33c7418caedf7509e2d6536e0cd86d042f4150c7
[deliverable/binutils-gdb.git] / gdb / f-exp.y
1 /* YACC parser for Fortran expressions, for GDB.
2 Copyright (C) 1986, 1989-1991, 1993-1996, 2000-2012 Free Software
3 Foundation, Inc.
4
5 Contributed by Motorola. Adapted from the C parser by Farooq Butt
6 (fmbutt@engage.sps.mot.com).
7
8 This file is part of GDB.
9
10 This program is free software; you can redistribute it and/or modify
11 it under the terms of the GNU General Public License as published by
12 the Free Software Foundation; either version 3 of the License, or
13 (at your option) any later version.
14
15 This program is distributed in the hope that it will be useful,
16 but WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 GNU General Public License for more details.
19
20 You should have received a copy of the GNU General Public License
21 along with this program. If not, see <http://www.gnu.org/licenses/>. */
22
23 /* This was blantantly ripped off the C expression parser, please
24 be aware of that as you look at its basic structure -FMB */
25
26 /* Parse a F77 expression from text in a string,
27 and return the result as a struct expression pointer.
28 That structure contains arithmetic operations in reverse polish,
29 with constants represented by operations that are followed by special data.
30 See expression.h for the details of the format.
31 What is important here is that it can be built up sequentially
32 during the process of parsing; the lower levels of the tree always
33 come first in the result.
34
35 Note that malloc's and realloc's in this file are transformed to
36 xmalloc and xrealloc respectively by the same sed command in the
37 makefile that remaps any other malloc/realloc inserted by the parser
38 generator. Doing this with #defines and trying to control the interaction
39 with include files (<malloc.h> and <stdlib.h> for example) just became
40 too messy, particularly when such includes can be inserted at random
41 times by the parser generator. */
42
43 %{
44
45 #include "defs.h"
46 #include "gdb_string.h"
47 #include "expression.h"
48 #include "value.h"
49 #include "parser-defs.h"
50 #include "language.h"
51 #include "f-lang.h"
52 #include "bfd.h" /* Required by objfiles.h. */
53 #include "symfile.h" /* Required by objfiles.h. */
54 #include "objfiles.h" /* For have_full_symbols and have_partial_symbols */
55 #include "block.h"
56 #include <ctype.h>
57
58 #define parse_type builtin_type (parse_gdbarch)
59 #define parse_f_type builtin_f_type (parse_gdbarch)
60
61 /* Remap normal yacc parser interface names (yyparse, yylex, yyerror, etc),
62 as well as gratuitiously global symbol names, so we can have multiple
63 yacc generated parsers in gdb. Note that these are only the variables
64 produced by yacc. If other parser generators (bison, byacc, etc) produce
65 additional global names that conflict at link time, then those parser
66 generators need to be fixed instead of adding those names to this list. */
67
68 #define yymaxdepth f_maxdepth
69 #define yyparse f_parse
70 #define yylex f_lex
71 #define yyerror f_error
72 #define yylval f_lval
73 #define yychar f_char
74 #define yydebug f_debug
75 #define yypact f_pact
76 #define yyr1 f_r1
77 #define yyr2 f_r2
78 #define yydef f_def
79 #define yychk f_chk
80 #define yypgo f_pgo
81 #define yyact f_act
82 #define yyexca f_exca
83 #define yyerrflag f_errflag
84 #define yynerrs f_nerrs
85 #define yyps f_ps
86 #define yypv f_pv
87 #define yys f_s
88 #define yy_yys f_yys
89 #define yystate f_state
90 #define yytmp f_tmp
91 #define yyv f_v
92 #define yy_yyv f_yyv
93 #define yyval f_val
94 #define yylloc f_lloc
95 #define yyreds f_reds /* With YYDEBUG defined */
96 #define yytoks f_toks /* With YYDEBUG defined */
97 #define yyname f_name /* With YYDEBUG defined */
98 #define yyrule f_rule /* With YYDEBUG defined */
99 #define yylhs f_yylhs
100 #define yylen f_yylen
101 #define yydefred f_yydefred
102 #define yydgoto f_yydgoto
103 #define yysindex f_yysindex
104 #define yyrindex f_yyrindex
105 #define yygindex f_yygindex
106 #define yytable f_yytable
107 #define yycheck f_yycheck
108 #define yyss f_yyss
109 #define yysslim f_yysslim
110 #define yyssp f_yyssp
111 #define yystacksize f_yystacksize
112 #define yyvs f_yyvs
113 #define yyvsp f_yyvsp
114
115 #ifndef YYDEBUG
116 #define YYDEBUG 1 /* Default to yydebug support */
117 #endif
118
119 #define YYFPRINTF parser_fprintf
120
121 int yyparse (void);
122
123 static int yylex (void);
124
125 void yyerror (char *);
126
127 static void growbuf_by_size (int);
128
129 static int match_string_literal (void);
130
131 %}
132
133 /* Although the yacc "value" of an expression is not used,
134 since the result is stored in the structure being created,
135 other node types do have values. */
136
137 %union
138 {
139 LONGEST lval;
140 struct {
141 LONGEST val;
142 struct type *type;
143 } typed_val;
144 DOUBLEST dval;
145 struct symbol *sym;
146 struct type *tval;
147 struct stoken sval;
148 struct ttype tsym;
149 struct symtoken ssym;
150 int voidval;
151 struct block *bval;
152 enum exp_opcode opcode;
153 struct internalvar *ivar;
154
155 struct type **tvec;
156 int *ivec;
157 }
158
159 %{
160 /* YYSTYPE gets defined by %union */
161 static int parse_number (char *, int, int, YYSTYPE *);
162 %}
163
164 %type <voidval> exp type_exp start variable
165 %type <tval> type typebase
166 %type <tvec> nonempty_typelist
167 /* %type <bval> block */
168
169 /* Fancy type parsing. */
170 %type <voidval> func_mod direct_abs_decl abs_decl
171 %type <tval> ptype
172
173 %token <typed_val> INT
174 %token <dval> FLOAT
175
176 /* Both NAME and TYPENAME tokens represent symbols in the input,
177 and both convey their data as strings.
