]> git.saurik.com Git - bison.git/blob - src/scan-gram.l
235272f1f90addccca08d84365d8f7dab039ec72
[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 YY_STEP;
111 %}
112
113
114 /*----------------------------.
115 | Scanning Bison directives. |
116 `----------------------------*/
117 <INITIAL>
118 {
119 "%binary" return PERCENT_NONASSOC;
120 "%debug" return PERCENT_DEBUG;
121 "%define" return PERCENT_DEFINE;
122 "%defines" return PERCENT_DEFINES;
123 "%destructor" return PERCENT_DESTRUCTOR;
124 "%dprec" return PERCENT_DPREC;
125 "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE;
126 "%expect" return PERCENT_EXPECT;
127 "%file-prefix" return PERCENT_FILE_PREFIX;
128 "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC;
129 "%glr-parser" return PERCENT_GLR_PARSER;
130 "%left" return PERCENT_LEFT;
131 "%locations" return PERCENT_LOCATIONS;
132 "%merge" return PERCENT_MERGE;
133 "%name"[-_]"prefix" return PERCENT_NAME_PREFIX;
134 "%no"[-_]"lines" return PERCENT_NO_LINES;
135 "%nonassoc" return PERCENT_NONASSOC;
136 "%nterm" return PERCENT_NTERM;
137 "%output" return PERCENT_OUTPUT;
138 "%parse-param" return PERCENT_PARSE_PARAM;
139 "%prec" { rule_length--; return PERCENT_PREC; }
140 "%printer" return PERCENT_PRINTER;
141 "%pure"[-_]"parser" return PERCENT_PURE_PARSER;
142 "%right" return PERCENT_RIGHT;
143 "%lex-param" return PERCENT_LEX_PARAM;
144 "%skeleton" return PERCENT_SKELETON;
145 "%start" return PERCENT_START;
146 "%term" return PERCENT_TOKEN;
147 "%token" return PERCENT_TOKEN;
148 "%token"[-_]"table" return PERCENT_TOKEN_TABLE;
149 "%type" return PERCENT_TYPE;
150 "%union" return PERCENT_UNION;
151 "%verbose" return PERCENT_VERBOSE;
152 "%yacc" return PERCENT_YACC;
153
154 "=" return EQUAL;
155 ":" { rule_length = 0; return COLON; }
156 "|" { rule_length = 0; return PIPE; }
157 "," return COMMA;
158 ";" return SEMICOLON;
159
160 {eols} YY_LINES; YY_STEP;
161 {blanks} YY_STEP;
162 {id} {
163 yylval->symbol = symbol_get (yytext, *yylloc);
164 rule_length++;
165 return ID;
166 }
167
168 {int} yylval->integer = strtol (yytext, 0, 10); return INT;
169
170 /* Characters. We don't check there is only one. */
171 "'" YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER);
172
173 /* Strings. */
174 "\"" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING);
175
176 /* Comments. */
177 "/*" yy_push_state (SC_COMMENT);
178 "//".* YY_STEP;
179
180 /* Prologue. */
181 "%{" yy_push_state (SC_PROLOGUE);
182
183 /* Code in between braces. */
184 "{" YY_OBS_GROW; ++braces_level; yy_push_state (SC_BRACED_CODE);
185
186 /* A type. */
187 "<"[^>]+">" {
188 obstack_grow (&string_obstack, yytext + 1, yyleng - 2);
189 YY_OBS_FINISH;
190 yylval->string = last_string;
191 return TYPE;
192 }
193
194
195 "%%" {
196 if (++percent_percent_count == 2)
197 yy_push_state (SC_EPILOGUE);
198 return PERCENT_PERCENT;
199 }
200
201 . {
202 LOCATION_PRINT (stderr, *yylloc);
203 fprintf (stderr, _(": invalid character: `%c'\n"), *yytext);
204 YY_STEP;
205 }
206 }
207
208
209 /*------------------------------------------------------------.
