]> git.saurik.com Git - bison.git/blob - src/scan-gram.l
b89ab688b107ad93811bb72df33aec906ada6c2e
[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 /* This is only to avoid GCC warnings. */ \
38 if (yycontrol) {;}; \
39 } while (0)
40
41 #define YY_USER_ACTION LOCATION_COLUMNS (*yylloc, yyleng)
42 #define YY_LINES LOCATION_LINES (*yylloc, yyleng); lineno += yyleng;
43 #define YY_STEP LOCATION_STEP (*yylloc)
44
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 static void handle_dollar PARAMS ((braced_code_t code_kind,
83 char *cp, location_t location));
84 static void handle_at PARAMS ((braced_code_t code_kind,
85 char *cp, location_t location));
86
87 %}
88 %x SC_COMMENT
89 %x SC_STRING SC_CHARACTER
90 %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
91 %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE
92
93 id [.a-zA-Z][.a-zA-Z_0-9]*
94 int [0-9]+
95 eols (\n|\r|\n\r|\r\n)+
96 blanks [ \t\f]+
97
98 %%
99 %{
100 /* At each yylex invocation, mark the current position as the
101 start of the next token. */
102 #define TR_POS 0
103 #if TR_POS
104 fprintf (stderr, "FOO1: %p: ", yylloc);
105 LOCATION_PRINT (stderr, *yylloc);
106 fprintf (stderr, "\n");
107 #endif
108 YY_STEP;
109 #if TR_POS
110 fprintf (stderr, "BAR1: ");
111 LOCATION_PRINT (stderr, *yylloc);
112 fprintf (stderr, "\n");
113 #endif
114 %}
115
116
117 /*----------------------------.
118 | Scanning Bison directives. |
119 `----------------------------*/
120 <INITIAL>
121 {
122 "%binary" return PERCENT_NONASSOC;
123 "%debug" return PERCENT_DEBUG;
124 "%define" return PERCENT_DEFINE;
125 "%defines" return PERCENT_DEFINES;
126 "%destructor" return PERCENT_DESTRUCTOR;
127 "%dprec" return PERCENT_DPREC;
128 "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE;
129 "%expect" return PERCENT_EXPECT;
130 "%file-prefix" return PERCENT_FILE_PREFIX;
131 "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC;
132 "%glr"[-_]"parser" return PERCENT_GLR_PARSER;
133 "%left" return PERCENT_LEFT;
134 "%locations" return PERCENT_LOCATIONS;
135 "%merge" return PERCENT_MERGE;
136 "%name"[-_]"prefix" return PERCENT_NAME_PREFIX;
137 "%no"[-_]"lines" return PERCENT_NO_LINES;
138 "%nonassoc" return PERCENT_NONASSOC;
139 "%nterm" return PERCENT_NTERM;
140 "%output" return PERCENT_OUTPUT;
141 "%prec" return PERCENT_PREC;
142 "%printer" return PERCENT_PRINTER;
143 "%pure"[-_]"parser" return PERCENT_PURE_PARSER;
144 "%right" return PERCENT_RIGHT;
145 "%skeleton" return PERCENT_SKELETON;
146 "%start" return PERCENT_START;
147 "%term" return PERCENT_TOKEN;
148 "%token" return PERCENT_TOKEN;
149 "%token"[-_]"table" return PERCENT_TOKEN_TABLE;
150 "%type" return PERCENT_TYPE;
151 "%union" return PERCENT_UNION;
152 "%verbose" return PERCENT_VERBOSE;
153 "%yacc" return PERCENT_YACC;
154
155 "=" return EQUAL;
156 ":" return COLON;
157 "|" return PIPE;
158 ";" return SEMICOLON;
159
160 {eols} YY_LINES; YY_STEP;
161 {blanks} YY_STEP;
162 {id} {
163 yylval->symbol = symbol_get (yytext, *yylloc);
164 return ID;
165 }
166
167 {int} yylval->integer = strtol (yytext, 0, 10); return INT;
168
169 /* Characters. We don't check there is only one. */
170 \' YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER);
171
172 /* Strings. */
173 \" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING);
174
175 /* Comments. */
176 "/*" yy_push_state (SC_COMMENT);
177 "//".* YY_STEP;
178
179 /* Prologue. */
180 "%{" yy_push_state (SC_PROLOGUE);
181
182 /* Code in between braces. */
183 "{" YY_OBS_GROW; ++braces_level; yy_push_state (SC_BRACED_CODE);
184
185 /* A type. */
186 "<"[^>]+">" {
187 obstack_grow (&string_obstack, yytext + 1, yyleng - 2);
188 YY_OBS_FINISH;
189 yylval->string = last_string;
190 return TYPE;
191 }
192
193
194 "%%" {
195 if (++percent_percent_count == 2)
196 yy_push_state (SC_EPILOGUE);
197 return PERCENT_PERCENT;
198 }
199
200 . {
201 LOCATION_PRINT (stderr, *yylloc);
202 fprintf (stderr, ": invalid character: `%c'\n", *yytext);
203 YY_STEP;
204 }
205 }
206
207
208 /*------------------------------------------------------------.
