gdb: add target_ops::supports_displaced_step
[deliverable/binutils-gdb.git] / gdb / f-exp.y
1
2 /* YACC parser for Fortran expressions, for GDB.
3 Copyright (C) 1986-2020 Free Software 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 "expression.h"
47 #include "value.h"
48 #include "parser-defs.h"
49 #include "language.h"
50 #include "f-lang.h"
51 #include "bfd.h" /* Required by objfiles.h. */
52 #include "symfile.h" /* Required by objfiles.h. */
53 #include "objfiles.h" /* For have_full_symbols and have_partial_symbols */
54 #include "block.h"
55 #include <ctype.h>
56 #include <algorithm>
57 #include "type-stack.h"
58
59 #define parse_type(ps) builtin_type (ps->gdbarch ())
60 #define parse_f_type(ps) builtin_f_type (ps->gdbarch ())
61
62 /* Remap normal yacc parser interface names (yyparse, yylex, yyerror,
63 etc). */
64 #define GDB_YY_REMAP_PREFIX f_
65 #include "yy-remap.h"
66
67 /* The state of the parser, used internally when we are parsing the
68 expression. */
69
70 static struct parser_state *pstate = NULL;
71
72 /* Depth of parentheses. */
73 static int paren_depth;
74
75 /* The current type stack. */
76 static struct type_stack *type_stack;
77
78 int yyparse (void);
79
80 static int yylex (void);
81
82 static void yyerror (const char *);
83
84 static void growbuf_by_size (int);
85
86 static int match_string_literal (void);
87
88 static void push_kind_type (LONGEST val, struct type *type);
89
90 static struct type *convert_to_kind_type (struct type *basetype, int kind);
91
92 %}
93
94 /* Although the yacc "value" of an expression is not used,
95 since the result is stored in the structure being created,
96 other node types do have values. */
97
98 %union
99 {
100 LONGEST lval;
101 struct {
102 LONGEST val;
103 struct type *type;
104 } typed_val;
105 struct {
106 gdb_byte val[16];
107 struct type *type;
108 } typed_val_float;
109 struct symbol *sym;
110 struct type *tval;
111 struct stoken sval;
112 struct ttype tsym;
113 struct symtoken ssym;
114 int voidval;
115 enum exp_opcode opcode;
116 struct internalvar *ivar;
117
118 struct type **tvec;
119 int *ivec;
120 }
121
122 %{
123 /* YYSTYPE gets defined by %union */
124 static int parse_number (struct parser_state *, const char *, int,
125 int, YYSTYPE *);
126 %}
127
128 %type <voidval> exp type_exp start variable
129 %type <tval> type typebase
130 %type <tvec> nonempty_typelist
131 /* %type <bval> block */
132
133 /* Fancy type parsing. */
134 %type <voidval> func_mod direct_abs_decl abs_decl
135 %type <tval> ptype
136
137 %token <typed_val> INT
138 %token <typed_val_float> FLOAT
139
140 /* Both NAME and TYPENAME tokens represent symbols in the input,
141 and both convey their data as strings.
142 But a TYPENAME is a string that happens to be defined as a typedef
143 or builtin type name (such as int or char)
144 and a NAME is any other symbol.
145 Contexts where this distinction is not important can use the
146 nonterminal "name", which matches either NAME or TYPENAME. */
147
148 %token <sval> STRING_LITERAL
149 %token <lval> BOOLEAN_LITERAL
150 %token <ssym> NAME
151 %token <tsym> TYPENAME
152 %type <sval> name
153 %type <ssym> name_not_typename
154
155 /* A NAME_OR_INT is a symbol which is not known in the symbol table,
156 but which would parse as a valid number in the current input radix.
157 E.g. "c" when input_radix==16. Depending on the parse, it will be
158 turned into a name or into a number. */
159
160 %token <ssym> NAME_OR_INT
161
162 %token SIZEOF KIND
163 %token ERROR
164
165 /* Special type cases, put in to allow the parser to distinguish different
166 legal basetypes. */
167 %token INT_KEYWORD INT_S2_KEYWORD LOGICAL_S1_KEYWORD LOGICAL_S2_KEYWORD
168 %token LOGICAL_S8_KEYWORD
169 %token LOGICAL_KEYWORD REAL_KEYWORD REAL_S8_KEYWORD REAL_S16_KEYWORD
170 %token COMPLEX_KEYWORD
171 %token COMPLEX_S8_KEYWORD COMPLEX_S16_KEYWORD COMPLEX_S32_KEYWORD
172 %token BOOL_AND BOOL_OR BOOL_NOT
173 %token SINGLE DOUBLE PRECISION
174 %token <lval> CHARACTER
175
176 %token <voidval> DOLLAR_VARIABLE
177
178 %token <opcode> ASSIGN_MODIFY
179 %token <opcode> UNOP_INTRINSIC BINOP_INTRINSIC
180
181 %left ','
182 %left ABOVE_COMMA
183 %right '=' ASSIGN_MODIFY
184 %right '?'