210 | Whatever the start condition (but those which correspond to |
211 | entity `swallowed' by Bison: SC_ESCAPED_STRING and |
212 | SC_ESCAPED_CHARACTER), no M4 character must escape as is. |
213 `------------------------------------------------------------*/
214
215 <SC_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
216 {
217 \[ if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@<:@");
218 \] if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@:>@");
219 }
220
221
222
223 /*-----------------------------------------------------------.
224 | Scanning a C comment. The initial `/ *' is already eaten. |
225 `-----------------------------------------------------------*/
226
227 <SC_COMMENT>
228 {
229 "*/" { /* End of the comment. */
230 if (yy_top_state () == INITIAL)
231 {
232 YY_STEP;
233 }
234 else
235 {
236 YY_OBS_GROW;
237 }
238 yy_pop_state ();
239 }
240
241 [^\[\]*\n\r]+ if (yy_top_state () != INITIAL) YY_OBS_GROW;
242 {eols} if (yy_top_state () != INITIAL) YY_OBS_GROW; YY_LINES;
243 . /* Stray `*'. */if (yy_top_state () != INITIAL) YY_OBS_GROW;
244
245 <<EOF>> {
246 LOCATION_PRINT (stderr, *yylloc);
247 fprintf (stderr, _(": unexpected end of file in a comment\n"));
248 yy_pop_state ();
249 }
250 }
251
252
253 /*----------------------------------------------------------------.
254 | Scanning a C string, including its escapes. The initial `"' is |
255 | already eaten. |
256 `----------------------------------------------------------------*/
257
258 <SC_ESCAPED_STRING>
259 {
260 "\"" {
261 assert (yy_top_state () == INITIAL);
262 YY_OBS_GROW;
263 YY_OBS_FINISH;
264 yylval->string = last_string;
265 yy_pop_state ();
266 rule_length++;
267 return STRING;
268 }
269
270 [^\"\n\r\\]+ YY_OBS_GROW;
271
272 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
273
274 <<EOF>> {
275 LOCATION_PRINT (stderr, *yylloc);
276 fprintf (stderr, _(": unexpected end of file in a string\n"));
277 assert (yy_top_state () == INITIAL);
278 YY_OBS_FINISH;
279 yylval->string = last_string;
280 yy_pop_state ();
281 return STRING;
282 }
283 }
284
285 /*---------------------------------------------------------------.
286 | Scanning a C character, decoding its escapes. The initial "'" |
287 | is already eaten. |
288 `---------------------------------------------------------------*/
289
290 <SC_ESCAPED_CHARACTER>
291 {
292 "'" {
293 YY_OBS_GROW;
294 assert (yy_top_state () == INITIAL);
295 {
296 YY_OBS_FINISH;
297 yylval->symbol = symbol_get (last_string, *yylloc);
298 symbol_class_set (yylval->symbol, token_sym, *yylloc);
299 symbol_user_token_number_set (yylval->symbol,
300 (unsigned char) last_string[1], *yylloc);
301 YY_OBS_FREE;
302 yy_pop_state ();
303 rule_length++;
304 return ID;
305 }
306 }
307
308 [^\n\r\\] YY_OBS_GROW;
309
310 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
311
312 <<EOF>> {
313 LOCATION_PRINT (stderr, *yylloc);
314 fprintf (stderr, _(": unexpected end of file in a character\n"));
315 assert (yy_top_state () == INITIAL);
316 YY_OBS_FINISH;
317 yylval->string = last_string;
318 yy_pop_state ();
319 return CHARACTER;
320 }
321 }
322
323
324 /*----------------------------.
