]>
Commit | Line | Data |
---|---|---|
1 | /* Compute look-ahead criteria for bison, | |
2 | Copyright (C) 1984, 1986, 1989, 2000, 2001, 2002 | |
3 | Free Software Foundation, Inc. | |
4 | ||
5 | This file is part of Bison, the GNU Compiler Compiler. | |
6 | ||
7 | Bison is free software; you can redistribute it and/or modify | |
8 | it under the terms of the GNU General Public License as published by | |
9 | the Free Software Foundation; either version 2, or (at your option) | |
10 | any later version. | |
11 | ||
12 | Bison is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
18 | along with Bison; see the file COPYING. If not, write to | |
19 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
20 | Boston, MA 02111-1307, USA. */ | |
21 | ||
22 | ||
23 | /* Compute how to make the finite state machine deterministic; find | |
24 | which rules need lookahead in each state, and which lookahead | |
25 | tokens they accept. */ | |
26 | ||
27 | #include "system.h" | |
28 | #include "bitset.h" | |
29 | #include "bitsetv.h" | |
30 | #include "relation.h" | |
31 | #include "quotearg.h" | |
32 | #include "symtab.h" | |
33 | #include "gram.h" | |
34 | #include "reader.h" | |
35 | #include "LR0.h" | |
36 | #include "complain.h" | |
37 | #include "lalr.h" | |
38 | #include "nullable.h" | |
39 | #include "derives.h" | |
40 | #include "getargs.h" | |
41 | ||
42 | goto_number_t *goto_map = NULL; | |
43 | static goto_number_t ngotos = 0; | |
44 | state_number_t *from_state = NULL; | |
45 | state_number_t *to_state = NULL; | |
46 | ||
47 | /* Linked list of goto numbers. */ | |
48 | typedef struct goto_list_s | |
49 | { | |
50 | struct goto_list_s *next; | |
51 | goto_number_t value; | |
52 | } goto_list_t; | |
53 | ||
54 | ||
55 | /* LA is a LR by NTOKENS matrix of bits. LA[l, i] is 1 if the rule | |
56 | LArule[l] is applicable in the appropriate state when the next | |
57 | token is symbol i. If LA[l, i] and LA[l, j] are both 1 for i != j, | |
58 | it is a conflict. */ | |
59 | ||
60 | static bitsetv LA = NULL; | |
61 | size_t nLA; | |
62 | ||
63 | ||
64 | /* And for the famous F variable, which name is so descriptive that a | |
65 | comment is hardly needed. <grin>. */ | |
66 | static bitsetv F = NULL; | |
67 | ||
68 | static goto_number_t **includes; | |
69 | static goto_list_t **lookback; | |
70 | ||
71 | ||
72 | ||
73 | ||
74 | static void | |
75 | set_goto_map (void) | |
76 | { | |
77 | state_number_t state; | |
78 | goto_number_t *temp_map; | |
79 | ||
80 | goto_map = XCALLOC (goto_number_t, nvars + 1) - ntokens; | |
81 | temp_map = XCALLOC (goto_number_t, nvars + 1) - ntokens; | |
82 | ||
83 | ngotos = 0; | |
84 | for (state = 0; state < nstates; ++state) | |
85 | { | |
86 | transitions_t *sp = states[state]->transitions; | |
87 | int i; | |
88 | for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i) | |
89 | { | |
90 | if (ngotos == GOTO_NUMBER_MAX) | |
91 | fatal (_("too many gotos (max %d)"), GOTO_NUMBER_MAX); | |
92 | ||
93 | ngotos++; | |
94 | goto_map[TRANSITION_SYMBOL (sp, i)]++; | |
95 | } | |
96 | } | |
97 | ||
98 | { | |
99 | int k = 0; | |
100 | int i; | |
101 | for (i = ntokens; i < nsyms; i++) | |
102 | { | |
103 | temp_map[i] = k; | |
104 | k += goto_map[i]; | |
105 | } | |
106 | ||
107 | for (i = ntokens; i < nsyms; i++) | |
108 | goto_map[i] = temp_map[i]; | |