185 %left BOOL_OR
186 %right BOOL_NOT
187 %left BOOL_AND
188 %left '|'
189 %left '^'
190 %left '&'
191 %left EQUAL NOTEQUAL
192 %left LESSTHAN GREATERTHAN LEQ GEQ
193 %left LSH RSH
194 %left '@'
195 %left '+' '-'
196 %left '*' '/'
197 %right STARSTAR
198 %right '%'
199 %right UNARY
200 %right '('
201
202 \f
203 %%
204
205 start : exp
206 | type_exp
207 ;
208
209 type_exp: type
210 { write_exp_elt_opcode (pstate, OP_TYPE);
211 write_exp_elt_type (pstate, $1);
212 write_exp_elt_opcode (pstate, OP_TYPE); }
213 ;
214
215 exp : '(' exp ')'
216 { }
217 ;
218
219 /* Expressions, not including the comma operator. */
220 exp : '*' exp %prec UNARY
221 { write_exp_elt_opcode (pstate, UNOP_IND); }
222 ;
223
224 exp : '&' exp %prec UNARY
225 { write_exp_elt_opcode (pstate, UNOP_ADDR); }
226 ;
227
228 exp : '-' exp %prec UNARY
229 { write_exp_elt_opcode (pstate, UNOP_NEG); }
230 ;
231
232 exp : BOOL_NOT exp %prec UNARY
233 { write_exp_elt_opcode (pstate, UNOP_LOGICAL_NOT); }
234 ;
235
236 exp : '~' exp %prec UNARY
237 { write_exp_elt_opcode (pstate, UNOP_COMPLEMENT); }
238 ;
239
240 exp : SIZEOF exp %prec UNARY
241 { write_exp_elt_opcode (pstate, UNOP_SIZEOF); }
242 ;
243
244 exp : KIND '(' exp ')' %prec UNARY
245 { write_exp_elt_opcode (pstate, UNOP_FORTRAN_KIND); }
246 ;
247
248 /* No more explicit array operators, we treat everything in F77 as
249 a function call. The disambiguation as to whether we are
250 doing a subscript operation or a function call is done
251 later in eval.c. */
252
253 exp : exp '('
254 { pstate->start_arglist (); }
255 arglist ')'
256 { write_exp_elt_opcode (pstate,
257 OP_F77_UNDETERMINED_ARGLIST);
258 write_exp_elt_longcst (pstate,
259 pstate->end_arglist ());
260 write_exp_elt_opcode (pstate,
261 OP_F77_UNDETERMINED_ARGLIST); }
262 ;
263
264 exp : UNOP_INTRINSIC '(' exp ')'
265 { write_exp_elt_opcode (pstate, $1); }
266 ;
267
268 exp : BINOP_INTRINSIC '(' exp ',' exp ')'
269 { write_exp_elt_opcode (pstate, $1); }
270 ;
271
272 arglist :
273 ;
274
275 arglist : exp
276 { pstate->arglist_len = 1; }
277 ;
278
279 arglist : subrange
280 { pstate->arglist_len = 1; }
281 ;
282
283 arglist : arglist ',' exp %prec ABOVE_COMMA
284 { pstate->arglist_len++; }
285 ;
286
287 /* There are four sorts of subrange types in F90. */
288
289 subrange: exp ':' exp %prec ABOVE_COMMA
290 { write_exp_elt_opcode (pstate, OP_RANGE);
291 write_exp_elt_longcst (pstate, NONE_BOUND_DEFAULT);
292 write_exp_elt_opcode (pstate, OP_RANGE); }
293 ;
294
295 subrange: exp ':' %prec ABOVE_COMMA
296 { write_exp_elt_opcode (pstate, OP_RANGE);
297 write_exp_elt_longcst (pstate, HIGH_BOUND_DEFAULT);
298 write_exp_elt_opcode (pstate, OP_RANGE); }
299 ;
300
301 subrange: ':' exp %prec ABOVE_COMMA
302 { write_exp_elt_opcode (pstate, OP_RANGE);
303 write_exp_elt_longcst (pstate, LOW_BOUND_DEFAULT);
304 write_exp_elt_opcode (pstate, OP_RANGE); }
305 ;
306
307 subrange: ':' %prec ABOVE_COMMA
308 { write_exp_elt_opcode (pstate, OP_RANGE);
309 write_exp_elt_longcst (pstate, BOTH_BOUND_DEFAULT);
310 write_exp_elt_opcode (pstate, OP_RANGE); }
311 ;
312
313 complexnum: exp ',' exp
314 { }
315 ;
316
317 exp : '(' complexnum ')'
318 { write_exp_elt_opcode (pstate, OP_COMPLEX);
319 write_exp_elt_type (pstate,
320 parse_f_type (pstate)
321 ->builtin_complex_s16);
322 write_exp_elt_opcode (pstate, OP_COMPLEX); }
323 ;
324
325 exp : '(' type ')' exp %prec UNARY
326 { write_exp_elt_opcode (pstate, UNOP_CAST);
327 write_exp_elt_type (pstate, $2);
328 write_exp_elt_opcode (pstate, UNOP_CAST); }
329 ;
330
331 exp : exp '%' name
332 { write_exp_elt_opcode (pstate, STRUCTOP_STRUCT);
333 write_exp_string (pstate, $3);
334 write_exp_elt_opcode (pstate, STRUCTOP_STRUCT); }
335 ;
336
337 /* Binary operators in order of decreasing precedence. */
338
339 exp : exp '@' exp
340 { write_exp_elt_opcode (pstate, BINOP_REPEAT); }
341 ;
342
343 exp : exp STARSTAR exp
344 { write_exp_elt_opcode (pstate, BINOP_EXP); }
345 ;
346
347 exp : exp '*' exp
348 { write_exp_elt_opcode (pstate, BINOP_MUL); }
349 ;
350
351 exp : exp '/' exp
352 { write_exp_elt_opcode (pstate, BINOP_DIV); }
353 ;
354
355 exp : exp '+' exp
356 { write_exp_elt_opcode (pstate, BINOP_ADD); }
