]> git.saurik.com Git - bison.git/blob - src/scan-gram.l
3408f98d80773d386dd6fe05c8e1e1a316cab9b9
[bison.git] / src / scan-gram.l
1 /* Bison Grammar Scanner -*- C -*-
2 Copyright (C) 2002 Free Software Foundation, Inc.
3
4 This file is part of Bison, the GNU Compiler Compiler.
5
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
10
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
15
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software
18 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
19 02111-1307 USA
20 */
21
22 %option debug nodefault noyywrap nounput never-interactive stack
23 %option prefix="gram_" outfile="lex.yy.c"
24
25 %{
26 #include "system.h"
27 #include "complain.h"
28 #include "quote.h"
29 #include "getargs.h"
30 #include "gram.h"
31 #include "reader.h"
32
33 /* Each time we match a string, move the end cursor to its end. */
34 #define YY_USER_INIT \
35 do { \
36 LOCATION_RESET (*yylloc); \
37 yylloc->file = infile; \
38 /* This is only to avoid GCC warnings. */ \
39 if (yycontrol) {;}; \
40 } while (0)
41
42 #define YY_USER_ACTION LOCATION_COLUMNS (*yylloc, yyleng);
43 #define YY_LINES LOCATION_LINES (*yylloc, yyleng);
44 #define YY_STEP LOCATION_STEP (*yylloc);
45
46 /* STRING_OBSTACK -- Used to store all the characters that we need to
47 keep (to construct ID, STRINGS etc.). Use the following macros to
48 use it.
49
50 Use YY_OBS_GROW to append what has just been matched, and
51 YY_OBS_FINISH to end the string (it puts the ending 0).
52 YY_OBS_FINISH also stores this string in LAST_STRING, which can be
53 used, and which is used by YY_OBS_FREE to free the last string. */
54
55 static struct obstack string_obstack;
56 char *last_string;
57
58 #define YY_OBS_GROW \
59 obstack_grow (&string_obstack, yytext, yyleng)
60
61 #define YY_OBS_FINISH \
62 do { \
63 obstack_1grow (&string_obstack, '\0'); \
64 last_string = obstack_finish (&string_obstack); \
65 } while (0)
66
67 #define YY_OBS_FREE \
68 do { \
69 obstack_free (&string_obstack, last_string); \
70 } while (0)
71
72 void
73 scanner_last_string_free (void)
74 {
75 YY_OBS_FREE;
76 }
77
78
79 static int braces_level = 0;
80 static int percent_percent_count = 0;
81
82 /* Within well-formed rules, RULE_LENGTH is the number of values in
83 the current rule so far, which says where to find `$0' with respect
84 to the top of the stack. It is not the same as the rule->length in
85 the case of mid rule actions.
86
87 Outside of well-formed rules, RULE_LENGTH has an undefined value. */
88 static int rule_length;
89
90 static void handle_dollar PARAMS ((braced_code_t code_kind,
91 char *cp, location_t location));
92 static void handle_at PARAMS ((braced_code_t code_kind,
93 char *cp, location_t location));
94
95 %}
96 %x SC_COMMENT
97 %x SC_STRING SC_CHARACTER
98 %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
99 %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE
100
101 id [.a-zA-Z_][.a-zA-Z_0-9]*
102 int [0-9]+
103 eols (\n|\r|\n\r|\r\n)+
104 blanks [ \t\f]+
105
106 %%
107 %{
108 /* At each yylex invocation, mark the current position as the
109 start of the next token. */
110 #define TR_POS 0
111 #if TR_POS
112 fprintf (stderr, "FOO1: %p: ", yylloc);
113 LOCATION_PRINT (stderr, *yylloc);
114 fprintf (stderr, "\n");
115 #endif
116 YY_STEP;
117 #if TR_POS
118 fprintf (stderr, "BAR1: ");
119 LOCATION_PRINT (stderr, *yylloc);
120 fprintf (stderr, "\n");
121 #endif
122 %}
123
124
125 /*----------------------------.