178 But a TYPENAME is a string that happens to be defined as a typedef
179 or builtin type name (such as int or char)
180 and a NAME is any other symbol.
181 Contexts where this distinction is not important can use the
182 nonterminal "name", which matches either NAME or TYPENAME. */
183
184 %token <sval> STRING_LITERAL
185 %token <lval> BOOLEAN_LITERAL
186 %token <ssym> NAME
187 %token <tsym> TYPENAME
188 %type <sval> name
189 %type <ssym> name_not_typename
190
191 /* A NAME_OR_INT is a symbol which is not known in the symbol table,
192 but which would parse as a valid number in the current input radix.
193 E.g. "c" when input_radix==16. Depending on the parse, it will be
194 turned into a name or into a number. */
195
196 %token <ssym> NAME_OR_INT
197
198 %token SIZEOF
199 %token ERROR
200
201 /* Special type cases, put in to allow the parser to distinguish different
202 legal basetypes. */
203 %token INT_KEYWORD INT_S2_KEYWORD LOGICAL_S1_KEYWORD LOGICAL_S2_KEYWORD
204 %token LOGICAL_S8_KEYWORD
205 %token LOGICAL_KEYWORD REAL_KEYWORD REAL_S8_KEYWORD REAL_S16_KEYWORD
206 %token COMPLEX_S8_KEYWORD COMPLEX_S16_KEYWORD COMPLEX_S32_KEYWORD
207 %token BOOL_AND BOOL_OR BOOL_NOT
208 %token <lval> CHARACTER
209
210 %token <voidval> VARIABLE
211
212 %token <opcode> ASSIGN_MODIFY
213
214 %left ','
215 %left ABOVE_COMMA
216 %right '=' ASSIGN_MODIFY
217 %right '?'
218 %left BOOL_OR
219 %right BOOL_NOT
220 %left BOOL_AND
221 %left '|'
222 %left '^'
223 %left '&'
224 %left EQUAL NOTEQUAL
225 %left LESSTHAN GREATERTHAN LEQ GEQ
226 %left LSH RSH
227 %left '@'
228 %left '+' '-'
229 %left '*' '/'
230 %right STARSTAR
231 %right '%'
232 %right UNARY
233 %right '('
234
235 \f
236 %%
237
238 start : exp
239 | type_exp
240 ;
241
242 type_exp: type
243 { write_exp_elt_opcode(OP_TYPE);
244 write_exp_elt_type($1);
245 write_exp_elt_opcode(OP_TYPE); }
246 ;
247
248 exp : '(' exp ')'
249 { }
250 ;
251
252 /* Expressions, not including the comma operator. */
253 exp : '*' exp %prec UNARY
254 { write_exp_elt_opcode (UNOP_IND); }
255 ;
256
257 exp : '&' exp %prec UNARY
258 { write_exp_elt_opcode (UNOP_ADDR); }
259 ;
260
261 exp : '-' exp %prec UNARY
262 { write_exp_elt_opcode (UNOP_NEG); }
263 ;
264
265 exp : BOOL_NOT exp %prec UNARY
266 { write_exp_elt_opcode (UNOP_LOGICAL_NOT); }
267 ;
268
269 exp : '~' exp %prec UNARY
270 { write_exp_elt_opcode (UNOP_COMPLEMENT); }
271 ;
272
273 exp : SIZEOF exp %prec UNARY
274 { write_exp_elt_opcode (UNOP_SIZEOF); }
275 ;
276
277 /* No more explicit array operators, we treat everything in F77 as
278 a function call. The disambiguation as to whether we are
279 doing a subscript operation or a function call is done
280 later in eval.c. */
281
282 exp : exp '('
283 { start_arglist (); }
284 arglist ')'
285 { write_exp_elt_opcode (OP_F77_UNDETERMINED_ARGLIST);
286 write_exp_elt_longcst ((LONGEST) end_arglist ());
287 write_exp_elt_opcode (OP_F77_UNDETERMINED_ARGLIST); }
288 ;
289
290 arglist :
291 ;
292
293 arglist : exp
294 { arglist_len = 1; }
295 ;
296
297 arglist : subrange
298 { arglist_len = 1; }
299 ;
300
301 arglist : arglist ',' exp %prec ABOVE_COMMA
302 { arglist_len++; }
303 ;
304
305 /* There are four sorts of subrange types in F90. */
306
307 subrange: exp ':' exp %prec ABOVE_COMMA
308 { write_exp_elt_opcode (OP_F90_RANGE);
309 write_exp_elt_longcst (NONE_BOUND_DEFAULT);
310 write_exp_elt_opcode (OP_F90_RANGE); }
311 ;
312
313 subrange: exp ':' %prec ABOVE_COMMA
314 { write_exp_elt_opcode (OP_F90_RANGE);
315 write_exp_elt_longcst (HIGH_BOUND_DEFAULT);
316 write_exp_elt_opcode (OP_F90_RANGE); }
317 ;
318
319 subrange: ':' exp %prec ABOVE_COMMA