325 | Decode escaped characters. |
326 `----------------------------*/
327
328 <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
329 {
330 \\[0-7]{3} {
331 long c = strtol (yytext + 1, 0, 8);
332 if (c > 255)
333 {
334 LOCATION_PRINT (stderr, *yylloc);
335 fprintf (stderr, _(": invalid escape: %s\n"), quote (yytext));
336 YY_STEP;
337 }
338 else
339 obstack_1grow (&string_obstack, c);
340 }
341
342 \\x[0-9a-fA-F]{2} {
343 obstack_1grow (&string_obstack, strtol (yytext + 2, 0, 16));
344 }
345
346 \\a obstack_1grow (&string_obstack, '\a');
347 \\b obstack_1grow (&string_obstack, '\b');
348 \\f obstack_1grow (&string_obstack, '\f');
349 \\n obstack_1grow (&string_obstack, '\n');
350 \\r obstack_1grow (&string_obstack, '\r');
351 \\t obstack_1grow (&string_obstack, '\t');
352 \\v obstack_1grow (&string_obstack, '\v');
353 \\[\\""''] obstack_1grow (&string_obstack, yytext[1]);
354 \\(.|\n) {
355 LOCATION_PRINT (stderr, *yylloc);
356 fprintf (stderr, _(": unrecognized escape: %s\n"), quote (yytext));
357 YY_OBS_GROW;
358 }
359 /* FLex wants this rule, in case of a `\<<EOF>>'. */
360 \\ YY_OBS_GROW;
361 }
362
363
364 /*----------------------------------------------------------.
365 | Scanning a C character without decoding its escapes. The |
366 | initial "'" is already eaten. |
367 `----------------------------------------------------------*/
368
369 <SC_CHARACTER>
370 {
371 "'" {
372 YY_OBS_GROW;
373 assert (yy_top_state () != INITIAL);
374 yy_pop_state ();
375 }
376
377 [^\[\]\'\n\r\\]+ YY_OBS_GROW;
378 \\(.|\n) YY_OBS_GROW;
379 /* FLex wants this rule, in case of a `\<<EOF>>'. */
380 \\ YY_OBS_GROW;
381
382 {eols} YY_OBS_GROW; YY_LINES;
383
384 <<EOF>> {
385 LOCATION_PRINT (stderr, *yylloc);
386 fprintf (stderr, _(": unexpected end of file in a character\n"));
387 assert (yy_top_state () != INITIAL);
388 yy_pop_state ();
389 }
390 }
391
392
393 /*----------------------------------------------------------------.
394 | Scanning a C string, without decoding its escapes. The initial |
395 | `"' is already eaten. |
396 `----------------------------------------------------------------*/
397
398 <SC_STRING>
399 {
400 "\"" {
401 assert (yy_top_state () != INITIAL);
402 YY_OBS_GROW;
403 yy_pop_state ();
404 }
405
406 [^\[\]\"\n\r\\]+ YY_OBS_GROW;
407 \\(.|\n) YY_OBS_GROW;
408 /* FLex wants this rule, in case of a `\<<EOF>>'. */
409 \\ YY_OBS_GROW;
410
411 {eols} YY_OBS_GROW; YY_LINES;
412
413 <<EOF>> {
414 LOCATION_PRINT (stderr, *yylloc);
415 fprintf (stderr, _(": unexpected end of file in a string\n"));
416 assert (yy_top_state () != INITIAL);
417 yy_pop_state ();
418 }
419 }
420
421
422 /*---------------------------------------------------.
423 | Strings, comments etc. can be found in user code. |
424 `---------------------------------------------------*/
425
426 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
427 {
428 /* Characters. We don't check there is only one. */
429 "'" YY_OBS_GROW; yy_push_state (SC_CHARACTER);
430
431 /* Strings. */
432 "\"" YY_OBS_GROW; yy_push_state (SC_STRING);
433
434 /* Comments. */
435 "/*" YY_OBS_GROW; yy_push_state (SC_COMMENT);
436 "//".* YY_OBS_GROW;
437
438 /* Not comments. */
439 "/" YY_OBS_GROW;
440 }
441
442
443 /*---------------------------------------------------------------.
