2 | ||
3 | Copyright (C) 1984, 1986, 1989, 2000-2010 Free Software Foundation, | |
4 | Inc. | |
5 | ||
6 | This file is part of Bison, the GNU Compiler Compiler. | |
7 | ||
8 | This program is free software: you can redistribute it and/or modify | |
9 | it under the terms of the GNU General Public License as published by | |
10 | the Free Software Foundation, either version 3 of the License, or | |
11 | (at your option) any later version. | |
12 | ||
13 | This program is distributed in the hope that it will be useful, | |
14 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
15 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
16 | GNU General Public License for more details. | |
17 | ||
18 | You should have received a copy of the GNU General Public License | |
19 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ | |
20 | ||
21 | ||
22 | /* Find which rules need lookahead in each state, and which lookahead | |
23 | tokens they accept. */ | |
24 | ||
25 | #include <config.h> | |
26 | #include "system.h" | |
27 | ||
28 | #include <bitset.h> | |
29 | #include <bitsetv.h> | |
30 | #include <quotearg.h> | |
31 | ||
32 | #include "LR0.h" | |
33 | #include "complain.h" | |
34 | #include "derives.h" | |
35 | #include "getargs.h" | |
36 | #include "gram.h" | |
37 | #include "lalr.h" | |
38 | #include "muscle-tab.h" | |
39 | #include "nullable.h" | |
40 | #include "reader.h" | |
41 | #include "relation.h" | |
42 | #include "symtab.h" | |
43 | ||
44 | goto_number *goto_map; | |
45 | goto_number ngotos; | |
46 | state_number *from_state; | |
47 | state_number *to_state; | |
48 | bitsetv goto_follows = NULL; | |
49 | ||
50 | /* Linked list of goto numbers. */ | |
51 | typedef struct goto_list | |
52 | { | |
53 | struct goto_list *next; | |
54 | goto_number value; | |
55 | } goto_list; | |
56 | ||
57 | ||
58 | /* LA is an NLA by NTOKENS matrix of bits. LA[l, i] is 1 if the rule | |
59 | LArule[l] is applicable in the appropriate state when the next | |
60 | token is symbol i. If LA[l, i] and LA[l, j] are both 1 for i != j, | |
61 | it is a conflict. */ | |
62 | ||
63 | static bitsetv LA = NULL; | |
64 | size_t nLA; | |
65 | ||
66 | ||
67 | static goto_number **includes; | |
68 | static goto_list **lookback; | |
69 | ||
70 | ||
71 | ||
72 | ||
73 | void | |
74 | set_goto_map (void) | |
75 | { | |
76 | state_number s; | |
77 | goto_number *temp_map; | |
78 | ||
79 | goto_map = xcalloc (nvars + 1, sizeof *goto_map); | |
80 | temp_map = xnmalloc (nvars + 1, sizeof *temp_map); | |
81 | ||
82 | ngotos = 0; | |
83 | for (s = 0; s < nstates; ++s) | |
84 | { | |
85 | transitions *sp = states[s]->transitions; | |
86 | int i; | |
87 | for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i) | |
88 | { | |
89 | ngotos++; | |
90 | ||
91 | /* Abort if (ngotos + 1) would overflow. */ | |
92 | aver (ngotos != GOTO_NUMBER_MAXIMUM); | |
93 | ||
94 | goto_map[TRANSITION_SYMBOL (sp, i) - ntokens]++; | |
95 | } | |
96 | } | |
97 | ||
98 | { | |
99 | goto_number k = 0; | |
100 | int i; | |
101 | for (i = ntokens; i < nsyms; i++) | |
102 | { | |
103 | temp_map[i - ntokens] = k; | |