320 { write_exp_elt_opcode (OP_F90_RANGE);
321 write_exp_elt_longcst (LOW_BOUND_DEFAULT);
322 write_exp_elt_opcode (OP_F90_RANGE); }
323 ;
324
325 subrange: ':' %prec ABOVE_COMMA
326 { write_exp_elt_opcode (OP_F90_RANGE);
327 write_exp_elt_longcst (BOTH_BOUND_DEFAULT);
328 write_exp_elt_opcode (OP_F90_RANGE); }
329 ;
330
331 complexnum: exp ',' exp
332 { }
333 ;
334
335 exp : '(' complexnum ')'
336 { write_exp_elt_opcode(OP_COMPLEX);
337 write_exp_elt_type (parse_f_type->builtin_complex_s16);
338 write_exp_elt_opcode(OP_COMPLEX); }
339 ;
340
341 exp : '(' type ')' exp %prec UNARY
342 { write_exp_elt_opcode (UNOP_CAST);
343 write_exp_elt_type ($2);
344 write_exp_elt_opcode (UNOP_CAST); }
345 ;
346
347 exp : exp '%' name
348 { write_exp_elt_opcode (STRUCTOP_STRUCT);
349 write_exp_string ($3);
350 write_exp_elt_opcode (STRUCTOP_STRUCT); }
351 ;
352
353 /* Binary operators in order of decreasing precedence. */
354
355 exp : exp '@' exp
356 { write_exp_elt_opcode (BINOP_REPEAT); }
357 ;
358
359 exp : exp STARSTAR exp
360 { write_exp_elt_opcode (BINOP_EXP); }
361 ;
362
363 exp : exp '*' exp
364 { write_exp_elt_opcode (BINOP_MUL); }
365 ;
366
367 exp : exp '/' exp
368 { write_exp_elt_opcode (BINOP_DIV); }
369 ;
370
371 exp : exp '+' exp
372 { write_exp_elt_opcode (BINOP_ADD); }
373 ;
374
375 exp : exp '-' exp
376 { write_exp_elt_opcode (BINOP_SUB); }
377 ;
378
379 exp : exp LSH exp
380 { write_exp_elt_opcode (BINOP_LSH); }
381 ;
382
383 exp : exp RSH exp
384 { write_exp_elt_opcode (BINOP_RSH); }
385 ;
386
387 exp : exp EQUAL exp
388 { write_exp_elt_opcode (BINOP_EQUAL); }
389 ;
390
391 exp : exp NOTEQUAL exp
392 { write_exp_elt_opcode (BINOP_NOTEQUAL); }
393 ;
394
395 exp : exp LEQ exp
396 { write_exp_elt_opcode (BINOP_LEQ); }
397 ;
398
399 exp : exp GEQ exp
400 { write_exp_elt_opcode (BINOP_GEQ); }
401 ;
402
403 exp : exp LESSTHAN exp
404 { write_exp_elt_opcode (BINOP_LESS); }
405 ;
406
407 exp : exp GREATERTHAN exp
408 { write_exp_elt_opcode (BINOP_GTR); }
409 ;
410
411 exp : exp '&' exp
412 { write_exp_elt_opcode (BINOP_BITWISE_AND); }
413 ;
414
415 exp : exp '^' exp
416 { write_exp_elt_opcode (BINOP_BITWISE_XOR); }
417 ;
418
419 exp : exp '|' exp
420 { write_exp_elt_opcode (BINOP_BITWISE_IOR); }
421 ;
422
423 exp : exp BOOL_AND exp
424 { write_exp_elt_opcode (BINOP_LOGICAL_AND); }
425 ;
426
427
428 exp : exp BOOL_OR exp
429 { write_exp_elt_opcode (BINOP_LOGICAL_OR); }
430 ;
431
432 exp : exp '=' exp
433 { write_exp_elt_opcode (BINOP_ASSIGN); }
434 ;
435
436 exp : exp ASSIGN_MODIFY exp
437 { write_exp_elt_opcode (BINOP_ASSIGN_MODIFY);
438 write_exp_elt_opcode ($2);
439 write_exp_elt_opcode (BINOP_ASSIGN_MODIFY); }
440 ;
441
442 exp : INT
443 { write_exp_elt_opcode (OP_LONG);
444 write_exp_elt_type ($1.type);
445 write_exp_elt_longcst ((LONGEST)($1.val));
446 write_exp_elt_opcode (OP_LONG); }
447 ;
448
449 exp : NAME_OR_INT
450 { YYSTYPE val;
451 parse_number ($1.stoken.ptr, $1.stoken.length, 0, &val);
452 write_exp_elt_opcode (OP_LONG);
453 write_exp_elt_type (val.typed_val.type);
454 write_exp_elt_longcst ((LONGEST)val.typed_val.val);
455 write_exp_elt_opcode (OP_LONG); }
456 ;
457
458 exp : FLOAT
459 { write_exp_elt_opcode (OP_DOUBLE);
460 write_exp_elt_type (parse_f_type->builtin_real_s8);
461 write_exp_elt_dblcst ($1);
462 write_exp_elt_opcode (OP_DOUBLE); }
463 ;
464
465 exp : variable
466 ;
467
468 exp : VARIABLE
469 ;
470
471 exp : SIZEOF '(' type ')' %prec UNARY
472 { write_exp_elt_opcode (OP_LONG);
473 write_exp_elt_type (parse_f_type->builtin_integer);
474 CHECK_TYPEDEF ($3);
475 write_exp_elt_longcst ((LONGEST) TYPE_LENGTH ($3));
476 write_exp_elt_opcode (OP_LONG); }
