aarch64: Use gdb_sys_no_syscall enum instead of -1
[deliverable/binutils-gdb.git] / gdb / ada-lang.c
1 /* Ada language support routines for GDB, the GNU debugger.
2
3 Copyright (C) 1992-2015 Free Software Foundation, Inc.
4
5 This file is part of GDB.
6
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
11
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License
18 along with this program. If not, see <http://www.gnu.org/licenses/>. */
19
20
21 #include "defs.h"
22 #include <ctype.h>
23 #include "demangle.h"
24 #include "gdb_regex.h"
25 #include "frame.h"
26 #include "symtab.h"
27 #include "gdbtypes.h"
28 #include "gdbcmd.h"
29 #include "expression.h"
30 #include "parser-defs.h"
31 #include "language.h"
32 #include "varobj.h"
33 #include "c-lang.h"
34 #include "inferior.h"
35 #include "symfile.h"
36 #include "objfiles.h"
37 #include "breakpoint.h"
38 #include "gdbcore.h"
39 #include "hashtab.h"
40 #include "gdb_obstack.h"
41 #include "ada-lang.h"
42 #include "completer.h"
43 #include <sys/stat.h>
44 #include "ui-out.h"
45 #include "block.h"
46 #include "infcall.h"
47 #include "dictionary.h"
48 #include "annotate.h"
49 #include "valprint.h"
50 #include "source.h"
51 #include "observer.h"
52 #include "vec.h"
53 #include "stack.h"
54 #include "gdb_vecs.h"
55 #include "typeprint.h"
56 #include "namespace.h"
57
58 #include "psymtab.h"
59 #include "value.h"
60 #include "mi/mi-common.h"
61 #include "arch-utils.h"
62 #include "cli/cli-utils.h"
63
64 /* Define whether or not the C operator '/' truncates towards zero for
65 differently signed operands (truncation direction is undefined in C).
66 Copied from valarith.c. */
67
68 #ifndef TRUNCATION_TOWARDS_ZERO
69 #define TRUNCATION_TOWARDS_ZERO ((-5 / 2) == -2)
70 #endif
71
72 static struct type *desc_base_type (struct type *);
73
74 static struct type *desc_bounds_type (struct type *);
75
76 static struct value *desc_bounds (struct value *);
77
78 static int fat_pntr_bounds_bitpos (struct type *);
79
80 static int fat_pntr_bounds_bitsize (struct type *);
81
82 static struct type *desc_data_target_type (struct type *);
83
84 static struct value *desc_data (struct value *);
85
86 static int fat_pntr_data_bitpos (struct type *);
87
88 static int fat_pntr_data_bitsize (struct type *);
89
90 static struct value *desc_one_bound (struct value *, int, int);
91
92 static int desc_bound_bitpos (struct type *, int, int);
93
94 static int desc_bound_bitsize (struct type *, int, int);
95
96 static struct type *desc_index_type (struct type *, int);
97
98 static int desc_arity (struct type *);
99
100 static int ada_type_match (struct type *, struct type *, int);
101
102 static int ada_args_match (struct symbol *, struct value **, int);
103
104 static int full_match (const char *, const char *);
105
106 static struct value *make_array_descriptor (struct type *, struct value *);
107
108 static void ada_add_block_symbols (struct obstack *,
109 const struct block *, const char *,
110 domain_enum, struct objfile *, int);
111
112 static void ada_add_all_symbols (struct obstack *, const struct block *,
113 const char *, domain_enum, int, int *);
114
115 static int is_nonfunction (struct block_symbol *, int);
116
117 static void add_defn_to_vec (struct obstack *, struct symbol *,
118 const struct block *);
119
120 static int num_defns_collected (struct obstack *);
121
122 static struct block_symbol *defns_collected (struct obstack *, int);
123
124 static struct value *resolve_subexp (struct expression **, int *, int,
125 struct type *);
126
127 static void replace_operator_with_call (struct expression **, int, int, int,
128 struct symbol *, const struct block *);
129
130 static int possible_user_operator_p (enum exp_opcode, struct value **);
131
132 static char *ada_op_name (enum exp_opcode);
133
134 static const char *ada_decoded_op_name (enum exp_opcode);
135
136 static int numeric_type_p (struct type *);
137
138 static int integer_type_p (struct type *);
139
140 static int scalar_type_p (struct type *);
141
142 static int discrete_type_p (struct type *);
143
144 static enum ada_renaming_category parse_old_style_renaming (struct type *,
145 const char **,
146 int *,
147 const char **);
148
149 static struct symbol *find_old_style_renaming_symbol (const char *,
150 const struct block *);
151
152 static struct type *ada_lookup_struct_elt_type (struct type *, char *,
153 int, int, int *);
154
155 static struct value *evaluate_subexp_type (struct expression *, int *);
156
157 static struct type *ada_find_parallel_type_with_name (struct type *,
158 const char *);
159
160 static int is_dynamic_field (struct type *, int);
161
162 static struct type *to_fixed_variant_branch_type (struct type *,
163 const gdb_byte *,
164 CORE_ADDR, struct value *);
165
166 static struct type *to_fixed_array_type (struct type *, struct value *, int);
167
168 static struct type *to_fixed_range_type (struct type *, struct value *);
169
170 static struct type *to_static_fixed_type (struct type *);
171 static struct type *static_unwrap_type (struct type *type);
172
173 static struct value *unwrap_value (struct value *);
174
175 static struct type *constrained_packed_array_type (struct type *, long *);
176
177 static struct type *decode_constrained_packed_array_type (struct type *);
178
179 static long decode_packed_array_bitsize (struct type *);
180
181 static struct value *decode_constrained_packed_array (struct value *);
182
183 static int ada_is_packed_array_type (struct type *);
184
185 static int ada_is_unconstrained_packed_array_type (struct type *);
186
187 static struct value *value_subscript_packed (struct value *, int,
188 struct value **);
189
190 static void move_bits (gdb_byte *, int, const gdb_byte *, int, int, int);
191
192 static struct value *coerce_unspec_val_to_type (struct value *,
193 struct type *);
194
195 static struct value *get_var_value (char *, char *);
196
197 static int lesseq_defined_than (struct symbol *, struct symbol *);
198
199 static int equiv_types (struct type *, struct type *);
200
201 static int is_name_suffix (const char *);
202
203 static int advance_wild_match (const char **, const char *, int);
204
205 static int wild_match (const char *, const char *);
206
207 static struct value *ada_coerce_ref (struct value *);
208
209 static LONGEST pos_atr (struct value *);
210
211 static struct value *value_pos_atr (struct type *, struct value *);
212
213 static struct value *value_val_atr (struct type *, struct value *);
214
215 static struct symbol *standard_lookup (const char *, const struct block *,
216 domain_enum);
217
218 static struct value *ada_search_struct_field (const char *, struct value *, int,
219 struct type *);
220
221 static struct value *ada_value_primitive_field (struct value *, int, int,
222 struct type *);
223
224 static int find_struct_field (const char *, struct type *, int,
225 struct type **, int *, int *, int *, int *);
226
227 static struct value *ada_to_fixed_value_create (struct type *, CORE_ADDR,
228 struct value *);
229
230 static int ada_resolve_function (struct block_symbol *, int,
231 struct value **, int, const char *,
232 struct type *);
233
234 static int ada_is_direct_array_type (struct type *);
235
236 static void ada_language_arch_info (struct gdbarch *,
237 struct language_arch_info *);
238
239 static struct value *ada_index_struct_field (int, struct value *, int,
240 struct type *);
241
242 static struct value *assign_aggregate (struct value *, struct value *,
243 struct expression *,
244 int *, enum noside);
245
246 static void aggregate_assign_from_choices (struct value *, struct value *,
247 struct expression *,
248 int *, LONGEST *, int *,
249 int, LONGEST, LONGEST);
250
251 static void aggregate_assign_positional (struct value *, struct value *,
252 struct expression *,
253 int *, LONGEST *, int *, int,
254 LONGEST, LONGEST);
255
256
257 static void aggregate_assign_others (struct value *, struct value *,
258 struct expression *,
259 int *, LONGEST *, int, LONGEST, LONGEST);
260
261
262 static void add_component_interval (LONGEST, LONGEST, LONGEST *, int *, int);
263
264
265 static struct value *ada_evaluate_subexp (struct type *, struct expression *,
266 int *, enum noside);
267
268 static void ada_forward_operator_length (struct expression *, int, int *,
269 int *);
270
271 static struct type *ada_find_any_type (const char *name);
272 \f
273
274 /* The result of a symbol lookup to be stored in our symbol cache. */
275
276 struct cache_entry
277 {
278 /* The name used to perform the lookup. */
279 const char *name;
280 /* The namespace used during the lookup. */
281 domain_enum domain;
282 /* The symbol returned by the lookup, or NULL if no matching symbol
283 was found. */
284 struct symbol *sym;
285 /* The block where the symbol was found, or NULL if no matching
286 symbol was found. */
287 const struct block *block;
288 /* A pointer to the next entry with the same hash. */
289 struct cache_entry *next;
290 };
291
292 /* The Ada symbol cache, used to store the result of Ada-mode symbol
293 lookups in the course of executing the user's commands.
294
295 The cache is implemented using a simple, fixed-sized hash.
296 The size is fixed on the grounds that there are not likely to be
297 all that many symbols looked up during any given session, regardless
298 of the size of the symbol table. If we decide to go to a resizable
299 table, let's just use the stuff from libiberty instead. */
300
301 #define HASH_SIZE 1009
302
303 struct ada_symbol_cache
304 {
305 /* An obstack used to store the entries in our cache. */
306 struct obstack cache_space;
307
308 /* The root of the hash table used to implement our symbol cache. */
309 struct cache_entry *root[HASH_SIZE];
310 };
311
312 static void ada_free_symbol_cache (struct ada_symbol_cache *sym_cache);
313
314 /* Maximum-sized dynamic type. */
315 static unsigned int varsize_limit;
316
317 /* FIXME: brobecker/2003-09-17: No longer a const because it is
318 returned by a function that does not return a const char *. */
319 static char *ada_completer_word_break_characters =
320 #ifdef VMS
321 " \t\n!@#%^&*()+=|~`}{[]\";:?/,-";
322 #else
323 " \t\n!@#$%^&*()+=|~`}{[]\";:?/,-";
324 #endif
325
326 /* The name of the symbol to use to get the name of the main subprogram. */
327 static const char ADA_MAIN_PROGRAM_SYMBOL_NAME[]
328 = "__gnat_ada_main_program_name";
329
330 /* Limit on the number of warnings to raise per expression evaluation. */
331 static int warning_limit = 2;
332
333 /* Number of warning messages issued; reset to 0 by cleanups after
334 expression evaluation. */
335 static int warnings_issued = 0;
336
337 static const char *known_runtime_file_name_patterns[] = {
338 ADA_KNOWN_RUNTIME_FILE_NAME_PATTERNS NULL
339 };
340
341 static const char *known_auxiliary_function_name_patterns[] = {
342 ADA_KNOWN_AUXILIARY_FUNCTION_NAME_PATTERNS NULL
343 };
344
345 /* Space for allocating results of ada_lookup_symbol_list. */
346 static struct obstack symbol_list_obstack;
347
348 /* Maintenance-related settings for this module. */
349
350 static struct cmd_list_element *maint_set_ada_cmdlist;
351 static struct cmd_list_element *maint_show_ada_cmdlist;
352
353 /* Implement the "maintenance set ada" (prefix) command. */
354
355 static void
356 maint_set_ada_cmd (char *args, int from_tty)
357 {
358 help_list (maint_set_ada_cmdlist, "maintenance set ada ", all_commands,
359 gdb_stdout);
360 }
361
362 /* Implement the "maintenance show ada" (prefix) command. */
363
364 static void
365 maint_show_ada_cmd (char *args, int from_tty)
366 {
367 cmd_show_list (maint_show_ada_cmdlist, from_tty, "");
368 }
369
370 /* The "maintenance ada set/show ignore-descriptive-type" value. */
371
372 static int ada_ignore_descriptive_types_p = 0;
373
374 /* Inferior-specific data. */
375
376 /* Per-inferior data for this module. */
377
378 struct ada_inferior_data
379 {
380 /* The ada__tags__type_specific_data type, which is used when decoding
381 tagged types. With older versions of GNAT, this type was directly
382 accessible through a component ("tsd") in the object tag. But this
383 is no longer the case, so we cache it for each inferior. */
384 struct type *tsd_type;
385
386 /* The exception_support_info data. This data is used to determine
387 how to implement support for Ada exception catchpoints in a given
388 inferior. */
389 const struct exception_support_info *exception_info;
390 };
391
392 /* Our key to this module's inferior data. */
393 static const struct inferior_data *ada_inferior_data;
394
395 /* A cleanup routine for our inferior data. */
396 static void
397 ada_inferior_data_cleanup (struct inferior *inf, void *arg)
398 {
399 struct ada_inferior_data *data;
400
401 data = (struct ada_inferior_data *) inferior_data (inf, ada_inferior_data);
402 if (data != NULL)
403 xfree (data);
404 }
405
406 /* Return our inferior data for the given inferior (INF).
407
408 This function always returns a valid pointer to an allocated
409 ada_inferior_data structure. If INF's inferior data has not
410 been previously set, this functions creates a new one with all
411 fields set to zero, sets INF's inferior to it, and then returns
412 a pointer to that newly allocated ada_inferior_data. */
413
414 static struct ada_inferior_data *
415 get_ada_inferior_data (struct inferior *inf)
416 {
417 struct ada_inferior_data *data;
418
419 data = (struct ada_inferior_data *) inferior_data (inf, ada_inferior_data);
420 if (data == NULL)
421 {
422 data = XCNEW (struct ada_inferior_data);
423 set_inferior_data (inf, ada_inferior_data, data);
424 }
425
426 return data;
427 }
428
429 /* Perform all necessary cleanups regarding our module's inferior data
430 that is required after the inferior INF just exited. */
431
432 static void
433 ada_inferior_exit (struct inferior *inf)
434 {
435 ada_inferior_data_cleanup (inf, NULL);
436 set_inferior_data (inf, ada_inferior_data, NULL);
437 }
438
439
440 /* program-space-specific data. */
441
442 /* This module's per-program-space data. */
443 struct ada_pspace_data
444 {
445 /* The Ada symbol cache. */
446 struct ada_symbol_cache *sym_cache;
447 };
448
449 /* Key to our per-program-space data. */
450 static const struct program_space_data *ada_pspace_data_handle;
451
452 /* Return this module's data for the given program space (PSPACE).
453 If not is found, add a zero'ed one now.
454
455 This function always returns a valid object. */
456
457 static struct ada_pspace_data *
458 get_ada_pspace_data (struct program_space *pspace)
459 {
460 struct ada_pspace_data *data;
461
462 data = ((struct ada_pspace_data *)
463 program_space_data (pspace, ada_pspace_data_handle));
464 if (data == NULL)
465 {
466 data = XCNEW (struct ada_pspace_data);
467 set_program_space_data (pspace, ada_pspace_data_handle, data);
468 }
469
470 return data;
471 }
472
473 /* The cleanup callback for this module's per-program-space data. */
474
475 static void
476 ada_pspace_data_cleanup (struct program_space *pspace, void *data)
477 {
478 struct ada_pspace_data *pspace_data = (struct ada_pspace_data *) data;
479
480 if (pspace_data->sym_cache != NULL)
481 ada_free_symbol_cache (pspace_data->sym_cache);
482 xfree (pspace_data);
483 }
484
485 /* Utilities */
486
487 /* If TYPE is a TYPE_CODE_TYPEDEF type, return the target type after
488 all typedef layers have been peeled. Otherwise, return TYPE.
489
490 Normally, we really expect a typedef type to only have 1 typedef layer.
491 In other words, we really expect the target type of a typedef type to be
492 a non-typedef type. This is particularly true for Ada units, because
493 the language does not have a typedef vs not-typedef distinction.
494 In that respect, the Ada compiler has been trying to eliminate as many
495 typedef definitions in the debugging information, since they generally
496 do not bring any extra information (we still use typedef under certain
497 circumstances related mostly to the GNAT encoding).
498
499 Unfortunately, we have seen situations where the debugging information
500 generated by the compiler leads to such multiple typedef layers. For
501 instance, consider the following example with stabs:
502
503 .stabs "pck__float_array___XUP:Tt(0,46)=s16P_ARRAY:(0,47)=[...]"[...]
504 .stabs "pck__float_array___XUP:t(0,36)=(0,46)",128,0,6,0
505
506 This is an error in the debugging information which causes type
507 pck__float_array___XUP to be defined twice, and the second time,
508 it is defined as a typedef of a typedef.
509
510 This is on the fringe of legality as far as debugging information is
511 concerned, and certainly unexpected. But it is easy to handle these
512 situations correctly, so we can afford to be lenient in this case. */
513
514 static struct type *
515 ada_typedef_target_type (struct type *type)
516 {
517 while (TYPE_CODE (type) == TYPE_CODE_TYPEDEF)
518 type = TYPE_TARGET_TYPE (type);
519 return type;
520 }
521
522 /* Given DECODED_NAME a string holding a symbol name in its
523 decoded form (ie using the Ada dotted notation), returns
524 its unqualified name. */
525
526 static const char *
527 ada_unqualified_name (const char *decoded_name)
528 {
529 const char *result;
530
531 /* If the decoded name starts with '<', it means that the encoded
532 name does not follow standard naming conventions, and thus that
533 it is not your typical Ada symbol name. Trying to unqualify it
534 is therefore pointless and possibly erroneous. */
535 if (decoded_name[0] == '<')
536 return decoded_name;
537
538 result = strrchr (decoded_name, '.');
539 if (result != NULL)
540 result++; /* Skip the dot... */
541 else
542 result = decoded_name;
543
544 return result;
545 }
546
547 /* Return a string starting with '<', followed by STR, and '>'.
548 The result is good until the next call. */
549
550 static char *
551 add_angle_brackets (const char *str)
552 {
553 static char *result = NULL;
554
555 xfree (result);
556 result = xstrprintf ("<%s>", str);
557 return result;
558 }
559
560 static char *
561 ada_get_gdb_completer_word_break_characters (void)
562 {
563 return ada_completer_word_break_characters;
564 }
565
566 /* Print an array element index using the Ada syntax. */
567
568 static void
569 ada_print_array_index (struct value *index_value, struct ui_file *stream,
570 const struct value_print_options *options)
571 {
572 LA_VALUE_PRINT (index_value, stream, options);
573 fprintf_filtered (stream, " => ");
574 }
575
576 /* Assuming VECT points to an array of *SIZE objects of size
577 ELEMENT_SIZE, grow it to contain at least MIN_SIZE objects,
578 updating *SIZE as necessary and returning the (new) array. */
579
580 void *
581 grow_vect (void *vect, size_t *size, size_t min_size, int element_size)
582 {
583 if (*size < min_size)
584 {
585 *size *= 2;
586 if (*size < min_size)
587 *size = min_size;
588 vect = xrealloc (vect, *size * element_size);
589 }
590 return vect;
591 }
592
593 /* True (non-zero) iff TARGET matches FIELD_NAME up to any trailing
594 suffix of FIELD_NAME beginning "___". */
595
596 static int
597 field_name_match (const char *field_name, const char *target)
598 {
599 int len = strlen (target);
600
601 return
602 (strncmp (field_name, target, len) == 0
603 && (field_name[len] == '\0'
604 || (startswith (field_name + len, "___")
605 && strcmp (field_name + strlen (field_name) - 6,
606 "___XVN") != 0)));
607 }
608
609
610 /* Assuming TYPE is a TYPE_CODE_STRUCT or a TYPE_CODE_TYPDEF to
611 a TYPE_CODE_STRUCT, find the field whose name matches FIELD_NAME,
612 and return its index. This function also handles fields whose name
613 have ___ suffixes because the compiler sometimes alters their name
614 by adding such a suffix to represent fields with certain constraints.
615 If the field could not be found, return a negative number if
616 MAYBE_MISSING is set. Otherwise raise an error. */
617
618 int
619 ada_get_field_index (const struct type *type, const char *field_name,
620 int maybe_missing)
621 {
622 int fieldno;
623 struct type *struct_type = check_typedef ((struct type *) type);
624
625 for (fieldno = 0; fieldno < TYPE_NFIELDS (struct_type); fieldno++)
626 if (field_name_match (TYPE_FIELD_NAME (struct_type, fieldno), field_name))
627 return fieldno;
628
629 if (!maybe_missing)
630 error (_("Unable to find field %s in struct %s. Aborting"),
631 field_name, TYPE_NAME (struct_type));
632
633 return -1;
634 }
635
636 /* The length of the prefix of NAME prior to any "___" suffix. */
637
638 int
639 ada_name_prefix_len (const char *name)
640 {
641 if (name == NULL)
642 return 0;
643 else
644 {
645 const char *p = strstr (name, "___");
646
647 if (p == NULL)
648 return strlen (name);
649 else
650 return p - name;
651 }
652 }
653
654 /* Return non-zero if SUFFIX is a suffix of STR.
655 Return zero if STR is null. */
656
657 static int
658 is_suffix (const char *str, const char *suffix)
659 {
660 int len1, len2;
661
662 if (str == NULL)
663 return 0;
664 len1 = strlen (str);
665 len2 = strlen (suffix);
666 return (len1 >= len2 && strcmp (str + len1 - len2, suffix) == 0);
667 }
668
669 /* The contents of value VAL, treated as a value of type TYPE. The
670 result is an lval in memory if VAL is. */
671
672 static struct value *
673 coerce_unspec_val_to_type (struct value *val, struct type *type)
674 {
675 type = ada_check_typedef (type);
676 if (value_type (val) == type)
677 return val;
678 else
679 {
680 struct value *result;
681
682 /* Make sure that the object size is not unreasonable before
683 trying to allocate some memory for it. */
684 ada_ensure_varsize_limit (type);
685
686 if (value_lazy (val)
687 || TYPE_LENGTH (type) > TYPE_LENGTH (value_type (val)))
688 result = allocate_value_lazy (type);
689 else
690 {
691 result = allocate_value (type);
692 value_contents_copy_raw (result, 0, val, 0, TYPE_LENGTH (type));
693 }
694 set_value_component_location (result, val);
695 set_value_bitsize (result, value_bitsize (val));
696 set_value_bitpos (result, value_bitpos (val));
697 set_value_address (result, value_address (val));
698 return result;
699 }
700 }
701
702 static const gdb_byte *
703 cond_offset_host (const gdb_byte *valaddr, long offset)
704 {
705 if (valaddr == NULL)
706 return NULL;
707 else
708 return valaddr + offset;
709 }
710
711 static CORE_ADDR
712 cond_offset_target (CORE_ADDR address, long offset)
713 {
714 if (address == 0)
715 return 0;
716 else
717 return address + offset;
718 }
719
720 /* Issue a warning (as for the definition of warning in utils.c, but
721 with exactly one argument rather than ...), unless the limit on the
722 number of warnings has passed during the evaluation of the current
723 expression. */
724
725 /* FIXME: cagney/2004-10-10: This function is mimicking the behavior
726 provided by "complaint". */
727 static void lim_warning (const char *format, ...) ATTRIBUTE_PRINTF (1, 2);
728
729 static void
730 lim_warning (const char *format, ...)
731 {
732 va_list args;
733
734 va_start (args, format);
735 warnings_issued += 1;
736 if (warnings_issued <= warning_limit)
737 vwarning (format, args);
738
739 va_end (args);
740 }
741
742 /* Issue an error if the size of an object of type T is unreasonable,
743 i.e. if it would be a bad idea to allocate a value of this type in
744 GDB. */
745
746 void
747 ada_ensure_varsize_limit (const struct type *type)
748 {
749 if (TYPE_LENGTH (type) > varsize_limit)
750 error (_("object size is larger than varsize-limit"));
751 }
752
753 /* Maximum value of a SIZE-byte signed integer type. */
754 static LONGEST
755 max_of_size (int size)
756 {
757 LONGEST top_bit = (LONGEST) 1 << (size * 8 - 2);
758
759 return top_bit | (top_bit - 1);
760 }
761
762 /* Minimum value of a SIZE-byte signed integer type. */
763 static LONGEST
764 min_of_size (int size)
765 {
766 return -max_of_size (size) - 1;
767 }
768
769 /* Maximum value of a SIZE-byte unsigned integer type. */
770 static ULONGEST
771 umax_of_size (int size)
772 {
773 ULONGEST top_bit = (ULONGEST) 1 << (size * 8 - 1);
774
775 return top_bit | (top_bit - 1);
776 }
777
778 /* Maximum value of integral type T, as a signed quantity. */
779 static LONGEST
780 max_of_type (struct type *t)
781 {
782 if (TYPE_UNSIGNED (t))
783 return (LONGEST) umax_of_size (TYPE_LENGTH (t));
784 else
785 return max_of_size (TYPE_LENGTH (t));
786 }
787
788 /* Minimum value of integral type T, as a signed quantity. */
789 static LONGEST
790 min_of_type (struct type *t)
791 {
792 if (TYPE_UNSIGNED (t))
793 return 0;
794 else
795 return min_of_size (TYPE_LENGTH (t));
796 }
797
798 /* The largest value in the domain of TYPE, a discrete type, as an integer. */
799 LONGEST
800 ada_discrete_type_high_bound (struct type *type)
801 {
802 type = resolve_dynamic_type (type, NULL, 0);
803 switch (TYPE_CODE (type))
804 {
805 case TYPE_CODE_RANGE:
806 return TYPE_HIGH_BOUND (type);
807 case TYPE_CODE_ENUM:
808 return TYPE_FIELD_ENUMVAL (type, TYPE_NFIELDS (type) - 1);
809 case TYPE_CODE_BOOL:
810 return 1;
811 case TYPE_CODE_CHAR:
812 case TYPE_CODE_INT:
813 return max_of_type (type);
814 default:
815 error (_("Unexpected type in ada_discrete_type_high_bound."));
816 }
817 }
818
819 /* The smallest value in the domain of TYPE, a discrete type, as an integer. */
820 LONGEST
821 ada_discrete_type_low_bound (struct type *type)
822 {
823 type = resolve_dynamic_type (type, NULL, 0);
824 switch (TYPE_CODE (type))
825 {
826 case TYPE_CODE_RANGE:
827 return TYPE_LOW_BOUND (type);
828 case TYPE_CODE_ENUM:
829 return TYPE_FIELD_ENUMVAL (type, 0);
830 case TYPE_CODE_BOOL:
831 return 0;
832 case TYPE_CODE_CHAR:
833 case TYPE_CODE_INT:
834 return min_of_type (type);
835 default:
836 error (_("Unexpected type in ada_discrete_type_low_bound."));
837 }
838 }
839
840 /* The identity on non-range types. For range types, the underlying
841 non-range scalar type. */
842
843 static struct type *
844 get_base_type (struct type *type)
845 {
846 while (type != NULL && TYPE_CODE (type) == TYPE_CODE_RANGE)
847 {
848 if (type == TYPE_TARGET_TYPE (type) || TYPE_TARGET_TYPE (type) == NULL)
849 return type;
850 type = TYPE_TARGET_TYPE (type);
851 }
852 return type;
853 }
854
855 /* Return a decoded version of the given VALUE. This means returning
856 a value whose type is obtained by applying all the GNAT-specific
857 encondings, making the resulting type a static but standard description
858 of the initial type. */
859
860 struct value *
861 ada_get_decoded_value (struct value *value)
862 {
863 struct type *type = ada_check_typedef (value_type (value));
864
865 if (ada_is_array_descriptor_type (type)
866 || (ada_is_constrained_packed_array_type (type)
867 && TYPE_CODE (type) != TYPE_CODE_PTR))
868 {
869 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF) /* array access type. */
870 value = ada_coerce_to_simple_array_ptr (value);
871 else
872 value = ada_coerce_to_simple_array (value);
873 }
874 else
875 value = ada_to_fixed_value (value);
876
877 return value;
878 }
879
880 /* Same as ada_get_decoded_value, but with the given TYPE.
881 Because there is no associated actual value for this type,
882 the resulting type might be a best-effort approximation in
883 the case of dynamic types. */
884
885 struct type *
886 ada_get_decoded_type (struct type *type)
887 {
888 type = to_static_fixed_type (type);
889 if (ada_is_constrained_packed_array_type (type))
890 type = ada_coerce_to_simple_array_type (type);
891 return type;
892 }
893
894 \f
895
896 /* Language Selection */
897
898 /* If the main program is in Ada, return language_ada, otherwise return LANG
899 (the main program is in Ada iif the adainit symbol is found). */
900
901 enum language
902 ada_update_initial_language (enum language lang)
903 {
904 if (lookup_minimal_symbol ("adainit", (const char *) NULL,
905 (struct objfile *) NULL).minsym != NULL)
906 return language_ada;
907
908 return lang;
909 }
910
911 /* If the main procedure is written in Ada, then return its name.
912 The result is good until the next call. Return NULL if the main
913 procedure doesn't appear to be in Ada. */
914
915 char *
916 ada_main_name (void)
917 {
918 struct bound_minimal_symbol msym;
919 static char *main_program_name = NULL;
920
921 /* For Ada, the name of the main procedure is stored in a specific
922 string constant, generated by the binder. Look for that symbol,
923 extract its address, and then read that string. If we didn't find
924 that string, then most probably the main procedure is not written
925 in Ada. */
926 msym = lookup_minimal_symbol (ADA_MAIN_PROGRAM_SYMBOL_NAME, NULL, NULL);
927
928 if (msym.minsym != NULL)
929 {
930 CORE_ADDR main_program_name_addr;
931 int err_code;
932
933 main_program_name_addr = BMSYMBOL_VALUE_ADDRESS (msym);
934 if (main_program_name_addr == 0)
935 error (_("Invalid address for Ada main program name."));
936
937 xfree (main_program_name);
938 target_read_string (main_program_name_addr, &main_program_name,
939 1024, &err_code);
940
941 if (err_code != 0)
942 return NULL;
943 return main_program_name;
944 }
945
946 /* The main procedure doesn't seem to be in Ada. */
947 return NULL;
948 }
949 \f
950 /* Symbols */
951
952 /* Table of Ada operators and their GNAT-encoded names. Last entry is pair
953 of NULLs. */
954
955 const struct ada_opname_map ada_opname_table[] = {
956 {"Oadd", "\"+\"", BINOP_ADD},
957 {"Osubtract", "\"-\"", BINOP_SUB},
958 {"Omultiply", "\"*\"", BINOP_MUL},
959 {"Odivide", "\"/\"", BINOP_DIV},
960 {"Omod", "\"mod\"", BINOP_MOD},
961 {"Orem", "\"rem\"", BINOP_REM},
962 {"Oexpon", "\"**\"", BINOP_EXP},
963 {"Olt", "\"<\"", BINOP_LESS},
964 {"Ole", "\"<=\"", BINOP_LEQ},
965 {"Ogt", "\">\"", BINOP_GTR},
966 {"Oge", "\">=\"", BINOP_GEQ},
967 {"Oeq", "\"=\"", BINOP_EQUAL},
968 {"One", "\"/=\"", BINOP_NOTEQUAL},
969 {"Oand", "\"and\"", BINOP_BITWISE_AND},
970 {"Oor", "\"or\"", BINOP_BITWISE_IOR},
971 {"Oxor", "\"xor\"", BINOP_BITWISE_XOR},
972 {"Oconcat", "\"&\"", BINOP_CONCAT},
973 {"Oabs", "\"abs\"", UNOP_ABS},
974 {"Onot", "\"not\"", UNOP_LOGICAL_NOT},
975 {"Oadd", "\"+\"", UNOP_PLUS},
976 {"Osubtract", "\"-\"", UNOP_NEG},
977 {NULL, NULL}
978 };
979
980 /* The "encoded" form of DECODED, according to GNAT conventions.
981 The result is valid until the next call to ada_encode. */
982
983 char *
984 ada_encode (const char *decoded)
985 {
986 static char *encoding_buffer = NULL;
987 static size_t encoding_buffer_size = 0;
988 const char *p;
989 int k;
990
991 if (decoded == NULL)
992 return NULL;
993
994 GROW_VECT (encoding_buffer, encoding_buffer_size,
995 2 * strlen (decoded) + 10);
996
997 k = 0;
998 for (p = decoded; *p != '\0'; p += 1)
999 {
1000 if (*p == '.')
1001 {
1002 encoding_buffer[k] = encoding_buffer[k + 1] = '_';
1003 k += 2;
1004 }
1005 else if (*p == '"')
1006 {
1007 const struct ada_opname_map *mapping;
1008
1009 for (mapping = ada_opname_table;
1010 mapping->encoded != NULL
1011 && !startswith (p, mapping->decoded); mapping += 1)
1012 ;
1013 if (mapping->encoded == NULL)
1014 error (_("invalid Ada operator name: %s"), p);
1015 strcpy (encoding_buffer + k, mapping->encoded);
1016 k += strlen (mapping->encoded);
1017 break;
1018 }
1019 else
1020 {
1021 encoding_buffer[k] = *p;
1022 k += 1;
1023 }
1024 }
1025
1026 encoding_buffer[k] = '\0';
1027 return encoding_buffer;
1028 }
1029
1030 /* Return NAME folded to lower case, or, if surrounded by single
1031 quotes, unfolded, but with the quotes stripped away. Result good
1032 to next call. */
1033
1034 char *
1035 ada_fold_name (const char *name)
1036 {
1037 static char *fold_buffer = NULL;
1038 static size_t fold_buffer_size = 0;
1039
1040 int len = strlen (name);
1041 GROW_VECT (fold_buffer, fold_buffer_size, len + 1);
1042
1043 if (name[0] == '\'')
1044 {
1045 strncpy (fold_buffer, name + 1, len - 2);
1046 fold_buffer[len - 2] = '\000';
1047 }
1048 else
1049 {
1050 int i;
1051
1052 for (i = 0; i <= len; i += 1)
1053 fold_buffer[i] = tolower (name[i]);
1054 }
1055
1056 return fold_buffer;
1057 }
1058
1059 /* Return nonzero if C is either a digit or a lowercase alphabet character. */
1060
1061 static int
1062 is_lower_alphanum (const char c)
1063 {
1064 return (isdigit (c) || (isalpha (c) && islower (c)));
1065 }
1066
1067 /* ENCODED is the linkage name of a symbol and LEN contains its length.
1068 This function saves in LEN the length of that same symbol name but
1069 without either of these suffixes:
1070 . .{DIGIT}+
1071 . ${DIGIT}+
1072 . ___{DIGIT}+
1073 . __{DIGIT}+.
1074
1075 These are suffixes introduced by the compiler for entities such as
1076 nested subprogram for instance, in order to avoid name clashes.
1077 They do not serve any purpose for the debugger. */
1078
1079 static void
1080 ada_remove_trailing_digits (const char *encoded, int *len)
1081 {
1082 if (*len > 1 && isdigit (encoded[*len - 1]))
1083 {
1084 int i = *len - 2;
1085
1086 while (i > 0 && isdigit (encoded[i]))
1087 i--;
1088 if (i >= 0 && encoded[i] == '.')
1089 *len = i;
1090 else if (i >= 0 && encoded[i] == '$')
1091 *len = i;
1092 else if (i >= 2 && startswith (encoded + i - 2, "___"))
1093 *len = i - 2;
1094 else if (i >= 1 && startswith (encoded + i - 1, "__"))
1095 *len = i - 1;
1096 }
1097 }
1098
1099 /* Remove the suffix introduced by the compiler for protected object
1100 subprograms. */
1101
1102 static void
1103 ada_remove_po_subprogram_suffix (const char *encoded, int *len)
1104 {
1105 /* Remove trailing N. */
1106
1107 /* Protected entry subprograms are broken into two
1108 separate subprograms: The first one is unprotected, and has
1109 a 'N' suffix; the second is the protected version, and has
1110 the 'P' suffix. The second calls the first one after handling
1111 the protection. Since the P subprograms are internally generated,
1112 we leave these names undecoded, giving the user a clue that this
1113 entity is internal. */
1114
1115 if (*len > 1
1116 && encoded[*len - 1] == 'N'
1117 && (isdigit (encoded[*len - 2]) || islower (encoded[*len - 2])))
1118 *len = *len - 1;
1119 }
1120
1121 /* Remove trailing X[bn]* suffixes (indicating names in package bodies). */
1122
1123 static void
1124 ada_remove_Xbn_suffix (const char *encoded, int *len)
1125 {
1126 int i = *len - 1;
1127
1128 while (i > 0 && (encoded[i] == 'b' || encoded[i] == 'n'))
1129 i--;
1130
1131 if (encoded[i] != 'X')
1132 return;
1133
1134 if (i == 0)
1135 return;
1136
1137 if (isalnum (encoded[i-1]))
1138 *len = i;
1139 }
1140
1141 /* If ENCODED follows the GNAT entity encoding conventions, then return
1142 the decoded form of ENCODED. Otherwise, return "<%s>" where "%s" is
1143 replaced by ENCODED.
1144
1145 The resulting string is valid until the next call of ada_decode.
1146 If the string is unchanged by decoding, the original string pointer
1147 is returned. */
1148
1149 const char *
1150 ada_decode (const char *encoded)
1151 {
1152 int i, j;
1153 int len0;
1154 const char *p;
1155 char *decoded;
1156 int at_start_name;
1157 static char *decoding_buffer = NULL;
1158 static size_t decoding_buffer_size = 0;
1159
1160 /* The name of the Ada main procedure starts with "_ada_".
1161 This prefix is not part of the decoded name, so skip this part
1162 if we see this prefix. */
1163 if (startswith (encoded, "_ada_"))
1164 encoded += 5;
1165
1166 /* If the name starts with '_', then it is not a properly encoded
1167 name, so do not attempt to decode it. Similarly, if the name
1168 starts with '<', the name should not be decoded. */
1169 if (encoded[0] == '_' || encoded[0] == '<')
1170 goto Suppress;
1171
1172 len0 = strlen (encoded);
1173
1174 ada_remove_trailing_digits (encoded, &len0);
1175 ada_remove_po_subprogram_suffix (encoded, &len0);
1176
1177 /* Remove the ___X.* suffix if present. Do not forget to verify that
1178 the suffix is located before the current "end" of ENCODED. We want
1179 to avoid re-matching parts of ENCODED that have previously been
1180 marked as discarded (by decrementing LEN0). */
1181 p = strstr (encoded, "___");
1182 if (p != NULL && p - encoded < len0 - 3)
1183 {
1184 if (p[3] == 'X')
1185 len0 = p - encoded;
1186 else
1187 goto Suppress;
1188 }
1189
1190 /* Remove any trailing TKB suffix. It tells us that this symbol
1191 is for the body of a task, but that information does not actually
1192 appear in the decoded name. */
1193
1194 if (len0 > 3 && startswith (encoded + len0 - 3, "TKB"))
1195 len0 -= 3;
1196
1197 /* Remove any trailing TB suffix. The TB suffix is slightly different
1198 from the TKB suffix because it is used for non-anonymous task
1199 bodies. */
1200
1201 if (len0 > 2 && startswith (encoded + len0 - 2, "TB"))
1202 len0 -= 2;
1203
1204 /* Remove trailing "B" suffixes. */
1205 /* FIXME: brobecker/2006-04-19: Not sure what this are used for... */
1206
1207 if (len0 > 1 && startswith (encoded + len0 - 1, "B"))
1208 len0 -= 1;
1209
1210 /* Make decoded big enough for possible expansion by operator name. */
1211
1212 GROW_VECT (decoding_buffer, decoding_buffer_size, 2 * len0 + 1);
1213 decoded = decoding_buffer;
1214
1215 /* Remove trailing __{digit}+ or trailing ${digit}+. */
1216
1217 if (len0 > 1 && isdigit (encoded[len0 - 1]))
1218 {
1219 i = len0 - 2;
1220 while ((i >= 0 && isdigit (encoded[i]))
1221 || (i >= 1 && encoded[i] == '_' && isdigit (encoded[i - 1])))
1222 i -= 1;
1223 if (i > 1 && encoded[i] == '_' && encoded[i - 1] == '_')
1224 len0 = i - 1;
1225 else if (encoded[i] == '$')
1226 len0 = i;
1227 }
1228
1229 /* The first few characters that are not alphabetic are not part
1230 of any encoding we use, so we can copy them over verbatim. */
1231
1232 for (i = 0, j = 0; i < len0 && !isalpha (encoded[i]); i += 1, j += 1)
1233 decoded[j] = encoded[i];
1234
1235 at_start_name = 1;
1236 while (i < len0)
1237 {
1238 /* Is this a symbol function? */
1239 if (at_start_name && encoded[i] == 'O')
1240 {
1241 int k;
1242
1243 for (k = 0; ada_opname_table[k].encoded != NULL; k += 1)
1244 {
1245 int op_len = strlen (ada_opname_table[k].encoded);
1246 if ((strncmp (ada_opname_table[k].encoded + 1, encoded + i + 1,
1247 op_len - 1) == 0)
1248 && !isalnum (encoded[i + op_len]))
1249 {
1250 strcpy (decoded + j, ada_opname_table[k].decoded);
1251 at_start_name = 0;
1252 i += op_len;
1253 j += strlen (ada_opname_table[k].decoded);
1254 break;
1255 }
1256 }
1257 if (ada_opname_table[k].encoded != NULL)
1258 continue;
1259 }
1260 at_start_name = 0;
1261
1262 /* Replace "TK__" with "__", which will eventually be translated
1263 into "." (just below). */
1264
1265 if (i < len0 - 4 && startswith (encoded + i, "TK__"))
1266 i += 2;
1267
1268 /* Replace "__B_{DIGITS}+__" sequences by "__", which will eventually
1269 be translated into "." (just below). These are internal names
1270 generated for anonymous blocks inside which our symbol is nested. */
1271
1272 if (len0 - i > 5 && encoded [i] == '_' && encoded [i+1] == '_'
1273 && encoded [i+2] == 'B' && encoded [i+3] == '_'
1274 && isdigit (encoded [i+4]))
1275 {
1276 int k = i + 5;
1277
1278 while (k < len0 && isdigit (encoded[k]))
1279 k++; /* Skip any extra digit. */
1280
1281 /* Double-check that the "__B_{DIGITS}+" sequence we found
1282 is indeed followed by "__". */
1283 if (len0 - k > 2 && encoded [k] == '_' && encoded [k+1] == '_')
1284 i = k;
1285 }
1286
1287 /* Remove _E{DIGITS}+[sb] */
1288
1289 /* Just as for protected object subprograms, there are 2 categories
1290 of subprograms created by the compiler for each entry. The first
1291 one implements the actual entry code, and has a suffix following
1292 the convention above; the second one implements the barrier and
1293 uses the same convention as above, except that the 'E' is replaced
1294 by a 'B'.
1295
1296 Just as above, we do not decode the name of barrier functions
1297 to give the user a clue that the code he is debugging has been
1298 internally generated. */
1299
1300 if (len0 - i > 3 && encoded [i] == '_' && encoded[i+1] == 'E'
1301 && isdigit (encoded[i+2]))
1302 {
1303 int k = i + 3;
1304
1305 while (k < len0 && isdigit (encoded[k]))
1306 k++;
1307
1308 if (k < len0
1309 && (encoded[k] == 'b' || encoded[k] == 's'))
1310 {
1311 k++;
1312 /* Just as an extra precaution, make sure that if this
1313 suffix is followed by anything else, it is a '_'.
1314 Otherwise, we matched this sequence by accident. */
1315 if (k == len0
1316 || (k < len0 && encoded[k] == '_'))
1317 i = k;
1318 }
1319 }
1320
1321 /* Remove trailing "N" in [a-z0-9]+N__. The N is added by
1322 the GNAT front-end in protected object subprograms. */
1323
1324 if (i < len0 + 3
1325 && encoded[i] == 'N' && encoded[i+1] == '_' && encoded[i+2] == '_')
1326 {
1327 /* Backtrack a bit up until we reach either the begining of
1328 the encoded name, or "__". Make sure that we only find
1329 digits or lowercase characters. */
1330 const char *ptr = encoded + i - 1;
1331
1332 while (ptr >= encoded && is_lower_alphanum (ptr[0]))
1333 ptr--;
1334 if (ptr < encoded
1335 || (ptr > encoded && ptr[0] == '_' && ptr[-1] == '_'))
1336 i++;
1337 }
1338
1339 if (encoded[i] == 'X' && i != 0 && isalnum (encoded[i - 1]))
1340 {
1341 /* This is a X[bn]* sequence not separated from the previous
1342 part of the name with a non-alpha-numeric character (in other
1343 words, immediately following an alpha-numeric character), then
1344 verify that it is placed at the end of the encoded name. If
1345 not, then the encoding is not valid and we should abort the
1346 decoding. Otherwise, just skip it, it is used in body-nested
1347 package names. */
1348 do
1349 i += 1;
1350 while (i < len0 && (encoded[i] == 'b' || encoded[i] == 'n'));
1351 if (i < len0)
1352 goto Suppress;
1353 }
1354 else if (i < len0 - 2 && encoded[i] == '_' && encoded[i + 1] == '_')
1355 {
1356 /* Replace '__' by '.'. */
1357 decoded[j] = '.';
1358 at_start_name = 1;
1359 i += 2;
1360 j += 1;
1361 }
1362 else
1363 {
1364 /* It's a character part of the decoded name, so just copy it
1365 over. */
1366 decoded[j] = encoded[i];
1367 i += 1;
1368 j += 1;
1369 }
1370 }
1371 decoded[j] = '\000';
1372
1373 /* Decoded names should never contain any uppercase character.
1374 Double-check this, and abort the decoding if we find one. */
1375
1376 for (i = 0; decoded[i] != '\0'; i += 1)
1377 if (isupper (decoded[i]) || decoded[i] == ' ')
1378 goto Suppress;
1379
1380 if (strcmp (decoded, encoded) == 0)
1381 return encoded;
1382 else
1383 return decoded;
1384
1385 Suppress:
1386 GROW_VECT (decoding_buffer, decoding_buffer_size, strlen (encoded) + 3);
1387 decoded = decoding_buffer;
1388 if (encoded[0] == '<')
1389 strcpy (decoded, encoded);
1390 else
1391 xsnprintf (decoded, decoding_buffer_size, "<%s>", encoded);
1392 return decoded;
1393
1394 }
1395
1396 /* Table for keeping permanent unique copies of decoded names. Once
1397 allocated, names in this table are never released. While this is a
1398 storage leak, it should not be significant unless there are massive
1399 changes in the set of decoded names in successive versions of a
1400 symbol table loaded during a single session. */
1401 static struct htab *decoded_names_store;
1402
1403 /* Returns the decoded name of GSYMBOL, as for ada_decode, caching it
1404 in the language-specific part of GSYMBOL, if it has not been
1405 previously computed. Tries to save the decoded name in the same
1406 obstack as GSYMBOL, if possible, and otherwise on the heap (so that,
1407 in any case, the decoded symbol has a lifetime at least that of
1408 GSYMBOL).
1409 The GSYMBOL parameter is "mutable" in the C++ sense: logically
1410 const, but nevertheless modified to a semantically equivalent form
1411 when a decoded name is cached in it. */
1412
1413 const char *
1414 ada_decode_symbol (const struct general_symbol_info *arg)
1415 {
1416 struct general_symbol_info *gsymbol = (struct general_symbol_info *) arg;
1417 const char **resultp =
1418 &gsymbol->language_specific.demangled_name;
1419
1420 if (!gsymbol->ada_mangled)
1421 {
1422 const char *decoded = ada_decode (gsymbol->name);
1423 struct obstack *obstack = gsymbol->language_specific.obstack;
1424
1425 gsymbol->ada_mangled = 1;
1426
1427 if (obstack != NULL)
1428 *resultp
1429 = (const char *) obstack_copy0 (obstack, decoded, strlen (decoded));
1430 else
1431 {
1432 /* Sometimes, we can't find a corresponding objfile, in
1433 which case, we put the result on the heap. Since we only
1434 decode when needed, we hope this usually does not cause a
1435 significant memory leak (FIXME). */
1436
1437 char **slot = (char **) htab_find_slot (decoded_names_store,
1438 decoded, INSERT);
1439
1440 if (*slot == NULL)
1441 *slot = xstrdup (decoded);
1442 *resultp = *slot;
1443 }
1444 }
1445
1446 return *resultp;
1447 }
1448
1449 static char *
1450 ada_la_decode (const char *encoded, int options)
1451 {
1452 return xstrdup (ada_decode (encoded));
1453 }
1454
1455 /* Returns non-zero iff SYM_NAME matches NAME, ignoring any trailing
1456 suffixes that encode debugging information or leading _ada_ on
1457 SYM_NAME (see is_name_suffix commentary for the debugging
1458 information that is ignored). If WILD, then NAME need only match a
1459 suffix of SYM_NAME minus the same suffixes. Also returns 0 if
1460 either argument is NULL. */
1461
1462 static int
1463 match_name (const char *sym_name, const char *name, int wild)
1464 {
1465 if (sym_name == NULL || name == NULL)
1466 return 0;
1467 else if (wild)
1468 return wild_match (sym_name, name) == 0;
1469 else
1470 {
1471 int len_name = strlen (name);
1472
1473 return (strncmp (sym_name, name, len_name) == 0
1474 && is_name_suffix (sym_name + len_name))
1475 || (startswith (sym_name, "_ada_")
1476 && strncmp (sym_name + 5, name, len_name) == 0
1477 && is_name_suffix (sym_name + len_name + 5));
1478 }
1479 }
1480 \f
1481
1482 /* Arrays */
1483
1484 /* Assuming that INDEX_DESC_TYPE is an ___XA structure, a structure
1485 generated by the GNAT compiler to describe the index type used
1486 for each dimension of an array, check whether it follows the latest
1487 known encoding. If not, fix it up to conform to the latest encoding.
1488 Otherwise, do nothing. This function also does nothing if
1489 INDEX_DESC_TYPE is NULL.
1490
1491 The GNAT encoding used to describle the array index type evolved a bit.
1492 Initially, the information would be provided through the name of each
1493 field of the structure type only, while the type of these fields was
1494 described as unspecified and irrelevant. The debugger was then expected
1495 to perform a global type lookup using the name of that field in order
1496 to get access to the full index type description. Because these global
1497 lookups can be very expensive, the encoding was later enhanced to make
1498 the global lookup unnecessary by defining the field type as being
1499 the full index type description.
1500
1501 The purpose of this routine is to allow us to support older versions
1502 of the compiler by detecting the use of the older encoding, and by
1503 fixing up the INDEX_DESC_TYPE to follow the new one (at this point,
1504 we essentially replace each field's meaningless type by the associated
1505 index subtype). */
1506
1507 void
1508 ada_fixup_array_indexes_type (struct type *index_desc_type)
1509 {
1510 int i;
1511
1512 if (index_desc_type == NULL)
1513 return;
1514 gdb_assert (TYPE_NFIELDS (index_desc_type) > 0);
1515
1516 /* Check if INDEX_DESC_TYPE follows the older encoding (it is sufficient
1517 to check one field only, no need to check them all). If not, return
1518 now.
1519
1520 If our INDEX_DESC_TYPE was generated using the older encoding,
1521 the field type should be a meaningless integer type whose name
1522 is not equal to the field name. */
1523 if (TYPE_NAME (TYPE_FIELD_TYPE (index_desc_type, 0)) != NULL
1524 && strcmp (TYPE_NAME (TYPE_FIELD_TYPE (index_desc_type, 0)),
1525 TYPE_FIELD_NAME (index_desc_type, 0)) == 0)
1526 return;
1527
1528 /* Fixup each field of INDEX_DESC_TYPE. */
1529 for (i = 0; i < TYPE_NFIELDS (index_desc_type); i++)
1530 {
1531 const char *name = TYPE_FIELD_NAME (index_desc_type, i);
1532 struct type *raw_type = ada_check_typedef (ada_find_any_type (name));
1533
1534 if (raw_type)
1535 TYPE_FIELD_TYPE (index_desc_type, i) = raw_type;
1536 }
1537 }
1538
1539 /* Names of MAX_ADA_DIMENS bounds in P_BOUNDS fields of array descriptors. */
1540
1541 static char *bound_name[] = {
1542 "LB0", "UB0", "LB1", "UB1", "LB2", "UB2", "LB3", "UB3",
1543 "LB4", "UB4", "LB5", "UB5", "LB6", "UB6", "LB7", "UB7"
1544 };
1545
1546 /* Maximum number of array dimensions we are prepared to handle. */
1547
1548 #define MAX_ADA_DIMENS (sizeof(bound_name) / (2*sizeof(char *)))
1549
1550
1551 /* The desc_* routines return primitive portions of array descriptors
1552 (fat pointers). */
1553
1554 /* The descriptor or array type, if any, indicated by TYPE; removes
1555 level of indirection, if needed. */
1556
1557 static struct type *
1558 desc_base_type (struct type *type)
1559 {
1560 if (type == NULL)
1561 return NULL;
1562 type = ada_check_typedef (type);
1563 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF)
1564 type = ada_typedef_target_type (type);
1565
1566 if (type != NULL
1567 && (TYPE_CODE (type) == TYPE_CODE_PTR
1568 || TYPE_CODE (type) == TYPE_CODE_REF))
1569 return ada_check_typedef (TYPE_TARGET_TYPE (type));
1570 else
1571 return type;
1572 }
1573
1574 /* True iff TYPE indicates a "thin" array pointer type. */
1575
1576 static int
1577 is_thin_pntr (struct type *type)
1578 {
1579 return
1580 is_suffix (ada_type_name (desc_base_type (type)), "___XUT")
1581 || is_suffix (ada_type_name (desc_base_type (type)), "___XUT___XVE");
1582 }
1583
1584 /* The descriptor type for thin pointer type TYPE. */
1585
1586 static struct type *
1587 thin_descriptor_type (struct type *type)
1588 {
1589 struct type *base_type = desc_base_type (type);
1590
1591 if (base_type == NULL)
1592 return NULL;
1593 if (is_suffix (ada_type_name (base_type), "___XVE"))
1594 return base_type;
1595 else
1596 {
1597 struct type *alt_type = ada_find_parallel_type (base_type, "___XVE");
1598
1599 if (alt_type == NULL)
1600 return base_type;
1601 else
1602 return alt_type;
1603 }
1604 }
1605
1606 /* A pointer to the array data for thin-pointer value VAL. */
1607
1608 static struct value *
1609 thin_data_pntr (struct value *val)
1610 {
1611 struct type *type = ada_check_typedef (value_type (val));
1612 struct type *data_type = desc_data_target_type (thin_descriptor_type (type));
1613
1614 data_type = lookup_pointer_type (data_type);
1615
1616 if (TYPE_CODE (type) == TYPE_CODE_PTR)
1617 return value_cast (data_type, value_copy (val));
1618 else
1619 return value_from_longest (data_type, value_address (val));
1620 }
1621
1622 /* True iff TYPE indicates a "thick" array pointer type. */
1623
1624 static int
1625 is_thick_pntr (struct type *type)
1626 {
1627 type = desc_base_type (type);
1628 return (type != NULL && TYPE_CODE (type) == TYPE_CODE_STRUCT
1629 && lookup_struct_elt_type (type, "P_BOUNDS", 1) != NULL);
1630 }
1631
1632 /* If TYPE is the type of an array descriptor (fat or thin pointer) or a
1633 pointer to one, the type of its bounds data; otherwise, NULL. */
1634
1635 static struct type *
1636 desc_bounds_type (struct type *type)
1637 {
1638 struct type *r;
1639
1640 type = desc_base_type (type);
1641
1642 if (type == NULL)
1643 return NULL;
1644 else if (is_thin_pntr (type))
1645 {
1646 type = thin_descriptor_type (type);
1647 if (type == NULL)
1648 return NULL;
1649 r = lookup_struct_elt_type (type, "BOUNDS", 1);
1650 if (r != NULL)
1651 return ada_check_typedef (r);
1652 }
1653 else if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
1654 {
1655 r = lookup_struct_elt_type (type, "P_BOUNDS", 1);
1656 if (r != NULL)
1657 return ada_check_typedef (TYPE_TARGET_TYPE (ada_check_typedef (r)));
1658 }
1659 return NULL;
1660 }
1661
1662 /* If ARR is an array descriptor (fat or thin pointer), or pointer to
1663 one, a pointer to its bounds data. Otherwise NULL. */
1664
1665 static struct value *
1666 desc_bounds (struct value *arr)
1667 {
1668 struct type *type = ada_check_typedef (value_type (arr));
1669
1670 if (is_thin_pntr (type))
1671 {
1672 struct type *bounds_type =
1673 desc_bounds_type (thin_descriptor_type (type));
1674 LONGEST addr;
1675
1676 if (bounds_type == NULL)
1677 error (_("Bad GNAT array descriptor"));
1678
1679 /* NOTE: The following calculation is not really kosher, but
1680 since desc_type is an XVE-encoded type (and shouldn't be),
1681 the correct calculation is a real pain. FIXME (and fix GCC). */
1682 if (TYPE_CODE (type) == TYPE_CODE_PTR)
1683 addr = value_as_long (arr);
1684 else
1685 addr = value_address (arr);
1686
1687 return
1688 value_from_longest (lookup_pointer_type (bounds_type),
1689 addr - TYPE_LENGTH (bounds_type));
1690 }
1691
1692 else if (is_thick_pntr (type))
1693 {
1694 struct value *p_bounds = value_struct_elt (&arr, NULL, "P_BOUNDS", NULL,
1695 _("Bad GNAT array descriptor"));
1696 struct type *p_bounds_type = value_type (p_bounds);
1697
1698 if (p_bounds_type
1699 && TYPE_CODE (p_bounds_type) == TYPE_CODE_PTR)
1700 {
1701 struct type *target_type = TYPE_TARGET_TYPE (p_bounds_type);
1702
1703 if (TYPE_STUB (target_type))
1704 p_bounds = value_cast (lookup_pointer_type
1705 (ada_check_typedef (target_type)),
1706 p_bounds);
1707 }
1708 else
1709 error (_("Bad GNAT array descriptor"));
1710
1711 return p_bounds;
1712 }
1713 else
1714 return NULL;
1715 }
1716
1717 /* If TYPE is the type of an array-descriptor (fat pointer), the bit
1718 position of the field containing the address of the bounds data. */
1719
1720 static int
1721 fat_pntr_bounds_bitpos (struct type *type)
1722 {
1723 return TYPE_FIELD_BITPOS (desc_base_type (type), 1);
1724 }
1725
1726 /* If TYPE is the type of an array-descriptor (fat pointer), the bit
1727 size of the field containing the address of the bounds data. */
1728
1729 static int
1730 fat_pntr_bounds_bitsize (struct type *type)
1731 {
1732 type = desc_base_type (type);
1733
1734 if (TYPE_FIELD_BITSIZE (type, 1) > 0)
1735 return TYPE_FIELD_BITSIZE (type, 1);
1736 else
1737 return 8 * TYPE_LENGTH (ada_check_typedef (TYPE_FIELD_TYPE (type, 1)));
1738 }
1739
1740 /* If TYPE is the type of an array descriptor (fat or thin pointer) or a
1741 pointer to one, the type of its array data (a array-with-no-bounds type);
1742 otherwise, NULL. Use ada_type_of_array to get an array type with bounds
1743 data. */
1744
1745 static struct type *
1746 desc_data_target_type (struct type *type)
1747 {
1748 type = desc_base_type (type);
1749
1750 /* NOTE: The following is bogus; see comment in desc_bounds. */
1751 if (is_thin_pntr (type))
1752 return desc_base_type (TYPE_FIELD_TYPE (thin_descriptor_type (type), 1));
1753 else if (is_thick_pntr (type))
1754 {
1755 struct type *data_type = lookup_struct_elt_type (type, "P_ARRAY", 1);
1756
1757 if (data_type
1758 && TYPE_CODE (ada_check_typedef (data_type)) == TYPE_CODE_PTR)
1759 return ada_check_typedef (TYPE_TARGET_TYPE (data_type));
1760 }
1761
1762 return NULL;
1763 }
1764
1765 /* If ARR is an array descriptor (fat or thin pointer), a pointer to
1766 its array data. */
1767
1768 static struct value *
1769 desc_data (struct value *arr)
1770 {
1771 struct type *type = value_type (arr);
1772
1773 if (is_thin_pntr (type))
1774 return thin_data_pntr (arr);
1775 else if (is_thick_pntr (type))
1776 return value_struct_elt (&arr, NULL, "P_ARRAY", NULL,
1777 _("Bad GNAT array descriptor"));
1778 else
1779 return NULL;
1780 }
1781
1782
1783 /* If TYPE is the type of an array-descriptor (fat pointer), the bit
1784 position of the field containing the address of the data. */
1785
1786 static int
1787 fat_pntr_data_bitpos (struct type *type)
1788 {
1789 return TYPE_FIELD_BITPOS (desc_base_type (type), 0);
1790 }
1791
1792 /* If TYPE is the type of an array-descriptor (fat pointer), the bit
1793 size of the field containing the address of the data. */
1794
1795 static int
1796 fat_pntr_data_bitsize (struct type *type)
1797 {
1798 type = desc_base_type (type);
1799
1800 if (TYPE_FIELD_BITSIZE (type, 0) > 0)
1801 return TYPE_FIELD_BITSIZE (type, 0);
1802 else
1803 return TARGET_CHAR_BIT * TYPE_LENGTH (TYPE_FIELD_TYPE (type, 0));
1804 }
1805
1806 /* If BOUNDS is an array-bounds structure (or pointer to one), return
1807 the Ith lower bound stored in it, if WHICH is 0, and the Ith upper
1808 bound, if WHICH is 1. The first bound is I=1. */
1809
1810 static struct value *
1811 desc_one_bound (struct value *bounds, int i, int which)
1812 {
1813 return value_struct_elt (&bounds, NULL, bound_name[2 * i + which - 2], NULL,
1814 _("Bad GNAT array descriptor bounds"));
1815 }
1816
1817 /* If BOUNDS is an array-bounds structure type, return the bit position
1818 of the Ith lower bound stored in it, if WHICH is 0, and the Ith upper
1819 bound, if WHICH is 1. The first bound is I=1. */
1820
1821 static int
1822 desc_bound_bitpos (struct type *type, int i, int which)
1823 {
1824 return TYPE_FIELD_BITPOS (desc_base_type (type), 2 * i + which - 2);
1825 }
1826
1827 /* If BOUNDS is an array-bounds structure type, return the bit field size
1828 of the Ith lower bound stored in it, if WHICH is 0, and the Ith upper
1829 bound, if WHICH is 1. The first bound is I=1. */
1830
1831 static int
1832 desc_bound_bitsize (struct type *type, int i, int which)
1833 {
1834 type = desc_base_type (type);
1835
1836 if (TYPE_FIELD_BITSIZE (type, 2 * i + which - 2) > 0)
1837 return TYPE_FIELD_BITSIZE (type, 2 * i + which - 2);
1838 else
1839 return 8 * TYPE_LENGTH (TYPE_FIELD_TYPE (type, 2 * i + which - 2));
1840 }
1841
1842 /* If TYPE is the type of an array-bounds structure, the type of its
1843 Ith bound (numbering from 1). Otherwise, NULL. */
1844
1845 static struct type *
1846 desc_index_type (struct type *type, int i)
1847 {
1848 type = desc_base_type (type);
1849
1850 if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
1851 return lookup_struct_elt_type (type, bound_name[2 * i - 2], 1);
1852 else
1853 return NULL;
1854 }
1855
1856 /* The number of index positions in the array-bounds type TYPE.