126 | Scanning Bison directives. |
127 `----------------------------*/
128 <INITIAL>
129 {
130 "%binary" return PERCENT_NONASSOC;
131 "%debug" return PERCENT_DEBUG;
132 "%define" return PERCENT_DEFINE;
133 "%defines" return PERCENT_DEFINES;
134 "%destructor" return PERCENT_DESTRUCTOR;
135 "%dprec" return PERCENT_DPREC;
136 "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE;
137 "%expect" return PERCENT_EXPECT;
138 "%file-prefix" return PERCENT_FILE_PREFIX;
139 "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC;
140 "%glr"[-_]"parser" return PERCENT_GLR_PARSER;
141 "%left" return PERCENT_LEFT;
142 "%locations" return PERCENT_LOCATIONS;
143 "%merge" return PERCENT_MERGE;
144 "%name"[-_]"prefix" return PERCENT_NAME_PREFIX;
145 "%no"[-_]"lines" return PERCENT_NO_LINES;
146 "%nonassoc" return PERCENT_NONASSOC;
147 "%nterm" return PERCENT_NTERM;
148 "%output" return PERCENT_OUTPUT;
149 "%prec" { rule_length--; return PERCENT_PREC; }
150 "%printer" return PERCENT_PRINTER;
151 "%pure"[-_]"parser" return PERCENT_PURE_PARSER;
152 "%right" return PERCENT_RIGHT;
153 "%skeleton" return PERCENT_SKELETON;
154 "%start" return PERCENT_START;
155 "%term" return PERCENT_TOKEN;
156 "%token" return PERCENT_TOKEN;
157 "%token"[-_]"table" return PERCENT_TOKEN_TABLE;
158 "%type" return PERCENT_TYPE;
159 "%union" return PERCENT_UNION;
160 "%verbose" return PERCENT_VERBOSE;
161 "%yacc" return PERCENT_YACC;
162
163 "=" return EQUAL;
164 ":" { rule_length = 0; return COLON; }
165 "|" { rule_length = 0; return PIPE; }
166 ";" return SEMICOLON;
167
168 {eols} YY_LINES; YY_STEP;
169 {blanks} YY_STEP;
170 {id} {
171 yylval->symbol = symbol_get (yytext, *yylloc);
172 rule_length++;
173 return ID;
174 }
175
176 {int} yylval->integer = strtol (yytext, 0, 10); return INT;
177
178 /* Characters. We don't check there is only one. */
179 "'" YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER);
180
181 /* Strings. */
182 "\"" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING);
183
184 /* Comments. */
185 "/*" yy_push_state (SC_COMMENT);
186 "//".* YY_STEP;
187
188 /* Prologue. */
189 "%{" yy_push_state (SC_PROLOGUE);
190
191 /* Code in between braces. */
192 "{" YY_OBS_GROW; ++braces_level; yy_push_state (SC_BRACED_CODE);
193
194 /* A type. */
195 "<"[^>]+">" {
196 obstack_grow (&string_obstack, yytext + 1, yyleng - 2);
197 YY_OBS_FINISH;
198 yylval->string = last_string;
199 return TYPE;
200 }
201
202
203 "%%" {
204 if (++percent_percent_count == 2)
205 yy_push_state (SC_EPILOGUE);
206 return PERCENT_PERCENT;
207 }
208
209 . {
210 LOCATION_PRINT (stderr, *yylloc);
211 fprintf (stderr, _(": invalid character: `%c'\n"), *yytext);
212 YY_STEP;
213 }
214 }
215
216
217 /*------------------------------------------------------------.
218 | Whatever the start condition (but those which correspond to |
219 | entity `swallowed' by Bison: SC_ESCAPED_STRING and |
220 | SC_ESCAPED_CHARACTER), no M4 character must escape as is. |
221 `------------------------------------------------------------*/
222
223 <SC_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
224 {
225 \[ if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@<:@");
226 \] if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@:>@");
227 }
228
229
230
231 /*-----------------------------------------------------------.