209 | Whatever the start condition (but those which correspond to |
210 | entity `swallowed' by Bison: SC_ESCAPED_STRING and |
211 | SC_ESCAPED_CHARACTER), no M4 character must escape as is. |
212 `------------------------------------------------------------*/
213
214 <SC_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
215 {
216 \[ if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@<:@");
217 \] if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@:>@");
218 }
219
220
221
222 /*-----------------------------------------------------------.
223 | Scanning a C comment. The initial `/ *' is already eaten. |
224 `-----------------------------------------------------------*/
225
226 <SC_COMMENT>
227 {
228 "*/" { /* End of the comment. */
229 if (yy_top_state () == INITIAL)
230 {
231 YY_STEP;
232 }
233 else
234 {
235 YY_OBS_GROW;
236 }
237 yy_pop_state ();
238 }
239
240 [^\[\]*\n\r]+ if (yy_top_state () != INITIAL) YY_OBS_GROW;
241 {eols} if (yy_top_state () != INITIAL) YY_OBS_GROW; YY_LINES;
242 . /* Stray `*'. */if (yy_top_state () != INITIAL) YY_OBS_GROW;
243
244 <<EOF>> {
245 LOCATION_PRINT (stderr, *yylloc);
246 fprintf (stderr, ": unexpected end of file in a comment\n");
247 yy_pop_state ();
248 }
249 }
250
251
252 /*----------------------------------------------------------------.
253 | Scanning a C string, including its escapes. The initial `"' is |
254 | already eaten. |
255 `----------------------------------------------------------------*/
256
257 <SC_ESCAPED_STRING>
258 {
259 \" {
260 assert (yy_top_state () == INITIAL);
261 YY_OBS_GROW;
262 YY_OBS_FINISH;
263 yylval->string = last_string;
264 yy_pop_state ();
265 return STRING;
266 }
267
268 [^\"\n\r\\]+ YY_OBS_GROW;
269
270 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
271
272 <<EOF>> {
273 LOCATION_PRINT (stderr, *yylloc);
274 fprintf (stderr, ": unexpected end of file in a string\n");
275 assert (yy_top_state () == INITIAL);
276 YY_OBS_FINISH;
277 yylval->string = last_string;
278 yy_pop_state ();
279 return STRING;
280 }
281 }
282
283 /*---------------------------------------------------------------.
284 | Scanning a C character, decoding its escapes. The initial "'" |
285 | is already eaten. |
286 `---------------------------------------------------------------*/
287
288 <SC_ESCAPED_CHARACTER>
289 {
290 \' {
291 YY_OBS_GROW;
292 assert (yy_top_state () == INITIAL);
293 {
294 YY_OBS_FINISH;
295 yylval->symbol = symbol_get (last_string, *yylloc);
296 symbol_class_set (yylval->symbol, token_sym, *yylloc);
297 symbol_user_token_number_set (yylval->symbol,
298 (unsigned char) last_string[1], *yylloc);
299 YY_OBS_FREE;
300 yy_pop_state ();
301 return ID;
302 }
303 }
304
305 [^\'\n\r\\] YY_OBS_GROW;
306
307 {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES;
308
309 <<EOF>> {
310 LOCATION_PRINT (stderr, *yylloc);
311 fprintf (stderr, ": unexpected end of file in a character\n");
312 assert (yy_top_state () == INITIAL);
313 YY_OBS_FINISH;
314 yylval->string = last_string;
315 yy_pop_state ();
316 return CHARACTER;
317 }
318 }
319
320
321 /*----------------------------.