1857 Return 0 if TYPE is NULL. */
1858
1859 static int
1860 desc_arity (struct type *type)
1861 {
1862 type = desc_base_type (type);
1863
1864 if (type != NULL)
1865 return TYPE_NFIELDS (type) / 2;
1866 return 0;
1867 }
1868
1869 /* Non-zero iff TYPE is a simple array type (not a pointer to one) or
1870 an array descriptor type (representing an unconstrained array
1871 type). */
1872
1873 static int
1874 ada_is_direct_array_type (struct type *type)
1875 {
1876 if (type == NULL)
1877 return 0;
1878 type = ada_check_typedef (type);
1879 return (TYPE_CODE (type) == TYPE_CODE_ARRAY
1880 || ada_is_array_descriptor_type (type));
1881 }
1882
1883 /* Non-zero iff TYPE represents any kind of array in Ada, or a pointer
1884 * to one. */
1885
1886 static int
1887 ada_is_array_type (struct type *type)
1888 {
1889 while (type != NULL
1890 && (TYPE_CODE (type) == TYPE_CODE_PTR
1891 || TYPE_CODE (type) == TYPE_CODE_REF))
1892 type = TYPE_TARGET_TYPE (type);
1893 return ada_is_direct_array_type (type);
1894 }
1895
1896 /* Non-zero iff TYPE is a simple array type or pointer to one. */
1897
1898 int
1899 ada_is_simple_array_type (struct type *type)
1900 {
1901 if (type == NULL)
1902 return 0;
1903 type = ada_check_typedef (type);
1904 return (TYPE_CODE (type) == TYPE_CODE_ARRAY
1905 || (TYPE_CODE (type) == TYPE_CODE_PTR
1906 && TYPE_CODE (ada_check_typedef (TYPE_TARGET_TYPE (type)))
1907 == TYPE_CODE_ARRAY));
1908 }
1909
1910 /* Non-zero iff TYPE belongs to a GNAT array descriptor. */
1911
1912 int
1913 ada_is_array_descriptor_type (struct type *type)
1914 {
1915 struct type *data_type = desc_data_target_type (type);
1916
1917 if (type == NULL)
1918 return 0;
1919 type = ada_check_typedef (type);
1920 return (data_type != NULL
1921 && TYPE_CODE (data_type) == TYPE_CODE_ARRAY
1922 && desc_arity (desc_bounds_type (type)) > 0);
1923 }
1924
1925 /* Non-zero iff type is a partially mal-formed GNAT array
1926 descriptor. FIXME: This is to compensate for some problems with
1927 debugging output from GNAT. Re-examine periodically to see if it
1928 is still needed. */
1929
1930 int
1931 ada_is_bogus_array_descriptor (struct type *type)
1932 {
1933 return
1934 type != NULL
1935 && TYPE_CODE (type) == TYPE_CODE_STRUCT
1936 && (lookup_struct_elt_type (type, "P_BOUNDS", 1) != NULL
1937 || lookup_struct_elt_type (type, "P_ARRAY", 1) != NULL)
1938 && !ada_is_array_descriptor_type (type);
1939 }
1940
1941
1942 /* If ARR has a record type in the form of a standard GNAT array descriptor,
1943 (fat pointer) returns the type of the array data described---specifically,
1944 a pointer-to-array type. If BOUNDS is non-zero, the bounds data are filled
1945 in from the descriptor; otherwise, they are left unspecified. If
1946 the ARR denotes a null array descriptor and BOUNDS is non-zero,
1947 returns NULL. The result is simply the type of ARR if ARR is not
1948 a descriptor. */
1949 struct type *
1950 ada_type_of_array (struct value *arr, int bounds)
1951 {
1952 if (ada_is_constrained_packed_array_type (value_type (arr)))
1953 return decode_constrained_packed_array_type (value_type (arr));
1954
1955 if (!ada_is_array_descriptor_type (value_type (arr)))
1956 return value_type (arr);
1957
1958 if (!bounds)
1959 {
1960 struct type *array_type =
1961 ada_check_typedef (desc_data_target_type (value_type (arr)));
1962
1963 if (ada_is_unconstrained_packed_array_type (value_type (arr)))
1964 TYPE_FIELD_BITSIZE (array_type, 0) =
1965 decode_packed_array_bitsize (value_type (arr));
1966
1967 return array_type;
1968 }
1969 else
1970 {
1971 struct type *elt_type;
1972 int arity;
1973 struct value *descriptor;
1974
1975 elt_type = ada_array_element_type (value_type (arr), -1);
1976 arity = ada_array_arity (value_type (arr));
1977
1978 if (elt_type == NULL || arity == 0)
1979 return ada_check_typedef (value_type (arr));
1980
1981 descriptor = desc_bounds (arr);
1982 if (value_as_long (descriptor) == 0)
1983 return NULL;
1984 while (arity > 0)
1985 {
1986 struct type *range_type = alloc_type_copy (value_type (arr));
1987 struct type *array_type = alloc_type_copy (value_type (arr));
1988 struct value *low = desc_one_bound (descriptor, arity, 0);
1989 struct value *high = desc_one_bound (descriptor, arity, 1);
1990
1991 arity -= 1;
1992 create_static_range_type (range_type, value_type (low),
1993 longest_to_int (value_as_long (low)),
1994 longest_to_int (value_as_long (high)));
1995 elt_type = create_array_type (array_type, elt_type, range_type);
1996
1997 if (ada_is_unconstrained_packed_array_type (value_type (arr)))
1998 {
1999 /* We need to store the element packed bitsize, as well as
2000 recompute the array size, because it was previously
2001 computed based on the unpacked element size. */
2002 LONGEST lo = value_as_long (low);
2003 LONGEST hi = value_as_long (high);
2004
2005 TYPE_FIELD_BITSIZE (elt_type, 0) =
2006 decode_packed_array_bitsize (value_type (arr));
2007 /* If the array has no element, then the size is already
2008 zero, and does not need to be recomputed. */
2009 if (lo < hi)
2010 {
2011 int array_bitsize =
2012 (hi - lo + 1) * TYPE_FIELD_BITSIZE (elt_type, 0);
2013
2014 TYPE_LENGTH (array_type) = (array_bitsize + 7) / 8;
2015 }
2016 }
2017 }
2018
2019 return lookup_pointer_type (elt_type);
2020 }
2021 }
2022
2023 /* If ARR does not represent an array, returns ARR unchanged.
2024 Otherwise, returns either a standard GDB array with bounds set
2025 appropriately or, if ARR is a non-null fat pointer, a pointer to a standard
2026 GDB array. Returns NULL if ARR is a null fat pointer. */
2027
2028 struct value *
2029 ada_coerce_to_simple_array_ptr (struct value *arr)
2030 {
2031 if (ada_is_array_descriptor_type (value_type (arr)))
2032 {
2033 struct type *arrType = ada_type_of_array (arr, 1);
2034
2035 if (arrType == NULL)
2036 return NULL;
2037 return value_cast (arrType, value_copy (desc_data (arr)));
2038 }
2039 else if (ada_is_constrained_packed_array_type (value_type (arr)))
2040 return decode_constrained_packed_array (arr);
2041 else
2042 return arr;
2043 }
2044
2045 /* If ARR does not represent an array, returns ARR unchanged.
2046 Otherwise, returns a standard GDB array describing ARR (which may
2047 be ARR itself if it already is in the proper form). */
2048
2049 struct value *
2050 ada_coerce_to_simple_array (struct value *arr)
2051 {
2052 if (ada_is_array_descriptor_type (value_type (arr)))
2053 {
2054 struct value *arrVal = ada_coerce_to_simple_array_ptr (arr);
2055
2056 if (arrVal == NULL)
2057 error (_("Bounds unavailable for null array pointer."));
2058 ada_ensure_varsize_limit (TYPE_TARGET_TYPE (value_type (arrVal)));
2059 return value_ind (arrVal);
2060 }
2061 else if (ada_is_constrained_packed_array_type (value_type (arr)))
2062 return decode_constrained_packed_array (arr);
2063 else
2064 return arr;
2065 }
2066
2067 /* If TYPE represents a GNAT array type, return it translated to an
2068 ordinary GDB array type (possibly with BITSIZE fields indicating
2069 packing). For other types, is the identity. */
2070
2071 struct type *
2072 ada_coerce_to_simple_array_type (struct type *type)
2073 {
2074 if (ada_is_constrained_packed_array_type (type))
2075 return decode_constrained_packed_array_type (type);
2076
2077 if (ada_is_array_descriptor_type (type))
2078 return ada_check_typedef (desc_data_target_type (type));
2079
2080 return type;
2081 }
2082
2083 /* Non-zero iff TYPE represents a standard GNAT packed-array type. */
2084
2085 static int
2086 ada_is_packed_array_type (struct type *type)
2087 {
2088 if (type == NULL)
2089 return 0;
2090 type = desc_base_type (type);
2091 type = ada_check_typedef (type);
2092 return
2093 ada_type_name (type) != NULL
2094 && strstr (ada_type_name (type), "___XP") != NULL;
2095 }
2096
2097 /* Non-zero iff TYPE represents a standard GNAT constrained
2098 packed-array type. */
2099
2100 int
2101 ada_is_constrained_packed_array_type (struct type *type)
2102 {
2103 return ada_is_packed_array_type (type)
2104 && !ada_is_array_descriptor_type (type);
2105 }
2106
2107 /* Non-zero iff TYPE represents an array descriptor for a
2108 unconstrained packed-array type. */
2109
2110 static int
2111 ada_is_unconstrained_packed_array_type (struct type *type)
2112 {
2113 return ada_is_packed_array_type (type)
2114 && ada_is_array_descriptor_type (type);
2115 }
2116
2117 /* Given that TYPE encodes a packed array type (constrained or unconstrained),
2118 return the size of its elements in bits. */
2119
2120 static long
2121 decode_packed_array_bitsize (struct type *type)
2122 {
2123 const char *raw_name;
2124 const char *tail;
2125 long bits;
2126
2127 /* Access to arrays implemented as fat pointers are encoded as a typedef
2128 of the fat pointer type. We need the name of the fat pointer type
2129 to do the decoding, so strip the typedef layer. */
2130 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF)
2131 type = ada_typedef_target_type (type);
2132
2133 raw_name = ada_type_name (ada_check_typedef (type));
2134 if (!raw_name)
2135 raw_name = ada_type_name (desc_base_type (type));
2136
2137 if (!raw_name)
2138 return 0;
2139
2140 tail = strstr (raw_name, "___XP");
2141 gdb_assert (tail != NULL);
2142
2143 if (sscanf (tail + sizeof ("___XP") - 1, "%ld", &bits) != 1)
2144 {
2145 lim_warning
2146 (_("could not understand bit size information on packed array"));
2147 return 0;
2148 }
2149
2150 return bits;
2151 }
2152
2153 /* Given that TYPE is a standard GDB array type with all bounds filled
2154 in, and that the element size of its ultimate scalar constituents
2155 (that is, either its elements, or, if it is an array of arrays, its
2156 elements' elements, etc.) is *ELT_BITS, return an identical type,
2157 but with the bit sizes of its elements (and those of any
2158 constituent arrays) recorded in the BITSIZE components of its
2159 TYPE_FIELD_BITSIZE values, and with *ELT_BITS set to its total size
2160 in bits.
2161
2162 Note that, for arrays whose index type has an XA encoding where
2163 a bound references a record discriminant, getting that discriminant,
2164 and therefore the actual value of that bound, is not possible
2165 because none of the given parameters gives us access to the record.
2166 This function assumes that it is OK in the context where it is being
2167 used to return an array whose bounds are still dynamic and where
2168 the length is arbitrary. */
2169
2170 static struct type *
2171 constrained_packed_array_type (struct type *type, long *elt_bits)
2172 {
2173 struct type *new_elt_type;
2174 struct type *new_type;
2175 struct type *index_type_desc;
2176 struct type *index_type;
2177 LONGEST low_bound, high_bound;
2178
2179 type = ada_check_typedef (type);
2180 if (TYPE_CODE (type) != TYPE_CODE_ARRAY)
2181 return type;
2182
2183 index_type_desc = ada_find_parallel_type (type, "___XA");
2184 if (index_type_desc)
2185 index_type = to_fixed_range_type (TYPE_FIELD_TYPE (index_type_desc, 0),
2186 NULL);
2187 else
2188 index_type = TYPE_INDEX_TYPE (type);
2189
2190 new_type = alloc_type_copy (type);
2191 new_elt_type =
2192 constrained_packed_array_type (ada_check_typedef (TYPE_TARGET_TYPE (type)),
2193 elt_bits);
2194 create_array_type (new_type, new_elt_type, index_type);
2195 TYPE_FIELD_BITSIZE (new_type, 0) = *elt_bits;
2196 TYPE_NAME (new_type) = ada_type_name (type);
2197
2198 if ((TYPE_CODE (check_typedef (index_type)) == TYPE_CODE_RANGE
2199 && is_dynamic_type (check_typedef (index_type)))
2200 || get_discrete_bounds (index_type, &low_bound, &high_bound) < 0)
2201 low_bound = high_bound = 0;
2202 if (high_bound < low_bound)
2203 *elt_bits = TYPE_LENGTH (new_type) = 0;
2204 else
2205 {
2206 *elt_bits *= (high_bound - low_bound + 1);
2207 TYPE_LENGTH (new_type) =
2208 (*elt_bits + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT;
2209 }
2210
2211 TYPE_FIXED_INSTANCE (new_type) = 1;
2212 return new_type;
2213 }
2214
2215 /* The array type encoded by TYPE, where
2216 ada_is_constrained_packed_array_type (TYPE). */
2217
2218 static struct type *
2219 decode_constrained_packed_array_type (struct type *type)
2220 {
2221 const char *raw_name = ada_type_name (ada_check_typedef (type));
2222 char *name;
2223 const char *tail;
2224 struct type *shadow_type;
2225 long bits;
2226
2227 if (!raw_name)
2228 raw_name = ada_type_name (desc_base_type (type));
2229
2230 if (!raw_name)
2231 return NULL;
2232
2233 name = (char *) alloca (strlen (raw_name) + 1);
2234 tail = strstr (raw_name, "___XP");
2235 type = desc_base_type (type);
2236
2237 memcpy (name, raw_name, tail - raw_name);
2238 name[tail - raw_name] = '\000';
2239
2240 shadow_type = ada_find_parallel_type_with_name (type, name);
2241
2242 if (shadow_type == NULL)
2243 {
2244 lim_warning (_("could not find bounds information on packed array"));
2245 return NULL;
2246 }
2247 shadow_type = check_typedef (shadow_type);
2248
2249 if (TYPE_CODE (shadow_type) != TYPE_CODE_ARRAY)
2250 {
2251 lim_warning (_("could not understand bounds "
2252 "information on packed array"));
2253 return NULL;
2254 }
2255
2256 bits = decode_packed_array_bitsize (type);
2257 return constrained_packed_array_type (shadow_type, &bits);
2258 }
2259
2260 /* Given that ARR is a struct value *indicating a GNAT constrained packed
2261 array, returns a simple array that denotes that array. Its type is a
2262 standard GDB array type except that the BITSIZEs of the array
2263 target types are set to the number of bits in each element, and the
2264 type length is set appropriately. */
2265
2266 static struct value *
2267 decode_constrained_packed_array (struct value *arr)
2268 {
2269 struct type *type;
2270
2271 /* If our value is a pointer, then dereference it. Likewise if
2272 the value is a reference. Make sure that this operation does not
2273 cause the target type to be fixed, as this would indirectly cause
2274 this array to be decoded. The rest of the routine assumes that
2275 the array hasn't been decoded yet, so we use the basic "coerce_ref"
2276 and "value_ind" routines to perform the dereferencing, as opposed
2277 to using "ada_coerce_ref" or "ada_value_ind". */
2278 arr = coerce_ref (arr);
2279 if (TYPE_CODE (ada_check_typedef (value_type (arr))) == TYPE_CODE_PTR)
2280 arr = value_ind (arr);
2281
2282 type = decode_constrained_packed_array_type (value_type (arr));
2283 if (type == NULL)
2284 {
2285 error (_("can't unpack array"));
2286 return NULL;
2287 }
2288
2289 if (gdbarch_bits_big_endian (get_type_arch (value_type (arr)))
2290 && ada_is_modular_type (value_type (arr)))
2291 {
2292 /* This is a (right-justified) modular type representing a packed
2293 array with no wrapper. In order to interpret the value through
2294 the (left-justified) packed array type we just built, we must
2295 first left-justify it. */
2296 int bit_size, bit_pos;
2297 ULONGEST mod;
2298
2299 mod = ada_modulus (value_type (arr)) - 1;
2300 bit_size = 0;
2301 while (mod > 0)
2302 {
2303 bit_size += 1;
2304 mod >>= 1;
2305 }
2306 bit_pos = HOST_CHAR_BIT * TYPE_LENGTH (value_type (arr)) - bit_size;
2307 arr = ada_value_primitive_packed_val (arr, NULL,
2308 bit_pos / HOST_CHAR_BIT,
2309 bit_pos % HOST_CHAR_BIT,
2310 bit_size,
2311 type);
2312 }
2313
2314 return coerce_unspec_val_to_type (arr, type);
2315 }
2316
2317
2318 /* The value of the element of packed array ARR at the ARITY indices
2319 given in IND. ARR must be a simple array. */
2320
2321 static struct value *
2322 value_subscript_packed (struct value *arr, int arity, struct value **ind)
2323 {
2324 int i;
2325 int bits, elt_off, bit_off;
2326 long elt_total_bit_offset;
2327 struct type *elt_type;
2328 struct value *v;
2329
2330 bits = 0;
2331 elt_total_bit_offset = 0;
2332 elt_type = ada_check_typedef (value_type (arr));
2333 for (i = 0; i < arity; i += 1)
2334 {
2335 if (TYPE_CODE (elt_type) != TYPE_CODE_ARRAY
2336 || TYPE_FIELD_BITSIZE (elt_type, 0) == 0)
2337 error
2338 (_("attempt to do packed indexing of "
2339 "something other than a packed array"));
2340 else
2341 {
2342 struct type *range_type = TYPE_INDEX_TYPE (elt_type);
2343 LONGEST lowerbound, upperbound;
2344 LONGEST idx;
2345
2346 if (get_discrete_bounds (range_type, &lowerbound, &upperbound) < 0)
2347 {
2348 lim_warning (_("don't know bounds of array"));
2349 lowerbound = upperbound = 0;
2350 }
2351
2352 idx = pos_atr (ind[i]);
2353 if (idx < lowerbound || idx > upperbound)
2354 lim_warning (_("packed array index %ld out of bounds"),
2355 (long) idx);
2356 bits = TYPE_FIELD_BITSIZE (elt_type, 0);
2357 elt_total_bit_offset += (idx - lowerbound) * bits;
2358 elt_type = ada_check_typedef (TYPE_TARGET_TYPE (elt_type));
2359 }
2360 }
2361 elt_off = elt_total_bit_offset / HOST_CHAR_BIT;
2362 bit_off = elt_total_bit_offset % HOST_CHAR_BIT;
2363
2364 v = ada_value_primitive_packed_val (arr, NULL, elt_off, bit_off,
2365 bits, elt_type);
2366 return v;
2367 }
2368
2369 /* Non-zero iff TYPE includes negative integer values. */
2370
2371 static int
2372 has_negatives (struct type *type)
2373 {
2374 switch (TYPE_CODE (type))
2375 {
2376 default:
2377 return 0;
2378 case TYPE_CODE_INT:
2379 return !TYPE_UNSIGNED (type);
2380 case TYPE_CODE_RANGE:
2381 return TYPE_LOW_BOUND (type) < 0;
2382 }
2383 }
2384
2385 /* With SRC being a buffer containing BIT_SIZE bits of data at BIT_OFFSET,
2386 unpack that data into UNPACKED. UNPACKED_LEN is the size in bytes of
2387 the unpacked buffer.
2388
2389 The size of the unpacked buffer (UNPACKED_LEN) is expected to be large
2390 enough to contain at least BIT_OFFSET bits. If not, an error is raised.
2391
2392 IS_BIG_ENDIAN is nonzero if the data is stored in big endian mode,
2393 zero otherwise.
2394
2395 IS_SIGNED_TYPE is nonzero if the data corresponds to a signed type.
2396
2397 IS_SCALAR is nonzero if the data corresponds to a signed type. */
2398
2399 static void
2400 ada_unpack_from_contents (const gdb_byte *src, int bit_offset, int bit_size,
2401 gdb_byte *unpacked, int unpacked_len,
2402 int is_big_endian, int is_signed_type,
2403 int is_scalar)
2404 {
2405 int src_len = (bit_size + bit_offset + HOST_CHAR_BIT - 1) / 8;
2406 int src_idx; /* Index into the source area */
2407 int src_bytes_left; /* Number of source bytes left to process. */
2408 int srcBitsLeft; /* Number of source bits left to move */
2409 int unusedLS; /* Number of bits in next significant
2410 byte of source that are unused */
2411
2412 int unpacked_idx; /* Index into the unpacked buffer */
2413 int unpacked_bytes_left; /* Number of bytes left to set in unpacked. */
2414
2415 unsigned long accum; /* Staging area for bits being transferred */
2416 int accumSize; /* Number of meaningful bits in accum */
2417 unsigned char sign;
2418
2419 /* Transmit bytes from least to most significant; delta is the direction
2420 the indices move. */
2421 int delta = is_big_endian ? -1 : 1;
2422
2423 /* Make sure that unpacked is large enough to receive the BIT_SIZE
2424 bits from SRC. .*/
2425 if ((bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT > unpacked_len)
2426 error (_("Cannot unpack %d bits into buffer of %d bytes"),
2427 bit_size, unpacked_len);
2428
2429 srcBitsLeft = bit_size;
2430 src_bytes_left = src_len;
2431 unpacked_bytes_left = unpacked_len;
2432 sign = 0;
2433
2434 if (is_big_endian)
2435 {
2436 src_idx = src_len - 1;
2437 if (is_signed_type
2438 && ((src[0] << bit_offset) & (1 << (HOST_CHAR_BIT - 1))))
2439 sign = ~0;
2440
2441 unusedLS =
2442 (HOST_CHAR_BIT - (bit_size + bit_offset) % HOST_CHAR_BIT)
2443 % HOST_CHAR_BIT;
2444
2445 if (is_scalar)
2446 {
2447 accumSize = 0;
2448 unpacked_idx = unpacked_len - 1;
2449 }
2450 else
2451 {
2452 /* Non-scalar values must be aligned at a byte boundary... */
2453 accumSize =
2454 (HOST_CHAR_BIT - bit_size % HOST_CHAR_BIT) % HOST_CHAR_BIT;
2455 /* ... And are placed at the beginning (most-significant) bytes
2456 of the target. */
2457 unpacked_idx = (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT - 1;
2458 unpacked_bytes_left = unpacked_idx + 1;
2459 }
2460 }
2461 else
2462 {
2463 int sign_bit_offset = (bit_size + bit_offset - 1) % 8;
2464
2465 src_idx = unpacked_idx = 0;
2466 unusedLS = bit_offset;
2467 accumSize = 0;
2468
2469 if (is_signed_type && (src[src_len - 1] & (1 << sign_bit_offset)))
2470 sign = ~0;
2471 }
2472
2473 accum = 0;
2474 while (src_bytes_left > 0)
2475 {
2476 /* Mask for removing bits of the next source byte that are not
2477 part of the value. */
2478 unsigned int unusedMSMask =
2479 (1 << (srcBitsLeft >= HOST_CHAR_BIT ? HOST_CHAR_BIT : srcBitsLeft)) -
2480 1;
2481 /* Sign-extend bits for this byte. */
2482 unsigned int signMask = sign & ~unusedMSMask;
2483
2484 accum |=
2485 (((src[src_idx] >> unusedLS) & unusedMSMask) | signMask) << accumSize;
2486 accumSize += HOST_CHAR_BIT - unusedLS;
2487 if (accumSize >= HOST_CHAR_BIT)
2488 {
2489 unpacked[unpacked_idx] = accum & ~(~0L << HOST_CHAR_BIT);
2490 accumSize -= HOST_CHAR_BIT;
2491 accum >>= HOST_CHAR_BIT;
2492 unpacked_bytes_left -= 1;
2493 unpacked_idx += delta;
2494 }
2495 srcBitsLeft -= HOST_CHAR_BIT - unusedLS;
2496 unusedLS = 0;
2497 src_bytes_left -= 1;
2498 src_idx += delta;
2499 }
2500 while (unpacked_bytes_left > 0)
2501 {
2502 accum |= sign << accumSize;
2503 unpacked[unpacked_idx] = accum & ~(~0L << HOST_CHAR_BIT);
2504 accumSize -= HOST_CHAR_BIT;
2505 if (accumSize < 0)
2506 accumSize = 0;
2507 accum >>= HOST_CHAR_BIT;
2508 unpacked_bytes_left -= 1;
2509 unpacked_idx += delta;
2510 }
2511 }
2512
2513 /* Create a new value of type TYPE from the contents of OBJ starting
2514 at byte OFFSET, and bit offset BIT_OFFSET within that byte,
2515 proceeding for BIT_SIZE bits. If OBJ is an lval in memory, then
2516 assigning through the result will set the field fetched from.
2517 VALADDR is ignored unless OBJ is NULL, in which case,
2518 VALADDR+OFFSET must address the start of storage containing the
2519 packed value. The value returned in this case is never an lval.
2520 Assumes 0 <= BIT_OFFSET < HOST_CHAR_BIT. */
2521
2522 struct value *
2523 ada_value_primitive_packed_val (struct value *obj, const gdb_byte *valaddr,
2524 long offset, int bit_offset, int bit_size,
2525 struct type *type)
2526 {
2527 struct value *v;
2528 gdb_byte *src; /* First byte containing data to unpack */
2529 gdb_byte *unpacked;
2530 const int is_scalar = is_scalar_type (type);
2531 const int is_big_endian = gdbarch_bits_big_endian (get_type_arch (type));
2532 gdb_byte *staging = NULL;
2533 int staging_len = 0;
2534 struct cleanup *old_chain = make_cleanup (null_cleanup, NULL);
2535
2536 type = ada_check_typedef (type);
2537
2538 if (obj == NULL)
2539 src = (gdb_byte *) valaddr + offset;
2540 else
2541 src = (gdb_byte *) value_contents (obj) + offset;
2542
2543 if (is_dynamic_type (type))
2544 {
2545 /* The length of TYPE might by dynamic, so we need to resolve
2546 TYPE in order to know its actual size, which we then use
2547 to create the contents buffer of the value we return.
2548 The difficulty is that the data containing our object is
2549 packed, and therefore maybe not at a byte boundary. So, what
2550 we do, is unpack the data into a byte-aligned buffer, and then
2551 use that buffer as our object's value for resolving the type. */
2552 staging_len = (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT;
2553 staging = malloc (staging_len);
2554 make_cleanup (xfree, staging);
2555
2556 ada_unpack_from_contents (src, bit_offset, bit_size,
2557 staging, staging_len,
2558 is_big_endian, has_negatives (type),
2559 is_scalar);
2560 type = resolve_dynamic_type (type, staging, 0);
2561 if (TYPE_LENGTH (type) < (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT)
2562 {
2563 /* This happens when the length of the object is dynamic,
2564 and is actually smaller than the space reserved for it.
2565 For instance, in an array of variant records, the bit_size
2566 we're given is the array stride, which is constant and
2567 normally equal to the maximum size of its element.
2568 But, in reality, each element only actually spans a portion
2569 of that stride. */
2570 bit_size = TYPE_LENGTH (type) * HOST_CHAR_BIT;
2571 }
2572 }
2573
2574 if (obj == NULL)
2575 {
2576 v = allocate_value (type);
2577 src = (gdb_byte *) valaddr + offset;
2578 }
2579 else if (VALUE_LVAL (obj) == lval_memory && value_lazy (obj))
2580 {
2581 int src_len = (bit_size + bit_offset + HOST_CHAR_BIT - 1) / 8;
2582
2583 v = value_at (type, value_address (obj) + offset);
2584 src = alloca (src_len);
2585 read_memory (value_address (v), src, src_len);
2586 }
2587 else
2588 {
2589 v = allocate_value (type);
2590 src = (gdb_byte *) value_contents (obj) + offset;
2591 }
2592
2593 if (obj != NULL)
2594 {
2595 long new_offset = offset;
2596
2597 set_value_component_location (v, obj);
2598 set_value_bitpos (v, bit_offset + value_bitpos (obj));
2599 set_value_bitsize (v, bit_size);
2600 if (value_bitpos (v) >= HOST_CHAR_BIT)
2601 {
2602 ++new_offset;
2603 set_value_bitpos (v, value_bitpos (v) - HOST_CHAR_BIT);
2604 }
2605 set_value_offset (v, new_offset);
2606
2607 /* Also set the parent value. This is needed when trying to
2608 assign a new value (in inferior memory). */
2609 set_value_parent (v, obj);
2610 }
2611 else
2612 set_value_bitsize (v, bit_size);
2613 unpacked = (gdb_byte *) value_contents (v);
2614
2615 if (bit_size == 0)
2616 {
2617 memset (unpacked, 0, TYPE_LENGTH (type));
2618 do_cleanups (old_chain);
2619 return v;
2620 }
2621
2622 if (staging != NULL && staging_len == TYPE_LENGTH (type))
2623 {
2624 /* Small short-cut: If we've unpacked the data into a buffer
2625 of the same size as TYPE's length, then we can reuse that,
2626 instead of doing the unpacking again. */
2627 memcpy (unpacked, staging, staging_len);
2628 }
2629 else
2630 ada_unpack_from_contents (src, bit_offset, bit_size,
2631 unpacked, TYPE_LENGTH (type),
2632 is_big_endian, has_negatives (type), is_scalar);
2633
2634 do_cleanups (old_chain);
2635 return v;
2636 }
2637
2638 /* Move N bits from SOURCE, starting at bit offset SRC_OFFSET to
2639 TARGET, starting at bit offset TARG_OFFSET. SOURCE and TARGET must
2640 not overlap. */
2641 static void
2642 move_bits (gdb_byte *target, int targ_offset, const gdb_byte *source,
2643 int src_offset, int n, int bits_big_endian_p)
2644 {
2645 unsigned int accum, mask;
2646 int accum_bits, chunk_size;
2647
2648 target += targ_offset / HOST_CHAR_BIT;
2649 targ_offset %= HOST_CHAR_BIT;
2650 source += src_offset / HOST_CHAR_BIT;
2651 src_offset %= HOST_CHAR_BIT;
2652 if (bits_big_endian_p)
2653 {
2654 accum = (unsigned char) *source;
2655 source += 1;
2656 accum_bits = HOST_CHAR_BIT - src_offset;
2657
2658 while (n > 0)
2659 {
2660 int unused_right;
2661
2662 accum = (accum << HOST_CHAR_BIT) + (unsigned char) *source;
2663 accum_bits += HOST_CHAR_BIT;
2664 source += 1;
2665 chunk_size = HOST_CHAR_BIT - targ_offset;
2666 if (chunk_size > n)
2667 chunk_size = n;
2668 unused_right = HOST_CHAR_BIT - (chunk_size + targ_offset);
2669 mask = ((1 << chunk_size) - 1) << unused_right;
2670 *target =
2671 (*target & ~mask)
2672 | ((accum >> (accum_bits - chunk_size - unused_right)) & mask);
2673 n -= chunk_size;
2674 accum_bits -= chunk_size;
2675 target += 1;
2676 targ_offset = 0;
2677 }
2678 }
2679 else
2680 {
2681 accum = (unsigned char) *source >> src_offset;
2682 source += 1;
2683 accum_bits = HOST_CHAR_BIT - src_offset;
2684
2685 while (n > 0)
2686 {
2687 accum = accum + ((unsigned char) *source << accum_bits);
2688 accum_bits += HOST_CHAR_BIT;
2689 source += 1;
2690 chunk_size = HOST_CHAR_BIT - targ_offset;
2691 if (chunk_size > n)
2692 chunk_size = n;
2693 mask = ((1 << chunk_size) - 1) << targ_offset;
2694 *target = (*target & ~mask) | ((accum << targ_offset) & mask);
2695 n -= chunk_size;
2696 accum_bits -= chunk_size;
2697 accum >>= chunk_size;
2698 target += 1;
2699 targ_offset = 0;
2700 }
2701 }
2702 }
2703
2704 /* Store the contents of FROMVAL into the location of TOVAL.
2705 Return a new value with the location of TOVAL and contents of
2706 FROMVAL. Handles assignment into packed fields that have
2707 floating-point or non-scalar types. */
2708
2709 static struct value *
2710 ada_value_assign (struct value *toval, struct value *fromval)
2711 {
2712 struct type *type = value_type (toval);
2713 int bits = value_bitsize (toval);
2714
2715 toval = ada_coerce_ref (toval);
2716 fromval = ada_coerce_ref (fromval);
2717
2718 if (ada_is_direct_array_type (value_type (toval)))
2719 toval = ada_coerce_to_simple_array (toval);
2720 if (ada_is_direct_array_type (value_type (fromval)))
2721 fromval = ada_coerce_to_simple_array (fromval);
2722
2723 if (!deprecated_value_modifiable (toval))
2724 error (_("Left operand of assignment is not a modifiable lvalue."));
2725
2726 if (VALUE_LVAL (toval) == lval_memory
2727 && bits > 0
2728 && (TYPE_CODE (type) == TYPE_CODE_FLT
2729 || TYPE_CODE (type) == TYPE_CODE_STRUCT))
2730 {
2731 int len = (value_bitpos (toval)
2732 + bits + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT;
2733 int from_size;
2734 gdb_byte *buffer = (gdb_byte *) alloca (len);
2735 struct value *val;
2736 CORE_ADDR to_addr = value_address (toval);
2737
2738 if (TYPE_CODE (type) == TYPE_CODE_FLT)
2739 fromval = value_cast (type, fromval);
2740
2741 read_memory (to_addr, buffer, len);
2742 from_size = value_bitsize (fromval);
2743 if (from_size == 0)
2744 from_size = TYPE_LENGTH (value_type (fromval)) * TARGET_CHAR_BIT;
2745 if (gdbarch_bits_big_endian (get_type_arch (type)))
2746 move_bits (buffer, value_bitpos (toval),
2747 value_contents (fromval), from_size - bits, bits, 1);
2748 else
2749 move_bits (buffer, value_bitpos (toval),
2750 value_contents (fromval), 0, bits, 0);
2751 write_memory_with_notification (to_addr, buffer, len);
2752
2753 val = value_copy (toval);
2754 memcpy (value_contents_raw (val), value_contents (fromval),
2755 TYPE_LENGTH (type));
2756 deprecated_set_value_type (val, type);
2757
2758 return val;
2759 }
2760
2761 return value_assign (toval, fromval);
2762 }
2763
2764
2765 /* Given that COMPONENT is a memory lvalue that is part of the lvalue
2766 CONTAINER, assign the contents of VAL to COMPONENTS's place in
2767 CONTAINER. Modifies the VALUE_CONTENTS of CONTAINER only, not
2768 COMPONENT, and not the inferior's memory. The current contents
2769 of COMPONENT are ignored.
2770
2771 Although not part of the initial design, this function also works
2772 when CONTAINER and COMPONENT are not_lval's: it works as if CONTAINER
2773 had a null address, and COMPONENT had an address which is equal to
2774 its offset inside CONTAINER. */
2775
2776 static void
2777 value_assign_to_component (struct value *container, struct value *component,
2778 struct value *val)
2779 {
2780 LONGEST offset_in_container =
2781 (LONGEST) (value_address (component) - value_address (container));
2782 int bit_offset_in_container =
2783 value_bitpos (component) - value_bitpos (container);
2784 int bits;
2785
2786 val = value_cast (value_type (component), val);
2787
2788 if (value_bitsize (component) == 0)
2789 bits = TARGET_CHAR_BIT * TYPE_LENGTH (value_type (component));
2790 else
2791 bits = value_bitsize (component);
2792
2793 if (gdbarch_bits_big_endian (get_type_arch (value_type (container))))
2794 move_bits (value_contents_writeable (container) + offset_in_container,
2795 value_bitpos (container) + bit_offset_in_container,
2796 value_contents (val),
2797 TYPE_LENGTH (value_type (component)) * TARGET_CHAR_BIT - bits,
2798 bits, 1);
2799 else
2800 move_bits (value_contents_writeable (container) + offset_in_container,
2801 value_bitpos (container) + bit_offset_in_container,
2802 value_contents (val), 0, bits, 0);
2803 }
2804
2805 /* The value of the element of array ARR at the ARITY indices given in IND.
2806 ARR may be either a simple array, GNAT array descriptor, or pointer
2807 thereto. */
2808
2809 struct value *
2810 ada_value_subscript (struct value *arr, int arity, struct value **ind)
2811 {
2812 int k;
2813 struct value *elt;
2814 struct type *elt_type;
2815
2816 elt = ada_coerce_to_simple_array (arr);
2817
2818 elt_type = ada_check_typedef (value_type (elt));
2819 if (TYPE_CODE (elt_type) == TYPE_CODE_ARRAY
2820 && TYPE_FIELD_BITSIZE (elt_type, 0) > 0)
2821 return value_subscript_packed (elt, arity, ind);
2822
2823 for (k = 0; k < arity; k += 1)
2824 {
2825 if (TYPE_CODE (elt_type) != TYPE_CODE_ARRAY)
2826 error (_("too many subscripts (%d expected)"), k);
2827 elt = value_subscript (elt, pos_atr (ind[k]));
2828 }
2829 return elt;
2830 }
2831
2832 /* Assuming ARR is a pointer to a GDB array, the value of the element
2833 of *ARR at the ARITY indices given in IND.
2834 Does not read the entire array into memory.
2835
2836 Note: Unlike what one would expect, this function is used instead of
2837 ada_value_subscript for basically all non-packed array types. The reason
2838 for this is that a side effect of doing our own pointer arithmetics instead
2839 of relying on value_subscript is that there is no implicit typedef peeling.
2840 This is important for arrays of array accesses, where it allows us to
2841 preserve the fact that the array's element is an array access, where the
2842 access part os encoded in a typedef layer. */
2843
2844 static struct value *
2845 ada_value_ptr_subscript (struct value *arr, int arity, struct value **ind)
2846 {
2847 int k;
2848 struct value *array_ind = ada_value_ind (arr);
2849 struct type *type
2850 = check_typedef (value_enclosing_type (array_ind));
2851
2852 if (TYPE_CODE (type) == TYPE_CODE_ARRAY
2853 && TYPE_FIELD_BITSIZE (type, 0) > 0)
2854 return value_subscript_packed (array_ind, arity, ind);
2855
2856 for (k = 0; k < arity; k += 1)
2857 {
2858 LONGEST lwb, upb;
2859 struct value *lwb_value;
2860
2861 if (TYPE_CODE (type) != TYPE_CODE_ARRAY)
2862 error (_("too many subscripts (%d expected)"), k);
2863 arr = value_cast (lookup_pointer_type (TYPE_TARGET_TYPE (type)),
2864 value_copy (arr));
2865 get_discrete_bounds (TYPE_INDEX_TYPE (type), &lwb, &upb);
2866 lwb_value = value_from_longest (value_type(ind[k]), lwb);
2867 arr = value_ptradd (arr, pos_atr (ind[k]) - pos_atr (lwb_value));
2868 type = TYPE_TARGET_TYPE (type);
2869 }
2870
2871 return value_ind (arr);
2872 }
2873
2874 /* Given that ARRAY_PTR is a pointer or reference to an array of type TYPE (the
2875 actual type of ARRAY_PTR is ignored), returns the Ada slice of
2876 HIGH'Pos-LOW'Pos+1 elements starting at index LOW. The lower bound of
2877 this array is LOW, as per Ada rules. */
2878 static struct value *
2879 ada_value_slice_from_ptr (struct value *array_ptr, struct type *type,
2880 int low, int high)
2881 {
2882 struct type *type0 = ada_check_typedef (type);
2883 struct type *base_index_type = TYPE_TARGET_TYPE (TYPE_INDEX_TYPE (type0));
2884 struct type *index_type
2885 = create_static_range_type (NULL, base_index_type, low, high);
2886 struct type *slice_type =
2887 create_array_type (NULL, TYPE_TARGET_TYPE (type0), index_type);
2888 int base_low = ada_discrete_type_low_bound (TYPE_INDEX_TYPE (type0));
2889 LONGEST base_low_pos, low_pos;
2890 CORE_ADDR base;
2891
2892 if (!discrete_position (base_index_type, low, &low_pos)
2893 || !discrete_position (base_index_type, base_low, &base_low_pos))
2894 {
2895 warning (_("unable to get positions in slice, use bounds instead"));
2896 low_pos = low;
2897 base_low_pos = base_low;
2898 }
2899
2900 base = value_as_address (array_ptr)
2901 + ((low_pos - base_low_pos)
2902 * TYPE_LENGTH (TYPE_TARGET_TYPE (type0)));
2903 return value_at_lazy (slice_type, base);
2904 }
2905
2906
2907 static struct value *
2908 ada_value_slice (struct value *array, int low, int high)
2909 {
2910 struct type *type = ada_check_typedef (value_type (array));
2911 struct type *base_index_type = TYPE_TARGET_TYPE (TYPE_INDEX_TYPE (type));
2912 struct type *index_type
2913 = create_static_range_type (NULL, TYPE_INDEX_TYPE (type), low, high);
2914 struct type *slice_type =
2915 create_array_type (NULL, TYPE_TARGET_TYPE (type), index_type);
2916 LONGEST low_pos, high_pos;
2917
2918 if (!discrete_position (base_index_type, low, &low_pos)
2919 || !discrete_position (base_index_type, high, &high_pos))
2920 {
2921 warning (_("unable to get positions in slice, use bounds instead"));
2922 low_pos = low;
2923 high_pos = high;
2924 }
2925
2926 return value_cast (slice_type,
2927 value_slice (array, low, high_pos - low_pos + 1));
2928 }
2929
2930 /* If type is a record type in the form of a standard GNAT array
2931 descriptor, returns the number of dimensions for type. If arr is a
2932 simple array, returns the number of "array of"s that prefix its
2933 type designation. Otherwise, returns 0. */
2934
2935 int
2936 ada_array_arity (struct type *type)
2937 {
2938 int arity;
2939
2940 if (type == NULL)
2941 return 0;
2942
2943 type = desc_base_type (type);
2944
2945 arity = 0;
2946 if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
2947 return desc_arity (desc_bounds_type (type));
2948 else
2949 while (TYPE_CODE (type) == TYPE_CODE_ARRAY)
2950 {
2951 arity += 1;
2952 type = ada_check_typedef (TYPE_TARGET_TYPE (type));
2953 }
2954
2955 return arity;
2956 }
2957
2958 /* If TYPE is a record type in the form of a standard GNAT array
2959 descriptor or a simple array type, returns the element type for
2960 TYPE after indexing by NINDICES indices, or by all indices if
2961 NINDICES is -1. Otherwise, returns NULL. */
2962
2963 struct type *
2964 ada_array_element_type (struct type *type, int nindices)
2965 {
2966 type = desc_base_type (type);
2967
2968 if (TYPE_CODE (type) == TYPE_CODE_STRUCT)
2969 {
2970 int k;
2971 struct type *p_array_type;
2972
2973 p_array_type = desc_data_target_type (type);
2974
2975 k = ada_array_arity (type);
2976 if (k == 0)
2977 return NULL;
2978
2979 /* Initially p_array_type = elt_type(*)[]...(k times)...[]. */
2980 if (nindices >= 0 && k > nindices)
2981 k = nindices;
2982 while (k > 0 && p_array_type != NULL)
2983 {
2984 p_array_type = ada_check_typedef (TYPE_TARGET_TYPE (p_array_type));
2985 k -= 1;
2986 }
2987 return p_array_type;
2988 }
2989 else if (TYPE_CODE (type) == TYPE_CODE_ARRAY)
2990 {
2991 while (nindices != 0 && TYPE_CODE (type) == TYPE_CODE_ARRAY)
2992 {
2993 type = TYPE_TARGET_TYPE (type);
2994 nindices -= 1;
2995 }
2996 return type;
2997 }
2998
2999 return NULL;
3000 }
3001
3002 /* The type of nth index in arrays of given type (n numbering from 1).
3003 Does not examine memory. Throws an error if N is invalid or TYPE
3004 is not an array type. NAME is the name of the Ada attribute being
3005 evaluated ('range, 'first, 'last, or 'length); it is used in building
3006 the error message. */
3007
3008 static struct type *
3009 ada_index_type (struct type *type, int n, const char *name)
3010 {
3011 struct type *result_type;
3012
3013 type = desc_base_type (type);
3014
3015 if (n < 0 || n > ada_array_arity (type))
3016 error (_("invalid dimension number to '%s"), name);
3017
3018 if (ada_is_simple_array_type (type))
3019 {
3020 int i;
3021
3022 for (i = 1; i < n; i += 1)
3023 type = TYPE_TARGET_TYPE (type);
3024 result_type = TYPE_TARGET_TYPE (TYPE_INDEX_TYPE (type));
3025 /* FIXME: The stabs type r(0,0);bound;bound in an array type
3026 has a target type of TYPE_CODE_UNDEF. We compensate here, but
3027 perhaps stabsread.c would make more sense. */
3028 if (result_type && TYPE_CODE (result_type) == TYPE_CODE_UNDEF)
3029 result_type = NULL;
3030 }
3031 else
3032 {
3033 result_type = desc_index_type (desc_bounds_type (type), n);
3034 if (result_type == NULL)
3035 error (_("attempt to take bound of something that is not an array"));
3036 }
3037
3038 return result_type;
3039 }
3040
3041 /* Given that arr is an array type, returns the lower bound of the
3042 Nth index (numbering from 1) if WHICH is 0, and the upper bound if
3043 WHICH is 1. This returns bounds 0 .. -1 if ARR_TYPE is an
3044 array-descriptor type. It works for other arrays with bounds supplied
3045 by run-time quantities other than discriminants. */
3046
3047 static LONGEST
3048 ada_array_bound_from_type (struct type *arr_type, int n, int which)
3049 {
3050 struct type *type, *index_type_desc, *index_type;
3051 int i;
3052
3053 gdb_assert (which == 0 || which == 1);
3054
3055 if (ada_is_constrained_packed_array_type (arr_type))
3056 arr_type = decode_constrained_packed_array_type (arr_type);
3057
3058 if (arr_type == NULL || !ada_is_simple_array_type (arr_type))
3059 return (LONGEST) - which;
3060
3061 if (TYPE_CODE (arr_type) == TYPE_CODE_PTR)
3062 type = TYPE_TARGET_TYPE (arr_type);
3063 else
3064 type = arr_type;
3065
3066 if (TYPE_FIXED_INSTANCE (type))
3067 {
3068 /* The array has already been fixed, so we do not need to
3069 check the parallel ___XA type again. That encoding has
3070 already been applied, so ignore it now. */
3071 index_type_desc = NULL;
3072 }
3073 else
3074 {
3075 index_type_desc = ada_find_parallel_type (type, "___XA");
3076 ada_fixup_array_indexes_type (index_type_desc);
3077 }
3078
3079 if (index_type_desc != NULL)
3080 index_type = to_fixed_range_type (TYPE_FIELD_TYPE (index_type_desc, n - 1),
3081 NULL);
3082 else
3083 {
3084 struct type *elt_type = check_typedef (type);
3085
3086 for (i = 1; i < n; i++)
3087 elt_type = check_typedef (TYPE_TARGET_TYPE (elt_type));
3088
3089 index_type = TYPE_INDEX_TYPE (elt_type);
3090 }
3091
3092 return
3093 (LONGEST) (which == 0
3094 ? ada_discrete_type_low_bound (index_type)
3095 : ada_discrete_type_high_bound (index_type));
3096 }
3097
3098 /* Given that arr is an array value, returns the lower bound of the
3099 nth index (numbering from 1) if WHICH is 0, and the upper bound if
3100 WHICH is 1. This routine will also work for arrays with bounds
3101 supplied by run-time quantities other than discriminants. */
3102
3103 static LONGEST
3104 ada_array_bound (struct value *arr, int n, int which)
3105 {
3106 struct type *arr_type;
3107
3108 if (TYPE_CODE (check_typedef (value_type (arr))) == TYPE_CODE_PTR)
3109 arr = value_ind (arr);
3110 arr_type = value_enclosing_type (arr);
3111
3112 if (ada_is_constrained_packed_array_type (arr_type))
3113 return ada_array_bound (decode_constrained_packed_array (arr), n, which);
3114 else if (ada_is_simple_array_type (arr_type))
3115 return ada_array_bound_from_type (arr_type, n, which);
3116 else
3117 return value_as_long (desc_one_bound (desc_bounds (arr), n, which));
3118 }
3119
3120 /* Given that arr is an array value, returns the length of the
3121 nth index. This routine will also work for arrays with bounds
3122 supplied by run-time quantities other than discriminants.
3123 Does not work for arrays indexed by enumeration types with representation
3124 clauses at the moment. */
3125
3126 static LONGEST
3127 ada_array_length (struct value *arr, int n)
3128 {
3129 struct type *arr_type, *index_type;
3130 int low, high;
3131
3132 if (TYPE_CODE (check_typedef (value_type (arr))) == TYPE_CODE_PTR)
3133 arr = value_ind (arr);
3134 arr_type = value_enclosing_type (arr);
3135
3136 if (ada_is_constrained_packed_array_type (arr_type))
3137 return ada_array_length (decode_constrained_packed_array (arr), n);
3138
3139 if (ada_is_simple_array_type (arr_type))
3140 {
3141 low = ada_array_bound_from_type (arr_type, n, 0);
3142 high = ada_array_bound_from_type (arr_type, n, 1);
3143 }
3144 else
3145 {
3146 low = value_as_long (desc_one_bound (desc_bounds (arr), n, 0));
3147 high = value_as_long (desc_one_bound (desc_bounds (arr), n, 1));
3148 }
3149
3150 arr_type = check_typedef (arr_type);
3151 index_type = TYPE_INDEX_TYPE (arr_type);
3152 if (index_type != NULL)
3153 {
3154 struct type *base_type;
3155 if (TYPE_CODE (index_type) == TYPE_CODE_RANGE)
3156 base_type = TYPE_TARGET_TYPE (index_type);
3157 else
3158 base_type = index_type;
3159
3160 low = pos_atr (value_from_longest (base_type, low));
3161 high = pos_atr (value_from_longest (base_type, high));
3162 }
3163 return high - low + 1;
3164 }
3165
3166 /* An empty array whose type is that of ARR_TYPE (an array type),
3167 with bounds LOW to LOW-1. */
3168
3169 static struct value *
3170 empty_array (struct type *arr_type, int low)
3171 {
3172 struct type *arr_type0 = ada_check_typedef (arr_type);
3173 struct type *index_type
3174 = create_static_range_type
3175 (NULL, TYPE_TARGET_TYPE (TYPE_INDEX_TYPE (arr_type0)), low, low - 1);
3176 struct type *elt_type = ada_array_element_type (arr_type0, 1);
3177
3178 return allocate_value (create_array_type (NULL, elt_type, index_type));
3179 }
3180 \f
3181
3182 /* Name resolution */
3183
3184 /* The "decoded" name for the user-definable Ada operator corresponding
3185 to OP. */
3186
3187 static const char *
3188 ada_decoded_op_name (enum exp_opcode op)
3189 {
3190 int i;
3191
3192 for (i = 0; ada_opname_table[i].encoded != NULL; i += 1)
3193 {
3194 if (ada_opname_table[i].op == op)
3195 return ada_opname_table[i].decoded;
3196 }
3197 error (_("Could not find operator name for opcode"));
3198 }
3199
3200
3201 /* Same as evaluate_type (*EXP), but resolves ambiguous symbol
3202 references (marked by OP_VAR_VALUE nodes in which the symbol has an
3203 undefined namespace) and converts operators that are
3204 user-defined into appropriate function calls. If CONTEXT_TYPE is
3205 non-null, it provides a preferred result type [at the moment, only
3206 type void has any effect---causing procedures to be preferred over
3207 functions in calls]. A null CONTEXT_TYPE indicates that a non-void
3208 return type is preferred. May change (expand) *EXP. */
3209
3210 static void
3211 resolve (struct expression **expp, int void_context_p)
3212 {
3213 struct type *context_type = NULL;
3214 int pc = 0;
3215
3216 if (void_context_p)
3217 context_type = builtin_type ((*expp)->gdbarch)->builtin_void;
3218
3219 resolve_subexp (expp, &pc, 1, context_type);
3220 }
3221
3222 /* Resolve the operator of the subexpression beginning at
3223 position *POS of *EXPP. "Resolving" consists of replacing
3224 the symbols that have undefined namespaces in OP_VAR_VALUE nodes
3225 with their resolutions, replacing built-in operators with
3226 function calls to user-defined operators, where appropriate, and,
3227 when DEPROCEDURE_P is non-zero, converting function-valued variables
3228 into parameterless calls. May expand *EXPP. The CONTEXT_TYPE functions
3229 are as in ada_resolve, above. */
3230
3231 static struct value *
3232 resolve_subexp (struct expression **expp, int *pos, int deprocedure_p,
3233 struct type *context_type)
3234 {
3235 int pc = *pos;
3236 int i;
3237 struct expression *exp; /* Convenience: == *expp. */
3238 enum exp_opcode op = (*expp)->elts[pc].opcode;
3239 struct value **argvec; /* Vector of operand types (alloca'ed). */
3240 int nargs; /* Number of operands. */
3241 int oplen;
3242
3243 argvec = NULL;
3244 nargs = 0;
3245 exp = *expp;
3246
3247 /* Pass one: resolve operands, saving their types and updating *pos,
3248 if needed. */
3249 switch (op)
3250 {
3251 case OP_FUNCALL:
3252 if (exp->elts[pc + 3].opcode == OP_VAR_VALUE
3253 && SYMBOL_DOMAIN (exp->elts[pc + 5].symbol) == UNDEF_DOMAIN)
3254 *pos += 7;
3255 else
3256 {
3257 *pos += 3;
3258 resolve_subexp (expp, pos, 0, NULL);
3259 }
3260 nargs = longest_to_int (exp->elts[pc + 1].longconst);
3261 break;
3262
3263 case UNOP_ADDR:
3264 *pos += 1;
3265 resolve_subexp (expp, pos, 0, NULL);
3266 break;
3267
3268 case UNOP_QUAL:
3269 *pos += 3;
3270 resolve_subexp (expp, pos, 1, check_typedef (exp->elts[pc + 1].type));
3271 break;
3272
3273 case OP_ATR_MODULUS:
3274 case OP_ATR_SIZE:
3275 case OP_ATR_TAG:
3276 case OP_ATR_FIRST:
3277 case OP_ATR_LAST:
3278 case OP_ATR_LENGTH:
3279 case OP_ATR_POS:
3280 case OP_ATR_VAL:
3281 case OP_ATR_MIN:
3282 case OP_ATR_MAX:
3283 case TERNOP_IN_RANGE:
3284 case BINOP_IN_BOUNDS:
3285 case UNOP_IN_RANGE:
3286 case OP_AGGREGATE:
3287 case OP_OTHERS:
3288 case OP_CHOICES:
3289 case OP_POSITIONAL:
3290 case OP_DISCRETE_RANGE:
3291 case OP_NAME:
3292 ada_forward_operator_length (exp, pc, &oplen, &nargs);
3293 *pos += oplen;
3294 break;
3295
3296 case BINOP_ASSIGN:
3297 {
3298 struct value *arg1;
3299
3300 *pos += 1;
3301 arg1 = resolve_subexp (expp, pos, 0, NULL);
3302 if (arg1 == NULL)
3303 resolve_subexp (expp, pos, 1, NULL);
3304 else
3305 resolve_subexp (expp, pos, 1, value_type (arg1));
3306 break;
3307 }
3308
3309 case UNOP_CAST:
3310 *pos += 3;
3311 nargs = 1;
3312 break;
3313
3314 case BINOP_ADD:
3315 case BINOP_SUB:
3316 case BINOP_MUL:
3317 case BINOP_DIV:
3318 case BINOP_REM:
3319 case BINOP_MOD:
3320 case BINOP_EXP:
3321 case BINOP_CONCAT:
3322 case BINOP_LOGICAL_AND:
3323 case BINOP_LOGICAL_OR:
3324 case BINOP_BITWISE_AND:
3325 case BINOP_BITWISE_IOR:
3326 case BINOP_BITWISE_XOR:
3327
3328 case BINOP_EQUAL:
3329 case BINOP_NOTEQUAL:
3330 case BINOP_LESS:
3331 case BINOP_GTR:
3332 case BINOP_LEQ:
3333 case BINOP_GEQ:
3334
3335 case BINOP_REPEAT:
3336 case BINOP_SUBSCRIPT:
3337 case BINOP_COMMA:
3338 *pos += 1;
3339 nargs = 2;
3340 break;
3341
3342 case UNOP_NEG:
3343 case UNOP_PLUS:
3344 case UNOP_LOGICAL_NOT:
3345 case UNOP_ABS:
3346 case UNOP_IND:
3347 *pos += 1;
3348 nargs = 1;
3349 break;
3350
3351 case OP_LONG:
3352 case OP_DOUBLE:
3353 case OP_VAR_VALUE:
3354 *pos += 4;
3355 break;
3356
3357 case OP_TYPE:
3358 case OP_BOOL:
3359 case OP_LAST:
3360 case OP_INTERNALVAR:
3361 *pos += 3;
3362 break;
3363
3364 case UNOP_MEMVAL:
3365 *pos += 3;
3366 nargs = 1;
3367 break;
3368
3369 case OP_REGISTER:
3370 *pos += 4 + BYTES_TO_EXP_ELEM (exp->elts[pc + 1].longconst + 1);
3371 break;
3372
3373 case STRUCTOP_STRUCT:
3374 *pos += 4 + BYTES_TO_EXP_ELEM (exp->elts[pc + 1].longconst + 1);
3375 nargs = 1;
3376 break;
3377
3378 case TERNOP_SLICE:
3379 *pos += 1;
3380 nargs = 3;
3381 break;
3382
3383 case OP_STRING:
3384 break;
3385
3386 default:
3387 error (_("Unexpected operator during name resolution"));
3388 }
3389
3390 argvec = XALLOCAVEC (struct value *, nargs + 1);
3391 for (i = 0; i < nargs; i += 1)
3392 argvec[i] = resolve_subexp (expp, pos, 1, NULL);
3393 argvec[i] = NULL;
3394 exp = *expp;
3395
3396 /* Pass two: perform any resolution on principal operator. */
3397 switch (op)
3398 {
3399 default:
3400 break;
3401
3402 case OP_VAR_VALUE:
3403 if (SYMBOL_DOMAIN (exp->elts[pc + 2].symbol) == UNDEF_DOMAIN)
3404 {
3405 struct block_symbol *candidates;
3406 int n_candidates;
3407
3408 n_candidates =
3409 ada_lookup_symbol_list (SYMBOL_LINKAGE_NAME
3410 (exp->elts[pc + 2].symbol),
3411 exp->elts[pc + 1].block, VAR_DOMAIN,
3412 &candidates);
3413
3414 if (n_candidates > 1)
3415 {
3416 /* Types tend to get re-introduced locally, so if there
3417 are any local symbols that are not types, first filter
3418 out all types. */
3419 int j;
3420 for (j = 0; j < n_candidates; j += 1)
3421 switch (SYMBOL_CLASS (candidates[j].symbol))
3422 {
3423 case LOC_REGISTER:
3424 case LOC_ARG:
3425 case LOC_REF_ARG:
3426 case LOC_REGPARM_ADDR:
3427 case LOC_LOCAL:
3428 case LOC_COMPUTED:
3429 goto FoundNonType;
3430 default:
3431 break;
3432 }
3433 FoundNonType:
3434 if (j < n_candidates)
3435 {
3436 j = 0;
3437 while (j < n_candidates)
3438 {
3439 if (SYMBOL_CLASS (candidates[j].symbol) == LOC_TYPEDEF)
3440 {
3441 candidates[j] = candidates[n_candidates - 1];
3442 n_candidates -= 1;
3443 }
3444 else
3445 j += 1;
3446 }
3447 }
3448 }
3449
3450 if (n_candidates == 0)
3451 error (_("No definition found for %s"),
3452 SYMBOL_PRINT_NAME (exp->elts[pc + 2].symbol));
3453 else if (n_candidates == 1)
3454 i = 0;
3455 else if (deprocedure_p
3456 && !is_nonfunction (candidates, n_candidates))
3457 {
3458 i = ada_resolve_function
3459 (candidates, n_candidates, NULL, 0,
3460 SYMBOL_LINKAGE_NAME (exp->elts[pc + 2].symbol),
3461 context_type);
3462 if (i < 0)
3463 error (_("Could not find a match for %s"),
3464 SYMBOL_PRINT_NAME (exp->elts[pc + 2].symbol));
3465 }
3466 else
3467 {
3468 printf_filtered (_("Multiple matches for %s\n"),
3469 SYMBOL_PRINT_NAME (exp->elts[pc + 2].symbol));
3470 user_select_syms (candidates, n_candidates, 1);
3471 i = 0;
3472 }
3473
3474 exp->elts[pc + 1].block = candidates[i].block;
3475 exp->elts[pc + 2].symbol = candidates[i].symbol;
3476 if (innermost_block == NULL
3477 || contained_in (candidates[i].block, innermost_block))
3478 innermost_block = candidates[i].block;
3479 }
3480
3481 if (deprocedure_p
3482 && (TYPE_CODE (SYMBOL_TYPE (exp->elts[pc + 2].symbol))
3483 == TYPE_CODE_FUNC))
3484 {
3485 replace_operator_with_call (expp, pc, 0, 0,
3486 exp->elts[pc + 2].symbol,
3487 exp->elts[pc + 1].block);
3488 exp = *expp;
3489 }
3490 break;
3491
3492 case OP_FUNCALL:
3493 {
3494 if (exp->elts[pc + 3].opcode == OP_VAR_VALUE
3495 && SYMBOL_DOMAIN (exp->elts[pc + 5].symbol) == UNDEF_DOMAIN)
3496 {
3497 struct block_symbol *candidates;
3498 int n_candidates;
3499
3500 n_candidates =
3501 ada_lookup_symbol_list (SYMBOL_LINKAGE_NAME
3502 (exp->elts[pc + 5].symbol),
3503 exp->elts[pc + 4].block, VAR_DOMAIN,
3504 &candidates);
3505 if (n_candidates == 1)
3506 i = 0;
3507 else
3508 {
3509 i = ada_resolve_function
3510 (candidates, n_candidates,
3511 argvec, nargs,
3512 SYMBOL_LINKAGE_NAME (exp->elts[pc + 5].symbol),
3513 context_type);
3514 if (i < 0)
3515 error (_("Could not find a match for %s"),
3516 SYMBOL_PRINT_NAME (exp->elts[pc + 5].symbol));
3517 }
3518
3519 exp->elts[pc + 4].block = candidates[i].block;
3520 exp->elts[pc + 5].symbol = candidates[i].symbol;
3521 if (innermost_block == NULL
3522 || contained_in (candidates[i].block, innermost_block))
3523 innermost_block = candidates[i].block;
3524 }
3525 }
3526 break;
3527 case BINOP_ADD:
3528 case BINOP_SUB:
3529 case BINOP_MUL:
3530 case BINOP_DIV:
3531 case BINOP_REM:
3532 case BINOP_MOD:
3533 case BINOP_CONCAT:
3534 case BINOP_BITWISE_AND:
3535 case BINOP_BITWISE_IOR:
3536 case BINOP_BITWISE_XOR:
3537 case BINOP_EQUAL:
3538 case BINOP_NOTEQUAL:
3539 case BINOP_LESS:
3540 case BINOP_GTR:
3541 case BINOP_LEQ:
3542 case BINOP_GEQ:
3543 case BINOP_EXP:
3544 case UNOP_NEG:
3545 case UNOP_PLUS:
3546 case UNOP_LOGICAL_NOT:
3547 case UNOP_ABS:
3548 if (possible_user_operator_p (op, argvec))
3549 {
3550 struct block_symbol *candidates;
3551 int n_candidates;
3552
3553 n_candidates =
3554 ada_lookup_symbol_list (ada_encode (ada_decoded_op_name (op)),
3555 (struct block *) NULL, VAR_DOMAIN,
3556 &candidates);
3557 i = ada_resolve_function (candidates, n_candidates, argvec, nargs,
3558 ada_decoded_op_name (op), NULL);
3559 if (i < 0)
3560 break;
3561
3562 replace_operator_with_call (expp, pc, nargs, 1,
3563 candidates[i].symbol,
3564 candidates[i].block);
3565 exp = *expp;
3566 }
3567 break;
3568
3569 case OP_TYPE:
3570 case OP_REGISTER:
3571 return NULL;
3572 }
3573
3574 *pos = pc;
3575 return evaluate_subexp_type (exp, pos);
3576 }
3577
3578 /* Return non-zero if formal type FTYPE matches actual type ATYPE. If
3579 MAY_DEREF is non-zero, the formal may be a pointer and the actual
3580 a non-pointer. */
3581 /* The term "match" here is rather loose. The match is heuristic and
3582 liberal. */
3583
3584 static int
3585 ada_type_match (struct type *ftype, struct type *atype, int may_deref)
3586 {
3587 ftype = ada_check_typedef (ftype);
3588 atype = ada_check_typedef (atype);
3589
3590 if (TYPE_CODE (ftype) == TYPE_CODE_REF)
3591 ftype = TYPE_TARGET_TYPE (ftype);
3592 if (TYPE_CODE (atype) == TYPE_CODE_REF)
3593 atype = TYPE_TARGET_TYPE (atype);
3594
3595 switch (TYPE_CODE (ftype))
3596 {
3597 default:
3598 return TYPE_CODE (ftype) == TYPE_CODE (atype);
3599 case TYPE_CODE_PTR:
3600 if (TYPE_CODE (atype) == TYPE_CODE_PTR)
3601 return ada_type_match (TYPE_TARGET_TYPE (ftype),
3602 TYPE_TARGET_TYPE (atype), 0);
3603 else
3604 return (may_deref
3605 && ada_type_match (TYPE_TARGET_TYPE (ftype), atype, 0));
3606 case TYPE_CODE_INT:
3607 case TYPE_CODE_ENUM:
3608 case TYPE_CODE_RANGE:
3609 switch (TYPE_CODE (atype))
3610 {
3611 case TYPE_CODE_INT:
3612 case TYPE_CODE_ENUM:
3613 case TYPE_CODE_RANGE:
3614 return 1;
3615 default:
3616 return 0;
3617 }
3618
3619 case TYPE_CODE_ARRAY:
3620 return (TYPE_CODE (atype) == TYPE_CODE_ARRAY
3621 || ada_is_array_descriptor_type (atype));
3622
3623 case TYPE_CODE_STRUCT:
3624 if (ada_is_array_descriptor_type (ftype))
3625 return (TYPE_CODE (atype) == TYPE_CODE_ARRAY
3626 || ada_is_array_descriptor_type (atype));
3627 else
3628 return (TYPE_CODE (atype) == TYPE_CODE_STRUCT
3629 && !ada_is_array_descriptor_type (atype));
3630
3631 case TYPE_CODE_UNION:
3632 case TYPE_CODE_FLT:
3633 return (TYPE_CODE (atype) == TYPE_CODE (ftype));
3634 }
3635 }
3636
3637 /* Return non-zero if the formals of FUNC "sufficiently match" the
3638 vector of actual argument types ACTUALS of size N_ACTUALS. FUNC
3639 may also be an enumeral, in which case it is treated as a 0-
3640 argument function. */
3641
3642 static int
3643 ada_args_match (struct symbol *func, struct value **actuals, int n_actuals)
3644 {
3645 int i;
3646 struct type *func_type = SYMBOL_TYPE (func);
3647
3648 if (SYMBOL_CLASS (func) == LOC_CONST
3649 && TYPE_CODE (func_type) == TYPE_CODE_ENUM)
3650 return (n_actuals == 0);
3651 else if (func_type == NULL || TYPE_CODE (func_type) != TYPE_CODE_FUNC)
3652 return 0;
3653
3654 if (TYPE_NFIELDS (func_type) != n_actuals)
3655 return 0;
3656
3657 for (i = 0; i < n_actuals; i += 1)
3658 {
3659 if (actuals[i] == NULL)
3660 return 0;
3661 else
3662 {
3663 struct type *ftype = ada_check_typedef (TYPE_FIELD_TYPE (func_type,
3664 i));
3665 struct type *atype = ada_check_typedef (value_type (actuals[i]));
3666
3667 if (!ada_type_match (ftype, atype, 1))
3668 return 0;
3669 }
3670 }
3671 return 1;
3672 }
3673
3674 /* False iff function type FUNC_TYPE definitely does not produce a value
3675 compatible with type CONTEXT_TYPE. Conservatively returns 1 if
3676 FUNC_TYPE is not a valid function type with a non-null return type
3677 or an enumerated type. A null CONTEXT_TYPE indicates any non-void type. */
3678
3679 static int
3680 return_match (struct type *func_type, struct type *context_type)
3681 {
3682 struct type *return_type;
3683
3684 if (func_type == NULL)
3685 return 1;
3686
3687 if (TYPE_CODE (func_type) == TYPE_CODE_FUNC)
3688 return_type = get_base_type (TYPE_TARGET_TYPE (func_type));
3689 else
3690 return_type = get_base_type (func_type);
3691 if (return_type == NULL)
3692 return 1;
3693
3694 context_type = get_base_type (context_type);
3695
3696 if (TYPE_CODE (return_type) == TYPE_CODE_ENUM)
3697 return context_type == NULL || return_type == context_type;
3698 else if (context_type == NULL)
3699 return TYPE_CODE (return_type) != TYPE_CODE_VOID;
3700 else
3701 return TYPE_CODE (return_type) == TYPE_CODE (context_type);
3702 }
3703
3704
3705 /* Returns the index in SYMS[0..NSYMS-1] that contains the symbol for the
3706 function (if any) that matches the types of the NARGS arguments in
3707 ARGS. If CONTEXT_TYPE is non-null and there is at least one match
3708 that returns that type, then eliminate matches that don't. If
3709 CONTEXT_TYPE is void and there is at least one match that does not
3710 return void, eliminate all matches that do.