357 ;
358
359 exp : exp '-' exp
360 { write_exp_elt_opcode (pstate, BINOP_SUB); }
361 ;
362
363 exp : exp LSH exp
364 { write_exp_elt_opcode (pstate, BINOP_LSH); }
365 ;
366
367 exp : exp RSH exp
368 { write_exp_elt_opcode (pstate, BINOP_RSH); }
369 ;
370
371 exp : exp EQUAL exp
372 { write_exp_elt_opcode (pstate, BINOP_EQUAL); }
373 ;
374
375 exp : exp NOTEQUAL exp
376 { write_exp_elt_opcode (pstate, BINOP_NOTEQUAL); }
377 ;
378
379 exp : exp LEQ exp
380 { write_exp_elt_opcode (pstate, BINOP_LEQ); }
381 ;
382
383 exp : exp GEQ exp
384 { write_exp_elt_opcode (pstate, BINOP_GEQ); }
385 ;
386
387 exp : exp LESSTHAN exp
388 { write_exp_elt_opcode (pstate, BINOP_LESS); }
389 ;
390
391 exp : exp GREATERTHAN exp
392 { write_exp_elt_opcode (pstate, BINOP_GTR); }
393 ;
394
395 exp : exp '&' exp
396 { write_exp_elt_opcode (pstate, BINOP_BITWISE_AND); }
397 ;
398
399 exp : exp '^' exp
400 { write_exp_elt_opcode (pstate, BINOP_BITWISE_XOR); }
401 ;
402
403 exp : exp '|' exp
404 { write_exp_elt_opcode (pstate, BINOP_BITWISE_IOR); }
405 ;
406
407 exp : exp BOOL_AND exp
408 { write_exp_elt_opcode (pstate, BINOP_LOGICAL_AND); }
409 ;
410
411
412 exp : exp BOOL_OR exp
413 { write_exp_elt_opcode (pstate, BINOP_LOGICAL_OR); }
414 ;
415
416 exp : exp '=' exp
417 { write_exp_elt_opcode (pstate, BINOP_ASSIGN); }
418 ;
419
420 exp : exp ASSIGN_MODIFY exp
421 { write_exp_elt_opcode (pstate, BINOP_ASSIGN_MODIFY);
422 write_exp_elt_opcode (pstate, $2);
423 write_exp_elt_opcode (pstate, BINOP_ASSIGN_MODIFY); }
424 ;
425
426 exp : INT
427 { write_exp_elt_opcode (pstate, OP_LONG);
428 write_exp_elt_type (pstate, $1.type);
429 write_exp_elt_longcst (pstate, (LONGEST) ($1.val));
430 write_exp_elt_opcode (pstate, OP_LONG); }
431 ;
432
433 exp : NAME_OR_INT
434 { YYSTYPE val;
435 parse_number (pstate, $1.stoken.ptr,
436 $1.stoken.length, 0, &val);
437 write_exp_elt_opcode (pstate, OP_LONG);
438 write_exp_elt_type (pstate, val.typed_val.type);
439 write_exp_elt_longcst (pstate,
440 (LONGEST)val.typed_val.val);
441 write_exp_elt_opcode (pstate, OP_LONG); }
442 ;
443
444 exp : FLOAT
445 { write_exp_elt_opcode (pstate, OP_FLOAT);
446 write_exp_elt_type (pstate, $1.type);
447 write_exp_elt_floatcst (pstate, $1.val);
448 write_exp_elt_opcode (pstate, OP_FLOAT); }
449 ;
450
451 exp : variable
452 ;
453
454 exp : DOLLAR_VARIABLE
455 ;
456
457 exp : SIZEOF '(' type ')' %prec UNARY
458 { write_exp_elt_opcode (pstate, OP_LONG);
459 write_exp_elt_type (pstate,
460 parse_f_type (pstate)
461 ->builtin_integer);
462 $3 = check_typedef ($3);
463 write_exp_elt_longcst (pstate,
464 (LONGEST) TYPE_LENGTH ($3));
465 write_exp_elt_opcode (pstate, OP_LONG); }
466 ;
467
468 exp : BOOLEAN_LITERAL
469 { write_exp_elt_opcode (pstate, OP_BOOL);
470 write_exp_elt_longcst (pstate, (LONGEST) $1);
471 write_exp_elt_opcode (pstate, OP_BOOL);
472 }
473 ;
474
475 exp : STRING_LITERAL
476 {
477 write_exp_elt_opcode (pstate, OP_STRING);
478 write_exp_string (pstate, $1);
479 write_exp_elt_opcode (pstate, OP_STRING);
480 }
481 ;
482
483 variable: name_not_typename
484 { struct block_symbol sym = $1.sym;
485
486 if (sym.symbol)
487 {
488 if (symbol_read_needs_frame (sym.symbol))
489 pstate->block_tracker->update (sym);
490 write_exp_elt_opcode (pstate, OP_VAR_VALUE);
491 write_exp_elt_block (pstate, sym.block);
492 write_exp_elt_sym (pstate, sym.symbol);
493 write_exp_elt_opcode (pstate, OP_VAR_VALUE);
494 break;
495 }
496 else
497 {
498 struct bound_minimal_symbol msymbol;
499 std::string arg = copy_name ($1.stoken);
500
501 msymbol =
502 lookup_bound_minimal_symbol (arg.c_str ());
503 if (msymbol.minsym != NULL)
504 write_exp_msymbol (pstate, msymbol);
505 else if (!have_full_symbols () && !have_partial_symbols ())
506 error (_("No symbol table is loaded. Use the \"file\" command."));
507 else
508 error (_("No symbol \"%s\" in current context."),
509 arg.c_str ());
510 }
511 }
512 ;
513
514
515 type : ptype
516 ;
517
518 ptype : typebase
519 | typebase abs_decl
520 {
521 /* This is where the interesting stuff happens. */
522 int done = 0;
523 int array_size;
524 struct type *follow_type = $1;
525 struct type *range_type;