109 | ||
110 | goto_map[nsyms] = ngotos; | |
111 | temp_map[nsyms] = ngotos; | |
112 | } | |
113 | ||
114 | from_state = XCALLOC (state_number_t, ngotos); | |
115 | to_state = XCALLOC (state_number_t, ngotos); | |
116 | ||
117 | for (state = 0; state < nstates; ++state) | |
118 | { | |
119 | transitions_t *sp = states[state]->transitions; | |
120 | int i; | |
121 | for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i) | |
122 | { | |
123 | int k = temp_map[TRANSITION_SYMBOL (sp, i)]++; | |
124 | from_state[k] = state; | |
125 | to_state[k] = sp->states[i]->number; | |
126 | } | |
127 | } | |
128 | ||
129 | XFREE (temp_map + ntokens); | |
130 | } | |
131 | ||
132 | ||
133 | ||
134 | /*----------------------------------------------------------. | |
135 | | Map a state/symbol pair into its numeric representation. | | |
136 | `----------------------------------------------------------*/ | |
137 | ||
138 | static int | |
139 | map_goto (state_number_t state, symbol_number_t symbol) | |
140 | { | |
141 | int high; | |
142 | int low; | |
143 | int middle; | |
144 | state_number_t s; | |
145 | ||
146 | low = goto_map[symbol]; | |
147 | high = goto_map[symbol + 1] - 1; | |
148 | ||
149 | while (low <= high) | |
150 | { | |
151 | middle = (low + high) / 2; | |
152 | s = from_state[middle]; | |
153 | if (s == state) | |
154 | return middle; | |
155 | else if (s < state) | |
156 | low = middle + 1; | |
157 | else | |
158 | high = middle - 1; | |
159 | } | |
160 | ||
161 | assert (0); | |
162 | /* NOTREACHED */ | |
163 | return 0; | |
164 | } | |
165 | ||
166 | ||
167 | static void | |
168 | initialize_F (void) | |
169 | { | |
170 | goto_number_t **reads = XCALLOC (goto_number_t *, ngotos); | |
171 | goto_number_t *edge = XCALLOC (goto_number_t, ngotos + 1); | |
172 | int nedges = 0; | |
173 | ||
174 | int i; | |
175 | ||
176 | F = bitsetv_create (ngotos, ntokens, BITSET_FIXED); | |
177 | ||
178 | for (i = 0; i < ngotos; i++) | |
179 | { | |
180 | state_number_t stateno = to_state[i]; | |
181 | transitions_t *sp = states[stateno]->transitions; | |
182 | ||
183 | int j; | |
184 | FOR_EACH_SHIFT (sp, j) | |
185 | bitset_set (F[i], TRANSITION_SYMBOL (sp, j)); | |
186 | ||
187 | for (; j < sp->num; j++) | |
188 | { | |
189 | symbol_number_t symbol = TRANSITION_SYMBOL (sp, j); | |
190 | if (nullable[symbol]) | |
191 | edge[nedges++] = map_goto (stateno, symbol); | |
192 | } | |
193 | ||
194 | if (nedges) | |
195 | { | |
196 | reads[i] = XCALLOC (goto_number_t, nedges + 1); | |
197 | memcpy (reads[i], edge, nedges * sizeof (edge[0])); | |
198 | reads[i][nedges] = -1; | |
199 | nedges = 0; | |
200 | } | |
201 | } | |
202 | ||
203 | relation_digraph (reads, ngotos, &F); | |
204 | ||
205 | for (i = 0; i < ngotos; i++) | |
206 | XFREE (reads[i]); | |
207 | ||
208 | XFREE (reads); | |
209 | XFREE (edge); | |
210 | } | |
211 | ||
212 | ||
213 | static void | |
214 | add_lookback_edge (state_t *state, rule_t *rule, int gotono) | |
215 | { | |
216 | int r = state_reduction_find (state, rule); | |
217 | goto_list_t *sp = XCALLOC (goto_list_t, 1); | |
218 | sp->next = lookback[(state->reductions->lookaheads - LA) + r]; | |
219 | sp->value = gotono; | |
220 | lookback[(state->reductions->lookaheads - LA) + r] = sp; | |
221 | } | |
222 | ||
223 | ||
224 | ||
225 | static void | |
226 | build_relations (void) | |