3711
3712 Asks the user if there is more than one match remaining. Returns -1
3713 if there is no such symbol or none is selected. NAME is used
3714 solely for messages. May re-arrange and modify SYMS in
3715 the process; the index returned is for the modified vector. */
3716
3717 static int
3718 ada_resolve_function (struct block_symbol syms[],
3719 int nsyms, struct value **args, int nargs,
3720 const char *name, struct type *context_type)
3721 {
3722 int fallback;
3723 int k;
3724 int m; /* Number of hits */
3725
3726 m = 0;
3727 /* In the first pass of the loop, we only accept functions matching
3728 context_type. If none are found, we add a second pass of the loop
3729 where every function is accepted. */
3730 for (fallback = 0; m == 0 && fallback < 2; fallback++)
3731 {
3732 for (k = 0; k < nsyms; k += 1)
3733 {
3734 struct type *type = ada_check_typedef (SYMBOL_TYPE (syms[k].symbol));
3735
3736 if (ada_args_match (syms[k].symbol, args, nargs)
3737 && (fallback || return_match (type, context_type)))
3738 {
3739 syms[m] = syms[k];
3740 m += 1;
3741 }
3742 }
3743 }
3744
3745 /* If we got multiple matches, ask the user which one to use. Don't do this
3746 interactive thing during completion, though, as the purpose of the
3747 completion is providing a list of all possible matches. Prompting the
3748 user to filter it down would be completely unexpected in this case. */
3749 if (m == 0)
3750 return -1;
3751 else if (m > 1 && !parse_completion)
3752 {
3753 printf_filtered (_("Multiple matches for %s\n"), name);
3754 user_select_syms (syms, m, 1);
3755 return 0;
3756 }
3757 return 0;
3758 }
3759
3760 /* Returns true (non-zero) iff decoded name N0 should appear before N1
3761 in a listing of choices during disambiguation (see sort_choices, below).
3762 The idea is that overloadings of a subprogram name from the
3763 same package should sort in their source order. We settle for ordering
3764 such symbols by their trailing number (__N or $N). */
3765
3766 static int
3767 encoded_ordered_before (const char *N0, const char *N1)
3768 {
3769 if (N1 == NULL)
3770 return 0;
3771 else if (N0 == NULL)
3772 return 1;
3773 else
3774 {
3775 int k0, k1;
3776
3777 for (k0 = strlen (N0) - 1; k0 > 0 && isdigit (N0[k0]); k0 -= 1)
3778 ;
3779 for (k1 = strlen (N1) - 1; k1 > 0 && isdigit (N1[k1]); k1 -= 1)
3780 ;
3781 if ((N0[k0] == '_' || N0[k0] == '$') && N0[k0 + 1] != '\000'
3782 && (N1[k1] == '_' || N1[k1] == '$') && N1[k1 + 1] != '\000')
3783 {
3784 int n0, n1;
3785
3786 n0 = k0;
3787 while (N0[n0] == '_' && n0 > 0 && N0[n0 - 1] == '_')
3788 n0 -= 1;
3789 n1 = k1;
3790 while (N1[n1] == '_' && n1 > 0 && N1[n1 - 1] == '_')
3791 n1 -= 1;
3792 if (n0 == n1 && strncmp (N0, N1, n0) == 0)
3793 return (atoi (N0 + k0 + 1) < atoi (N1 + k1 + 1));
3794 }
3795 return (strcmp (N0, N1) < 0);
3796 }
3797 }
3798
3799 /* Sort SYMS[0..NSYMS-1] to put the choices in a canonical order by the
3800 encoded names. */
3801
3802 static void
3803 sort_choices (struct block_symbol syms[], int nsyms)
3804 {
3805 int i;
3806
3807 for (i = 1; i < nsyms; i += 1)
3808 {
3809 struct block_symbol sym = syms[i];
3810 int j;
3811
3812 for (j = i - 1; j >= 0; j -= 1)
3813 {
3814 if (encoded_ordered_before (SYMBOL_LINKAGE_NAME (syms[j].symbol),
3815 SYMBOL_LINKAGE_NAME (sym.symbol)))
3816 break;
3817 syms[j + 1] = syms[j];
3818 }
3819 syms[j + 1] = sym;
3820 }
3821 }
3822
3823 /* Given a list of NSYMS symbols in SYMS, select up to MAX_RESULTS>0
3824 by asking the user (if necessary), returning the number selected,
3825 and setting the first elements of SYMS items. Error if no symbols
3826 selected. */
3827
3828 /* NOTE: Adapted from decode_line_2 in symtab.c, with which it ought
3829 to be re-integrated one of these days. */
3830
3831 int
3832 user_select_syms (struct block_symbol *syms, int nsyms, int max_results)
3833 {
3834 int i;
3835 int *chosen = XALLOCAVEC (int , nsyms);
3836 int n_chosen;
3837 int first_choice = (max_results == 1) ? 1 : 2;
3838 const char *select_mode = multiple_symbols_select_mode ();
3839
3840 if (max_results < 1)
3841 error (_("Request to select 0 symbols!"));
3842 if (nsyms <= 1)
3843 return nsyms;
3844
3845 if (select_mode == multiple_symbols_cancel)
3846 error (_("\
3847 canceled because the command is ambiguous\n\
3848 See set/show multiple-symbol."));
3849
3850 /* If select_mode is "all", then return all possible symbols.
3851 Only do that if more than one symbol can be selected, of course.
3852 Otherwise, display the menu as usual. */
3853 if (select_mode == multiple_symbols_all && max_results > 1)
3854 return nsyms;
3855
3856 printf_unfiltered (_("[0] cancel\n"));
3857 if (max_results > 1)
3858 printf_unfiltered (_("[1] all\n"));
3859
3860 sort_choices (syms, nsyms);
3861
3862 for (i = 0; i < nsyms; i += 1)
3863 {
3864 if (syms[i].symbol == NULL)
3865 continue;
3866
3867 if (SYMBOL_CLASS (syms[i].symbol) == LOC_BLOCK)
3868 {
3869 struct symtab_and_line sal =
3870 find_function_start_sal (syms[i].symbol, 1);
3871
3872 if (sal.symtab == NULL)
3873 printf_unfiltered (_("[%d] %s at <no source file available>:%d\n"),
3874 i + first_choice,
3875 SYMBOL_PRINT_NAME (syms[i].symbol),
3876 sal.line);
3877 else
3878 printf_unfiltered (_("[%d] %s at %s:%d\n"), i + first_choice,
3879 SYMBOL_PRINT_NAME (syms[i].symbol),
3880 symtab_to_filename_for_display (sal.symtab),
3881 sal.line);
3882 continue;
3883 }
3884 else
3885 {
3886 int is_enumeral =
3887 (SYMBOL_CLASS (syms[i].symbol) == LOC_CONST
3888 && SYMBOL_TYPE (syms[i].symbol) != NULL
3889 && TYPE_CODE (SYMBOL_TYPE (syms[i].symbol)) == TYPE_CODE_ENUM);
3890 struct symtab *symtab = NULL;
3891
3892 if (SYMBOL_OBJFILE_OWNED (syms[i].symbol))
3893 symtab = symbol_symtab (syms[i].symbol);
3894
3895 if (SYMBOL_LINE (syms[i].symbol) != 0 && symtab != NULL)
3896 printf_unfiltered (_("[%d] %s at %s:%d\n"),
3897 i + first_choice,
3898 SYMBOL_PRINT_NAME (syms[i].symbol),
3899 symtab_to_filename_for_display (symtab),
3900 SYMBOL_LINE (syms[i].symbol));
3901 else if (is_enumeral
3902 && TYPE_NAME (SYMBOL_TYPE (syms[i].symbol)) != NULL)
3903 {
3904 printf_unfiltered (("[%d] "), i + first_choice);
3905 ada_print_type (SYMBOL_TYPE (syms[i].symbol), NULL,
3906 gdb_stdout, -1, 0, &type_print_raw_options);
3907 printf_unfiltered (_("'(%s) (enumeral)\n"),
3908 SYMBOL_PRINT_NAME (syms[i].symbol));
3909 }
3910 else if (symtab != NULL)
3911 printf_unfiltered (is_enumeral
3912 ? _("[%d] %s in %s (enumeral)\n")
3913 : _("[%d] %s at %s:?\n"),
3914 i + first_choice,
3915 SYMBOL_PRINT_NAME (syms[i].symbol),
3916 symtab_to_filename_for_display (symtab));
3917 else
3918 printf_unfiltered (is_enumeral
3919 ? _("[%d] %s (enumeral)\n")
3920 : _("[%d] %s at ?\n"),
3921 i + first_choice,
3922 SYMBOL_PRINT_NAME (syms[i].symbol));
3923 }
3924 }
3925
3926 n_chosen = get_selections (chosen, nsyms, max_results, max_results > 1,
3927 "overload-choice");
3928
3929 for (i = 0; i < n_chosen; i += 1)
3930 syms[i] = syms[chosen[i]];
3931
3932 return n_chosen;
3933 }
3934
3935 /* Read and validate a set of numeric choices from the user in the
3936 range 0 .. N_CHOICES-1. Place the results in increasing
3937 order in CHOICES[0 .. N-1], and return N.
3938
3939 The user types choices as a sequence of numbers on one line
3940 separated by blanks, encoding them as follows:
3941
3942 + A choice of 0 means to cancel the selection, throwing an error.
3943 + If IS_ALL_CHOICE, a choice of 1 selects the entire set 0 .. N_CHOICES-1.
3944 + The user chooses k by typing k+IS_ALL_CHOICE+1.
3945
3946 The user is not allowed to choose more than MAX_RESULTS values.
3947
3948 ANNOTATION_SUFFIX, if present, is used to annotate the input
3949 prompts (for use with the -f switch). */
3950
3951 int
3952 get_selections (int *choices, int n_choices, int max_results,
3953 int is_all_choice, char *annotation_suffix)
3954 {
3955 char *args;
3956 char *prompt;
3957 int n_chosen;
3958 int first_choice = is_all_choice ? 2 : 1;
3959
3960 prompt = getenv ("PS2");
3961 if (prompt == NULL)
3962 prompt = "> ";
3963
3964 args = command_line_input (prompt, 0, annotation_suffix);
3965
3966 if (args == NULL)
3967 error_no_arg (_("one or more choice numbers"));
3968
3969 n_chosen = 0;
3970
3971 /* Set choices[0 .. n_chosen-1] to the users' choices in ascending
3972 order, as given in args. Choices are validated. */
3973 while (1)
3974 {
3975 char *args2;
3976 int choice, j;
3977
3978 args = skip_spaces (args);
3979 if (*args == '\0' && n_chosen == 0)
3980 error_no_arg (_("one or more choice numbers"));
3981 else if (*args == '\0')
3982 break;
3983
3984 choice = strtol (args, &args2, 10);
3985 if (args == args2 || choice < 0
3986 || choice > n_choices + first_choice - 1)
3987 error (_("Argument must be choice number"));
3988 args = args2;
3989
3990 if (choice == 0)
3991 error (_("cancelled"));
3992
3993 if (choice < first_choice)
3994 {
3995 n_chosen = n_choices;
3996 for (j = 0; j < n_choices; j += 1)
3997 choices[j] = j;
3998 break;
3999 }
4000 choice -= first_choice;
4001
4002 for (j = n_chosen - 1; j >= 0 && choice < choices[j]; j -= 1)
4003 {
4004 }
4005
4006 if (j < 0 || choice != choices[j])
4007 {
4008 int k;
4009
4010 for (k = n_chosen - 1; k > j; k -= 1)
4011 choices[k + 1] = choices[k];
4012 choices[j + 1] = choice;
4013 n_chosen += 1;
4014 }
4015 }
4016
4017 if (n_chosen > max_results)
4018 error (_("Select no more than %d of the above"), max_results);
4019
4020 return n_chosen;
4021 }
4022
4023 /* Replace the operator of length OPLEN at position PC in *EXPP with a call
4024 on the function identified by SYM and BLOCK, and taking NARGS
4025 arguments. Update *EXPP as needed to hold more space. */
4026
4027 static void
4028 replace_operator_with_call (struct expression **expp, int pc, int nargs,
4029 int oplen, struct symbol *sym,
4030 const struct block *block)
4031 {
4032 /* A new expression, with 6 more elements (3 for funcall, 4 for function
4033 symbol, -oplen for operator being replaced). */
4034 struct expression *newexp = (struct expression *)
4035 xzalloc (sizeof (struct expression)
4036 + EXP_ELEM_TO_BYTES ((*expp)->nelts + 7 - oplen));
4037 struct expression *exp = *expp;
4038
4039 newexp->nelts = exp->nelts + 7 - oplen;
4040 newexp->language_defn = exp->language_defn;
4041 newexp->gdbarch = exp->gdbarch;
4042 memcpy (newexp->elts, exp->elts, EXP_ELEM_TO_BYTES (pc));
4043 memcpy (newexp->elts + pc + 7, exp->elts + pc + oplen,
4044 EXP_ELEM_TO_BYTES (exp->nelts - pc - oplen));
4045
4046 newexp->elts[pc].opcode = newexp->elts[pc + 2].opcode = OP_FUNCALL;
4047 newexp->elts[pc + 1].longconst = (LONGEST) nargs;
4048
4049 newexp->elts[pc + 3].opcode = newexp->elts[pc + 6].opcode = OP_VAR_VALUE;
4050 newexp->elts[pc + 4].block = block;
4051 newexp->elts[pc + 5].symbol = sym;
4052
4053 *expp = newexp;
4054 xfree (exp);
4055 }
4056
4057 /* Type-class predicates */
4058
4059 /* True iff TYPE is numeric (i.e., an INT, RANGE (of numeric type),
4060 or FLOAT). */
4061
4062 static int
4063 numeric_type_p (struct type *type)
4064 {
4065 if (type == NULL)
4066 return 0;
4067 else
4068 {
4069 switch (TYPE_CODE (type))
4070 {
4071 case TYPE_CODE_INT:
4072 case TYPE_CODE_FLT:
4073 return 1;
4074 case TYPE_CODE_RANGE:
4075 return (type == TYPE_TARGET_TYPE (type)
4076 || numeric_type_p (TYPE_TARGET_TYPE (type)));
4077 default:
4078 return 0;
4079 }
4080 }
4081 }
4082
4083 /* True iff TYPE is integral (an INT or RANGE of INTs). */
4084
4085 static int
4086 integer_type_p (struct type *type)
4087 {
4088 if (type == NULL)
4089 return 0;
4090 else
4091 {
4092 switch (TYPE_CODE (type))
4093 {
4094 case TYPE_CODE_INT:
4095 return 1;
4096 case TYPE_CODE_RANGE:
4097 return (type == TYPE_TARGET_TYPE (type)
4098 || integer_type_p (TYPE_TARGET_TYPE (type)));
4099 default:
4100 return 0;
4101 }
4102 }
4103 }
4104
4105 /* True iff TYPE is scalar (INT, RANGE, FLOAT, ENUM). */
4106
4107 static int
4108 scalar_type_p (struct type *type)
4109 {
4110 if (type == NULL)
4111 return 0;
4112 else
4113 {
4114 switch (TYPE_CODE (type))
4115 {
4116 case TYPE_CODE_INT:
4117 case TYPE_CODE_RANGE:
4118 case TYPE_CODE_ENUM:
4119 case TYPE_CODE_FLT:
4120 return 1;
4121 default:
4122 return 0;
4123 }
4124 }
4125 }
4126
4127 /* True iff TYPE is discrete (INT, RANGE, ENUM). */
4128
4129 static int
4130 discrete_type_p (struct type *type)
4131 {
4132 if (type == NULL)
4133 return 0;
4134 else
4135 {
4136 switch (TYPE_CODE (type))
4137 {
4138 case TYPE_CODE_INT:
4139 case TYPE_CODE_RANGE:
4140 case TYPE_CODE_ENUM:
4141 case TYPE_CODE_BOOL:
4142 return 1;
4143 default:
4144 return 0;
4145 }
4146 }
4147 }
4148
4149 /* Returns non-zero if OP with operands in the vector ARGS could be
4150 a user-defined function. Errs on the side of pre-defined operators
4151 (i.e., result 0). */
4152
4153 static int
4154 possible_user_operator_p (enum exp_opcode op, struct value *args[])
4155 {
4156 struct type *type0 =
4157 (args[0] == NULL) ? NULL : ada_check_typedef (value_type (args[0]));
4158 struct type *type1 =
4159 (args[1] == NULL) ? NULL : ada_check_typedef (value_type (args[1]));
4160
4161 if (type0 == NULL)
4162 return 0;
4163
4164 switch (op)
4165 {
4166 default:
4167 return 0;
4168
4169 case BINOP_ADD:
4170 case BINOP_SUB:
4171 case BINOP_MUL:
4172 case BINOP_DIV:
4173 return (!(numeric_type_p (type0) && numeric_type_p (type1)));
4174
4175 case BINOP_REM:
4176 case BINOP_MOD:
4177 case BINOP_BITWISE_AND:
4178 case BINOP_BITWISE_IOR:
4179 case BINOP_BITWISE_XOR:
4180 return (!(integer_type_p (type0) && integer_type_p (type1)));
4181
4182 case BINOP_EQUAL:
4183 case BINOP_NOTEQUAL:
4184 case BINOP_LESS:
4185 case BINOP_GTR:
4186 case BINOP_LEQ:
4187 case BINOP_GEQ:
4188 return (!(scalar_type_p (type0) && scalar_type_p (type1)));
4189
4190 case BINOP_CONCAT:
4191 return !ada_is_array_type (type0) || !ada_is_array_type (type1);
4192
4193 case BINOP_EXP:
4194 return (!(numeric_type_p (type0) && integer_type_p (type1)));
4195
4196 case UNOP_NEG:
4197 case UNOP_PLUS:
4198 case UNOP_LOGICAL_NOT:
4199 case UNOP_ABS:
4200 return (!numeric_type_p (type0));
4201
4202 }
4203 }
4204 \f
4205 /* Renaming */
4206
4207 /* NOTES:
4208
4209 1. In the following, we assume that a renaming type's name may
4210 have an ___XD suffix. It would be nice if this went away at some
4211 point.
4212 2. We handle both the (old) purely type-based representation of
4213 renamings and the (new) variable-based encoding. At some point,
4214 it is devoutly to be hoped that the former goes away
4215 (FIXME: hilfinger-2007-07-09).
4216 3. Subprogram renamings are not implemented, although the XRS
4217 suffix is recognized (FIXME: hilfinger-2007-07-09). */
4218
4219 /* If SYM encodes a renaming,
4220
4221 <renaming> renames <renamed entity>,
4222
4223 sets *LEN to the length of the renamed entity's name,
4224 *RENAMED_ENTITY to that name (not null-terminated), and *RENAMING_EXPR to
4225 the string describing the subcomponent selected from the renamed
4226 entity. Returns ADA_NOT_RENAMING if SYM does not encode a renaming
4227 (in which case, the values of *RENAMED_ENTITY, *LEN, and *RENAMING_EXPR
4228 are undefined). Otherwise, returns a value indicating the category
4229 of entity renamed: an object (ADA_OBJECT_RENAMING), exception
4230 (ADA_EXCEPTION_RENAMING), package (ADA_PACKAGE_RENAMING), or
4231 subprogram (ADA_SUBPROGRAM_RENAMING). Does no allocation; the
4232 strings returned in *RENAMED_ENTITY and *RENAMING_EXPR should not be
4233 deallocated. The values of RENAMED_ENTITY, LEN, or RENAMING_EXPR
4234 may be NULL, in which case they are not assigned.
4235
4236 [Currently, however, GCC does not generate subprogram renamings.] */
4237
4238 enum ada_renaming_category
4239 ada_parse_renaming (struct symbol *sym,
4240 const char **renamed_entity, int *len,
4241 const char **renaming_expr)
4242 {
4243 enum ada_renaming_category kind;
4244 const char *info;
4245 const char *suffix;
4246
4247 if (sym == NULL)
4248 return ADA_NOT_RENAMING;
4249 switch (SYMBOL_CLASS (sym))
4250 {
4251 default:
4252 return ADA_NOT_RENAMING;
4253 case LOC_TYPEDEF:
4254 return parse_old_style_renaming (SYMBOL_TYPE (sym),
4255 renamed_entity, len, renaming_expr);
4256 case LOC_LOCAL:
4257 case LOC_STATIC:
4258 case LOC_COMPUTED:
4259 case LOC_OPTIMIZED_OUT:
4260 info = strstr (SYMBOL_LINKAGE_NAME (sym), "___XR");
4261 if (info == NULL)
4262 return ADA_NOT_RENAMING;
4263 switch (info[5])
4264 {
4265 case '_':
4266 kind = ADA_OBJECT_RENAMING;
4267 info += 6;
4268 break;
4269 case 'E':
4270 kind = ADA_EXCEPTION_RENAMING;
4271 info += 7;
4272 break;
4273 case 'P':
4274 kind = ADA_PACKAGE_RENAMING;
4275 info += 7;
4276 break;
4277 case 'S':
4278 kind = ADA_SUBPROGRAM_RENAMING;
4279 info += 7;
4280 break;
4281 default:
4282 return ADA_NOT_RENAMING;
4283 }
4284 }
4285
4286 if (renamed_entity != NULL)
4287 *renamed_entity = info;
4288 suffix = strstr (info, "___XE");
4289 if (suffix == NULL || suffix == info)
4290 return ADA_NOT_RENAMING;
4291 if (len != NULL)
4292 *len = strlen (info) - strlen (suffix);
4293 suffix += 5;
4294 if (renaming_expr != NULL)
4295 *renaming_expr = suffix;
4296 return kind;
4297 }
4298
4299 /* Assuming TYPE encodes a renaming according to the old encoding in
4300 exp_dbug.ads, returns details of that renaming in *RENAMED_ENTITY,
4301 *LEN, and *RENAMING_EXPR, as for ada_parse_renaming, above. Returns
4302 ADA_NOT_RENAMING otherwise. */
4303 static enum ada_renaming_category
4304 parse_old_style_renaming (struct type *type,
4305 const char **renamed_entity, int *len,
4306 const char **renaming_expr)
4307 {
4308 enum ada_renaming_category kind;
4309 const char *name;
4310 const char *info;
4311 const char *suffix;
4312
4313 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_ENUM
4314 || TYPE_NFIELDS (type) != 1)
4315 return ADA_NOT_RENAMING;
4316
4317 name = type_name_no_tag (type);
4318 if (name == NULL)
4319 return ADA_NOT_RENAMING;
4320
4321 name = strstr (name, "___XR");
4322 if (name == NULL)
4323 return ADA_NOT_RENAMING;
4324 switch (name[5])
4325 {
4326 case '\0':
4327 case '_':
4328 kind = ADA_OBJECT_RENAMING;
4329 break;
4330 case 'E':
4331 kind = ADA_EXCEPTION_RENAMING;
4332 break;
4333 case 'P':
4334 kind = ADA_PACKAGE_RENAMING;
4335 break;
4336 case 'S':
4337 kind = ADA_SUBPROGRAM_RENAMING;
4338 break;
4339 default:
4340 return ADA_NOT_RENAMING;
4341 }
4342
4343 info = TYPE_FIELD_NAME (type, 0);
4344 if (info == NULL)
4345 return ADA_NOT_RENAMING;
4346 if (renamed_entity != NULL)
4347 *renamed_entity = info;
4348 suffix = strstr (info, "___XE");
4349 if (renaming_expr != NULL)
4350 *renaming_expr = suffix + 5;
4351 if (suffix == NULL || suffix == info)
4352 return ADA_NOT_RENAMING;
4353 if (len != NULL)
4354 *len = suffix - info;
4355 return kind;
4356 }
4357
4358 /* Compute the value of the given RENAMING_SYM, which is expected to
4359 be a symbol encoding a renaming expression. BLOCK is the block
4360 used to evaluate the renaming. */
4361
4362 static struct value *
4363 ada_read_renaming_var_value (struct symbol *renaming_sym,
4364 const struct block *block)
4365 {
4366 const char *sym_name;
4367 struct expression *expr;
4368 struct value *value;
4369 struct cleanup *old_chain = NULL;
4370
4371 sym_name = SYMBOL_LINKAGE_NAME (renaming_sym);
4372 expr = parse_exp_1 (&sym_name, 0, block, 0);
4373 old_chain = make_cleanup (free_current_contents, &expr);
4374 value = evaluate_expression (expr);
4375
4376 do_cleanups (old_chain);
4377 return value;
4378 }
4379 \f
4380
4381 /* Evaluation: Function Calls */
4382
4383 /* Return an lvalue containing the value VAL. This is the identity on
4384 lvalues, and otherwise has the side-effect of allocating memory
4385 in the inferior where a copy of the value contents is copied. */
4386
4387 static struct value *
4388 ensure_lval (struct value *val)
4389 {
4390 if (VALUE_LVAL (val) == not_lval
4391 || VALUE_LVAL (val) == lval_internalvar)
4392 {
4393 int len = TYPE_LENGTH (ada_check_typedef (value_type (val)));
4394 const CORE_ADDR addr =
4395 value_as_long (value_allocate_space_in_inferior (len));
4396
4397 set_value_address (val, addr);
4398 VALUE_LVAL (val) = lval_memory;
4399 write_memory (addr, value_contents (val), len);
4400 }
4401
4402 return val;
4403 }
4404
4405 /* Return the value ACTUAL, converted to be an appropriate value for a
4406 formal of type FORMAL_TYPE. Use *SP as a stack pointer for
4407 allocating any necessary descriptors (fat pointers), or copies of
4408 values not residing in memory, updating it as needed. */
4409
4410 struct value *
4411 ada_convert_actual (struct value *actual, struct type *formal_type0)
4412 {
4413 struct type *actual_type = ada_check_typedef (value_type (actual));
4414 struct type *formal_type = ada_check_typedef (formal_type0);
4415 struct type *formal_target =
4416 TYPE_CODE (formal_type) == TYPE_CODE_PTR
4417 ? ada_check_typedef (TYPE_TARGET_TYPE (formal_type)) : formal_type;
4418 struct type *actual_target =
4419 TYPE_CODE (actual_type) == TYPE_CODE_PTR
4420 ? ada_check_typedef (TYPE_TARGET_TYPE (actual_type)) : actual_type;
4421
4422 if (ada_is_array_descriptor_type (formal_target)
4423 && TYPE_CODE (actual_target) == TYPE_CODE_ARRAY)
4424 return make_array_descriptor (formal_type, actual);
4425 else if (TYPE_CODE (formal_type) == TYPE_CODE_PTR
4426 || TYPE_CODE (formal_type) == TYPE_CODE_REF)
4427 {
4428 struct value *result;
4429
4430 if (TYPE_CODE (formal_target) == TYPE_CODE_ARRAY
4431 && ada_is_array_descriptor_type (actual_target))
4432 result = desc_data (actual);
4433 else if (TYPE_CODE (actual_type) != TYPE_CODE_PTR)
4434 {
4435 if (VALUE_LVAL (actual) != lval_memory)
4436 {
4437 struct value *val;
4438
4439 actual_type = ada_check_typedef (value_type (actual));
4440 val = allocate_value (actual_type);
4441 memcpy ((char *) value_contents_raw (val),
4442 (char *) value_contents (actual),
4443 TYPE_LENGTH (actual_type));
4444 actual = ensure_lval (val);
4445 }
4446 result = value_addr (actual);
4447 }
4448 else
4449 return actual;
4450 return value_cast_pointers (formal_type, result, 0);
4451 }
4452 else if (TYPE_CODE (actual_type) == TYPE_CODE_PTR)
4453 return ada_value_ind (actual);
4454 else if (ada_is_aligner_type (formal_type))
4455 {
4456 /* We need to turn this parameter into an aligner type
4457 as well. */
4458 struct value *aligner = allocate_value (formal_type);
4459 struct value *component = ada_value_struct_elt (aligner, "F", 0);
4460
4461 value_assign_to_component (aligner, component, actual);
4462 return aligner;
4463 }
4464
4465 return actual;
4466 }
4467
4468 /* Convert VALUE (which must be an address) to a CORE_ADDR that is a pointer of
4469 type TYPE. This is usually an inefficient no-op except on some targets
4470 (such as AVR) where the representation of a pointer and an address
4471 differs. */
4472
4473 static CORE_ADDR
4474 value_pointer (struct value *value, struct type *type)
4475 {
4476 struct gdbarch *gdbarch = get_type_arch (type);
4477 unsigned len = TYPE_LENGTH (type);
4478 gdb_byte *buf = (gdb_byte *) alloca (len);
4479 CORE_ADDR addr;
4480
4481 addr = value_address (value);
4482 gdbarch_address_to_pointer (gdbarch, type, buf, addr);
4483 addr = extract_unsigned_integer (buf, len, gdbarch_byte_order (gdbarch));
4484 return addr;
4485 }
4486
4487
4488 /* Push a descriptor of type TYPE for array value ARR on the stack at
4489 *SP, updating *SP to reflect the new descriptor. Return either
4490 an lvalue representing the new descriptor, or (if TYPE is a pointer-
4491 to-descriptor type rather than a descriptor type), a struct value *
4492 representing a pointer to this descriptor. */
4493
4494 static struct value *
4495 make_array_descriptor (struct type *type, struct value *arr)
4496 {
4497 struct type *bounds_type = desc_bounds_type (type);
4498 struct type *desc_type = desc_base_type (type);
4499 struct value *descriptor = allocate_value (desc_type);
4500 struct value *bounds = allocate_value (bounds_type);
4501 int i;
4502
4503 for (i = ada_array_arity (ada_check_typedef (value_type (arr)));
4504 i > 0; i -= 1)
4505 {
4506 modify_field (value_type (bounds), value_contents_writeable (bounds),
4507 ada_array_bound (arr, i, 0),
4508 desc_bound_bitpos (bounds_type, i, 0),
4509 desc_bound_bitsize (bounds_type, i, 0));
4510 modify_field (value_type (bounds), value_contents_writeable (bounds),
4511 ada_array_bound (arr, i, 1),
4512 desc_bound_bitpos (bounds_type, i, 1),
4513 desc_bound_bitsize (bounds_type, i, 1));
4514 }
4515
4516 bounds = ensure_lval (bounds);
4517
4518 modify_field (value_type (descriptor),
4519 value_contents_writeable (descriptor),
4520 value_pointer (ensure_lval (arr),
4521 TYPE_FIELD_TYPE (desc_type, 0)),
4522 fat_pntr_data_bitpos (desc_type),
4523 fat_pntr_data_bitsize (desc_type));
4524
4525 modify_field (value_type (descriptor),
4526 value_contents_writeable (descriptor),
4527 value_pointer (bounds,
4528 TYPE_FIELD_TYPE (desc_type, 1)),
4529 fat_pntr_bounds_bitpos (desc_type),
4530 fat_pntr_bounds_bitsize (desc_type));
4531
4532 descriptor = ensure_lval (descriptor);
4533
4534 if (TYPE_CODE (type) == TYPE_CODE_PTR)
4535 return value_addr (descriptor);
4536 else
4537 return descriptor;
4538 }
4539 \f
4540 /* Symbol Cache Module */
4541
4542 /* Performance measurements made as of 2010-01-15 indicate that
4543 this cache does bring some noticeable improvements. Depending
4544 on the type of entity being printed, the cache can make it as much
4545 as an order of magnitude faster than without it.
4546
4547 The descriptive type DWARF extension has significantly reduced
4548 the need for this cache, at least when DWARF is being used. However,
4549 even in this case, some expensive name-based symbol searches are still
4550 sometimes necessary - to find an XVZ variable, mostly. */
4551
4552 /* Initialize the contents of SYM_CACHE. */
4553
4554 static void
4555 ada_init_symbol_cache (struct ada_symbol_cache *sym_cache)
4556 {
4557 obstack_init (&sym_cache->cache_space);
4558 memset (sym_cache->root, '\000', sizeof (sym_cache->root));
4559 }
4560
4561 /* Free the memory used by SYM_CACHE. */
4562
4563 static void
4564 ada_free_symbol_cache (struct ada_symbol_cache *sym_cache)
4565 {
4566 obstack_free (&sym_cache->cache_space, NULL);
4567 xfree (sym_cache);
4568 }
4569
4570 /* Return the symbol cache associated to the given program space PSPACE.
4571 If not allocated for this PSPACE yet, allocate and initialize one. */
4572
4573 static struct ada_symbol_cache *
4574 ada_get_symbol_cache (struct program_space *pspace)
4575 {
4576 struct ada_pspace_data *pspace_data = get_ada_pspace_data (pspace);
4577
4578 if (pspace_data->sym_cache == NULL)
4579 {
4580 pspace_data->sym_cache = XCNEW (struct ada_symbol_cache);
4581 ada_init_symbol_cache (pspace_data->sym_cache);
4582 }
4583
4584 return pspace_data->sym_cache;
4585 }
4586
4587 /* Clear all entries from the symbol cache. */
4588
4589 static void
4590 ada_clear_symbol_cache (void)
4591 {
4592 struct ada_symbol_cache *sym_cache
4593 = ada_get_symbol_cache (current_program_space);
4594
4595 obstack_free (&sym_cache->cache_space, NULL);
4596 ada_init_symbol_cache (sym_cache);
4597 }
4598
4599 /* Search our cache for an entry matching NAME and DOMAIN.
4600 Return it if found, or NULL otherwise. */
4601
4602 static struct cache_entry **
4603 find_entry (const char *name, domain_enum domain)
4604 {
4605 struct ada_symbol_cache *sym_cache
4606 = ada_get_symbol_cache (current_program_space);
4607 int h = msymbol_hash (name) % HASH_SIZE;
4608 struct cache_entry **e;
4609
4610 for (e = &sym_cache->root[h]; *e != NULL; e = &(*e)->next)
4611 {
4612 if (domain == (*e)->domain && strcmp (name, (*e)->name) == 0)
4613 return e;
4614 }
4615 return NULL;
4616 }
4617
4618 /* Search the symbol cache for an entry matching NAME and DOMAIN.
4619 Return 1 if found, 0 otherwise.
4620
4621 If an entry was found and SYM is not NULL, set *SYM to the entry's
4622 SYM. Same principle for BLOCK if not NULL. */
4623
4624 static int
4625 lookup_cached_symbol (const char *name, domain_enum domain,
4626 struct symbol **sym, const struct block **block)
4627 {
4628 struct cache_entry **e = find_entry (name, domain);
4629
4630 if (e == NULL)
4631 return 0;
4632 if (sym != NULL)
4633 *sym = (*e)->sym;
4634 if (block != NULL)
4635 *block = (*e)->block;
4636 return 1;
4637 }
4638
4639 /* Assuming that (SYM, BLOCK) is the result of the lookup of NAME
4640 in domain DOMAIN, save this result in our symbol cache. */
4641
4642 static void
4643 cache_symbol (const char *name, domain_enum domain, struct symbol *sym,
4644 const struct block *block)
4645 {
4646 struct ada_symbol_cache *sym_cache
4647 = ada_get_symbol_cache (current_program_space);
4648 int h;
4649 char *copy;
4650 struct cache_entry *e;
4651
4652 /* Symbols for builtin types don't have a block.
4653 For now don't cache such symbols. */
4654 if (sym != NULL && !SYMBOL_OBJFILE_OWNED (sym))
4655 return;
4656
4657 /* If the symbol is a local symbol, then do not cache it, as a search
4658 for that symbol depends on the context. To determine whether
4659 the symbol is local or not, we check the block where we found it
4660 against the global and static blocks of its associated symtab. */
4661 if (sym
4662 && BLOCKVECTOR_BLOCK (SYMTAB_BLOCKVECTOR (symbol_symtab (sym)),
4663 GLOBAL_BLOCK) != block
4664 && BLOCKVECTOR_BLOCK (SYMTAB_BLOCKVECTOR (symbol_symtab (sym)),
4665 STATIC_BLOCK) != block)
4666 return;
4667
4668 h = msymbol_hash (name) % HASH_SIZE;
4669 e = (struct cache_entry *) obstack_alloc (&sym_cache->cache_space,
4670 sizeof (*e));
4671 e->next = sym_cache->root[h];
4672 sym_cache->root[h] = e;
4673 e->name = copy
4674 = (char *) obstack_alloc (&sym_cache->cache_space, strlen (name) + 1);
4675 strcpy (copy, name);
4676 e->sym = sym;
4677 e->domain = domain;
4678 e->block = block;
4679 }
4680 \f
4681 /* Symbol Lookup */
4682
4683 /* Return nonzero if wild matching should be used when searching for
4684 all symbols matching LOOKUP_NAME.
4685
4686 LOOKUP_NAME is expected to be a symbol name after transformation
4687 for Ada lookups (see ada_name_for_lookup). */
4688
4689 static int
4690 should_use_wild_match (const char *lookup_name)
4691 {
4692 return (strstr (lookup_name, "__") == NULL);
4693 }
4694
4695 /* Return the result of a standard (literal, C-like) lookup of NAME in
4696 given DOMAIN, visible from lexical block BLOCK. */
4697
4698 static struct symbol *
4699 standard_lookup (const char *name, const struct block *block,
4700 domain_enum domain)
4701 {
4702 /* Initialize it just to avoid a GCC false warning. */
4703 struct block_symbol sym = {NULL, NULL};
4704
4705 if (lookup_cached_symbol (name, domain, &sym.symbol, NULL))
4706 return sym.symbol;
4707 sym = lookup_symbol_in_language (name, block, domain, language_c, 0);
4708 cache_symbol (name, domain, sym.symbol, sym.block);
4709 return sym.symbol;
4710 }
4711
4712
4713 /* Non-zero iff there is at least one non-function/non-enumeral symbol
4714 in the symbol fields of SYMS[0..N-1]. We treat enumerals as functions,
4715 since they contend in overloading in the same way. */
4716 static int
4717 is_nonfunction (struct block_symbol syms[], int n)
4718 {
4719 int i;
4720
4721 for (i = 0; i < n; i += 1)
4722 if (TYPE_CODE (SYMBOL_TYPE (syms[i].symbol)) != TYPE_CODE_FUNC
4723 && (TYPE_CODE (SYMBOL_TYPE (syms[i].symbol)) != TYPE_CODE_ENUM
4724 || SYMBOL_CLASS (syms[i].symbol) != LOC_CONST))
4725 return 1;
4726
4727 return 0;
4728 }
4729
4730 /* If true (non-zero), then TYPE0 and TYPE1 represent equivalent
4731 struct types. Otherwise, they may not. */
4732
4733 static int
4734 equiv_types (struct type *type0, struct type *type1)
4735 {
4736 if (type0 == type1)
4737 return 1;
4738 if (type0 == NULL || type1 == NULL
4739 || TYPE_CODE (type0) != TYPE_CODE (type1))
4740 return 0;
4741 if ((TYPE_CODE (type0) == TYPE_CODE_STRUCT
4742 || TYPE_CODE (type0) == TYPE_CODE_ENUM)
4743 && ada_type_name (type0) != NULL && ada_type_name (type1) != NULL
4744 && strcmp (ada_type_name (type0), ada_type_name (type1)) == 0)
4745 return 1;
4746
4747 return 0;
4748 }
4749
4750 /* True iff SYM0 represents the same entity as SYM1, or one that is
4751 no more defined than that of SYM1. */
4752
4753 static int
4754 lesseq_defined_than (struct symbol *sym0, struct symbol *sym1)
4755 {
4756 if (sym0 == sym1)
4757 return 1;
4758 if (SYMBOL_DOMAIN (sym0) != SYMBOL_DOMAIN (sym1)
4759 || SYMBOL_CLASS (sym0) != SYMBOL_CLASS (sym1))
4760 return 0;
4761
4762 switch (SYMBOL_CLASS (sym0))
4763 {
4764 case LOC_UNDEF:
4765 return 1;
4766 case LOC_TYPEDEF:
4767 {
4768 struct type *type0 = SYMBOL_TYPE (sym0);
4769 struct type *type1 = SYMBOL_TYPE (sym1);
4770 const char *name0 = SYMBOL_LINKAGE_NAME (sym0);
4771 const char *name1 = SYMBOL_LINKAGE_NAME (sym1);
4772 int len0 = strlen (name0);
4773
4774 return
4775 TYPE_CODE (type0) == TYPE_CODE (type1)
4776 && (equiv_types (type0, type1)
4777 || (len0 < strlen (name1) && strncmp (name0, name1, len0) == 0
4778 && startswith (name1 + len0, "___XV")));
4779 }
4780 case LOC_CONST:
4781 return SYMBOL_VALUE (sym0) == SYMBOL_VALUE (sym1)
4782 && equiv_types (SYMBOL_TYPE (sym0), SYMBOL_TYPE (sym1));
4783 default:
4784 return 0;
4785 }
4786 }
4787
4788 /* Append (SYM,BLOCK,SYMTAB) to the end of the array of struct block_symbol
4789 records in OBSTACKP. Do nothing if SYM is a duplicate. */
4790
4791 static void
4792 add_defn_to_vec (struct obstack *obstackp,
4793 struct symbol *sym,
4794 const struct block *block)
4795 {
4796 int i;
4797 struct block_symbol *prevDefns = defns_collected (obstackp, 0);
4798
4799 /* Do not try to complete stub types, as the debugger is probably
4800 already scanning all symbols matching a certain name at the
4801 time when this function is called. Trying to replace the stub
4802 type by its associated full type will cause us to restart a scan
4803 which may lead to an infinite recursion. Instead, the client
4804 collecting the matching symbols will end up collecting several
4805 matches, with at least one of them complete. It can then filter
4806 out the stub ones if needed. */
4807
4808 for (i = num_defns_collected (obstackp) - 1; i >= 0; i -= 1)
4809 {
4810 if (lesseq_defined_than (sym, prevDefns[i].symbol))
4811 return;
4812 else if (lesseq_defined_than (prevDefns[i].symbol, sym))
4813 {
4814 prevDefns[i].symbol = sym;
4815 prevDefns[i].block = block;
4816 return;
4817 }
4818 }
4819
4820 {
4821 struct block_symbol info;
4822
4823 info.symbol = sym;
4824 info.block = block;
4825 obstack_grow (obstackp, &info, sizeof (struct block_symbol));
4826 }
4827 }
4828
4829 /* Number of block_symbol structures currently collected in current vector in
4830 OBSTACKP. */
4831
4832 static int
4833 num_defns_collected (struct obstack *obstackp)
4834 {
4835 return obstack_object_size (obstackp) / sizeof (struct block_symbol);
4836 }
4837
4838 /* Vector of block_symbol structures currently collected in current vector in
4839 OBSTACKP. If FINISH, close off the vector and return its final address. */
4840
4841 static struct block_symbol *
4842 defns_collected (struct obstack *obstackp, int finish)
4843 {
4844 if (finish)
4845 return (struct block_symbol *) obstack_finish (obstackp);
4846 else
4847 return (struct block_symbol *) obstack_base (obstackp);
4848 }
4849
4850 /* Return a bound minimal symbol matching NAME according to Ada
4851 decoding rules. Returns an invalid symbol if there is no such
4852 minimal symbol. Names prefixed with "standard__" are handled
4853 specially: "standard__" is first stripped off, and only static and
4854 global symbols are searched. */
4855
4856 struct bound_minimal_symbol
4857 ada_lookup_simple_minsym (const char *name)
4858 {
4859 struct bound_minimal_symbol result;
4860 struct objfile *objfile;
4861 struct minimal_symbol *msymbol;
4862 const int wild_match_p = should_use_wild_match (name);
4863
4864 memset (&result, 0, sizeof (result));
4865
4866 /* Special case: If the user specifies a symbol name inside package
4867 Standard, do a non-wild matching of the symbol name without
4868 the "standard__" prefix. This was primarily introduced in order
4869 to allow the user to specifically access the standard exceptions
4870 using, for instance, Standard.Constraint_Error when Constraint_Error
4871 is ambiguous (due to the user defining its own Constraint_Error
4872 entity inside its program). */
4873 if (startswith (name, "standard__"))
4874 name += sizeof ("standard__") - 1;
4875
4876 ALL_MSYMBOLS (objfile, msymbol)
4877 {
4878 if (match_name (MSYMBOL_LINKAGE_NAME (msymbol), name, wild_match_p)
4879 && MSYMBOL_TYPE (msymbol) != mst_solib_trampoline)
4880 {
4881 result.minsym = msymbol;
4882 result.objfile = objfile;
4883 break;
4884 }
4885 }
4886
4887 return result;
4888 }
4889
4890 /* For all subprograms that statically enclose the subprogram of the
4891 selected frame, add symbols matching identifier NAME in DOMAIN
4892 and their blocks to the list of data in OBSTACKP, as for
4893 ada_add_block_symbols (q.v.). If WILD_MATCH_P, treat as NAME
4894 with a wildcard prefix. */
4895
4896 static void
4897 add_symbols_from_enclosing_procs (struct obstack *obstackp,
4898 const char *name, domain_enum domain,
4899 int wild_match_p)
4900 {
4901 }
4902
4903 /* True if TYPE is definitely an artificial type supplied to a symbol
4904 for which no debugging information was given in the symbol file. */
4905
4906 static int
4907 is_nondebugging_type (struct type *type)
4908 {
4909 const char *name = ada_type_name (type);
4910
4911 return (name != NULL && strcmp (name, "<variable, no debug info>") == 0);
4912 }
4913
4914 /* Return nonzero if TYPE1 and TYPE2 are two enumeration types
4915 that are deemed "identical" for practical purposes.
4916
4917 This function assumes that TYPE1 and TYPE2 are both TYPE_CODE_ENUM
4918 types and that their number of enumerals is identical (in other
4919 words, TYPE_NFIELDS (type1) == TYPE_NFIELDS (type2)). */
4920
4921 static int
4922 ada_identical_enum_types_p (struct type *type1, struct type *type2)
4923 {
4924 int i;
4925
4926 /* The heuristic we use here is fairly conservative. We consider
4927 that 2 enumerate types are identical if they have the same
4928 number of enumerals and that all enumerals have the same
4929 underlying value and name. */
4930
4931 /* All enums in the type should have an identical underlying value. */
4932 for (i = 0; i < TYPE_NFIELDS (type1); i++)
4933 if (TYPE_FIELD_ENUMVAL (type1, i) != TYPE_FIELD_ENUMVAL (type2, i))
4934 return 0;
4935
4936 /* All enumerals should also have the same name (modulo any numerical
4937 suffix). */
4938 for (i = 0; i < TYPE_NFIELDS (type1); i++)
4939 {
4940 const char *name_1 = TYPE_FIELD_NAME (type1, i);
4941 const char *name_2 = TYPE_FIELD_NAME (type2, i);
4942 int len_1 = strlen (name_1);
4943 int len_2 = strlen (name_2);
4944
4945 ada_remove_trailing_digits (TYPE_FIELD_NAME (type1, i), &len_1);
4946 ada_remove_trailing_digits (TYPE_FIELD_NAME (type2, i), &len_2);
4947 if (len_1 != len_2
4948 || strncmp (TYPE_FIELD_NAME (type1, i),
4949 TYPE_FIELD_NAME (type2, i),
4950 len_1) != 0)
4951 return 0;
4952 }
4953
4954 return 1;
4955 }
4956
4957 /* Return nonzero if all the symbols in SYMS are all enumeral symbols
4958 that are deemed "identical" for practical purposes. Sometimes,
4959 enumerals are not strictly identical, but their types are so similar
4960 that they can be considered identical.
4961
4962 For instance, consider the following code:
4963
4964 type Color is (Black, Red, Green, Blue, White);
4965 type RGB_Color is new Color range Red .. Blue;
4966
4967 Type RGB_Color is a subrange of an implicit type which is a copy
4968 of type Color. If we call that implicit type RGB_ColorB ("B" is
4969 for "Base Type"), then type RGB_ColorB is a copy of type Color.
4970 As a result, when an expression references any of the enumeral
4971 by name (Eg. "print green"), the expression is technically
4972 ambiguous and the user should be asked to disambiguate. But
4973 doing so would only hinder the user, since it wouldn't matter
4974 what choice he makes, the outcome would always be the same.
4975 So, for practical purposes, we consider them as the same. */
4976
4977 static int
4978 symbols_are_identical_enums (struct block_symbol *syms, int nsyms)
4979 {
4980 int i;
4981
4982 /* Before performing a thorough comparison check of each type,
4983 we perform a series of inexpensive checks. We expect that these
4984 checks will quickly fail in the vast majority of cases, and thus
4985 help prevent the unnecessary use of a more expensive comparison.
4986 Said comparison also expects us to make some of these checks
4987 (see ada_identical_enum_types_p). */
4988
4989 /* Quick check: All symbols should have an enum type. */
4990 for (i = 0; i < nsyms; i++)
4991 if (TYPE_CODE (SYMBOL_TYPE (syms[i].symbol)) != TYPE_CODE_ENUM)
4992 return 0;
4993
4994 /* Quick check: They should all have the same value. */
4995 for (i = 1; i < nsyms; i++)
4996 if (SYMBOL_VALUE (syms[i].symbol) != SYMBOL_VALUE (syms[0].symbol))
4997 return 0;
4998
4999 /* Quick check: They should all have the same number of enumerals. */
5000 for (i = 1; i < nsyms; i++)
5001 if (TYPE_NFIELDS (SYMBOL_TYPE (syms[i].symbol))
5002 != TYPE_NFIELDS (SYMBOL_TYPE (syms[0].symbol)))
5003 return 0;
5004
5005 /* All the sanity checks passed, so we might have a set of
5006 identical enumeration types. Perform a more complete
5007 comparison of the type of each symbol. */
5008 for (i = 1; i < nsyms; i++)
5009 if (!ada_identical_enum_types_p (SYMBOL_TYPE (syms[i].symbol),
5010 SYMBOL_TYPE (syms[0].symbol)))
5011 return 0;
5012
5013 return 1;
5014 }
5015
5016 /* Remove any non-debugging symbols in SYMS[0 .. NSYMS-1] that definitely
5017 duplicate other symbols in the list (The only case I know of where
5018 this happens is when object files containing stabs-in-ecoff are
5019 linked with files containing ordinary ecoff debugging symbols (or no
5020 debugging symbols)). Modifies SYMS to squeeze out deleted entries.
5021 Returns the number of items in the modified list. */
5022
5023 static int
5024 remove_extra_symbols (struct block_symbol *syms, int nsyms)
5025 {
5026 int i, j;
5027
5028 /* We should never be called with less than 2 symbols, as there
5029 cannot be any extra symbol in that case. But it's easy to
5030 handle, since we have nothing to do in that case. */
5031 if (nsyms < 2)
5032 return nsyms;
5033
5034 i = 0;
5035 while (i < nsyms)
5036 {
5037 int remove_p = 0;
5038
5039 /* If two symbols have the same name and one of them is a stub type,
5040 the get rid of the stub. */
5041
5042 if (TYPE_STUB (SYMBOL_TYPE (syms[i].symbol))
5043 && SYMBOL_LINKAGE_NAME (syms[i].symbol) != NULL)
5044 {
5045 for (j = 0; j < nsyms; j++)
5046 {
5047 if (j != i
5048 && !TYPE_STUB (SYMBOL_TYPE (syms[j].symbol))
5049 && SYMBOL_LINKAGE_NAME (syms[j].symbol) != NULL
5050 && strcmp (SYMBOL_LINKAGE_NAME (syms[i].symbol),
5051 SYMBOL_LINKAGE_NAME (syms[j].symbol)) == 0)
5052 remove_p = 1;
5053 }
5054 }
5055
5056 /* Two symbols with the same name, same class and same address
5057 should be identical. */
5058
5059 else if (SYMBOL_LINKAGE_NAME (syms[i].symbol) != NULL
5060 && SYMBOL_CLASS (syms[i].symbol) == LOC_STATIC
5061 && is_nondebugging_type (SYMBOL_TYPE (syms[i].symbol)))
5062 {
5063 for (j = 0; j < nsyms; j += 1)
5064 {
5065 if (i != j
5066 && SYMBOL_LINKAGE_NAME (syms[j].symbol) != NULL
5067 && strcmp (SYMBOL_LINKAGE_NAME (syms[i].symbol),
5068 SYMBOL_LINKAGE_NAME (syms[j].symbol)) == 0
5069 && SYMBOL_CLASS (syms[i].symbol)
5070 == SYMBOL_CLASS (syms[j].symbol)
5071 && SYMBOL_VALUE_ADDRESS (syms[i].symbol)
5072 == SYMBOL_VALUE_ADDRESS (syms[j].symbol))
5073 remove_p = 1;
5074 }
5075 }
5076
5077 if (remove_p)
5078 {
5079 for (j = i + 1; j < nsyms; j += 1)
5080 syms[j - 1] = syms[j];
5081 nsyms -= 1;
5082 }
5083
5084 i += 1;
5085 }
5086
5087 /* If all the remaining symbols are identical enumerals, then
5088 just keep the first one and discard the rest.
5089
5090 Unlike what we did previously, we do not discard any entry
5091 unless they are ALL identical. This is because the symbol
5092 comparison is not a strict comparison, but rather a practical
5093 comparison. If all symbols are considered identical, then
5094 we can just go ahead and use the first one and discard the rest.
5095 But if we cannot reduce the list to a single element, we have
5096 to ask the user to disambiguate anyways. And if we have to
5097 present a multiple-choice menu, it's less confusing if the list
5098 isn't missing some choices that were identical and yet distinct. */
5099 if (symbols_are_identical_enums (syms, nsyms))
5100 nsyms = 1;
5101
5102 return nsyms;
5103 }
5104
5105 /* Given a type that corresponds to a renaming entity, use the type name
5106 to extract the scope (package name or function name, fully qualified,
5107 and following the GNAT encoding convention) where this renaming has been
5108 defined. The string returned needs to be deallocated after use. */
5109
5110 static char *
5111 xget_renaming_scope (struct type *renaming_type)
5112 {
5113 /* The renaming types adhere to the following convention:
5114 <scope>__<rename>___<XR extension>.
5115 So, to extract the scope, we search for the "___XR" extension,
5116 and then backtrack until we find the first "__". */
5117
5118 const char *name = type_name_no_tag (renaming_type);
5119 const char *suffix = strstr (name, "___XR");
5120 const char *last;
5121 int scope_len;
5122 char *scope;
5123
5124 /* Now, backtrack a bit until we find the first "__". Start looking
5125 at suffix - 3, as the <rename> part is at least one character long. */
5126
5127 for (last = suffix - 3; last > name; last--)
5128 if (last[0] == '_' && last[1] == '_')
5129 break;
5130
5131 /* Make a copy of scope and return it. */
5132
5133 scope_len = last - name;
5134 scope = (char *) xmalloc ((scope_len + 1) * sizeof (char));
5135
5136 strncpy (scope, name, scope_len);
5137 scope[scope_len] = '\0';
5138
5139 return scope;
5140 }
5141
5142 /* Return nonzero if NAME corresponds to a package name. */
5143
5144 static int
5145 is_package_name (const char *name)
5146 {
5147 /* Here, We take advantage of the fact that no symbols are generated
5148 for packages, while symbols are generated for each function.
5149 So the condition for NAME represent a package becomes equivalent
5150 to NAME not existing in our list of symbols. There is only one
5151 small complication with library-level functions (see below). */
5152
5153 char *fun_name;
5154
5155 /* If it is a function that has not been defined at library level,
5156 then we should be able to look it up in the symbols. */
5157 if (standard_lookup (name, NULL, VAR_DOMAIN) != NULL)
5158 return 0;
5159
5160 /* Library-level function names start with "_ada_". See if function
5161 "_ada_" followed by NAME can be found. */
5162
5163 /* Do a quick check that NAME does not contain "__", since library-level
5164 functions names cannot contain "__" in them. */
5165 if (strstr (name, "__") != NULL)
5166 return 0;
5167
5168 fun_name = xstrprintf ("_ada_%s", name);
5169
5170 return (standard_lookup (fun_name, NULL, VAR_DOMAIN) == NULL);
5171 }
5172
5173 /* Return nonzero if SYM corresponds to a renaming entity that is
5174 not visible from FUNCTION_NAME. */
5175
5176 static int
5177 old_renaming_is_invisible (const struct symbol *sym, const char *function_name)
5178 {
5179 char *scope;
5180 struct cleanup *old_chain;
5181
5182 if (SYMBOL_CLASS (sym) != LOC_TYPEDEF)
5183 return 0;
5184
5185 scope = xget_renaming_scope (SYMBOL_TYPE (sym));
5186 old_chain = make_cleanup (xfree, scope);
5187
5188 /* If the rename has been defined in a package, then it is visible. */
5189 if (is_package_name (scope))
5190 {
5191 do_cleanups (old_chain);
5192 return 0;
5193 }
5194
5195 /* Check that the rename is in the current function scope by checking
5196 that its name starts with SCOPE. */
5197
5198 /* If the function name starts with "_ada_", it means that it is
5199 a library-level function. Strip this prefix before doing the
5200 comparison, as the encoding for the renaming does not contain
5201 this prefix. */
5202 if (startswith (function_name, "_ada_"))
5203 function_name += 5;
5204
5205 {
5206 int is_invisible = !startswith (function_name, scope);
5207
5208 do_cleanups (old_chain);
5209 return is_invisible;
5210 }
5211 }
5212
5213 /* Remove entries from SYMS that corresponds to a renaming entity that
5214 is not visible from the function associated with CURRENT_BLOCK or
5215 that is superfluous due to the presence of more specific renaming
5216 information. Places surviving symbols in the initial entries of
5217 SYMS and returns the number of surviving symbols.
5218
5219 Rationale:
5220 First, in cases where an object renaming is implemented as a
5221 reference variable, GNAT may produce both the actual reference
5222 variable and the renaming encoding. In this case, we discard the
5223 latter.
5224
5225 Second, GNAT emits a type following a specified encoding for each renaming
5226 entity. Unfortunately, STABS currently does not support the definition
5227 of types that are local to a given lexical block, so all renamings types
5228 are emitted at library level. As a consequence, if an application
5229 contains two renaming entities using the same name, and a user tries to
5230 print the value of one of these entities, the result of the ada symbol
5231 lookup will also contain the wrong renaming type.
5232
5233 This function partially covers for this limitation by attempting to
5234 remove from the SYMS list renaming symbols that should be visible
5235 from CURRENT_BLOCK. However, there does not seem be a 100% reliable
5236 method with the current information available. The implementation
5237 below has a couple of limitations (FIXME: brobecker-2003-05-12):
5238
5239 - When the user tries to print a rename in a function while there
5240 is another rename entity defined in a package: Normally, the
5241 rename in the function has precedence over the rename in the
5242 package, so the latter should be removed from the list. This is
5243 currently not the case.
5244
5245 - This function will incorrectly remove valid renames if
5246 the CURRENT_BLOCK corresponds to a function which symbol name
5247 has been changed by an "Export" pragma. As a consequence,
5248 the user will be unable to print such rename entities. */
5249
5250 static int
5251 remove_irrelevant_renamings (struct block_symbol *syms,
5252 int nsyms, const struct block *current_block)
5253 {
5254 struct symbol *current_function;
5255 const char *current_function_name;
5256 int i;
5257 int is_new_style_renaming;
5258
5259 /* If there is both a renaming foo___XR... encoded as a variable and
5260 a simple variable foo in the same block, discard the latter.
5261 First, zero out such symbols, then compress. */
5262 is_new_style_renaming = 0;
5263 for (i = 0; i < nsyms; i += 1)
5264 {
5265 struct symbol *sym = syms[i].symbol;
5266 const struct block *block = syms[i].block;
5267 const char *name;
5268 const char *suffix;
5269
5270 if (sym == NULL || SYMBOL_CLASS (sym) == LOC_TYPEDEF)
5271 continue;
5272 name = SYMBOL_LINKAGE_NAME (sym);
5273 suffix = strstr (name, "___XR");
5274
5275 if (suffix != NULL)
5276 {
5277 int name_len = suffix - name;
5278 int j;
5279
5280 is_new_style_renaming = 1;
5281 for (j = 0; j < nsyms; j += 1)
5282 if (i != j && syms[j].symbol != NULL
5283 && strncmp (name, SYMBOL_LINKAGE_NAME (syms[j].symbol),
5284 name_len) == 0
5285 && block == syms[j].block)
5286 syms[j].symbol = NULL;
5287 }
5288 }
5289 if (is_new_style_renaming)
5290 {
5291 int j, k;
5292
5293 for (j = k = 0; j < nsyms; j += 1)
5294 if (syms[j].symbol != NULL)
5295 {
5296 syms[k] = syms[j];
5297 k += 1;
5298 }
5299 return k;
5300 }
5301
5302 /* Extract the function name associated to CURRENT_BLOCK.
5303 Abort if unable to do so. */
5304
5305 if (current_block == NULL)
5306 return nsyms;
5307
5308 current_function = block_linkage_function (current_block);
5309 if (current_function == NULL)
5310 return nsyms;
5311
5312 current_function_name = SYMBOL_LINKAGE_NAME (current_function);
5313 if (current_function_name == NULL)
5314 return nsyms;
5315
5316 /* Check each of the symbols, and remove it from the list if it is
5317 a type corresponding to a renaming that is out of the scope of
5318 the current block. */
5319
5320 i = 0;
5321 while (i < nsyms)
5322 {
5323 if (ada_parse_renaming (syms[i].symbol, NULL, NULL, NULL)
5324 == ADA_OBJECT_RENAMING
5325 && old_renaming_is_invisible (syms[i].symbol, current_function_name))
5326 {
5327 int j;
5328
5329 for (j = i + 1; j < nsyms; j += 1)
5330 syms[j - 1] = syms[j];
5331 nsyms -= 1;
5332 }
5333 else
5334 i += 1;
5335 }
5336
5337 return nsyms;
5338 }
5339
5340 /* Add to OBSTACKP all symbols from BLOCK (and its super-blocks)
5341 whose name and domain match NAME and DOMAIN respectively.