526
527 while (!done)
528 switch (type_stack->pop ())
529 {
530 case tp_end:
531 done = 1;
532 break;
533 case tp_pointer:
534 follow_type = lookup_pointer_type (follow_type);
535 break;
536 case tp_reference:
537 follow_type = lookup_lvalue_reference_type (follow_type);
538 break;
539 case tp_array:
540 array_size = type_stack->pop_int ();
541 if (array_size != -1)
542 {
543 range_type =
544 create_static_range_type ((struct type *) NULL,
545 parse_f_type (pstate)
546 ->builtin_integer,
547 0, array_size - 1);
548 follow_type =
549 create_array_type ((struct type *) NULL,
550 follow_type, range_type);
551 }
552 else
553 follow_type = lookup_pointer_type (follow_type);
554 break;
555 case tp_function:
556 follow_type = lookup_function_type (follow_type);
557 break;
558 case tp_kind:
559 {
560 int kind_val = type_stack->pop_int ();
561 follow_type
562 = convert_to_kind_type (follow_type, kind_val);
563 }
564 break;
565 }
566 $$ = follow_type;
567 }
568 ;
569
570 abs_decl: '*'
571 { type_stack->push (tp_pointer); $$ = 0; }
572 | '*' abs_decl
573 { type_stack->push (tp_pointer); $$ = $2; }
574 | '&'
575 { type_stack->push (tp_reference); $$ = 0; }
576 | '&' abs_decl
577 { type_stack->push (tp_reference); $$ = $2; }
578 | direct_abs_decl
579 ;
580
581 direct_abs_decl: '(' abs_decl ')'
582 { $$ = $2; }
583 | '(' KIND '=' INT ')'
584 { push_kind_type ($4.val, $4.type); }
585 | '*' INT
586 { push_kind_type ($2.val, $2.type); }
587 | direct_abs_decl func_mod
588 { type_stack->push (tp_function); }
589 | func_mod
590 { type_stack->push (tp_function); }
591 ;
592
593 func_mod: '(' ')'
594 { $$ = 0; }
595 | '(' nonempty_typelist ')'
596 { free ($2); $$ = 0; }
597 ;
598
599 typebase /* Implements (approximately): (type-qualifier)* type-specifier */
600 : TYPENAME
601 { $$ = $1.type; }
602 | INT_KEYWORD
603 { $$ = parse_f_type (pstate)->builtin_integer; }
604 | INT_S2_KEYWORD
605 { $$ = parse_f_type (pstate)->builtin_integer_s2; }
606 | CHARACTER
607 { $$ = parse_f_type (pstate)->builtin_character; }
608 | LOGICAL_S8_KEYWORD
609 { $$ = parse_f_type (pstate)->builtin_logical_s8; }
610 | LOGICAL_KEYWORD
611 { $$ = parse_f_type (pstate)->builtin_logical; }
612 | LOGICAL_S2_KEYWORD
613 { $$ = parse_f_type (pstate)->builtin_logical_s2; }
614 | LOGICAL_S1_KEYWORD
615 { $$ = parse_f_type (pstate)->builtin_logical_s1; }
616 | REAL_KEYWORD
617 { $$ = parse_f_type (pstate)->builtin_real; }
618 | REAL_S8_KEYWORD
619 { $$ = parse_f_type (pstate)->builtin_real_s8; }
620 | REAL_S16_KEYWORD
621 { $$ = parse_f_type (pstate)->builtin_real_s16; }
622 | COMPLEX_KEYWORD
623 { $$ = parse_f_type (pstate)->builtin_complex_s8; }
624 | COMPLEX_S8_KEYWORD
625 { $$ = parse_f_type (pstate)->builtin_complex_s8; }
626 | COMPLEX_S16_KEYWORD
627 { $$ = parse_f_type (pstate)->builtin_complex_s16; }
628 | COMPLEX_S32_KEYWORD
629 { $$ = parse_f_type (pstate)->builtin_complex_s32; }
630 | SINGLE PRECISION
631 { $$ = parse_f_type (pstate)->builtin_real;}
632 | DOUBLE PRECISION
633 { $$ = parse_f_type (pstate)->builtin_real_s8;}
634 | SINGLE COMPLEX_KEYWORD
635 { $$ = parse_f_type (pstate)->builtin_complex_s8;}
636 | DOUBLE COMPLEX_KEYWORD
637 { $$ = parse_f_type (pstate)->builtin_complex_s16;}
638 ;
639
640 nonempty_typelist
641 : type
642 { $$ = (struct type **) malloc (sizeof (struct type *) * 2);
643 $<ivec>$[0] = 1; /* Number of types in vector */
644 $$[1] = $1;
645 }
646 | nonempty_typelist ',' type
647 { int len = sizeof (struct type *) * (++($<ivec>1[0]) + 1);
648 $$ = (struct type **) realloc ((char *) $1, len);
649 $$[$<ivec>$[0]] = $3;
650 }
651 ;
652
653 name : NAME
654 { $$ = $1.stoken; }
655 ;
656
657 name_not_typename : NAME
658 /* These would be useful if name_not_typename was useful, but it is just
659 a fake for "variable", so these cause reduce/reduce conflicts because
660 the parser can't tell whether NAME_OR_INT is a name_not_typename (=variable,
661 =exp) or just an exp. If name_not_typename was ever used in an lvalue
662 context where only a name could occur, this might be useful.
663 | NAME_OR_INT
664 */
665 ;
666
667 %%
668
669 /* Take care of parsing a number (anything that starts with a digit).
