]>
Commit | Line | Data |
---|---|---|
40675e7c | 1 | /* Generate the nondeterministic finite state machine for bison, |
6fc82eaf PE |
2 | |
3 | Copyright (C) 1984, 1986, 1989, 2000, 2001, 2002 Free Software | |
4 | Foundation, Inc. | |
40675e7c | 5 | |
2fa6973e | 6 | This file is part of Bison, the GNU Compiler Compiler. |
40675e7c | 7 | |
2fa6973e AD |
8 | Bison 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 2, or (at your option) | |
11 | any later version. | |
40675e7c | 12 | |
2fa6973e AD |
13 | Bison 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. | |
40675e7c | 17 | |
2fa6973e AD |
18 | You should have received a copy of the GNU General Public License |
19 | along with Bison; see the file COPYING. If not, write to | |
20 | the Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
21 | Boston, MA 02111-1307, USA. */ | |
40675e7c DM |
22 | |
23 | ||
24 | /* See comments in state.h for the data structures that represent it. | |
25 | The entry point is generate_states. */ | |
26 | ||
40675e7c | 27 | #include "system.h" |
602bbf31 | 28 | #include "bitset.h" |
8b3df748 | 29 | #include "quotearg.h" |
0e78e603 | 30 | #include "symtab.h" |
5fbb0954 | 31 | #include "gram.h" |
9bfe901c | 32 | #include "getargs.h" |
c87d4863 | 33 | #include "reader.h" |
40675e7c DM |
34 | #include "gram.h" |
35 | #include "state.h" | |
a0f6b076 | 36 | #include "complain.h" |
2fa6973e | 37 | #include "closure.h" |
403b315b | 38 | #include "LR0.h" |
49701457 | 39 | #include "lalr.h" |
630e182b | 40 | #include "reduce.h" |
40675e7c | 41 | |
32e1e0a4 AD |
42 | typedef struct state_list_s |
43 | { | |
44 | struct state_list_s *next; | |
45 | state_t *state; | |
46 | } state_list_t; | |
47 | ||
48 | static state_list_t *first_state = NULL; | |
49 | static state_list_t *last_state = NULL; | |
50 | ||
8b752b00 AD |
51 | |
52 | /*------------------------------------------------------------------. | |
53 | | A state was just discovered from another state. Queue it for | | |
54 | | later examination, in order to find its transitions. Return it. | | |
55 | `------------------------------------------------------------------*/ | |
56 | ||
57 | static state_t * | |
58 | state_list_append (symbol_number_t symbol, | |
59 | size_t core_size, item_number_t *core) | |
32e1e0a4 AD |
60 | { |
61 | state_list_t *node = XMALLOC (state_list_t, 1); | |
8b752b00 AD |
62 | state_t *state = state_new (symbol, core_size, core); |
63 | ||
273a74fa | 64 | if (trace_flag & trace_automaton) |
427c0dda | 65 | fprintf (stderr, "state_list_append (state = %d, symbol = %d (%s))\n", |
8b752b00 AD |
66 | nstates, symbol, symbols[symbol]->tag); |
67 | ||
88bce5a2 | 68 | /* If this is the endtoken, and this is not the initial state, then |
8b752b00 AD |
69 | this is the final state. */ |
70 | if (symbol == 0 && first_state) | |
71 | final_state = state; | |
72 | ||
32e1e0a4 AD |
73 | node->next = NULL; |
74 | node->state = state; | |
40675e7c | 75 | |
32e1e0a4 AD |
76 | if (!first_state) |
77 | first_state = node; | |
78 | if (last_state) | |
79 | last_state->next = node; | |
80 | last_state = node; | |
8b752b00 AD |
81 | |
82 | return state; | |
32e1e0a4 | 83 | } |
40675e7c DM |
84 | |
85 | static int nshifts; | |
a49aecd5 | 86 | static symbol_number_t *shift_symbol = NULL; |
40675e7c | 87 | |
640748ee AD |
88 | static rule_t **redset = NULL; |
89 | static state_t **shiftset = NULL; | |
40675e7c | 90 | |
62a3e4f0 | 91 | static item_number_t **kernel_base = NULL; |
6255b435 | 92 | static int *kernel_size = NULL; |
62a3e4f0 | 93 | static item_number_t *kernel_items = NULL; |
40675e7c | 94 | |
2fa6973e | 95 | \f |
4a120d45 | 96 | static void |
d2729d44 | 97 | allocate_itemsets (void) |
40675e7c | 98 | { |
9222837b AD |
99 | symbol_number_t i; |
100 | rule_number_t r; | |