444 | Scanning some code in braces (%union and actions). The initial |
445 | "{" is already eaten. |
446 `---------------------------------------------------------------*/
447
448 <SC_BRACED_CODE>
449 {
450 "}" {
451 YY_OBS_GROW;
452 if (--braces_level == 0)
453 {
454 yy_pop_state ();
455 YY_OBS_FINISH;
456 yylval->string = last_string;
457 rule_length++;
458 return BRACED_CODE;
459 }
460 }
461
462 "{" YY_OBS_GROW; braces_level++;
463
464 "$"("<"[^>]+">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,
465 yytext, *yylloc); }
466 "@"(-?[0-9]+|"$") { handle_at (current_braced_code,
467 yytext, *yylloc); }
468
469 [^$@\[\]/\'\"\{\}\n\r]+ YY_OBS_GROW;
470 {eols} YY_OBS_GROW; YY_LINES;
471
472 /* A lose $, or /, or etc. */
473 . YY_OBS_GROW;
474
475 <<EOF>> {
476 LOCATION_PRINT (stderr, *yylloc);
477 fprintf (stderr, _(": unexpected end of file in a braced code\n"));
478 yy_pop_state ();
479 YY_OBS_FINISH;
480 yylval->string = last_string;
481 return BRACED_CODE;
482 }
483
484 }
485
486
487 /*--------------------------------------------------------------.
488 | Scanning some prologue: from "%{" (already scanned) to "%}". |
489 `--------------------------------------------------------------*/
490
491 <SC_PROLOGUE>
492 {
493 "%}" {
494 yy_pop_state ();
495 YY_OBS_FINISH;
496 yylval->string = last_string;
497 return PROLOGUE;
498 }
499
500 [^%\[\]/\'\"\n\r]+ YY_OBS_GROW;
501 "%" YY_OBS_GROW;
502 {eols} YY_OBS_GROW; YY_LINES;
503
504 <<EOF>> {
505 LOCATION_PRINT (stderr, *yylloc);
506 fprintf (stderr, _(": unexpected end of file in a prologue\n"));
507 yy_pop_state ();
508 YY_OBS_FINISH;
509 yylval->string = last_string;
510 return PROLOGUE;
511 }
512 }
513
514
515 /*---------------------------------------------------------------.
516 | Scanning the epilogue (everything after the second "%%", which |
517 | has already been eaten. |
518 `---------------------------------------------------------------*/
519
520 <SC_EPILOGUE>
521 {
522 ([^\[\]]|{eols})+ YY_OBS_GROW;
523
524 <<EOF>> {
525 yy_pop_state ();
526 YY_OBS_FINISH;
527 yylval->string = last_string;
528 return EPILOGUE;
529 }
530 }
531
532
533 %%
534
535 /*------------------------------------------------------------------.
536 | TEXT is pointing to a wannabee semantic value (i.e., a `$'). |
537 | |
538 | Possible inputs: $[<TYPENAME>]($|integer) |
539 | |
540 | Output to the STRING_OBSTACK a reference to this semantic value. |
541 `------------------------------------------------------------------*/
542
543 static inline void
544 handle_action_dollar (char *text, location_t location)
545 {
546 const char *type_name = NULL;
547 char *cp = text + 1;
548
549 /* Get the type name if explicit. */
550 if (*cp == '<')
551 {
552 type_name = ++cp;
553 while (*cp != '>')
554 ++cp;
555 *cp = '\0';
556 ++cp;
557 }
558
559 if (*cp == '$')
560 {
561 if (!type_name)
562 type_name = symbol_list_n_type_name_get (current_rule, location, 0);
563 if (!type_name && typed)
564 complain_at (location, _("$$ of `%s' has no declared type"),
565 current_rule->sym->tag);
566 if (!type_name)
567 type_name = "";
568 obstack_fgrow1 (&string_obstack,
569 "]b4_lhs_value([%s])[", type_name);
570 }
571 else if (('0' <= *cp && *cp <= '9') || *cp == '-')
572 {
573 int n = strtol (cp, &cp, 10);
574
575 if (n > rule_length)
576 complain_at (location, _("invalid value: %s%d"), "$", n);
577 else
578 {
579 if (!type_name && n > 0)
580 type_name = symbol_list_n_type_name_get (current_rule, location,
581 n);
582 if (!type_name && typed)
583 complain_at (location, _("$%d of `%s' has no declared type"),
584 n, current_rule->sym->tag);
585 if (!type_name)
586 type_name = "";
587 obstack_fgrow3 (&string_obstack,
588 "]b4_rhs_value([%d], [%d], [%s])[",
589 rule_length, n, type_name);
590 }
591 }
592 else
593 {
594 complain_at (location, _("%s is invalid"), quote (text));
595 }
596 }
597
598
599 /*---------------------------------------------------------------.