232 | Scanning a C comment. The initial `/ *' is already eaten. |
233 `-----------------------------------------------------------*/
234
235 <SC_COMMENT>
236 {
237 "*/" { /* End of the comment. */
238 if (yy_top_state () == INITIAL)
239 {
240 YY_STEP;
241 }
242 else
243 {
244 YY_OBS_GROW;
245 }
246 yy_pop_state ();
247 }
248
249 [^\[\]*\n\r]+ if (yy_top_state () != INITIAL) YY_OBS_GROW;
250 {eols} if (yy_top_state () != INITIAL) YY_OBS_GROW; YY_LINES;
251 . /* Stray `*'. */if (yy_top_state () != INITIAL) YY_OBS_GROW;
252
253 <<EOF>> {
254 LOCATION_PRINT (stderr, *yylloc);
255 fprintf (stderr, _(": unexpected end of file in a comment\n"));
256 yy_pop_state ();
257 }
258 }
259
260
261 /*----------------------------------------------------------------.
262 | Scanning a C string, including its escapes. The initial `"' is |
263 | already eaten. |
264 `----------------------------------------------------------------*/
265
266 <SC_ESCAPED_STRING>
267 {
268 "\"" {
269 assert (yy_top_state () == INITIAL);
270 YY_OBS_GROW;
271 YY_OBS_FINISH;
272 yylval->string = last_string;
273 yy_pop_state ();
274 rule_length++;
275 return STRING;
276 }
277
278 [^\"\n\r\\]+ YY_OBS_GROW;
279
280 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
281
282 <<EOF>> {
283 LOCATION_PRINT (stderr, *yylloc);
284 fprintf (stderr, _(": unexpected end of file in a string\n"));
285 assert (yy_top_state () == INITIAL);
286 YY_OBS_FINISH;
287 yylval->string = last_string;
288 yy_pop_state ();
289 return STRING;
290 }
291 }
292
293 /*---------------------------------------------------------------.
294 | Scanning a C character, decoding its escapes. The initial "'" |
295 | is already eaten. |
296 `---------------------------------------------------------------*/
297
298 <SC_ESCAPED_CHARACTER>
299 {
300 "'" {
301 YY_OBS_GROW;
302 assert (yy_top_state () == INITIAL);
303 {
304 YY_OBS_FINISH;
305 yylval->symbol = symbol_get (last_string, *yylloc);
306 symbol_class_set (yylval->symbol, token_sym, *yylloc);
307 symbol_user_token_number_set (yylval->symbol,
308 (unsigned char) last_string[1], *yylloc);
309 YY_OBS_FREE;
310 yy_pop_state ();
311 rule_length++;
312 return ID;
313 }
314 }
315
316 [^\n\r\\] YY_OBS_GROW;
317
318 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
319
320 <<EOF>> {
321 LOCATION_PRINT (stderr, *yylloc);
322 fprintf (stderr, _(": unexpected end of file in a character\n"));
323 assert (yy_top_state () == INITIAL);
324 YY_OBS_FINISH;
325 yylval->string = last_string;
326 yy_pop_state ();
327 return CHARACTER;
328 }
329 }
330
331
332 /*----------------------------.