477 ;
478
479 exp : BOOLEAN_LITERAL
480 { write_exp_elt_opcode (OP_BOOL);
481 write_exp_elt_longcst ((LONGEST) $1);
482 write_exp_elt_opcode (OP_BOOL);
483 }
484 ;
485
486 exp : STRING_LITERAL
487 {
488 write_exp_elt_opcode (OP_STRING);
489 write_exp_string ($1);
490 write_exp_elt_opcode (OP_STRING);
491 }
492 ;
493
494 variable: name_not_typename
495 { struct symbol *sym = $1.sym;
496
497 if (sym)
498 {
499 if (symbol_read_needs_frame (sym))
500 {
501 if (innermost_block == 0
502 || contained_in (block_found,
503 innermost_block))
504 innermost_block = block_found;
505 }
506 write_exp_elt_opcode (OP_VAR_VALUE);
507 /* We want to use the selected frame, not
508 another more inner frame which happens to
509 be in the same block. */
510 write_exp_elt_block (NULL);
511 write_exp_elt_sym (sym);
512 write_exp_elt_opcode (OP_VAR_VALUE);
513 break;
514 }
515 else
516 {
517 struct minimal_symbol *msymbol;
518 char *arg = copy_name ($1.stoken);
519
520 msymbol =
521 lookup_minimal_symbol (arg, NULL, NULL);
522 if (msymbol != NULL)
523 write_exp_msymbol (msymbol);
524 else if (!have_full_symbols () && !have_partial_symbols ())
525 error (_("No symbol table is loaded. Use the \"file\" command."));
526 else
527 error (_("No symbol \"%s\" in current context."),
528 copy_name ($1.stoken));
529 }
530 }
531 ;
532
533
534 type : ptype
535 ;
536
537 ptype : typebase
538 | typebase abs_decl
539 {
540 /* This is where the interesting stuff happens. */
541 int done = 0;
542 int array_size;
543 struct type *follow_type = $1;
544 struct type *range_type;
545
546 while (!done)
547 switch (pop_type ())
548 {
549 case tp_end:
550 done = 1;
551 break;
552 case tp_pointer:
553 follow_type = lookup_pointer_type (follow_type);
554 break;
555 case tp_reference:
556 follow_type = lookup_reference_type (follow_type);
557 break;
558 case tp_array:
559 array_size = pop_type_int ();
560 if (array_size != -1)
561 {
562 range_type =
563 create_range_type ((struct type *) NULL,
564 parse_f_type->builtin_integer,
565 0, array_size - 1);
566 follow_type =
567 create_array_type ((struct type *) NULL,
568 follow_type, range_type);
569 }
570 else
571 follow_type = lookup_pointer_type (follow_type);
572 break;
573 case tp_function:
574 follow_type = lookup_function_type (follow_type);
575 break;
576 }
577 $$ = follow_type;
578 }
579 ;
580
581 abs_decl: '*'
582 { push_type (tp_pointer); $$ = 0; }
583 | '*' abs_decl
584 { push_type (tp_pointer); $$ = $2; }
585 | '&'
586 { push_type (tp_reference); $$ = 0; }
587 | '&' abs_decl
588 { push_type (tp_reference); $$ = $2; }
589 | direct_abs_decl
590 ;
591
592 direct_abs_decl: '(' abs_decl ')'
593 { $$ = $2; }
594 | direct_abs_decl func_mod
595 { push_type (tp_function); }
596 | func_mod
597 { push_type (tp_function); }
598 ;
599
600 func_mod: '(' ')'
601 { $$ = 0; }
602 | '(' nonempty_typelist ')'
603 { free ($2); $$ = 0; }
604 ;
605
606 typebase /* Implements (approximately): (type-qualifier)* type-specifier */
607 : TYPENAME
608 { $$ = $1.type; }
609 | INT_KEYWORD
610 { $$ = parse_f_type->builtin_integer; }
611 | INT_S2_KEYWORD
612 { $$ = parse_f_type->builtin_integer_s2; }
613 | CHARACTER
614 { $$ = parse_f_type->builtin_character; }
615 | LOGICAL_S8_KEYWORD
616 { $$ = parse_f_type->builtin_logical_s8; }
617 | LOGICAL_KEYWORD
618 { $$ = parse_f_type->builtin_logical; }
619 | LOGICAL_S2_KEYWORD
620 { $$ = parse_f_type->builtin_logical_s2; }
621 | LOGICAL_S1_KEYWORD
622 { $$ = parse_f_type->builtin_logical_s1; }
623 | REAL_KEYWORD
624 { $$ = parse_f_type->builtin_real; }
625 | REAL_S8_KEYWORD
626 { $$ = parse_f_type->builtin_real_s8; }
627 | REAL_S16_KEYWORD