670 Set yylval and return the token type; update lexptr.
671 LEN is the number of characters in it. */
672
673 /*** Needs some error checking for the float case ***/
674
675 static int
676 parse_number (struct parser_state *par_state,
677 const char *p, int len, int parsed_float, YYSTYPE *putithere)
678 {
679 LONGEST n = 0;
680 LONGEST prevn = 0;
681 int c;
682 int base = input_radix;
683 int unsigned_p = 0;
684 int long_p = 0;
685 ULONGEST high_bit;
686 struct type *signed_type;
687 struct type *unsigned_type;
688
689 if (parsed_float)
690 {
691 /* It's a float since it contains a point or an exponent. */
692 /* [dD] is not understood as an exponent by parse_float,
693 change it to 'e'. */
694 char *tmp, *tmp2;
695
696 tmp = xstrdup (p);
697 for (tmp2 = tmp; *tmp2; ++tmp2)
698 if (*tmp2 == 'd' || *tmp2 == 'D')
699 *tmp2 = 'e';
700
701 /* FIXME: Should this use different types? */
702 putithere->typed_val_float.type = parse_f_type (pstate)->builtin_real_s8;
703 bool parsed = parse_float (tmp, len,
704 putithere->typed_val_float.type,
705 putithere->typed_val_float.val);
706 free (tmp);
707 return parsed? FLOAT : ERROR;
708 }
709
710 /* Handle base-switching prefixes 0x, 0t, 0d, 0 */
711 if (p[0] == '0')
712 switch (p[1])
713 {
714 case 'x':
715 case 'X':
716 if (len >= 3)
717 {
718 p += 2;
719 base = 16;
720 len -= 2;
721 }
722 break;
723
724 case 't':
725 case 'T':
726 case 'd':
727 case 'D':
728 if (len >= 3)
729 {
730 p += 2;
731 base = 10;
732 len -= 2;
733 }
734 break;
735
736 default:
737 base = 8;
738 break;
739 }
740
741 while (len-- > 0)
742 {
743 c = *p++;
744 if (isupper (c))
745 c = tolower (c);
746 if (len == 0 && c == 'l')
747 long_p = 1;
748 else if (len == 0 && c == 'u')
749 unsigned_p = 1;
750 else
751 {
752 int i;
753 if (c >= '0' && c <= '9')
754 i = c - '0';
755 else if (c >= 'a' && c <= 'f')
756 i = c - 'a' + 10;
757 else
758 return ERROR; /* Char not a digit */
759 if (i >= base)
760 return ERROR; /* Invalid digit in this base */
761 n *= base;
762 n += i;
763 }
764 /* Portably test for overflow (only works for nonzero values, so make
765 a second check for zero). */
766 if ((prevn >= n) && n != 0)
767 unsigned_p=1; /* Try something unsigned */
768 /* If range checking enabled, portably test for unsigned overflow. */
769 if (RANGE_CHECK && n != 0)
770 {
771 if ((unsigned_p && (unsigned)prevn >= (unsigned)n))
772 range_error (_("Overflow on numeric constant."));
773 }
774 prevn = n;
775 }
776
777 /* If the number is too big to be an int, or it's got an l suffix
778 then it's a long. Work out if this has to be a long by
779 shifting right and seeing if anything remains, and the
780 target int size is different to the target long size.
781
782 In the expression below, we could have tested
783 (n >> gdbarch_int_bit (parse_gdbarch))
784 to see if it was zero,
785 but too many compilers warn about that, when ints and longs
786 are the same size. So we shift it twice, with fewer bits
787 each time, for the same result. */
788
789 if ((gdbarch_int_bit (par_state->gdbarch ())
790 != gdbarch_long_bit (par_state->gdbarch ())
791 && ((n >> 2)
792 >> (gdbarch_int_bit (par_state->gdbarch ())-2))) /* Avoid
793 shift warning */
794 || long_p)
795 {
796 high_bit = ((ULONGEST)1)
797 << (gdbarch_long_bit (par_state->gdbarch ())-1);
798 unsigned_type = parse_type (par_state)->builtin_unsigned_long;
799 signed_type = parse_type (par_state)->builtin_long;
800 }
801 else
802 {
803 high_bit =
804 ((ULONGEST)1) << (gdbarch_int_bit (par_state->gdbarch ()) - 1);
805 unsigned_type = parse_type (par_state)->builtin_unsigned_int;
806 signed_type = parse_type (par_state)->builtin_int;
807 }
808
809 putithere->typed_val.val = n;
810
811 /* If the high bit of the worked out type is set then this number
812 has to be unsigned. */
813
814 if (unsigned_p || (n & high_bit))
815 putithere->typed_val.type = unsigned_type;
816 else
817 putithere->typed_val.type = signed_type;
818
819 return INT;
820 }
821
822 /* Called to setup the type stack when we encounter a '(kind=N)' type
823 modifier, performs some bounds checking on 'N' and then pushes this to
824 the type stack followed by the 'tp_kind' marker. */
825 static void
826 push_kind_type (LONGEST val, struct type *type)
827 {
828 int ival;
829
830 if (TYPE_UNSIGNED (type))
831 {
832 ULONGEST uval = static_cast <ULONGEST> (val);
833 if (uval > INT_MAX)
834 error (_("kind value out of range"));
835 ival = static_cast <int> (uval);
836 }
837 else
838 {
839 if (val > INT_MAX || val < 0)
840 error (_("kind value out of range"));
841 ival = static_cast <int> (val);
842 }
843
844 type_stack->push (ival);
845 type_stack->push (tp_kind);
846 }
847
848 /* Called when a type has a '(kind=N)' modifier after it, for example
849 'character(kind=1)'. The BASETYPE is the type described by 'character'
850 in our example, and KIND is the integer '1'. This function returns a
851 new type that represents the basetype of a specific kind. */
852 static struct type *
853 convert_to_kind_type (struct type *basetype, int kind)
854 {
855 if (basetype == parse_f_type (pstate)->builtin_character)
856 {
857 /* Character of kind 1 is a special case, this is the same as the
858 base character type. */
859 if (kind == 1)
860 return parse_f_type (pstate)->builtin_character;
861 }
862 else if (basetype == parse_f_type (pstate)->builtin_complex_s8)
863 {
864 if (kind == 4)
865 return parse_f_type (pstate)->builtin_complex_s8;
866 else if (kind == 8)