104 | k += goto_map[i - ntokens]; | |
105 | } | |
106 | ||
107 | for (i = ntokens; i < nsyms; i++) | |
108 | goto_map[i - ntokens] = temp_map[i - ntokens]; | |
109 | ||
110 | goto_map[nsyms - ntokens] = ngotos; | |
111 | temp_map[nsyms - ntokens] = ngotos; | |
112 | } | |
113 | ||
114 | from_state = xcalloc (ngotos, sizeof *from_state); | |
115 | to_state = xcalloc (ngotos, sizeof *to_state); | |
116 | ||
117 | for (s = 0; s < nstates; ++s) | |
118 | { | |
119 | transitions *sp = states[s]->transitions; | |
120 | int i; | |
121 | for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i) | |
122 | { | |
123 | goto_number k = temp_map[TRANSITION_SYMBOL (sp, i) - ntokens]++; | |
124 | from_state[k] = s; | |
125 | to_state[k] = sp->states[i]->number; | |
126 | } | |
127 | } | |
128 | ||
129 | free (temp_map); | |
130 | } | |
131 | ||
132 | ||
133 | goto_number | |
134 | map_goto (state_number s0, symbol_number sym) | |
135 | { | |
136 | goto_number high; | |
137 | goto_number low; | |
138 | goto_number middle; | |
139 | state_number s; | |
140 | ||
141 | low = goto_map[sym - ntokens]; | |
142 | high = goto_map[sym - ntokens + 1] - 1; | |
143 | ||
144 | for (;;) | |
145 | { | |
146 | aver (low <= high); | |
147 | middle = (low + high) / 2; | |
148 | s = from_state[middle]; | |
149 | if (s == s0) | |
150 | return middle; | |
151 | else if (s < s0) | |
152 | low = middle + 1; | |
153 | else | |
154 | high = middle - 1; | |
155 | } | |
156 | } | |
157 | ||
158 | ||
159 | static void | |
160 | initialize_F (void) | |
161 | { | |
162 | goto_number **reads = xnmalloc (ngotos, sizeof *reads); | |
163 | goto_number *edge = xnmalloc (ngotos + 1, sizeof *edge); | |
164 | goto_number nedges = 0; | |
165 | ||
166 | goto_number i; | |
167 | ||
168 | goto_follows = bitsetv_create (ngotos, ntokens, BITSET_FIXED); | |
169 | ||
170 | for (i = 0; i < ngotos; i++) | |
171 | { | |
172 | state_number stateno = to_state[i]; | |
173 | transitions *sp = states[stateno]->transitions; | |
174 | ||
175 | int j; | |
176 | FOR_EACH_SHIFT (sp, j) | |
177 | bitset_set (goto_follows[i], TRANSITION_SYMBOL (sp, j)); | |
178 | ||
179 | for (; j < sp->num; j++) | |
180 | { | |
181 | symbol_number sym = TRANSITION_SYMBOL (sp, j); | |
182 | if (nullable[sym - ntokens]) | |
183 | edge[nedges++] = map_goto (stateno, sym); | |
184 | } | |
185 | ||
186 | if (nedges == 0) | |
187 | reads[i] = NULL; | |
188 | else | |
189 | { | |
190 | reads[i] = xnmalloc (nedges + 1, sizeof reads[i][0]); | |
191 | memcpy (reads[i], edge, nedges * sizeof edge[0]); | |
192 | reads[i][nedges] = END_NODE; | |
193 | nedges = 0; | |
194 | } | |
195 | } | |
196 | ||
197 | relation_digraph (reads, ngotos, &goto_follows); | |
198 | ||
199 | for (i = 0; i < ngotos; i++) | |
200 | free (reads[i]); | |
201 | ||
202 | free (reads); | |
203 | free (edge); | |
204 | } | |
205 | ||
206 | ||
207 | static void | |
208 | add_lookback_edge (state *s, rule *r, goto_number gotono) | |
209 | { | |
210 | int ri = state_reduction_find (s, r); | |
211 | goto_list *sp = xmalloc (sizeof *sp); | |
212 | sp->next = lookback[(s->reductions->lookahead_tokens - LA) + ri]; | |