333 | Decode escaped characters. |
334 `----------------------------*/
335
336 <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
337 {
338 \\[0-7]{3} {
339 long c = strtol (yytext + 1, 0, 8);
340 if (c > 255)
341 {
342 LOCATION_PRINT (stderr, *yylloc);
343 fprintf (stderr, _(": invalid escape: %s\n"), quote (yytext));
344 YY_STEP;
345 }
346 else
347 obstack_1grow (&string_obstack, c);
348 }
349
350 \\x[0-9a-fA-F]{2} {
351 obstack_1grow (&string_obstack, strtol (yytext + 2, 0, 16));
352 }
353
354 \\a obstack_1grow (&string_obstack, '\a');
355 \\b obstack_1grow (&string_obstack, '\b');
356 \\f obstack_1grow (&string_obstack, '\f');
357 \\n obstack_1grow (&string_obstack, '\n');
358 \\r obstack_1grow (&string_obstack, '\r');
359 \\t obstack_1grow (&string_obstack, '\t');
360 \\v obstack_1grow (&string_obstack, '\v');
361 \\[\\""''] obstack_1grow (&string_obstack, yytext[1]);
362 \\(.|\n) {
363 LOCATION_PRINT (stderr, *yylloc);
364 fprintf (stderr, _(": unrecognized escape: %s\n"), quote (yytext));
365 YY_OBS_GROW;
366 }
367 /* FLex wants this rule, in case of a `\<<EOF>>'. */
368 \\ YY_OBS_GROW;
369 }
370
371
372 /*----------------------------------------------------------.
373 | Scanning a C character without decoding its escapes. The |
374 | initial "'" is already eaten. |
375 `----------------------------------------------------------*/
376
377 <SC_CHARACTER>
378 {
379 "'" {
380 YY_OBS_GROW;
381 assert (yy_top_state () != INITIAL);
382 yy_pop_state ();
383 }
384
385 [^\[\]\'\n\r\\]+ YY_OBS_GROW;
386 \\(.|\n) YY_OBS_GROW;
387 /* FLex wants this rule, in case of a `\<<EOF>>'. */
388 \\ YY_OBS_GROW;
389
390 {eols} YY_OBS_GROW; YY_LINES;
391
392 <<EOF>> {
393 LOCATION_PRINT (stderr, *yylloc);
394 fprintf (stderr, _(": unexpected end of file in a character\n"));
395 assert (yy_top_state () != INITIAL);
396 yy_pop_state ();
397 }
398 }
399
400
401 /*----------------------------------------------------------------.
402 | Scanning a C string, without decoding its escapes. The initial |
403 | `"' is already eaten. |
404 `----------------------------------------------------------------*/
405
406 <SC_STRING>
407 {
408 "\"" {
409 assert (yy_top_state () != INITIAL);
410 YY_OBS_GROW;
411 yy_pop_state ();
412 }
413
414 [^\[\]\"\n\r\\]+ YY_OBS_GROW;
415 \\(.|\n) YY_OBS_GROW;
416 /* FLex wants this rule, in case of a `\<<EOF>>'. */
417 \\ YY_OBS_GROW;
418
419 {eols} YY_OBS_GROW; YY_LINES;
420
421 <<EOF>> {
422 LOCATION_PRINT (stderr, *yylloc);
423 fprintf (stderr, _(": unexpected end of file in a string\n"));
424 assert (yy_top_state () != INITIAL);
425 yy_pop_state ();
426 }
427 }
428
429
430 /*---------------------------------------------------.
431 | Strings, comments etc. can be found in user code. |
432 `---------------------------------------------------*/
433
434 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
435 {
436 /* Characters. We don't check there is only one. */
437 "'" YY_OBS_GROW; yy_push_state (SC_CHARACTER);
438
439 /* Strings. */
440 "\"" YY_OBS_GROW; yy_push_state (SC_STRING);
441
442 /* Comments. */
443 "/*" YY_OBS_GROW; yy_push_state (SC_COMMENT);
444 "//".* YY_OBS_GROW;
445
446 /* Not comments. */
447 "/" YY_OBS_GROW;
448 }
449
450
451 /*---------------------------------------------------------------.