867 return parse_f_type (pstate)->builtin_complex_s16;
868 else if (kind == 16)
869 return parse_f_type (pstate)->builtin_complex_s32;
870 }
871 else if (basetype == parse_f_type (pstate)->builtin_real)
872 {
873 if (kind == 4)
874 return parse_f_type (pstate)->builtin_real;
875 else if (kind == 8)
876 return parse_f_type (pstate)->builtin_real_s8;
877 else if (kind == 16)
878 return parse_f_type (pstate)->builtin_real_s16;
879 }
880 else if (basetype == parse_f_type (pstate)->builtin_logical)
881 {
882 if (kind == 1)
883 return parse_f_type (pstate)->builtin_logical_s1;
884 else if (kind == 2)
885 return parse_f_type (pstate)->builtin_logical_s2;
886 else if (kind == 4)
887 return parse_f_type (pstate)->builtin_logical;
888 else if (kind == 8)
889 return parse_f_type (pstate)->builtin_logical_s8;
890 }
891 else if (basetype == parse_f_type (pstate)->builtin_integer)
892 {
893 if (kind == 2)
894 return parse_f_type (pstate)->builtin_integer_s2;
895 else if (kind == 4)
896 return parse_f_type (pstate)->builtin_integer;
897 else if (kind == 8)
898 return parse_f_type (pstate)->builtin_integer_s8;
899 }
900
901 error (_("unsupported kind %d for type %s"),
902 kind, TYPE_SAFE_NAME (basetype));
903
904 /* Should never get here. */
905 return nullptr;
906 }
907
908 struct token
909 {
910 /* The string to match against. */
911 const char *oper;
912
913 /* The lexer token to return. */
914 int token;
915
916 /* The expression opcode to embed within the token. */
917 enum exp_opcode opcode;
918
919 /* When this is true the string in OPER is matched exactly including
920 case, when this is false OPER is matched case insensitively. */
921 bool case_sensitive;
922 };
923
924 static const struct token dot_ops[] =
925 {
926 { ".and.", BOOL_AND, BINOP_END, false },
927 { ".or.", BOOL_OR, BINOP_END, false },
928 { ".not.", BOOL_NOT, BINOP_END, false },
929 { ".eq.", EQUAL, BINOP_END, false },
930 { ".eqv.", EQUAL, BINOP_END, false },
931 { ".neqv.", NOTEQUAL, BINOP_END, false },
932 { ".ne.", NOTEQUAL, BINOP_END, false },
933 { ".le.", LEQ, BINOP_END, false },
934 { ".ge.", GEQ, BINOP_END, false },
935 { ".gt.", GREATERTHAN, BINOP_END, false },
936 { ".lt.", LESSTHAN, BINOP_END, false },
937 };
938
939 /* Holds the Fortran representation of a boolean, and the integer value we
940 substitute in when one of the matching strings is parsed. */
941 struct f77_boolean_val
942 {
943 /* The string representing a Fortran boolean. */
944 const char *name;
945
946 /* The integer value to replace it with. */
947 int value;
948 };
949
950 /* The set of Fortran booleans. These are matched case insensitively. */
951 static const struct f77_boolean_val boolean_values[] =
952 {
953 { ".true.", 1 },
954 { ".false.", 0 }
955 };
956
957 static const struct token f77_keywords[] =
958 {
959 /* Historically these have always been lowercase only in GDB. */
960 { "complex_16", COMPLEX_S16_KEYWORD, BINOP_END, true },
961 { "complex_32", COMPLEX_S32_KEYWORD, BINOP_END, true },
962 { "character", CHARACTER, BINOP_END, true },
963 { "integer_2", INT_S2_KEYWORD, BINOP_END, true },
964 { "logical_1", LOGICAL_S1_KEYWORD, BINOP_END, true },
965 { "logical_2", LOGICAL_S2_KEYWORD, BINOP_END, true },
966 { "logical_8", LOGICAL_S8_KEYWORD, BINOP_END, true },
967 { "complex_8", COMPLEX_S8_KEYWORD, BINOP_END, true },
968 { "integer", INT_KEYWORD, BINOP_END, true },
969 { "logical", LOGICAL_KEYWORD, BINOP_END, true },
970 { "real_16", REAL_S16_KEYWORD, BINOP_END, true },
971 { "complex", COMPLEX_KEYWORD, BINOP_END, true },
972 { "sizeof", SIZEOF, BINOP_END, true },
973 { "real_8", REAL_S8_KEYWORD, BINOP_END, true },
974 { "real", REAL_KEYWORD, BINOP_END, true },
975 { "single", SINGLE, BINOP_END, true },
976 { "double", DOUBLE, BINOP_END, true },
977 { "precision", PRECISION, BINOP_END, true },
978 /* The following correspond to actual functions in Fortran and are case
979 insensitive. */
980 { "kind", KIND, BINOP_END, false },
981 { "abs", UNOP_INTRINSIC, UNOP_ABS, false },
982 { "mod", BINOP_INTRINSIC, BINOP_MOD, false },
983 { "floor", UNOP_INTRINSIC, UNOP_FORTRAN_FLOOR, false },
984 { "ceiling", UNOP_INTRINSIC, UNOP_FORTRAN_CEILING, false },
985 { "modulo", BINOP_INTRINSIC, BINOP_FORTRAN_MODULO, false },
986 { "cmplx", BINOP_INTRINSIC, BINOP_FORTRAN_CMPLX, false },
987 };
988
989 /* Implementation of a dynamically expandable buffer for processing input
990 characters acquired through lexptr and building a value to return in
991 yylval. Ripped off from ch-exp.y */
992
993 static char *tempbuf; /* Current buffer contents */
994 static int tempbufsize; /* Size of allocated buffer */
995 static int tempbufindex; /* Current index into buffer */
996
997 #define GROWBY_MIN_SIZE 64 /* Minimum amount to grow buffer by */
998
999 #define CHECKBUF(size) \
1000 do { \
1001 if (tempbufindex + (size) >= tempbufsize) \
1002 { \
1003 growbuf_by_size (size); \
1004 } \
1005 } while (0);
1006
1007
1008 /* Grow the static temp buffer if necessary, including allocating the
1009 first one on demand. */
1010
1011 static void
1012 growbuf_by_size (int count)
1013 {
1014 int growby;
1015
1016 growby = std::max (count, GROWBY_MIN_SIZE);
1017 tempbufsize += growby;
1018 if (tempbuf == NULL)
1019 tempbuf = (char *) malloc (tempbufsize);
1020 else
1021 tempbuf = (char *) realloc (tempbuf, tempbufsize);
1022 }
1023
1024 /* Blatantly ripped off from ch-exp.y. This routine recognizes F77
1025 string-literals.