62a3e4f0 | 101 | item_number_t *rhsp; |
40675e7c | 102 | |
630e182b AD |
103 | /* Count the number of occurrences of all the symbols in RITEMS. |
104 | Note that useless productions (hence useless nonterminals) are | |
105 | browsed too, hence we need to allocate room for _all_ the | |
106 | symbols. */ | |
107 | int count = 0; | |
108 | short *symbol_count = XCALLOC (short, nsyms + nuseless_nonterminals); | |
40675e7c | 109 | |
4b3d3a8e | 110 | for (r = 0; r < nrules; ++r) |
b4c4ccc2 | 111 | for (rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp) |
c87d4863 AD |
112 | { |
113 | count++; | |
b4c4ccc2 | 114 | symbol_count[*rhsp]++; |
c87d4863 | 115 | } |
40675e7c | 116 | |
2fa6973e AD |
117 | /* See comments before new_itemsets. All the vectors of items |
118 | live inside KERNEL_ITEMS. The number of active items after | |
40675e7c | 119 | some symbol cannot be more than the number of times that symbol |
8a731ca8 | 120 | appears as an item, which is SYMBOL_COUNT[SYMBOL]. |
40675e7c DM |
121 | We allocate that much space for each symbol. */ |
122 | ||
62a3e4f0 | 123 | kernel_base = XCALLOC (item_number_t *, nsyms); |
342b8b6e | 124 | if (count) |
62a3e4f0 | 125 | kernel_items = XCALLOC (item_number_t, count); |
40675e7c DM |
126 | |
127 | count = 0; | |
128 | for (i = 0; i < nsyms; i++) | |
129 | { | |
130 | kernel_base[i] = kernel_items + count; | |
131 | count += symbol_count[i]; | |
132 | } | |
133 | ||
630e182b | 134 | free (symbol_count); |
0e41b407 | 135 | kernel_size = XCALLOC (int, nsyms); |
40675e7c DM |
136 | } |
137 | ||
138 | ||
4a120d45 | 139 | static void |
d2729d44 | 140 | allocate_storage (void) |
40675e7c | 141 | { |
2fa6973e | 142 | allocate_itemsets (); |
40675e7c | 143 | |
640748ee AD |
144 | shiftset = XCALLOC (state_t *, nsyms); |
145 | redset = XCALLOC (rule_t *, nrules); | |
c7ca99d4 | 146 | state_hash_new (); |
a49aecd5 | 147 | shift_symbol = XCALLOC (symbol_number_t, nsyms); |
40675e7c DM |
148 | } |
149 | ||
150 | ||
4a120d45 | 151 | static void |
d2729d44 | 152 | free_storage (void) |
40675e7c | 153 | { |
630e182b AD |
154 | free (shift_symbol); |
155 | free (redset); | |
156 | free (shiftset); | |
157 | free (kernel_base); | |
158 | free (kernel_size); | |
d7913476 | 159 | XFREE (kernel_items); |
c7ca99d4 | 160 | state_hash_free (); |
40675e7c DM |
161 | } |
162 | ||
163 | ||
164 | ||
40675e7c | 165 | |
32e1e0a4 AD |
166 | /*---------------------------------------------------------------. |
167 | | Find which symbols can be shifted in STATE, and for each one | | |
168 | | record which items would be active after that shift. Uses the | | |
169 | | contents of itemset. | | |
170 | | | | |
171 | | shift_symbol is set to a vector of the symbols that can be | | |
172 | | shifted. For each symbol in the grammar, kernel_base[symbol] | | |
173 | | points to a vector of item numbers activated if that symbol is | | |
174 | | shifted, and kernel_size[symbol] is their numbers. | | |
175 | `---------------------------------------------------------------*/ | |
40675e7c | 176 | |
4a120d45 | 177 | static void |
32e1e0a4 | 178 | new_itemsets (state_t *state) |
40675e7c | 179 | { |
2fa6973e | 180 | int i; |
2fa6973e | 181 | |
273a74fa | 182 | if (trace_flag & trace_automaton) |
427c0dda | 183 | fprintf (stderr, "Entering new_itemsets, state = %d\n", |
32e1e0a4 | 184 | state->number); |
40675e7c DM |
185 | |
186 | for (i = 0; i < nsyms; i++) | |
125ecb56 | 187 | kernel_size[i] = 0; |
40675e7c | 188 | |
b2872512 | 189 | nshifts = 0; |
40675e7c | 190 | |
5123689b | 191 | for (i = 0; i < nritemset; ++i) |
5fbb0954 AD |
192 | if (ritem[itemset[i]] >= 0) |
193 | { | |
a49aecd5 AD |
194 | symbol_number_t symbol |
195 | = item_number_as_symbol_number (ritem[itemset[i]]); | |
5fbb0954 AD |
196 | if (!kernel_size[symbol]) |
