1 /* Ada language support routines for GDB, the GNU debugger. 2 3 Copyright (C) 1992-2023 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 "gdbsupport/gdb_regex.h" 24 #include "frame.h" 25 #include "symtab.h" 26 #include "gdbtypes.h" 27 #include "gdbcmd.h" 28 #include "expression.h" 29 #include "parser-defs.h" 30 #include "language.h" 31 #include "varobj.h" 32 #include "inferior.h" 33 #include "symfile.h" 34 #include "objfiles.h" 35 #include "breakpoint.h" 36 #include "gdbcore.h" 37 #include "hashtab.h" 38 #include "gdbsupport/gdb_obstack.h" 39 #include "ada-lang.h" 40 #include "completer.h" 41 #include "ui-out.h" 42 #include "block.h" 43 #include "infcall.h" 44 #include "annotate.h" 45 #include "valprint.h" 46 #include "source.h" 47 #include "observable.h" 48 #include "stack.h" 49 #include "typeprint.h" 50 #include "namespace.h" 51 #include "cli/cli-style.h" 52 #include "cli/cli-decode.h" 53 54 #include "value.h" 55 #include "mi/mi-common.h" 56 #include "arch-utils.h" 57 #include "cli/cli-utils.h" 58 #include "gdbsupport/function-view.h" 59 #include "gdbsupport/byte-vector.h" 60 #include <algorithm> 61 #include "ada-exp.h" 62 #include "charset.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_args_match (struct symbol *, struct value **, int); 101 102 static struct value *make_array_descriptor (struct type *, struct value *); 103 104 static void ada_add_block_symbols (std::vector<struct block_symbol> &, 105 const struct block *, 106 const lookup_name_info &lookup_name, 107 domain_enum, struct objfile *); 108 109 static void ada_add_all_symbols (std::vector<struct block_symbol> &, 110 const struct block *, 111 const lookup_name_info &lookup_name, 112 domain_enum, int, int *); 113 114 static int is_nonfunction (const std::vector<struct block_symbol> &); 115 116 static void add_defn_to_vec (std::vector<struct block_symbol> &, 117 struct symbol *, 118 const struct block *); 119 120 static int possible_user_operator_p (enum exp_opcode, struct value **); 121 122 static const char *ada_decoded_op_name (enum exp_opcode); 123 124 static int numeric_type_p (struct type *); 125 126 static int integer_type_p (struct type *); 127 128 static int scalar_type_p (struct type *); 129 130 static int discrete_type_p (struct type *); 131 132 static struct type *ada_lookup_struct_elt_type (struct type *, const char *, 133 int, int); 134 135 static struct type *ada_find_parallel_type_with_name (struct type *, 136 const char *); 137 138 static int is_dynamic_field (struct type *, int); 139 140 static struct type *to_fixed_variant_branch_type (struct type *, 141 const gdb_byte *, 142 CORE_ADDR, struct value *); 143 144 static struct type *to_fixed_array_type (struct type *, struct value *, int); 145 146 static struct type *to_fixed_range_type (struct type *, struct value *); 147 148 static struct type *to_static_fixed_type (struct type *); 149 static struct type *static_unwrap_type (struct type *type); 150 151 static struct value *unwrap_value (struct value *); 152 153 static struct type *constrained_packed_array_type (struct type *, long *); 154 155 static struct type *decode_constrained_packed_array_type (struct type *); 156 157 static long decode_packed_array_bitsize (struct type *); 158 159 static struct value *decode_constrained_packed_array (struct value *); 160 161 static int ada_is_unconstrained_packed_array_type (struct type *); 162 163 static struct value *value_subscript_packed (struct value *, int, 164 struct value **); 165 166 static struct value *coerce_unspec_val_to_type (struct value *, 167 struct type *); 168 169 static int lesseq_defined_than (struct symbol *, struct symbol *); 170 171 static int equiv_types (struct type *, struct type *); 172 173 static int is_name_suffix (const char *); 174 175 static int advance_wild_match (const char **, const char *, char); 176 177 static bool wild_match (const char *name, const char *patn); 178 179 static struct value *ada_coerce_ref (struct value *); 180 181 static LONGEST pos_atr (struct value *); 182 183 static struct value *val_atr (struct type *, LONGEST); 184 185 static struct symbol *standard_lookup (const char *, const struct block *, 186 domain_enum); 187 188 static struct value *ada_search_struct_field (const char *, struct value *, int, 189 struct type *); 190 191 static int find_struct_field (const char *, struct type *, int, 192 struct type **, int *, int *, int *, int *); 193 194 static int ada_resolve_function (std::vector<struct block_symbol> &, 195 struct value **, int, const char *, 196 struct type *, bool); 197 198 static int ada_is_direct_array_type (struct type *); 199 200 static struct value *ada_index_struct_field (int, struct value *, int, 201 struct type *); 202 203 static void add_component_interval (LONGEST, LONGEST, std::vector<LONGEST> &); 204 205 206 static struct type *ada_find_any_type (const char *name); 207 208 static symbol_name_matcher_ftype *ada_get_symbol_name_matcher 209 (const lookup_name_info &lookup_name); 210 211 212 213 /* The character set used for source files. */ 214 static const char *ada_source_charset; 215 216 /* The string "UTF-8". This is here so we can check for the UTF-8 217 charset using == rather than strcmp. */ 218 static const char ada_utf8[] = "UTF-8"; 219 220 /* Each entry in the UTF-32 case-folding table is of this form. */ 221 struct utf8_entry 222 { 223 /* The start and end, inclusive, of this range of codepoints. */ 224 uint32_t start, end; 225 /* The delta to apply to get the upper-case form. 0 if this is 226 already upper-case. */ 227 int upper_delta; 228 /* The delta to apply to get the lower-case form. 0 if this is 229 already lower-case. */ 230 int lower_delta; 231 232 bool operator< (uint32_t val) const 233 { 234 return end < val; 235 } 236 }; 237 238 static const utf8_entry ada_case_fold[] = 239 { 240 #include "ada-casefold.h" 241 }; 242 243 244 245 /* The result of a symbol lookup to be stored in our symbol cache. */ 246 247 struct cache_entry 248 { 249 /* The name used to perform the lookup. */ 250 const char *name; 251 /* The namespace used during the lookup. */ 252 domain_enum domain; 253 /* The symbol returned by the lookup, or NULL if no matching symbol 254 was found. */ 255 struct symbol *sym; 256 /* The block where the symbol was found, or NULL if no matching 257 symbol was found. */ 258 const struct block *block; 259 /* A pointer to the next entry with the same hash. */ 260 struct cache_entry *next; 261 }; 262 263 /* The Ada symbol cache, used to store the result of Ada-mode symbol 264 lookups in the course of executing the user's commands. 265 266 The cache is implemented using a simple, fixed-sized hash. 267 The size is fixed on the grounds that there are not likely to be 268 all that many symbols looked up during any given session, regardless 269 of the size of the symbol table. If we decide to go to a resizable 270 table, let's just use the stuff from libiberty instead. */ 271 272 #define HASH_SIZE 1009 273 274 struct ada_symbol_cache 275 { 276 /* An obstack used to store the entries in our cache. */ 277 struct auto_obstack cache_space; 278 279 /* The root of the hash table used to implement our symbol cache. */ 280 struct cache_entry *root[HASH_SIZE] {}; 281 }; 282 283 static const char ada_completer_word_break_characters[] = 284 #ifdef VMS 285 " \t\n!@#%^&*()+=|~`}{[]\";:?/,-"; 286 #else 287 " \t\n!@#$%^&*()+=|~`}{[]\";:?