628 { $$ = parse_f_type->builtin_real_s16; }
629 | COMPLEX_S8_KEYWORD
630 { $$ = parse_f_type->builtin_complex_s8; }
631 | COMPLEX_S16_KEYWORD
632 { $$ = parse_f_type->builtin_complex_s16; }
633 | COMPLEX_S32_KEYWORD
634 { $$ = parse_f_type->builtin_complex_s32; }
635 ;
636
637 nonempty_typelist
638 : type
639 { $$ = (struct type **) malloc (sizeof (struct type *) * 2);
640 $<ivec>$[0] = 1; /* Number of types in vector */
641 $$[1] = $1;
642 }
643 | nonempty_typelist ',' type
644 { int len = sizeof (struct type *) * (++($<ivec>1[0]) + 1);
645 $$ = (struct type **) realloc ((char *) $1, len);
646 $$[$<ivec>$[0]] = $3;
647 }
648 ;
649
650 name : NAME
651 { $$ = $1.stoken; }
652 ;
653
654 name_not_typename : NAME
655 /* These would be useful if name_not_typename was useful, but it is just
656 a fake for "variable", so these cause reduce/reduce conflicts because
657 the parser can't tell whether NAME_OR_INT is a name_not_typename (=variable,
658 =exp) or just an exp. If name_not_typename was ever used in an lvalue
659 context where only a name could occur, this might be useful.
660 | NAME_OR_INT
661 */
662 ;
663
664 %%
665
666 /* Take care of parsing a number (anything that starts with a digit).
667 Set yylval and return the token type; update lexptr.
668 LEN is the number of characters in it. */
669
670 /*** Needs some error checking for the float case ***/
671
672 static int
673 parse_number (char *p, int len, int parsed_float, YYSTYPE *putithere)
674 {
675 LONGEST n = 0;
676 LONGEST prevn = 0;
677 int c;
678 int base = input_radix;
679 int unsigned_p = 0;
680 int long_p = 0;
681 ULONGEST high_bit;
682 struct type *signed_type;
683 struct type *unsigned_type;
684
685 if (parsed_float)
686 {
687 /* It's a float since it contains a point or an exponent. */
688 /* [dD] is not understood as an exponent by atof, change it to 'e'. */
689 char *tmp, *tmp2;
690
691 tmp = xstrdup (p);
692 for (tmp2 = tmp; *tmp2; ++tmp2)
693 if (*tmp2 == 'd' || *tmp2 == 'D')
694 *tmp2 = 'e';
695 putithere->dval = atof (tmp);
696 free (tmp);
697 return FLOAT;
698 }
699
700 /* Handle base-switching prefixes 0x, 0t, 0d, 0 */
701 if (p[0] == '0')
702 switch (p[1])
703 {
704 case 'x':
705 case 'X':
706 if (len >= 3)
707 {
708 p += 2;
709 base = 16;
710 len -= 2;
711 }
712 break;
713
714 case 't':
715 case 'T':
716 case 'd':
717 case 'D':
718 if (len >= 3)
719 {
720 p += 2;
721 base = 10;
722 len -= 2;
723 }
724 break;
725
726 default:
727 base = 8;
728 break;
729 }
730
731 while (len-- > 0)
732 {
733 c = *p++;
734 if (isupper (c))
735 c = tolower (c);
736 if (len == 0 && c == 'l')
737 long_p = 1;
738 else if (len == 0 && c == 'u')
739 unsigned_p = 1;
740 else
741 {
742 int i;
743 if (c >= '0' && c <= '9')
744 i = c - '0';
745 else if (c >= 'a' && c <= 'f')
746 i = c - 'a' + 10;
747 else
748 return ERROR; /* Char not a digit */
749 if (i >= base)
750 return ERROR; /* Invalid digit in this base */
751 n *= base;
752 n += i;
753 }
754 /* Portably test for overflow (only works for nonzero values, so make
755 a second check for zero). */
756 if ((prevn >= n) && n != 0)
757 unsigned_p=1; /* Try something unsigned */
758 /* If range checking enabled, portably test for unsigned overflow. */
759 if (RANGE_CHECK && n != 0)
760 {
761 if ((unsigned_p && (unsigned)prevn >= (unsigned)n))
762 range_error (_("Overflow on numeric constant."));
763 }
764 prevn = n;
765 }
766
767 /* If the number is too big to be an int, or it's got an l suffix
768 then it's a long. Work out if this has to be a long by
769 shifting right and seeing if anything remains, and the
770 target int size is different to the target long size.