197 | { | |
198 | shift_symbol[nshifts] = symbol; | |
199 | nshifts++; | |
200 | } | |
201 | ||
202 | kernel_base[symbol][kernel_size[symbol]] = itemset[i] + 1; | |
203 | kernel_size[symbol]++; | |
204 | } | |
40675e7c DM |
205 | } |
206 | ||
207 | ||
208 | ||
640748ee AD |
209 | /*-----------------------------------------------------------------. |
210 | | Find the state we would get to (from the current state) by | | |
211 | | shifting SYMBOL. Create a new state if no equivalent one exists | | |
212 | | already. Used by append_states. | | |
213 | `-----------------------------------------------------------------*/ | |
40675e7c | 214 | |
640748ee | 215 | static state_t * |
a49aecd5 | 216 | get_state (symbol_number_t symbol, size_t core_size, item_number_t *core) |
40675e7c | 217 | { |
f693ad14 | 218 | state_t *sp; |
40675e7c | 219 | |
273a74fa | 220 | if (trace_flag & trace_automaton) |
427c0dda | 221 | fprintf (stderr, "Entering get_state, symbol = %d (%s)\n", |
97650f4e | 222 | symbol, symbols[symbol]->tag); |
40675e7c | 223 | |
c7ca99d4 AD |
224 | sp = state_hash_lookup (core_size, core); |
225 | if (!sp) | |
8b752b00 | 226 | sp = state_list_append (symbol, core_size, core); |
40675e7c | 227 | |
273a74fa | 228 | if (trace_flag & trace_automaton) |
427c0dda | 229 | fprintf (stderr, "Exiting get_state => %d\n", sp->number); |
c87d4863 | 230 | |
640748ee | 231 | return sp; |
40675e7c DM |
232 | } |
233 | ||
640748ee AD |
234 | /*---------------------------------------------------------------. |
235 | | Use the information computed by new_itemsets to find the state | | |
236 | | numbers reached by each shift transition from STATE. | | |
237 | | | | |
238 | | SHIFTSET is set up as a vector of those states. | | |
239 | `---------------------------------------------------------------*/ | |
40675e7c | 240 | |
2fa6973e | 241 | static void |
32e1e0a4 | 242 | append_states (state_t *state) |
40675e7c | 243 | { |
2fa6973e AD |
244 | int i; |
245 | int j; | |
a49aecd5 | 246 | symbol_number_t symbol; |
40675e7c | 247 | |
273a74fa | 248 | if (trace_flag & trace_automaton) |
427c0dda | 249 | fprintf (stderr, "Entering append_states, state = %d\n", |
32e1e0a4 | 250 | state->number); |
40675e7c | 251 | |
2fa6973e | 252 | /* first sort shift_symbol into increasing order */ |
40675e7c | 253 | |
2fa6973e AD |
254 | for (i = 1; i < nshifts; i++) |
255 | { | |
256 | symbol = shift_symbol[i]; | |
257 | j = i; | |
258 | while (j > 0 && shift_symbol[j - 1] > symbol) | |
259 | { | |
260 | shift_symbol[j] = shift_symbol[j - 1]; | |
261 | j--; | |
262 | } | |
263 | shift_symbol[j] = symbol; | |
264 | } | |
40675e7c | 265 | |
2fa6973e | 266 | for (i = 0; i < nshifts; i++) |
458be8e0 AD |
267 | { |
268 | symbol = shift_symbol[i]; | |
269 | shiftset[i] = get_state (symbol, | |
270 | kernel_size[symbol], kernel_base[symbol]); | |
271 | } | |
40675e7c DM |
272 | } |
273 | ||
274 | ||
2fa6973e AD |
275 | /*----------------------------------------------------------------. |
276 | | Find which rules can be used for reduction transitions from the | | |
277 | | current state and make a reductions structure for the state to | | |
278 | | record their rule numbers. | | |
279 | `----------------------------------------------------------------*/ | |
280 | ||
4a120d45 | 281 | static void |
32e1e0a4 | 282 | save_reductions (state_t *state) |
40675e7c | 283 | { |
30171f79 | 284 | int count = 0; |
fb908786 | 285 | int i; |
40675e7c | 286 | |
30171f79 | 287 | /* Find and count the active items that represent ends of rules. */ |
5123689b | 288 | for (i = 0; i < nritemset; ++i) |
2fa6973e | 289 | { |
fb908786 | 290 | int item = ritem[itemset[i]]; |
2fa6973e | 291 | if (item < 0) |
640748ee | 292 | redset[count++] = &rules[item_number_as_rule_number (item)]; |