213 | sp->value = gotono; | |
214 | lookback[(s->reductions->lookahead_tokens - LA) + ri] = sp; | |
215 | } | |
216 | ||
217 | ||
218 | ||
219 | static void | |
220 | build_relations (void) | |
221 | { | |
222 | goto_number *edge = xnmalloc (ngotos + 1, sizeof *edge); | |
223 | state_number *states1 = xnmalloc (ritem_longest_rhs () + 1, sizeof *states1); | |
224 | goto_number i; | |
225 | ||
226 | includes = xnmalloc (ngotos, sizeof *includes); | |
227 | ||
228 | for (i = 0; i < ngotos; i++) | |
229 | { | |
230 | int nedges = 0; | |
231 | symbol_number symbol1 = states[to_state[i]]->accessing_symbol; | |
232 | rule **rulep; | |
233 | ||
234 | for (rulep = derives[symbol1 - ntokens]; *rulep; rulep++) | |
235 | { | |
236 | bool done; | |
237 | int length = 1; | |
238 | item_number const *rp; | |
239 | state *s = states[from_state[i]]; | |
240 | states1[0] = s->number; | |
241 | ||
242 | for (rp = (*rulep)->rhs; ! item_number_is_rule_number (*rp); rp++) | |
243 | { | |
244 | s = transitions_to (s->transitions, | |
245 | item_number_as_symbol_number (*rp)); | |
246 | states1[length++] = s->number; | |
247 | } | |
248 | ||
249 | if (!s->consistent) | |
250 | add_lookback_edge (s, *rulep, i); | |
251 | ||
252 | length--; | |
253 | done = false; | |
254 | while (!done) | |
255 | { | |
256 | done = true; | |
257 | /* Each rhs ends in a rule number, and there is a | |
258 | sentinel (ritem[-1]=0) before the first rhs, so it is safe to | |
259 | decrement RP here. */ | |
260 | rp--; | |
261 | if (ISVAR (*rp)) | |
262 | { | |
263 | /* Downcasting from item_number to symbol_number. */ | |
264 | edge[nedges++] = map_goto (states1[--length], | |
265 | item_number_as_symbol_number (*rp)); | |
266 | if (nullable[*rp - ntokens]) | |
267 | done = false; | |
268 | } | |
269 | } | |
270 | } | |
271 | ||
272 | if (nedges == 0) | |
273 | includes[i] = NULL; | |
274 | else | |
275 | { | |
276 | int j; | |
277 | includes[i] = xnmalloc (nedges + 1, sizeof includes[i][0]); | |
278 | for (j = 0; j < nedges; j++) | |
279 | includes[i][j] = edge[j]; | |
280 | includes[i][nedges] = END_NODE; | |
281 | } | |
282 | } | |
283 | ||
284 | free (edge); | |
285 | free (states1); | |
286 | ||
287 | relation_transpose (&includes, ngotos); | |
288 | } | |
289 | ||
290 | ||
291 | ||
292 | static void | |
293 | compute_FOLLOWS (void) | |
294 | { | |
295 | goto_number i; | |
296 | ||
297 | relation_digraph (includes, ngotos, &goto_follows); | |
298 | ||
299 | for (i = 0; i < ngotos; i++) | |
300 | free (includes[i]); | |
301 | ||
302 | free (includes); | |
303 | } | |
304 | ||
305 | ||
306 | static void | |
307 | compute_lookahead_tokens (void) | |
308 | { | |
309 | size_t i; | |
310 | goto_list *sp; | |
311 | ||
312 | for (i = 0; i < nLA; i++) | |
313 | for (sp = lookback[i]; sp; sp = sp->next) | |
314 | bitset_or (LA[i], LA[i], goto_follows[sp->value]); | |
315 | ||
316 | /* Free LOOKBACK. */ | |
317 | for (i = 0; i < nLA; i++) | |
318 | LIST_FREE (goto_list, lookback[i]); | |
319 | ||
320 | free (lookback); | |
321 | } | |
322 | ||
323 | ||
324 | /*----------------------------------------------------. | |