1026
1027 Recognize a string literal. A string literal is a nonzero sequence
1028 of characters enclosed in matching single quotes, except that
1029 a single character inside single quotes is a character literal, which
1030 we reject as a string literal. To embed the terminator character inside
1031 a string, it is simply doubled (I.E. 'this''is''one''string') */
1032
1033 static int
1034 match_string_literal (void)
1035 {
1036 const char *tokptr = pstate->lexptr;
1037
1038 for (tempbufindex = 0, tokptr++; *tokptr != '\0'; tokptr++)
1039 {
1040 CHECKBUF (1);
1041 if (*tokptr == *pstate->lexptr)
1042 {
1043 if (*(tokptr + 1) == *pstate->lexptr)
1044 tokptr++;
1045 else
1046 break;
1047 }
1048 tempbuf[tempbufindex++] = *tokptr;
1049 }
1050 if (*tokptr == '\0' /* no terminator */
1051 || tempbufindex == 0) /* no string */
1052 return 0;
1053 else
1054 {
1055 tempbuf[tempbufindex] = '\0';
1056 yylval.sval.ptr = tempbuf;
1057 yylval.sval.length = tempbufindex;
1058 pstate->lexptr = ++tokptr;
1059 return STRING_LITERAL;
1060 }
1061 }
1062
1063 /* Read one token, getting characters through lexptr. */
1064
1065 static int
1066 yylex (void)
1067 {
1068 int c;
1069 int namelen;
1070 unsigned int token;
1071 const char *tokstart;
1072
1073 retry:
1074
1075 pstate->prev_lexptr = pstate->lexptr;
1076
1077 tokstart = pstate->lexptr;
1078
1079 /* First of all, let us make sure we are not dealing with the
1080 special tokens .true. and .false. which evaluate to 1 and 0. */
1081
1082 if (*pstate->lexptr == '.')
1083 {
1084 for (int i = 0; i < ARRAY_SIZE (boolean_values); i++)
1085 {
1086 if (strncasecmp (tokstart, boolean_values[i].name,
1087 strlen (boolean_values[i].name)) == 0)
1088 {
1089 pstate->lexptr += strlen (boolean_values[i].name);
1090 yylval.lval = boolean_values[i].value;
1091 return BOOLEAN_LITERAL;
1092 }
1093 }
1094 }
1095
1096 /* See if it is a special .foo. operator. */
1097 for (int i = 0; i < ARRAY_SIZE (dot_ops); i++)
1098 if (strncasecmp (tokstart, dot_ops[i].oper,
1099 strlen (dot_ops[i].oper)) == 0)
1100 {
1101 gdb_assert (!dot_ops[i].case_sensitive);
1102 pstate->lexptr += strlen (dot_ops[i].oper);
1103 yylval.opcode = dot_ops[i].opcode;
1104 return dot_ops[i].token;
1105 }
1106
1107 /* See if it is an exponentiation operator. */
1108
1109 if (strncmp (tokstart, "**", 2) == 0)
1110 {
1111 pstate->lexptr += 2;
1112 yylval.opcode = BINOP_EXP;
1113 return STARSTAR;
1114 }
1115
1116 switch (c = *tokstart)
1117 {
1118 case 0:
1119 return 0;
1120
1121 case ' ':
1122 case '\t':
1123 case '\n':
1124 pstate->lexptr++;
1125 goto retry;
1126
1127 case '\'':
1128 token = match_string_literal ();
1129 if (token != 0)
1130 return (token);
1131 break;
1132
1133 case '(':
1134 paren_depth++;
1135 pstate->lexptr++;
1136 return c;
1137
1138 case ')':
1139 if (paren_depth == 0)
1140 return 0;
1141 paren_depth--;
1142 pstate->lexptr++;
1143 return c;
1144
1145 case ',':
1146 if (pstate->comma_terminates && paren_depth == 0)
1147 return 0;
1148 pstate->lexptr++;
1149 return c;
1150
1151 case '.':
1152 /* Might be a floating point number. */
1153 if (pstate->lexptr[1] < '0' || pstate->lexptr[1] > '9')
1154 goto symbol; /* Nope, must be a symbol. */
1155 /* FALL THRU. */
1156
1157 case '0':
1158 case '1':
1159 case '2':
1160 case '3':
1161 case '4':
1162 case '5':
1163 case '6':
1164 case '7':
1165 case '8':
1166 case '9':
1167 {
1168 /* It's a number. */
1169 int got_dot = 0, got_e = 0, got_d = 0, toktype;
1170 const char *p = tokstart;
1171 int hex = input_radix > 10;
1172
1173 if (c == '0' && (p[1] == 'x' || p[1] == 'X'))
1174 {
1175 p += 2;
1176 hex = 1;
1177 }
1178 else if (c == '0' && (p[1]=='t' || p[1]=='T'
1179 || p[1]=='d' || p[1]=='D'))
1180 {
1181 p += 2;
1182 hex = 0;
1183 }
1184
1185 for (;; ++p)
1186 {
1187 if (!hex && !got_e && (*p == 'e' || *p == 'E'))
1188 got_dot = got_e = 1;
1189 else if (!hex && !got_d && (*p == 'd' || *p == 'D'))
1190 got_dot = got_d = 1;
1191 else if (!hex && !got_dot && *p == '.')