5342 If no match was found, then extend the search to "enclosing"
5343 routines (in other words, if we're inside a nested function,
5344 search the symbols defined inside the enclosing functions).
5345 If WILD_MATCH_P is nonzero, perform the naming matching in
5346 "wild" mode (see function "wild_match" for more info).
5347
5348 Note: This function assumes that OBSTACKP has 0 (zero) element in it. */
5349
5350 static void
5351 ada_add_local_symbols (struct obstack *obstackp, const char *name,
5352 const struct block *block, domain_enum domain,
5353 int wild_match_p)
5354 {
5355 int block_depth = 0;
5356
5357 while (block != NULL)
5358 {
5359 block_depth += 1;
5360 ada_add_block_symbols (obstackp, block, name, domain, NULL,
5361 wild_match_p);
5362
5363 /* If we found a non-function match, assume that's the one. */
5364 if (is_nonfunction (defns_collected (obstackp, 0),
5365 num_defns_collected (obstackp)))
5366 return;
5367
5368 block = BLOCK_SUPERBLOCK (block);
5369 }
5370
5371 /* If no luck so far, try to find NAME as a local symbol in some lexically
5372 enclosing subprogram. */
5373 if (num_defns_collected (obstackp) == 0 && block_depth > 2)
5374 add_symbols_from_enclosing_procs (obstackp, name, domain, wild_match_p);
5375 }
5376
5377 /* An object of this type is used as the user_data argument when
5378 calling the map_matching_symbols method. */
5379
5380 struct match_data
5381 {
5382 struct objfile *objfile;
5383 struct obstack *obstackp;
5384 struct symbol *arg_sym;
5385 int found_sym;
5386 };
5387
5388 /* A callback for add_nonlocal_symbols that adds SYM, found in BLOCK,
5389 to a list of symbols. DATA0 is a pointer to a struct match_data *
5390 containing the obstack that collects the symbol list, the file that SYM
5391 must come from, a flag indicating whether a non-argument symbol has
5392 been found in the current block, and the last argument symbol
5393 passed in SYM within the current block (if any). When SYM is null,
5394 marking the end of a block, the argument symbol is added if no
5395 other has been found. */
5396
5397 static int
5398 aux_add_nonlocal_symbols (struct block *block, struct symbol *sym, void *data0)
5399 {
5400 struct match_data *data = (struct match_data *) data0;
5401
5402 if (sym == NULL)
5403 {
5404 if (!data->found_sym && data->arg_sym != NULL)
5405 add_defn_to_vec (data->obstackp,
5406 fixup_symbol_section (data->arg_sym, data->objfile),
5407 block);
5408 data->found_sym = 0;
5409 data->arg_sym = NULL;
5410 }
5411 else
5412 {
5413 if (SYMBOL_CLASS (sym) == LOC_UNRESOLVED)
5414 return 0;
5415 else if (SYMBOL_IS_ARGUMENT (sym))
5416 data->arg_sym = sym;
5417 else
5418 {
5419 data->found_sym = 1;
5420 add_defn_to_vec (data->obstackp,
5421 fixup_symbol_section (sym, data->objfile),
5422 block);
5423 }
5424 }
5425 return 0;
5426 }
5427
5428 /* Helper for add_nonlocal_symbols. Find symbols in DOMAIN which are targetted
5429 by renamings matching NAME in BLOCK. Add these symbols to OBSTACKP. If
5430 WILD_MATCH_P is nonzero, perform the naming matching in "wild" mode (see
5431 function "wild_match" for more information). Return whether we found such
5432 symbols. */
5433
5434 static int
5435 ada_add_block_renamings (struct obstack *obstackp,
5436 const struct block *block,
5437 const char *name,
5438 domain_enum domain,
5439 int wild_match_p)
5440 {
5441 struct using_direct *renaming;
5442 int defns_mark = num_defns_collected (obstackp);
5443
5444 for (renaming = block_using (block);
5445 renaming != NULL;
5446 renaming = renaming->next)
5447 {
5448 const char *r_name;
5449 int name_match;
5450
5451 /* Avoid infinite recursions: skip this renaming if we are actually
5452 already traversing it.
5453
5454 Currently, symbol lookup in Ada don't use the namespace machinery from
5455 C++/Fortran support: skip namespace imports that use them. */
5456 if (renaming->searched
5457 || (renaming->import_src != NULL
5458 && renaming->import_src[0] != '\0')
5459 || (renaming->import_dest != NULL
5460 && renaming->import_dest[0] != '\0'))
5461 continue;
5462 renaming->searched = 1;
5463
5464 /* TODO: here, we perform another name-based symbol lookup, which can
5465 pull its own multiple overloads. In theory, we should be able to do
5466 better in this case since, in DWARF, DW_AT_import is a DIE reference,
5467 not a simple name. But in order to do this, we would need to enhance
5468 the DWARF reader to associate a symbol to this renaming, instead of a
5469 name. So, for now, we do something simpler: re-use the C++/Fortran
5470 namespace machinery. */
5471 r_name = (renaming->alias != NULL
5472 ? renaming->alias
5473 : renaming->declaration);
5474 name_match
5475 = wild_match_p ? wild_match (r_name, name) : strcmp (r_name, name);
5476 if (name_match == 0)
5477 ada_add_all_symbols (obstackp, block, renaming->declaration, domain,
5478 1, NULL);
5479 renaming->searched = 0;
5480 }
5481 return num_defns_collected (obstackp) != defns_mark;
5482 }
5483
5484 /* Implements compare_names, but only applying the comparision using
5485 the given CASING. */
5486
5487 static int
5488 compare_names_with_case (const char *string1, const char *string2,
5489 enum case_sensitivity casing)
5490 {
5491 while (*string1 != '\0' && *string2 != '\0')
5492 {
5493 char c1, c2;
5494
5495 if (isspace (*string1) || isspace (*string2))
5496 return strcmp_iw_ordered (string1, string2);
5497
5498 if (casing == case_sensitive_off)
5499 {
5500 c1 = tolower (*string1);
5501 c2 = tolower (*string2);
5502 }
5503 else
5504 {
5505 c1 = *string1;
5506 c2 = *string2;
5507 }
5508 if (c1 != c2)
5509 break;
5510
5511 string1 += 1;
5512 string2 += 1;
5513 }
5514
5515 switch (*string1)
5516 {
5517 case '(':
5518 return strcmp_iw_ordered (string1, string2);
5519 case '_':
5520 if (*string2 == '\0')
5521 {
5522 if (is_name_suffix (string1))
5523 return 0;
5524 else
5525 return 1;
5526 }
5527 /* FALLTHROUGH */
5528 default:
5529 if (*string2 == '(')
5530 return strcmp_iw_ordered (string1, string2);
5531 else
5532 {
5533 if (casing == case_sensitive_off)
5534 return tolower (*string1) - tolower (*string2);
5535 else
5536 return *string1 - *string2;
5537 }
5538 }
5539 }
5540
5541 /* Compare STRING1 to STRING2, with results as for strcmp.
5542 Compatible with strcmp_iw_ordered in that...
5543
5544 strcmp_iw_ordered (STRING1, STRING2) <= 0
5545
5546 ... implies...
5547
5548 compare_names (STRING1, STRING2) <= 0
5549
5550 (they may differ as to what symbols compare equal). */
5551
5552 static int
5553 compare_names (const char *string1, const char *string2)
5554 {
5555 int result;
5556
5557 /* Similar to what strcmp_iw_ordered does, we need to perform
5558 a case-insensitive comparison first, and only resort to
5559 a second, case-sensitive, comparison if the first one was
5560 not sufficient to differentiate the two strings. */
5561
5562 result = compare_names_with_case (string1, string2, case_sensitive_off);
5563 if (result == 0)
5564 result = compare_names_with_case (string1, string2, case_sensitive_on);
5565
5566 return result;
5567 }
5568
5569 /* Add to OBSTACKP all non-local symbols whose name and domain match
5570 NAME and DOMAIN respectively. The search is performed on GLOBAL_BLOCK
5571 symbols if GLOBAL is non-zero, or on STATIC_BLOCK symbols otherwise. */
5572
5573 static void
5574 add_nonlocal_symbols (struct obstack *obstackp, const char *name,
5575 domain_enum domain, int global,
5576 int is_wild_match)
5577 {
5578 struct objfile *objfile;
5579 struct compunit_symtab *cu;
5580 struct match_data data;
5581
5582 memset (&data, 0, sizeof data);
5583 data.obstackp = obstackp;
5584
5585 ALL_OBJFILES (objfile)
5586 {
5587 data.objfile = objfile;
5588
5589 if (is_wild_match)
5590 objfile->sf->qf->map_matching_symbols (objfile, name, domain, global,
5591 aux_add_nonlocal_symbols, &data,
5592 wild_match, NULL);
5593 else
5594 objfile->sf->qf->map_matching_symbols (objfile, name, domain, global,
5595 aux_add_nonlocal_symbols, &data,
5596 full_match, compare_names);
5597
5598 ALL_OBJFILE_COMPUNITS (objfile, cu)
5599 {
5600 const struct block *global_block
5601 = BLOCKVECTOR_BLOCK (COMPUNIT_BLOCKVECTOR (cu), GLOBAL_BLOCK);
5602
5603 if (ada_add_block_renamings (obstackp, global_block , name, domain,
5604 is_wild_match))
5605 data.found_sym = 1;
5606 }
5607 }
5608
5609 if (num_defns_collected (obstackp) == 0 && global && !is_wild_match)
5610 {
5611 ALL_OBJFILES (objfile)
5612 {
5613 char *name1 = (char *) alloca (strlen (name) + sizeof ("_ada_"));
5614 strcpy (name1, "_ada_");
5615 strcpy (name1 + sizeof ("_ada_") - 1, name);
5616 data.objfile = objfile;
5617 objfile->sf->qf->map_matching_symbols (objfile, name1, domain,
5618 global,
5619 aux_add_nonlocal_symbols,
5620 &data,
5621 full_match, compare_names);
5622 }
5623 }
5624 }
5625
5626 /* Find symbols in DOMAIN matching NAME, in BLOCK and, if FULL_SEARCH is
5627 non-zero, enclosing scope and in global scopes, returning the number of
5628 matches. Add these to OBSTACKP.
5629
5630 When FULL_SEARCH is non-zero, any non-function/non-enumeral
5631 symbol match within the nest of blocks whose innermost member is BLOCK,
5632 is the one match returned (no other matches in that or
5633 enclosing blocks is returned). If there are any matches in or
5634 surrounding BLOCK, then these alone are returned.
5635
5636 Names prefixed with "standard__" are handled specially: "standard__"
5637 is first stripped off, and only static and global symbols are searched.
5638
5639 If MADE_GLOBAL_LOOKUP_P is non-null, set it before return to whether we had
5640 to lookup global symbols. */
5641
5642 static void
5643 ada_add_all_symbols (struct obstack *obstackp,
5644 const struct block *block,
5645 const char *name,
5646 domain_enum domain,
5647 int full_search,
5648 int *made_global_lookup_p)
5649 {
5650 struct symbol *sym;
5651 const int wild_match_p = should_use_wild_match (name);
5652
5653 if (made_global_lookup_p)
5654 *made_global_lookup_p = 0;
5655
5656 /* Special case: If the user specifies a symbol name inside package
5657 Standard, do a non-wild matching of the symbol name without
5658 the "standard__" prefix. This was primarily introduced in order
5659 to allow the user to specifically access the standard exceptions
5660 using, for instance, Standard.Constraint_Error when Constraint_Error
5661 is ambiguous (due to the user defining its own Constraint_Error
5662 entity inside its program). */
5663 if (startswith (name, "standard__"))
5664 {
5665 block = NULL;
5666 name = name + sizeof ("standard__") - 1;
5667 }
5668
5669 /* Check the non-global symbols. If we have ANY match, then we're done. */
5670
5671 if (block != NULL)
5672 {
5673 if (full_search)
5674 ada_add_local_symbols (obstackp, name, block, domain, wild_match_p);
5675 else
5676 {
5677 /* In the !full_search case we're are being called by
5678 ada_iterate_over_symbols, and we don't want to search
5679 superblocks. */
5680 ada_add_block_symbols (obstackp, block, name, domain, NULL,
5681 wild_match_p);
5682 }
5683 if (num_defns_collected (obstackp) > 0 || !full_search)
5684 return;
5685 }
5686
5687 /* No non-global symbols found. Check our cache to see if we have
5688 already performed this search before. If we have, then return
5689 the same result. */
5690
5691 if (lookup_cached_symbol (name, domain, &sym, &block))
5692 {
5693 if (sym != NULL)
5694 add_defn_to_vec (obstackp, sym, block);
5695 return;
5696 }
5697
5698 if (made_global_lookup_p)
5699 *made_global_lookup_p = 1;
5700
5701 /* Search symbols from all global blocks. */
5702
5703 add_nonlocal_symbols (obstackp, name, domain, 1, wild_match_p);
5704
5705 /* Now add symbols from all per-file blocks if we've gotten no hits
5706 (not strictly correct, but perhaps better than an error). */
5707
5708 if (num_defns_collected (obstackp) == 0)
5709 add_nonlocal_symbols (obstackp, name, domain, 0, wild_match_p);
5710 }
5711
5712 /* Find symbols in DOMAIN matching NAME, in BLOCK and, if full_search is
5713 non-zero, enclosing scope and in global scopes, returning the number of
5714 matches.
5715 Sets *RESULTS to point to a vector of (SYM,BLOCK) tuples,
5716 indicating the symbols found and the blocks and symbol tables (if
5717 any) in which they were found. This vector is transient---good only to
5718 the next call of ada_lookup_symbol_list.
5719
5720 When full_search is non-zero, any non-function/non-enumeral
5721 symbol match within the nest of blocks whose innermost member is BLOCK,
5722 is the one match returned (no other matches in that or
5723 enclosing blocks is returned). If there are any matches in or
5724 surrounding BLOCK, then these alone are returned.
5725
5726 Names prefixed with "standard__" are handled specially: "standard__"
5727 is first stripped off, and only static and global symbols are searched. */
5728
5729 static int
5730 ada_lookup_symbol_list_worker (const char *name, const struct block *block,
5731 domain_enum domain,
5732 struct block_symbol **results,
5733 int full_search)
5734 {
5735 const int wild_match_p = should_use_wild_match (name);
5736 int syms_from_global_search;
5737 int ndefns;
5738
5739 obstack_free (&symbol_list_obstack, NULL);
5740 obstack_init (&symbol_list_obstack);
5741 ada_add_all_symbols (&symbol_list_obstack, block, name, domain,
5742 full_search, &syms_from_global_search);
5743
5744 ndefns = num_defns_collected (&symbol_list_obstack);
5745 *results = defns_collected (&symbol_list_obstack, 1);
5746
5747 ndefns = remove_extra_symbols (*results, ndefns);
5748
5749 if (ndefns == 0 && full_search && syms_from_global_search)
5750 cache_symbol (name, domain, NULL, NULL);
5751
5752 if (ndefns == 1 && full_search && syms_from_global_search)
5753 cache_symbol (name, domain, (*results)[0].symbol, (*results)[0].block);
5754
5755 ndefns = remove_irrelevant_renamings (*results, ndefns, block);
5756 return ndefns;
5757 }
5758
5759 /* Find symbols in DOMAIN matching NAME0, in BLOCK0 and enclosing scope and
5760 in global scopes, returning the number of matches, and setting *RESULTS
5761 to a vector of (SYM,BLOCK) tuples.
5762 See ada_lookup_symbol_list_worker for further details. */
5763
5764 int
5765 ada_lookup_symbol_list (const char *name0, const struct block *block0,
5766 domain_enum domain, struct block_symbol **results)
5767 {
5768 return ada_lookup_symbol_list_worker (name0, block0, domain, results, 1);
5769 }
5770
5771 /* Implementation of the la_iterate_over_symbols method. */
5772
5773 static void
5774 ada_iterate_over_symbols (const struct block *block,
5775 const char *name, domain_enum domain,
5776 symbol_found_callback_ftype *callback,
5777 void *data)
5778 {
5779 int ndefs, i;
5780 struct block_symbol *results;
5781
5782 ndefs = ada_lookup_symbol_list_worker (name, block, domain, &results, 0);
5783 for (i = 0; i < ndefs; ++i)
5784 {
5785 if (! (*callback) (results[i].symbol, data))
5786 break;
5787 }
5788 }
5789
5790 /* If NAME is the name of an entity, return a string that should
5791 be used to look that entity up in Ada units. This string should
5792 be deallocated after use using xfree.
5793
5794 NAME can have any form that the "break" or "print" commands might
5795 recognize. In other words, it does not have to be the "natural"
5796 name, or the "encoded" name. */
5797
5798 char *
5799 ada_name_for_lookup (const char *name)
5800 {
5801 char *canon;
5802 int nlen = strlen (name);
5803
5804 if (name[0] == '<' && name[nlen - 1] == '>')
5805 {
5806 canon = (char *) xmalloc (nlen - 1);
5807 memcpy (canon, name + 1, nlen - 2);
5808 canon[nlen - 2] = '\0';
5809 }
5810 else
5811 canon = xstrdup (ada_encode (ada_fold_name (name)));
5812 return canon;
5813 }
5814
5815 /* The result is as for ada_lookup_symbol_list with FULL_SEARCH set
5816 to 1, but choosing the first symbol found if there are multiple
5817 choices.
5818
5819 The result is stored in *INFO, which must be non-NULL.
5820 If no match is found, INFO->SYM is set to NULL. */
5821
5822 void
5823 ada_lookup_encoded_symbol (const char *name, const struct block *block,
5824 domain_enum domain,
5825 struct block_symbol *info)
5826 {
5827 struct block_symbol *candidates;
5828 int n_candidates;
5829
5830 gdb_assert (info != NULL);
5831 memset (info, 0, sizeof (struct block_symbol));
5832
5833 n_candidates = ada_lookup_symbol_list (name, block, domain, &candidates);
5834 if (n_candidates == 0)
5835 return;
5836
5837 *info = candidates[0];
5838 info->symbol = fixup_symbol_section (info->symbol, NULL);
5839 }
5840
5841 /* Return a symbol in DOMAIN matching NAME, in BLOCK0 and enclosing
5842 scope and in global scopes, or NULL if none. NAME is folded and
5843 encoded first. Otherwise, the result is as for ada_lookup_symbol_list,
5844 choosing the first symbol if there are multiple choices.
5845 If IS_A_FIELD_OF_THIS is not NULL, it is set to zero. */
5846
5847 struct block_symbol
5848 ada_lookup_symbol (const char *name, const struct block *block0,
5849 domain_enum domain, int *is_a_field_of_this)
5850 {
5851 struct block_symbol info;
5852
5853 if (is_a_field_of_this != NULL)
5854 *is_a_field_of_this = 0;
5855
5856 ada_lookup_encoded_symbol (ada_encode (ada_fold_name (name)),
5857 block0, domain, &info);
5858 return info;
5859 }
5860
5861 static struct block_symbol
5862 ada_lookup_symbol_nonlocal (const struct language_defn *langdef,
5863 const char *name,
5864 const struct block *block,
5865 const domain_enum domain)
5866 {
5867 struct block_symbol sym;
5868
5869 sym = ada_lookup_symbol (name, block_static_block (block), domain, NULL);
5870 if (sym.symbol != NULL)
5871 return sym;
5872
5873 /* If we haven't found a match at this point, try the primitive
5874 types. In other languages, this search is performed before
5875 searching for global symbols in order to short-circuit that
5876 global-symbol search if it happens that the name corresponds
5877 to a primitive type. But we cannot do the same in Ada, because
5878 it is perfectly legitimate for a program to declare a type which
5879 has the same name as a standard type. If looking up a type in
5880 that situation, we have traditionally ignored the primitive type
5881 in favor of user-defined types. This is why, unlike most other
5882 languages, we search the primitive types this late and only after
5883 having searched the global symbols without success. */
5884
5885 if (domain == VAR_DOMAIN)
5886 {
5887 struct gdbarch *gdbarch;
5888
5889 if (block == NULL)
5890 gdbarch = target_gdbarch ();
5891 else
5892 gdbarch = block_gdbarch (block);
5893 sym.symbol = language_lookup_primitive_type_as_symbol (langdef, gdbarch, name);
5894 if (sym.symbol != NULL)
5895 return sym;
5896 }
5897
5898 return (struct block_symbol) {NULL, NULL};
5899 }
5900
5901
5902 /* True iff STR is a possible encoded suffix of a normal Ada name
5903 that is to be ignored for matching purposes. Suffixes of parallel
5904 names (e.g., XVE) are not included here. Currently, the possible suffixes
5905 are given by any of the regular expressions:
5906
5907 [.$][0-9]+ [nested subprogram suffix, on platforms such as GNU/Linux]
5908 ___[0-9]+ [nested subprogram suffix, on platforms such as HP/UX]
5909 TKB [subprogram suffix for task bodies]
5910 _E[0-9]+[bs]$ [protected object entry suffixes]
5911 (X[nb]*)?((\$|__)[0-9](_?[0-9]+)|___(JM|LJM|X([FDBUP].*|R[^T]?)))?$
5912
5913 Also, any leading "__[0-9]+" sequence is skipped before the suffix
5914 match is performed. This sequence is used to differentiate homonyms,
5915 is an optional part of a valid name suffix. */
5916
5917 static int
5918 is_name_suffix (const char *str)
5919 {
5920 int k;
5921 const char *matching;
5922 const int len = strlen (str);
5923
5924 /* Skip optional leading __[0-9]+. */
5925
5926 if (len > 3 && str[0] == '_' && str[1] == '_' && isdigit (str[2]))
5927 {
5928 str += 3;
5929 while (isdigit (str[0]))
5930 str += 1;
5931 }
5932
5933 /* [.$][0-9]+ */
5934
5935 if (str[0] == '.' || str[0] == '$')
5936 {
5937 matching = str + 1;
5938 while (isdigit (matching[0]))
5939 matching += 1;
5940 if (matching[0] == '\0')
5941 return 1;
5942 }
5943
5944 /* ___[0-9]+ */
5945
5946 if (len > 3 && str[0] == '_' && str[1] == '_' && str[2] == '_')
5947 {
5948 matching = str + 3;
5949 while (isdigit (matching[0]))
5950 matching += 1;
5951 if (matching[0] == '\0')
5952 return 1;
5953 }
5954
5955 /* "TKB" suffixes are used for subprograms implementing task bodies. */
5956
5957 if (strcmp (str, "TKB") == 0)
5958 return 1;
5959
5960 #if 0
5961 /* FIXME: brobecker/2005-09-23: Protected Object subprograms end
5962 with a N at the end. Unfortunately, the compiler uses the same
5963 convention for other internal types it creates. So treating
5964 all entity names that end with an "N" as a name suffix causes
5965 some regressions. For instance, consider the case of an enumerated
5966 type. To support the 'Image attribute, it creates an array whose
5967 name ends with N.
5968 Having a single character like this as a suffix carrying some
5969 information is a bit risky. Perhaps we should change the encoding
5970 to be something like "_N" instead. In the meantime, do not do
5971 the following check. */
5972 /* Protected Object Subprograms */
5973 if (len == 1 && str [0] == 'N')
5974 return 1;
5975 #endif
5976
5977 /* _E[0-9]+[bs]$ */
5978 if (len > 3 && str[0] == '_' && str [1] == 'E' && isdigit (str[2]))
5979 {
5980 matching = str + 3;
5981 while (isdigit (matching[0]))
5982 matching += 1;
5983 if ((matching[0] == 'b' || matching[0] == 's')
5984 && matching [1] == '\0')
5985 return 1;
5986 }
5987
5988 /* ??? We should not modify STR directly, as we are doing below. This
5989 is fine in this case, but may become problematic later if we find
5990 that this alternative did not work, and want to try matching
5991 another one from the begining of STR. Since we modified it, we
5992 won't be able to find the begining of the string anymore! */
5993 if (str[0] == 'X')
5994 {
5995 str += 1;
5996 while (str[0] != '_' && str[0] != '\0')
5997 {
5998 if (str[0] != 'n' && str[0] != 'b')
5999 return 0;
6000 str += 1;
6001 }
6002 }
6003
6004 if (str[0] == '\000')
6005 return 1;
6006
6007 if (str[0] == '_')
6008 {
6009 if (str[1] != '_' || str[2] == '\000')
6010 return 0;
6011 if (str[2] == '_')
6012 {
6013 if (strcmp (str + 3, "JM") == 0)
6014 return 1;
6015 /* FIXME: brobecker/2004-09-30: GNAT will soon stop using
6016 the LJM suffix in favor of the JM one. But we will
6017 still accept LJM as a valid suffix for a reasonable
6018 amount of time, just to allow ourselves to debug programs
6019 compiled using an older version of GNAT. */
6020 if (strcmp (str + 3, "LJM") == 0)
6021 return 1;
6022 if (str[3] != 'X')
6023 return 0;
6024 if (str[4] == 'F' || str[4] == 'D' || str[4] == 'B'
6025 || str[4] == 'U' || str[4] == 'P')
6026 return 1;
6027 if (str[4] == 'R' && str[5] != 'T')
6028 return 1;
6029 return 0;
6030 }
6031 if (!isdigit (str[2]))
6032 return 0;
6033 for (k = 3; str[k] != '\0'; k += 1)
6034 if (!isdigit (str[k]) && str[k] != '_')
6035 return 0;
6036 return 1;
6037 }
6038 if (str[0] == '$' && isdigit (str[1]))
6039 {
6040 for (k = 2; str[k] != '\0'; k += 1)
6041 if (!isdigit (str[k]) && str[k] != '_')
6042 return 0;
6043 return 1;
6044 }
6045 return 0;
6046 }
6047
6048 /* Return non-zero if the string starting at NAME and ending before
6049 NAME_END contains no capital letters. */
6050
6051 static int
6052 is_valid_name_for_wild_match (const char *name0)
6053 {
6054 const char *decoded_name = ada_decode (name0);
6055 int i;
6056
6057 /* If the decoded name starts with an angle bracket, it means that
6058 NAME0 does not follow the GNAT encoding format. It should then
6059 not be allowed as a possible wild match. */
6060 if (decoded_name[0] == '<')
6061 return 0;
6062
6063 for (i=0; decoded_name[i] != '\0'; i++)
6064 if (isalpha (decoded_name[i]) && !islower (decoded_name[i]))
6065 return 0;
6066
6067 return 1;
6068 }
6069
6070 /* Advance *NAMEP to next occurrence of TARGET0 in the string NAME0
6071 that could start a simple name. Assumes that *NAMEP points into
6072 the string beginning at NAME0. */
6073
6074 static int
6075 advance_wild_match (const char **namep, const char *name0, int target0)
6076 {
6077 const char *name = *namep;
6078
6079 while (1)
6080 {
6081 int t0, t1;
6082
6083 t0 = *name;
6084 if (t0 == '_')
6085 {
6086 t1 = name[1];
6087 if ((t1 >= 'a' && t1 <= 'z') || (t1 >= '0' && t1 <= '9'))
6088 {
6089 name += 1;
6090 if (name == name0 + 5 && startswith (name0, "_ada"))
6091 break;
6092 else
6093 name += 1;
6094 }
6095 else if (t1 == '_' && ((name[2] >= 'a' && name[2] <= 'z')
6096 || name[2] == target0))
6097 {
6098 name += 2;
6099 break;
6100 }
6101 else
6102 return 0;
6103 }
6104 else if ((t0 >= 'a' && t0 <= 'z') || (t0 >= '0' && t0 <= '9'))
6105 name += 1;
6106 else
6107 return 0;
6108 }
6109
6110 *namep = name;
6111 return 1;
6112 }
6113
6114 /* Return 0 iff NAME encodes a name of the form prefix.PATN. Ignores any
6115 informational suffixes of NAME (i.e., for which is_name_suffix is
6116 true). Assumes that PATN is a lower-cased Ada simple name. */
6117
6118 static int
6119 wild_match (const char *name, const char *patn)
6120 {
6121 const char *p;
6122 const char *name0 = name;
6123
6124 while (1)
6125 {
6126 const char *match = name;
6127
6128 if (*name == *patn)
6129 {
6130 for (name += 1, p = patn + 1; *p != '\0'; name += 1, p += 1)
6131 if (*p != *name)
6132 break;
6133 if (*p == '\0' && is_name_suffix (name))
6134 return match != name0 && !is_valid_name_for_wild_match (name0);
6135
6136 if (name[-1] == '_')
6137 name -= 1;
6138 }
6139 if (!advance_wild_match (&name, name0, *patn))
6140 return 1;
6141 }
6142 }
6143
6144 /* Returns 0 iff symbol name SYM_NAME matches SEARCH_NAME, apart from
6145 informational suffix. */
6146
6147 static int
6148 full_match (const char *sym_name, const char *search_name)
6149 {
6150 return !match_name (sym_name, search_name, 0);
6151 }
6152
6153
6154 /* Add symbols from BLOCK matching identifier NAME in DOMAIN to
6155 vector *defn_symbols, updating the list of symbols in OBSTACKP
6156 (if necessary). If WILD, treat as NAME with a wildcard prefix.
6157 OBJFILE is the section containing BLOCK. */
6158
6159 static void
6160 ada_add_block_symbols (struct obstack *obstackp,
6161 const struct block *block, const char *name,
6162 domain_enum domain, struct objfile *objfile,
6163 int wild)
6164 {
6165 struct block_iterator iter;
6166 int name_len = strlen (name);
6167 /* A matching argument symbol, if any. */
6168 struct symbol *arg_sym;
6169 /* Set true when we find a matching non-argument symbol. */
6170 int found_sym;
6171 struct symbol *sym;
6172
6173 arg_sym = NULL;
6174 found_sym = 0;
6175 if (wild)
6176 {
6177 for (sym = block_iter_match_first (block, name, wild_match, &iter);
6178 sym != NULL; sym = block_iter_match_next (name, wild_match, &iter))
6179 {
6180 if (symbol_matches_domain (SYMBOL_LANGUAGE (sym),
6181 SYMBOL_DOMAIN (sym), domain)
6182 && wild_match (SYMBOL_LINKAGE_NAME (sym), name) == 0)
6183 {
6184 if (SYMBOL_CLASS (sym) == LOC_UNRESOLVED)
6185 continue;
6186 else if (SYMBOL_IS_ARGUMENT (sym))
6187 arg_sym = sym;
6188 else
6189 {
6190 found_sym = 1;
6191 add_defn_to_vec (obstackp,
6192 fixup_symbol_section (sym, objfile),
6193 block);
6194 }
6195 }
6196 }
6197 }
6198 else
6199 {
6200 for (sym = block_iter_match_first (block, name, full_match, &iter);
6201 sym != NULL; sym = block_iter_match_next (name, full_match, &iter))
6202 {
6203 if (symbol_matches_domain (SYMBOL_LANGUAGE (sym),
6204 SYMBOL_DOMAIN (sym), domain))
6205 {
6206 if (SYMBOL_CLASS (sym) != LOC_UNRESOLVED)
6207 {
6208 if (SYMBOL_IS_ARGUMENT (sym))
6209 arg_sym = sym;
6210 else
6211 {
6212 found_sym = 1;
6213 add_defn_to_vec (obstackp,
6214 fixup_symbol_section (sym, objfile),
6215 block);
6216 }
6217 }
6218 }
6219 }
6220 }
6221
6222 /* Handle renamings. */
6223
6224 if (ada_add_block_renamings (obstackp, block, name, domain, wild))
6225 found_sym = 1;
6226
6227 if (!found_sym && arg_sym != NULL)
6228 {
6229 add_defn_to_vec (obstackp,
6230 fixup_symbol_section (arg_sym, objfile),
6231 block);
6232 }
6233
6234 if (!wild)
6235 {
6236 arg_sym = NULL;
6237 found_sym = 0;
6238
6239 ALL_BLOCK_SYMBOLS (block, iter, sym)
6240 {
6241 if (symbol_matches_domain (SYMBOL_LANGUAGE (sym),
6242 SYMBOL_DOMAIN (sym), domain))
6243 {
6244 int cmp;
6245
6246 cmp = (int) '_' - (int) SYMBOL_LINKAGE_NAME (sym)[0];
6247 if (cmp == 0)
6248 {
6249 cmp = !startswith (SYMBOL_LINKAGE_NAME (sym), "_ada_");
6250 if (cmp == 0)
6251 cmp = strncmp (name, SYMBOL_LINKAGE_NAME (sym) + 5,
6252 name_len);
6253 }
6254
6255 if (cmp == 0
6256 && is_name_suffix (SYMBOL_LINKAGE_NAME (sym) + name_len + 5))
6257 {
6258 if (SYMBOL_CLASS (sym) != LOC_UNRESOLVED)
6259 {
6260 if (SYMBOL_IS_ARGUMENT (sym))
6261 arg_sym = sym;
6262 else
6263 {
6264 found_sym = 1;
6265 add_defn_to_vec (obstackp,
6266 fixup_symbol_section (sym, objfile),
6267 block);
6268 }
6269 }
6270 }
6271 }
6272 }
6273
6274 /* NOTE: This really shouldn't be needed for _ada_ symbols.
6275 They aren't parameters, right? */
6276 if (!found_sym && arg_sym != NULL)
6277 {
6278 add_defn_to_vec (obstackp,
6279 fixup_symbol_section (arg_sym, objfile),
6280 block);
6281 }
6282 }
6283 }
6284 \f
6285
6286 /* Symbol Completion */
6287
6288 /* If SYM_NAME is a completion candidate for TEXT, return this symbol
6289 name in a form that's appropriate for the completion. The result
6290 does not need to be deallocated, but is only good until the next call.
6291
6292 TEXT_LEN is equal to the length of TEXT.
6293 Perform a wild match if WILD_MATCH_P is set.
6294 ENCODED_P should be set if TEXT represents the start of a symbol name
6295 in its encoded form. */
6296
6297 static const char *
6298 symbol_completion_match (const char *sym_name,
6299 const char *text, int text_len,
6300 int wild_match_p, int encoded_p)
6301 {
6302 const int verbatim_match = (text[0] == '<');
6303 int match = 0;
6304
6305 if (verbatim_match)
6306 {
6307 /* Strip the leading angle bracket. */
6308 text = text + 1;
6309 text_len--;
6310 }
6311
6312 /* First, test against the fully qualified name of the symbol. */
6313
6314 if (strncmp (sym_name, text, text_len) == 0)
6315 match = 1;
6316
6317 if (match && !encoded_p)
6318 {
6319 /* One needed check before declaring a positive match is to verify
6320 that iff we are doing a verbatim match, the decoded version
6321 of the symbol name starts with '<'. Otherwise, this symbol name
6322 is not a suitable completion. */
6323 const char *sym_name_copy = sym_name;
6324 int has_angle_bracket;
6325
6326 sym_name = ada_decode (sym_name);
6327 has_angle_bracket = (sym_name[0] == '<');
6328 match = (has_angle_bracket == verbatim_match);
6329 sym_name = sym_name_copy;
6330 }
6331
6332 if (match && !verbatim_match)
6333 {
6334 /* When doing non-verbatim match, another check that needs to
6335 be done is to verify that the potentially matching symbol name
6336 does not include capital letters, because the ada-mode would
6337 not be able to understand these symbol names without the
6338 angle bracket notation. */
6339 const char *tmp;
6340
6341 for (tmp = sym_name; *tmp != '\0' && !isupper (*tmp); tmp++);
6342 if (*tmp != '\0')
6343 match = 0;
6344 }
6345
6346 /* Second: Try wild matching... */
6347
6348 if (!match && wild_match_p)
6349 {
6350 /* Since we are doing wild matching, this means that TEXT
6351 may represent an unqualified symbol name. We therefore must
6352 also compare TEXT against the unqualified name of the symbol. */
6353 sym_name = ada_unqualified_name (ada_decode (sym_name));
6354
6355 if (strncmp (sym_name, text, text_len) == 0)
6356 match = 1;
6357 }
6358
6359 /* Finally: If we found a mach, prepare the result to return. */
6360
6361 if (!match)
6362 return NULL;
6363
6364 if (verbatim_match)
6365 sym_name = add_angle_brackets (sym_name);
6366
6367 if (!encoded_p)
6368 sym_name = ada_decode (sym_name);
6369
6370 return sym_name;
6371 }
6372
6373 /* A companion function to ada_make_symbol_completion_list().
6374 Check if SYM_NAME represents a symbol which name would be suitable
6375 to complete TEXT (TEXT_LEN is the length of TEXT), in which case
6376 it is appended at the end of the given string vector SV.
6377
6378 ORIG_TEXT is the string original string from the user command
6379 that needs to be completed. WORD is the entire command on which
6380 completion should be performed. These two parameters are used to
6381 determine which part of the symbol name should be added to the
6382 completion vector.
6383 if WILD_MATCH_P is set, then wild matching is performed.
6384 ENCODED_P should be set if TEXT represents a symbol name in its
6385 encoded formed (in which case the completion should also be
6386 encoded). */
6387
6388 static void
6389 symbol_completion_add (VEC(char_ptr) **sv,
6390 const char *sym_name,
6391 const char *text, int text_len,
6392 const char *orig_text, const char *word,
6393 int wild_match_p, int encoded_p)
6394 {
6395 const char *match = symbol_completion_match (sym_name, text, text_len,
6396 wild_match_p, encoded_p);
6397 char *completion;
6398
6399 if (match == NULL)
6400 return;
6401
6402 /* We found a match, so add the appropriate completion to the given
6403 string vector. */
6404
6405 if (word == orig_text)
6406 {
6407 completion = (char *) xmalloc (strlen (match) + 5);
6408 strcpy (completion, match);
6409 }
6410 else if (word > orig_text)
6411 {
6412 /* Return some portion of sym_name. */
6413 completion = (char *) xmalloc (strlen (match) + 5);
6414 strcpy (completion, match + (word - orig_text));
6415 }
6416 else
6417 {
6418 /* Return some of ORIG_TEXT plus sym_name. */
6419 completion = (char *) xmalloc (strlen (match) + (orig_text - word) + 5);
6420 strncpy (completion, word, orig_text - word);
6421 completion[orig_text - word] = '\0';
6422 strcat (completion, match);
6423 }
6424
6425 VEC_safe_push (char_ptr, *sv, completion);
6426 }
6427
6428 /* An object of this type is passed as the user_data argument to the
6429 expand_symtabs_matching method. */
6430 struct add_partial_datum
6431 {
6432 VEC(char_ptr) **completions;
6433 const char *text;
6434 int text_len;
6435 const char *text0;
6436 const char *word;
6437 int wild_match;
6438 int encoded;
6439 };
6440
6441 /* A callback for expand_symtabs_matching. */
6442
6443 static int
6444 ada_complete_symbol_matcher (const char *name, void *user_data)
6445 {
6446 struct add_partial_datum *data = (struct add_partial_datum *) user_data;
6447
6448 return symbol_completion_match (name, data->text, data->text_len,
6449 data->wild_match, data->encoded) != NULL;
6450 }
6451
6452 /* Return a list of possible symbol names completing TEXT0. WORD is
6453 the entire command on which completion is made. */
6454
6455 static VEC (char_ptr) *
6456 ada_make_symbol_completion_list (const char *text0, const char *word,
6457 enum type_code code)
6458 {
6459 char *text;
6460 int text_len;
6461 int wild_match_p;
6462 int encoded_p;
6463 VEC(char_ptr) *completions = VEC_alloc (char_ptr, 128);
6464 struct symbol *sym;
6465 struct compunit_symtab *s;
6466 struct minimal_symbol *msymbol;
6467 struct objfile *objfile;
6468 const struct block *b, *surrounding_static_block = 0;
6469 int i;
6470 struct block_iterator iter;
6471 struct cleanup *old_chain = make_cleanup (null_cleanup, NULL);
6472
6473 gdb_assert (code == TYPE_CODE_UNDEF);
6474
6475 if (text0[0] == '<')
6476 {
6477 text = xstrdup (text0);
6478 make_cleanup (xfree, text);
6479 text_len = strlen (text);
6480 wild_match_p = 0;
6481 encoded_p = 1;
6482 }
6483 else
6484 {
6485 text = xstrdup (ada_encode (text0));
6486 make_cleanup (xfree, text);
6487 text_len = strlen (text);
6488 for (i = 0; i < text_len; i++)
6489 text[i] = tolower (text[i]);
6490
6491 encoded_p = (strstr (text0, "__") != NULL);
6492 /* If the name contains a ".", then the user is entering a fully
6493 qualified entity name, and the match must not be done in wild
6494 mode. Similarly, if the user wants to complete what looks like
6495 an encoded name, the match must not be done in wild mode. */
6496 wild_match_p = (strchr (text0, '.') == NULL && !encoded_p);
6497 }
6498
6499 /* First, look at the partial symtab symbols. */
6500 {
6501 struct add_partial_datum data;
6502
6503 data.completions = &completions;
6504 data.text = text;
6505 data.text_len = text_len;
6506 data.text0 = text0;
6507 data.word = word;
6508 data.wild_match = wild_match_p;
6509 data.encoded = encoded_p;
6510 expand_symtabs_matching (NULL, ada_complete_symbol_matcher, NULL,
6511 ALL_DOMAIN, &data);
6512 }
6513
6514 /* At this point scan through the misc symbol vectors and add each
6515 symbol you find to the list. Eventually we want to ignore
6516 anything that isn't a text symbol (everything else will be
6517 handled by the psymtab code above). */
6518
6519 ALL_MSYMBOLS (objfile, msymbol)
6520 {
6521 QUIT;
6522 symbol_completion_add (&completions, MSYMBOL_LINKAGE_NAME (msymbol),
6523 text, text_len, text0, word, wild_match_p,
6524 encoded_p);
6525 }
6526
6527 /* Search upwards from currently selected frame (so that we can
6528 complete on local vars. */
6529
6530 for (b = get_selected_block (0); b != NULL; b = BLOCK_SUPERBLOCK (b))
6531 {
6532 if (!BLOCK_SUPERBLOCK (b))
6533 surrounding_static_block = b; /* For elmin of dups */
6534
6535 ALL_BLOCK_SYMBOLS (b, iter, sym)
6536 {
6537 symbol_completion_add (&completions, SYMBOL_LINKAGE_NAME (sym),
6538 text, text_len, text0, word,
6539 wild_match_p, encoded_p);
6540 }
6541 }
6542
6543 /* Go through the symtabs and check the externs and statics for
6544 symbols which match. */
6545
6546 ALL_COMPUNITS (objfile, s)
6547 {
6548 QUIT;
6549 b = BLOCKVECTOR_BLOCK (COMPUNIT_BLOCKVECTOR (s), GLOBAL_BLOCK);
6550 ALL_BLOCK_SYMBOLS (b, iter, sym)
6551 {
6552 symbol_completion_add (&completions, SYMBOL_LINKAGE_NAME (sym),
6553 text, text_len, text0, word,
6554 wild_match_p, encoded_p);
6555 }
6556 }
6557
6558 ALL_COMPUNITS (objfile, s)
6559 {
6560 QUIT;
6561 b = BLOCKVECTOR_BLOCK (COMPUNIT_BLOCKVECTOR (s), STATIC_BLOCK);
6562 /* Don't do this block twice. */
6563 if (b == surrounding_static_block)
6564 continue;
6565 ALL_BLOCK_SYMBOLS (b, iter, sym)
6566 {
6567 symbol_completion_add (&completions, SYMBOL_LINKAGE_NAME (sym),
6568 text, text_len, text0, word,
6569 wild_match_p, encoded_p);
6570 }
6571 }
6572
6573 do_cleanups (old_chain);
6574 return completions;
6575 }
6576
6577 /* Field Access */
6578
6579 /* Return non-zero if TYPE is a pointer to the GNAT dispatch table used
6580 for tagged types. */
6581
6582 static int
6583 ada_is_dispatch_table_ptr_type (struct type *type)
6584 {
6585 const char *name;
6586
6587 if (TYPE_CODE (type) != TYPE_CODE_PTR)
6588 return 0;
6589
6590 name = TYPE_NAME (TYPE_TARGET_TYPE (type));
6591 if (name == NULL)
6592 return 0;
6593
6594 return (strcmp (name, "ada__tags__dispatch_table") == 0);
6595 }
6596
6597 /* Return non-zero if TYPE is an interface tag. */
6598
6599 static int
6600 ada_is_interface_tag (struct type *type)
6601 {
6602 const char *name = TYPE_NAME (type);
6603
6604 if (name == NULL)
6605 return 0;
6606
6607 return (strcmp (name, "ada__tags__interface_tag") == 0);
6608 }
6609
6610 /* True if field number FIELD_NUM in struct or union type TYPE is supposed
6611 to be invisible to users. */
6612
6613 int
6614 ada_is_ignored_field (struct type *type, int field_num)
6615 {
6616 if (field_num < 0 || field_num > TYPE_NFIELDS (type))
6617 return 1;
6618
6619 /* Check the name of that field. */
6620 {
6621 const char *name = TYPE_FIELD_NAME (type, field_num);
6622
6623 /* Anonymous field names should not be printed.
6624 brobecker/2007-02-20: I don't think this can actually happen
6625 but we don't want to print the value of annonymous fields anyway. */
6626 if (name == NULL)
6627 return 1;
6628
6629 /* Normally, fields whose name start with an underscore ("_")
6630 are fields that have been internally generated by the compiler,
6631 and thus should not be printed. The "_parent" field is special,
6632 however: This is a field internally generated by the compiler
6633 for tagged types, and it contains the components inherited from
6634 the parent type. This field should not be printed as is, but
6635 should not be ignored either. */
6636 if (name[0] == '_' && !startswith (name, "_parent"))
6637 return 1;
6638 }
6639
6640 /* If this is the dispatch table of a tagged type or an interface tag,
6641 then ignore. */
6642 if (ada_is_tagged_type (type, 1)
6643 && (ada_is_dispatch_table_ptr_type (TYPE_FIELD_TYPE (type, field_num))
6644 || ada_is_interface_tag (TYPE_FIELD_TYPE (type, field_num))))
6645 return 1;
6646
6647 /* Not a special field, so it should not be ignored. */
6648 return 0;
6649 }
6650
6651 /* True iff TYPE has a tag field. If REFOK, then TYPE may also be a
6652 pointer or reference type whose ultimate target has a tag field. */
6653
6654 int
6655 ada_is_tagged_type (struct type *type, int refok)
6656 {
6657 return (ada_lookup_struct_elt_type (type, "_tag", refok, 1, NULL) != NULL);
6658 }
6659
6660 /* True iff TYPE represents the type of X'Tag */
6661
6662 int
6663 ada_is_tag_type (struct type *type)
6664 {
6665 type = ada_check_typedef (type);
6666
6667 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_PTR)
6668 return 0;
6669 else
6670 {
6671 const char *name = ada_type_name (TYPE_TARGET_TYPE (type));
6672
6673 return (name != NULL
6674 && strcmp (name, "ada__tags__dispatch_table") == 0);
6675 }
6676 }
6677
6678 /* The type of the tag on VAL. */
6679
6680 struct type *
6681 ada_tag_type (struct value *val)
6682 {
6683 return ada_lookup_struct_elt_type (value_type (val), "_tag", 1, 0, NULL);
6684 }
6685
6686 /* Return 1 if TAG follows the old scheme for Ada tags (used for Ada 95,
6687 retired at Ada 05). */
6688
6689 static int
6690 is_ada95_tag (struct value *tag)
6691 {
6692 return ada_value_struct_elt (tag, "tsd", 1) != NULL;
6693 }
6694
6695 /* The value of the tag on VAL. */
6696
6697 struct value *
6698 ada_value_tag (struct value *val)
6699 {
6700 return ada_value_struct_elt (val, "_tag", 0);
6701 }
6702
6703 /* The value of the tag on the object of type TYPE whose contents are
6704 saved at VALADDR, if it is non-null, or is at memory address
6705 ADDRESS. */
6706
6707 static struct value *
6708 value_tag_from_contents_and_address (struct type *type,
6709 const gdb_byte *valaddr,
6710 CORE_ADDR address)
6711 {
6712 int tag_byte_offset;
6713 struct type *tag_type;
6714
6715 if (find_struct_field ("_tag", type, 0, &tag_type, &tag_byte_offset,
6716 NULL, NULL, NULL))
6717 {
6718 const gdb_byte *valaddr1 = ((valaddr == NULL)
6719 ? NULL
6720 : valaddr + tag_byte_offset);
6721 CORE_ADDR address1 = (address == 0) ? 0 : address + tag_byte_offset;
6722
6723 return value_from_contents_and_address (tag_type, valaddr1, address1);
6724 }
6725 return NULL;
6726 }
6727
6728 static struct type *
6729 type_from_tag (struct value *tag)
6730 {
6731 const char *type_name = ada_tag_name (tag);
6732
6733 if (type_name != NULL)
6734 return ada_find_any_type (ada_encode (type_name));
6735 return NULL;
6736 }
6737
6738 /* Given a value OBJ of a tagged type, return a value of this
6739 type at the base address of the object. The base address, as
6740 defined in Ada.Tags, it is the address of the primary tag of
6741 the object, and therefore where the field values of its full
6742 view can be fetched. */
6743
6744 struct value *
6745 ada_tag_value_at_base_address (struct value *obj)
6746 {
6747 struct value *val;
6748 LONGEST offset_to_top = 0;
6749 struct type *ptr_type, *obj_type;
6750 struct value *tag;
6751 CORE_ADDR base_address;
6752
6753 obj_type = value_type (obj);
6754
6755 /* It is the responsability of the caller to deref pointers. */
6756
6757 if (TYPE_CODE (obj_type) == TYPE_CODE_PTR
6758 || TYPE_CODE (obj_type) == TYPE_CODE_REF)
6759 return obj;
6760
6761 tag = ada_value_tag (obj);
6762 if (!tag)
6763 return obj;
6764
6765 /* Base addresses only appeared with Ada 05 and multiple inheritance. */
6766
6767 if (is_ada95_tag (tag))
6768 return obj;
6769
6770 ptr_type = builtin_type (target_gdbarch ())->builtin_data_ptr;
6771 ptr_type = lookup_pointer_type (ptr_type);
6772 val = value_cast (ptr_type, tag);
6773 if (!val)
6774 return obj;
6775
6776 /* It is perfectly possible that an exception be raised while
6777 trying to determine the base address, just like for the tag;
6778 see ada_tag_name for more details. We do not print the error
6779 message for the same reason. */
6780
6781 TRY
6782 {
6783 offset_to_top = value_as_long (value_ind (value_ptradd (val, -2)));
6784 }
6785
6786 CATCH (e, RETURN_MASK_ERROR)
6787 {
6788 return obj;
6789 }
6790 END_CATCH
6791
6792 /* If offset is null, nothing to do. */
6793
6794 if (offset_to_top == 0)
6795 return obj;
6796
6797 /* -1 is a special case in Ada.Tags; however, what should be done
6798 is not quite clear from the documentation. So do nothing for
6799 now. */
6800
6801 if (offset_to_top == -1)
6802 return obj;
6803
6804 base_address = value_address (obj) - offset_to_top;
6805 tag = value_tag_from_contents_and_address (obj_type, NULL, base_address);
6806
6807 /* Make sure that we have a proper tag at the new address.
6808 Otherwise, offset_to_top is bogus (which can happen when
6809 the object is not initialized yet). */
6810
6811 if (!tag)
6812 return obj;
6813
6814 obj_type = type_from_tag (tag);
6815
6816 if (!obj_type)
6817 return obj;
6818
6819 return value_from_contents_and_address (obj_type, NULL, base_address);
6820 }
6821
6822 /* Return the "ada__tags__type_specific_data" type. */
6823
6824 static struct type *
6825 ada_get_tsd_type (struct inferior *inf)
6826 {
6827 struct ada_inferior_data *data = get_ada_inferior_data (inf);
6828
6829 if (data->tsd_type == 0)
6830 data->tsd_type = ada_find_any_type ("ada__tags__type_specific_data");
6831 return data->tsd_type;
6832 }
6833
6834 /* Return the TSD (type-specific data) associated to the given TAG.
6835 TAG is assumed to be the tag of a tagged-type entity.
6836
6837 May return NULL if we are unable to get the TSD. */
6838
6839 static struct value *
6840 ada_get_tsd_from_tag (struct value *tag)
6841 {
6842 struct value *val;
6843 struct type *type;
6844
6845 /* First option: The TSD is simply stored as a field of our TAG.
6846 Only older versions of GNAT would use this format, but we have
6847 to test it first, because there are no visible markers for
6848 the current approach except the absence of that field. */
6849
6850 val = ada_value_struct_elt (tag, "tsd", 1);
6851 if (val)
6852 return val;
6853
6854 /* Try the second representation for the dispatch table (in which
6855 there is no explicit 'tsd' field in the referent of the tag pointer,
6856 and instead the tsd pointer is stored just before the dispatch
6857 table. */
6858
6859 type = ada_get_tsd_type (current_inferior());
6860 if (type == NULL)
6861 return NULL;
6862 type = lookup_pointer_type (lookup_pointer_type (type));
6863 val = value_cast (type, tag);
6864 if (val == NULL)
6865 return NULL;
6866 return value_ind (value_ptradd (val, -1));
6867 }
6868
6869 /* Given the TSD of a tag (type-specific data), return a string
6870 containing the name of the associated type.
6871
6872 The returned value is good until the next call. May return NULL
6873 if we are unable to determine the tag name. */
6874
6875 static char *
6876 ada_tag_name_from_tsd (struct value *tsd)
6877 {
6878 static char name[1024];
6879 char *p;
6880 struct value *val;
6881
6882 val = ada_value_struct_elt (tsd, "expanded_name", 1);
6883 if (val == NULL)
6884 return NULL;
6885 read_memory_string (value_as_address (val), name, sizeof (name) - 1);
6886 for (p = name; *p != '\0'; p += 1)
6887 if (isalpha (*p))
6888 *p = tolower (*p);
6889 return name;
6890 }
6891
6892 /* The type name of the dynamic type denoted by the 'tag value TAG, as
6893 a C string.
6894
6895 Return NULL if the TAG is not an Ada tag, or if we were unable to
6896 determine the name of that tag. The result is good until the next
6897 call. */
6898
6899 const char *
6900 ada_tag_name (struct value *tag)
6901 {
6902 char *name = NULL;
6903
6904 if (!ada_is_tag_type (value_type (tag)))
6905 return NULL;
6906
6907 /* It is perfectly possible that an exception be raised while trying
6908 to determine the TAG's name, even under normal circumstances:
6909 The associated variable may be uninitialized or corrupted, for
6910 instance. We do not let any exception propagate past this point.
6911 instead we return NULL.
6912
6913 We also do not print the error message either (which often is very
6914 low-level (Eg: "Cannot read memory at 0x[...]"), but instead let
6915 the caller print a more meaningful message if necessary. */
6916 TRY
6917 {
6918 struct value *tsd = ada_get_tsd_from_tag (tag);
6919
6920 if (tsd != NULL)
6921 name = ada_tag_name_from_tsd (tsd);
6922 }
6923 CATCH (e, RETURN_MASK_ERROR)
6924 {
6925 }
6926 END_CATCH
6927
6928 return name;
6929 }
6930
6931 /* The parent type of TYPE, or NULL if none. */
6932
6933 struct type *
6934 ada_parent_type (struct type *type)
6935 {
6936 int i;
6937
6938 type = ada_check_typedef (type);
6939
6940 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_STRUCT)
6941 return NULL;
6942
6943 for (i = 0; i < TYPE_NFIELDS (type); i += 1)
6944 if (ada_is_parent_field (type, i))
6945 {
6946 struct type *parent_type = TYPE_FIELD_TYPE (type, i);
6947
6948 /* If the _parent field is a pointer, then dereference it. */
6949 if (TYPE_CODE (parent_type) == TYPE_CODE_PTR)
6950 parent_type = TYPE_TARGET_TYPE (parent_type);
6951 /* If there is a parallel XVS type, get the actual base type. */
6952 parent_type = ada_get_base_type (parent_type);
6953
6954 return ada_check_typedef (parent_type);
6955 }
6956
6957 return NULL;
6958 }
6959
6960 /* True iff field number FIELD_NUM of structure type TYPE contains the
6961 parent-type (inherited) fields of a derived type. Assumes TYPE is
6962 a structure type with at least FIELD_NUM+1 fields. */
6963
6964 int
6965 ada_is_parent_field (struct type *type, int field_num)
6966 {
6967 const char *name = TYPE_FIELD_NAME (ada_check_typedef (type), field_num);
6968
6969 return (name != NULL
6970 && (startswith (name, "PARENT")
6971 || startswith (name, "_parent")));
6972 }
6973
6974 /* True iff field number FIELD_NUM of structure type TYPE is a
6975 transparent wrapper field (which should be silently traversed when doing
6976 field selection and flattened when printing). Assumes TYPE is a
6977 structure type with at least FIELD_NUM+1 fields. Such fields are always
6978 structures. */
6979
6980 int
6981 ada_is_wrapper_field (struct type *type, int field_num)
6982 {
6983 const char *name = TYPE_FIELD_NAME (type, field_num);
6984
6985 return (name != NULL
6986 && (startswith (name, "PARENT")
6987 || strcmp (name, "REP") == 0
6988 || startswith (name, "_parent")
6989 || name[0] == 'S' || name[0] == 'R' || name[0] == 'O'));
6990 }
6991
6992 /* True iff field number FIELD_NUM of structure or union type TYPE
6993 is a variant wrapper. Assumes TYPE is a structure type with at least
6994 FIELD_NUM+1 fields. */
6995
6996 int
6997 ada_is_variant_part (struct type *type, int field_num)
6998 {
6999 struct type *field_type = TYPE_FIELD_TYPE (type, field_num);
7000
7001 return (TYPE_CODE (field_type) == TYPE_CODE_UNION
7002 || (is_dynamic_field (type, field_num)
7003 && (TYPE_CODE (TYPE_TARGET_TYPE (field_type))
7004 == TYPE_CODE_UNION)));
7005 }
7006
7007 /* Assuming that VAR_TYPE is a variant wrapper (type of the variant part)
7008 whose discriminants are contained in the record type OUTER_TYPE,
7009 returns the type of the controlling discriminant for the variant.
7010 May return NULL if the type could not be found. */
7011
7012 struct type *
7013 ada_variant_discrim_type (struct type *var_type, struct type *outer_type)
7014 {
7015 char *name = ada_variant_discrim_name (var_type);
7016
7017 return ada_lookup_struct_elt_type (outer_type, name, 1, 1, NULL);
7018 }
7019
7020 /* Assuming that TYPE is the type of a variant wrapper, and FIELD_NUM is a
7021 valid field number within it, returns 1 iff field FIELD_NUM of TYPE
7022 represents a 'when others' clause; otherwise 0. */
7023
7024 int
7025 ada_is_others_clause (struct type *type, int field_num)
7026 {
7027 const char *name = TYPE_FIELD_NAME (type, field_num);
7028
7029 return (name != NULL && name[0] == 'O');
7030 }
7031
7032 /* Assuming that TYPE0 is the type of the variant part of a record,
7033 returns the name of the discriminant controlling the variant.
7034 The value is valid until the next call to ada_variant_discrim_name. */
7035
7036 char *
7037 ada_variant_discrim_name (struct type *type0)
7038 {
7039 static char *result = NULL;
7040 static size_t result_len = 0;
7041 struct type *type;
7042 const char *name;
7043 const char *discrim_end;
7044 const char *discrim_start;
7045
7046 if (TYPE_CODE (type0) == TYPE_CODE_PTR)
7047 type = TYPE_TARGET_TYPE (type0);
7048 else
7049 type = type0;
7050
7051 name = ada_type_name (type);
7052
7053 if (name == NULL || name[0] == '\000')
7054 return "";
7055
7056 for (discrim_end = name + strlen (name) - 6; discrim_end != name;
7057 discrim_end -= 1)
7058 {
7059 if (startswith (discrim_end, "___XVN"))
7060 break;
7061 }
7062 if (discrim_end == name)
7063 return "";
7064
7065 for (discrim_start = discrim_end; discrim_start != name + 3;
7066 discrim_start -= 1)
7067 {
7068 if (discrim_start == name + 1)
7069 return "";
7070 if ((discrim_start > name + 3
7071 && startswith (discrim_start - 3, "___"))
7072 || discrim_start[-1] == '.')
7073 break;
7074 }
7075
7076 GROW_VECT (result, result_len, discrim_end - discrim_start + 1);
7077 strncpy (result, discrim_start, discrim_end - discrim_start);
7078 result[discrim_end - discrim_start] = '\0';
7079 return result;
7080 }
7081
7082 /* Scan STR for a subtype-encoded number, beginning at position K.
7083 Put the position of the character just past the number scanned in
7084 *NEW_K, if NEW_K!=NULL. Put the scanned number in *R, if R!=NULL.
7085 Return 1 if there was a valid number at the given position, and 0
7086 otherwise. A "subtype-encoded" number consists of the absolute value
7087 in decimal, followed by the letter 'm' to indicate a negative number.
7088 Assumes 0m does not occur. */
7089
7090 int
7091 ada_scan_number (const char str[], int k, LONGEST * R, int *new_k)
7092 {
7093 ULONGEST RU;
7094
7095 if (!isdigit (str[k]))
7096 return 0;
7097
7098 /* Do it the hard way so as not to make any assumption about
7099 the relationship of unsigned long (%lu scan format code) and
7100 LONGEST. */
7101 RU = 0;
7102 while (isdigit (str[k]))
7103 {
7104 RU = RU * 10 + (str[k] - '0');
7105 k += 1;
7106 }
7107
7108 if (str[k] == 'm')
7109 {
7110 if (R != NULL)
7111 *R = (-(LONGEST) (RU - 1)) - 1;
7112 k += 1;
7113 }
7114 else if (R != NULL)
7115 *R = (LONGEST) RU;
7116
7117 /* NOTE on the above: Technically, C does not say what the results of
7118 - (LONGEST) RU or (LONGEST) -RU are for RU == largest positive
7119 number representable as a LONGEST (although either would probably work
7120 in most implementations). When RU>0, the locution in the then branch
7121 above is always equivalent to the negative of RU. */
7122
7123 if (new_k != NULL)
7124 *new_k = k;
7125 return 1;
7126 }
7127
7128 /* Assuming that TYPE is a variant part wrapper type (a VARIANTS field),
7129 and FIELD_NUM is a valid field number within it, returns 1 iff VAL is
7130 in the range encoded by field FIELD_NUM of TYPE; otherwise 0. */
7131
7132 int
7133 ada_in_variant (LONGEST val, struct type *type, int field_num)
7134 {
7135 const char *name = TYPE_FIELD_NAME (type, field_num);
7136 int p;
7137
7138 p = 0;
7139 while (1)
7140 {
7141 switch (name[p])
7142 {
7143 case '\0':
7144 return 0;
7145 case 'S':
7146 {
7147 LONGEST W;
7148
7149 if (!ada_scan_number (name, p + 1, &W, &p))
7150 return 0;
7151 if (val == W)
7152 return 1;
7153 break;
7154 }
7155 case 'R':
7156 {
7157 LONGEST L, U;
7158
7159 if (!ada_scan_number (name, p + 1, &L, &p)
7160 || name[p] != 'T' || !ada_scan_number (name, p + 1, &U, &p))
7161 return 0;
7162 if (val >= L && val <= U)
7163 return 1;
7164 break;
7165 }
7166 case 'O':
7167 return 1;
7168 default:
7169 return 0;
7170 }
7171 }
7172 }
7173
7174 /* FIXME: Lots of redundancy below. Try to consolidate. */
7175
7176 /* Given a value ARG1 (offset by OFFSET bytes) of a struct or union type
7177 ARG_TYPE, extract and return the value of one of its (non-static)
7178 fields. FIELDNO says which field. Differs from value_primitive_field
7179 only in that it can handle packed values of arbitrary type. */
7180
7181 static struct value *
7182 ada_value_primitive_field (struct value *arg1, int offset, int fieldno,
7183 struct type *arg_type)
7184 {
7185 struct type *type;
7186
7187 arg_type = ada_check_typedef (arg_type);
7188 type = TYPE_FIELD_TYPE (arg_type, fieldno);
7189
7190 /* Handle packed fields. */
7191
7192 if (TYPE_FIELD_BITSIZE (arg_type, fieldno) != 0)
7193 {
7194 int bit_pos = TYPE_FIELD_BITPOS (arg_type, fieldno);
7195 int bit_size = TYPE_FIELD_BITSIZE (arg_type, fieldno);
7196
7197 return ada_value_primitive_packed_val (arg1, value_contents (arg1),
7198 offset + bit_pos / 8,
7199 bit_pos % 8, bit_size, type);
7200 }
7201 else
7202 return value_primitive_field (arg1, offset, fieldno, arg_type);
7203 }
7204
7205 /* Find field with name NAME in object of type TYPE. If found,
7206 set the following for each argument that is non-null:
7207 - *FIELD_TYPE_P to the field's type;
7208 - *BYTE_OFFSET_P to OFFSET + the byte offset of the field within
7209 an object of that type;
7210 - *BIT_OFFSET_P to the bit offset modulo byte size of the field;
7211 - *BIT_SIZE_P to its size in bits if the field is packed, and
7212 0 otherwise;
7213 If INDEX_P is non-null, increment *INDEX_P by the number of source-visible
7214 fields up to but not including the desired field, or by the total
7215 number of fields if not found. A NULL value of NAME never
7216 matches; the function just counts visible fields in this case.