771
772 In the expression below, we could have tested
773 (n >> gdbarch_int_bit (parse_gdbarch))
774 to see if it was zero,
775 but too many compilers warn about that, when ints and longs
776 are the same size. So we shift it twice, with fewer bits
777 each time, for the same result. */
778
779 if ((gdbarch_int_bit (parse_gdbarch) != gdbarch_long_bit (parse_gdbarch)
780 && ((n >> 2)
781 >> (gdbarch_int_bit (parse_gdbarch)-2))) /* Avoid shift warning */
782 || long_p)
783 {
784 high_bit = ((ULONGEST)1) << (gdbarch_long_bit (parse_gdbarch)-1);
785 unsigned_type = parse_type->builtin_unsigned_long;
786 signed_type = parse_type->builtin_long;
787 }
788 else
789 {
790 high_bit = ((ULONGEST)1) << (gdbarch_int_bit (parse_gdbarch)-1);
791 unsigned_type = parse_type->builtin_unsigned_int;
792 signed_type = parse_type->builtin_int;
793 }
794
795 putithere->typed_val.val = n;
796
797 /* If the high bit of the worked out type is set then this number
798 has to be unsigned. */
799
800 if (unsigned_p || (n & high_bit))
801 putithere->typed_val.type = unsigned_type;
802 else
803 putithere->typed_val.type = signed_type;
804
805 return INT;
806 }
807
808 struct token
809 {
810 char *operator;
811 int token;
812 enum exp_opcode opcode;
813 };
814
815 static const struct token dot_ops[] =
816 {
817 { ".and.", BOOL_AND, BINOP_END },
818 { ".AND.", BOOL_AND, BINOP_END },
819 { ".or.", BOOL_OR, BINOP_END },
820 { ".OR.", BOOL_OR, BINOP_END },
821 { ".not.", BOOL_NOT, BINOP_END },
822 { ".NOT.", BOOL_NOT, BINOP_END },
823 { ".eq.", EQUAL, BINOP_END },
824 { ".EQ.", EQUAL, BINOP_END },
825 { ".eqv.", EQUAL, BINOP_END },
826 { ".NEQV.", NOTEQUAL, BINOP_END },
827 { ".neqv.", NOTEQUAL, BINOP_END },
828 { ".EQV.", EQUAL, BINOP_END },
829 { ".ne.", NOTEQUAL, BINOP_END },
830 { ".NE.", NOTEQUAL, BINOP_END },
831 { ".le.", LEQ, BINOP_END },
832 { ".LE.", LEQ, BINOP_END },
833 { ".ge.", GEQ, BINOP_END },
834 { ".GE.", GEQ, BINOP_END },
835 { ".gt.", GREATERTHAN, BINOP_END },
836 { ".GT.", GREATERTHAN, BINOP_END },
837 { ".lt.", LESSTHAN, BINOP_END },
838 { ".LT.", LESSTHAN, BINOP_END },
839 { NULL, 0, 0 }
840 };
841
842 struct f77_boolean_val
843 {
844 char *name;
845 int value;
846 };
847
848 static const struct f77_boolean_val boolean_values[] =
849 {
850 { ".true.", 1 },
851 { ".TRUE.", 1 },
852 { ".false.", 0 },
853 { ".FALSE.", 0 },
854 { NULL, 0 }
855 };
856
857 static const struct token f77_keywords[] =
858 {
859 { "complex_16", COMPLEX_S16_KEYWORD, BINOP_END },
860 { "complex_32", COMPLEX_S32_KEYWORD, BINOP_END },
861 { "character", CHARACTER, BINOP_END },
862 { "integer_2", INT_S2_KEYWORD, BINOP_END },
863 { "logical_1", LOGICAL_S1_KEYWORD, BINOP_END },
864 { "logical_2", LOGICAL_S2_KEYWORD, BINOP_END },
865 { "logical_8", LOGICAL_S8_KEYWORD, BINOP_END },
866 { "complex_8", COMPLEX_S8_KEYWORD, BINOP_END },
867 { "integer", INT_KEYWORD, BINOP_END },
868 { "logical", LOGICAL_KEYWORD, BINOP_END },
869 { "real_16", REAL_S16_KEYWORD, BINOP_END },
870 { "complex", COMPLEX_S8_KEYWORD, BINOP_END },
871 { "sizeof", SIZEOF, BINOP_END },
872 { "real_8", REAL_S8_KEYWORD, BINOP_END },
873 { "real", REAL_KEYWORD, BINOP_END },
874 { NULL, 0, 0 }
875 };
876
877 /* Implementation of a dynamically expandable buffer for processing input
878 characters acquired through lexptr and building a value to return in
879 yylval. Ripped off from ch-exp.y */
880
881 static char *tempbuf; /* Current buffer contents */
882 static int tempbufsize; /* Size of allocated buffer */
883 static int tempbufindex; /* Current index into buffer */
884
885 #define GROWBY_MIN_SIZE 64 /* Minimum amount to grow buffer by */
886
887 #define CHECKBUF(size) \
888 do { \
889 if (tempbufindex + (size) >= tempbufsize) \
890 { \
891 growbuf_by_size (size); \
892 } \
893 } while (0);
894
895
896 /* Grow the static temp buffer if necessary, including allocating the
897 first one on demand. */
898
899 static void
900 growbuf_by_size (int count)
901 {
902 int growby;
903
904 growby = max (count, GROWBY_MIN_SIZE);
905 tempbufsize += growby;
906 if (tempbuf == NULL)
907 tempbuf = (char *) malloc (tempbufsize);
908 else
909 tempbuf = (char *) realloc (tempbuf, tempbufsize);
910 }
911
912 /* Blatantly ripped off from ch-exp.y. This routine recognizes F77
913 string-literals.