227 | { | |
228 | goto_number_t *edge = XCALLOC (goto_number_t, ngotos + 1); | |
229 | state_number_t *states1 = XCALLOC (state_number_t, ritem_longest_rhs () + 1); | |
230 | int i; | |
231 | ||
232 | includes = XCALLOC (goto_number_t *, ngotos); | |
233 | ||
234 | for (i = 0; i < ngotos; i++) | |
235 | { | |
236 | int nedges = 0; | |
237 | symbol_number_t symbol1 = states[to_state[i]]->accessing_symbol; | |
238 | rule_t **rulep; | |
239 | ||
240 | for (rulep = derives[symbol1]; *rulep; rulep++) | |
241 | { | |
242 | int done; | |
243 | int length = 1; | |
244 | item_number_t *rp; | |
245 | state_t *state = states[from_state[i]]; | |
246 | states1[0] = state->number; | |
247 | ||
248 | for (rp = (*rulep)->rhs; *rp >= 0; rp++) | |
249 | { | |
250 | state = transitions_to (state->transitions, | |
251 | item_number_as_symbol_number (*rp)); | |
252 | states1[length++] = state->number; | |
253 | } | |
254 | ||
255 | if (!state->consistent) | |
256 | add_lookback_edge (state, *rulep, i); | |
257 | ||
258 | length--; | |
259 | done = 0; | |
260 | while (!done) | |
261 | { | |
262 | done = 1; | |
263 | rp--; | |
264 | /* JF added rp>=ritem && I hope to god its right! */ | |
265 | if (rp >= ritem && ISVAR (*rp)) | |
266 | { | |
267 | /* Downcasting from item_number_t to symbol_number_t. */ | |
268 | edge[nedges++] = map_goto (states1[--length], | |
269 | item_number_as_symbol_number (*rp)); | |
270 | if (nullable[*rp]) | |
271 | done = 0; | |
272 | } | |
273 | } | |
274 | } | |
275 | ||
276 | if (nedges) | |
277 | { | |
278 | int j; | |
279 | includes[i] = XCALLOC (goto_number_t, nedges + 1); | |
280 | for (j = 0; j < nedges; j++) | |
281 | includes[i][j] = edge[j]; | |
282 | includes[i][nedges] = -1; | |
283 | } | |
284 | } | |
285 | ||
286 | XFREE (edge); | |
287 | XFREE (states1); | |
288 | ||
289 | relation_transpose (&includes, ngotos); | |
290 | } | |
291 | ||
292 | ||
293 | ||
294 | static void | |
295 | compute_FOLLOWS (void) | |
296 | { | |
297 | int i; | |
298 | ||
299 | relation_digraph (includes, ngotos, &F); | |
300 | ||
301 | for (i = 0; i < ngotos; i++) | |
302 | XFREE (includes[i]); | |
303 | ||
304 | XFREE (includes); | |
305 | } | |
306 | ||
307 | ||
308 | static void | |
309 | compute_lookaheads (void) | |
310 | { | |
311 | size_t i; | |
312 | goto_list_t *sp; | |
313 | ||
314 | for (i = 0; i < nLA; i++) | |
315 | for (sp = lookback[i]; sp; sp = sp->next) | |
316 | bitset_or (LA[i], LA[i], F[sp->value]); | |
317 | ||
318 | /* Free LOOKBACK. */ | |
319 | for (i = 0; i < nLA; i++) | |
320 | LIST_FREE (goto_list_t, lookback[i]); | |
321 | ||
322 | XFREE (lookback); | |
323 | bitsetv_free (F); | |
324 | } | |
325 | ||
326 | ||
327 | /*---------------------------------------------------------------. | |
328 | | Count the number of lookaheads required for STATE (NLOOKAHEADS | | |
329 | | member). | | |
330 | `---------------------------------------------------------------*/ | |
331 | ||
332 | static int | |
333 | state_lookaheads_count (state_t *state) | |
334 | { | |
335 | int k; | |
336 | int nlookaheads = 0; | |
337 | reductions_t *rp = state->reductions; | |
338 | transitions_t *sp = state->transitions; | |
339 | ||
340 | /* We need a lookahead either to distinguish different | |
341 | reductions (i.e., there are two or more), or to distinguish a | |