7217
7218 Returns 1 if found, 0 otherwise. */
7219
7220 static int
7221 find_struct_field (const char *name, struct type *type, int offset,
7222 struct type **field_type_p,
7223 int *byte_offset_p, int *bit_offset_p, int *bit_size_p,
7224 int *index_p)
7225 {
7226 int i;
7227
7228 type = ada_check_typedef (type);
7229
7230 if (field_type_p != NULL)
7231 *field_type_p = NULL;
7232 if (byte_offset_p != NULL)
7233 *byte_offset_p = 0;
7234 if (bit_offset_p != NULL)
7235 *bit_offset_p = 0;
7236 if (bit_size_p != NULL)
7237 *bit_size_p = 0;
7238
7239 for (i = 0; i < TYPE_NFIELDS (type); i += 1)
7240 {
7241 int bit_pos = TYPE_FIELD_BITPOS (type, i);
7242 int fld_offset = offset + bit_pos / 8;
7243 const char *t_field_name = TYPE_FIELD_NAME (type, i);
7244
7245 if (t_field_name == NULL)
7246 continue;
7247
7248 else if (name != NULL && field_name_match (t_field_name, name))
7249 {
7250 int bit_size = TYPE_FIELD_BITSIZE (type, i);
7251
7252 if (field_type_p != NULL)
7253 *field_type_p = TYPE_FIELD_TYPE (type, i);
7254 if (byte_offset_p != NULL)
7255 *byte_offset_p = fld_offset;
7256 if (bit_offset_p != NULL)
7257 *bit_offset_p = bit_pos % 8;
7258 if (bit_size_p != NULL)
7259 *bit_size_p = bit_size;
7260 return 1;
7261 }
7262 else if (ada_is_wrapper_field (type, i))
7263 {
7264 if (find_struct_field (name, TYPE_FIELD_TYPE (type, i), fld_offset,
7265 field_type_p, byte_offset_p, bit_offset_p,
7266 bit_size_p, index_p))
7267 return 1;
7268 }
7269 else if (ada_is_variant_part (type, i))
7270 {
7271 /* PNH: Wait. Do we ever execute this section, or is ARG always of
7272 fixed type?? */
7273 int j;
7274 struct type *field_type
7275 = ada_check_typedef (TYPE_FIELD_TYPE (type, i));
7276
7277 for (j = 0; j < TYPE_NFIELDS (field_type); j += 1)
7278 {
7279 if (find_struct_field (name, TYPE_FIELD_TYPE (field_type, j),
7280 fld_offset
7281 + TYPE_FIELD_BITPOS (field_type, j) / 8,
7282 field_type_p, byte_offset_p,
7283 bit_offset_p, bit_size_p, index_p))
7284 return 1;
7285 }
7286 }
7287 else if (index_p != NULL)
7288 *index_p += 1;
7289 }
7290 return 0;
7291 }
7292
7293 /* Number of user-visible fields in record type TYPE. */
7294
7295 static int
7296 num_visible_fields (struct type *type)
7297 {
7298 int n;
7299
7300 n = 0;
7301 find_struct_field (NULL, type, 0, NULL, NULL, NULL, NULL, &n);
7302 return n;
7303 }
7304
7305 /* Look for a field NAME in ARG. Adjust the address of ARG by OFFSET bytes,
7306 and search in it assuming it has (class) type TYPE.
7307 If found, return value, else return NULL.
7308
7309 Searches recursively through wrapper fields (e.g., '_parent'). */
7310
7311 static struct value *
7312 ada_search_struct_field (const char *name, struct value *arg, int offset,
7313 struct type *type)
7314 {
7315 int i;
7316
7317 type = ada_check_typedef (type);
7318 for (i = 0; i < TYPE_NFIELDS (type); i += 1)
7319 {
7320 const char *t_field_name = TYPE_FIELD_NAME (type, i);
7321
7322 if (t_field_name == NULL)
7323 continue;
7324
7325 else if (field_name_match (t_field_name, name))
7326 return ada_value_primitive_field (arg, offset, i, type);
7327
7328 else if (ada_is_wrapper_field (type, i))
7329 {
7330 struct value *v = /* Do not let indent join lines here. */
7331 ada_search_struct_field (name, arg,
7332 offset + TYPE_FIELD_BITPOS (type, i) / 8,
7333 TYPE_FIELD_TYPE (type, i));
7334
7335 if (v != NULL)
7336 return v;
7337 }
7338
7339 else if (ada_is_variant_part (type, i))
7340 {
7341 /* PNH: Do we ever get here? See find_struct_field. */
7342 int j;
7343 struct type *field_type = ada_check_typedef (TYPE_FIELD_TYPE (type,
7344 i));
7345 int var_offset = offset + TYPE_FIELD_BITPOS (type, i) / 8;
7346
7347 for (j = 0; j < TYPE_NFIELDS (field_type); j += 1)
7348 {
7349 struct value *v = ada_search_struct_field /* Force line
7350 break. */
7351 (name, arg,
7352 var_offset + TYPE_FIELD_BITPOS (field_type, j) / 8,
7353 TYPE_FIELD_TYPE (field_type, j));
7354
7355 if (v != NULL)
7356 return v;
7357 }
7358 }
7359 }
7360 return NULL;
7361 }
7362
7363 static struct value *ada_index_struct_field_1 (int *, struct value *,
7364 int, struct type *);
7365
7366
7367 /* Return field #INDEX in ARG, where the index is that returned by
7368 * find_struct_field through its INDEX_P argument. Adjust the address
7369 * of ARG by OFFSET bytes, and search in it assuming it has (class) type TYPE.
7370 * If found, return value, else return NULL. */
7371
7372 static struct value *
7373 ada_index_struct_field (int index, struct value *arg, int offset,
7374 struct type *type)
7375 {
7376 return ada_index_struct_field_1 (&index, arg, offset, type);
7377 }
7378
7379
7380 /* Auxiliary function for ada_index_struct_field. Like
7381 * ada_index_struct_field, but takes index from *INDEX_P and modifies
7382 * *INDEX_P. */
7383
7384 static struct value *
7385 ada_index_struct_field_1 (int *index_p, struct value *arg, int offset,
7386 struct type *type)
7387 {
7388 int i;
7389 type = ada_check_typedef (type);
7390
7391 for (i = 0; i < TYPE_NFIELDS (type); i += 1)
7392 {
7393 if (TYPE_FIELD_NAME (type, i) == NULL)
7394 continue;
7395 else if (ada_is_wrapper_field (type, i))
7396 {
7397 struct value *v = /* Do not let indent join lines here. */
7398 ada_index_struct_field_1 (index_p, arg,
7399 offset + TYPE_FIELD_BITPOS (type, i) / 8,
7400 TYPE_FIELD_TYPE (type, i));
7401
7402 if (v != NULL)
7403 return v;
7404 }
7405
7406 else if (ada_is_variant_part (type, i))
7407 {
7408 /* PNH: Do we ever get here? See ada_search_struct_field,
7409 find_struct_field. */
7410 error (_("Cannot assign this kind of variant record"));
7411 }
7412 else if (*index_p == 0)
7413 return ada_value_primitive_field (arg, offset, i, type);
7414 else
7415 *index_p -= 1;
7416 }
7417 return NULL;
7418 }
7419
7420 /* Given ARG, a value of type (pointer or reference to a)*
7421 structure/union, extract the component named NAME from the ultimate
7422 target structure/union and return it as a value with its
7423 appropriate type.
7424
7425 The routine searches for NAME among all members of the structure itself
7426 and (recursively) among all members of any wrapper members
7427 (e.g., '_parent').
7428
7429 If NO_ERR, then simply return NULL in case of error, rather than
7430 calling error. */
7431
7432 struct value *
7433 ada_value_struct_elt (struct value *arg, char *name, int no_err)
7434 {
7435 struct type *t, *t1;
7436 struct value *v;
7437
7438 v = NULL;
7439 t1 = t = ada_check_typedef (value_type (arg));
7440 if (TYPE_CODE (t) == TYPE_CODE_REF)
7441 {
7442 t1 = TYPE_TARGET_TYPE (t);
7443 if (t1 == NULL)
7444 goto BadValue;
7445 t1 = ada_check_typedef (t1);
7446 if (TYPE_CODE (t1) == TYPE_CODE_PTR)
7447 {
7448 arg = coerce_ref (arg);
7449 t = t1;
7450 }
7451 }
7452
7453 while (TYPE_CODE (t) == TYPE_CODE_PTR)
7454 {
7455 t1 = TYPE_TARGET_TYPE (t);
7456 if (t1 == NULL)
7457 goto BadValue;
7458 t1 = ada_check_typedef (t1);
7459 if (TYPE_CODE (t1) == TYPE_CODE_PTR)
7460 {
7461 arg = value_ind (arg);
7462 t = t1;
7463 }
7464 else
7465 break;
7466 }
7467
7468 if (TYPE_CODE (t1) != TYPE_CODE_STRUCT && TYPE_CODE (t1) != TYPE_CODE_UNION)
7469 goto BadValue;
7470
7471 if (t1 == t)
7472 v = ada_search_struct_field (name, arg, 0, t);
7473 else
7474 {
7475 int bit_offset, bit_size, byte_offset;
7476 struct type *field_type;
7477 CORE_ADDR address;
7478
7479 if (TYPE_CODE (t) == TYPE_CODE_PTR)
7480 address = value_address (ada_value_ind (arg));
7481 else
7482 address = value_address (ada_coerce_ref (arg));
7483
7484 t1 = ada_to_fixed_type (ada_get_base_type (t1), NULL, address, NULL, 1);
7485 if (find_struct_field (name, t1, 0,
7486 &field_type, &byte_offset, &bit_offset,
7487 &bit_size, NULL))
7488 {
7489 if (bit_size != 0)
7490 {
7491 if (TYPE_CODE (t) == TYPE_CODE_REF)
7492 arg = ada_coerce_ref (arg);
7493 else
7494 arg = ada_value_ind (arg);
7495 v = ada_value_primitive_packed_val (arg, NULL, byte_offset,
7496 bit_offset, bit_size,
7497 field_type);
7498 }
7499 else
7500 v = value_at_lazy (field_type, address + byte_offset);
7501 }
7502 }
7503
7504 if (v != NULL || no_err)
7505 return v;
7506 else
7507 error (_("There is no member named %s."), name);
7508
7509 BadValue:
7510 if (no_err)
7511 return NULL;
7512 else
7513 error (_("Attempt to extract a component of "
7514 "a value that is not a record."));
7515 }
7516
7517 /* Given a type TYPE, look up the type of the component of type named NAME.
7518 If DISPP is non-null, add its byte displacement from the beginning of a
7519 structure (pointed to by a value) of type TYPE to *DISPP (does not
7520 work for packed fields).
7521
7522 Matches any field whose name has NAME as a prefix, possibly
7523 followed by "___".
7524
7525 TYPE can be either a struct or union. If REFOK, TYPE may also
7526 be a (pointer or reference)+ to a struct or union, and the
7527 ultimate target type will be searched.
7528
7529 Looks recursively into variant clauses and parent types.
7530
7531 If NOERR is nonzero, return NULL if NAME is not suitably defined or
7532 TYPE is not a type of the right kind. */
7533
7534 static struct type *
7535 ada_lookup_struct_elt_type (struct type *type, char *name, int refok,
7536 int noerr, int *dispp)
7537 {
7538 int i;
7539
7540 if (name == NULL)
7541 goto BadName;
7542
7543 if (refok && type != NULL)
7544 while (1)
7545 {
7546 type = ada_check_typedef (type);
7547 if (TYPE_CODE (type) != TYPE_CODE_PTR
7548 && TYPE_CODE (type) != TYPE_CODE_REF)
7549 break;
7550 type = TYPE_TARGET_TYPE (type);
7551 }
7552
7553 if (type == NULL
7554 || (TYPE_CODE (type) != TYPE_CODE_STRUCT
7555 && TYPE_CODE (type) != TYPE_CODE_UNION))
7556 {
7557 if (noerr)
7558 return NULL;
7559 else
7560 {
7561 target_terminal_ours ();
7562 gdb_flush (gdb_stdout);
7563 if (type == NULL)
7564 error (_("Type (null) is not a structure or union type"));
7565 else
7566 {
7567 /* XXX: type_sprint */
7568 fprintf_unfiltered (gdb_stderr, _("Type "));
7569 type_print (type, "", gdb_stderr, -1);
7570 error (_(" is not a structure or union type"));
7571 }
7572 }
7573 }
7574
7575 type = to_static_fixed_type (type);
7576
7577 for (i = 0; i < TYPE_NFIELDS (type); i += 1)
7578 {
7579 const char *t_field_name = TYPE_FIELD_NAME (type, i);
7580 struct type *t;
7581 int disp;
7582
7583 if (t_field_name == NULL)
7584 continue;
7585
7586 else if (field_name_match (t_field_name, name))
7587 {
7588 if (dispp != NULL)
7589 *dispp += TYPE_FIELD_BITPOS (type, i) / 8;
7590 return TYPE_FIELD_TYPE (type, i);
7591 }
7592
7593 else if (ada_is_wrapper_field (type, i))
7594 {
7595 disp = 0;
7596 t = ada_lookup_struct_elt_type (TYPE_FIELD_TYPE (type, i), name,
7597 0, 1, &disp);
7598 if (t != NULL)
7599 {
7600 if (dispp != NULL)
7601 *dispp += disp + TYPE_FIELD_BITPOS (type, i) / 8;
7602 return t;
7603 }
7604 }
7605
7606 else if (ada_is_variant_part (type, i))
7607 {
7608 int j;
7609 struct type *field_type = ada_check_typedef (TYPE_FIELD_TYPE (type,
7610 i));
7611
7612 for (j = TYPE_NFIELDS (field_type) - 1; j >= 0; j -= 1)
7613 {
7614 /* FIXME pnh 2008/01/26: We check for a field that is
7615 NOT wrapped in a struct, since the compiler sometimes
7616 generates these for unchecked variant types. Revisit
7617 if the compiler changes this practice. */
7618 const char *v_field_name = TYPE_FIELD_NAME (field_type, j);
7619 disp = 0;
7620 if (v_field_name != NULL
7621 && field_name_match (v_field_name, name))
7622 t = TYPE_FIELD_TYPE (field_type, j);
7623 else
7624 t = ada_lookup_struct_elt_type (TYPE_FIELD_TYPE (field_type,
7625 j),
7626 name, 0, 1, &disp);
7627
7628 if (t != NULL)
7629 {
7630 if (dispp != NULL)
7631 *dispp += disp + TYPE_FIELD_BITPOS (type, i) / 8;
7632 return t;
7633 }
7634 }
7635 }
7636
7637 }
7638
7639 BadName:
7640 if (!noerr)
7641 {
7642 target_terminal_ours ();
7643 gdb_flush (gdb_stdout);
7644 if (name == NULL)
7645 {
7646 /* XXX: type_sprint */
7647 fprintf_unfiltered (gdb_stderr, _("Type "));
7648 type_print (type, "", gdb_stderr, -1);
7649 error (_(" has no component named <null>"));
7650 }
7651 else
7652 {
7653 /* XXX: type_sprint */
7654 fprintf_unfiltered (gdb_stderr, _("Type "));
7655 type_print (type, "", gdb_stderr, -1);
7656 error (_(" has no component named %s"), name);
7657 }
7658 }
7659
7660 return NULL;
7661 }
7662
7663 /* Assuming that VAR_TYPE is the type of a variant part of a record (a union),
7664 within a value of type OUTER_TYPE, return true iff VAR_TYPE
7665 represents an unchecked union (that is, the variant part of a
7666 record that is named in an Unchecked_Union pragma). */
7667
7668 static int
7669 is_unchecked_variant (struct type *var_type, struct type *outer_type)
7670 {
7671 char *discrim_name = ada_variant_discrim_name (var_type);
7672
7673 return (ada_lookup_struct_elt_type (outer_type, discrim_name, 0, 1, NULL)
7674 == NULL);
7675 }
7676
7677
7678 /* Assuming that VAR_TYPE is the type of a variant part of a record (a union),
7679 within a value of type OUTER_TYPE that is stored in GDB at
7680 OUTER_VALADDR, determine which variant clause (field number in VAR_TYPE,
7681 numbering from 0) is applicable. Returns -1 if none are. */
7682
7683 int
7684 ada_which_variant_applies (struct type *var_type, struct type *outer_type,
7685 const gdb_byte *outer_valaddr)
7686 {
7687 int others_clause;
7688 int i;
7689 char *discrim_name = ada_variant_discrim_name (var_type);
7690 struct value *outer;
7691 struct value *discrim;
7692 LONGEST discrim_val;
7693
7694 /* Using plain value_from_contents_and_address here causes problems
7695 because we will end up trying to resolve a type that is currently
7696 being constructed. */
7697 outer = value_from_contents_and_address_unresolved (outer_type,
7698 outer_valaddr, 0);
7699 discrim = ada_value_struct_elt (outer, discrim_name, 1);
7700 if (discrim == NULL)
7701 return -1;
7702 discrim_val = value_as_long (discrim);
7703
7704 others_clause = -1;
7705 for (i = 0; i < TYPE_NFIELDS (var_type); i += 1)
7706 {
7707 if (ada_is_others_clause (var_type, i))
7708 others_clause = i;
7709 else if (ada_in_variant (discrim_val, var_type, i))
7710 return i;
7711 }
7712
7713 return others_clause;
7714 }
7715 \f
7716
7717
7718 /* Dynamic-Sized Records */
7719
7720 /* Strategy: The type ostensibly attached to a value with dynamic size
7721 (i.e., a size that is not statically recorded in the debugging
7722 data) does not accurately reflect the size or layout of the value.
7723 Our strategy is to convert these values to values with accurate,
7724 conventional types that are constructed on the fly. */
7725
7726 /* There is a subtle and tricky problem here. In general, we cannot
7727 determine the size of dynamic records without its data. However,
7728 the 'struct value' data structure, which GDB uses to represent
7729 quantities in the inferior process (the target), requires the size
7730 of the type at the time of its allocation in order to reserve space
7731 for GDB's internal copy of the data. That's why the
7732 'to_fixed_xxx_type' routines take (target) addresses as parameters,
7733 rather than struct value*s.
7734
7735 However, GDB's internal history variables ($1, $2, etc.) are
7736 struct value*s containing internal copies of the data that are not, in
7737 general, the same as the data at their corresponding addresses in
7738 the target. Fortunately, the types we give to these values are all
7739 conventional, fixed-size types (as per the strategy described
7740 above), so that we don't usually have to perform the
7741 'to_fixed_xxx_type' conversions to look at their values.
7742 Unfortunately, there is one exception: if one of the internal
7743 history variables is an array whose elements are unconstrained
7744 records, then we will need to create distinct fixed types for each
7745 element selected. */
7746
7747 /* The upshot of all of this is that many routines take a (type, host
7748 address, target address) triple as arguments to represent a value.
7749 The host address, if non-null, is supposed to contain an internal
7750 copy of the relevant data; otherwise, the program is to consult the
7751 target at the target address. */
7752
7753 /* Assuming that VAL0 represents a pointer value, the result of
7754 dereferencing it. Differs from value_ind in its treatment of
7755 dynamic-sized types. */
7756
7757 struct value *
7758 ada_value_ind (struct value *val0)
7759 {
7760 struct value *val = value_ind (val0);
7761
7762 if (ada_is_tagged_type (value_type (val), 0))
7763 val = ada_tag_value_at_base_address (val);
7764
7765 return ada_to_fixed_value (val);
7766 }
7767
7768 /* The value resulting from dereferencing any "reference to"
7769 qualifiers on VAL0. */
7770
7771 static struct value *
7772 ada_coerce_ref (struct value *val0)
7773 {
7774 if (TYPE_CODE (value_type (val0)) == TYPE_CODE_REF)
7775 {
7776 struct value *val = val0;
7777
7778 val = coerce_ref (val);
7779
7780 if (ada_is_tagged_type (value_type (val), 0))
7781 val = ada_tag_value_at_base_address (val);
7782
7783 return ada_to_fixed_value (val);
7784 }
7785 else
7786 return val0;
7787 }
7788
7789 /* Return OFF rounded upward if necessary to a multiple of
7790 ALIGNMENT (a power of 2). */
7791
7792 static unsigned int
7793 align_value (unsigned int off, unsigned int alignment)
7794 {
7795 return (off + alignment - 1) & ~(alignment - 1);
7796 }
7797
7798 /* Return the bit alignment required for field #F of template type TYPE. */
7799
7800 static unsigned int
7801 field_alignment (struct type *type, int f)
7802 {
7803 const char *name = TYPE_FIELD_NAME (type, f);
7804 int len;
7805 int align_offset;
7806
7807 /* The field name should never be null, unless the debugging information
7808 is somehow malformed. In this case, we assume the field does not
7809 require any alignment. */
7810 if (name == NULL)
7811 return 1;
7812
7813 len = strlen (name);
7814
7815 if (!isdigit (name[len - 1]))
7816 return 1;
7817
7818 if (isdigit (name[len - 2]))
7819 align_offset = len - 2;
7820 else
7821 align_offset = len - 1;
7822
7823 if (align_offset < 7 || !startswith (name + align_offset - 6, "___XV"))
7824 return TARGET_CHAR_BIT;
7825
7826 return atoi (name + align_offset) * TARGET_CHAR_BIT;
7827 }
7828
7829 /* Find a typedef or tag symbol named NAME. Ignores ambiguity. */
7830
7831 static struct symbol *
7832 ada_find_any_type_symbol (const char *name)
7833 {
7834 struct symbol *sym;
7835
7836 sym = standard_lookup (name, get_selected_block (NULL), VAR_DOMAIN);
7837 if (sym != NULL && SYMBOL_CLASS (sym) == LOC_TYPEDEF)
7838 return sym;
7839
7840 sym = standard_lookup (name, NULL, STRUCT_DOMAIN);
7841 return sym;
7842 }
7843
7844 /* Find a type named NAME. Ignores ambiguity. This routine will look
7845 solely for types defined by debug info, it will not search the GDB
7846 primitive types. */
7847
7848 static struct type *
7849 ada_find_any_type (const char *name)
7850 {
7851 struct symbol *sym = ada_find_any_type_symbol (name);
7852
7853 if (sym != NULL)
7854 return SYMBOL_TYPE (sym);
7855
7856 return NULL;
7857 }
7858
7859 /* Given NAME_SYM and an associated BLOCK, find a "renaming" symbol
7860 associated with NAME_SYM's name. NAME_SYM may itself be a renaming
7861 symbol, in which case it is returned. Otherwise, this looks for
7862 symbols whose name is that of NAME_SYM suffixed with "___XR".
7863 Return symbol if found, and NULL otherwise. */
7864
7865 struct symbol *
7866 ada_find_renaming_symbol (struct symbol *name_sym, const struct block *block)
7867 {
7868 const char *name = SYMBOL_LINKAGE_NAME (name_sym);
7869 struct symbol *sym;
7870
7871 if (strstr (name, "___XR") != NULL)
7872 return name_sym;
7873
7874 sym = find_old_style_renaming_symbol (name, block);
7875
7876 if (sym != NULL)
7877 return sym;
7878
7879 /* Not right yet. FIXME pnh 7/20/2007. */
7880 sym = ada_find_any_type_symbol (name);
7881 if (sym != NULL && strstr (SYMBOL_LINKAGE_NAME (sym), "___XR") != NULL)
7882 return sym;
7883 else
7884 return NULL;
7885 }
7886
7887 static struct symbol *
7888 find_old_style_renaming_symbol (const char *name, const struct block *block)
7889 {
7890 const struct symbol *function_sym = block_linkage_function (block);
7891 char *rename;
7892
7893 if (function_sym != NULL)
7894 {
7895 /* If the symbol is defined inside a function, NAME is not fully
7896 qualified. This means we need to prepend the function name
7897 as well as adding the ``___XR'' suffix to build the name of
7898 the associated renaming symbol. */
7899 const char *function_name = SYMBOL_LINKAGE_NAME (function_sym);
7900 /* Function names sometimes contain suffixes used
7901 for instance to qualify nested subprograms. When building
7902 the XR type name, we need to make sure that this suffix is
7903 not included. So do not include any suffix in the function
7904 name length below. */
7905 int function_name_len = ada_name_prefix_len (function_name);
7906 const int rename_len = function_name_len + 2 /* "__" */
7907 + strlen (name) + 6 /* "___XR\0" */ ;
7908
7909 /* Strip the suffix if necessary. */
7910 ada_remove_trailing_digits (function_name, &function_name_len);
7911 ada_remove_po_subprogram_suffix (function_name, &function_name_len);
7912 ada_remove_Xbn_suffix (function_name, &function_name_len);
7913
7914 /* Library-level functions are a special case, as GNAT adds
7915 a ``_ada_'' prefix to the function name to avoid namespace
7916 pollution. However, the renaming symbols themselves do not
7917 have this prefix, so we need to skip this prefix if present. */
7918 if (function_name_len > 5 /* "_ada_" */
7919 && strstr (function_name, "_ada_") == function_name)
7920 {
7921 function_name += 5;
7922 function_name_len -= 5;
7923 }
7924
7925 rename = (char *) alloca (rename_len * sizeof (char));
7926 strncpy (rename, function_name, function_name_len);
7927 xsnprintf (rename + function_name_len, rename_len - function_name_len,
7928 "__%s___XR", name);
7929 }
7930 else
7931 {
7932 const int rename_len = strlen (name) + 6;
7933
7934 rename = (char *) alloca (rename_len * sizeof (char));
7935 xsnprintf (rename, rename_len * sizeof (char), "%s___XR", name);
7936 }
7937
7938 return ada_find_any_type_symbol (rename);
7939 }
7940
7941 /* Because of GNAT encoding conventions, several GDB symbols may match a
7942 given type name. If the type denoted by TYPE0 is to be preferred to
7943 that of TYPE1 for purposes of type printing, return non-zero;
7944 otherwise return 0. */
7945
7946 int
7947 ada_prefer_type (struct type *type0, struct type *type1)
7948 {
7949 if (type1 == NULL)
7950 return 1;
7951 else if (type0 == NULL)
7952 return 0;
7953 else if (TYPE_CODE (type1) == TYPE_CODE_VOID)
7954 return 1;
7955 else if (TYPE_CODE (type0) == TYPE_CODE_VOID)
7956 return 0;
7957 else if (TYPE_NAME (type1) == NULL && TYPE_NAME (type0) != NULL)
7958 return 1;
7959 else if (ada_is_constrained_packed_array_type (type0))
7960 return 1;
7961 else if (ada_is_array_descriptor_type (type0)
7962 && !ada_is_array_descriptor_type (type1))
7963 return 1;
7964 else
7965 {
7966 const char *type0_name = type_name_no_tag (type0);
7967 const char *type1_name = type_name_no_tag (type1);
7968
7969 if (type0_name != NULL && strstr (type0_name, "___XR") != NULL
7970 && (type1_name == NULL || strstr (type1_name, "___XR") == NULL))
7971 return 1;
7972 }
7973 return 0;
7974 }
7975
7976 /* The name of TYPE, which is either its TYPE_NAME, or, if that is
7977 null, its TYPE_TAG_NAME. Null if TYPE is null. */
7978
7979 const char *
7980 ada_type_name (struct type *type)
7981 {
7982 if (type == NULL)
7983 return NULL;
7984 else if (TYPE_NAME (type) != NULL)
7985 return TYPE_NAME (type);
7986 else
7987 return TYPE_TAG_NAME (type);
7988 }
7989
7990 /* Search the list of "descriptive" types associated to TYPE for a type
7991 whose name is NAME. */
7992
7993 static struct type *
7994 find_parallel_type_by_descriptive_type (struct type *type, const char *name)
7995 {
7996 struct type *result, *tmp;
7997
7998 if (ada_ignore_descriptive_types_p)
7999 return NULL;
8000
8001 /* If there no descriptive-type info, then there is no parallel type
8002 to be found. */
8003 if (!HAVE_GNAT_AUX_INFO (type))
8004 return NULL;
8005
8006 result = TYPE_DESCRIPTIVE_TYPE (type);
8007 while (result != NULL)
8008 {
8009 const char *result_name = ada_type_name (result);
8010
8011 if (result_name == NULL)
8012 {
8013 warning (_("unexpected null name on descriptive type"));
8014 return NULL;
8015 }
8016
8017 /* If the names match, stop. */
8018 if (strcmp (result_name, name) == 0)
8019 break;
8020
8021 /* Otherwise, look at the next item on the list, if any. */
8022 if (HAVE_GNAT_AUX_INFO (result))
8023 tmp = TYPE_DESCRIPTIVE_TYPE (result);
8024 else
8025 tmp = NULL;
8026
8027 /* If not found either, try after having resolved the typedef. */
8028 if (tmp != NULL)
8029 result = tmp;
8030 else
8031 {
8032 result = check_typedef (result);
8033 if (HAVE_GNAT_AUX_INFO (result))
8034 result = TYPE_DESCRIPTIVE_TYPE (result);
8035 else
8036 result = NULL;
8037 }
8038 }
8039
8040 /* If we didn't find a match, see whether this is a packed array. With
8041 older compilers, the descriptive type information is either absent or
8042 irrelevant when it comes to packed arrays so the above lookup fails.
8043 Fall back to using a parallel lookup by name in this case. */
8044 if (result == NULL && ada_is_constrained_packed_array_type (type))
8045 return ada_find_any_type (name);
8046
8047 return result;
8048 }
8049
8050 /* Find a parallel type to TYPE with the specified NAME, using the
8051 descriptive type taken from the debugging information, if available,
8052 and otherwise using the (slower) name-based method. */
8053
8054 static struct type *
8055 ada_find_parallel_type_with_name (struct type *type, const char *name)
8056 {
8057 struct type *result = NULL;
8058
8059 if (HAVE_GNAT_AUX_INFO (type))
8060 result = find_parallel_type_by_descriptive_type (type, name);
8061 else
8062 result = ada_find_any_type (name);
8063
8064 return result;
8065 }
8066
8067 /* Same as above, but specify the name of the parallel type by appending
8068 SUFFIX to the name of TYPE. */
8069
8070 struct type *
8071 ada_find_parallel_type (struct type *type, const char *suffix)
8072 {
8073 char *name;
8074 const char *type_name = ada_type_name (type);
8075 int len;
8076
8077 if (type_name == NULL)
8078 return NULL;
8079
8080 len = strlen (type_name);
8081
8082 name = (char *) alloca (len + strlen (suffix) + 1);
8083
8084 strcpy (name, type_name);
8085 strcpy (name + len, suffix);
8086
8087 return ada_find_parallel_type_with_name (type, name);
8088 }
8089
8090 /* If TYPE is a variable-size record type, return the corresponding template
8091 type describing its fields. Otherwise, return NULL. */
8092
8093 static struct type *
8094 dynamic_template_type (struct type *type)
8095 {
8096 type = ada_check_typedef (type);
8097
8098 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_STRUCT
8099 || ada_type_name (type) == NULL)
8100 return NULL;
8101 else
8102 {
8103 int len = strlen (ada_type_name (type));
8104
8105 if (len > 6 && strcmp (ada_type_name (type) + len - 6, "___XVE") == 0)
8106 return type;
8107 else
8108 return ada_find_parallel_type (type, "___XVE");
8109 }
8110 }
8111
8112 /* Assuming that TEMPL_TYPE is a union or struct type, returns
8113 non-zero iff field FIELD_NUM of TEMPL_TYPE has dynamic size. */
8114
8115 static int
8116 is_dynamic_field (struct type *templ_type, int field_num)
8117 {
8118 const char *name = TYPE_FIELD_NAME (templ_type, field_num);
8119
8120 return name != NULL
8121 && TYPE_CODE (TYPE_FIELD_TYPE (templ_type, field_num)) == TYPE_CODE_PTR
8122 && strstr (name, "___XVL") != NULL;
8123 }
8124
8125 /* The index of the variant field of TYPE, or -1 if TYPE does not
8126 represent a variant record type. */
8127
8128 static int
8129 variant_field_index (struct type *type)
8130 {
8131 int f;
8132
8133 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_STRUCT)
8134 return -1;
8135
8136 for (f = 0; f < TYPE_NFIELDS (type); f += 1)
8137 {
8138 if (ada_is_variant_part (type, f))
8139 return f;
8140 }
8141 return -1;
8142 }
8143
8144 /* A record type with no fields. */
8145
8146 static struct type *
8147 empty_record (struct type *templ)
8148 {
8149 struct type *type = alloc_type_copy (templ);
8150
8151 TYPE_CODE (type) = TYPE_CODE_STRUCT;
8152 TYPE_NFIELDS (type) = 0;
8153 TYPE_FIELDS (type) = NULL;
8154 INIT_CPLUS_SPECIFIC (type);
8155 TYPE_NAME (type) = "<empty>";
8156 TYPE_TAG_NAME (type) = NULL;
8157 TYPE_LENGTH (type) = 0;
8158 return type;
8159 }
8160
8161 /* An ordinary record type (with fixed-length fields) that describes
8162 the value of type TYPE at VALADDR or ADDRESS (see comments at
8163 the beginning of this section) VAL according to GNAT conventions.
8164 DVAL0 should describe the (portion of a) record that contains any
8165 necessary discriminants. It should be NULL if value_type (VAL) is
8166 an outer-level type (i.e., as opposed to a branch of a variant.) A
8167 variant field (unless unchecked) is replaced by a particular branch
8168 of the variant.
8169
8170 If not KEEP_DYNAMIC_FIELDS, then all fields whose position or
8171 length are not statically known are discarded. As a consequence,
8172 VALADDR, ADDRESS and DVAL0 are ignored.
8173
8174 NOTE: Limitations: For now, we assume that dynamic fields and
8175 variants occupy whole numbers of bytes. However, they need not be
8176 byte-aligned. */
8177
8178 struct type *
8179 ada_template_to_fixed_record_type_1 (struct type *type,
8180 const gdb_byte *valaddr,
8181 CORE_ADDR address, struct value *dval0,
8182 int keep_dynamic_fields)
8183 {
8184 struct value *mark = value_mark ();
8185 struct value *dval;
8186 struct type *rtype;
8187 int nfields, bit_len;
8188 int variant_field;
8189 long off;
8190 int fld_bit_len;
8191 int f;
8192
8193 /* Compute the number of fields in this record type that are going
8194 to be processed: unless keep_dynamic_fields, this includes only
8195 fields whose position and length are static will be processed. */
8196 if (keep_dynamic_fields)
8197 nfields = TYPE_NFIELDS (type);
8198 else
8199 {
8200 nfields = 0;
8201 while (nfields < TYPE_NFIELDS (type)
8202 && !ada_is_variant_part (type, nfields)
8203 && !is_dynamic_field (type, nfields))
8204 nfields++;
8205 }
8206
8207 rtype = alloc_type_copy (type);
8208 TYPE_CODE (rtype) = TYPE_CODE_STRUCT;
8209 INIT_CPLUS_SPECIFIC (rtype);
8210 TYPE_NFIELDS (rtype) = nfields;
8211 TYPE_FIELDS (rtype) = (struct field *)
8212 TYPE_ALLOC (rtype, nfields * sizeof (struct field));
8213 memset (TYPE_FIELDS (rtype), 0, sizeof (struct field) * nfields);
8214 TYPE_NAME (rtype) = ada_type_name (type);
8215 TYPE_TAG_NAME (rtype) = NULL;
8216 TYPE_FIXED_INSTANCE (rtype) = 1;
8217
8218 off = 0;
8219 bit_len = 0;
8220 variant_field = -1;
8221
8222 for (f = 0; f < nfields; f += 1)
8223 {
8224 off = align_value (off, field_alignment (type, f))
8225 + TYPE_FIELD_BITPOS (type, f);
8226 SET_FIELD_BITPOS (TYPE_FIELD (rtype, f), off);
8227 TYPE_FIELD_BITSIZE (rtype, f) = 0;
8228
8229 if (ada_is_variant_part (type, f))
8230 {
8231 variant_field = f;
8232 fld_bit_len = 0;
8233 }
8234 else if (is_dynamic_field (type, f))
8235 {
8236 const gdb_byte *field_valaddr = valaddr;
8237 CORE_ADDR field_address = address;
8238 struct type *field_type =
8239 TYPE_TARGET_TYPE (TYPE_FIELD_TYPE (type, f));
8240
8241 if (dval0 == NULL)
8242 {
8243 /* rtype's length is computed based on the run-time
8244 value of discriminants. If the discriminants are not
8245 initialized, the type size may be completely bogus and
8246 GDB may fail to allocate a value for it. So check the
8247 size first before creating the value. */
8248 ada_ensure_varsize_limit (rtype);
8249 /* Using plain value_from_contents_and_address here
8250 causes problems because we will end up trying to
8251 resolve a type that is currently being
8252 constructed. */
8253 dval = value_from_contents_and_address_unresolved (rtype,
8254 valaddr,
8255 address);
8256 rtype = value_type (dval);
8257 }
8258 else
8259 dval = dval0;
8260
8261 /* If the type referenced by this field is an aligner type, we need
8262 to unwrap that aligner type, because its size might not be set.
8263 Keeping the aligner type would cause us to compute the wrong
8264 size for this field, impacting the offset of the all the fields
8265 that follow this one. */
8266 if (ada_is_aligner_type (field_type))
8267 {
8268 long field_offset = TYPE_FIELD_BITPOS (field_type, f);
8269
8270 field_valaddr = cond_offset_host (field_valaddr, field_offset);
8271 field_address = cond_offset_target (field_address, field_offset);
8272 field_type = ada_aligned_type (field_type);
8273 }
8274
8275 field_valaddr = cond_offset_host (field_valaddr,
8276 off / TARGET_CHAR_BIT);
8277 field_address = cond_offset_target (field_address,
8278 off / TARGET_CHAR_BIT);
8279
8280 /* Get the fixed type of the field. Note that, in this case,
8281 we do not want to get the real type out of the tag: if
8282 the current field is the parent part of a tagged record,
8283 we will get the tag of the object. Clearly wrong: the real
8284 type of the parent is not the real type of the child. We
8285 would end up in an infinite loop. */
8286 field_type = ada_get_base_type (field_type);
8287 field_type = ada_to_fixed_type (field_type, field_valaddr,
8288 field_address, dval, 0);
8289 /* If the field size is already larger than the maximum
8290 object size, then the record itself will necessarily
8291 be larger than the maximum object size. We need to make
8292 this check now, because the size might be so ridiculously
8293 large (due to an uninitialized variable in the inferior)
8294 that it would cause an overflow when adding it to the
8295 record size. */
8296 ada_ensure_varsize_limit (field_type);
8297
8298 TYPE_FIELD_TYPE (rtype, f) = field_type;
8299 TYPE_FIELD_NAME (rtype, f) = TYPE_FIELD_NAME (type, f);
8300 /* The multiplication can potentially overflow. But because
8301 the field length has been size-checked just above, and
8302 assuming that the maximum size is a reasonable value,
8303 an overflow should not happen in practice. So rather than
8304 adding overflow recovery code to this already complex code,
8305 we just assume that it's not going to happen. */
8306 fld_bit_len =
8307 TYPE_LENGTH (TYPE_FIELD_TYPE (rtype, f)) * TARGET_CHAR_BIT;
8308 }
8309 else
8310 {
8311 /* Note: If this field's type is a typedef, it is important
8312 to preserve the typedef layer.
8313
8314 Otherwise, we might be transforming a typedef to a fat
8315 pointer (encoding a pointer to an unconstrained array),
8316 into a basic fat pointer (encoding an unconstrained
8317 array). As both types are implemented using the same
8318 structure, the typedef is the only clue which allows us
8319 to distinguish between the two options. Stripping it
8320 would prevent us from printing this field appropriately. */
8321 TYPE_FIELD_TYPE (rtype, f) = TYPE_FIELD_TYPE (type, f);
8322 TYPE_FIELD_NAME (rtype, f) = TYPE_FIELD_NAME (type, f);
8323 if (TYPE_FIELD_BITSIZE (type, f) > 0)
8324 fld_bit_len =
8325 TYPE_FIELD_BITSIZE (rtype, f) = TYPE_FIELD_BITSIZE (type, f);
8326 else
8327 {
8328 struct type *field_type = TYPE_FIELD_TYPE (type, f);
8329
8330 /* We need to be careful of typedefs when computing
8331 the length of our field. If this is a typedef,
8332 get the length of the target type, not the length
8333 of the typedef. */
8334 if (TYPE_CODE (field_type) == TYPE_CODE_TYPEDEF)
8335 field_type = ada_typedef_target_type (field_type);
8336
8337 fld_bit_len =
8338 TYPE_LENGTH (ada_check_typedef (field_type)) * TARGET_CHAR_BIT;
8339 }
8340 }
8341 if (off + fld_bit_len > bit_len)
8342 bit_len = off + fld_bit_len;
8343 off += fld_bit_len;
8344 TYPE_LENGTH (rtype) =
8345 align_value (bit_len, TARGET_CHAR_BIT) / TARGET_CHAR_BIT;
8346 }
8347
8348 /* We handle the variant part, if any, at the end because of certain
8349 odd cases in which it is re-ordered so as NOT to be the last field of
8350 the record. This can happen in the presence of representation
8351 clauses. */
8352 if (variant_field >= 0)
8353 {
8354 struct type *branch_type;
8355
8356 off = TYPE_FIELD_BITPOS (rtype, variant_field);
8357
8358 if (dval0 == NULL)
8359 {
8360 /* Using plain value_from_contents_and_address here causes
8361 problems because we will end up trying to resolve a type
8362 that is currently being constructed. */
8363 dval = value_from_contents_and_address_unresolved (rtype, valaddr,
8364 address);
8365 rtype = value_type (dval);
8366 }
8367 else
8368 dval = dval0;
8369
8370 branch_type =
8371 to_fixed_variant_branch_type
8372 (TYPE_FIELD_TYPE (type, variant_field),
8373 cond_offset_host (valaddr, off / TARGET_CHAR_BIT),
8374 cond_offset_target (address, off / TARGET_CHAR_BIT), dval);
8375 if (branch_type == NULL)
8376 {
8377 for (f = variant_field + 1; f < TYPE_NFIELDS (rtype); f += 1)
8378 TYPE_FIELDS (rtype)[f - 1] = TYPE_FIELDS (rtype)[f];
8379 TYPE_NFIELDS (rtype) -= 1;
8380 }
8381 else
8382 {
8383 TYPE_FIELD_TYPE (rtype, variant_field) = branch_type;
8384 TYPE_FIELD_NAME (rtype, variant_field) = "S";
8385 fld_bit_len =
8386 TYPE_LENGTH (TYPE_FIELD_TYPE (rtype, variant_field)) *
8387 TARGET_CHAR_BIT;
8388 if (off + fld_bit_len > bit_len)
8389 bit_len = off + fld_bit_len;
8390 TYPE_LENGTH (rtype) =
8391 align_value (bit_len, TARGET_CHAR_BIT) / TARGET_CHAR_BIT;
8392 }
8393 }
8394
8395 /* According to exp_dbug.ads, the size of TYPE for variable-size records
8396 should contain the alignment of that record, which should be a strictly
8397 positive value. If null or negative, then something is wrong, most
8398 probably in the debug info. In that case, we don't round up the size
8399 of the resulting type. If this record is not part of another structure,
8400 the current RTYPE length might be good enough for our purposes. */
8401 if (TYPE_LENGTH (type) <= 0)
8402 {
8403 if (TYPE_NAME (rtype))
8404 warning (_("Invalid type size for `%s' detected: %d."),
8405 TYPE_NAME (rtype), TYPE_LENGTH (type));
8406 else
8407 warning (_("Invalid type size for <unnamed> detected: %d."),
8408 TYPE_LENGTH (type));
8409 }
8410 else
8411 {
8412 TYPE_LENGTH (rtype) = align_value (TYPE_LENGTH (rtype),
8413 TYPE_LENGTH (type));
8414 }
8415
8416 value_free_to_mark (mark);
8417 if (TYPE_LENGTH (rtype) > varsize_limit)
8418 error (_("record type with dynamic size is larger than varsize-limit"));
8419 return rtype;
8420 }
8421
8422 /* As for ada_template_to_fixed_record_type_1 with KEEP_DYNAMIC_FIELDS
8423 of 1. */
8424
8425 static struct type *
8426 template_to_fixed_record_type (struct type *type, const gdb_byte *valaddr,
8427 CORE_ADDR address, struct value *dval0)
8428 {
8429 return ada_template_to_fixed_record_type_1 (type, valaddr,
8430 address, dval0, 1);
8431 }
8432
8433 /* An ordinary record type in which ___XVL-convention fields and
8434 ___XVU- and ___XVN-convention field types in TYPE0 are replaced with
8435 static approximations, containing all possible fields. Uses
8436 no runtime values. Useless for use in values, but that's OK,
8437 since the results are used only for type determinations. Works on both
8438 structs and unions. Representation note: to save space, we memorize
8439 the result of this function in the TYPE_TARGET_TYPE of the
8440 template type. */
8441
8442 static struct type *
8443 template_to_static_fixed_type (struct type *type0)
8444 {
8445 struct type *type;
8446 int nfields;
8447 int f;
8448
8449 /* No need no do anything if the input type is already fixed. */
8450 if (TYPE_FIXED_INSTANCE (type0))
8451 return type0;
8452
8453 /* Likewise if we already have computed the static approximation. */
8454 if (TYPE_TARGET_TYPE (type0) != NULL)
8455 return TYPE_TARGET_TYPE (type0);
8456
8457 /* Don't clone TYPE0 until we are sure we are going to need a copy. */
8458 type = type0;
8459 nfields = TYPE_NFIELDS (type0);
8460
8461 /* Whether or not we cloned TYPE0, cache the result so that we don't do
8462 recompute all over next time. */
8463 TYPE_TARGET_TYPE (type0) = type;
8464
8465 for (f = 0; f < nfields; f += 1)
8466 {
8467 struct type *field_type = TYPE_FIELD_TYPE (type0, f);
8468 struct type *new_type;
8469
8470 if (is_dynamic_field (type0, f))
8471 {
8472 field_type = ada_check_typedef (field_type);
8473 new_type = to_static_fixed_type (TYPE_TARGET_TYPE (field_type));
8474 }
8475 else
8476 new_type = static_unwrap_type (field_type);
8477
8478 if (new_type != field_type)
8479 {
8480 /* Clone TYPE0 only the first time we get a new field type. */
8481 if (type == type0)
8482 {
8483 TYPE_TARGET_TYPE (type0) = type = alloc_type_copy (type0);
8484 TYPE_CODE (type) = TYPE_CODE (type0);
8485 INIT_CPLUS_SPECIFIC (type);
8486 TYPE_NFIELDS (type) = nfields;
8487 TYPE_FIELDS (type) = (struct field *)
8488 TYPE_ALLOC (type, nfields * sizeof (struct field));
8489 memcpy (TYPE_FIELDS (type), TYPE_FIELDS (type0),
8490 sizeof (struct field) * nfields);
8491 TYPE_NAME (type) = ada_type_name (type0);
8492 TYPE_TAG_NAME (type) = NULL;
8493 TYPE_FIXED_INSTANCE (type) = 1;
8494 TYPE_LENGTH (type) = 0;
8495 }
8496 TYPE_FIELD_TYPE (type, f) = new_type;
8497 TYPE_FIELD_NAME (type, f) = TYPE_FIELD_NAME (type0, f);
8498 }
8499 }
8500
8501 return type;
8502 }
8503
8504 /* Given an object of type TYPE whose contents are at VALADDR and
8505 whose address in memory is ADDRESS, returns a revision of TYPE,
8506 which should be a non-dynamic-sized record, in which the variant
8507 part, if any, is replaced with the appropriate branch. Looks
8508 for discriminant values in DVAL0, which can be NULL if the record
8509 contains the necessary discriminant values. */
8510
8511 static struct type *
8512 to_record_with_fixed_variant_part (struct type *type, const gdb_byte *valaddr,
8513 CORE_ADDR address, struct value *dval0)
8514 {
8515 struct value *mark = value_mark ();
8516 struct value *dval;
8517 struct type *rtype;
8518 struct type *branch_type;
8519 int nfields = TYPE_NFIELDS (type);
8520 int variant_field = variant_field_index (type);
8521
8522 if (variant_field == -1)
8523 return type;
8524
8525 if (dval0 == NULL)
8526 {
8527 dval = value_from_contents_and_address (type, valaddr, address);
8528 type = value_type (dval);
8529 }
8530 else
8531 dval = dval0;
8532
8533 rtype = alloc_type_copy (type);
8534 TYPE_CODE (rtype) = TYPE_CODE_STRUCT;
8535 INIT_CPLUS_SPECIFIC (rtype);
8536 TYPE_NFIELDS (rtype) = nfields;
8537 TYPE_FIELDS (rtype) =
8538 (struct field *) TYPE_ALLOC (rtype, nfields * sizeof (struct field));
8539 memcpy (TYPE_FIELDS (rtype), TYPE_FIELDS (type),
8540 sizeof (struct field) * nfields);
8541 TYPE_NAME (rtype) = ada_type_name (type);
8542 TYPE_TAG_NAME (rtype) = NULL;
8543 TYPE_FIXED_INSTANCE (rtype) = 1;
8544 TYPE_LENGTH (rtype) = TYPE_LENGTH (type);
8545
8546 branch_type = to_fixed_variant_branch_type
8547 (TYPE_FIELD_TYPE (type, variant_field),
8548 cond_offset_host (valaddr,
8549 TYPE_FIELD_BITPOS (type, variant_field)
8550 / TARGET_CHAR_BIT),
8551 cond_offset_target (address,
8552 TYPE_FIELD_BITPOS (type, variant_field)
8553 / TARGET_CHAR_BIT), dval);
8554 if (branch_type == NULL)
8555 {
8556 int f;
8557
8558 for (f = variant_field + 1; f < nfields; f += 1)
8559 TYPE_FIELDS (rtype)[f - 1] = TYPE_FIELDS (rtype)[f];
8560 TYPE_NFIELDS (rtype) -= 1;
8561 }
8562 else
8563 {
8564 TYPE_FIELD_TYPE (rtype, variant_field) = branch_type;
8565 TYPE_FIELD_NAME (rtype, variant_field) = "S";
8566 TYPE_FIELD_BITSIZE (rtype, variant_field) = 0;
8567 TYPE_LENGTH (rtype) += TYPE_LENGTH (branch_type);
8568 }
8569 TYPE_LENGTH (rtype) -= TYPE_LENGTH (TYPE_FIELD_TYPE (type, variant_field));
8570
8571 value_free_to_mark (mark);
8572 return rtype;
8573 }
8574
8575 /* An ordinary record type (with fixed-length fields) that describes
8576 the value at (TYPE0, VALADDR, ADDRESS) [see explanation at
8577 beginning of this section]. Any necessary discriminants' values
8578 should be in DVAL, a record value; it may be NULL if the object
8579 at ADDR itself contains any necessary discriminant values.
8580 Additionally, VALADDR and ADDRESS may also be NULL if no discriminant
8581 values from the record are needed. Except in the case that DVAL,
8582 VALADDR, and ADDRESS are all 0 or NULL, a variant field (unless
8583 unchecked) is replaced by a particular branch of the variant.
8584
8585 NOTE: the case in which DVAL and VALADDR are NULL and ADDRESS is 0
8586 is questionable and may be removed. It can arise during the
8587 processing of an unconstrained-array-of-record type where all the
8588 variant branches have exactly the same size. This is because in
8589 such cases, the compiler does not bother to use the XVS convention
8590 when encoding the record. I am currently dubious of this
8591 shortcut and suspect the compiler should be altered. FIXME. */
8592
8593 static struct type *
8594 to_fixed_record_type (struct type *type0, const gdb_byte *valaddr,
8595 CORE_ADDR address, struct value *dval)
8596 {
8597 struct type *templ_type;
8598
8599 if (TYPE_FIXED_INSTANCE (type0))
8600 return type0;
8601
8602 templ_type = dynamic_template_type (type0);
8603
8604 if (templ_type != NULL)
8605 return template_to_fixed_record_type (templ_type, valaddr, address, dval);
8606 else if (variant_field_index (type0) >= 0)
8607 {
8608 if (dval == NULL && valaddr == NULL && address == 0)
8609 return type0;
8610 return to_record_with_fixed_variant_part (type0, valaddr, address,
8611 dval);
8612 }
8613 else
8614 {
8615 TYPE_FIXED_INSTANCE (type0) = 1;
8616 return type0;
8617 }
8618
8619 }
8620
8621 /* An ordinary record type (with fixed-length fields) that describes
8622 the value at (VAR_TYPE0, VALADDR, ADDRESS), where VAR_TYPE0 is a
8623 union type. Any necessary discriminants' values should be in DVAL,
8624 a record value. That is, this routine selects the appropriate
8625 branch of the union at ADDR according to the discriminant value
8626 indicated in the union's type name. Returns VAR_TYPE0 itself if
8627 it represents a variant subject to a pragma Unchecked_Union. */
8628
8629 static struct type *
8630 to_fixed_variant_branch_type (struct type *var_type0, const gdb_byte *valaddr,
8631 CORE_ADDR address, struct value *dval)
8632 {
8633 int which;
8634 struct type *templ_type;
8635 struct type *var_type;
8636
8637 if (TYPE_CODE (var_type0) == TYPE_CODE_PTR)
8638 var_type = TYPE_TARGET_TYPE (var_type0);
8639 else
8640 var_type = var_type0;
8641
8642 templ_type = ada_find_parallel_type (var_type, "___XVU");
8643
8644 if (templ_type != NULL)
8645 var_type = templ_type;
8646
8647 if (is_unchecked_variant (var_type, value_type (dval)))
8648 return var_type0;
8649 which =
8650 ada_which_variant_applies (var_type,
8651 value_type (dval), value_contents (dval));
8652
8653 if (which < 0)
8654 return empty_record (var_type);
8655 else if (is_dynamic_field (var_type, which))
8656 return to_fixed_record_type
8657 (TYPE_TARGET_TYPE (TYPE_FIELD_TYPE (var_type, which)),
8658 valaddr, address, dval);
8659 else if (variant_field_index (TYPE_FIELD_TYPE (var_type, which)) >= 0)
8660 return
8661 to_fixed_record_type
8662 (TYPE_FIELD_TYPE (var_type, which), valaddr, address, dval);
8663 else
8664 return TYPE_FIELD_TYPE (var_type, which);
8665 }
8666
8667 /* Assuming RANGE_TYPE is a TYPE_CODE_RANGE, return nonzero if
8668 ENCODING_TYPE, a type following the GNAT conventions for discrete
8669 type encodings, only carries redundant information. */
8670
8671 static int
8672 ada_is_redundant_range_encoding (struct type *range_type,
8673 struct type *encoding_type)
8674 {
8675 struct type *fixed_range_type;
8676 const char *bounds_str;
8677 int n;
8678 LONGEST lo, hi;
8679
8680 gdb_assert (TYPE_CODE (range_type) == TYPE_CODE_RANGE);
8681
8682 if (TYPE_CODE (get_base_type (range_type))
8683 != TYPE_CODE (get_base_type (encoding_type)))
8684 {
8685 /* The compiler probably used a simple base type to describe
8686 the range type instead of the range's actual base type,
8687 expecting us to get the real base type from the encoding
8688 anyway. In this situation, the encoding cannot be ignored
8689 as redundant. */
8690 return 0;
8691 }
8692
8693 if (is_dynamic_type (range_type))
8694 return 0;
8695
8696 if (TYPE_NAME (encoding_type) == NULL)
8697 return 0;
8698
8699 bounds_str = strstr (TYPE_NAME (encoding_type), "___XDLU_");
8700 if (bounds_str == NULL)
8701 return 0;
8702
8703 n = 8; /* Skip "___XDLU_". */
8704 if (!ada_scan_number (bounds_str, n, &lo, &n))
8705 return 0;
8706 if (TYPE_LOW_BOUND (range_type) != lo)
8707 return 0;
8708
8709 n += 2; /* Skip the "__" separator between the two bounds. */
8710 if (!ada_scan_number (bounds_str, n, &hi, &n))
8711 return 0;
8712 if (TYPE_HIGH_BOUND (range_type) != hi)
8713 return 0;
8714
8715 return 1;
8716 }
8717
8718 /* Given the array type ARRAY_TYPE, return nonzero if DESC_TYPE,
8719 a type following the GNAT encoding for describing array type
8720 indices, only carries redundant information. */
8721
8722 static int
8723 ada_is_redundant_index_type_desc (struct type *array_type,
8724 struct type *desc_type)
8725 {
8726 struct type *this_layer = check_typedef (array_type);
8727 int i;
8728
8729 for (i = 0; i < TYPE_NFIELDS (desc_type); i++)
8730 {
8731 if (!ada_is_redundant_range_encoding (TYPE_INDEX_TYPE (this_layer),
8732 TYPE_FIELD_TYPE (desc_type, i)))
8733 return 0;
8734 this_layer = check_typedef (TYPE_TARGET_TYPE (this_layer));
8735 }
8736
8737 return 1;
8738 }
8739
8740 /* Assuming that TYPE0 is an array type describing the type of a value
8741 at ADDR, and that DVAL describes a record containing any
8742 discriminants used in TYPE0, returns a type for the value that
8743 contains no dynamic components (that is, no components whose sizes
8744 are determined by run-time quantities). Unless IGNORE_TOO_BIG is
8745 true, gives an error message if the resulting type's size is over
8746 varsize_limit. */
8747
8748 static struct type *
8749 to_fixed_array_type (struct type *type0, struct value *dval,
8750 int ignore_too_big)
8751 {
8752 struct type *index_type_desc;
8753 struct type *result;
8754 int constrained_packed_array_p;
8755 static const char *xa_suffix = "___XA";
8756
8757 type0 = ada_check_typedef (type0);
8758 if (TYPE_FIXED_INSTANCE (type0))
8759 return type0;
8760
8761 constrained_packed_array_p = ada_is_constrained_packed_array_type (type0);
8762 if (constrained_packed_array_p)
8763 type0 = decode_constrained_packed_array_type (type0);
8764
8765 index_type_desc = ada_find_parallel_type (type0, xa_suffix);
8766
8767 /* As mentioned in exp_dbug.ads, for non bit-packed arrays an
8768 encoding suffixed with 'P' may still be generated. If so,
8769 it should be used to find the XA type. */
8770
8771 if (index_type_desc == NULL)
8772 {
8773 const char *type_name = ada_type_name (type0);
8774
8775 if (type_name != NULL)
8776 {
8777 const int len = strlen (type_name);
8778 char *name = (char *) alloca (len + strlen (xa_suffix));
8779
8780 if (type_name[len - 1] == 'P')
8781 {
8782 strcpy (name, type_name);
8783 strcpy (name + len - 1, xa_suffix);
8784 index_type_desc = ada_find_parallel_type_with_name (type0, name);
8785 }
8786 }
8787 }
8788
8789 ada_fixup_array_indexes_type (index_type_desc);
8790 if (index_type_desc != NULL
8791 && ada_is_redundant_index_type_desc (type0, index_type_desc))
8792 {
8793 /* Ignore this ___XA parallel type, as it does not bring any
8794 useful information. This allows us to avoid creating fixed
8795 versions of the array's index types, which would be identical
8796 to the original ones. This, in turn, can also help avoid
8797 the creation of fixed versions of the array itself. */
8798 index_type_desc = NULL;
8799 }
8800
8801 if (index_type_desc == NULL)
8802 {
8803 struct type *elt_type0 = ada_check_typedef (TYPE_TARGET_TYPE (type0));
8804
8805 /* NOTE: elt_type---the fixed version of elt_type0---should never
8806 depend on the contents of the array in properly constructed
8807 debugging data. */
8808 /* Create a fixed version of the array element type.
8809 We're not providing the address of an element here,
8810 and thus the actual object value cannot be inspected to do
8811 the conversion. This should not be a problem, since arrays of
8812 unconstrained objects are not allowed. In particular, all
8813 the elements of an array of a tagged type should all be of
8814 the same type specified in the debugging info. No need to
8815 consult the object tag. */
8816 struct type *elt_type = ada_to_fixed_type (elt_type0, 0, 0, dval, 1);
8817
8818 /* Make sure we always create a new array type when dealing with
8819 packed array types, since we're going to fix-up the array
8820 type length and element bitsize a little further down. */
8821 if (elt_type0 == elt_type && !constrained_packed_array_p)
8822 result = type0;
8823 else
8824 result = create_array_type (alloc_type_copy (type0),
8825 elt_type, TYPE_INDEX_TYPE (type0));
8826 }
8827 else
8828 {
8829 int i;
8830 struct type *elt_type0;
8831
8832 elt_type0 = type0;
8833 for (i = TYPE_NFIELDS (index_type_desc); i > 0; i -= 1)
8834 elt_type0 = TYPE_TARGET_TYPE (elt_type0);
8835
8836 /* NOTE: result---the fixed version of elt_type0---should never
8837 depend on the contents of the array in properly constructed
8838 debugging data. */
8839 /* Create a fixed version of the array element type.
8840 We're not providing the address of an element here,
8841 and thus the actual object value cannot be inspected to do
8842 the conversion. This should not be a problem, since arrays of
8843 unconstrained objects are not allowed. In particular, all
8844 the elements of an array of a tagged type should all be of
8845 the same type specified in the debugging info. No need to
8846 consult the object tag. */
8847 result =
8848 ada_to_fixed_type (ada_check_typedef (elt_type0), 0, 0, dval, 1);
8849
8850 elt_type0 = type0;
8851 for (i = TYPE_NFIELDS (index_type_desc) - 1; i >= 0; i -= 1)
8852 {
8853 struct type *range_type =
8854 to_fixed_range_type (TYPE_FIELD_TYPE (index_type_desc, i), dval);
8855
8856 result = create_array_type (alloc_type_copy (elt_type0),
8857 result, range_type);
8858 elt_type0 = TYPE_TARGET_TYPE (elt_type0);
8859 }
8860 if (!ignore_too_big && TYPE_LENGTH (result) > varsize_limit)
8861 error (_("array type with dynamic size is larger than varsize-limit"));
8862 }
8863
8864 /* We want to preserve the type name. This can be useful when
8865 trying to get the type name of a value that has already been
8866 printed (for instance, if the user did "print VAR; whatis $". */
8867 TYPE_NAME (result) = TYPE_NAME (type0);
8868
8869 if (constrained_packed_array_p)
8870 {
8871 /* So far, the resulting type has been created as if the original
8872 type was a regular (non-packed) array type. As a result, the
8873 bitsize of the array elements needs to be set again, and the array
8874 length needs to be recomputed based on that bitsize. */
8875 int len = TYPE_LENGTH (result) / TYPE_LENGTH (TYPE_TARGET_TYPE (result));
8876 int elt_bitsize = TYPE_FIELD_BITSIZE (type0, 0);
8877
8878 TYPE_FIELD_BITSIZE (result, 0) = TYPE_FIELD_BITSIZE (type0, 0);
8879 TYPE_LENGTH (result) = len * elt_bitsize / HOST_CHAR_BIT;
8880 if (TYPE_LENGTH (result) * HOST_CHAR_BIT < len * elt_bitsize)
8881 TYPE_LENGTH (result)++;
8882 }
8883
8884 TYPE_FIXED_INSTANCE (result) = 1;
8885 return result;
8886 }
8887
8888
8889 /* A standard type (containing no dynamically sized components)
8890 corresponding to TYPE for the value (TYPE, VALADDR, ADDRESS)
8891 DVAL describes a record containing any discriminants used in TYPE0,
8892 and may be NULL if there are none, or if the object of type TYPE at
8893 ADDRESS or in VALADDR contains these discriminants.
8894
8895 If CHECK_TAG is not null, in the case of tagged types, this function
8896 attempts to locate the object's tag and use it to compute the actual
8897 type. However, when ADDRESS is null, we cannot use it to determine the
8898 location of the tag, and therefore compute the tagged type's actual type.
8899 So we return the tagged type without consulting the tag. */
8900
8901 static struct type *
8902 ada_to_fixed_type_1 (struct type *type, const gdb_byte *valaddr,
8903 CORE_ADDR address, struct value *dval, int check_tag)
8904 {
8905 type = ada_check_typedef (type);
8906 switch (TYPE_CODE (type))
8907 {
8908 default:
8909 return type;
8910 case TYPE_CODE_STRUCT:
8911 {
8912 struct type *static_type = to_static_fixed_type (type);
8913 struct type *fixed_record_type =
8914 to_fixed_record_type (type, valaddr, address, NULL);
8915
8916 /* If STATIC_TYPE is a tagged type and we know the object's address,
8917 then we can determine its tag, and compute the object's actual
8918 type from there. Note that we have to use the fixed record
8919 type (the parent part of the record may have dynamic fields
8920 and the way the location of _tag is expressed may depend on
8921 them). */
8922
8923 if (check_tag && address != 0 && ada_is_tagged_type (static_type, 0))
8924 {
8925 struct value *tag =
8926 value_tag_from_contents_and_address
8927 (fixed_record_type,
8928 valaddr,
8929 address);
8930 struct type *real_type = type_from_tag (tag);
8931 struct value *obj =
8932 value_from_contents_and_address (fixed_record_type,
8933 valaddr,
8934 address);
8935 fixed_record_type = value_type (obj);
8936 if (real_type != NULL)
8937 return to_fixed_record_type
8938 (real_type, NULL,
8939 value_address (ada_tag_value_at_base_address (obj)), NULL);
8940 }
8941
8942 /* Check to see if there is a parallel ___XVZ variable.