325 | | Count the number of lookahead tokens required for S | | |
326 | | (N_LOOKAHEAD_TOKENS member). | | |
327 | `----------------------------------------------------*/ | |
328 | ||
329 | static int | |
330 | state_lookahead_tokens_count (state *s, bool default_reduction_only_for_accept) | |
331 | { | |
332 | int n_lookahead_tokens = 0; | |
333 | reductions *rp = s->reductions; | |
334 | transitions *sp = s->transitions; | |
335 | ||
336 | /* Transitions are only disabled during conflict resolution, and that | |
337 | hasn't happened yet, so there should be no need to check that | |
338 | transition 0 hasn't been disabled before checking if it is a shift. | |
339 | However, this check was performed at one time, so we leave it as an | |
340 | aver. */ | |
341 | aver (sp->num == 0 || !TRANSITION_IS_DISABLED (sp, 0)); | |
342 | ||
343 | /* We need a lookahead either to distinguish different reductions | |
344 | (i.e., there are two or more), or to distinguish a reduction from a | |
345 | shift. Otherwise, it is straightforward, and the state is | |
346 | `consistent'. However, do not treat a state with any reductions as | |
347 | consistent unless it is the accepting state (because there is never | |
348 | a lookahead token that makes sense there, and so no lookahead token | |
349 | should be read) if the user has otherwise disabled default | |
350 | reductions. */ | |
351 | if (rp->num > 1 | |
352 | || (rp->num == 1 && sp->num && TRANSITION_IS_SHIFT (sp, 0)) | |
353 | || (rp->num == 1 && rp->rules[0]->number != 0 | |
354 | && default_reduction_only_for_accept)) | |
355 | n_lookahead_tokens += rp->num; | |
356 | else | |
357 | s->consistent = 1; | |
358 | ||
359 | return n_lookahead_tokens; | |
360 | } | |
361 | ||
362 | ||
363 | /*----------------------------------------------------. | |
364 | | Compute LA, NLA, and the lookahead_tokens members. | | |
365 | `----------------------------------------------------*/ | |
366 | ||
367 | void | |
368 | initialize_LA (void) | |
369 | { | |
370 | state_number i; | |
371 | bitsetv pLA; | |
372 | bool default_reduction_only_for_accept; | |
373 | { | |
374 | char *default_reductions = | |
375 | muscle_percent_define_get ("lr.default-reductions"); | |
376 | default_reduction_only_for_accept = | |
377 | 0 == strcmp (default_reductions, "accepting"); | |
378 | free (default_reductions); | |
379 | } | |
380 | ||
381 | /* Compute the total number of reductions requiring a lookahead. */ | |
382 | nLA = 0; | |
383 | for (i = 0; i < nstates; i++) | |
384 | nLA += | |
385 | state_lookahead_tokens_count (states[i], | |
386 | default_reduction_only_for_accept); | |
387 | /* Avoid having to special case 0. */ | |
388 | if (!nLA) | |
389 | nLA = 1; | |
390 | ||
391 | pLA = LA = bitsetv_create (nLA, ntokens, BITSET_FIXED); | |
392 | ||
393 | /* Initialize the members LOOKAHEAD_TOKENS for each state whose reductions | |
394 | require lookahead tokens. */ | |
395 | for (i = 0; i < nstates; i++) | |
396 | { | |
397 | int count = | |
398 | state_lookahead_tokens_count (states[i], | |
399 | default_reduction_only_for_accept); | |
400 | if (count) | |
401 | { | |