342 | reduction from a shift. Otherwise, it is straightforward, | |
343 | and the state is `consistent'. */ | |
344 | if (rp->num > 1 | |
345 | || (rp->num == 1 && sp->num && | |
346 | !TRANSITION_IS_DISABLED (sp, 0) && TRANSITION_IS_SHIFT (sp, 0))) | |
347 | nlookaheads += rp->num; | |
348 | else | |
349 | state->consistent = 1; | |
350 | ||
351 | for (k = 0; k < sp->num; k++) | |
352 | if (!TRANSITION_IS_DISABLED (sp, k) && TRANSITION_IS_ERROR (sp, k)) | |
353 | { | |
354 | state->consistent = 0; | |
355 | break; | |
356 | } | |
357 | ||
358 | return nlookaheads; | |
359 | } | |
360 | ||
361 | ||
362 | /*----------------------------------------------. | |
363 | | Compute LA, NLA, and the lookaheads members. | | |
364 | `----------------------------------------------*/ | |
365 | ||
366 | static void | |
367 | initialize_LA (void) | |
368 | { | |
369 | state_number_t i; | |
370 | bitsetv pLA; | |
371 | ||
372 | /* Compute the total number of reductions requiring a lookahead. */ | |
373 | nLA = 0; | |
374 | for (i = 0; i < nstates; i++) | |
375 | nLA += state_lookaheads_count (states[i]); | |
376 | /* Avoid having to special case 0. */ | |
377 | if (!nLA) | |
378 | nLA = 1; | |
379 | ||
380 | pLA = LA = bitsetv_create (nLA, ntokens, BITSET_FIXED); | |
381 | lookback = XCALLOC (goto_list_t *, nLA); | |
382 | ||
383 | /* Initialize the members LOOKAHEADS for each state which reductions | |
384 | require lookaheads. */ | |
385 | for (i = 0; i < nstates; i++) | |
386 | { | |
387 | int count = state_lookaheads_count (states[i]); | |
388 | if (count) | |
389 | { | |
390 | states[i]->reductions->lookaheads = pLA; | |
391 | pLA += count; | |
392 | } | |
393 | } | |
394 | } | |
395 | ||
396 | ||
397 | /*---------------------------------------. | |
398 | | Output the lookaheads for each state. | | |
399 | `---------------------------------------*/ | |
400 | ||
401 | static void | |
402 | lookaheads_print (FILE *out) | |
403 | { | |
404 | state_number_t i; | |
405 | int j, k; | |
406 | fprintf (out, "Lookaheads: BEGIN\n"); | |
407 | for (i = 0; i < nstates; ++i) | |
408 | { | |
409 | reductions_t *reds = states[i]->reductions; | |
410 | bitset_iterator iter; | |
411 | int nlookaheads = 0; | |
412 | ||
413 | if (reds->lookaheads) | |
414 | for (k = 0; k < reds->num; ++k) | |
415 | if (reds->lookaheads[k]) | |
416 | ++nlookaheads; | |
417 | ||
418 | fprintf (out, "State %d: %d lookaheads\n", | |
419 | i, nlookaheads); | |
420 | ||
421 | if (reds->lookaheads) | |
422 | for (j = 0; j < reds->num; ++j) | |
423 | BITSET_FOR_EACH (iter, reds->lookaheads[j], k, 0) | |
424 | { | |
425 | fprintf (out, " on %d (%s) -> rule %d\n", | |
426 | k, symbols[k]->tag, | |
427 | reds->rules[j]->number); | |
428 | }; | |
429 | } | |
430 | fprintf (out, "Lookaheads: END\n"); | |
431 | } | |
432 | ||
433 | void | |
434 | lalr (void) | |
435 | { | |
436 | initialize_LA (); | |
437 | set_goto_map (); | |
438 | initialize_F (); | |
439 | build_relations (); | |
440 | compute_FOLLOWS (); | |
441 | compute_lookaheads (); | |
442 | ||
443 | if (trace_flag & trace_sets) | |
444 | lookaheads_print (stderr); | |
445 | } | |
446 | ||
447 | ||
448 | void | |
449 | lalr_free (void) | |
450 | { | |
451 | state_number_t s; | |
452 | for (s = 0; s < nstates; ++s) | |
453 | states[s]->reductions->lookaheads = NULL; | |
454 | bitsetv_free (LA); | |
455 | } |