452 | Scanning some code in braces (%union and actions). The initial |
453 | "{" is already eaten. |
454 `---------------------------------------------------------------*/
455
456 <SC_BRACED_CODE>
457 {
458 "}" {
459 YY_OBS_GROW;
460 if (--braces_level == 0)
461 {
462 yy_pop_state ();
463 YY_OBS_FINISH;
464 yylval->string = last_string;
465 rule_length++;
466 return BRACED_CODE;
467 }
468 }
469
470 "{" YY_OBS_GROW; braces_level++;
471
472 "$"("<"[^>]+">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,
473 yytext, *yylloc); }
474 "@"(-?[0-9]+|"$") { handle_at (current_braced_code,
475 yytext, *yylloc); }
476
477 [^$@\[\]/\'\"\{\}\n\r]+ YY_OBS_GROW;
478 {eols} YY_OBS_GROW; YY_LINES;
479
480 /* A lose $, or /, or etc. */
481 . YY_OBS_GROW;
482
483 <<EOF>> {
484 LOCATION_PRINT (stderr, *yylloc);
485 fprintf (stderr, _(": unexpected end of file in a braced code\n"));
486 yy_pop_state ();
487 YY_OBS_FINISH;
488 yylval->string = last_string;
489 return BRACED_CODE;
490 }
491
492 }
493
494
495 /*--------------------------------------------------------------.
496 | Scanning some prologue: from "%{" (already scanned) to "%}". |
497 `--------------------------------------------------------------*/
498
499 <SC_PROLOGUE>
500 {
501 "%}" {
502 yy_pop_state ();
503 YY_OBS_FINISH;
504 yylval->string = last_string;
505 return PROLOGUE;
506 }
507
508 [^%\[\]/\'\"\n\r]+ YY_OBS_GROW;
509 "%" YY_OBS_GROW;
510 {eols} YY_OBS_GROW; YY_LINES;
511
512 <<EOF>> {
513 LOCATION_PRINT (stderr, *yylloc);
514 fprintf (stderr, _(": unexpected end of file in a prologue\n"));
515 yy_pop_state ();
516 YY_OBS_FINISH;
517 yylval->string = last_string;
518 return PROLOGUE;
519 }
520 }
521
522
523 /*---------------------------------------------------------------.
524 | Scanning the epilogue (everything after the second "%%", which |
525 | has already been eaten. |
526 `---------------------------------------------------------------*/
527
528 <SC_EPILOGUE>
529 {
530 ([^\[\]]|{eols})+ YY_OBS_GROW;
531
532 <<EOF>> {
533 yy_pop_state ();
534 YY_OBS_FINISH;
535 yylval->string = last_string;
536 return EPILOGUE;
537 }
538 }
539
540
541 %%
542
543 /*------------------------------------------------------------------.
544 | TEXT is pointing to a wannabee semantic value (i.e., a `$'). |
545 | |
546 | Possible inputs: $[<TYPENAME>]($|integer) |
547 | |
548 | Output to the STRING_OBSTACK a reference to this semantic value. |
549 `------------------------------------------------------------------*/
550
551 static inline void
552 handle_action_dollar (char *text, location_t location)
553 {
554 const char *type_name = NULL;
555 char *cp = text + 1;
556
557 /* Get the type name if explicit. */
558 if (*cp == '<')
559 {
560 type_name = ++cp;
561 while (*cp != '>')
562 ++cp;
563 *cp = '\0';
564 ++cp;
565 }
566
567 if (*cp == '$')
568 {
569 if (!type_name)
570 type_name = symbol_list_n_type_name_get (current_rule, location, 0);
571 if (!type_name && typed)
572 complain_at (location, _("$$ of `%s' has no declared type"),
573 current_rule->sym->tag);
574 if (!type_name)
575 type_name = "";
576 obstack_fgrow1 (&string_obstack,
577 "]b4_lhs_value([%s])[", type_name);
578 }
579 else if (('0' <= *cp && *cp <= '9') || *cp == '-')
580 {
581 int n = strtol (cp, &cp, 10);
582
583 if (n > rule_length)
584 complain_at (location, _("invalid value: %s%d"), "$", n);
585 else
586 {
587 if (!type_name && n > 0)
588 type_name = symbol_list_n_type_name_get (current_rule, location,
589 n);
590 if (!type_name && typed)
591 complain_at (location, _("$%d of `%s' has no declared type"),
592 n, current_rule->sym->tag);
593 if (!type_name)
594 type_name = "";
595 obstack_fgrow3 (&string_obstack,
596 "]b4_rhs_value([%d], [%d], [%s])[",
597 rule_length, n, type_name);
598 }
599 }
600 else
601 {
602 complain_at (location, _("%s is invalid"), quote (text));
603 }
604 }
605
606
607 /*---------------------------------------------------------------.