402 | states[i]->reductions->lookahead_tokens = pLA; | |
403 | pLA += count; | |
404 | } | |
405 | } | |
406 | } | |
407 | ||
408 | ||
409 | /*---------------------------------------------. | |
410 | | Output the lookahead tokens for each state. | | |
411 | `---------------------------------------------*/ | |
412 | ||
413 | static void | |
414 | lookahead_tokens_print (FILE *out) | |
415 | { | |
416 | state_number i; | |
417 | int j, k; | |
418 | fprintf (out, "Lookahead tokens: BEGIN\n"); | |
419 | for (i = 0; i < nstates; ++i) | |
420 | { | |
421 | reductions *reds = states[i]->reductions; | |
422 | bitset_iterator iter; | |
423 | int n_lookahead_tokens = 0; | |
424 | ||
425 | if (reds->lookahead_tokens) | |
426 | for (k = 0; k < reds->num; ++k) | |
427 | if (reds->lookahead_tokens[k]) | |
428 | ++n_lookahead_tokens; | |
429 | ||
430 | fprintf (out, "State %d: %d lookahead tokens\n", | |
431 | i, n_lookahead_tokens); | |
432 | ||
433 | if (reds->lookahead_tokens) | |
434 | for (j = 0; j < reds->num; ++j) | |
435 | BITSET_FOR_EACH (iter, reds->lookahead_tokens[j], k, 0) | |
436 | { | |
437 | fprintf (out, " on %d (%s) -> rule %d\n", | |
438 | k, symbols[k]->tag, | |
439 | reds->rules[j]->number); | |
440 | }; | |
441 | } | |
442 | fprintf (out, "Lookahead tokens: END\n"); | |
443 | } | |
444 | ||
445 | void | |
446 | lalr (void) | |
447 | { | |
448 | initialize_LA (); | |
449 | set_goto_map (); | |
450 | initialize_F (); | |
451 | lookback = xcalloc (nLA, sizeof *lookback); | |
452 | build_relations (); | |
453 | compute_FOLLOWS (); | |
454 | compute_lookahead_tokens (); | |
455 | ||
456 | if (trace_flag & trace_sets) | |
457 | lookahead_tokens_print (stderr); | |
458 | } | |
459 | ||
460 | ||
461 | void | |
462 | lalr_update_state_numbers (state_number old_to_new[], state_number nstates_old) | |
463 | { | |
464 | goto_number ngotos_reachable = 0; | |
465 | symbol_number nonterminal = 0; | |
466 | aver (nsyms == nvars + ntokens); | |
467 | { | |
468 | goto_number i; | |
469 | for (i = 0; i < ngotos; ++i) | |
470 | { | |
471 | while (i == goto_map[nonterminal]) | |
472 | goto_map[nonterminal++] = ngotos_reachable; | |
473 | /* If old_to_new[from_state[i]] = nstates_old, remove this goto | |
474 | entry. */ | |
475 | if (old_to_new[from_state[i]] != nstates_old) | |
476 | { | |
477 | /* from_state[i] is not removed, so it and thus to_state[i] are | |
478 | reachable, so to_state[i] != nstates_old. */ | |
479 | aver (old_to_new[to_state[i]] != nstates_old); | |
480 | from_state[ngotos_reachable] = old_to_new[from_state[i]]; | |
481 | to_state[ngotos_reachable] = old_to_new[to_state[i]]; | |
482 | ++ngotos_reachable; | |
483 | } | |
484 | } | |
485 | } | |
486 | while (nonterminal <= nvars) | |
487 | { | |
488 | aver (ngotos == goto_map[nonterminal]); | |
489 | goto_map[nonterminal++] = ngotos_reachable; | |
490 | } | |
491 | ngotos = ngotos_reachable; | |
492 | } | |
493 | ||
494 | ||
495 | void | |
496 | lalr_free (void) | |
497 | { | |
498 | state_number s; | |
499 | for (s = 0; s < nstates; ++s) | |
500 | states[s]->reductions->lookahead_tokens = NULL; | |
501 | bitsetv_free (LA); | |
502 | } |