2fa6973e | 293 | } |
40675e7c | 294 | |
2fa6973e | 295 | /* Make a reductions structure and copy the data into it. */ |
8a731ca8 | 296 | state_reductions_set (state, count, redset); |
2fa6973e AD |
297 | } |
298 | ||
299 | \f | |
82841af7 | 300 | /*---------------. |
29e88316 | 301 | | Build STATES. | |
82841af7 | 302 | `---------------*/ |
6a164e0c AD |
303 | |
304 | static void | |
29e88316 | 305 | set_states (void) |
6a164e0c | 306 | { |
29e88316 | 307 | states = XCALLOC (state_t *, nstates); |
6a164e0c | 308 | |
32e1e0a4 | 309 | while (first_state) |
2cec70b9 | 310 | { |
32e1e0a4 AD |
311 | state_list_t *this = first_state; |
312 | ||
2cec70b9 | 313 | /* Pessimization, but simplification of the code: make sure all |
8b752b00 AD |
314 | the states have valid transitions and reductions members, |
315 | even if reduced to 0. It is too soon for errs, which are | |
316 | computed later, but set_conflicts. */ | |
32e1e0a4 | 317 | state_t *state = this->state; |
8b752b00 | 318 | if (!state->transitions) |
ccaf65bc | 319 | state_transitions_set (state, 0, 0); |
32e1e0a4 | 320 | if (!state->reductions) |
8a731ca8 | 321 | state_reductions_set (state, 0, 0); |
32e1e0a4 AD |
322 | |
323 | states[state->number] = state; | |
324 | ||
325 | first_state = this->next; | |
326 | free (this); | |
2cec70b9 | 327 | } |
32e1e0a4 AD |
328 | first_state = NULL; |
329 | last_state = NULL; | |
6a164e0c AD |
330 | } |
331 | ||
c7ca99d4 | 332 | |
2fa6973e AD |
333 | /*-------------------------------------------------------------------. |
334 | | Compute the nondeterministic finite state machine (see state.h for | | |
335 | | details) from the grammar. | | |
336 | `-------------------------------------------------------------------*/ | |
337 | ||
338 | void | |
339 | generate_states (void) | |
340 | { | |
32e1e0a4 | 341 | state_list_t *list = NULL; |
2fa6973e | 342 | allocate_storage (); |
9e7f6bbd | 343 | new_closure (nritems); |
8b752b00 AD |
344 | |
345 | /* Create the initial state. The 0 at the lhs is the index of the | |
346 | item of this initial rule. */ | |
347 | kernel_base[0][0] = 0; | |
348 | kernel_size[0] = 1; | |
349 | state_list_append (0, kernel_size[0], kernel_base[0]); | |
350 | ||
32e1e0a4 | 351 | list = first_state; |
2fa6973e | 352 | |
32e1e0a4 | 353 | while (list) |
2fa6973e | 354 | { |
32e1e0a4 | 355 | state_t *state = list->state; |
273a74fa | 356 | if (trace_flag & trace_automaton) |
427c0dda | 357 | fprintf (stderr, "Processing state %d (reached by %s)\n", |
32e1e0a4 | 358 | state->number, |
97650f4e | 359 | symbols[state->accessing_symbol]->tag); |
2fa6973e AD |
360 | /* Set up ruleset and itemset for the transitions out of this |
361 | state. ruleset gets a 1 bit for each rule that could reduce | |
362 | now. itemset gets a vector of all the items that could be | |
363 | accepted next. */ | |
32e1e0a4 AD |
364 | closure (state->items, state->nitems); |
365 | /* Record the reductions allowed out of this state. */ | |
366 | save_reductions (state); | |
367 | /* Find the itemsets of the states that shifts can reach. */ | |
368 | new_itemsets (state); | |
369 | /* Find or create the core structures for those states. */ | |
370 | append_states (state); | |
371 | ||
372 | /* Create the shifts structures for the shifts to those states, | |
373 | now that the state numbers transitioning to are known. */ | |
ccaf65bc | 374 | state_transitions_set (state, nshifts, shiftset); |
32e1e0a4 AD |
375 | |
376 | /* States are queued when they are created; process them all. | |
377 | */ | |
378 | list = list->next; | |
2fa6973e AD |
379 | } |
380 | ||
381 | /* discard various storage */ | |
382 | free_closure (); | |
383 | free_storage (); | |
384 | ||
29e88316 AD |
385 | /* Set up STATES. */ |
386 | set_states (); | |
40675e7c | 387 | } |