1192 got_dot = 1;
1193 else if (((got_e && (p[-1] == 'e' || p[-1] == 'E'))
1194 || (got_d && (p[-1] == 'd' || p[-1] == 'D')))
1195 && (*p == '-' || *p == '+'))
1196 /* This is the sign of the exponent, not the end of the
1197 number. */
1198 continue;
1199 /* We will take any letters or digits. parse_number will
1200 complain if past the radix, or if L or U are not final. */
1201 else if ((*p < '0' || *p > '9')
1202 && ((*p < 'a' || *p > 'z')
1203 && (*p < 'A' || *p > 'Z')))
1204 break;
1205 }
1206 toktype = parse_number (pstate, tokstart, p - tokstart,
1207 got_dot|got_e|got_d,
1208 &yylval);
1209 if (toktype == ERROR)
1210 {
1211 char *err_copy = (char *) alloca (p - tokstart + 1);
1212
1213 memcpy (err_copy, tokstart, p - tokstart);
1214 err_copy[p - tokstart] = 0;
1215 error (_("Invalid number \"%s\"."), err_copy);
1216 }
1217 pstate->lexptr = p;
1218 return toktype;
1219 }
1220
1221 case '+':
1222 case '-':
1223 case '*':
1224 case '/':
1225 case '%':
1226 case '|':
1227 case '&':
1228 case '^':
1229 case '~':
1230 case '!':
1231 case '@':
1232 case '<':
1233 case '>':
1234 case '[':
1235 case ']':
1236 case '?':
1237 case ':':
1238 case '=':
1239 case '{':
1240 case '}':
1241 symbol:
1242 pstate->lexptr++;
1243 return c;
1244 }
1245
1246 if (!(c == '_' || c == '$' || c ==':'
1247 || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z')))
1248 /* We must have come across a bad character (e.g. ';'). */
1249 error (_("Invalid character '%c' in expression."), c);
1250
1251 namelen = 0;
1252 for (c = tokstart[namelen];
1253 (c == '_' || c == '$' || c == ':' || (c >= '0' && c <= '9')
1254 || (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'));
1255 c = tokstart[++namelen]);
1256
1257 /* The token "if" terminates the expression and is NOT
1258 removed from the input stream. */
1259
1260 if (namelen == 2 && tokstart[0] == 'i' && tokstart[1] == 'f')
1261 return 0;
1262
1263 pstate->lexptr += namelen;
1264
1265 /* Catch specific keywords. */
1266
1267 for (int i = 0; i < ARRAY_SIZE (f77_keywords); i++)
1268 if (strlen (f77_keywords[i].oper) == namelen
1269 && ((!f77_keywords[i].case_sensitive
1270 && strncasecmp (tokstart, f77_keywords[i].oper, namelen) == 0)
1271 || (f77_keywords[i].case_sensitive
1272 && strncmp (tokstart, f77_keywords[i].oper, namelen) == 0)))
1273 {
1274 yylval.opcode = f77_keywords[i].opcode;
1275 return f77_keywords[i].token;
1276 }
1277
1278 yylval.sval.ptr = tokstart;
1279 yylval.sval.length = namelen;
1280
1281 if (*tokstart == '$')
1282 {
1283 write_dollar_variable (pstate, yylval.sval);
1284 return DOLLAR_VARIABLE;
1285 }
1286
1287 /* Use token-type TYPENAME for symbols that happen to be defined
1288 currently as names of types; NAME for other symbols.
1289 The caller is not constrained to care about the distinction. */
1290 {
1291 std::string tmp = copy_name (yylval.sval);
1292 struct block_symbol result;
1293 enum domain_enum_tag lookup_domains[] =
1294 {
1295 STRUCT_DOMAIN,
1296 VAR_DOMAIN,
1297 MODULE_DOMAIN
1298 };
1299 int hextype;
1300
1301 for (int i = 0; i < ARRAY_SIZE (lookup_domains); ++i)
1302 {
1303 result = lookup_symbol (tmp.c_str (), pstate->expression_context_block,
1304 lookup_domains[i], NULL);
1305 if (result.symbol && SYMBOL_CLASS (result.symbol) == LOC_TYPEDEF)
1306 {
1307 yylval.tsym.type = SYMBOL_TYPE (result.symbol);
1308 return TYPENAME;
1309 }
1310
1311 if (result.symbol)
1312 break;
1313 }
1314
1315 yylval.tsym.type
1316 = language_lookup_primitive_type (pstate->language (),
1317 pstate->gdbarch (), tmp.c_str ());
1318 if (yylval.tsym.type != NULL)
1319 return TYPENAME;
1320
1321 /* Input names that aren't symbols but ARE valid hex numbers,
1322 when the input radix permits them, can be names or numbers
1323 depending on the parse. Note we support radixes > 16 here. */
1324 if (!result.symbol
1325 && ((tokstart[0] >= 'a' && tokstart[0] < 'a' + input_radix - 10)
1326 || (tokstart[0] >= 'A' && tokstart[0] < 'A' + input_radix - 10)))
1327 {
1328 YYSTYPE newlval; /* Its value is ignored. */
1329 hextype = parse_number (pstate, tokstart, namelen, 0, &newlval);
1330 if (hextype == INT)
1331 {
1332 yylval.ssym.sym = result;
1333 yylval.ssym.is_a_field_of_this = false;
1334 return NAME_OR_INT;
1335 }
1336 }
1337
1338 /* Any other kind of symbol */
1339 yylval.ssym.sym = result;
1340 yylval.ssym.is_a_field_of_this = false;
1341 return NAME;
1342 }
1343 }
1344
1345 int
1346 f_parse (struct parser_state *par_state)
1347 {
1348 /* Setting up the parser state. */
1349 scoped_restore pstate_restore = make_scoped_restore (&pstate);
1350 scoped_restore restore_yydebug = make_scoped_restore (&yydebug,
1351 parser_debug);
1352 gdb_assert (par_state != NULL);
1353 pstate = par_state;
1354 paren_depth = 0;
1355
1356 struct type_stack stack;
1357 scoped_restore restore_type_stack = make_scoped_restore (&type_stack,
1358 &stack);
1359
1360 return yyparse ();
1361 }
1362
1363 static void
1364 yyerror (const char *msg)
1365 {
1366 if (pstate->prev_lexptr)
1367 pstate->lexptr = pstate->prev_lexptr;
1368
1369 error (_("A %s in expression, near `%s'."), msg, pstate->lexptr);
1370 }
This page took 0.059269 seconds and 4 git commands to generate.