914
915 Recognize a string literal. A string literal is a nonzero sequence
916 of characters enclosed in matching single quotes, except that
917 a single character inside single quotes is a character literal, which
918 we reject as a string literal. To embed the terminator character inside
919 a string, it is simply doubled (I.E. 'this''is''one''string') */
920
921 static int
922 match_string_literal (void)
923 {
924 char *tokptr = lexptr;
925
926 for (tempbufindex = 0, tokptr++; *tokptr != '\0'; tokptr++)
927 {
928 CHECKBUF (1);
929 if (*tokptr == *lexptr)
930 {
931 if (*(tokptr + 1) == *lexptr)
932 tokptr++;
933 else
934 break;
935 }
936 tempbuf[tempbufindex++] = *tokptr;
937 }
938 if (*tokptr == '\0' /* no terminator */
939 || tempbufindex == 0) /* no string */
940 return 0;
941 else
942 {
943 tempbuf[tempbufindex] = '\0';
944 yylval.sval.ptr = tempbuf;
945 yylval.sval.length = tempbufindex;
946 lexptr = ++tokptr;
947 return STRING_LITERAL;
948 }
949 }
950
951 /* Read one token, getting characters through lexptr. */
952
953 static int
954 yylex (void)
955 {
956 int c;
957 int namelen;
958 unsigned int i,token;
959 char *tokstart;
960
961 retry:
962
963 prev_lexptr = lexptr;
964
965 tokstart = lexptr;
966
967 /* First of all, let us make sure we are not dealing with the
968 special tokens .true. and .false. which evaluate to 1 and 0. */
969
970 if (*lexptr == '.')
971 {
972 for (i = 0; boolean_values[i].name != NULL; i++)
973 {
974 if (strncmp (tokstart, boolean_values[i].name,
975 strlen (boolean_values[i].name)) == 0)
976 {
977 lexptr += strlen (boolean_values[i].name);
978 yylval.lval = boolean_values[i].value;
979 return BOOLEAN_LITERAL;
980 }
981 }
982 }
983
984 /* See if it is a special .foo. operator. */
985
986 for (i = 0; dot_ops[i].operator != NULL; i++)
987 if (strncmp (tokstart, dot_ops[i].operator,
988 strlen (dot_ops[i].operator)) == 0)
989 {
990 lexptr += strlen (dot_ops[i].operator);
991 yylval.opcode = dot_ops[i].opcode;
992 return dot_ops[i].token;
993 }
994
995 /* See if it is an exponentiation operator. */
996
997 if (strncmp (tokstart, "**", 2) == 0)
998 {
999 lexptr += 2;
1000 yylval.opcode = BINOP_EXP;
1001 return STARSTAR;
1002 }
1003
1004 switch (c = *tokstart)
1005 {
1006 case 0:
1007 return 0;
1008
1009 case ' ':
1010 case '\t':
1011 case '\n':
1012 lexptr++;
1013 goto retry;
1014
1015 case '\'':
1016 token = match_string_literal ();
1017 if (token != 0)
1018 return (token);
1019 break;
1020
1021 case '(':
1022 paren_depth++;
1023 lexptr++;
1024 return c;
1025
1026 case ')':
1027 if (paren_depth == 0)
1028 return 0;
1029 paren_depth--;
1030 lexptr++;
1031 return c;
1032
1033 case ',':
1034 if (comma_terminates && paren_depth == 0)
1035 return 0;
1036 lexptr++;
1037 return c;
1038
1039 case '.':
1040 /* Might be a floating point number. */
1041 if (lexptr[1] < '0' || lexptr[1] > '9')
1042 goto symbol; /* Nope, must be a symbol. */
1043 /* FALL THRU into number case. */
1044
1045 case '0':
1046 case '1':
1047 case '2':
1048 case '3':
1049 case '4':
1050 case '5':
1051 case '6':
1052 case '7':
1053 case '8':
1054 case '9':
1055 {
1056 /* It's a number. */
1057 int got_dot = 0, got_e = 0, got_d = 0, toktype;
1058 char *p = tokstart;
1059 int hex = input_radix > 10;
1060
1061 if (c == '0' && (p[1] == 'x' || p[1] == 'X'))
1062 {
1063 p += 2;
1064 hex = 1;
1065 }
1066 else if (c == '0' && (p[1]=='t' || p[1]=='T'
1067 || p[1]=='d' || p[1]=='D'))
1068 {
1069 p += 2;
1070 hex = 0;
1071 }
1072
1073 for (;; ++p)
1074 {
1075 if (!hex && !got_e && (*p == 'e' || *p == 'E'))
1076 got_dot = got_e = 1;
1077 else if (!hex && !got_d && (*p == 'd' || *p == 'D'))
1078 got_dot = got_d = 1;
1079 else if (!hex && !got_dot && *p == '.')