608 | TEXT is expexted tp be $$ in some code associated to a symbol: |
609 | destructor or printer. |
610 `---------------------------------------------------------------*/
611
612 static inline void
613 handle_symbol_code_dollar (char *text, location_t location)
614 {
615 char *cp = text + 1;
616 if (*cp == '$')
617 obstack_sgrow (&string_obstack, "]b4_dollar_dollar[");
618 else
619 complain_at (location, _("%s is invalid"), quote (text));
620 }
621
622
623 /*-----------------------------------------------------------------.
624 | Dispatch onto handle_action_dollar, or handle_destructor_dollar, |
625 | depending upon CODE_KIND. |
626 `-----------------------------------------------------------------*/
627
628 static void
629 handle_dollar (braced_code_t braced_code_kind,
630 char *text, location_t location)
631 {
632 switch (braced_code_kind)
633 {
634 case action_braced_code:
635 handle_action_dollar (text, location);
636 break;
637
638 case destructor_braced_code:
639 case printer_braced_code:
640 handle_symbol_code_dollar (text, location);
641 break;
642 }
643 }
644
645
646 /*------------------------------------------------------.
647 | TEXT is a location token (i.e., a `@...'). Output to |
648 | STRING_OBSTACK a reference to this location. |
649 `------------------------------------------------------*/
650
651 static inline void
652 handle_action_at (char *text, location_t location)
653 {
654 char *cp = text + 1;
655 locations_flag = 1;
656
657 if (*cp == '$')
658 {
659 obstack_sgrow (&string_obstack, "]b4_lhs_location[");
660 }
661 else if (('0' <= *cp && *cp <= '9') || *cp == '-')
662 {
663 int n = strtol (cp, &cp, 10);
664
665 if (n > rule_length)
666 complain_at (location, _("invalid value: %s%d"), "@", n);
667 else
668 obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[",
669 rule_length, n);
670 }
671 else
672 {
673 complain_at (location, _("%s is invalid"), quote (text));
674 }
675 }
676
677
678 /*---------------------------------------------------------------.
679 | TEXT is expexted tp be @$ in some code associated to a symbol: |
680 | destructor or printer. |
681 `---------------------------------------------------------------*/
682
683 static inline void
684 handle_symbol_code_at (char *text, location_t location)
685 {
686 char *cp = text + 1;
687 if (*cp == '$')
688 obstack_sgrow (&string_obstack, "]b4_at_dollar[");
689 else
690 complain_at (location, _("%s is invalid"), quote (text));
691 }
692
693
694 /*-------------------------------------------------------------------.
695 | Dispatch onto handle_action_at, or handle_destructor_at, depending |
696 | upon CODE_KIND. |
697 `-------------------------------------------------------------------*/
698
699 static void
700 handle_at (braced_code_t braced_code_kind,
701 char *text, location_t location)
702 {
703 switch (braced_code_kind)
704 {
705 case action_braced_code:
706 handle_action_at (text, location);
707 break;
708
709 case destructor_braced_code:
710 case printer_braced_code:
711 handle_symbol_code_at (text, location);
712 break;
713 }
714 }
715
716
717 /*-------------------------.
718 | Initialize the scanner. |
719 `-------------------------*/
720
721 void
722 scanner_initialize (void)
723 {
724 obstack_init (&string_obstack);
725 }
726
727
728 /*-----------------------------------------------.
729 | Free all the memory allocated to the scanner. |
730 `-----------------------------------------------*/
731
732 void
733 scanner_free (void)
734 {
735 obstack_free (&string_obstack, 0);
736 /* Reclaim Flex's buffers. */
737 yy_delete_buffer (YY_CURRENT_BUFFER);
738 }