/,-"; 288 #endif 289 290 /* The name of the symbol to use to get the name of the main subprogram. */ 291 static const char ADA_MAIN_PROGRAM_SYMBOL_NAME[] 292 = "__gnat_ada_main_program_name"; 293 294 /* Limit on the number of warnings to raise per expression evaluation. */ 295 static int warning_limit = 2; 296 297 /* Number of warning messages issued; reset to 0 by cleanups after 298 expression evaluation. */ 299 static int warnings_issued = 0; 300 301 static const char * const known_runtime_file_name_patterns[] = { 302 ADA_KNOWN_RUNTIME_FILE_NAME_PATTERNS NULL 303 }; 304 305 static const char * const known_auxiliary_function_name_patterns[] = { 306 ADA_KNOWN_AUXILIARY_FUNCTION_NAME_PATTERNS NULL 307 }; 308 309 /* Maintenance-related settings for this module. */ 310 311 static struct cmd_list_element *maint_set_ada_cmdlist; 312 static struct cmd_list_element *maint_show_ada_cmdlist; 313 314 /* The "maintenance ada set/show ignore-descriptive-type" value. */ 315 316 static bool ada_ignore_descriptive_types_p = false; 317 318 /* Inferior-specific data. */ 319 320 /* Per-inferior data for this module. */ 321 322 struct ada_inferior_data 323 { 324 /* The ada__tags__type_specific_data type, which is used when decoding 325 tagged types. With older versions of GNAT, this type was directly 326 accessible through a component ("tsd") in the object tag. But this 327 is no longer the case, so we cache it for each inferior. */ 328 struct type *tsd_type = nullptr; 329 330 /* The exception_support_info data. This data is used to determine 331 how to implement support for Ada exception catchpoints in a given 332 inferior. */ 333 const struct exception_support_info *exception_info = nullptr; 334 }; 335 336 /* Our key to this module's inferior data. */ 337 static const registry<inferior>::key<ada_inferior_data> ada_inferior_data; 338 339 /* Return our inferior data for the given inferior (INF). 340 341 This function always returns a valid pointer to an allocated 342 ada_inferior_data structure. If INF's inferior data has not 343 been previously set, this functions creates a new one with all 344 fields set to zero, sets INF's inferior to it, and then returns 345 a pointer to that newly allocated ada_inferior_data. */ 346 347 static struct ada_inferior_data * 348 get_ada_inferior_data (struct inferior *inf) 349 { 350 struct ada_inferior_data *data; 351 352 data = ada_inferior_data.get (inf); 353 if (data == NULL) 354 data = ada_inferior_data.emplace (inf); 355 356 return data; 357 } 358 359 /* Perform all necessary cleanups regarding our module's inferior data 360 that is required after the inferior INF just exited. */ 361 362 static void 363 ada_inferior_exit (struct inferior *inf) 364 { 365 ada_inferior_data.clear (inf); 366 } 367 368 369 /* program-space-specific data. */ 370 371 /* This module's per-program-space data. */ 372 struct ada_pspace_data 373 { 374 /* The Ada symbol cache. */ 375 std::unique_ptr<ada_symbol_cache> sym_cache; 376 }; 377 378 /* Key to our per-program-space data. */ 379 static const registry<program_space>::key<ada_pspace_data> 380 ada_pspace_data_handle; 381 382 /* Return this module's data for the given program space (PSPACE). 383 If not is found, add a zero'ed one now. 384 385 This function always returns a valid object. */ 386 387 static struct ada_pspace_data * 388 get_ada_pspace_data (struct program_space *pspace) 389 { 390 struct ada_pspace_data *data; 391 392 data = ada_pspace_data_handle.get (pspace); 393 if (data == NULL) 394 data = ada_pspace_data_handle.emplace (pspace); 395 396 return data; 397 } 398 399 /* Utilities */ 400 401 /* If TYPE is a TYPE_CODE_TYPEDEF type, return the target type after 402 all typedef layers have been peeled. Otherwise, return TYPE. 403 404 Normally, we really expect a typedef type to only have 1 typedef layer. 405 In other words, we really expect the target type of a typedef type to be 406 a non-typedef type. This is particularly true for Ada units, because 407 the language does not have a typedef vs not-typedef distinction. 408 In that respect, the Ada compiler has been trying to eliminate as many 409 typedef definitions in the debugging information, since they generally 410 do not bring any extra information (we still use typedef under certain 411 circumstances related mostly to the GNAT encoding). 