8943 If there is, then it provides the actual size of our type. */
8944 else if (ada_type_name (fixed_record_type) != NULL)
8945 {
8946 const char *name = ada_type_name (fixed_record_type);
8947 char *xvz_name
8948 = (char *) alloca (strlen (name) + 7 /* "___XVZ\0" */);
8949 int xvz_found = 0;
8950 LONGEST size;
8951
8952 xsnprintf (xvz_name, strlen (name) + 7, "%s___XVZ", name);
8953 size = get_int_var_value (xvz_name, &xvz_found);
8954 if (xvz_found && TYPE_LENGTH (fixed_record_type) != size)
8955 {
8956 fixed_record_type = copy_type (fixed_record_type);
8957 TYPE_LENGTH (fixed_record_type) = size;
8958
8959 /* The FIXED_RECORD_TYPE may have be a stub. We have
8960 observed this when the debugging info is STABS, and
8961 apparently it is something that is hard to fix.
8962
8963 In practice, we don't need the actual type definition
8964 at all, because the presence of the XVZ variable allows us
8965 to assume that there must be a XVS type as well, which we
8966 should be able to use later, when we need the actual type
8967 definition.
8968
8969 In the meantime, pretend that the "fixed" type we are
8970 returning is NOT a stub, because this can cause trouble
8971 when using this type to create new types targeting it.
8972 Indeed, the associated creation routines often check
8973 whether the target type is a stub and will try to replace
8974 it, thus using a type with the wrong size. This, in turn,
8975 might cause the new type to have the wrong size too.
8976 Consider the case of an array, for instance, where the size
8977 of the array is computed from the number of elements in
8978 our array multiplied by the size of its element. */
8979 TYPE_STUB (fixed_record_type) = 0;
8980 }
8981 }
8982 return fixed_record_type;
8983 }
8984 case TYPE_CODE_ARRAY:
8985 return to_fixed_array_type (type, dval, 1);
8986 case TYPE_CODE_UNION:
8987 if (dval == NULL)
8988 return type;
8989 else
8990 return to_fixed_variant_branch_type (type, valaddr, address, dval);
8991 }
8992 }
8993
8994 /* The same as ada_to_fixed_type_1, except that it preserves the type
8995 if it is a TYPE_CODE_TYPEDEF of a type that is already fixed.
8996
8997 The typedef layer needs be preserved in order to differentiate between
8998 arrays and array pointers when both types are implemented using the same
8999 fat pointer. In the array pointer case, the pointer is encoded as
9000 a typedef of the pointer type. For instance, considering:
9001
9002 type String_Access is access String;
9003 S1 : String_Access := null;
9004
9005 To the debugger, S1 is defined as a typedef of type String. But
9006 to the user, it is a pointer. So if the user tries to print S1,
9007 we should not dereference the array, but print the array address
9008 instead.
9009
9010 If we didn't preserve the typedef layer, we would lose the fact that
9011 the type is to be presented as a pointer (needs de-reference before
9012 being printed). And we would also use the source-level type name. */
9013
9014 struct type *
9015 ada_to_fixed_type (struct type *type, const gdb_byte *valaddr,
9016 CORE_ADDR address, struct value *dval, int check_tag)
9017
9018 {
9019 struct type *fixed_type =
9020 ada_to_fixed_type_1 (type, valaddr, address, dval, check_tag);
9021
9022 /* If TYPE is a typedef and its target type is the same as the FIXED_TYPE,
9023 then preserve the typedef layer.
9024
9025 Implementation note: We can only check the main-type portion of
9026 the TYPE and FIXED_TYPE, because eliminating the typedef layer
9027 from TYPE now returns a type that has the same instance flags
9028 as TYPE. For instance, if TYPE is a "typedef const", and its
9029 target type is a "struct", then the typedef elimination will return
9030 a "const" version of the target type. See check_typedef for more
9031 details about how the typedef layer elimination is done.
9032
9033 brobecker/2010-11-19: It seems to me that the only case where it is
9034 useful to preserve the typedef layer is when dealing with fat pointers.
9035 Perhaps, we could add a check for that and preserve the typedef layer
9036 only in that situation. But this seems unecessary so far, probably
9037 because we call check_typedef/ada_check_typedef pretty much everywhere.
9038 */
9039 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF
9040 && (TYPE_MAIN_TYPE (ada_typedef_target_type (type))
9041 == TYPE_MAIN_TYPE (fixed_type)))
9042 return type;
9043
9044 return fixed_type;
9045 }
9046
9047 /* A standard (static-sized) type corresponding as well as possible to
9048 TYPE0, but based on no runtime data. */
9049
9050 static struct type *
9051 to_static_fixed_type (struct type *type0)
9052 {
9053 struct type *type;
9054
9055 if (type0 == NULL)
9056 return NULL;
9057
9058 if (TYPE_FIXED_INSTANCE (type0))
9059 return type0;
9060
9061 type0 = ada_check_typedef (type0);
9062
9063 switch (TYPE_CODE (type0))
9064 {
9065 default:
9066 return type0;
9067 case TYPE_CODE_STRUCT:
9068 type = dynamic_template_type (type0);
9069 if (type != NULL)
9070 return template_to_static_fixed_type (type);
9071 else
9072 return template_to_static_fixed_type (type0);
9073 case TYPE_CODE_UNION:
9074 type = ada_find_parallel_type (type0, "___XVU");
9075 if (type != NULL)
9076 return template_to_static_fixed_type (type);
9077 else
9078 return template_to_static_fixed_type (type0);
9079 }
9080 }
9081
9082 /* A static approximation of TYPE with all type wrappers removed. */
9083
9084 static struct type *
9085 static_unwrap_type (struct type *type)
9086 {
9087 if (ada_is_aligner_type (type))
9088 {
9089 struct type *type1 = TYPE_FIELD_TYPE (ada_check_typedef (type), 0);
9090 if (ada_type_name (type1) == NULL)
9091 TYPE_NAME (type1) = ada_type_name (type);
9092
9093 return static_unwrap_type (type1);
9094 }
9095 else
9096 {
9097 struct type *raw_real_type = ada_get_base_type (type);
9098
9099 if (raw_real_type == type)
9100 return type;
9101 else
9102 return to_static_fixed_type (raw_real_type);
9103 }
9104 }
9105
9106 /* In some cases, incomplete and private types require
9107 cross-references that are not resolved as records (for example,
9108 type Foo;
9109 type FooP is access Foo;
9110 V: FooP;
9111 type Foo is array ...;
9112 ). In these cases, since there is no mechanism for producing
9113 cross-references to such types, we instead substitute for FooP a
9114 stub enumeration type that is nowhere resolved, and whose tag is
9115 the name of the actual type. Call these types "non-record stubs". */
9116
9117 /* A type equivalent to TYPE that is not a non-record stub, if one
9118 exists, otherwise TYPE. */
9119
9120 struct type *
9121 ada_check_typedef (struct type *type)
9122 {
9123 if (type == NULL)
9124 return NULL;
9125
9126 /* If our type is a typedef type of a fat pointer, then we're done.
9127 We don't want to strip the TYPE_CODE_TYPDEF layer, because this is
9128 what allows us to distinguish between fat pointers that represent
9129 array types, and fat pointers that represent array access types
9130 (in both cases, the compiler implements them as fat pointers). */
9131 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF
9132 && is_thick_pntr (ada_typedef_target_type (type)))
9133 return type;
9134
9135 type = check_typedef (type);
9136 if (type == NULL || TYPE_CODE (type) != TYPE_CODE_ENUM
9137 || !TYPE_STUB (type)
9138 || TYPE_TAG_NAME (type) == NULL)
9139 return type;
9140 else
9141 {
9142 const char *name = TYPE_TAG_NAME (type);
9143 struct type *type1 = ada_find_any_type (name);
9144
9145 if (type1 == NULL)
9146 return type;
9147
9148 /* TYPE1 might itself be a TYPE_CODE_TYPEDEF (this can happen with
9149 stubs pointing to arrays, as we don't create symbols for array
9150 types, only for the typedef-to-array types). If that's the case,
9151 strip the typedef layer. */
9152 if (TYPE_CODE (type1) == TYPE_CODE_TYPEDEF)
9153 type1 = ada_check_typedef (type1);
9154
9155 return type1;
9156 }
9157 }
9158
9159 /* A value representing the data at VALADDR/ADDRESS as described by
9160 type TYPE0, but with a standard (static-sized) type that correctly
9161 describes it. If VAL0 is not NULL and TYPE0 already is a standard
9162 type, then return VAL0 [this feature is simply to avoid redundant
9163 creation of struct values]. */
9164
9165 static struct value *
9166 ada_to_fixed_value_create (struct type *type0, CORE_ADDR address,
9167 struct value *val0)
9168 {
9169 struct type *type = ada_to_fixed_type (type0, 0, address, NULL, 1);
9170
9171 if (type == type0 && val0 != NULL)
9172 return val0;
9173 else
9174 return value_from_contents_and_address (type, 0, address);
9175 }
9176
9177 /* A value representing VAL, but with a standard (static-sized) type
9178 that correctly describes it. Does not necessarily create a new
9179 value. */
9180
9181 struct value *
9182 ada_to_fixed_value (struct value *val)
9183 {
9184 val = unwrap_value (val);
9185 val = ada_to_fixed_value_create (value_type (val),
9186 value_address (val),
9187 val);
9188 return val;
9189 }
9190 \f
9191
9192 /* Attributes */
9193
9194 /* Table mapping attribute numbers to names.
9195 NOTE: Keep up to date with enum ada_attribute definition in ada-lang.h. */
9196
9197 static const char *attribute_names[] = {
9198 "<?>",
9199
9200 "first",
9201 "last",
9202 "length",
9203 "image",
9204 "max",
9205 "min",
9206 "modulus",
9207 "pos",
9208 "size",
9209 "tag",
9210 "val",
9211 0
9212 };
9213
9214 const char *
9215 ada_attribute_name (enum exp_opcode n)
9216 {
9217 if (n >= OP_ATR_FIRST && n <= (int) OP_ATR_VAL)
9218 return attribute_names[n - OP_ATR_FIRST + 1];
9219 else
9220 return attribute_names[0];
9221 }
9222
9223 /* Evaluate the 'POS attribute applied to ARG. */
9224
9225 static LONGEST
9226 pos_atr (struct value *arg)
9227 {
9228 struct value *val = coerce_ref (arg);
9229 struct type *type = value_type (val);
9230 LONGEST result;
9231
9232 if (!discrete_type_p (type))
9233 error (_("'POS only defined on discrete types"));
9234
9235 if (!discrete_position (type, value_as_long (val), &result))
9236 error (_("enumeration value is invalid: can't find 'POS"));
9237
9238 return result;
9239 }
9240
9241 static struct value *
9242 value_pos_atr (struct type *type, struct value *arg)
9243 {
9244 return value_from_longest (type, pos_atr (arg));
9245 }
9246
9247 /* Evaluate the TYPE'VAL attribute applied to ARG. */
9248
9249 static struct value *
9250 value_val_atr (struct type *type, struct value *arg)
9251 {
9252 if (!discrete_type_p (type))
9253 error (_("'VAL only defined on discrete types"));
9254 if (!integer_type_p (value_type (arg)))
9255 error (_("'VAL requires integral argument"));
9256
9257 if (TYPE_CODE (type) == TYPE_CODE_ENUM)
9258 {
9259 long pos = value_as_long (arg);
9260
9261 if (pos < 0 || pos >= TYPE_NFIELDS (type))
9262 error (_("argument to 'VAL out of range"));
9263 return value_from_longest (type, TYPE_FIELD_ENUMVAL (type, pos));
9264 }
9265 else
9266 return value_from_longest (type, value_as_long (arg));
9267 }
9268 \f
9269
9270 /* Evaluation */
9271
9272 /* True if TYPE appears to be an Ada character type.
9273 [At the moment, this is true only for Character and Wide_Character;
9274 It is a heuristic test that could stand improvement]. */
9275
9276 int
9277 ada_is_character_type (struct type *type)
9278 {
9279 const char *name;
9280
9281 /* If the type code says it's a character, then assume it really is,
9282 and don't check any further. */
9283 if (TYPE_CODE (type) == TYPE_CODE_CHAR)
9284 return 1;
9285
9286 /* Otherwise, assume it's a character type iff it is a discrete type
9287 with a known character type name. */
9288 name = ada_type_name (type);
9289 return (name != NULL
9290 && (TYPE_CODE (type) == TYPE_CODE_INT
9291 || TYPE_CODE (type) == TYPE_CODE_RANGE)
9292 && (strcmp (name, "character") == 0
9293 || strcmp (name, "wide_character") == 0
9294 || strcmp (name, "wide_wide_character") == 0
9295 || strcmp (name, "unsigned char") == 0));
9296 }
9297
9298 /* True if TYPE appears to be an Ada string type. */
9299
9300 int
9301 ada_is_string_type (struct type *type)
9302 {
9303 type = ada_check_typedef (type);
9304 if (type != NULL
9305 && TYPE_CODE (type) != TYPE_CODE_PTR
9306 && (ada_is_simple_array_type (type)
9307 || ada_is_array_descriptor_type (type))
9308 && ada_array_arity (type) == 1)
9309 {
9310 struct type *elttype = ada_array_element_type (type, 1);
9311
9312 return ada_is_character_type (elttype);
9313 }
9314 else
9315 return 0;
9316 }
9317
9318 /* The compiler sometimes provides a parallel XVS type for a given
9319 PAD type. Normally, it is safe to follow the PAD type directly,
9320 but older versions of the compiler have a bug that causes the offset
9321 of its "F" field to be wrong. Following that field in that case
9322 would lead to incorrect results, but this can be worked around
9323 by ignoring the PAD type and using the associated XVS type instead.
9324
9325 Set to True if the debugger should trust the contents of PAD types.
9326 Otherwise, ignore the PAD type if there is a parallel XVS type. */
9327 static int trust_pad_over_xvs = 1;
9328
9329 /* True if TYPE is a struct type introduced by the compiler to force the
9330 alignment of a value. Such types have a single field with a
9331 distinctive name. */
9332
9333 int
9334 ada_is_aligner_type (struct type *type)
9335 {
9336 type = ada_check_typedef (type);
9337
9338 if (!trust_pad_over_xvs && ada_find_parallel_type (type, "___XVS") != NULL)
9339 return 0;
9340
9341 return (TYPE_CODE (type) == TYPE_CODE_STRUCT
9342 && TYPE_NFIELDS (type) == 1
9343 && strcmp (TYPE_FIELD_NAME (type, 0), "F") == 0);
9344 }
9345
9346 /* If there is an ___XVS-convention type parallel to SUBTYPE, return
9347 the parallel type. */
9348
9349 struct type *
9350 ada_get_base_type (struct type *raw_type)
9351 {
9352 struct type *real_type_namer;
9353 struct type *raw_real_type;
9354
9355 if (raw_type == NULL || TYPE_CODE (raw_type) != TYPE_CODE_STRUCT)
9356 return raw_type;
9357
9358 if (ada_is_aligner_type (raw_type))
9359 /* The encoding specifies that we should always use the aligner type.
9360 So, even if this aligner type has an associated XVS type, we should
9361 simply ignore it.
9362
9363 According to the compiler gurus, an XVS type parallel to an aligner
9364 type may exist because of a stabs limitation. In stabs, aligner
9365 types are empty because the field has a variable-sized type, and
9366 thus cannot actually be used as an aligner type. As a result,
9367 we need the associated parallel XVS type to decode the type.
9368 Since the policy in the compiler is to not change the internal
9369 representation based on the debugging info format, we sometimes
9370 end up having a redundant XVS type parallel to the aligner type. */
9371 return raw_type;
9372
9373 real_type_namer = ada_find_parallel_type (raw_type, "___XVS");
9374 if (real_type_namer == NULL
9375 || TYPE_CODE (real_type_namer) != TYPE_CODE_STRUCT
9376 || TYPE_NFIELDS (real_type_namer) != 1)
9377 return raw_type;
9378
9379 if (TYPE_CODE (TYPE_FIELD_TYPE (real_type_namer, 0)) != TYPE_CODE_REF)
9380 {
9381 /* This is an older encoding form where the base type needs to be
9382 looked up by name. We prefer the newer enconding because it is
9383 more efficient. */
9384 raw_real_type = ada_find_any_type (TYPE_FIELD_NAME (real_type_namer, 0));
9385 if (raw_real_type == NULL)
9386 return raw_type;
9387 else
9388 return raw_real_type;
9389 }
9390
9391 /* The field in our XVS type is a reference to the base type. */
9392 return TYPE_TARGET_TYPE (TYPE_FIELD_TYPE (real_type_namer, 0));
9393 }
9394
9395 /* The type of value designated by TYPE, with all aligners removed. */
9396
9397 struct type *
9398 ada_aligned_type (struct type *type)
9399 {
9400 if (ada_is_aligner_type (type))
9401 return ada_aligned_type (TYPE_FIELD_TYPE (type, 0));
9402 else
9403 return ada_get_base_type (type);
9404 }
9405
9406
9407 /* The address of the aligned value in an object at address VALADDR
9408 having type TYPE. Assumes ada_is_aligner_type (TYPE). */
9409
9410 const gdb_byte *
9411 ada_aligned_value_addr (struct type *type, const gdb_byte *valaddr)
9412 {
9413 if (ada_is_aligner_type (type))
9414 return ada_aligned_value_addr (TYPE_FIELD_TYPE (type, 0),
9415 valaddr +
9416 TYPE_FIELD_BITPOS (type,
9417 0) / TARGET_CHAR_BIT);
9418 else
9419 return valaddr;
9420 }
9421
9422
9423
9424 /* The printed representation of an enumeration literal with encoded
9425 name NAME. The value is good to the next call of ada_enum_name. */
9426 const char *
9427 ada_enum_name (const char *name)
9428 {
9429 static char *result;
9430 static size_t result_len = 0;
9431 char *tmp;
9432
9433 /* First, unqualify the enumeration name:
9434 1. Search for the last '.' character. If we find one, then skip
9435 all the preceding characters, the unqualified name starts
9436 right after that dot.
9437 2. Otherwise, we may be debugging on a target where the compiler
9438 translates dots into "__". Search forward for double underscores,
9439 but stop searching when we hit an overloading suffix, which is
9440 of the form "__" followed by digits. */
9441
9442 tmp = strrchr (name, '.');
9443 if (tmp != NULL)
9444 name = tmp + 1;
9445 else
9446 {
9447 while ((tmp = strstr (name, "__")) != NULL)
9448 {
9449 if (isdigit (tmp[2]))
9450 break;
9451 else
9452 name = tmp + 2;
9453 }
9454 }
9455
9456 if (name[0] == 'Q')
9457 {
9458 int v;
9459
9460 if (name[1] == 'U' || name[1] == 'W')
9461 {
9462 if (sscanf (name + 2, "%x", &v) != 1)
9463 return name;
9464 }
9465 else
9466 return name;
9467
9468 GROW_VECT (result, result_len, 16);
9469 if (isascii (v) && isprint (v))
9470 xsnprintf (result, result_len, "'%c'", v);
9471 else if (name[1] == 'U')
9472 xsnprintf (result, result_len, "[\"%02x\"]", v);
9473 else
9474 xsnprintf (result, result_len, "[\"%04x\"]", v);
9475
9476 return result;
9477 }
9478 else
9479 {
9480 tmp = strstr (name, "__");
9481 if (tmp == NULL)
9482 tmp = strstr (name, "$");
9483 if (tmp != NULL)
9484 {
9485 GROW_VECT (result, result_len, tmp - name + 1);
9486 strncpy (result, name, tmp - name);
9487 result[tmp - name] = '\0';
9488 return result;
9489 }
9490
9491 return name;
9492 }
9493 }
9494
9495 /* Evaluate the subexpression of EXP starting at *POS as for
9496 evaluate_type, updating *POS to point just past the evaluated
9497 expression. */
9498
9499 static struct value *
9500 evaluate_subexp_type (struct expression *exp, int *pos)
9501 {
9502 return evaluate_subexp (NULL_TYPE, exp, pos, EVAL_AVOID_SIDE_EFFECTS);
9503 }
9504
9505 /* If VAL is wrapped in an aligner or subtype wrapper, return the
9506 value it wraps. */
9507
9508 static struct value *
9509 unwrap_value (struct value *val)
9510 {
9511 struct type *type = ada_check_typedef (value_type (val));
9512
9513 if (ada_is_aligner_type (type))
9514 {
9515 struct value *v = ada_value_struct_elt (val, "F", 0);
9516 struct type *val_type = ada_check_typedef (value_type (v));
9517
9518 if (ada_type_name (val_type) == NULL)
9519 TYPE_NAME (val_type) = ada_type_name (type);
9520
9521 return unwrap_value (v);
9522 }
9523 else
9524 {
9525 struct type *raw_real_type =
9526 ada_check_typedef (ada_get_base_type (type));
9527
9528 /* If there is no parallel XVS or XVE type, then the value is
9529 already unwrapped. Return it without further modification. */
9530 if ((type == raw_real_type)
9531 && ada_find_parallel_type (type, "___XVE") == NULL)
9532 return val;
9533
9534 return
9535 coerce_unspec_val_to_type
9536 (val, ada_to_fixed_type (raw_real_type, 0,
9537 value_address (val),
9538 NULL, 1));
9539 }
9540 }
9541
9542 static struct value *
9543 cast_to_fixed (struct type *type, struct value *arg)
9544 {
9545 LONGEST val;
9546
9547 if (type == value_type (arg))
9548 return arg;
9549 else if (ada_is_fixed_point_type (value_type (arg)))
9550 val = ada_float_to_fixed (type,
9551 ada_fixed_to_float (value_type (arg),
9552 value_as_long (arg)));
9553 else
9554 {
9555 DOUBLEST argd = value_as_double (arg);
9556
9557 val = ada_float_to_fixed (type, argd);
9558 }
9559
9560 return value_from_longest (type, val);
9561 }
9562
9563 static struct value *
9564 cast_from_fixed (struct type *type, struct value *arg)
9565 {
9566 DOUBLEST val = ada_fixed_to_float (value_type (arg),
9567 value_as_long (arg));
9568
9569 return value_from_double (type, val);
9570 }
9571
9572 /* Given two array types T1 and T2, return nonzero iff both arrays
9573 contain the same number of elements. */
9574
9575 static int
9576 ada_same_array_size_p (struct type *t1, struct type *t2)
9577 {
9578 LONGEST lo1, hi1, lo2, hi2;
9579
9580 /* Get the array bounds in order to verify that the size of
9581 the two arrays match. */
9582 if (!get_array_bounds (t1, &lo1, &hi1)
9583 || !get_array_bounds (t2, &lo2, &hi2))
9584 error (_("unable to determine array bounds"));
9585
9586 /* To make things easier for size comparison, normalize a bit
9587 the case of empty arrays by making sure that the difference
9588 between upper bound and lower bound is always -1. */
9589 if (lo1 > hi1)
9590 hi1 = lo1 - 1;
9591 if (lo2 > hi2)
9592 hi2 = lo2 - 1;
9593
9594 return (hi1 - lo1 == hi2 - lo2);
9595 }
9596
9597 /* Assuming that VAL is an array of integrals, and TYPE represents
9598 an array with the same number of elements, but with wider integral
9599 elements, return an array "casted" to TYPE. In practice, this
9600 means that the returned array is built by casting each element
9601 of the original array into TYPE's (wider) element type. */
9602
9603 static struct value *
9604 ada_promote_array_of_integrals (struct type *type, struct value *val)
9605 {
9606 struct type *elt_type = TYPE_TARGET_TYPE (type);
9607 LONGEST lo, hi;
9608 struct value *res;
9609 LONGEST i;
9610
9611 /* Verify that both val and type are arrays of scalars, and
9612 that the size of val's elements is smaller than the size
9613 of type's element. */
9614 gdb_assert (TYPE_CODE (type) == TYPE_CODE_ARRAY);
9615 gdb_assert (is_integral_type (TYPE_TARGET_TYPE (type)));
9616 gdb_assert (TYPE_CODE (value_type (val)) == TYPE_CODE_ARRAY);
9617 gdb_assert (is_integral_type (TYPE_TARGET_TYPE (value_type (val))));
9618 gdb_assert (TYPE_LENGTH (TYPE_TARGET_TYPE (type))
9619 > TYPE_LENGTH (TYPE_TARGET_TYPE (value_type (val))));
9620
9621 if (!get_array_bounds (type, &lo, &hi))
9622 error (_("unable to determine array bounds"));
9623
9624 res = allocate_value (type);
9625
9626 /* Promote each array element. */
9627 for (i = 0; i < hi - lo + 1; i++)
9628 {
9629 struct value *elt = value_cast (elt_type, value_subscript (val, lo + i));
9630
9631 memcpy (value_contents_writeable (res) + (i * TYPE_LENGTH (elt_type)),
9632 value_contents_all (elt), TYPE_LENGTH (elt_type));
9633 }
9634
9635 return res;
9636 }
9637
9638 /* Coerce VAL as necessary for assignment to an lval of type TYPE, and
9639 return the converted value. */
9640
9641 static struct value *
9642 coerce_for_assign (struct type *type, struct value *val)
9643 {
9644 struct type *type2 = value_type (val);
9645
9646 if (type == type2)
9647 return val;
9648
9649 type2 = ada_check_typedef (type2);
9650 type = ada_check_typedef (type);
9651
9652 if (TYPE_CODE (type2) == TYPE_CODE_PTR
9653 && TYPE_CODE (type) == TYPE_CODE_ARRAY)
9654 {
9655 val = ada_value_ind (val);
9656 type2 = value_type (val);
9657 }
9658
9659 if (TYPE_CODE (type2) == TYPE_CODE_ARRAY
9660 && TYPE_CODE (type) == TYPE_CODE_ARRAY)
9661 {
9662 if (!ada_same_array_size_p (type, type2))
9663 error (_("cannot assign arrays of different length"));
9664
9665 if (is_integral_type (TYPE_TARGET_TYPE (type))
9666 && is_integral_type (TYPE_TARGET_TYPE (type2))
9667 && TYPE_LENGTH (TYPE_TARGET_TYPE (type2))
9668 < TYPE_LENGTH (TYPE_TARGET_TYPE (type)))
9669 {
9670 /* Allow implicit promotion of the array elements to
9671 a wider type. */
9672 return ada_promote_array_of_integrals (type, val);
9673 }
9674
9675 if (TYPE_LENGTH (TYPE_TARGET_TYPE (type2))
9676 != TYPE_LENGTH (TYPE_TARGET_TYPE (type)))
9677 error (_("Incompatible types in assignment"));
9678 deprecated_set_value_type (val, type);
9679 }
9680 return val;
9681 }
9682
9683 static struct value *
9684 ada_value_binop (struct value *arg1, struct value *arg2, enum exp_opcode op)
9685 {
9686 struct value *val;
9687 struct type *type1, *type2;
9688 LONGEST v, v1, v2;
9689
9690 arg1 = coerce_ref (arg1);
9691 arg2 = coerce_ref (arg2);
9692 type1 = get_base_type (ada_check_typedef (value_type (arg1)));
9693 type2 = get_base_type (ada_check_typedef (value_type (arg2)));
9694
9695 if (TYPE_CODE (type1) != TYPE_CODE_INT
9696 || TYPE_CODE (type2) != TYPE_CODE_INT)
9697 return value_binop (arg1, arg2, op);
9698
9699 switch (op)
9700 {
9701 case BINOP_MOD:
9702 case BINOP_DIV:
9703 case BINOP_REM:
9704 break;
9705 default:
9706 return value_binop (arg1, arg2, op);
9707 }
9708
9709 v2 = value_as_long (arg2);
9710 if (v2 == 0)
9711 error (_("second operand of %s must not be zero."), op_string (op));
9712
9713 if (TYPE_UNSIGNED (type1) || op == BINOP_MOD)
9714 return value_binop (arg1, arg2, op);
9715
9716 v1 = value_as_long (arg1);
9717 switch (op)
9718 {
9719 case BINOP_DIV:
9720 v = v1 / v2;
9721 if (!TRUNCATION_TOWARDS_ZERO && v1 * (v1 % v2) < 0)
9722 v += v > 0 ? -1 : 1;
9723 break;
9724 case BINOP_REM:
9725 v = v1 % v2;
9726 if (v * v1 < 0)
9727 v -= v2;
9728 break;
9729 default:
9730 /* Should not reach this point. */
9731 v = 0;
9732 }
9733
9734 val = allocate_value (type1);
9735 store_unsigned_integer (value_contents_raw (val),
9736 TYPE_LENGTH (value_type (val)),
9737 gdbarch_byte_order (get_type_arch (type1)), v);
9738 return val;
9739 }
9740
9741 static int
9742 ada_value_equal (struct value *arg1, struct value *arg2)
9743 {
9744 if (ada_is_direct_array_type (value_type (arg1))
9745 || ada_is_direct_array_type (value_type (arg2)))
9746 {
9747 /* Automatically dereference any array reference before
9748 we attempt to perform the comparison. */
9749 arg1 = ada_coerce_ref (arg1);
9750 arg2 = ada_coerce_ref (arg2);
9751
9752 arg1 = ada_coerce_to_simple_array (arg1);
9753 arg2 = ada_coerce_to_simple_array (arg2);
9754 if (TYPE_CODE (value_type (arg1)) != TYPE_CODE_ARRAY
9755 || TYPE_CODE (value_type (arg2)) != TYPE_CODE_ARRAY)
9756 error (_("Attempt to compare array with non-array"));
9757 /* FIXME: The following works only for types whose
9758 representations use all bits (no padding or undefined bits)
9759 and do not have user-defined equality. */
9760 return
9761 TYPE_LENGTH (value_type (arg1)) == TYPE_LENGTH (value_type (arg2))
9762 && memcmp (value_contents (arg1), value_contents (arg2),
9763 TYPE_LENGTH (value_type (arg1))) == 0;
9764 }
9765 return value_equal (arg1, arg2);
9766 }
9767
9768 /* Total number of component associations in the aggregate starting at
9769 index PC in EXP. Assumes that index PC is the start of an
9770 OP_AGGREGATE. */
9771
9772 static int
9773 num_component_specs (struct expression *exp, int pc)
9774 {
9775 int n, m, i;
9776
9777 m = exp->elts[pc + 1].longconst;
9778 pc += 3;
9779 n = 0;
9780 for (i = 0; i < m; i += 1)
9781 {
9782 switch (exp->elts[pc].opcode)
9783 {
9784 default:
9785 n += 1;
9786 break;
9787 case OP_CHOICES:
9788 n += exp->elts[pc + 1].longconst;
9789 break;
9790 }
9791 ada_evaluate_subexp (NULL, exp, &pc, EVAL_SKIP);
9792 }
9793 return n;
9794 }
9795
9796 /* Assign the result of evaluating EXP starting at *POS to the INDEXth
9797 component of LHS (a simple array or a record), updating *POS past
9798 the expression, assuming that LHS is contained in CONTAINER. Does
9799 not modify the inferior's memory, nor does it modify LHS (unless
9800 LHS == CONTAINER). */
9801
9802 static void
9803 assign_component (struct value *container, struct value *lhs, LONGEST index,
9804 struct expression *exp, int *pos)
9805 {
9806 struct value *mark = value_mark ();
9807 struct value *elt;
9808
9809 if (TYPE_CODE (value_type (lhs)) == TYPE_CODE_ARRAY)
9810 {
9811 struct type *index_type = builtin_type (exp->gdbarch)->builtin_int;
9812 struct value *index_val = value_from_longest (index_type, index);
9813
9814 elt = unwrap_value (ada_value_subscript (lhs, 1, &index_val));
9815 }
9816 else
9817 {
9818 elt = ada_index_struct_field (index, lhs, 0, value_type (lhs));
9819 elt = ada_to_fixed_value (elt);
9820 }
9821
9822 if (exp->elts[*pos].opcode == OP_AGGREGATE)
9823 assign_aggregate (container, elt, exp, pos, EVAL_NORMAL);
9824 else
9825 value_assign_to_component (container, elt,
9826 ada_evaluate_subexp (NULL, exp, pos,
9827 EVAL_NORMAL));
9828
9829 value_free_to_mark (mark);
9830 }
9831
9832 /* Assuming that LHS represents an lvalue having a record or array
9833 type, and EXP->ELTS[*POS] is an OP_AGGREGATE, evaluate an assignment
9834 of that aggregate's value to LHS, advancing *POS past the
9835 aggregate. NOSIDE is as for evaluate_subexp. CONTAINER is an
9836 lvalue containing LHS (possibly LHS itself). Does not modify
9837 the inferior's memory, nor does it modify the contents of
9838 LHS (unless == CONTAINER). Returns the modified CONTAINER. */
9839
9840 static struct value *
9841 assign_aggregate (struct value *container,
9842 struct value *lhs, struct expression *exp,
9843 int *pos, enum noside noside)
9844 {
9845 struct type *lhs_type;
9846 int n = exp->elts[*pos+1].longconst;
9847 LONGEST low_index, high_index;
9848 int num_specs;
9849 LONGEST *indices;
9850 int max_indices, num_indices;
9851 int i;
9852
9853 *pos += 3;
9854 if (noside != EVAL_NORMAL)
9855 {
9856 for (i = 0; i < n; i += 1)
9857 ada_evaluate_subexp (NULL, exp, pos, noside);
9858 return container;
9859 }
9860
9861 container = ada_coerce_ref (container);
9862 if (ada_is_direct_array_type (value_type (container)))
9863 container = ada_coerce_to_simple_array (container);
9864 lhs = ada_coerce_ref (lhs);
9865 if (!deprecated_value_modifiable (lhs))
9866 error (_("Left operand of assignment is not a modifiable lvalue."));
9867
9868 lhs_type = value_type (lhs);
9869 if (ada_is_direct_array_type (lhs_type))
9870 {
9871 lhs = ada_coerce_to_simple_array (lhs);
9872 lhs_type = value_type (lhs);
9873 low_index = TYPE_ARRAY_LOWER_BOUND_VALUE (lhs_type);
9874 high_index = TYPE_ARRAY_UPPER_BOUND_VALUE (lhs_type);
9875 }
9876 else if (TYPE_CODE (lhs_type) == TYPE_CODE_STRUCT)
9877 {
9878 low_index = 0;
9879 high_index = num_visible_fields (lhs_type) - 1;
9880 }
9881 else
9882 error (_("Left-hand side must be array or record."));
9883
9884 num_specs = num_component_specs (exp, *pos - 3);
9885 max_indices = 4 * num_specs + 4;
9886 indices = XALLOCAVEC (LONGEST, max_indices);
9887 indices[0] = indices[1] = low_index - 1;
9888 indices[2] = indices[3] = high_index + 1;
9889 num_indices = 4;
9890
9891 for (i = 0; i < n; i += 1)
9892 {
9893 switch (exp->elts[*pos].opcode)
9894 {
9895 case OP_CHOICES:
9896 aggregate_assign_from_choices (container, lhs, exp, pos, indices,
9897 &num_indices, max_indices,
9898 low_index, high_index);
9899 break;
9900 case OP_POSITIONAL:
9901 aggregate_assign_positional (container, lhs, exp, pos, indices,
9902 &num_indices, max_indices,
9903 low_index, high_index);
9904 break;
9905 case OP_OTHERS:
9906 if (i != n-1)
9907 error (_("Misplaced 'others' clause"));
9908 aggregate_assign_others (container, lhs, exp, pos, indices,
9909 num_indices, low_index, high_index);
9910 break;
9911 default:
9912 error (_("Internal error: bad aggregate clause"));
9913 }
9914 }
9915
9916 return container;
9917 }
9918
9919 /* Assign into the component of LHS indexed by the OP_POSITIONAL
9920 construct at *POS, updating *POS past the construct, given that
9921 the positions are relative to lower bound LOW, where HIGH is the
9922 upper bound. Record the position in INDICES[0 .. MAX_INDICES-1]
9923 updating *NUM_INDICES as needed. CONTAINER is as for
9924 assign_aggregate. */
9925 static void
9926 aggregate_assign_positional (struct value *container,
9927 struct value *lhs, struct expression *exp,
9928 int *pos, LONGEST *indices, int *num_indices,
9929 int max_indices, LONGEST low, LONGEST high)
9930 {
9931 LONGEST ind = longest_to_int (exp->elts[*pos + 1].longconst) + low;
9932
9933 if (ind - 1 == high)
9934 warning (_("Extra components in aggregate ignored."));
9935 if (ind <= high)
9936 {
9937 add_component_interval (ind, ind, indices, num_indices, max_indices);
9938 *pos += 3;
9939 assign_component (container, lhs, ind, exp, pos);
9940 }
9941 else
9942 ada_evaluate_subexp (NULL, exp, pos, EVAL_SKIP);
9943 }
9944
9945 /* Assign into the components of LHS indexed by the OP_CHOICES
9946 construct at *POS, updating *POS past the construct, given that
9947 the allowable indices are LOW..HIGH. Record the indices assigned
9948 to in INDICES[0 .. MAX_INDICES-1], updating *NUM_INDICES as
9949 needed. CONTAINER is as for assign_aggregate. */
9950 static void
9951 aggregate_assign_from_choices (struct value *container,
9952 struct value *lhs, struct expression *exp,
9953 int *pos, LONGEST *indices, int *num_indices,
9954 int max_indices, LONGEST low, LONGEST high)
9955 {
9956 int j;
9957 int n_choices = longest_to_int (exp->elts[*pos+1].longconst);
9958 int choice_pos, expr_pc;
9959 int is_array = ada_is_direct_array_type (value_type (lhs));
9960
9961 choice_pos = *pos += 3;
9962
9963 for (j = 0; j < n_choices; j += 1)
9964 ada_evaluate_subexp (NULL, exp, pos, EVAL_SKIP);
9965 expr_pc = *pos;
9966 ada_evaluate_subexp (NULL, exp, pos, EVAL_SKIP);
9967
9968 for (j = 0; j < n_choices; j += 1)
9969 {
9970 LONGEST lower, upper;
9971 enum exp_opcode op = exp->elts[choice_pos].opcode;
9972
9973 if (op == OP_DISCRETE_RANGE)
9974 {
9975 choice_pos += 1;
9976 lower = value_as_long (ada_evaluate_subexp (NULL, exp, pos,
9977 EVAL_NORMAL));
9978 upper = value_as_long (ada_evaluate_subexp (NULL, exp, pos,
9979 EVAL_NORMAL));
9980 }
9981 else if (is_array)
9982 {
9983 lower = value_as_long (ada_evaluate_subexp (NULL, exp, &choice_pos,
9984 EVAL_NORMAL));
9985 upper = lower;
9986 }
9987 else
9988 {
9989 int ind;
9990 const char *name;
9991
9992 switch (op)
9993 {
9994 case OP_NAME:
9995 name = &exp->elts[choice_pos + 2].string;
9996 break;
9997 case OP_VAR_VALUE:
9998 name = SYMBOL_NATURAL_NAME (exp->elts[choice_pos + 2].symbol);
9999 break;
10000 default:
10001 error (_("Invalid record component association."));
10002 }
10003 ada_evaluate_subexp (NULL, exp, &choice_pos, EVAL_SKIP);
10004 ind = 0;
10005 if (! find_struct_field (name, value_type (lhs), 0,
10006 NULL, NULL, NULL, NULL, &ind))
10007 error (_("Unknown component name: %s."), name);
10008 lower = upper = ind;
10009 }
10010
10011 if (lower <= upper && (lower < low || upper > high))
10012 error (_("Index in component association out of bounds."));
10013
10014 add_component_interval (lower, upper, indices, num_indices,
10015 max_indices);
10016 while (lower <= upper)
10017 {
10018 int pos1;
10019
10020 pos1 = expr_pc;
10021 assign_component (container, lhs, lower, exp, &pos1);
10022 lower += 1;
10023 }
10024 }
10025 }
10026
10027 /* Assign the value of the expression in the OP_OTHERS construct in
10028 EXP at *POS into the components of LHS indexed from LOW .. HIGH that
10029 have not been previously assigned. The index intervals already assigned
10030 are in INDICES[0 .. NUM_INDICES-1]. Updates *POS to after the
10031 OP_OTHERS clause. CONTAINER is as for assign_aggregate. */
10032 static void
10033 aggregate_assign_others (struct value *container,
10034 struct value *lhs, struct expression *exp,
10035 int *pos, LONGEST *indices, int num_indices,
10036 LONGEST low, LONGEST high)
10037 {
10038 int i;
10039 int expr_pc = *pos + 1;
10040
10041 for (i = 0; i < num_indices - 2; i += 2)
10042 {
10043 LONGEST ind;
10044
10045 for (ind = indices[i + 1] + 1; ind < indices[i + 2]; ind += 1)
10046 {
10047 int localpos;
10048
10049 localpos = expr_pc;
10050 assign_component (container, lhs, ind, exp, &localpos);
10051 }
10052 }
10053 ada_evaluate_subexp (NULL, exp, pos, EVAL_SKIP);
10054 }
10055
10056 /* Add the interval [LOW .. HIGH] to the sorted set of intervals
10057 [ INDICES[0] .. INDICES[1] ],..., [ INDICES[*SIZE-2] .. INDICES[*SIZE-1] ],
10058 modifying *SIZE as needed. It is an error if *SIZE exceeds
10059 MAX_SIZE. The resulting intervals do not overlap. */
10060 static void
10061 add_component_interval (LONGEST low, LONGEST high,
10062 LONGEST* indices, int *size, int max_size)
10063 {
10064 int i, j;
10065
10066 for (i = 0; i < *size; i += 2) {
10067 if (high >= indices[i] && low <= indices[i + 1])
10068 {
10069 int kh;
10070
10071 for (kh = i + 2; kh < *size; kh += 2)
10072 if (high < indices[kh])
10073 break;
10074 if (low < indices[i])
10075 indices[i] = low;
10076 indices[i + 1] = indices[kh - 1];
10077 if (high > indices[i + 1])
10078 indices[i + 1] = high;
10079 memcpy (indices + i + 2, indices + kh, *size - kh);
10080 *size -= kh - i - 2;
10081 return;
10082 }
10083 else if (high < indices[i])
10084 break;
10085 }
10086
10087 if (*size == max_size)
10088 error (_("Internal error: miscounted aggregate components."));
10089 *size += 2;
10090 for (j = *size-1; j >= i+2; j -= 1)
10091 indices[j] = indices[j - 2];
10092 indices[i] = low;
10093 indices[i + 1] = high;
10094 }
10095
10096 /* Perform and Ada cast of ARG2 to type TYPE if the type of ARG2
10097 is different. */
10098
10099 static struct value *
10100 ada_value_cast (struct type *type, struct value *arg2, enum noside noside)
10101 {
10102 if (type == ada_check_typedef (value_type (arg2)))
10103 return arg2;
10104
10105 if (ada_is_fixed_point_type (type))
10106 return (cast_to_fixed (type, arg2));
10107
10108 if (ada_is_fixed_point_type (value_type (arg2)))
10109 return cast_from_fixed (type, arg2);
10110
10111 return value_cast (type, arg2);
10112 }
10113
10114 /* Evaluating Ada expressions, and printing their result.
10115 ------------------------------------------------------
10116
10117 1. Introduction:
10118 ----------------
10119
10120 We usually evaluate an Ada expression in order to print its value.
10121 We also evaluate an expression in order to print its type, which
10122 happens during the EVAL_AVOID_SIDE_EFFECTS phase of the evaluation,
10123 but we'll focus mostly on the EVAL_NORMAL phase. In practice, the
10124 EVAL_AVOID_SIDE_EFFECTS phase allows us to simplify certain aspects of
10125 the evaluation compared to the EVAL_NORMAL, but is otherwise very
10126 similar.
10127
10128 Evaluating expressions is a little more complicated for Ada entities
10129 than it is for entities in languages such as C. The main reason for
10130 this is that Ada provides types whose definition might be dynamic.
10131 One example of such types is variant records. Or another example
10132 would be an array whose bounds can only be known at run time.
10133
10134 The following description is a general guide as to what should be
10135 done (and what should NOT be done) in order to evaluate an expression
10136 involving such types, and when. This does not cover how the semantic
10137 information is encoded by GNAT as this is covered separatly. For the
10138 document used as the reference for the GNAT encoding, see exp_dbug.ads
10139 in the GNAT sources.
10140
10141 Ideally, we should embed each part of this description next to its
10142 associated code. Unfortunately, the amount of code is so vast right
10143 now that it's hard to see whether the code handling a particular
10144 situation might be duplicated or not. One day, when the code is
10145 cleaned up, this guide might become redundant with the comments
10146 inserted in the code, and we might want to remove it.
10147
10148 2. ``Fixing'' an Entity, the Simple Case:
10149 -----------------------------------------
10150
10151 When evaluating Ada expressions, the tricky issue is that they may
10152 reference entities whose type contents and size are not statically
10153 known. Consider for instance a variant record:
10154
10155 type Rec (Empty : Boolean := True) is record
10156 case Empty is
10157 when True => null;
10158 when False => Value : Integer;
10159 end case;
10160 end record;
10161 Yes : Rec := (Empty => False, Value => 1);
10162 No : Rec := (empty => True);
10163
10164 The size and contents of that record depends on the value of the
10165 descriminant (Rec.Empty). At this point, neither the debugging
10166 information nor the associated type structure in GDB are able to
10167 express such dynamic types. So what the debugger does is to create
10168 "fixed" versions of the type that applies to the specific object.
10169 We also informally refer to this opperation as "fixing" an object,
10170 which means creating its associated fixed type.
10171
10172 Example: when printing the value of variable "Yes" above, its fixed
10173 type would look like this:
10174
10175 type Rec is record
10176 Empty : Boolean;
10177 Value : Integer;
10178 end record;
10179
10180 On the other hand, if we printed the value of "No", its fixed type
10181 would become:
10182
10183 type Rec is record
10184 Empty : Boolean;
10185 end record;
10186
10187 Things become a little more complicated when trying to fix an entity
10188 with a dynamic type that directly contains another dynamic type,
10189 such as an array of variant records, for instance. There are
10190 two possible cases: Arrays, and records.
10191
10192 3. ``Fixing'' Arrays:
10193 ---------------------
10194
10195 The type structure in GDB describes an array in terms of its bounds,
10196 and the type of its elements. By design, all elements in the array
10197 have the same type and we cannot represent an array of variant elements
10198 using the current type structure in GDB. When fixing an array,
10199 we cannot fix the array element, as we would potentially need one
10200 fixed type per element of the array. As a result, the best we can do
10201 when fixing an array is to produce an array whose bounds and size
10202 are correct (allowing us to read it from memory), but without having
10203 touched its element type. Fixing each element will be done later,
10204 when (if) necessary.
10205
10206 Arrays are a little simpler to handle than records, because the same
10207 amount of memory is allocated for each element of the array, even if
10208 the amount of space actually used by each element differs from element
10209 to element. Consider for instance the following array of type Rec:
10210
10211 type Rec_Array is array (1 .. 2) of Rec;
10212
10213 The actual amount of memory occupied by each element might be different
10214 from element to element, depending on the value of their discriminant.
10215 But the amount of space reserved for each element in the array remains
10216 fixed regardless. So we simply need to compute that size using
10217 the debugging information available, from which we can then determine
10218 the array size (we multiply the number of elements of the array by
10219 the size of each element).
10220
10221 The simplest case is when we have an array of a constrained element
10222 type. For instance, consider the following type declarations:
10223
10224 type Bounded_String (Max_Size : Integer) is
10225 Length : Integer;
10226 Buffer : String (1 .. Max_Size);
10227 end record;
10228 type Bounded_String_Array is array (1 ..2) of Bounded_String (80);
10229
10230 In this case, the compiler describes the array as an array of
10231 variable-size elements (identified by its XVS suffix) for which
10232 the size can be read in the parallel XVZ variable.
10233
10234 In the case of an array of an unconstrained element type, the compiler
10235 wraps the array element inside a private PAD type. This type should not
10236 be shown to the user, and must be "unwrap"'ed before printing. Note
10237 that we also use the adjective "aligner" in our code to designate
10238 these wrapper types.
10239
10240 In some cases, the size allocated for each element is statically
10241 known. In that case, the PAD type already has the correct size,
10242 and the array element should remain unfixed.
10243
10244 But there are cases when this size is not statically known.
10245 For instance, assuming that "Five" is an integer variable:
10246
10247 type Dynamic is array (1 .. Five) of Integer;
10248 type Wrapper (Has_Length : Boolean := False) is record
10249 Data : Dynamic;
10250 case Has_Length is
10251 when True => Length : Integer;
10252 when False => null;
10253 end case;
10254 end record;
10255 type Wrapper_Array is array (1 .. 2) of Wrapper;
10256
10257 Hello : Wrapper_Array := (others => (Has_Length => True,
10258 Data => (others => 17),
10259 Length => 1));
10260
10261
10262 The debugging info would describe variable Hello as being an
10263 array of a PAD type. The size of that PAD type is not statically
10264 known, but can be determined using a parallel XVZ variable.
10265 In that case, a copy of the PAD type with the correct size should
10266 be used for the fixed array.
10267
10268 3. ``Fixing'' record type objects:
10269 ----------------------------------
10270
10271 Things are slightly different from arrays in the case of dynamic
10272 record types. In this case, in order to compute the associated
10273 fixed type, we need to determine the size and offset of each of
10274 its components. This, in turn, requires us to compute the fixed
10275 type of each of these components.
10276
10277 Consider for instance the example:
10278
10279 type Bounded_String (Max_Size : Natural) is record
10280 Str : String (1 .. Max_Size);
10281 Length : Natural;
10282 end record;
10283 My_String : Bounded_String (Max_Size => 10);
10284
10285 In that case, the position of field "Length" depends on the size
10286 of field Str, which itself depends on the value of the Max_Size
10287 discriminant. In order to fix the type of variable My_String,
10288 we need to fix the type of field Str. Therefore, fixing a variant
10289 record requires us to fix each of its components.
10290
10291 However, if a component does not have a dynamic size, the component
10292 should not be fixed. In particular, fields that use a PAD type
10293 should not fixed. Here is an example where this might happen
10294 (assuming type Rec above):
10295
10296 type Container (Big : Boolean) is record
10297 First : Rec;
10298 After : Integer;
10299 case Big is
10300 when True => Another : Integer;
10301 when False => null;
10302 end case;
10303 end record;
10304 My_Container : Container := (Big => False,
10305 First => (Empty => True),
10306 After => 42);
10307
10308 In that example, the compiler creates a PAD type for component First,
10309 whose size is constant, and then positions the component After just
10310 right after it. The offset of component After is therefore constant
10311 in this case.
10312
10313 The debugger computes the position of each field based on an algorithm
10314 that uses, among other things, the actual position and size of the field
10315 preceding it. Let's now imagine that the user is trying to print
10316 the value of My_Container. If the type fixing was recursive, we would
10317 end up computing the offset of field After based on the size of the
10318 fixed version of field First. And since in our example First has
10319 only one actual field, the size of the fixed type is actually smaller
10320 than the amount of space allocated to that field, and thus we would
10321 compute the wrong offset of field After.
10322
10323 To make things more complicated, we need to watch out for dynamic
10324 components of variant records (identified by the ___XVL suffix in
10325 the component name). Even if the target type is a PAD type, the size
10326 of that type might not be statically known. So the PAD type needs
10327 to be unwrapped and the resulting type needs to be fixed. Otherwise,
10328 we might end up with the wrong size for our component. This can be
10329 observed with the following type declarations:
10330
10331 type Octal is new Integer range 0 .. 7;
10332 type Octal_Array is array (Positive range <>) of Octal;
10333 pragma Pack (Octal_Array);
10334
10335 type Octal_Buffer (Size : Positive) is record
10336 Buffer : Octal_Array (1 .. Size);
10337 Length : Integer;
10338 end record;
10339
10340 In that case, Buffer is a PAD type whose size is unset and needs
10341 to be computed by fixing the unwrapped type.
10342
10343 4. When to ``Fix'' un-``Fixed'' sub-elements of an entity:
10344 ----------------------------------------------------------
10345
10346 Lastly, when should the sub-elements of an entity that remained unfixed
10347 thus far, be actually fixed?
10348
10349 The answer is: Only when referencing that element. For instance
10350 when selecting one component of a record, this specific component
10351 should be fixed at that point in time. Or when printing the value
10352 of a record, each component should be fixed before its value gets
10353 printed. Similarly for arrays, the element of the array should be
10354 fixed when printing each element of the array, or when extracting
10355 one element out of that array. On the other hand, fixing should
10356 not be performed on the elements when taking a slice of an array!
10357
10358 Note that one of the side-effects of miscomputing the offset and
10359 size of each field is that we end up also miscomputing the size
10360 of the containing type. This can have adverse results when computing
10361 the value of an entity. GDB fetches the value of an entity based
10362 on the size of its type, and thus a wrong size causes GDB to fetch
10363 the wrong amount of memory. In the case where the computed size is
10364 too small, GDB fetches too little data to print the value of our
10365 entiry. Results in this case as unpredicatble, as we usually read
10366 past the buffer containing the data =:-o. */
10367
10368 /* Implement the evaluate_exp routine in the exp_descriptor structure
10369 for the Ada language. */
10370
10371 static struct value *
10372 ada_evaluate_subexp (struct type *expect_type, struct expression *exp,
10373 int *pos, enum noside noside)
10374 {
10375 enum exp_opcode op;
10376 int tem;
10377 int pc;
10378 int preeval_pos;
10379 struct value *arg1 = NULL, *arg2 = NULL, *arg3;
10380 struct type *type;
10381 int nargs, oplen;
10382 struct value **argvec;
10383
10384 pc = *pos;
10385 *pos += 1;
10386 op = exp->elts[pc].opcode;
10387
10388 switch (op)
10389 {
10390 default:
10391 *pos -= 1;
10392 arg1 = evaluate_subexp_standard (expect_type, exp, pos, noside);
10393
10394 if (noside == EVAL_NORMAL)
10395 arg1 = unwrap_value (arg1);
10396
10397 /* If evaluating an OP_DOUBLE and an EXPECT_TYPE was provided,
10398 then we need to perform the conversion manually, because
10399 evaluate_subexp_standard doesn't do it. This conversion is
10400 necessary in Ada because the different kinds of float/fixed
10401 types in Ada have different representations.
10402
10403 Similarly, we need to perform the conversion from OP_LONG
10404 ourselves. */
10405 if ((op == OP_DOUBLE || op == OP_LONG) && expect_type != NULL)
10406 arg1 = ada_value_cast (expect_type, arg1, noside);
10407
10408 return arg1;
10409
10410 case OP_STRING:
10411 {
10412 struct value *result;
10413
10414 *pos -= 1;
10415 result = evaluate_subexp_standard (expect_type, exp, pos, noside);
10416 /* The result type will have code OP_STRING, bashed there from
10417 OP_ARRAY. Bash it back. */
10418 if (TYPE_CODE (value_type (result)) == TYPE_CODE_STRING)
10419 TYPE_CODE (value_type (result)) = TYPE_CODE_ARRAY;
10420 return result;
10421 }
10422
10423 case UNOP_CAST:
10424 (*pos) += 2;
10425 type = exp->elts[pc + 1].type;
10426 arg1 = evaluate_subexp (type, exp, pos, noside);
10427 if (noside == EVAL_SKIP)
10428 goto nosideret;
10429 arg1 = ada_value_cast (type, arg1, noside);
10430 return arg1;
10431
10432 case UNOP_QUAL:
10433 (*pos) += 2;
10434 type = exp->elts[pc + 1].type;
10435 return ada_evaluate_subexp (type, exp, pos, noside);
10436
10437 case BINOP_ASSIGN:
10438 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10439 if (exp->elts[*pos].opcode == OP_AGGREGATE)
10440 {
10441 arg1 = assign_aggregate (arg1, arg1, exp, pos, noside);
10442 if (noside == EVAL_SKIP || noside == EVAL_AVOID_SIDE_EFFECTS)
10443 return arg1;
10444 return ada_value_assign (arg1, arg1);
10445 }
10446 /* Force the evaluation of the rhs ARG2 to the type of the lhs ARG1,
10447 except if the lhs of our assignment is a convenience variable.
10448 In the case of assigning to a convenience variable, the lhs
10449 should be exactly the result of the evaluation of the rhs. */
10450 type = value_type (arg1);
10451 if (VALUE_LVAL (arg1) == lval_internalvar)
10452 type = NULL;
10453 arg2 = evaluate_subexp (type, exp, pos, noside);
10454 if (noside == EVAL_SKIP || noside == EVAL_AVOID_SIDE_EFFECTS)
10455 return arg1;
10456 if (ada_is_fixed_point_type (value_type (arg1)))
10457 arg2 = cast_to_fixed (value_type (arg1), arg2);
10458 else if (ada_is_fixed_point_type (value_type (arg2)))
10459 error
10460 (_("Fixed-point values must be assigned to fixed-point variables"));
10461 else
10462 arg2 = coerce_for_assign (value_type (arg1), arg2);
10463 return ada_value_assign (arg1, arg2);
10464
10465 case BINOP_ADD:
10466 arg1 = evaluate_subexp_with_coercion (exp, pos, noside);
10467 arg2 = evaluate_subexp_with_coercion (exp, pos, noside);
10468 if (noside == EVAL_SKIP)
10469 goto nosideret;
10470 if (TYPE_CODE (value_type (arg1)) == TYPE_CODE_PTR)
10471 return (value_from_longest
10472 (value_type (arg1),
10473 value_as_long (arg1) + value_as_long (arg2)));
10474 if (TYPE_CODE (value_type (arg2)) == TYPE_CODE_PTR)
10475 return (value_from_longest
10476 (value_type (arg2),
10477 value_as_long (arg1) + value_as_long (arg2)));
10478 if ((ada_is_fixed_point_type (value_type (arg1))
10479 || ada_is_fixed_point_type (value_type (arg2)))
10480 && value_type (arg1) != value_type (arg2))
10481 error (_("Operands of fixed-point addition must have the same type"));
10482 /* Do the addition, and cast the result to the type of the first
10483 argument. We cannot cast the result to a reference type, so if
10484 ARG1 is a reference type, find its underlying type. */
10485 type = value_type (arg1);
10486 while (TYPE_CODE (type) == TYPE_CODE_REF)
10487 type = TYPE_TARGET_TYPE (type);
10488 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10489 return value_cast (type, value_binop (arg1, arg2, BINOP_ADD));
10490
10491 case BINOP_SUB:
10492 arg1 = evaluate_subexp_with_coercion (exp, pos, noside);
10493 arg2 = evaluate_subexp_with_coercion (exp, pos, noside);
10494 if (noside == EVAL_SKIP)
10495 goto nosideret;
10496 if (TYPE_CODE (value_type (arg1)) == TYPE_CODE_PTR)
10497 return (value_from_longest
10498 (value_type (arg1),
10499 value_as_long (arg1) - value_as_long (arg2)));
10500 if (TYPE_CODE (value_type (arg2)) == TYPE_CODE_PTR)
10501 return (value_from_longest
10502 (value_type (arg2),
10503 value_as_long (arg1) - value_as_long (arg2)));
10504 if ((ada_is_fixed_point_type (value_type (arg1))
10505 || ada_is_fixed_point_type (value_type (arg2)))
10506 && value_type (arg1) != value_type (arg2))
10507 error (_("Operands of fixed-point subtraction "
10508 "must have the same type"));
10509 /* Do the substraction, and cast the result to the type of the first
10510 argument. We cannot cast the result to a reference type, so if
10511 ARG1 is a reference type, find its underlying type. */
10512 type = value_type (arg1);
10513 while (TYPE_CODE (type) == TYPE_CODE_REF)
10514 type = TYPE_TARGET_TYPE (type);
10515 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10516 return value_cast (type, value_binop (arg1, arg2, BINOP_SUB));
10517
10518 case BINOP_MUL:
10519 case BINOP_DIV:
10520 case BINOP_REM:
10521 case BINOP_MOD:
10522 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10523 arg2 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10524 if (noside == EVAL_SKIP)
10525 goto nosideret;
10526 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
10527 {
10528 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10529 return value_zero (value_type (arg1), not_lval);
10530 }
10531 else
10532 {
10533 type = builtin_type (exp->gdbarch)->builtin_double;
10534 if (ada_is_fixed_point_type (value_type (arg1)))
10535 arg1 = cast_from_fixed (type, arg1);
10536 if (ada_is_fixed_point_type (value_type (arg2)))
10537 arg2 = cast_from_fixed (type, arg2);
10538 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10539 return ada_value_binop (arg1, arg2, op);
10540 }
10541
10542 case BINOP_EQUAL:
10543 case BINOP_NOTEQUAL:
10544 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10545 arg2 = evaluate_subexp (value_type (arg1), exp, pos, noside);
10546 if (noside == EVAL_SKIP)
10547 goto nosideret;
10548 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10549 tem = 0;
10550 else
10551 {
10552 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10553 tem = ada_value_equal (arg1, arg2);
10554 }
10555 if (op == BINOP_NOTEQUAL)
10556 tem = !tem;
10557 type = language_bool_type (exp->language_defn, exp->gdbarch);
10558 return value_from_longest (type, (LONGEST) tem);
10559
10560 case UNOP_NEG:
10561 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10562 if (noside == EVAL_SKIP)
10563 goto nosideret;
10564 else if (ada_is_fixed_point_type (value_type (arg1)))
10565 return value_cast (value_type (arg1), value_neg (arg1));
10566 else
10567 {
10568 unop_promote (exp->language_defn, exp->gdbarch, &arg1);
10569 return value_neg (arg1);
10570 }
10571
10572 case BINOP_LOGICAL_AND:
10573 case BINOP_LOGICAL_OR:
10574 case UNOP_LOGICAL_NOT:
10575 {
10576 struct value *val;
10577
10578 *pos -= 1;
10579 val = evaluate_subexp_standard (expect_type, exp, pos, noside);
10580 type = language_bool_type (exp->language_defn, exp->gdbarch);
10581 return value_cast (type, val);
10582 }
10583
10584 case BINOP_BITWISE_AND:
10585 case BINOP_BITWISE_IOR:
10586 case BINOP_BITWISE_XOR:
10587 {
10588 struct value *val;
10589
10590 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, EVAL_AVOID_SIDE_EFFECTS);
10591 *pos = pc;
10592 val = evaluate_subexp_standard (expect_type, exp, pos, noside);
10593
10594 return value_cast (value_type (arg1), val);
10595 }
10596
10597 case OP_VAR_VALUE:
10598 *pos -= 1;
10599
10600 if (noside == EVAL_SKIP)
10601 {
10602 *pos += 4;
10603 goto nosideret;
10604 }
10605
10606 if (SYMBOL_DOMAIN (exp->elts[pc + 2].symbol) == UNDEF_DOMAIN)
10607 /* Only encountered when an unresolved symbol occurs in a
10608 context other than a function call, in which case, it is
10609 invalid. */
10610 error (_("Unexpected unresolved symbol, %s, during evaluation"),
10611 SYMBOL_PRINT_NAME (exp->elts[pc + 2].symbol));
10612
10613 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10614 {
10615 type = static_unwrap_type (SYMBOL_TYPE (exp->elts[pc + 2].symbol));
10616 /* Check to see if this is a tagged type. We also need to handle
10617 the case where the type is a reference to a tagged type, but
10618 we have to be careful to exclude pointers to tagged types.
10619 The latter should be shown as usual (as a pointer), whereas
10620 a reference should mostly be transparent to the user. */
10621 if (ada_is_tagged_type (type, 0)
10622 || (TYPE_CODE (type) == TYPE_CODE_REF
10623 && ada_is_tagged_type (TYPE_TARGET_TYPE (type), 0)))
10624 {
10625 /* Tagged types are a little special in the fact that the real
10626 type is dynamic and can only be determined by inspecting the
10627 object's tag. This means that we need to get the object's
10628 value first (EVAL_NORMAL) and then extract the actual object
10629 type from its tag.
10630
10631 Note that we cannot skip the final step where we extract
10632 the object type from its tag, because the EVAL_NORMAL phase
10633 results in dynamic components being resolved into fixed ones.
10634 This can cause problems when trying to print the type
10635 description of tagged types whose parent has a dynamic size:
10636 We use the type name of the "_parent" component in order
10637 to print the name of the ancestor type in the type description.