322 | Decode escaped characters. |
323 `----------------------------*/
324
325 <SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
326 {
327 \\[0-7]{3} {
328 long c = strtol (yytext + 1, 0, 8);
329 if (c > 255)
330 {
331 LOCATION_PRINT (stderr, *yylloc);
332 fprintf (stderr, ": invalid escape: %s\n", quote (yytext));
333 YY_STEP;
334 }
335 else
336 obstack_1grow (&string_obstack, c);
337 }
338
339 \\x[0-9a-fA-F]{2} {
340 obstack_1grow (&string_obstack, strtol (yytext + 2, 0, 16));
341 }
342
343 \\a obstack_1grow (&string_obstack, '\a');
344 \\b obstack_1grow (&string_obstack, '\b');
345 \\f obstack_1grow (&string_obstack, '\f');
346 \\n obstack_1grow (&string_obstack, '\n');
347 \\r obstack_1grow (&string_obstack, '\r');
348 \\t obstack_1grow (&string_obstack, '\t');
349 \\v obstack_1grow (&string_obstack, '\v');
350 \\[\\""] obstack_1grow (&string_obstack, yytext[1]);
351 \\(.|\n) {
352 LOCATION_PRINT (stderr, *yylloc);
353 fprintf (stderr, ": unrecognized escape: %s\n", quote (yytext));
354 YY_OBS_GROW;
355 }
356 /* FLex wants this rule, in case of a `\<<EOF>>'. */
357 \\ YY_OBS_GROW;
358 }
359
360
361 /*----------------------------------------------------------.
362 | Scanning a C character without decoding its escapes. The |
363 | initial "'" is already eaten. |
364 `----------------------------------------------------------*/
365
366 <SC_CHARACTER>
367 {
368 \' {
369 YY_OBS_GROW;
370 assert (yy_top_state () != INITIAL);
371 yy_pop_state ();
372 }
373
374 [^\[\]\'\n\r\\]+ YY_OBS_GROW;
375 \\(.|\n) YY_OBS_GROW;
376 /* FLex wants this rule, in case of a `\<<EOF>>'. */
377 \\ YY_OBS_GROW;
378
379 {eols} YY_OBS_GROW; YY_LINES;
380
381 <<EOF>> {
382 LOCATION_PRINT (stderr, *yylloc);
383 fprintf (stderr, ": unexpected end of file in a character\n");
384 assert (yy_top_state () != INITIAL);
385 yy_pop_state ();
386 }
387 }
388
389
390 /*----------------------------------------------------------------.
391 | Scanning a C string, without decoding its escapes. The initial |
392 | `"' is already eaten. |
393 `----------------------------------------------------------------*/
394
395 <SC_STRING>
396 {
397 \" {
398 assert (yy_top_state () != INITIAL);
399 YY_OBS_GROW;
400 yy_pop_state ();
401 }
402
403 [^\[\]\"\n\r\\]+ YY_OBS_GROW;
404 \\(.|\n) YY_OBS_GROW;
405 /* FLex wants this rule, in case of a `\<<EOF>>'. */
406 \\ YY_OBS_GROW;
407
408 {eols} YY_OBS_GROW; YY_LINES;
409
410 <<EOF>> {
411 LOCATION_PRINT (stderr, *yylloc);
412 fprintf (stderr, ": unexpected end of file in a string\n");
413 assert (yy_top_state () != INITIAL);
414 yy_pop_state ();
415 }
416 }
417
418
419 /*---------------------------------------------------.
420 | Strings, comments etc. can be found in user code. |
421 `---------------------------------------------------*/
422
423 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
424 {
425 /* Characters. We don't check there is only one. */
426 \' YY_OBS_GROW; yy_push_state (SC_CHARACTER);
427
428 /* Strings. */
429 \" YY_OBS_GROW; yy_push_state (SC_STRING);
430
431 /* Comments. */
432 "/*" YY_OBS_GROW; yy_push_state (SC_COMMENT);
433 "//".* YY_OBS_GROW;
434
435 /* Not comments. */
436 "/" YY_OBS_GROW;
437 }
438
439
440 /*---------------------------------------------------------------.