1080 got_dot = 1;
1081 else if (((got_e && (p[-1] == 'e' || p[-1] == 'E'))
1082 || (got_d && (p[-1] == 'd' || p[-1] == 'D')))
1083 && (*p == '-' || *p == '+'))
1084 /* This is the sign of the exponent, not the end of the
1085 number. */
1086 continue;
1087 /* We will take any letters or digits. parse_number will
1088 complain if past the radix, or if L or U are not final. */
1089 else if ((*p < '0' || *p > '9')
1090 && ((*p < 'a' || *p > 'z')
1091 && (*p < 'A' || *p > 'Z')))
1092 break;
1093 }
1094 toktype = parse_number (tokstart, p - tokstart, got_dot|got_e|got_d,
1095 &yylval);
1096 if (toktype == ERROR)
1097 {
1098 char *err_copy = (char *) alloca (p - tokstart + 1);
1099
1100 memcpy (err_copy, tokstart, p - tokstart);
1101 err_copy[p - tokstart] = 0;
1102 error (_("Invalid number \"%s\"."), err_copy);
1103 }
1104 lexptr = p;
1105 return toktype;
1106 }
1107
1108 case '+':
1109 case '-':
1110 case '*':
1111 case '/':
1112 case '%':
1113 case '|':
1114 case '&':
1115 case '^':
1116 case '~':
1117 case '!':
1118 case '@':
1119 case '<':
1120 case '>':
1121 case '[':
1122 case ']':
1123 case '?':
1124 case ':':
1125 case '=':
1126 case '{':
1127 case '}':
1128 symbol:
1129 lexptr++;
1130 return c;
1131 }
1132
1133 if (!(c == '_' || c == '$' || c ==':'
1134 || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z')))
1135 /* We must have come across a bad character (e.g. ';'). */
1136 error (_("Invalid character '%c' in expression."), c);
1137
1138 namelen = 0;
1139 for (c = tokstart[namelen];
1140 (c == '_' || c == '$' || c == ':' || (c >= '0' && c <= '9')
1141 || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'));
1142 c = tokstart[++namelen]);
1143
1144 /* The token "if" terminates the expression and is NOT
1145 removed from the input stream. */
1146
1147 if (namelen == 2 && tokstart[0] == 'i' && tokstart[1] == 'f')
1148 return 0;
1149
1150 lexptr += namelen;
1151
1152 /* Catch specific keywords. */
1153
1154 for (i = 0; f77_keywords[i].operator != NULL; i++)
1155 if (strlen (f77_keywords[i].operator) == namelen
1156 && strncmp (tokstart, f77_keywords[i].operator, namelen) == 0)
1157 {
1158 /* lexptr += strlen(f77_keywords[i].operator); */
1159 yylval.opcode = f77_keywords[i].opcode;
1160 return f77_keywords[i].token;
1161 }
1162
1163 yylval.sval.ptr = tokstart;
1164 yylval.sval.length = namelen;
1165
1166 if (*tokstart == '$')
1167 {
1168 write_dollar_variable (yylval.sval);
1169 return VARIABLE;
1170 }
1171
1172 /* Use token-type TYPENAME for symbols that happen to be defined
1173 currently as names of types; NAME for other symbols.
1174 The caller is not constrained to care about the distinction. */
1175 {
1176 char *tmp = copy_name (yylval.sval);
1177 struct symbol *sym;
1178 int is_a_field_of_this = 0;
1179 int hextype;
1180
1181 sym = lookup_symbol (tmp, expression_context_block,
1182 VAR_DOMAIN,
1183 parse_language->la_language == language_cplus
1184 ? &is_a_field_of_this : NULL);
1185 if (sym && SYMBOL_CLASS (sym) == LOC_TYPEDEF)
1186 {
1187 yylval.tsym.type = SYMBOL_TYPE (sym);
1188 return TYPENAME;
1189 }
1190 yylval.tsym.type
1191 = language_lookup_primitive_type_by_name (parse_language,
1192 parse_gdbarch, tmp);
1193 if (yylval.tsym.type != NULL)
1194 return TYPENAME;
1195
1196 /* Input names that aren't symbols but ARE valid hex numbers,
1197 when the input radix permits them, can be names or numbers
1198 depending on the parse. Note we support radixes > 16 here. */
1199 if (!sym
1200 && ((tokstart[0] >= 'a' && tokstart[0] < 'a' + input_radix - 10)
1201 || (tokstart[0] >= 'A' && tokstart[0] < 'A' + input_radix - 10)))
1202 {
1203 YYSTYPE newlval; /* Its value is ignored. */
1204 hextype = parse_number (tokstart, namelen, 0, &newlval);
1205 if (hextype == INT)
1206 {
1207 yylval.ssym.sym = sym;
1208 yylval.ssym.is_a_field_of_this = is_a_field_of_this;
1209 return NAME_OR_INT;
1210 }
1211 }
1212
1213 /* Any other kind of symbol */
1214 yylval.ssym.sym = sym;
1215 yylval.ssym.is_a_field_of_this = is_a_field_of_this;
1216 return NAME;
1217 }
1218 }
1219
1220 void
1221 yyerror (char *msg)
1222 {
1223 if (prev_lexptr)
1224 lexptr = prev_lexptr;
1225
1226 error (_("A %s in expression, near `%s'."), (msg ? msg : "error"), lexptr);
1227 }
This page took 0.061034 seconds and 4 git commands to generate.