10638 If that component had a dynamic size, the resolution into
10639 a fixed type would result in the loss of that type name,
10640 thus preventing us from printing the name of the ancestor
10641 type in the type description. */
10642 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, EVAL_NORMAL);
10643
10644 if (TYPE_CODE (type) != TYPE_CODE_REF)
10645 {
10646 struct type *actual_type;
10647
10648 actual_type = type_from_tag (ada_value_tag (arg1));
10649 if (actual_type == NULL)
10650 /* If, for some reason, we were unable to determine
10651 the actual type from the tag, then use the static
10652 approximation that we just computed as a fallback.
10653 This can happen if the debugging information is
10654 incomplete, for instance. */
10655 actual_type = type;
10656 return value_zero (actual_type, not_lval);
10657 }
10658 else
10659 {
10660 /* In the case of a ref, ada_coerce_ref takes care
10661 of determining the actual type. But the evaluation
10662 should return a ref as it should be valid to ask
10663 for its address; so rebuild a ref after coerce. */
10664 arg1 = ada_coerce_ref (arg1);
10665 return value_ref (arg1);
10666 }
10667 }
10668
10669 /* Records and unions for which GNAT encodings have been
10670 generated need to be statically fixed as well.
10671 Otherwise, non-static fixing produces a type where
10672 all dynamic properties are removed, which prevents "ptype"
10673 from being able to completely describe the type.
10674 For instance, a case statement in a variant record would be
10675 replaced by the relevant components based on the actual
10676 value of the discriminants. */
10677 if ((TYPE_CODE (type) == TYPE_CODE_STRUCT
10678 && dynamic_template_type (type) != NULL)
10679 || (TYPE_CODE (type) == TYPE_CODE_UNION
10680 && ada_find_parallel_type (type, "___XVU") != NULL))
10681 {
10682 *pos += 4;
10683 return value_zero (to_static_fixed_type (type), not_lval);
10684 }
10685 }
10686
10687 arg1 = evaluate_subexp_standard (expect_type, exp, pos, noside);
10688 return ada_to_fixed_value (arg1);
10689
10690 case OP_FUNCALL:
10691 (*pos) += 2;
10692
10693 /* Allocate arg vector, including space for the function to be
10694 called in argvec[0] and a terminating NULL. */
10695 nargs = longest_to_int (exp->elts[pc + 1].longconst);
10696 argvec = XALLOCAVEC (struct value *, nargs + 2);
10697
10698 if (exp->elts[*pos].opcode == OP_VAR_VALUE
10699 && SYMBOL_DOMAIN (exp->elts[pc + 5].symbol) == UNDEF_DOMAIN)
10700 error (_("Unexpected unresolved symbol, %s, during evaluation"),
10701 SYMBOL_PRINT_NAME (exp->elts[pc + 5].symbol));
10702 else
10703 {
10704 for (tem = 0; tem <= nargs; tem += 1)
10705 argvec[tem] = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10706 argvec[tem] = 0;
10707
10708 if (noside == EVAL_SKIP)
10709 goto nosideret;
10710 }
10711
10712 if (ada_is_constrained_packed_array_type
10713 (desc_base_type (value_type (argvec[0]))))
10714 argvec[0] = ada_coerce_to_simple_array (argvec[0]);
10715 else if (TYPE_CODE (value_type (argvec[0])) == TYPE_CODE_ARRAY
10716 && TYPE_FIELD_BITSIZE (value_type (argvec[0]), 0) != 0)
10717 /* This is a packed array that has already been fixed, and
10718 therefore already coerced to a simple array. Nothing further
10719 to do. */
10720 ;
10721 else if (TYPE_CODE (value_type (argvec[0])) == TYPE_CODE_REF)
10722 {
10723 /* Make sure we dereference references so that all the code below
10724 feels like it's really handling the referenced value. Wrapping
10725 types (for alignment) may be there, so make sure we strip them as
10726 well. */
10727 argvec[0] = ada_to_fixed_value (coerce_ref (argvec[0]));
10728 }
10729 else if (TYPE_CODE (value_type (argvec[0])) == TYPE_CODE_ARRAY
10730 && VALUE_LVAL (argvec[0]) == lval_memory)
10731 argvec[0] = value_addr (argvec[0]);
10732
10733 type = ada_check_typedef (value_type (argvec[0]));
10734
10735 /* Ada allows us to implicitly dereference arrays when subscripting
10736 them. So, if this is an array typedef (encoding use for array
10737 access types encoded as fat pointers), strip it now. */
10738 if (TYPE_CODE (type) == TYPE_CODE_TYPEDEF)
10739 type = ada_typedef_target_type (type);
10740
10741 if (TYPE_CODE (type) == TYPE_CODE_PTR)
10742 {
10743 switch (TYPE_CODE (ada_check_typedef (TYPE_TARGET_TYPE (type))))
10744 {
10745 case TYPE_CODE_FUNC:
10746 type = ada_check_typedef (TYPE_TARGET_TYPE (type));
10747 break;
10748 case TYPE_CODE_ARRAY:
10749 break;
10750 case TYPE_CODE_STRUCT:
10751 if (noside != EVAL_AVOID_SIDE_EFFECTS)
10752 argvec[0] = ada_value_ind (argvec[0]);
10753 type = ada_check_typedef (TYPE_TARGET_TYPE (type));
10754 break;
10755 default:
10756 error (_("cannot subscript or call something of type `%s'"),
10757 ada_type_name (value_type (argvec[0])));
10758 break;
10759 }
10760 }
10761
10762 switch (TYPE_CODE (type))
10763 {
10764 case TYPE_CODE_FUNC:
10765 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10766 {
10767 struct type *rtype = TYPE_TARGET_TYPE (type);
10768
10769 if (TYPE_GNU_IFUNC (type))
10770 return allocate_value (TYPE_TARGET_TYPE (rtype));
10771 return allocate_value (rtype);
10772 }
10773 return call_function_by_hand (argvec[0], nargs, argvec + 1);
10774 case TYPE_CODE_INTERNAL_FUNCTION:
10775 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10776 /* We don't know anything about what the internal
10777 function might return, but we have to return
10778 something. */
10779 return value_zero (builtin_type (exp->gdbarch)->builtin_int,
10780 not_lval);
10781 else
10782 return call_internal_function (exp->gdbarch, exp->language_defn,
10783 argvec[0], nargs, argvec + 1);
10784
10785 case TYPE_CODE_STRUCT:
10786 {
10787 int arity;
10788
10789 arity = ada_array_arity (type);
10790 type = ada_array_element_type (type, nargs);
10791 if (type == NULL)
10792 error (_("cannot subscript or call a record"));
10793 if (arity != nargs)
10794 error (_("wrong number of subscripts; expecting %d"), arity);
10795 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10796 return value_zero (ada_aligned_type (type), lval_memory);
10797 return
10798 unwrap_value (ada_value_subscript
10799 (argvec[0], nargs, argvec + 1));
10800 }
10801 case TYPE_CODE_ARRAY:
10802 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10803 {
10804 type = ada_array_element_type (type, nargs);
10805 if (type == NULL)
10806 error (_("element type of array unknown"));
10807 else
10808 return value_zero (ada_aligned_type (type), lval_memory);
10809 }
10810 return
10811 unwrap_value (ada_value_subscript
10812 (ada_coerce_to_simple_array (argvec[0]),
10813 nargs, argvec + 1));
10814 case TYPE_CODE_PTR: /* Pointer to array */
10815 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10816 {
10817 type = to_fixed_array_type (TYPE_TARGET_TYPE (type), NULL, 1);
10818 type = ada_array_element_type (type, nargs);
10819 if (type == NULL)
10820 error (_("element type of array unknown"));
10821 else
10822 return value_zero (ada_aligned_type (type), lval_memory);
10823 }
10824 return
10825 unwrap_value (ada_value_ptr_subscript (argvec[0],
10826 nargs, argvec + 1));
10827
10828 default:
10829 error (_("Attempt to index or call something other than an "
10830 "array or function"));
10831 }
10832
10833 case TERNOP_SLICE:
10834 {
10835 struct value *array = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10836 struct value *low_bound_val =
10837 evaluate_subexp (NULL_TYPE, exp, pos, noside);
10838 struct value *high_bound_val =
10839 evaluate_subexp (NULL_TYPE, exp, pos, noside);
10840 LONGEST low_bound;
10841 LONGEST high_bound;
10842
10843 low_bound_val = coerce_ref (low_bound_val);
10844 high_bound_val = coerce_ref (high_bound_val);
10845 low_bound = value_as_long (low_bound_val);
10846 high_bound = value_as_long (high_bound_val);
10847
10848 if (noside == EVAL_SKIP)
10849 goto nosideret;
10850
10851 /* If this is a reference to an aligner type, then remove all
10852 the aligners. */
10853 if (TYPE_CODE (value_type (array)) == TYPE_CODE_REF
10854 && ada_is_aligner_type (TYPE_TARGET_TYPE (value_type (array))))
10855 TYPE_TARGET_TYPE (value_type (array)) =
10856 ada_aligned_type (TYPE_TARGET_TYPE (value_type (array)));
10857
10858 if (ada_is_constrained_packed_array_type (value_type (array)))
10859 error (_("cannot slice a packed array"));
10860
10861 /* If this is a reference to an array or an array lvalue,
10862 convert to a pointer. */
10863 if (TYPE_CODE (value_type (array)) == TYPE_CODE_REF
10864 || (TYPE_CODE (value_type (array)) == TYPE_CODE_ARRAY
10865 && VALUE_LVAL (array) == lval_memory))
10866 array = value_addr (array);
10867
10868 if (noside == EVAL_AVOID_SIDE_EFFECTS
10869 && ada_is_array_descriptor_type (ada_check_typedef
10870 (value_type (array))))
10871 return empty_array (ada_type_of_array (array, 0), low_bound);
10872
10873 array = ada_coerce_to_simple_array_ptr (array);
10874
10875 /* If we have more than one level of pointer indirection,
10876 dereference the value until we get only one level. */
10877 while (TYPE_CODE (value_type (array)) == TYPE_CODE_PTR
10878 && (TYPE_CODE (TYPE_TARGET_TYPE (value_type (array)))
10879 == TYPE_CODE_PTR))
10880 array = value_ind (array);
10881
10882 /* Make sure we really do have an array type before going further,
10883 to avoid a SEGV when trying to get the index type or the target
10884 type later down the road if the debug info generated by
10885 the compiler is incorrect or incomplete. */
10886 if (!ada_is_simple_array_type (value_type (array)))
10887 error (_("cannot take slice of non-array"));
10888
10889 if (TYPE_CODE (ada_check_typedef (value_type (array)))
10890 == TYPE_CODE_PTR)
10891 {
10892 struct type *type0 = ada_check_typedef (value_type (array));
10893
10894 if (high_bound < low_bound || noside == EVAL_AVOID_SIDE_EFFECTS)
10895 return empty_array (TYPE_TARGET_TYPE (type0), low_bound);
10896 else
10897 {
10898 struct type *arr_type0 =
10899 to_fixed_array_type (TYPE_TARGET_TYPE (type0), NULL, 1);
10900
10901 return ada_value_slice_from_ptr (array, arr_type0,
10902 longest_to_int (low_bound),
10903 longest_to_int (high_bound));
10904 }
10905 }
10906 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
10907 return array;
10908 else if (high_bound < low_bound)
10909 return empty_array (value_type (array), low_bound);
10910 else
10911 return ada_value_slice (array, longest_to_int (low_bound),
10912 longest_to_int (high_bound));
10913 }
10914
10915 case UNOP_IN_RANGE:
10916 (*pos) += 2;
10917 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10918 type = check_typedef (exp->elts[pc + 1].type);
10919
10920 if (noside == EVAL_SKIP)
10921 goto nosideret;
10922
10923 switch (TYPE_CODE (type))
10924 {
10925 default:
10926 lim_warning (_("Membership test incompletely implemented; "
10927 "always returns true"));
10928 type = language_bool_type (exp->language_defn, exp->gdbarch);
10929 return value_from_longest (type, (LONGEST) 1);
10930
10931 case TYPE_CODE_RANGE:
10932 arg2 = value_from_longest (type, TYPE_LOW_BOUND (type));
10933 arg3 = value_from_longest (type, TYPE_HIGH_BOUND (type));
10934 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10935 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3);
10936 type = language_bool_type (exp->language_defn, exp->gdbarch);
10937 return
10938 value_from_longest (type,
10939 (value_less (arg1, arg3)
10940 || value_equal (arg1, arg3))
10941 && (value_less (arg2, arg1)
10942 || value_equal (arg2, arg1)));
10943 }
10944
10945 case BINOP_IN_BOUNDS:
10946 (*pos) += 2;
10947 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10948 arg2 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10949
10950 if (noside == EVAL_SKIP)
10951 goto nosideret;
10952
10953 if (noside == EVAL_AVOID_SIDE_EFFECTS)
10954 {
10955 type = language_bool_type (exp->language_defn, exp->gdbarch);
10956 return value_zero (type, not_lval);
10957 }
10958
10959 tem = longest_to_int (exp->elts[pc + 1].longconst);
10960
10961 type = ada_index_type (value_type (arg2), tem, "range");
10962 if (!type)
10963 type = value_type (arg1);
10964
10965 arg3 = value_from_longest (type, ada_array_bound (arg2, tem, 1));
10966 arg2 = value_from_longest (type, ada_array_bound (arg2, tem, 0));
10967
10968 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10969 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3);
10970 type = language_bool_type (exp->language_defn, exp->gdbarch);
10971 return
10972 value_from_longest (type,
10973 (value_less (arg1, arg3)
10974 || value_equal (arg1, arg3))
10975 && (value_less (arg2, arg1)
10976 || value_equal (arg2, arg1)));
10977
10978 case TERNOP_IN_RANGE:
10979 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10980 arg2 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10981 arg3 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
10982
10983 if (noside == EVAL_SKIP)
10984 goto nosideret;
10985
10986 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
10987 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3);
10988 type = language_bool_type (exp->language_defn, exp->gdbarch);
10989 return
10990 value_from_longest (type,
10991 (value_less (arg1, arg3)
10992 || value_equal (arg1, arg3))
10993 && (value_less (arg2, arg1)
10994 || value_equal (arg2, arg1)));
10995
10996 case OP_ATR_FIRST:
10997 case OP_ATR_LAST:
10998 case OP_ATR_LENGTH:
10999 {
11000 struct type *type_arg;
11001
11002 if (exp->elts[*pos].opcode == OP_TYPE)
11003 {
11004 evaluate_subexp (NULL_TYPE, exp, pos, EVAL_SKIP);
11005 arg1 = NULL;
11006 type_arg = check_typedef (exp->elts[pc + 2].type);
11007 }
11008 else
11009 {
11010 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11011 type_arg = NULL;
11012 }
11013
11014 if (exp->elts[*pos].opcode != OP_LONG)
11015 error (_("Invalid operand to '%s"), ada_attribute_name (op));
11016 tem = longest_to_int (exp->elts[*pos + 2].longconst);
11017 *pos += 4;
11018
11019 if (noside == EVAL_SKIP)
11020 goto nosideret;
11021
11022 if (type_arg == NULL)
11023 {
11024 arg1 = ada_coerce_ref (arg1);
11025
11026 if (ada_is_constrained_packed_array_type (value_type (arg1)))
11027 arg1 = ada_coerce_to_simple_array (arg1);
11028
11029 if (op == OP_ATR_LENGTH)
11030 type = builtin_type (exp->gdbarch)->builtin_int;
11031 else
11032 {
11033 type = ada_index_type (value_type (arg1), tem,
11034 ada_attribute_name (op));
11035 if (type == NULL)
11036 type = builtin_type (exp->gdbarch)->builtin_int;
11037 }
11038
11039 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11040 return allocate_value (type);
11041
11042 switch (op)
11043 {
11044 default: /* Should never happen. */
11045 error (_("unexpected attribute encountered"));
11046 case OP_ATR_FIRST:
11047 return value_from_longest
11048 (type, ada_array_bound (arg1, tem, 0));
11049 case OP_ATR_LAST:
11050 return value_from_longest
11051 (type, ada_array_bound (arg1, tem, 1));
11052 case OP_ATR_LENGTH:
11053 return value_from_longest
11054 (type, ada_array_length (arg1, tem));
11055 }
11056 }
11057 else if (discrete_type_p (type_arg))
11058 {
11059 struct type *range_type;
11060 const char *name = ada_type_name (type_arg);
11061
11062 range_type = NULL;
11063 if (name != NULL && TYPE_CODE (type_arg) != TYPE_CODE_ENUM)
11064 range_type = to_fixed_range_type (type_arg, NULL);
11065 if (range_type == NULL)
11066 range_type = type_arg;
11067 switch (op)
11068 {
11069 default:
11070 error (_("unexpected attribute encountered"));
11071 case OP_ATR_FIRST:
11072 return value_from_longest
11073 (range_type, ada_discrete_type_low_bound (range_type));
11074 case OP_ATR_LAST:
11075 return value_from_longest
11076 (range_type, ada_discrete_type_high_bound (range_type));
11077 case OP_ATR_LENGTH:
11078 error (_("the 'length attribute applies only to array types"));
11079 }
11080 }
11081 else if (TYPE_CODE (type_arg) == TYPE_CODE_FLT)
11082 error (_("unimplemented type attribute"));
11083 else
11084 {
11085 LONGEST low, high;
11086
11087 if (ada_is_constrained_packed_array_type (type_arg))
11088 type_arg = decode_constrained_packed_array_type (type_arg);
11089
11090 if (op == OP_ATR_LENGTH)
11091 type = builtin_type (exp->gdbarch)->builtin_int;
11092 else
11093 {
11094 type = ada_index_type (type_arg, tem, ada_attribute_name (op));
11095 if (type == NULL)
11096 type = builtin_type (exp->gdbarch)->builtin_int;
11097 }
11098
11099 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11100 return allocate_value (type);
11101
11102 switch (op)
11103 {
11104 default:
11105 error (_("unexpected attribute encountered"));
11106 case OP_ATR_FIRST:
11107 low = ada_array_bound_from_type (type_arg, tem, 0);
11108 return value_from_longest (type, low);
11109 case OP_ATR_LAST:
11110 high = ada_array_bound_from_type (type_arg, tem, 1);
11111 return value_from_longest (type, high);
11112 case OP_ATR_LENGTH:
11113 low = ada_array_bound_from_type (type_arg, tem, 0);
11114 high = ada_array_bound_from_type (type_arg, tem, 1);
11115 return value_from_longest (type, high - low + 1);
11116 }
11117 }
11118 }
11119
11120 case OP_ATR_TAG:
11121 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11122 if (noside == EVAL_SKIP)
11123 goto nosideret;
11124
11125 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11126 return value_zero (ada_tag_type (arg1), not_lval);
11127
11128 return ada_value_tag (arg1);
11129
11130 case OP_ATR_MIN:
11131 case OP_ATR_MAX:
11132 evaluate_subexp (NULL_TYPE, exp, pos, EVAL_SKIP);
11133 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11134 arg2 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11135 if (noside == EVAL_SKIP)
11136 goto nosideret;
11137 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
11138 return value_zero (value_type (arg1), not_lval);
11139 else
11140 {
11141 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
11142 return value_binop (arg1, arg2,
11143 op == OP_ATR_MIN ? BINOP_MIN : BINOP_MAX);
11144 }
11145
11146 case OP_ATR_MODULUS:
11147 {
11148 struct type *type_arg = check_typedef (exp->elts[pc + 2].type);
11149
11150 evaluate_subexp (NULL_TYPE, exp, pos, EVAL_SKIP);
11151 if (noside == EVAL_SKIP)
11152 goto nosideret;
11153
11154 if (!ada_is_modular_type (type_arg))
11155 error (_("'modulus must be applied to modular type"));
11156
11157 return value_from_longest (TYPE_TARGET_TYPE (type_arg),
11158 ada_modulus (type_arg));
11159 }
11160
11161
11162 case OP_ATR_POS:
11163 evaluate_subexp (NULL_TYPE, exp, pos, EVAL_SKIP);
11164 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11165 if (noside == EVAL_SKIP)
11166 goto nosideret;
11167 type = builtin_type (exp->gdbarch)->builtin_int;
11168 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11169 return value_zero (type, not_lval);
11170 else
11171 return value_pos_atr (type, arg1);
11172
11173 case OP_ATR_SIZE:
11174 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11175 type = value_type (arg1);
11176
11177 /* If the argument is a reference, then dereference its type, since
11178 the user is really asking for the size of the actual object,
11179 not the size of the pointer. */
11180 if (TYPE_CODE (type) == TYPE_CODE_REF)
11181 type = TYPE_TARGET_TYPE (type);
11182
11183 if (noside == EVAL_SKIP)
11184 goto nosideret;
11185 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
11186 return value_zero (builtin_type (exp->gdbarch)->builtin_int, not_lval);
11187 else
11188 return value_from_longest (builtin_type (exp->gdbarch)->builtin_int,
11189 TARGET_CHAR_BIT * TYPE_LENGTH (type));
11190
11191 case OP_ATR_VAL:
11192 evaluate_subexp (NULL_TYPE, exp, pos, EVAL_SKIP);
11193 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11194 type = exp->elts[pc + 2].type;
11195 if (noside == EVAL_SKIP)
11196 goto nosideret;
11197 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
11198 return value_zero (type, not_lval);
11199 else
11200 return value_val_atr (type, arg1);
11201
11202 case BINOP_EXP:
11203 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11204 arg2 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11205 if (noside == EVAL_SKIP)
11206 goto nosideret;
11207 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
11208 return value_zero (value_type (arg1), not_lval);
11209 else
11210 {
11211 /* For integer exponentiation operations,
11212 only promote the first argument. */
11213 if (is_integral_type (value_type (arg2)))
11214 unop_promote (exp->language_defn, exp->gdbarch, &arg1);
11215 else
11216 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2);
11217
11218 return value_binop (arg1, arg2, op);
11219 }
11220
11221 case UNOP_PLUS:
11222 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11223 if (noside == EVAL_SKIP)
11224 goto nosideret;
11225 else
11226 return arg1;
11227
11228 case UNOP_ABS:
11229 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11230 if (noside == EVAL_SKIP)
11231 goto nosideret;
11232 unop_promote (exp->language_defn, exp->gdbarch, &arg1);
11233 if (value_less (arg1, value_zero (value_type (arg1), not_lval)))
11234 return value_neg (arg1);
11235 else
11236 return arg1;
11237
11238 case UNOP_IND:
11239 preeval_pos = *pos;
11240 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11241 if (noside == EVAL_SKIP)
11242 goto nosideret;
11243 type = ada_check_typedef (value_type (arg1));
11244 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11245 {
11246 if (ada_is_array_descriptor_type (type))
11247 /* GDB allows dereferencing GNAT array descriptors. */
11248 {
11249 struct type *arrType = ada_type_of_array (arg1, 0);
11250
11251 if (arrType == NULL)
11252 error (_("Attempt to dereference null array pointer."));
11253 return value_at_lazy (arrType, 0);
11254 }
11255 else if (TYPE_CODE (type) == TYPE_CODE_PTR
11256 || TYPE_CODE (type) == TYPE_CODE_REF
11257 /* In C you can dereference an array to get the 1st elt. */
11258 || TYPE_CODE (type) == TYPE_CODE_ARRAY)
11259 {
11260 /* As mentioned in the OP_VAR_VALUE case, tagged types can
11261 only be determined by inspecting the object's tag.
11262 This means that we need to evaluate completely the
11263 expression in order to get its type. */
11264
11265 if ((TYPE_CODE (type) == TYPE_CODE_REF
11266 || TYPE_CODE (type) == TYPE_CODE_PTR)
11267 && ada_is_tagged_type (TYPE_TARGET_TYPE (type), 0))
11268 {
11269 arg1 = evaluate_subexp (NULL_TYPE, exp, &preeval_pos,
11270 EVAL_NORMAL);
11271 type = value_type (ada_value_ind (arg1));
11272 }
11273 else
11274 {
11275 type = to_static_fixed_type
11276 (ada_aligned_type
11277 (ada_check_typedef (TYPE_TARGET_TYPE (type))));
11278 }
11279 ada_ensure_varsize_limit (type);
11280 return value_zero (type, lval_memory);
11281 }
11282 else if (TYPE_CODE (type) == TYPE_CODE_INT)
11283 {
11284 /* GDB allows dereferencing an int. */
11285 if (expect_type == NULL)
11286 return value_zero (builtin_type (exp->gdbarch)->builtin_int,
11287 lval_memory);
11288 else
11289 {
11290 expect_type =
11291 to_static_fixed_type (ada_aligned_type (expect_type));
11292 return value_zero (expect_type, lval_memory);
11293 }
11294 }
11295 else
11296 error (_("Attempt to take contents of a non-pointer value."));
11297 }
11298 arg1 = ada_coerce_ref (arg1); /* FIXME: What is this for?? */
11299 type = ada_check_typedef (value_type (arg1));
11300
11301 if (TYPE_CODE (type) == TYPE_CODE_INT)
11302 /* GDB allows dereferencing an int. If we were given
11303 the expect_type, then use that as the target type.
11304 Otherwise, assume that the target type is an int. */
11305 {
11306 if (expect_type != NULL)
11307 return ada_value_ind (value_cast (lookup_pointer_type (expect_type),
11308 arg1));
11309 else
11310 return value_at_lazy (builtin_type (exp->gdbarch)->builtin_int,
11311 (CORE_ADDR) value_as_address (arg1));
11312 }
11313
11314 if (ada_is_array_descriptor_type (type))
11315 /* GDB allows dereferencing GNAT array descriptors. */
11316 return ada_coerce_to_simple_array (arg1);
11317 else
11318 return ada_value_ind (arg1);
11319
11320 case STRUCTOP_STRUCT:
11321 tem = longest_to_int (exp->elts[pc + 1].longconst);
11322 (*pos) += 3 + BYTES_TO_EXP_ELEM (tem + 1);
11323 preeval_pos = *pos;
11324 arg1 = evaluate_subexp (NULL_TYPE, exp, pos, noside);
11325 if (noside == EVAL_SKIP)
11326 goto nosideret;
11327 if (noside == EVAL_AVOID_SIDE_EFFECTS)
11328 {
11329 struct type *type1 = value_type (arg1);
11330
11331 if (ada_is_tagged_type (type1, 1))
11332 {
11333 type = ada_lookup_struct_elt_type (type1,
11334 &exp->elts[pc + 2].string,
11335 1, 1, NULL);
11336
11337 /* If the field is not found, check if it exists in the
11338 extension of this object's type. This means that we
11339 need to evaluate completely the expression. */
11340
11341 if (type == NULL)
11342 {
11343 arg1 = evaluate_subexp (NULL_TYPE, exp, &preeval_pos,
11344 EVAL_NORMAL);
11345 arg1 = ada_value_struct_elt (arg1,
11346 &exp->elts[pc + 2].string,
11347 0);
11348 arg1 = unwrap_value (arg1);
11349 type = value_type (ada_to_fixed_value (arg1));
11350 }
11351 }
11352 else
11353 type =
11354 ada_lookup_struct_elt_type (type1, &exp->elts[pc + 2].string, 1,
11355 0, NULL);
11356
11357 return value_zero (ada_aligned_type (type), lval_memory);
11358 }
11359 else
11360 arg1 = ada_value_struct_elt (arg1, &exp->elts[pc + 2].string, 0);
11361 arg1 = unwrap_value (arg1);
11362 return ada_to_fixed_value (arg1);
11363
11364 case OP_TYPE:
11365 /* The value is not supposed to be used. This is here to make it
11366 easier to accommodate expressions that contain types. */
11367 (*pos) += 2;
11368 if (noside == EVAL_SKIP)
11369 goto nosideret;
11370 else if (noside == EVAL_AVOID_SIDE_EFFECTS)
11371 return allocate_value (exp->elts[pc + 1].type);
11372 else
11373 error (_("Attempt to use a type name as an expression"));
11374
11375 case OP_AGGREGATE:
11376 case OP_CHOICES:
11377 case OP_OTHERS:
11378 case OP_DISCRETE_RANGE:
11379 case OP_POSITIONAL:
11380 case OP_NAME:
11381 if (noside == EVAL_NORMAL)
11382 switch (op)
11383 {
11384 case OP_NAME:
11385 error (_("Undefined name, ambiguous name, or renaming used in "
11386 "component association: %s."), &exp->elts[pc+2].string);
11387 case OP_AGGREGATE:
11388 error (_("Aggregates only allowed on the right of an assignment"));
11389 default:
11390 internal_error (__FILE__, __LINE__,
11391 _("aggregate apparently mangled"));
11392 }
11393
11394 ada_forward_operator_length (exp, pc, &oplen, &nargs);
11395 *pos += oplen - 1;
11396 for (tem = 0; tem < nargs; tem += 1)
11397 ada_evaluate_subexp (NULL, exp, pos, noside);
11398 goto nosideret;
11399 }
11400
11401 nosideret:
11402 return value_from_longest (builtin_type (exp->gdbarch)->builtin_int, 1);
11403 }
11404 \f
11405
11406 /* Fixed point */
11407
11408 /* If TYPE encodes an Ada fixed-point type, return the suffix of the
11409 type name that encodes the 'small and 'delta information.
11410 Otherwise, return NULL. */
11411
11412 static const char *
11413 fixed_type_info (struct type *type)
11414 {
11415 const char *name = ada_type_name (type);
11416 enum type_code code = (type == NULL) ? TYPE_CODE_UNDEF : TYPE_CODE (type);
11417
11418 if ((code == TYPE_CODE_INT || code == TYPE_CODE_RANGE) && name != NULL)
11419 {
11420 const char *tail = strstr (name, "___XF_");
11421
11422 if (tail == NULL)
11423 return NULL;
11424 else
11425 return tail + 5;
11426 }
11427 else if (code == TYPE_CODE_RANGE && TYPE_TARGET_TYPE (type) != type)
11428 return fixed_type_info (TYPE_TARGET_TYPE (type));
11429 else
11430 return NULL;
11431 }
11432
11433 /* Returns non-zero iff TYPE represents an Ada fixed-point type. */
11434
11435 int
11436 ada_is_fixed_point_type (struct type *type)
11437 {
11438 return fixed_type_info (type) != NULL;
11439 }
11440
11441 /* Return non-zero iff TYPE represents a System.Address type. */
11442
11443 int
11444 ada_is_system_address_type (struct type *type)
11445 {
11446 return (TYPE_NAME (type)
11447 && strcmp (TYPE_NAME (type), "system__address") == 0);
11448 }
11449
11450 /* Assuming that TYPE is the representation of an Ada fixed-point
11451 type, return its delta, or -1 if the type is malformed and the
11452 delta cannot be determined. */
11453
11454 DOUBLEST
11455 ada_delta (struct type *type)
11456 {
11457 const char *encoding = fixed_type_info (type);
11458 DOUBLEST num, den;
11459
11460 /* Strictly speaking, num and den are encoded as integer. However,
11461 they may not fit into a long, and they will have to be converted
11462 to DOUBLEST anyway. So scan them as DOUBLEST. */
11463 if (sscanf (encoding, "_%" DOUBLEST_SCAN_FORMAT "_%" DOUBLEST_SCAN_FORMAT,
11464 &num, &den) < 2)
11465 return -1.0;
11466 else
11467 return num / den;
11468 }
11469
11470 /* Assuming that ada_is_fixed_point_type (TYPE), return the scaling
11471 factor ('SMALL value) associated with the type. */
11472
11473 static DOUBLEST
11474 scaling_factor (struct type *type)
11475 {
11476 const char *encoding = fixed_type_info (type);
11477 DOUBLEST num0, den0, num1, den1;
11478 int n;
11479
11480 /* Strictly speaking, num's and den's are encoded as integer. However,
11481 they may not fit into a long, and they will have to be converted
11482 to DOUBLEST anyway. So scan them as DOUBLEST. */
11483 n = sscanf (encoding,
11484 "_%" DOUBLEST_SCAN_FORMAT "_%" DOUBLEST_SCAN_FORMAT
11485 "_%" DOUBLEST_SCAN_FORMAT "_%" DOUBLEST_SCAN_FORMAT,
11486 &num0, &den0, &num1, &den1);
11487
11488 if (n < 2)
11489 return 1.0;
11490 else if (n == 4)
11491 return num1 / den1;
11492 else
11493 return num0 / den0;
11494 }
11495
11496
11497 /* Assuming that X is the representation of a value of fixed-point
11498 type TYPE, return its floating-point equivalent. */
11499
11500 DOUBLEST
11501 ada_fixed_to_float (struct type *type, LONGEST x)
11502 {
11503 return (DOUBLEST) x *scaling_factor (type);
11504 }
11505
11506 /* The representation of a fixed-point value of type TYPE
11507 corresponding to the value X. */
11508
11509 LONGEST
11510 ada_float_to_fixed (struct type *type, DOUBLEST x)
11511 {
11512 return (LONGEST) (x / scaling_factor (type) + 0.5);
11513 }
11514
11515 \f
11516
11517 /* Range types */
11518
11519 /* Scan STR beginning at position K for a discriminant name, and
11520 return the value of that discriminant field of DVAL in *PX. If
11521 PNEW_K is not null, put the position of the character beyond the
11522 name scanned in *PNEW_K. Return 1 if successful; return 0 and do
11523 not alter *PX and *PNEW_K if unsuccessful. */
11524
11525 static int
11526 scan_discrim_bound (const char *str, int k, struct value *dval, LONGEST * px,
11527 int *pnew_k)
11528 {
11529 static char *bound_buffer = NULL;
11530 static size_t bound_buffer_len = 0;
11531 const char *pstart, *pend, *bound;
11532 struct value *bound_val;
11533
11534 if (dval == NULL || str == NULL || str[k] == '\0')
11535 return 0;
11536
11537 pstart = str + k;
11538 pend = strstr (pstart, "__");
11539 if (pend == NULL)
11540 {
11541 bound = pstart;
11542 k += strlen (bound);
11543 }
11544 else
11545 {
11546 int len = pend - pstart;
11547
11548 /* Strip __ and beyond. */
11549 GROW_VECT (bound_buffer, bound_buffer_len, len + 1);
11550 strncpy (bound_buffer, pstart, len);
11551 bound_buffer[len] = '\0';
11552
11553 bound = bound_buffer;
11554 k = pend - str;
11555 }
11556
11557 bound_val = ada_search_struct_field (bound, dval, 0, value_type (dval));
11558 if (bound_val == NULL)
11559 return 0;
11560
11561 *px = value_as_long (bound_val);
11562 if (pnew_k != NULL)
11563 *pnew_k = k;
11564 return 1;
11565 }
11566
11567 /* Value of variable named NAME in the current environment. If
11568 no such variable found, then if ERR_MSG is null, returns 0, and
11569 otherwise causes an error with message ERR_MSG. */
11570
11571 static struct value *
11572 get_var_value (char *name, char *err_msg)
11573 {
11574 struct block_symbol *syms;
11575 int nsyms;
11576
11577 nsyms = ada_lookup_symbol_list (name, get_selected_block (0), VAR_DOMAIN,
11578 &syms);
11579
11580 if (nsyms != 1)
11581 {
11582 if (err_msg == NULL)
11583 return 0;
11584 else
11585 error (("%s"), err_msg);
11586 }
11587
11588 return value_of_variable (syms[0].symbol, syms[0].block);
11589 }
11590
11591 /* Value of integer variable named NAME in the current environment. If
11592 no such variable found, returns 0, and sets *FLAG to 0. If
11593 successful, sets *FLAG to 1. */
11594
11595 LONGEST
11596 get_int_var_value (char *name, int *flag)
11597 {
11598 struct value *var_val = get_var_value (name, 0);
11599
11600 if (var_val == 0)
11601 {
11602 if (flag != NULL)
11603 *flag = 0;
11604 return 0;
11605 }
11606 else
11607 {
11608 if (flag != NULL)
11609 *flag = 1;
11610 return value_as_long (var_val);
11611 }
11612 }
11613
11614
11615 /* Return a range type whose base type is that of the range type named
11616 NAME in the current environment, and whose bounds are calculated
11617 from NAME according to the GNAT range encoding conventions.
11618 Extract discriminant values, if needed, from DVAL. ORIG_TYPE is the
11619 corresponding range type from debug information; fall back to using it
11620 if symbol lookup fails. If a new type must be created, allocate it
11621 like ORIG_TYPE was. The bounds information, in general, is encoded
11622 in NAME, the base type given in the named range type. */
11623
11624 static struct type *
11625 to_fixed_range_type (struct type *raw_type, struct value *dval)
11626 {
11627 const char *name;
11628 struct type *base_type;
11629 const char *subtype_info;
11630
11631 gdb_assert (raw_type != NULL);
11632 gdb_assert (TYPE_NAME (raw_type) != NULL);
11633
11634 if (TYPE_CODE (raw_type) == TYPE_CODE_RANGE)
11635 base_type = TYPE_TARGET_TYPE (raw_type);
11636 else
11637 base_type = raw_type;
11638
11639 name = TYPE_NAME (raw_type);
11640 subtype_info = strstr (name, "___XD");
11641 if (subtype_info == NULL)
11642 {
11643 LONGEST L = ada_discrete_type_low_bound (raw_type);
11644 LONGEST U = ada_discrete_type_high_bound (raw_type);
11645
11646 if (L < INT_MIN || U > INT_MAX)
11647 return raw_type;
11648 else
11649 return create_static_range_type (alloc_type_copy (raw_type), raw_type,
11650 L, U);
11651 }
11652 else
11653 {
11654 static char *name_buf = NULL;
11655 static size_t name_len = 0;
11656 int prefix_len = subtype_info - name;
11657 LONGEST L, U;
11658 struct type *type;
11659 const char *bounds_str;
11660 int n;
11661
11662 GROW_VECT (name_buf, name_len, prefix_len + 5);
11663 strncpy (name_buf, name, prefix_len);
11664 name_buf[prefix_len] = '\0';
11665
11666 subtype_info += 5;
11667 bounds_str = strchr (subtype_info, '_');
11668 n = 1;
11669
11670 if (*subtype_info == 'L')
11671 {
11672 if (!ada_scan_number (bounds_str, n, &L, &n)
11673 && !scan_discrim_bound (bounds_str, n, dval, &L, &n))
11674 return raw_type;
11675 if (bounds_str[n] == '_')
11676 n += 2;
11677 else if (bounds_str[n] == '.') /* FIXME? SGI Workshop kludge. */
11678 n += 1;
11679 subtype_info += 1;
11680 }
11681 else
11682 {
11683 int ok;
11684
11685 strcpy (name_buf + prefix_len, "___L");
11686 L = get_int_var_value (name_buf, &ok);
11687 if (!ok)
11688 {
11689 lim_warning (_("Unknown lower bound, using 1."));
11690 L = 1;
11691 }
11692 }
11693
11694 if (*subtype_info == 'U')
11695 {
11696 if (!ada_scan_number (bounds_str, n, &U, &n)
11697 && !scan_discrim_bound (bounds_str, n, dval, &U, &n))
11698 return raw_type;
11699 }
11700 else
11701 {
11702 int ok;
11703
11704 strcpy (name_buf + prefix_len, "___U");
11705 U = get_int_var_value (name_buf, &ok);
11706 if (!ok)
11707 {
11708 lim_warning (_("Unknown upper bound, using %ld."), (long) L);
11709 U = L;
11710 }
11711 }
11712
11713 type = create_static_range_type (alloc_type_copy (raw_type),
11714 base_type, L, U);
11715 TYPE_NAME (type) = name;
11716 return type;
11717 }
11718 }
11719
11720 /* True iff NAME is the name of a range type. */
11721
11722 int
11723 ada_is_range_type_name (const char *name)
11724 {
11725 return (name != NULL && strstr (name, "___XD"));
11726 }
11727 \f
11728
11729 /* Modular types */
11730
11731 /* True iff TYPE is an Ada modular type. */
11732
11733 int
11734 ada_is_modular_type (struct type *type)
11735 {
11736 struct type *subranged_type = get_base_type (type);
11737
11738 return (subranged_type != NULL && TYPE_CODE (type) == TYPE_CODE_RANGE
11739 && TYPE_CODE (subranged_type) == TYPE_CODE_INT
11740 && TYPE_UNSIGNED (subranged_type));
11741 }
11742
11743 /* Assuming ada_is_modular_type (TYPE), the modulus of TYPE. */
11744
11745 ULONGEST
11746 ada_modulus (struct type *type)
11747 {
11748 return (ULONGEST) TYPE_HIGH_BOUND (type) + 1;
11749 }
11750 \f
11751
11752 /* Ada exception catchpoint support:
11753 ---------------------------------
11754
11755 We support 3 kinds of exception catchpoints:
11756 . catchpoints on Ada exceptions
11757 . catchpoints on unhandled Ada exceptions
11758 . catchpoints on failed assertions
11759
11760 Exceptions raised during failed assertions, or unhandled exceptions
11761 could perfectly be caught with the general catchpoint on Ada exceptions.
11762 However, we can easily differentiate these two special cases, and having
11763 the option to distinguish these two cases from the rest can be useful
11764 to zero-in on certain situations.
11765
11766 Exception catchpoints are a specialized form of breakpoint,
11767 since they rely on inserting breakpoints inside known routines
11768 of the GNAT runtime. The implementation therefore uses a standard
11769 breakpoint structure of the BP_BREAKPOINT type, but with its own set
11770 of breakpoint_ops.
11771
11772 Support in the runtime for exception catchpoints have been changed
11773 a few times already, and these changes affect the implementation
11774 of these catchpoints. In order to be able to support several
11775 variants of the runtime, we use a sniffer that will determine
11776 the runtime variant used by the program being debugged. */
11777
11778 /* Ada's standard exceptions.
11779
11780 The Ada 83 standard also defined Numeric_Error. But there so many
11781 situations where it was unclear from the Ada 83 Reference Manual
11782 (RM) whether Constraint_Error or Numeric_Error should be raised,
11783 that the ARG (Ada Rapporteur Group) eventually issued a Binding
11784 Interpretation saying that anytime the RM says that Numeric_Error
11785 should be raised, the implementation may raise Constraint_Error.
11786 Ada 95 went one step further and pretty much removed Numeric_Error
11787 from the list of standard exceptions (it made it a renaming of
11788 Constraint_Error, to help preserve compatibility when compiling
11789 an Ada83 compiler). As such, we do not include Numeric_Error from
11790 this list of standard exceptions. */
11791
11792 static char *standard_exc[] = {
11793 "constraint_error",
11794 "program_error",
11795 "storage_error",
11796 "tasking_error"
11797 };
11798
11799 typedef CORE_ADDR (ada_unhandled_exception_name_addr_ftype) (void);
11800
11801 /* A structure that describes how to support exception catchpoints
11802 for a given executable. */
11803
11804 struct exception_support_info
11805 {
11806 /* The name of the symbol to break on in order to insert
11807 a catchpoint on exceptions. */
11808 const char *catch_exception_sym;
11809
11810 /* The name of the symbol to break on in order to insert
11811 a catchpoint on unhandled exceptions. */
11812 const char *catch_exception_unhandled_sym;
11813
11814 /* The name of the symbol to break on in order to insert
11815 a catchpoint on failed assertions. */
11816 const char *catch_assert_sym;
11817
11818 /* Assuming that the inferior just triggered an unhandled exception
11819 catchpoint, this function is responsible for returning the address
11820 in inferior memory where the name of that exception is stored.
11821 Return zero if the address could not be computed. */
11822 ada_unhandled_exception_name_addr_ftype *unhandled_exception_name_addr;
11823 };
11824
11825 static CORE_ADDR ada_unhandled_exception_name_addr (void);
11826 static CORE_ADDR ada_unhandled_exception_name_addr_from_raise (void);
11827
11828 /* The following exception support info structure describes how to
11829 implement exception catchpoints with the latest version of the
11830 Ada runtime (as of 2007-03-06). */
11831
11832 static const struct exception_support_info default_exception_support_info =
11833 {
11834 "__gnat_debug_raise_exception", /* catch_exception_sym */
11835 "__gnat_unhandled_exception", /* catch_exception_unhandled_sym */
11836 "__gnat_debug_raise_assert_failure", /* catch_assert_sym */
11837 ada_unhandled_exception_name_addr
11838 };
11839
11840 /* The following exception support info structure describes how to
11841 implement exception catchpoints with a slightly older version
11842 of the Ada runtime. */
11843
11844 static const struct exception_support_info exception_support_info_fallback =
11845 {
11846 "__gnat_raise_nodefer_with_msg", /* catch_exception_sym */
11847 "__gnat_unhandled_exception", /* catch_exception_unhandled_sym */
11848 "system__assertions__raise_assert_failure", /* catch_assert_sym */
11849 ada_unhandled_exception_name_addr_from_raise
11850 };
11851
11852 /* Return nonzero if we can detect the exception support routines
11853 described in EINFO.
11854
11855 This function errors out if an abnormal situation is detected
11856 (for instance, if we find the exception support routines, but
11857 that support is found to be incomplete). */
11858
11859 static int
11860 ada_has_this_exception_support (const struct exception_support_info *einfo)
11861 {
11862 struct symbol *sym;
11863
11864 /* The symbol we're looking up is provided by a unit in the GNAT runtime
11865 that should be compiled with debugging information. As a result, we
11866 expect to find that symbol in the symtabs. */
11867
11868 sym = standard_lookup (einfo->catch_exception_sym, NULL, VAR_DOMAIN);
11869 if (sym == NULL)
11870 {
11871 /* Perhaps we did not find our symbol because the Ada runtime was
11872 compiled without debugging info, or simply stripped of it.
11873 It happens on some GNU/Linux distributions for instance, where
11874 users have to install a separate debug package in order to get
11875 the runtime's debugging info. In that situation, let the user
11876 know why we cannot insert an Ada exception catchpoint.
11877
11878 Note: Just for the purpose of inserting our Ada exception
11879 catchpoint, we could rely purely on the associated minimal symbol.
11880 But we would be operating in degraded mode anyway, since we are
11881 still lacking the debugging info needed later on to extract
11882 the name of the exception being raised (this name is printed in
11883 the catchpoint message, and is also used when trying to catch
11884 a specific exception). We do not handle this case for now. */
11885 struct bound_minimal_symbol msym
11886 = lookup_minimal_symbol (einfo->catch_exception_sym, NULL, NULL);
11887
11888 if (msym.minsym && MSYMBOL_TYPE (msym.minsym) != mst_solib_trampoline)
11889 error (_("Your Ada runtime appears to be missing some debugging "
11890 "information.\nCannot insert Ada exception catchpoint "
11891 "in this configuration."));
11892
11893 return 0;
11894 }
11895
11896 /* Make sure that the symbol we found corresponds to a function. */
11897
11898 if (SYMBOL_CLASS (sym) != LOC_BLOCK)
11899 error (_("Symbol \"%s\" is not a function (class = %d)"),
11900 SYMBOL_LINKAGE_NAME (sym), SYMBOL_CLASS (sym));
11901
11902 return 1;
11903 }
11904
11905 /* Inspect the Ada runtime and determine which exception info structure
11906 should be used to provide support for exception catchpoints.
11907
11908 This function will always set the per-inferior exception_info,
11909 or raise an error. */
11910
11911 static void
11912 ada_exception_support_info_sniffer (void)
11913 {
11914 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ());
11915
11916 /* If the exception info is already known, then no need to recompute it. */
11917 if (data->exception_info != NULL)
11918 return;
11919
11920 /* Check the latest (default) exception support info. */
11921 if (ada_has_this_exception_support (&default_exception_support_info))
11922 {
11923 data->exception_info = &default_exception_support_info;
11924 return;
11925 }
11926
11927 /* Try our fallback exception suport info. */
11928 if (ada_has_this_exception_support (&exception_support_info_fallback))
11929 {
11930 data->exception_info = &exception_support_info_fallback;
11931 return;
11932 }
11933
11934 /* Sometimes, it is normal for us to not be able to find the routine
11935 we are looking for. This happens when the program is linked with
11936 the shared version of the GNAT runtime, and the program has not been
11937 started yet. Inform the user of these two possible causes if
11938 applicable. */
11939
11940 if (ada_update_initial_language (language_unknown) != language_ada)
11941 error (_("Unable to insert catchpoint. Is this an Ada main program?"));
11942
11943 /* If the symbol does not exist, then check that the program is
11944 already started, to make sure that shared libraries have been
11945 loaded. If it is not started, this may mean that the symbol is
11946 in a shared library. */
11947
11948 if (ptid_get_pid (inferior_ptid) == 0)
11949 error (_("Unable to insert catchpoint. Try to start the program first."));
11950
11951 /* At this point, we know that we are debugging an Ada program and
11952 that the inferior has been started, but we still are not able to
11953 find the run-time symbols. That can mean that we are in
11954 configurable run time mode, or that a-except as been optimized
11955 out by the linker... In any case, at this point it is not worth
11956 supporting this feature. */
11957
11958 error (_("Cannot insert Ada exception catchpoints in this configuration."));
11959 }
11960
11961 /* True iff FRAME is very likely to be that of a function that is
11962 part of the runtime system. This is all very heuristic, but is
11963 intended to be used as advice as to what frames are uninteresting
11964 to most users. */
11965
11966 static int
11967 is_known_support_routine (struct frame_info *frame)
11968 {
11969 struct symtab_and_line sal;
11970 char *func_name;
11971 enum language func_lang;
11972 int i;
11973 const char *fullname;
11974
11975 /* If this code does not have any debugging information (no symtab),
11976 This cannot be any user code. */
11977
11978 find_frame_sal (frame, &sal);
11979 if (sal.symtab == NULL)
11980 return 1;
11981
11982 /* If there is a symtab, but the associated source file cannot be
11983 located, then assume this is not user code: Selecting a frame
11984 for which we cannot display the code would not be very helpful
11985 for the user. This should also take care of case such as VxWorks
11986 where the kernel has some debugging info provided for a few units. */
11987
11988 fullname = symtab_to_fullname (sal.symtab);
11989 if (access (fullname, R_OK) != 0)
11990 return 1;
11991
11992 /* Check the unit filename againt the Ada runtime file naming.
11993 We also check the name of the objfile against the name of some
11994 known system libraries that sometimes come with debugging info
11995 too. */
11996
11997 for (i = 0; known_runtime_file_name_patterns[i] != NULL; i += 1)
11998 {
11999 re_comp (known_runtime_file_name_patterns[i]);
12000 if (re_exec (lbasename (sal.symtab->filename)))
12001 return 1;
12002 if (SYMTAB_OBJFILE (sal.symtab) != NULL
12003 && re_exec (objfile_name (SYMTAB_OBJFILE (sal.symtab))))
12004 return 1;
12005 }
12006
12007 /* Check whether the function is a GNAT-generated entity. */
12008
12009 find_frame_funname (frame, &func_name, &func_lang, NULL);
12010 if (func_name == NULL)
12011 return 1;
12012
12013 for (i = 0; known_auxiliary_function_name_patterns[i] != NULL; i += 1)
12014 {
12015 re_comp (known_auxiliary_function_name_patterns[i]);
12016 if (re_exec (func_name))
12017 {
12018 xfree (func_name);
12019 return 1;
12020 }
12021 }
12022
12023 xfree (func_name);
12024 return 0;
12025 }
12026
12027 /* Find the first frame that contains debugging information and that is not
12028 part of the Ada run-time, starting from FI and moving upward. */
12029
12030 void
12031 ada_find_printable_frame (struct frame_info *fi)
12032 {
12033 for (; fi != NULL; fi = get_prev_frame (fi))
12034 {
12035 if (!is_known_support_routine (fi))
12036 {
12037 select_frame (fi);
12038 break;
12039 }
12040 }
12041
12042 }
12043
12044 /* Assuming that the inferior just triggered an unhandled exception
12045 catchpoint, return the address in inferior memory where the name
12046 of the exception is stored.
12047
12048 Return zero if the address could not be computed. */
12049
12050 static CORE_ADDR
12051 ada_unhandled_exception_name_addr (void)
12052 {
12053 return parse_and_eval_address ("e.full_name");
12054 }
12055
12056 /* Same as ada_unhandled_exception_name_addr, except that this function
12057 should be used when the inferior uses an older version of the runtime,
12058 where the exception name needs to be extracted from a specific frame
12059 several frames up in the callstack. */
12060
12061 static CORE_ADDR
12062 ada_unhandled_exception_name_addr_from_raise (void)
12063 {
12064 int frame_level;
12065 struct frame_info *fi;
12066 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ());
12067 struct cleanup *old_chain;
12068
12069 /* To determine the name of this exception, we need to select
12070 the frame corresponding to RAISE_SYM_NAME. This frame is
12071 at least 3 levels up, so we simply skip the first 3 frames
12072 without checking the name of their associated function. */
12073 fi = get_current_frame ();
12074 for (frame_level = 0; frame_level < 3; frame_level += 1)
12075 if (fi != NULL)
12076 fi = get_prev_frame (fi);
12077
12078 old_chain = make_cleanup (null_cleanup, NULL);
12079 while (fi != NULL)
12080 {
12081 char *func_name;
12082 enum language func_lang;
12083
12084 find_frame_funname (fi, &func_name, &func_lang, NULL);
12085 if (func_name != NULL)
12086 {
12087 make_cleanup (xfree, func_name);
12088
12089 if (strcmp (func_name,
12090 data->exception_info->catch_exception_sym) == 0)
12091 break; /* We found the frame we were looking for... */
12092 fi = get_prev_frame (fi);
12093 }
12094 }
12095 do_cleanups (old_chain);
12096
12097 if (fi == NULL)
12098 return 0;
12099
12100 select_frame (fi);
12101 return parse_and_eval_address ("id.full_name");
12102 }
12103
12104 /* Assuming the inferior just triggered an Ada exception catchpoint
12105 (of any type), return the address in inferior memory where the name
12106 of the exception is stored, if applicable.
12107
12108 Return zero if the address could not be computed, or if not relevant. */
12109
12110 static CORE_ADDR
12111 ada_exception_name_addr_1 (enum ada_exception_catchpoint_kind ex,
12112 struct breakpoint *b)
12113 {
12114 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ());
12115
12116 switch (ex)
12117 {
12118 case ada_catch_exception:
12119 return (parse_and_eval_address ("e.full_name"));
12120 break;
12121
12122 case ada_catch_exception_unhandled:
12123 return data->exception_info->unhandled_exception_name_addr ();
12124 break;
12125
12126 case ada_catch_assert:
12127 return 0; /* Exception name is not relevant in this case. */
12128 break;
12129
12130 default:
12131 internal_error (__FILE__, __LINE__, _("unexpected catchpoint type"));
12132 break;
12133 }
12134
12135 return 0; /* Should never be reached. */
12136 }
12137
12138 /* Same as ada_exception_name_addr_1, except that it intercepts and contains
12139 any error that ada_exception_name_addr_1 might cause to be thrown.
12140 When an error is intercepted, a warning with the error message is printed,
12141 and zero is returned. */
12142
12143 static CORE_ADDR
12144 ada_exception_name_addr (enum ada_exception_catchpoint_kind ex,
12145 struct breakpoint *b)
12146 {
12147 CORE_ADDR result = 0;
12148
12149 TRY
12150 {
12151 result = ada_exception_name_addr_1 (ex, b);
12152 }
12153
12154 CATCH (e, RETURN_MASK_ERROR)
12155 {
12156 warning (_("failed to get exception name: %s"), e.message);
12157 return 0;
12158 }
12159 END_CATCH
12160
12161 return result;
12162 }
12163
12164 static char *ada_exception_catchpoint_cond_string (const char *excep_string);
12165
12166 /* Ada catchpoints.
12167
12168 In the case of catchpoints on Ada exceptions, the catchpoint will
12169 stop the target on every exception the program throws. When a user
12170 specifies the name of a specific exception, we translate this
12171 request into a condition expression (in text form), and then parse
12172 it into an expression stored in each of the catchpoint's locations.
12173 We then use this condition to check whether the exception that was
12174 raised is the one the user is interested in. If not, then the
12175 target is resumed again. We store the name of the requested
12176 exception, in order to be able to re-set the condition expression
12177 when symbols change. */
12178
12179 /* An instance of this type is used to represent an Ada catchpoint
12180 breakpoint location. It includes a "struct bp_location" as a kind
12181 of base class; users downcast to "struct bp_location *" when
12182 needed. */
12183
12184 struct ada_catchpoint_location
12185 {
12186 /* The base class. */
12187 struct bp_location base;
12188
12189 /* The condition that checks whether the exception that was raised
12190 is the specific exception the user specified on catchpoint
12191 creation. */
12192 struct expression *excep_cond_expr;
12193 };
12194
12195 /* Implement the DTOR method in the bp_location_ops structure for all
12196 Ada exception catchpoint kinds. */
12197
12198 static void
12199 ada_catchpoint_location_dtor (struct bp_location *bl)
12200 {
12201 struct ada_catchpoint_location *al = (struct ada_catchpoint_location *) bl;
12202
12203 xfree (al->excep_cond_expr);
12204 }
12205
12206 /* The vtable to be used in Ada catchpoint locations. */
12207
12208 static const struct bp_location_ops ada_catchpoint_location_ops =
12209 {
12210 ada_catchpoint_location_dtor
12211 };
12212
12213 /* An instance of this type is used to represent an Ada catchpoint.
12214 It includes a "struct breakpoint" as a kind of base class; users
12215 downcast to "struct breakpoint *" when needed. */
12216
12217 struct ada_catchpoint
12218 {
12219 /* The base class. */
12220 struct breakpoint base;
12221
12222 /* The name of the specific exception the user specified. */
12223 char *excep_string;
12224 };
12225
12226 /* Parse the exception condition string in the context of each of the
12227 catchpoint's locations, and store them for later evaluation. */
12228
12229 static void
12230 create_excep_cond_exprs (struct ada_catchpoint *c)
12231 {
12232 struct cleanup *old_chain;
12233 struct bp_location *bl;
12234 char *cond_string;
12235
12236 /* Nothing to do if there's no specific exception to catch. */
12237 if (c->excep_string == NULL)
12238 return;
12239
12240 /* Same if there are no locations... */
12241 if (c->base.loc == NULL)
12242 return;
12243
12244 /* Compute the condition expression in text form, from the specific
12245 expection we want to catch. */
12246 cond_string = ada_exception_catchpoint_cond_string (c->excep_string);
12247 old_chain = make_cleanup (xfree, cond_string);
12248
12249 /* Iterate over all the catchpoint's locations, and parse an
12250 expression for each. */
12251 for (bl = c->base.loc; bl != NULL; bl = bl->next)
12252 {
12253 struct ada_catchpoint_location *ada_loc
12254 = (struct ada_catchpoint_location *) bl;
12255 struct expression *exp = NULL;
12256
12257 if (!bl->shlib_disabled)
12258 {
12259 const char *s;
12260
12261 s = cond_string;
12262 TRY
12263 {
12264 exp = parse_exp_1 (&s, bl->address,
12265 block_for_pc (bl->address), 0);
12266 }
12267 CATCH (e, RETURN_MASK_ERROR)
12268 {
12269 warning (_("failed to reevaluate internal exception condition "
12270 "for catchpoint %d: %s"),
12271 c->base.number, e.message);
12272 /* There is a bug in GCC on sparc-solaris when building with
12273 optimization which causes EXP to change unexpectedly
12274 (http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56982).
12275 The problem should be fixed starting with GCC 4.9.
12276 In the meantime, work around it by forcing EXP back
12277 to NULL. */
12278 exp = NULL;
12279 }
12280 END_CATCH
12281 }
12282
12283 ada_loc->excep_cond_expr = exp;
12284 }
12285
12286 do_cleanups (old_chain);
12287 }
12288
12289 /* Implement the DTOR method in the breakpoint_ops structure for all
12290 exception catchpoint kinds. */
12291
12292 static void
12293 dtor_exception (enum ada_exception_catchpoint_kind ex, struct breakpoint *b)
12294 {
12295 struct ada_catchpoint *c = (struct ada_catchpoint *) b;
12296
12297 xfree (c->excep_string);
12298
12299 bkpt_breakpoint_ops.dtor (b);
12300 }
12301
12302 /* Implement the ALLOCATE_LOCATION method in the breakpoint_ops
12303 structure for all exception catchpoint kinds. */
12304
12305 static struct bp_location *
12306 allocate_location_exception (enum ada_exception_catchpoint_kind ex,
12307 struct breakpoint *self)
12308 {
12309 struct ada_catchpoint_location *loc;
12310
12311 loc = XNEW (struct ada_catchpoint_location);
12312 init_bp_location (&loc->base, &ada_catchpoint_location_ops, self);
12313 loc->excep_cond_expr = NULL;
12314 return &loc->base;
12315 }
12316
12317 /* Implement the RE_SET method in the breakpoint_ops structure for all
12318 exception catchpoint kinds. */
12319
12320 static void
12321 re_set_exception (enum ada_exception_catchpoint_kind ex, struct breakpoint *b)
12322 {
12323 struct ada_catchpoint *c = (struct ada_catchpoint *) b;
12324
12325 /* Call the base class's method. This updates the catchpoint's
12326 locations. */
12327 bkpt_breakpoint_ops.re_set (b);
12328
12329 /* Reparse the exception conditional expressions. One for each
12330 location. */
12331 create_excep_cond_exprs (c);
12332 }
12333
12334 /* Returns true if we should stop for this breakpoint hit. If the
12335 user specified a specific exception, we only want to cause a stop
12336 if the program thrown that exception. */
12337
12338 static int
12339 should_stop_exception (const struct bp_location *bl)
12340 {
12341 struct ada_catchpoint *c = (struct ada_catchpoint *) bl->owner;
12342 const struct ada_catchpoint_location *ada_loc
12343 = (const struct ada_catchpoint_location *) bl;
12344 int stop;
12345
12346 /* With no specific exception, should always stop. */
12347 if (c->excep_string == NULL)
12348 return 1;
12349
12350 if (ada_loc->excep_cond_expr == NULL)
12351 {
12352 /* We will have a NULL expression if back when we were creating
12353 the expressions, this location's had failed to parse. */
12354 return 1;
12355 }
12356
12357 stop = 1;
12358 TRY
12359 {
12360 struct value *mark;
12361
12362 mark = value_mark ();
12363 stop = value_true (evaluate_expression (ada_loc->excep_cond_expr));
12364 value_free_to_mark (mark);
12365 }
12366 CATCH (ex, RETURN_MASK_ALL)
12367 {
12368 exception_fprintf (gdb_stderr, ex,
12369 _("Error in testing exception condition:\n"));
12370 }
12371 END_CATCH
12372
12373 return stop;
12374 }
12375
12376 /* Implement the CHECK_STATUS method in the breakpoint_ops structure
12377 for all exception catchpoint kinds. */
12378
12379 static void
12380 check_status_exception (enum ada_exception_catchpoint_kind ex, bpstat bs)
12381 {
12382 bs->stop = should_stop_exception (bs->bp_location_at);
12383 }
12384
12385 /* Implement the PRINT_IT method in the breakpoint_ops structure
12386 for all exception catchpoint kinds. */
12387
12388 static enum print_stop_action
12389 print_it_exception (enum ada_exception_catchpoint_kind ex, bpstat bs)
12390 {
12391 struct ui_out *uiout = current_uiout;
12392 struct breakpoint *b = bs->breakpoint_at;
12393
12394 annotate_catchpoint (b->number);
12395
12396 if (ui_out_is_mi_like_p (uiout))
12397 {
12398 ui_out_field_string (uiout, "reason",
12399 async_reason_lookup (EXEC_ASYNC_BREAKPOINT_HIT));
12400 ui_out_field_string (uiout, "disp", bpdisp_text (b->disposition));
12401 }
12402
12403 ui_out_text (uiout,
12404 b->disposition == disp_del ? "\nTemporary catchpoint "
12405 : "\nCatchpoint ");
12406 ui_out_field_int (uiout, "bkptno", b->number);
12407 ui_out_text (uiout, ", ");
12408
12409 switch (ex)
12410 {
12411 case ada_catch_exception:
12412 case ada_catch_exception_unhandled:
12413 {
12414 const CORE_ADDR addr = ada_exception_name_addr (ex, b);
12415 char exception_name[256];
12416
12417 if (addr != 0)
12418 {
12419 read_memory (addr, (gdb_byte *) exception_name,
12420 sizeof (exception_name) - 1);
12421 exception_name [sizeof (exception_name) - 1] = '\0';
12422 }
12423 else
12424 {
12425 /* For some reason, we were unable to read the exception
12426 name. This could happen if the Runtime was compiled
12427 without debugging info, for instance. In that case,
12428 just replace the exception name by the generic string
12429 "exception" - it will read as "an exception" in the
12430 notification we are about to print. */
12431 memcpy (exception_name, "exception", sizeof ("exception"));
12432 }
12433 /* In the case of unhandled exception breakpoints, we print
12434 the exception name as "unhandled EXCEPTION_NAME", to make
12435 it clearer to the user which kind of catchpoint just got
12436 hit. We used ui_out_text to make sure that this extra
12437 info does not pollute the exception name in the MI case. */
12438 if (ex == ada_catch_exception_unhandled)
12439 ui_out_text (uiout, "unhandled ");
12440 ui_out_field_string (uiout, "exception-name", exception_name);
12441 }
12442 break;
12443 case ada_catch_assert:
12444 /* In this case, the name of the exception is not really
12445 important. Just print "failed assertion" to make it clearer
12446 that his program just hit an assertion-failure catchpoint.