441 | Scanning some code in braces (%union and actions). The initial |
442 | "{" is already eaten. |
443 `---------------------------------------------------------------*/
444
445 <SC_BRACED_CODE>
446 {
447 "}" {
448 YY_OBS_GROW;
449 if (--braces_level == 0)
450 {
451 yy_pop_state ();
452 YY_OBS_FINISH;
453 yylval->string = last_string;
454 return BRACED_CODE;
455 }
456 }
457
458 "{" YY_OBS_GROW; braces_level++;
459
460 "$"("<"[^>]+">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,
461 yytext, *yylloc); }
462 "@"(-?[0-9]+|"$") { handle_at (current_braced_code,
463 yytext, *yylloc); }
464
465 [^$@\[\]/\'\"\{\}\n\r]+ YY_OBS_GROW;
466 {eols} YY_OBS_GROW; YY_LINES;
467
468 /* A lose $, or /, or etc. */
469 . YY_OBS_GROW;
470
471 <<EOF>> {
472 LOCATION_PRINT (stderr, *yylloc);
473 fprintf (stderr, ": unexpected end of file in a braced code\n");
474 yy_pop_state ();
475 YY_OBS_FINISH;
476 yylval->string = last_string;
477 return BRACED_CODE;
478 }
479
480 }
481
482
483 /*--------------------------------------------------------------.
484 | Scanning some prologue: from "%{" (already scanned) to "%}". |
485 `--------------------------------------------------------------*/
486
487 <SC_PROLOGUE>
488 {
489 "%}" {
490 yy_pop_state ();
491 YY_OBS_FINISH;
492 yylval->string = last_string;
493 return PROLOGUE;
494 }
495
496 [^%\[\]/\'\"\n\r]+ YY_OBS_GROW;
497 "%" YY_OBS_GROW;
498 {eols} YY_OBS_GROW; YY_LINES;
499
500 <<EOF>> {
501 LOCATION_PRINT (stderr, *yylloc);
502 fprintf (stderr, ": unexpected end of file in a prologue\n");
503 yy_pop_state ();
504 YY_OBS_FINISH;
505 yylval->string = last_string;
506 return PROLOGUE;
507 }
508 }
509
510
511 /*---------------------------------------------------------------.
512 | Scanning the epilogue (everything after the second "%%", which |
513 | has already been eaten. |
514 `---------------------------------------------------------------*/
515
516 <SC_EPILOGUE>
517 {
518 ([^\[\]]|{eols})+ YY_OBS_GROW;
519
520 <<EOF>> {
521 yy_pop_state ();
522 YY_OBS_FINISH;
523 yylval->string = last_string;
524 return EPILOGUE;
525 }
526 }
527
528
529 %%
530
531 /*------------------------------------------------------------------.
532 | TEXT is pointing to a wannabee semantic value (i.e., a `$'). |
533 | |
534 | Possible inputs: $[<TYPENAME>]($|integer) |
535 | |
536 | Output to the STRING_OBSTACK a reference to this semantic value. |
537 `------------------------------------------------------------------*/
538
539 static inline void
540 handle_action_dollar (char *text, location_t location)
541 {
542 const char *type_name = NULL;
543 char *cp = text + 1;
544
545 /* Get the type name if explicit. */
546 if (*cp == '<')
547 {
548 type_name = ++cp;
549 while (*cp != '>')
550 ++cp;
551 *cp = '\0';
552 ++cp;
553 }
554
555 if (*cp == '$')
556 {
557 if (!type_name)
558 type_name = symbol_list_n_type_name_get (current_rule, location, 0);
559 if (!type_name && typed)
560 complain_at (location, _("$$ of `%s' has no declared type"),
561 current_rule->sym->tag);
562 if (!type_name)
563 type_name = "";
564 obstack_fgrow1 (&string_obstack,
565 "]b4_lhs_value([%s])[", type_name);
566 }
567 else if (isdigit (*cp) || *cp == '-')
568 {
569 /* RULE_LENGTH is the number of values in the current rule so
570 far, which says where to find `$0' with respect to the top of
571 the stack. It is not the same as the rule->length in the
572 case of mid rule actions. */
573 int rule_length = symbol_list_length (current_rule->next);
574 int n = strtol (cp, &cp, 10);
575
576 if (n > rule_length)
577 complain_at (location, _("invalid value: %s%d"), "$", n);
578 else
579 {
580 if (!type_name && n > 0)
581 type_name = symbol_list_n_type_name_get (current_rule, location,
582 n);
583 if (!type_name && typed)
584 complain_at (location, _("$%d of `%s' has no declared type"),
585 n, current_rule->sym->tag);
586 if (!type_name)
587 type_name = "";
588 obstack_fgrow3 (&string_obstack,
589 "]b4_rhs_value([%d], [%d], [%s])[",
590 rule_length, n, type_name);
591 }
592 }
593 else
594 {
595 complain_at (location, _("%s is invalid"), quote (text));
596 }
597 }
598
599
600 /*---------------------------------------------------------------.