600 | TEXT is expexted tp be $$ in some code associated to a symbol: |
601 | destructor or printer. |
602 `---------------------------------------------------------------*/
603
604 static inline void
605 handle_symbol_code_dollar (char *text, location_t location)
606 {
607 char *cp = text + 1;
608 if (*cp == '$')
609 obstack_sgrow (&string_obstack, "]b4_dollar_dollar[");
610 else
611 complain_at (location, _("%s is invalid"), quote (text));
612 }
613
614
615 /*-----------------------------------------------------------------.
616 | Dispatch onto handle_action_dollar, or handle_destructor_dollar, |
617 | depending upon CODE_KIND. |
618 `-----------------------------------------------------------------*/
619
620 static void
621 handle_dollar (braced_code_t braced_code_kind,
622 char *text, location_t location)
623 {
624 switch (braced_code_kind)
625 {
626 case action_braced_code:
627 handle_action_dollar (text, location);
628 break;
629
630 case destructor_braced_code:
631 case printer_braced_code:
632 handle_symbol_code_dollar (text, location);
633 break;
634 }
635 }
636
637
638 /*------------------------------------------------------.
639 | TEXT is a location token (i.e., a `@...'). Output to |
640 | STRING_OBSTACK a reference to this location. |
641 `------------------------------------------------------*/
642
643 static inline void
644 handle_action_at (char *text, location_t location)
645 {
646 char *cp = text + 1;
647 locations_flag = 1;
648
649 if (*cp == '$')
650 {
651 obstack_sgrow (&string_obstack, "]b4_lhs_location[");
652 }
653 else if (('0' <= *cp && *cp <= '9') || *cp == '-')
654 {
655 int n = strtol (cp, &cp, 10);
656
657 if (n > rule_length)
658 complain_at (location, _("invalid value: %s%d"), "@", n);
659 else
660 obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[",
661 rule_length, n);
662 }
663 else
664 {
665 complain_at (location, _("%s is invalid"), quote (text));
666 }
667 }
668
669
670 /*---------------------------------------------------------------.
671 | TEXT is expexted tp be @$ in some code associated to a symbol: |
672 | destructor or printer. |
673 `---------------------------------------------------------------*/
674
675 static inline void
676 handle_symbol_code_at (char *text, location_t location)
677 {
678 char *cp = text + 1;
679 if (*cp == '$')
680 obstack_sgrow (&string_obstack, "]b4_at_dollar[");
681 else
682 complain_at (location, _("%s is invalid"), quote (text));
683 }
684
685
686 /*-------------------------------------------------------------------.
687 | Dispatch onto handle_action_at, or handle_destructor_at, depending |
688 | upon CODE_KIND. |
689 `-------------------------------------------------------------------*/
690
691 static void
692 handle_at (braced_code_t braced_code_kind,
693 char *text, location_t location)
694 {
695 switch (braced_code_kind)
696 {
697 case action_braced_code:
698 handle_action_at (text, location);
699 break;
700
701 case destructor_braced_code:
702 case printer_braced_code:
703 handle_symbol_code_at (text, location);
704 break;
705 }
706 }
707
708
709 /*-------------------------.
710 | Initialize the scanner. |
711 `-------------------------*/
712
713 void
714 scanner_initialize (void)
715 {
716 obstack_init (&string_obstack);
717 }
718
719
720 /*-----------------------------------------------.
721 | Free all the memory allocated to the scanner. |
722 `-----------------------------------------------*/
723
724 void
725 scanner_free (void)
726 {
727 obstack_free (&string_obstack, 0);
728 /* Reclaim Flex's buffers. */
729 yy_delete_buffer (YY_CURRENT_BUFFER);
730 }