12447 We used ui_out_text because this info does not belong in
12448 the MI output. */
12449 ui_out_text (uiout, "failed assertion");
12450 break;
12451 }
12452 ui_out_text (uiout, " at ");
12453 ada_find_printable_frame (get_current_frame ());
12454
12455 return PRINT_SRC_AND_LOC;
12456 }
12457
12458 /* Implement the PRINT_ONE method in the breakpoint_ops structure
12459 for all exception catchpoint kinds. */
12460
12461 static void
12462 print_one_exception (enum ada_exception_catchpoint_kind ex,
12463 struct breakpoint *b, struct bp_location **last_loc)
12464 {
12465 struct ui_out *uiout = current_uiout;
12466 struct ada_catchpoint *c = (struct ada_catchpoint *) b;
12467 struct value_print_options opts;
12468
12469 get_user_print_options (&opts);
12470 if (opts.addressprint)
12471 {
12472 annotate_field (4);
12473 ui_out_field_core_addr (uiout, "addr", b->loc->gdbarch, b->loc->address);
12474 }
12475
12476 annotate_field (5);
12477 *last_loc = b->loc;
12478 switch (ex)
12479 {
12480 case ada_catch_exception:
12481 if (c->excep_string != NULL)
12482 {
12483 char *msg = xstrprintf (_("`%s' Ada exception"), c->excep_string);
12484
12485 ui_out_field_string (uiout, "what", msg);
12486 xfree (msg);
12487 }
12488 else
12489 ui_out_field_string (uiout, "what", "all Ada exceptions");
12490
12491 break;
12492
12493 case ada_catch_exception_unhandled:
12494 ui_out_field_string (uiout, "what", "unhandled Ada exceptions");
12495 break;
12496
12497 case ada_catch_assert:
12498 ui_out_field_string (uiout, "what", "failed Ada assertions");
12499 break;
12500
12501 default:
12502 internal_error (__FILE__, __LINE__, _("unexpected catchpoint type"));
12503 break;
12504 }
12505 }
12506
12507 /* Implement the PRINT_MENTION method in the breakpoint_ops structure
12508 for all exception catchpoint kinds. */
12509
12510 static void
12511 print_mention_exception (enum ada_exception_catchpoint_kind ex,
12512 struct breakpoint *b)
12513 {
12514 struct ada_catchpoint *c = (struct ada_catchpoint *) b;
12515 struct ui_out *uiout = current_uiout;
12516
12517 ui_out_text (uiout, b->disposition == disp_del ? _("Temporary catchpoint ")
12518 : _("Catchpoint "));
12519 ui_out_field_int (uiout, "bkptno", b->number);
12520 ui_out_text (uiout, ": ");
12521
12522 switch (ex)
12523 {
12524 case ada_catch_exception:
12525 if (c->excep_string != NULL)
12526 {
12527 char *info = xstrprintf (_("`%s' Ada exception"), c->excep_string);
12528 struct cleanup *old_chain = make_cleanup (xfree, info);
12529
12530 ui_out_text (uiout, info);
12531 do_cleanups (old_chain);
12532 }
12533 else
12534 ui_out_text (uiout, _("all Ada exceptions"));
12535 break;
12536
12537 case ada_catch_exception_unhandled:
12538 ui_out_text (uiout, _("unhandled Ada exceptions"));
12539 break;
12540
12541 case ada_catch_assert:
12542 ui_out_text (uiout, _("failed Ada assertions"));
12543 break;
12544
12545 default:
12546 internal_error (__FILE__, __LINE__, _("unexpected catchpoint type"));
12547 break;
12548 }
12549 }
12550
12551 /* Implement the PRINT_RECREATE method in the breakpoint_ops structure
12552 for all exception catchpoint kinds. */
12553
12554 static void
12555 print_recreate_exception (enum ada_exception_catchpoint_kind ex,
12556 struct breakpoint *b, struct ui_file *fp)
12557 {
12558 struct ada_catchpoint *c = (struct ada_catchpoint *) b;
12559
12560 switch (ex)
12561 {
12562 case ada_catch_exception:
12563 fprintf_filtered (fp, "catch exception");
12564 if (c->excep_string != NULL)
12565 fprintf_filtered (fp, " %s", c->excep_string);
12566 break;
12567
12568 case ada_catch_exception_unhandled:
12569 fprintf_filtered (fp, "catch exception unhandled");
12570 break;
12571
12572 case ada_catch_assert:
12573 fprintf_filtered (fp, "catch assert");
12574 break;
12575
12576 default:
12577 internal_error (__FILE__, __LINE__, _("unexpected catchpoint type"));
12578 }
12579 print_recreate_thread (b, fp);
12580 }
12581
12582 /* Virtual table for "catch exception" breakpoints. */
12583
12584 static void
12585 dtor_catch_exception (struct breakpoint *b)
12586 {
12587 dtor_exception (ada_catch_exception, b);
12588 }
12589
12590 static struct bp_location *
12591 allocate_location_catch_exception (struct breakpoint *self)
12592 {
12593 return allocate_location_exception (ada_catch_exception, self);
12594 }
12595
12596 static void
12597 re_set_catch_exception (struct breakpoint *b)
12598 {
12599 re_set_exception (ada_catch_exception, b);
12600 }
12601
12602 static void
12603 check_status_catch_exception (bpstat bs)
12604 {
12605 check_status_exception (ada_catch_exception, bs);
12606 }
12607
12608 static enum print_stop_action
12609 print_it_catch_exception (bpstat bs)
12610 {
12611 return print_it_exception (ada_catch_exception, bs);
12612 }
12613
12614 static void
12615 print_one_catch_exception (struct breakpoint *b, struct bp_location **last_loc)
12616 {
12617 print_one_exception (ada_catch_exception, b, last_loc);
12618 }
12619
12620 static void
12621 print_mention_catch_exception (struct breakpoint *b)
12622 {
12623 print_mention_exception (ada_catch_exception, b);
12624 }
12625
12626 static void
12627 print_recreate_catch_exception (struct breakpoint *b, struct ui_file *fp)
12628 {
12629 print_recreate_exception (ada_catch_exception, b, fp);
12630 }
12631
12632 static struct breakpoint_ops catch_exception_breakpoint_ops;
12633
12634 /* Virtual table for "catch exception unhandled" breakpoints. */
12635
12636 static void
12637 dtor_catch_exception_unhandled (struct breakpoint *b)
12638 {
12639 dtor_exception (ada_catch_exception_unhandled, b);
12640 }
12641
12642 static struct bp_location *
12643 allocate_location_catch_exception_unhandled (struct breakpoint *self)
12644 {
12645 return allocate_location_exception (ada_catch_exception_unhandled, self);
12646 }
12647
12648 static void
12649 re_set_catch_exception_unhandled (struct breakpoint *b)
12650 {
12651 re_set_exception (ada_catch_exception_unhandled, b);
12652 }
12653
12654 static void
12655 check_status_catch_exception_unhandled (bpstat bs)
12656 {
12657 check_status_exception (ada_catch_exception_unhandled, bs);
12658 }
12659
12660 static enum print_stop_action
12661 print_it_catch_exception_unhandled (bpstat bs)
12662 {
12663 return print_it_exception (ada_catch_exception_unhandled, bs);
12664 }
12665
12666 static void
12667 print_one_catch_exception_unhandled (struct breakpoint *b,
12668 struct bp_location **last_loc)
12669 {
12670 print_one_exception (ada_catch_exception_unhandled, b, last_loc);
12671 }
12672
12673 static void
12674 print_mention_catch_exception_unhandled (struct breakpoint *b)
12675 {
12676 print_mention_exception (ada_catch_exception_unhandled, b);
12677 }
12678
12679 static void
12680 print_recreate_catch_exception_unhandled (struct breakpoint *b,
12681 struct ui_file *fp)
12682 {
12683 print_recreate_exception (ada_catch_exception_unhandled, b, fp);
12684 }
12685
12686 static struct breakpoint_ops catch_exception_unhandled_breakpoint_ops;
12687
12688 /* Virtual table for "catch assert" breakpoints. */
12689
12690 static void
12691 dtor_catch_assert (struct breakpoint *b)
12692 {
12693 dtor_exception (ada_catch_assert, b);
12694 }
12695
12696 static struct bp_location *
12697 allocate_location_catch_assert (struct breakpoint *self)
12698 {
12699 return allocate_location_exception (ada_catch_assert, self);
12700 }
12701
12702 static void
12703 re_set_catch_assert (struct breakpoint *b)
12704 {
12705 re_set_exception (ada_catch_assert, b);
12706 }
12707
12708 static void
12709 check_status_catch_assert (bpstat bs)
12710 {
12711 check_status_exception (ada_catch_assert, bs);
12712 }
12713
12714 static enum print_stop_action
12715 print_it_catch_assert (bpstat bs)
12716 {
12717 return print_it_exception (ada_catch_assert, bs);
12718 }
12719
12720 static void
12721 print_one_catch_assert (struct breakpoint *b, struct bp_location **last_loc)
12722 {
12723 print_one_exception (ada_catch_assert, b, last_loc);
12724 }
12725
12726 static void
12727 print_mention_catch_assert (struct breakpoint *b)
12728 {
12729 print_mention_exception (ada_catch_assert, b);
12730 }
12731
12732 static void
12733 print_recreate_catch_assert (struct breakpoint *b, struct ui_file *fp)
12734 {
12735 print_recreate_exception (ada_catch_assert, b, fp);
12736 }
12737
12738 static struct breakpoint_ops catch_assert_breakpoint_ops;
12739
12740 /* Return a newly allocated copy of the first space-separated token
12741 in ARGSP, and then adjust ARGSP to point immediately after that
12742 token.
12743
12744 Return NULL if ARGPS does not contain any more tokens. */
12745
12746 static char *
12747 ada_get_next_arg (char **argsp)
12748 {
12749 char *args = *argsp;
12750 char *end;
12751 char *result;
12752
12753 args = skip_spaces (args);
12754 if (args[0] == '\0')
12755 return NULL; /* No more arguments. */
12756
12757 /* Find the end of the current argument. */
12758
12759 end = skip_to_space (args);
12760
12761 /* Adjust ARGSP to point to the start of the next argument. */
12762
12763 *argsp = end;
12764
12765 /* Make a copy of the current argument and return it. */
12766
12767 result = (char *) xmalloc (end - args + 1);
12768 strncpy (result, args, end - args);
12769 result[end - args] = '\0';
12770
12771 return result;
12772 }
12773
12774 /* Split the arguments specified in a "catch exception" command.
12775 Set EX to the appropriate catchpoint type.
12776 Set EXCEP_STRING to the name of the specific exception if
12777 specified by the user.
12778 If a condition is found at the end of the arguments, the condition
12779 expression is stored in COND_STRING (memory must be deallocated
12780 after use). Otherwise COND_STRING is set to NULL. */
12781
12782 static void
12783 catch_ada_exception_command_split (char *args,
12784 enum ada_exception_catchpoint_kind *ex,
12785 char **excep_string,
12786 char **cond_string)
12787 {
12788 struct cleanup *old_chain = make_cleanup (null_cleanup, NULL);
12789 char *exception_name;
12790 char *cond = NULL;
12791
12792 exception_name = ada_get_next_arg (&args);
12793 if (exception_name != NULL && strcmp (exception_name, "if") == 0)
12794 {
12795 /* This is not an exception name; this is the start of a condition
12796 expression for a catchpoint on all exceptions. So, "un-get"
12797 this token, and set exception_name to NULL. */
12798 xfree (exception_name);
12799 exception_name = NULL;
12800 args -= 2;
12801 }
12802 make_cleanup (xfree, exception_name);
12803
12804 /* Check to see if we have a condition. */
12805
12806 args = skip_spaces (args);
12807 if (startswith (args, "if")
12808 && (isspace (args[2]) || args[2] == '\0'))
12809 {
12810 args += 2;
12811 args = skip_spaces (args);
12812
12813 if (args[0] == '\0')
12814 error (_("Condition missing after `if' keyword"));
12815 cond = xstrdup (args);
12816 make_cleanup (xfree, cond);
12817
12818 args += strlen (args);
12819 }
12820
12821 /* Check that we do not have any more arguments. Anything else
12822 is unexpected. */
12823
12824 if (args[0] != '\0')
12825 error (_("Junk at end of expression"));
12826
12827 discard_cleanups (old_chain);
12828
12829 if (exception_name == NULL)
12830 {
12831 /* Catch all exceptions. */
12832 *ex = ada_catch_exception;
12833 *excep_string = NULL;
12834 }
12835 else if (strcmp (exception_name, "unhandled") == 0)
12836 {
12837 /* Catch unhandled exceptions. */
12838 *ex = ada_catch_exception_unhandled;
12839 *excep_string = NULL;
12840 }
12841 else
12842 {
12843 /* Catch a specific exception. */
12844 *ex = ada_catch_exception;
12845 *excep_string = exception_name;
12846 }
12847 *cond_string = cond;
12848 }
12849
12850 /* Return the name of the symbol on which we should break in order to
12851 implement a catchpoint of the EX kind. */
12852
12853 static const char *
12854 ada_exception_sym_name (enum ada_exception_catchpoint_kind ex)
12855 {
12856 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ());
12857
12858 gdb_assert (data->exception_info != NULL);
12859
12860 switch (ex)
12861 {
12862 case ada_catch_exception:
12863 return (data->exception_info->catch_exception_sym);
12864 break;
12865 case ada_catch_exception_unhandled:
12866 return (data->exception_info->catch_exception_unhandled_sym);
12867 break;
12868 case ada_catch_assert:
12869 return (data->exception_info->catch_assert_sym);
12870 break;
12871 default:
12872 internal_error (__FILE__, __LINE__,
12873 _("unexpected catchpoint kind (%d)"), ex);
12874 }
12875 }
12876
12877 /* Return the breakpoint ops "virtual table" used for catchpoints
12878 of the EX kind. */
12879
12880 static const struct breakpoint_ops *
12881 ada_exception_breakpoint_ops (enum ada_exception_catchpoint_kind ex)
12882 {
12883 switch (ex)
12884 {
12885 case ada_catch_exception:
12886 return (&catch_exception_breakpoint_ops);
12887 break;
12888 case ada_catch_exception_unhandled:
12889 return (&catch_exception_unhandled_breakpoint_ops);
12890 break;
12891 case ada_catch_assert:
12892 return (&catch_assert_breakpoint_ops);
12893 break;
12894 default:
12895 internal_error (__FILE__, __LINE__,
12896 _("unexpected catchpoint kind (%d)"), ex);
12897 }
12898 }
12899
12900 /* Return the condition that will be used to match the current exception
12901 being raised with the exception that the user wants to catch. This
12902 assumes that this condition is used when the inferior just triggered
12903 an exception catchpoint.
12904
12905 The string returned is a newly allocated string that needs to be
12906 deallocated later. */
12907
12908 static char *
12909 ada_exception_catchpoint_cond_string (const char *excep_string)
12910 {
12911 int i;
12912
12913 /* The standard exceptions are a special case. They are defined in
12914 runtime units that have been compiled without debugging info; if
12915 EXCEP_STRING is the not-fully-qualified name of a standard
12916 exception (e.g. "constraint_error") then, during the evaluation
12917 of the condition expression, the symbol lookup on this name would
12918 *not* return this standard exception. The catchpoint condition
12919 may then be set only on user-defined exceptions which have the
12920 same not-fully-qualified name (e.g. my_package.constraint_error).
12921
12922 To avoid this unexcepted behavior, these standard exceptions are
12923 systematically prefixed by "standard". This means that "catch
12924 exception constraint_error" is rewritten into "catch exception
12925 standard.constraint_error".
12926
12927 If an exception named contraint_error is defined in another package of
12928 the inferior program, then the only way to specify this exception as a
12929 breakpoint condition is to use its fully-qualified named:
12930 e.g. my_package.constraint_error. */
12931
12932 for (i = 0; i < sizeof (standard_exc) / sizeof (char *); i++)
12933 {
12934 if (strcmp (standard_exc [i], excep_string) == 0)
12935 {
12936 return xstrprintf ("long_integer (e) = long_integer (&standard.%s)",
12937 excep_string);
12938 }
12939 }
12940 return xstrprintf ("long_integer (e) = long_integer (&%s)", excep_string);
12941 }
12942
12943 /* Return the symtab_and_line that should be used to insert an exception
12944 catchpoint of the TYPE kind.
12945
12946 EXCEP_STRING should contain the name of a specific exception that
12947 the catchpoint should catch, or NULL otherwise.
12948
12949 ADDR_STRING returns the name of the function where the real
12950 breakpoint that implements the catchpoints is set, depending on the
12951 type of catchpoint we need to create. */
12952
12953 static struct symtab_and_line
12954 ada_exception_sal (enum ada_exception_catchpoint_kind ex, char *excep_string,
12955 char **addr_string, const struct breakpoint_ops **ops)
12956 {
12957 const char *sym_name;
12958 struct symbol *sym;
12959
12960 /* First, find out which exception support info to use. */
12961 ada_exception_support_info_sniffer ();
12962
12963 /* Then lookup the function on which we will break in order to catch
12964 the Ada exceptions requested by the user. */
12965 sym_name = ada_exception_sym_name (ex);
12966 sym = standard_lookup (sym_name, NULL, VAR_DOMAIN);
12967
12968 /* We can assume that SYM is not NULL at this stage. If the symbol
12969 did not exist, ada_exception_support_info_sniffer would have
12970 raised an exception.
12971
12972 Also, ada_exception_support_info_sniffer should have already
12973 verified that SYM is a function symbol. */
12974 gdb_assert (sym != NULL);
12975 gdb_assert (SYMBOL_CLASS (sym) == LOC_BLOCK);
12976
12977 /* Set ADDR_STRING. */
12978 *addr_string = xstrdup (sym_name);
12979
12980 /* Set OPS. */
12981 *ops = ada_exception_breakpoint_ops (ex);
12982
12983 return find_function_start_sal (sym, 1);
12984 }
12985
12986 /* Create an Ada exception catchpoint.
12987
12988 EX_KIND is the kind of exception catchpoint to be created.
12989
12990 If EXCEPT_STRING is NULL, this catchpoint is expected to trigger
12991 for all exceptions. Otherwise, EXCEPT_STRING indicates the name
12992 of the exception to which this catchpoint applies. When not NULL,
12993 the string must be allocated on the heap, and its deallocation
12994 is no longer the responsibility of the caller.
12995
12996 COND_STRING, if not NULL, is the catchpoint condition. This string
12997 must be allocated on the heap, and its deallocation is no longer
12998 the responsibility of the caller.
12999
13000 TEMPFLAG, if nonzero, means that the underlying breakpoint
13001 should be temporary.
13002
13003 FROM_TTY is the usual argument passed to all commands implementations. */
13004
13005 void
13006 create_ada_exception_catchpoint (struct gdbarch *gdbarch,
13007 enum ada_exception_catchpoint_kind ex_kind,
13008 char *excep_string,
13009 char *cond_string,
13010 int tempflag,
13011 int disabled,
13012 int from_tty)
13013 {
13014 struct ada_catchpoint *c;
13015 char *addr_string = NULL;
13016 const struct breakpoint_ops *ops = NULL;
13017 struct symtab_and_line sal
13018 = ada_exception_sal (ex_kind, excep_string, &addr_string, &ops);
13019
13020 c = XNEW (struct ada_catchpoint);
13021 init_ada_exception_breakpoint (&c->base, gdbarch, sal, addr_string,
13022 ops, tempflag, disabled, from_tty);
13023 c->excep_string = excep_string;
13024 create_excep_cond_exprs (c);
13025 if (cond_string != NULL)
13026 set_breakpoint_condition (&c->base, cond_string, from_tty);
13027 install_breakpoint (0, &c->base, 1);
13028 }
13029
13030 /* Implement the "catch exception" command. */
13031
13032 static void
13033 catch_ada_exception_command (char *arg, int from_tty,
13034 struct cmd_list_element *command)
13035 {
13036 struct gdbarch *gdbarch = get_current_arch ();
13037 int tempflag;
13038 enum ada_exception_catchpoint_kind ex_kind;
13039 char *excep_string = NULL;
13040 char *cond_string = NULL;
13041
13042 tempflag = get_cmd_context (command) == CATCH_TEMPORARY;
13043
13044 if (!arg)
13045 arg = "";
13046 catch_ada_exception_command_split (arg, &ex_kind, &excep_string,
13047 &cond_string);
13048 create_ada_exception_catchpoint (gdbarch, ex_kind,
13049 excep_string, cond_string,
13050 tempflag, 1 /* enabled */,
13051 from_tty);
13052 }
13053
13054 /* Split the arguments specified in a "catch assert" command.
13055
13056 ARGS contains the command's arguments (or the empty string if
13057 no arguments were passed).
13058
13059 If ARGS contains a condition, set COND_STRING to that condition
13060 (the memory needs to be deallocated after use). */
13061
13062 static void
13063 catch_ada_assert_command_split (char *args, char **cond_string)
13064 {
13065 args = skip_spaces (args);
13066
13067 /* Check whether a condition was provided. */
13068 if (startswith (args, "if")
13069 && (isspace (args[2]) || args[2] == '\0'))
13070 {
13071 args += 2;
13072 args = skip_spaces (args);
13073 if (args[0] == '\0')
13074 error (_("condition missing after `if' keyword"));
13075 *cond_string = xstrdup (args);
13076 }
13077
13078 /* Otherwise, there should be no other argument at the end of
13079 the command. */
13080 else if (args[0] != '\0')
13081 error (_("Junk at end of arguments."));
13082 }
13083
13084 /* Implement the "catch assert" command. */
13085
13086 static void
13087 catch_assert_command (char *arg, int from_tty,
13088 struct cmd_list_element *command)
13089 {
13090 struct gdbarch *gdbarch = get_current_arch ();
13091 int tempflag;
13092 char *cond_string = NULL;
13093
13094 tempflag = get_cmd_context (command) == CATCH_TEMPORARY;
13095
13096 if (!arg)
13097 arg = "";
13098 catch_ada_assert_command_split (arg, &cond_string);
13099 create_ada_exception_catchpoint (gdbarch, ada_catch_assert,
13100 NULL, cond_string,
13101 tempflag, 1 /* enabled */,
13102 from_tty);
13103 }
13104
13105 /* Return non-zero if the symbol SYM is an Ada exception object. */
13106
13107 static int
13108 ada_is_exception_sym (struct symbol *sym)
13109 {
13110 const char *type_name = type_name_no_tag (SYMBOL_TYPE (sym));
13111
13112 return (SYMBOL_CLASS (sym) != LOC_TYPEDEF
13113 && SYMBOL_CLASS (sym) != LOC_BLOCK
13114 && SYMBOL_CLASS (sym) != LOC_CONST
13115 && SYMBOL_CLASS (sym) != LOC_UNRESOLVED
13116 && type_name != NULL && strcmp (type_name, "exception") == 0);
13117 }
13118
13119 /* Given a global symbol SYM, return non-zero iff SYM is a non-standard
13120 Ada exception object. This matches all exceptions except the ones
13121 defined by the Ada language. */
13122
13123 static int
13124 ada_is_non_standard_exception_sym (struct symbol *sym)
13125 {
13126 int i;
13127
13128 if (!ada_is_exception_sym (sym))
13129 return 0;
13130
13131 for (i = 0; i < ARRAY_SIZE (standard_exc); i++)
13132 if (strcmp (SYMBOL_LINKAGE_NAME (sym), standard_exc[i]) == 0)
13133 return 0; /* A standard exception. */
13134
13135 /* Numeric_Error is also a standard exception, so exclude it.
13136 See the STANDARD_EXC description for more details as to why
13137 this exception is not listed in that array. */
13138 if (strcmp (SYMBOL_LINKAGE_NAME (sym), "numeric_error") == 0)
13139 return 0;
13140
13141 return 1;
13142 }
13143
13144 /* A helper function for qsort, comparing two struct ada_exc_info
13145 objects.
13146
13147 The comparison is determined first by exception name, and then
13148 by exception address. */
13149
13150 static int
13151 compare_ada_exception_info (const void *a, const void *b)
13152 {
13153 const struct ada_exc_info *exc_a = (struct ada_exc_info *) a;
13154 const struct ada_exc_info *exc_b = (struct ada_exc_info *) b;
13155 int result;
13156
13157 result = strcmp (exc_a->name, exc_b->name);
13158 if (result != 0)
13159 return result;
13160
13161 if (exc_a->addr < exc_b->addr)
13162 return -1;
13163 if (exc_a->addr > exc_b->addr)
13164 return 1;
13165
13166 return 0;
13167 }
13168
13169 /* Sort EXCEPTIONS using compare_ada_exception_info as the comparison
13170 routine, but keeping the first SKIP elements untouched.
13171
13172 All duplicates are also removed. */
13173
13174 static void
13175 sort_remove_dups_ada_exceptions_list (VEC(ada_exc_info) **exceptions,
13176 int skip)
13177 {
13178 struct ada_exc_info *to_sort
13179 = VEC_address (ada_exc_info, *exceptions) + skip;
13180 int to_sort_len
13181 = VEC_length (ada_exc_info, *exceptions) - skip;
13182 int i, j;
13183
13184 qsort (to_sort, to_sort_len, sizeof (struct ada_exc_info),
13185 compare_ada_exception_info);
13186
13187 for (i = 1, j = 1; i < to_sort_len; i++)
13188 if (compare_ada_exception_info (&to_sort[i], &to_sort[j - 1]) != 0)
13189 to_sort[j++] = to_sort[i];
13190 to_sort_len = j;
13191 VEC_truncate(ada_exc_info, *exceptions, skip + to_sort_len);
13192 }
13193
13194 /* A function intended as the "name_matcher" callback in the struct
13195 quick_symbol_functions' expand_symtabs_matching method.
13196
13197 SEARCH_NAME is the symbol's search name.
13198
13199 If USER_DATA is not NULL, it is a pointer to a regext_t object
13200 used to match the symbol (by natural name). Otherwise, when USER_DATA
13201 is null, no filtering is performed, and all symbols are a positive
13202 match. */
13203
13204 static int
13205 ada_exc_search_name_matches (const char *search_name, void *user_data)
13206 {
13207 regex_t *preg = (regex_t *) user_data;
13208
13209 if (preg == NULL)
13210 return 1;
13211
13212 /* In Ada, the symbol "search name" is a linkage name, whereas
13213 the regular expression used to do the matching refers to
13214 the natural name. So match against the decoded name. */
13215 return (regexec (preg, ada_decode (search_name), 0, NULL, 0) == 0);
13216 }
13217
13218 /* Add all exceptions defined by the Ada standard whose name match
13219 a regular expression.
13220
13221 If PREG is not NULL, then this regexp_t object is used to
13222 perform the symbol name matching. Otherwise, no name-based
13223 filtering is performed.
13224
13225 EXCEPTIONS is a vector of exceptions to which matching exceptions
13226 gets pushed. */
13227
13228 static void
13229 ada_add_standard_exceptions (regex_t *preg, VEC(ada_exc_info) **exceptions)
13230 {
13231 int i;
13232
13233 for (i = 0; i < ARRAY_SIZE (standard_exc); i++)
13234 {
13235 if (preg == NULL
13236 || regexec (preg, standard_exc[i], 0, NULL, 0) == 0)
13237 {
13238 struct bound_minimal_symbol msymbol
13239 = ada_lookup_simple_minsym (standard_exc[i]);
13240
13241 if (msymbol.minsym != NULL)
13242 {
13243 struct ada_exc_info info
13244 = {standard_exc[i], BMSYMBOL_VALUE_ADDRESS (msymbol)};
13245
13246 VEC_safe_push (ada_exc_info, *exceptions, &info);
13247 }
13248 }
13249 }
13250 }
13251
13252 /* Add all Ada exceptions defined locally and accessible from the given
13253 FRAME.
13254
13255 If PREG is not NULL, then this regexp_t object is used to
13256 perform the symbol name matching. Otherwise, no name-based
13257 filtering is performed.
13258
13259 EXCEPTIONS is a vector of exceptions to which matching exceptions
13260 gets pushed. */
13261
13262 static void
13263 ada_add_exceptions_from_frame (regex_t *preg, struct frame_info *frame,
13264 VEC(ada_exc_info) **exceptions)
13265 {
13266 const struct block *block = get_frame_block (frame, 0);
13267
13268 while (block != 0)
13269 {
13270 struct block_iterator iter;
13271 struct symbol *sym;
13272
13273 ALL_BLOCK_SYMBOLS (block, iter, sym)
13274 {
13275 switch (SYMBOL_CLASS (sym))
13276 {
13277 case LOC_TYPEDEF:
13278 case LOC_BLOCK:
13279 case LOC_CONST:
13280 break;
13281 default:
13282 if (ada_is_exception_sym (sym))
13283 {
13284 struct ada_exc_info info = {SYMBOL_PRINT_NAME (sym),
13285 SYMBOL_VALUE_ADDRESS (sym)};
13286
13287 VEC_safe_push (ada_exc_info, *exceptions, &info);
13288 }
13289 }
13290 }
13291 if (BLOCK_FUNCTION (block) != NULL)
13292 break;
13293 block = BLOCK_SUPERBLOCK (block);
13294 }
13295 }
13296
13297 /* Add all exceptions defined globally whose name name match
13298 a regular expression, excluding standard exceptions.
13299
13300 The reason we exclude standard exceptions is that they need
13301 to be handled separately: Standard exceptions are defined inside
13302 a runtime unit which is normally not compiled with debugging info,
13303 and thus usually do not show up in our symbol search. However,
13304 if the unit was in fact built with debugging info, we need to
13305 exclude them because they would duplicate the entry we found
13306 during the special loop that specifically searches for those
13307 standard exceptions.
13308
13309 If PREG is not NULL, then this regexp_t object is used to
13310 perform the symbol name matching. Otherwise, no name-based
13311 filtering is performed.
13312
13313 EXCEPTIONS is a vector of exceptions to which matching exceptions
13314 gets pushed. */
13315
13316 static void
13317 ada_add_global_exceptions (regex_t *preg, VEC(ada_exc_info) **exceptions)
13318 {
13319 struct objfile *objfile;
13320 struct compunit_symtab *s;
13321
13322 expand_symtabs_matching (NULL, ada_exc_search_name_matches, NULL,
13323 VARIABLES_DOMAIN, preg);
13324
13325 ALL_COMPUNITS (objfile, s)
13326 {
13327 const struct blockvector *bv = COMPUNIT_BLOCKVECTOR (s);
13328 int i;
13329
13330 for (i = GLOBAL_BLOCK; i <= STATIC_BLOCK; i++)
13331 {
13332 struct block *b = BLOCKVECTOR_BLOCK (bv, i);
13333 struct block_iterator iter;
13334 struct symbol *sym;
13335
13336 ALL_BLOCK_SYMBOLS (b, iter, sym)
13337 if (ada_is_non_standard_exception_sym (sym)
13338 && (preg == NULL
13339 || regexec (preg, SYMBOL_NATURAL_NAME (sym),
13340 0, NULL, 0) == 0))
13341 {
13342 struct ada_exc_info info
13343 = {SYMBOL_PRINT_NAME (sym), SYMBOL_VALUE_ADDRESS (sym)};
13344
13345 VEC_safe_push (ada_exc_info, *exceptions, &info);
13346 }
13347 }
13348 }
13349 }
13350
13351 /* Implements ada_exceptions_list with the regular expression passed
13352 as a regex_t, rather than a string.
13353
13354 If not NULL, PREG is used to filter out exceptions whose names
13355 do not match. Otherwise, all exceptions are listed. */
13356
13357 static VEC(ada_exc_info) *
13358 ada_exceptions_list_1 (regex_t *preg)
13359 {
13360 VEC(ada_exc_info) *result = NULL;
13361 struct cleanup *old_chain
13362 = make_cleanup (VEC_cleanup (ada_exc_info), &result);
13363 int prev_len;
13364
13365 /* First, list the known standard exceptions. These exceptions
13366 need to be handled separately, as they are usually defined in
13367 runtime units that have been compiled without debugging info. */
13368
13369 ada_add_standard_exceptions (preg, &result);
13370
13371 /* Next, find all exceptions whose scope is local and accessible
13372 from the currently selected frame. */
13373
13374 if (has_stack_frames ())
13375 {
13376 prev_len = VEC_length (ada_exc_info, result);
13377 ada_add_exceptions_from_frame (preg, get_selected_frame (NULL),
13378 &result);
13379 if (VEC_length (ada_exc_info, result) > prev_len)
13380 sort_remove_dups_ada_exceptions_list (&result, prev_len);
13381 }
13382
13383 /* Add all exceptions whose scope is global. */
13384
13385 prev_len = VEC_length (ada_exc_info, result);
13386 ada_add_global_exceptions (preg, &result);
13387 if (VEC_length (ada_exc_info, result) > prev_len)
13388 sort_remove_dups_ada_exceptions_list (&result, prev_len);
13389
13390 discard_cleanups (old_chain);
13391 return result;
13392 }
13393
13394 /* Return a vector of ada_exc_info.
13395
13396 If REGEXP is NULL, all exceptions are included in the result.
13397 Otherwise, it should contain a valid regular expression,
13398 and only the exceptions whose names match that regular expression
13399 are included in the result.
13400
13401 The exceptions are sorted in the following order:
13402 - Standard exceptions (defined by the Ada language), in
13403 alphabetical order;
13404 - Exceptions only visible from the current frame, in
13405 alphabetical order;
13406 - Exceptions whose scope is global, in alphabetical order. */
13407
13408 VEC(ada_exc_info) *
13409 ada_exceptions_list (const char *regexp)
13410 {
13411 VEC(ada_exc_info) *result = NULL;
13412 struct cleanup *old_chain = NULL;
13413 regex_t reg;
13414
13415 if (regexp != NULL)
13416 old_chain = compile_rx_or_error (&reg, regexp,
13417 _("invalid regular expression"));
13418
13419 result = ada_exceptions_list_1 (regexp != NULL ? &reg : NULL);
13420
13421 if (old_chain != NULL)
13422 do_cleanups (old_chain);
13423 return result;
13424 }
13425
13426 /* Implement the "info exceptions" command. */
13427
13428 static void
13429 info_exceptions_command (char *regexp, int from_tty)
13430 {
13431 VEC(ada_exc_info) *exceptions;
13432 struct cleanup *cleanup;
13433 struct gdbarch *gdbarch = get_current_arch ();
13434 int ix;
13435 struct ada_exc_info *info;
13436
13437 exceptions = ada_exceptions_list (regexp);
13438 cleanup = make_cleanup (VEC_cleanup (ada_exc_info), &exceptions);
13439
13440 if (regexp != NULL)
13441 printf_filtered
13442 (_("All Ada exceptions matching regular expression \"%s\":\n"), regexp);
13443 else
13444 printf_filtered (_("All defined Ada exceptions:\n"));
13445
13446 for (ix = 0; VEC_iterate(ada_exc_info, exceptions, ix, info); ix++)
13447 printf_filtered ("%s: %s\n", info->name, paddress (gdbarch, info->addr));
13448
13449 do_cleanups (cleanup);
13450 }
13451
13452 /* Operators */
13453 /* Information about operators given special treatment in functions
13454 below. */
13455 /* Format: OP_DEFN (<operator>, <operator length>, <# args>, <binop>). */
13456
13457 #define ADA_OPERATORS \
13458 OP_DEFN (OP_VAR_VALUE, 4, 0, 0) \
13459 OP_DEFN (BINOP_IN_BOUNDS, 3, 2, 0) \
13460 OP_DEFN (TERNOP_IN_RANGE, 1, 3, 0) \
13461 OP_DEFN (OP_ATR_FIRST, 1, 2, 0) \
13462 OP_DEFN (OP_ATR_LAST, 1, 2, 0) \
13463 OP_DEFN (OP_ATR_LENGTH, 1, 2, 0) \
13464 OP_DEFN (OP_ATR_IMAGE, 1, 2, 0) \
13465 OP_DEFN (OP_ATR_MAX, 1, 3, 0) \
13466 OP_DEFN (OP_ATR_MIN, 1, 3, 0) \
13467 OP_DEFN (OP_ATR_MODULUS, 1, 1, 0) \
13468 OP_DEFN (OP_ATR_POS, 1, 2, 0) \
13469 OP_DEFN (OP_ATR_SIZE, 1, 1, 0) \
13470 OP_DEFN (OP_ATR_TAG, 1, 1, 0) \
13471 OP_DEFN (OP_ATR_VAL, 1, 2, 0) \
13472 OP_DEFN (UNOP_QUAL, 3, 1, 0) \
13473 OP_DEFN (UNOP_IN_RANGE, 3, 1, 0) \
13474 OP_DEFN (OP_OTHERS, 1, 1, 0) \
13475 OP_DEFN (OP_POSITIONAL, 3, 1, 0) \
13476 OP_DEFN (OP_DISCRETE_RANGE, 1, 2, 0)
13477
13478 static void
13479 ada_operator_length (const struct expression *exp, int pc, int *oplenp,
13480 int *argsp)
13481 {
13482 switch (exp->elts[pc - 1].opcode)
13483 {
13484 default:
13485 operator_length_standard (exp, pc, oplenp, argsp);
13486 break;
13487
13488 #define OP_DEFN(op, len, args, binop) \
13489 case op: *oplenp = len; *argsp = args; break;
13490 ADA_OPERATORS;
13491 #undef OP_DEFN
13492
13493 case OP_AGGREGATE:
13494 *oplenp = 3;
13495 *argsp = longest_to_int (exp->elts[pc - 2].longconst);
13496 break;
13497
13498 case OP_CHOICES:
13499 *oplenp = 3;
13500 *argsp = longest_to_int (exp->elts[pc - 2].longconst) + 1;
13501 break;
13502 }
13503 }
13504
13505 /* Implementation of the exp_descriptor method operator_check. */
13506
13507 static int
13508 ada_operator_check (struct expression *exp, int pos,
13509 int (*objfile_func) (struct objfile *objfile, void *data),
13510 void *data)
13511 {
13512 const union exp_element *const elts = exp->elts;
13513 struct type *type = NULL;
13514
13515 switch (elts[pos].opcode)
13516 {
13517 case UNOP_IN_RANGE:
13518 case UNOP_QUAL:
13519 type = elts[pos + 1].type;
13520 break;
13521
13522 default:
13523 return operator_check_standard (exp, pos, objfile_func, data);
13524 }
13525
13526 /* Invoke callbacks for TYPE and OBJFILE if they were set as non-NULL. */
13527
13528 if (type && TYPE_OBJFILE (type)
13529 && (*objfile_func) (TYPE_OBJFILE (type), data))
13530 return 1;
13531
13532 return 0;
13533 }
13534
13535 static char *
13536 ada_op_name (enum exp_opcode opcode)
13537 {
13538 switch (opcode)
13539 {
13540 default:
13541 return op_name_standard (opcode);
13542
13543 #define OP_DEFN(op, len, args, binop) case op: return #op;
13544 ADA_OPERATORS;
13545 #undef OP_DEFN
13546
13547 case OP_AGGREGATE:
13548 return "OP_AGGREGATE";
13549 case OP_CHOICES:
13550 return "OP_CHOICES";
13551 case OP_NAME:
13552 return "OP_NAME";
13553 }
13554 }
13555
13556 /* As for operator_length, but assumes PC is pointing at the first
13557 element of the operator, and gives meaningful results only for the
13558 Ada-specific operators, returning 0 for *OPLENP and *ARGSP otherwise. */
13559
13560 static void
13561 ada_forward_operator_length (struct expression *exp, int pc,
13562 int *oplenp, int *argsp)
13563 {
13564 switch (exp->elts[pc].opcode)
13565 {
13566 default:
13567 *oplenp = *argsp = 0;
13568 break;
13569
13570 #define OP_DEFN(op, len, args, binop) \
13571 case op: *oplenp = len; *argsp = args; break;
13572 ADA_OPERATORS;
13573 #undef OP_DEFN
13574
13575 case OP_AGGREGATE:
13576 *oplenp = 3;
13577 *argsp = longest_to_int (exp->elts[pc + 1].longconst);
13578 break;
13579
13580 case OP_CHOICES:
13581 *oplenp = 3;
13582 *argsp = longest_to_int (exp->elts[pc + 1].longconst) + 1;
13583 break;
13584
13585 case OP_STRING:
13586 case OP_NAME:
13587 {
13588 int len = longest_to_int (exp->elts[pc + 1].longconst);
13589
13590 *oplenp = 4 + BYTES_TO_EXP_ELEM (len + 1);
13591 *argsp = 0;
13592 break;
13593 }
13594 }
13595 }
13596
13597 static int
13598 ada_dump_subexp_body (struct expression *exp, struct ui_file *stream, int elt)
13599 {
13600 enum exp_opcode op = exp->elts[elt].opcode;
13601 int oplen, nargs;
13602 int pc = elt;
13603 int i;
13604
13605 ada_forward_operator_length (exp, elt, &oplen, &nargs);
13606
13607 switch (op)
13608 {
13609 /* Ada attributes ('Foo). */
13610 case OP_ATR_FIRST:
13611 case OP_ATR_LAST:
13612 case OP_ATR_LENGTH:
13613 case OP_ATR_IMAGE:
13614 case OP_ATR_MAX:
13615 case OP_ATR_MIN:
13616 case OP_ATR_MODULUS:
13617 case OP_ATR_POS:
13618 case OP_ATR_SIZE:
13619 case OP_ATR_TAG:
13620 case OP_ATR_VAL:
13621 break;
13622
13623 case UNOP_IN_RANGE:
13624 case UNOP_QUAL:
13625 /* XXX: gdb_sprint_host_address, type_sprint */
13626 fprintf_filtered (stream, _("Type @"));
13627 gdb_print_host_address (exp->elts[pc + 1].type, stream);
13628 fprintf_filtered (stream, " (");
13629 type_print (exp->elts[pc + 1].type, NULL, stream, 0);
13630 fprintf_filtered (stream, ")");
13631 break;
13632 case BINOP_IN_BOUNDS:
13633 fprintf_filtered (stream, " (%d)",
13634 longest_to_int (exp->elts[pc + 2].longconst));
13635 break;
13636 case TERNOP_IN_RANGE:
13637 break;
13638
13639 case OP_AGGREGATE:
13640 case OP_OTHERS:
13641 case OP_DISCRETE_RANGE:
13642 case OP_POSITIONAL:
13643 case OP_CHOICES:
13644 break;
13645
13646 case OP_NAME:
13647 case OP_STRING:
13648 {
13649 char *name = &exp->elts[elt + 2].string;
13650 int len = longest_to_int (exp->elts[elt + 1].longconst);
13651
13652 fprintf_filtered (stream, "Text: `%.*s'", len, name);
13653 break;
13654 }
13655
13656 default:
13657 return dump_subexp_body_standard (exp, stream, elt);
13658 }
13659
13660 elt += oplen;
13661 for (i = 0; i < nargs; i += 1)
13662 elt = dump_subexp (exp, stream, elt);
13663
13664 return elt;
13665 }
13666
13667 /* The Ada extension of print_subexp (q.v.). */
13668
13669 static void
13670 ada_print_subexp (struct expression *exp, int *pos,
13671 struct ui_file *stream, enum precedence prec)
13672 {
13673 int oplen, nargs, i;
13674 int pc = *pos;
13675 enum exp_opcode op = exp->elts[pc].opcode;
13676
13677 ada_forward_operator_length (exp, pc, &oplen, &nargs);
13678
13679 *pos += oplen;
13680 switch (op)
13681 {
13682 default:
13683 *pos -= oplen;
13684 print_subexp_standard (exp, pos, stream, prec);
13685 return;
13686
13687 case OP_VAR_VALUE:
13688 fputs_filtered (SYMBOL_NATURAL_NAME (exp->elts[pc + 2].symbol), stream);
13689 return;
13690
13691 case BINOP_IN_BOUNDS:
13692 /* XXX: sprint_subexp */
13693 print_subexp (exp, pos, stream, PREC_SUFFIX);
13694 fputs_filtered (" in ", stream);
13695 print_subexp (exp, pos, stream, PREC_SUFFIX);
13696 fputs_filtered ("'range", stream);
13697 if (exp->elts[pc + 1].longconst > 1)
13698 fprintf_filtered (stream, "(%ld)",
13699 (long) exp->elts[pc + 1].longconst);
13700 return;
13701
13702 case TERNOP_IN_RANGE:
13703 if (prec >= PREC_EQUAL)
13704 fputs_filtered ("(", stream);
13705 /* XXX: sprint_subexp */
13706 print_subexp (exp, pos, stream, PREC_SUFFIX);
13707 fputs_filtered (" in ", stream);
13708 print_subexp (exp, pos, stream, PREC_EQUAL);
13709 fputs_filtered (" .. ", stream);
13710 print_subexp (exp, pos, stream, PREC_EQUAL);
13711 if (prec >= PREC_EQUAL)
13712 fputs_filtered (")", stream);
13713 return;
13714
13715 case OP_ATR_FIRST:
13716 case OP_ATR_LAST:
13717 case OP_ATR_LENGTH:
13718 case OP_ATR_IMAGE:
13719 case OP_ATR_MAX:
13720 case OP_ATR_MIN:
13721 case OP_ATR_MODULUS:
13722 case OP_ATR_POS:
13723 case OP_ATR_SIZE:
13724 case OP_ATR_TAG:
13725 case OP_ATR_VAL:
13726 if (exp->elts[*pos].opcode == OP_TYPE)
13727 {
13728 if (TYPE_CODE (exp->elts[*pos + 1].type) != TYPE_CODE_VOID)
13729 LA_PRINT_TYPE (exp->elts[*pos + 1].type, "", stream, 0, 0,
13730 &type_print_raw_options);
13731 *pos += 3;
13732 }
13733 else
13734 print_subexp (exp, pos, stream, PREC_SUFFIX);
13735 fprintf_filtered (stream, "'%s", ada_attribute_name (op));
13736 if (nargs > 1)
13737 {
13738 int tem;
13739
13740 for (tem = 1; tem < nargs; tem += 1)
13741 {
13742 fputs_filtered ((tem == 1) ? " (" : ", ", stream);
13743 print_subexp (exp, pos, stream, PREC_ABOVE_COMMA);
13744 }
13745 fputs_filtered (")", stream);
13746 }
13747 return;
13748
13749 case UNOP_QUAL:
13750 type_print (exp->elts[pc + 1].type, "", stream, 0);
13751 fputs_filtered ("'(", stream);
13752 print_subexp (exp, pos, stream, PREC_PREFIX);
13753 fputs_filtered (")", stream);
13754 return;
13755
13756 case UNOP_IN_RANGE:
13757 /* XXX: sprint_subexp */
13758 print_subexp (exp, pos, stream, PREC_SUFFIX);
13759 fputs_filtered (" in ", stream);
13760 LA_PRINT_TYPE (exp->elts[pc + 1].type, "", stream, 1, 0,
13761 &type_print_raw_options);
13762 return;
13763
13764 case OP_DISCRETE_RANGE:
13765 print_subexp (exp, pos, stream, PREC_SUFFIX);
13766 fputs_filtered ("..", stream);
13767 print_subexp (exp, pos, stream, PREC_SUFFIX);
13768 return;
13769
13770 case OP_OTHERS:
13771 fputs_filtered ("others => ", stream);
13772 print_subexp (exp, pos, stream, PREC_SUFFIX);
13773 return;
13774
13775 case OP_CHOICES:
13776 for (i = 0; i < nargs-1; i += 1)
13777 {
13778 if (i > 0)
13779 fputs_filtered ("|", stream);
13780 print_subexp (exp, pos, stream, PREC_SUFFIX);
13781 }
13782 fputs_filtered (" => ", stream);
13783 print_subexp (exp, pos, stream, PREC_SUFFIX);
13784 return;
13785
13786 case OP_POSITIONAL:
13787 print_subexp (exp, pos, stream, PREC_SUFFIX);
13788 return;
13789
13790 case OP_AGGREGATE:
13791 fputs_filtered ("(", stream);
13792 for (i = 0; i < nargs; i += 1)
13793 {
13794 if (i > 0)
13795 fputs_filtered (", ", stream);
13796 print_subexp (exp, pos, stream, PREC_SUFFIX);
13797 }
13798 fputs_filtered (")", stream);
13799 return;
13800 }
13801 }
13802
13803 /* Table mapping opcodes into strings for printing operators
13804 and precedences of the operators. */
13805
13806 static const struct op_print ada_op_print_tab[] = {
13807 {":=", BINOP_ASSIGN, PREC_ASSIGN, 1},
13808 {"or else", BINOP_LOGICAL_OR, PREC_LOGICAL_OR, 0},
13809 {"and then", BINOP_LOGICAL_AND, PREC_LOGICAL_AND, 0},
13810 {"or", BINOP_BITWISE_IOR, PREC_BITWISE_IOR, 0},
13811 {"xor", BINOP_BITWISE_XOR, PREC_BITWISE_XOR, 0},
13812 {"and", BINOP_BITWISE_AND, PREC_BITWISE_AND, 0},
13813 {"=", BINOP_EQUAL, PREC_EQUAL, 0},
13814 {"/=", BINOP_NOTEQUAL, PREC_EQUAL, 0},
13815 {"<=", BINOP_LEQ, PREC_ORDER, 0},
13816 {">=", BINOP_GEQ, PREC_ORDER, 0},
13817 {">", BINOP_GTR, PREC_ORDER, 0},
13818 {"<", BINOP_LESS, PREC_ORDER, 0},
13819 {">>", BINOP_RSH, PREC_SHIFT, 0},
13820 {"<<", BINOP_LSH, PREC_SHIFT, 0},
13821 {"+", BINOP_ADD, PREC_ADD, 0},
13822 {"-", BINOP_SUB, PREC_ADD, 0},
13823 {"&", BINOP_CONCAT, PREC_ADD, 0},
13824 {"*", BINOP_MUL, PREC_MUL, 0},
13825 {"/", BINOP_DIV, PREC_MUL, 0},
13826 {"rem", BINOP_REM, PREC_MUL, 0},
13827 {"mod", BINOP_MOD, PREC_MUL, 0},
13828 {"**", BINOP_EXP, PREC_REPEAT, 0},
13829 {"@", BINOP_REPEAT, PREC_REPEAT, 0},
13830 {"-", UNOP_NEG, PREC_PREFIX, 0},
13831 {"+", UNOP_PLUS, PREC_PREFIX, 0},
13832 {"not ", UNOP_LOGICAL_NOT, PREC_PREFIX, 0},
13833 {"not ", UNOP_COMPLEMENT, PREC_PREFIX, 0},
13834 {"abs ", UNOP_ABS, PREC_PREFIX, 0},
13835 {".all", UNOP_IND, PREC_SUFFIX, 1},
13836 {"'access", UNOP_ADDR, PREC_SUFFIX, 1},
13837 {"'size", OP_ATR_SIZE, PREC_SUFFIX, 1},
13838 {NULL, OP_NULL, PREC_SUFFIX, 0}
13839 };
13840 \f
13841 enum ada_primitive_types {
13842 ada_primitive_type_int,
13843 ada_primitive_type_long,
13844 ada_primitive_type_short,
13845 ada_primitive_type_char,
13846 ada_primitive_type_float,
13847 ada_primitive_type_double,
13848 ada_primitive_type_void,
13849 ada_primitive_type_long_long,
13850 ada_primitive_type_long_double,
13851 ada_primitive_type_natural,
13852 ada_primitive_type_positive,
13853 ada_primitive_type_system_address,
13854 nr_ada_primitive_types
13855 };
13856
13857 static void
13858 ada_language_arch_info (struct gdbarch *gdbarch,
13859 struct language_arch_info *lai)
13860 {
13861 const struct builtin_type *builtin = builtin_type (gdbarch);
13862
13863 lai->primitive_type_vector
13864 = GDBARCH_OBSTACK_CALLOC (gdbarch, nr_ada_primitive_types + 1,
13865 struct type *);
13866
13867 lai->primitive_type_vector [ada_primitive_type_int]
13868 = arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch),
13869 0, "integer");
13870 lai->primitive_type_vector [ada_primitive_type_long]
13871 = arch_integer_type (gdbarch, gdbarch_long_bit (gdbarch),
13872 0, "long_integer");
13873 lai->primitive_type_vector [ada_primitive_type_short]
13874 = arch_integer_type (gdbarch, gdbarch_short_bit (gdbarch),
13875 0, "short_integer");
13876 lai->string_char_type
13877 = lai->primitive_type_vector [ada_primitive_type_char]
13878 = arch_character_type (gdbarch, TARGET_CHAR_BIT, 0, "character");
13879 lai->primitive_type_vector [ada_primitive_type_float]
13880 = arch_float_type (gdbarch, gdbarch_float_bit (gdbarch),
13881 "float", NULL);
13882 lai->primitive_type_vector [ada_primitive_type_double]
13883 = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
13884 "long_float", NULL);
13885 lai->primitive_type_vector [ada_primitive_type_long_long]
13886 = arch_integer_type (gdbarch, gdbarch_long_long_bit (gdbarch),
13887 0, "long_long_integer");
13888 lai->primitive_type_vector [ada_primitive_type_long_double]
13889 = arch_float_type (gdbarch, gdbarch_double_bit (gdbarch),
13890 "long_long_float", NULL);
13891 lai->primitive_type_vector [ada_primitive_type_natural]
13892 = arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch),
13893 0, "natural");
13894 lai->primitive_type_vector [ada_primitive_type_positive]
13895 = arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch),
13896 0, "positive");
13897 lai->primitive_type_vector [ada_primitive_type_void]
13898 = builtin->builtin_void;
13899
13900 lai->primitive_type_vector [ada_primitive_type_system_address]
13901 = lookup_pointer_type (arch_type (gdbarch, TYPE_CODE_VOID, 1, "void"));
13902 TYPE_NAME (lai->primitive_type_vector [ada_primitive_type_system_address])
13903 = "system__address";
13904
13905 lai->bool_type_symbol = NULL;
13906 lai->bool_type_default = builtin->builtin_bool;
13907 }
13908 \f
13909 /* Language vector */
13910
13911 /* Not really used, but needed in the ada_language_defn. */
13912
13913 static void
13914 emit_char (int c, struct type *type, struct ui_file *stream, int quoter)
13915 {
13916 ada_emit_char (c, type, stream, quoter, 1);
13917 }
13918
13919 static int
13920 parse (struct parser_state *ps)
13921 {
13922 warnings_issued = 0;
13923 return ada_parse (ps);
13924 }
13925
13926 static const struct exp_descriptor ada_exp_descriptor = {
13927 ada_print_subexp,
13928 ada_operator_length,
13929 ada_operator_check,
13930 ada_op_name,
13931 ada_dump_subexp_body,
13932 ada_evaluate_subexp
13933 };
13934
13935 /* Implement the "la_get_symbol_name_cmp" language_defn method
13936 for Ada. */
13937
13938 static symbol_name_cmp_ftype
13939 ada_get_symbol_name_cmp (const char *lookup_name)
13940 {
13941 if (should_use_wild_match (lookup_name))
13942 return wild_match;
13943 else
13944 return compare_names;
13945 }
13946
13947 /* Implement the "la_read_var_value" language_defn method for Ada. */
13948
13949 static struct value *
13950 ada_read_var_value (struct symbol *var, const struct block *var_block,
13951 struct frame_info *frame)
13952 {
13953 const struct block *frame_block = NULL;
13954 struct symbol *renaming_sym = NULL;
13955
13956 /* The only case where default_read_var_value is not sufficient
13957 is when VAR is a renaming... */
13958 if (frame)
13959 frame_block = get_frame_block (frame, NULL);
13960 if (frame_block)
13961 renaming_sym = ada_find_renaming_symbol (var, frame_block);
13962 if (renaming_sym != NULL)
13963 return ada_read_renaming_var_value (renaming_sym, frame_block);
13964
13965 /* This is a typical case where we expect the default_read_var_value
13966 function to work. */
13967 return default_read_var_value (var, var_block, frame);
13968 }
13969
13970 const struct language_defn ada_language_defn = {
13971 "ada", /* Language name */
13972 "Ada",
13973 language_ada,
13974 range_check_off,
13975 case_sensitive_on, /* Yes, Ada is case-insensitive, but
13976 that's not quite what this means. */
13977 array_row_major,
13978 macro_expansion_no,
13979 &ada_exp_descriptor,
13980 parse,
13981 ada_error,
13982 resolve,
13983 ada_printchar, /* Print a character constant */
13984 ada_printstr, /* Function to print string constant */
13985 emit_char, /* Function to print single char (not used) */
13986 ada_print_type, /* Print a type using appropriate syntax */
13987 ada_print_typedef, /* Print a typedef using appropriate syntax */
13988 ada_val_print, /* Print a value using appropriate syntax */
13989 ada_value_print, /* Print a top-level value */
13990 ada_read_var_value, /* la_read_var_value */
13991 NULL, /* Language specific skip_trampoline */
13992 NULL, /* name_of_this */
13993 ada_lookup_symbol_nonlocal, /* Looking up non-local symbols. */
13994 basic_lookup_transparent_type, /* lookup_transparent_type */
13995 ada_la_decode, /* Language specific symbol demangler */
13996 NULL, /* Language specific
13997 class_name_from_physname */
13998 ada_op_print_tab, /* expression operators for printing */
13999 0, /* c-style arrays */
14000 1, /* String lower bound */
14001 ada_get_gdb_completer_word_break_characters,
14002 ada_make_symbol_completion_list,
14003 ada_language_arch_info,
14004 ada_print_array_index,
14005 default_pass_by_reference,
14006 c_get_string,
14007 ada_get_symbol_name_cmp, /* la_get_symbol_name_cmp */
14008 ada_iterate_over_symbols,
14009 &ada_varobj_ops,
14010 NULL,
14011 NULL,
14012 LANG_MAGIC
14013 };
14014
14015 /* Provide a prototype to silence -Wmissing-prototypes. */
14016 extern initialize_file_ftype _initialize_ada_language;
14017
14018 /* Command-list for the "set/show ada" prefix command. */
14019 static struct cmd_list_element *set_ada_list;
14020 static struct cmd_list_element *show_ada_list;
14021
14022 /* Implement the "set ada" prefix command. */
14023
14024 static void
14025 set_ada_command (char *arg, int from_tty)
14026 {
14027 printf_unfiltered (_(\
14028 "\"set ada\" must be followed by the name of a setting.\n"));
14029 help_list (set_ada_list, "set ada ", all_commands, gdb_stdout);
14030 }
14031
14032 /* Implement the "show ada" prefix command. */
14033
14034 static void
14035 show_ada_command (char *args, int from_tty)
14036 {
14037 cmd_show_list (show_ada_list, from_tty, "");
14038 }
14039
14040 static void
14041 initialize_ada_catchpoint_ops (void)
14042 {
14043 struct breakpoint_ops *ops;
14044
14045 initialize_breakpoint_ops ();
14046
14047 ops = &catch_exception_breakpoint_ops;
14048 *ops = bkpt_breakpoint_ops;
14049 ops->dtor = dtor_catch_exception;
14050 ops->allocate_location = allocate_location_catch_exception;
14051 ops->re_set = re_set_catch_exception;
14052 ops->check_status = check_status_catch_exception;
14053 ops->print_it = print_it_catch_exception;
14054 ops->print_one = print_one_catch_exception;
14055 ops->print_mention = print_mention_catch_exception;
14056 ops->print_recreate = print_recreate_catch_exception;
14057
14058 ops = &catch_exception_unhandled_breakpoint_ops;
14059 *ops = bkpt_breakpoint_ops;
14060 ops->dtor = dtor_catch_exception_unhandled;
14061 ops->allocate_location = allocate_location_catch_exception_unhandled;
14062 ops->re_set = re_set_catch_exception_unhandled;
14063 ops->check_status = check_status_catch_exception_unhandled;
14064 ops->print_it = print_it_catch_exception_unhandled;
14065 ops->print_one = print_one_catch_exception_unhandled;
14066 ops->print_mention = print_mention_catch_exception_unhandled;
14067 ops->print_recreate = print_recreate_catch_exception_unhandled;
14068
14069 ops = &catch_assert_breakpoint_ops;
14070 *ops = bkpt_breakpoint_ops;
14071 ops->dtor = dtor_catch_assert;
14072 ops->allocate_location = allocate_location_catch_assert;
14073 ops->re_set = re_set_catch_assert;
14074 ops->check_status = check_status_catch_assert;
14075 ops->print_it = print_it_catch_assert;
14076 ops->print_one = print_one_catch_assert;
14077 ops->print_mention = print_mention_catch_assert;
14078 ops->print_recreate = print_recreate_catch_assert;
14079 }
14080
14081 /* This module's 'new_objfile' observer. */
14082
14083 static void
14084 ada_new_objfile_observer (struct objfile *objfile)
14085 {
14086 ada_clear_symbol_cache ();
14087 }
14088
14089 /* This module's 'free_objfile' observer. */
14090
14091 static void
14092 ada_free_objfile_observer (struct objfile *objfile)
14093 {
14094 ada_clear_symbol_cache ();
14095 }
14096
14097 void
14098 _initialize_ada_language (void)
14099 {
14100 add_language (&ada_language_defn);
14101
14102 initialize_ada_catchpoint_ops ();
14103
14104 add_prefix_cmd ("ada", no_class, set_ada_command,
14105 _("Prefix command for changing Ada-specfic settings"),
14106 &set_ada_list, "set ada ", 0, &setlist);
14107
14108 add_prefix_cmd ("ada", no_class, show_ada_command,
14109 _("Generic command for showing Ada-specific settings."),
14110 &show_ada_list, "show ada ", 0, &showlist);
14111
14112 add_setshow_boolean_cmd ("trust-PAD-over-XVS", class_obscure,
14113 &trust_pad_over_xvs, _("\
14114 Enable or disable an optimization trusting PAD types over XVS types"), _("\
14115 Show whether an optimization trusting PAD types over XVS types is activated"),
14116 _("\
14117 This is related to the encoding used by the GNAT compiler. The debugger\n\
14118 should normally trust the contents of PAD types, but certain older versions\n\
14119 of GNAT have a bug that sometimes causes the information in the PAD type\n\
14120 to be incorrect. Turning this setting \"off\" allows the debugger to\n\
14121 work around this bug. It is always safe to turn this option \"off\", but\n\
14122 this incurs a slight performance penalty, so it is recommended to NOT change\n\
14123 this option to \"off\" unless necessary."),
14124 NULL, NULL, &set_ada_list, &show_ada_list);
14125
14126 add_catch_command ("exception", _("\
14127 Catch Ada exceptions, when raised.\n\
14128 With an argument, catch only exceptions with the given name."),
14129 catch_ada_exception_command,
14130 NULL,
14131 CATCH_PERMANENT,
14132 CATCH_TEMPORARY);
14133 add_catch_command ("assert", _("\
14134 Catch failed Ada assertions, when raised.\n\
14135 With an argument, catch only exceptions with the given name."),
14136 catch_assert_command,
14137 NULL,
14138 CATCH_PERMANENT,
14139 CATCH_TEMPORARY);
14140
14141 varsize_limit = 65536;
14142
14143 add_info ("exceptions", info_exceptions_command,
14144 _("\
14145 List all Ada exception names.\n\
14146 If a regular expression is passed as an argument, only those matching\n\
14147 the regular expression are listed."));
14148
14149 add_prefix_cmd ("ada", class_maintenance, maint_set_ada_cmd,
14150 _("Set Ada maintenance-related variables."),
14151 &maint_set_ada_cmdlist, "maintenance set ada ",
14152 0/*allow-unknown*/, &maintenance_set_cmdlist);
14153
14154 add_prefix_cmd ("ada", class_maintenance, maint_show_ada_cmd,
14155 _("Show Ada maintenance-related variables"),
14156 &maint_show_ada_cmdlist, "maintenance show ada ",
14157 0/*allow-unknown*/, &maintenance_show_cmdlist);
14158
14159 add_setshow_boolean_cmd
14160 ("ignore-descriptive-types", class_maintenance,
14161 &ada_ignore_descriptive_types_p,
14162 _("Set whether descriptive types generated by GNAT should be ignored."),
14163 _("Show whether descriptive types generated by GNAT should be ignored."),
14164 _("\
14165 When enabled, the debugger will stop using the DW_AT_GNAT_descriptive_type\n\
14166 DWARF attribute."),
14167 NULL, NULL, &maint_set_ada_cmdlist, &maint_show_ada_cmdlist);
14168
14169 obstack_init (&symbol_list_obstack);
14170
14171 decoded_names_store = htab_create_alloc
14172 (256, htab_hash_string, (int (*)(const void *, const void *)) streq,
14173 NULL, xcalloc, xfree);
14174
14175 /* The ada-lang observers. */
14176 observer_attach_new_objfile (ada_new_objfile_observer);
14177 observer_attach_free_objfile (ada_free_objfile_observer);
14178 observer_attach_inferior_exit (ada_inferior_exit);
14179
14180 /* Setup various context-specific data. */
14181 ada_inferior_data
14182 = register_inferior_data_with_cleanup (NULL, ada_inferior_data_cleanup);
14183 ada_pspace_data_handle
14184 = register_program_space_data_with_cleanup (NULL, ada_pspace_data_cleanup);
14185 }
This page took 0.33253 seconds and 4 git commands to generate.