601 | TEXT is expexted tp be $$ in some code associated to a symbol: |
602 | destructor or printer. |
603 `---------------------------------------------------------------*/
604
605 static inline void
606 handle_symbol_code_dollar (char *text, location_t location)
607 {
608 char *cp = text + 1;
609 if (*cp == '$')
610 obstack_sgrow (&string_obstack, "]b4_dollar_dollar[");
611 else
612 complain_at (location, _("%s is invalid"), quote (text));
613 }
614
615
616 /*-----------------------------------------------------------------.
617 | Dispatch onto handle_action_dollar, or handle_destructor_dollar, |
618 | depending upon CODE_KIND. |
619 `-----------------------------------------------------------------*/
620
621 static void
622 handle_dollar (braced_code_t braced_code_kind,
623 char *text, location_t location)
624 {
625 switch (braced_code_kind)
626 {
627 case action_braced_code:
628 handle_action_dollar (text, location);
629 break;
630
631 case destructor_braced_code:
632 case printer_braced_code:
633 handle_symbol_code_dollar (text, location);
634 break;
635 }
636 }
637
638
639 /*------------------------------------------------------.
640 | TEXT is a location token (i.e., a `@...'). Output to |
641 | STRING_OBSTACK a reference to this location. |
642 `------------------------------------------------------*/
643
644 static inline void
645 handle_action_at (char *text, location_t location)
646 {
647 char *cp = text + 1;
648 locations_flag = 1;
649
650 if (*cp == '$')
651 {
652 obstack_sgrow (&string_obstack, "]b4_lhs_location[");
653 }
654 else if (isdigit (*cp) || *cp == '-')
655 {
656 /* RULE_LENGTH is the number of values in the current rule so
657 far, which says where to find `$0' with respect to the top of
658 the stack. It is not the same as the rule->length in the
659 case of mid rule actions. */
660 int rule_length = symbol_list_length (current_rule->next);
661 int n = strtol (cp, &cp, 10);
662
663 if (n > rule_length)
664 complain_at (location, _("invalid value: %s%d"), "@", n);
665 else
666 obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[",
667 rule_length, n);
668 }
669 else
670 {
671 complain_at (location, _("%s is invalid"), quote (text));
672 }
673 }
674
675
676 /*---------------------------------------------------------------.
677 | TEXT is expexted tp be @$ in some code associated to a symbol: |
678 | destructor or printer. |
679 `---------------------------------------------------------------*/
680
681 static inline void
682 handle_symbol_code_at (char *text, location_t location)
683 {
684 char *cp = text + 1;
685 if (*cp == '$')
686 obstack_sgrow (&string_obstack, "]b4_at_dollar[");
687 else
688 complain_at (location, _("%s is invalid"), quote (text));
689 }
690
691
692 /*-------------------------------------------------------------------.
693 | Dispatch onto handle_action_at, or handle_destructor_at, depending |
694 | upon CODE_KIND. |
695 `-------------------------------------------------------------------*/
696
697 static void
698 handle_at (braced_code_t braced_code_kind,
699 char *text, location_t location)
700 {
701 switch (braced_code_kind)
702 {
703 case action_braced_code:
704 handle_action_at (text, location);
705 break;
706
707 case destructor_braced_code:
708 case printer_braced_code:
709 handle_symbol_code_at (text, location);
710 break;
711 }
712 }
713
714
715 /*-------------------------.
716 | Initialize the scanner. |
717 `-------------------------*/
718
719 void
720 scanner_initialize (void)
721 {
722 obstack_init (&string_obstack);
723 }
724
725
726 /*-----------------------------------------------.
727 | Free all the memory allocated to the scanner. |
728 `-----------------------------------------------*/
729
730 void
731 scanner_free (void)
732 {
733 obstack_free (&string_obstack, 0);
734 }