412 413 Unfortunately, we have seen situations where the debugging information 414 generated by the compiler leads to such multiple typedef layers. For 415 instance, consider the following example with stabs: 416 417 .stabs "pck__float_array___XUP:Tt(0,46)=s16P_ARRAY:(0,47)=[...]"[...] 418 .stabs "pck__float_array___XUP:t(0,36)=(0,46)",128,0,6,0 419 420 This is an error in the debugging information which causes type 421 pck__float_array___XUP to be defined twice, and the second time, 422 it is defined as a typedef of a typedef. 423 424 This is on the fringe of legality as far as debugging information is 425 concerned, and certainly unexpected. But it is easy to handle these 426 situations correctly, so we can afford to be lenient in this case. */ 427 428 static struct type * 429 ada_typedef_target_type (struct type *type) 430 { 431 while (type->code () == TYPE_CODE_TYPEDEF) 432 type = type->target_type (); 433 return type; 434 } 435 436 /* Given DECODED_NAME a string holding a symbol name in its 437 decoded form (ie using the Ada dotted notation), returns 438 its unqualified name. */ 439 440 static const char * 441 ada_unqualified_name (const char *decoded_name) 442 { 443 const char *result; 444 445 /* If the decoded name starts with '<', it means that the encoded 446 name does not follow standard naming conventions, and thus that 447 it is not your typical Ada symbol name. Trying to unqualify it 448 is therefore pointless and possibly erroneous. */ 449 if (decoded_name[0] == '<') 450 return decoded_name; 451 452 result = strrchr (decoded_name, '.'); 453 if (result != NULL) 454 result++; /* Skip the dot... */ 455 else 456 result = decoded_name; 457 458 return result; 459 } 460 461 /* Return a string starting with '<', followed by STR, and '>'. */ 462 463 static std::string 464 add_angle_brackets (const char *str) 465 { 466 return string_printf ("<%s>", str); 467 } 468 469 /* True (non-zero) iff TARGET matches FIELD_NAME up to any trailing 470 suffix of FIELD_NAME beginning "___". */ 471 472 static int 473 field_name_match (const char *field_name, const char *target) 474 { 475 int len = strlen (target); 476 477 return 478 (strncmp (field_name, target, len) == 0 479 && (field_name[len] == '\0' 480 || (startswith (field_name + len, "___") 481 && strcmp (field_name + strlen (field_name) - 6, 482 "___XVN") != 0))); 483 } 484 485 486 /* Assuming TYPE is a TYPE_CODE_STRUCT or a TYPE_CODE_TYPDEF to 487 a TYPE_CODE_STRUCT, find the field whose name matches FIELD_NAME, 488 and return its index. This function also handles fields whose name 489 have ___ suffixes because the compiler sometimes alters their name 490 by adding such a suffix to represent fields with certain constraints. 491 If the field could not be found, return a negative number if 492 MAYBE_MISSING is set. Otherwise raise an error. */ 493 494 int 495 ada_get_field_index (const struct type *type, const char *field_name, 496 int maybe_missing) 497 { 498 int fieldno; 499 struct type *struct_type = check_typedef ((struct type *) type); 500 501 for (fieldno = 0; fieldno < struct_type->num_fields (); fieldno++) 502 if (field_name_match (struct_type->field (fieldno).name (), field_name)) 503 return fieldno; 504 505 if (!maybe_missing) 506 error (_("Unable to find field %s in struct %s. Aborting"), 507 field_name, struct_type->name ()); 508 509 return -1; 510 } 511 512 /* The length of the prefix of NAME prior to any "___" suffix. */ 513 514 int 515 ada_name_prefix_len (const char *name) 516 { 517 if (name == NULL) 518 return 0; 519 else 520 { 521 const char *p = strstr (name, "___"); 522 523 if (p == NULL) 524 return strlen (name); 525 else 526 return p - name; 527 } 528 } 529 530 /* Return non-zero if SUFFIX is a suffix of STR. 531 Return zero if STR is null. */ 532 533 static int 534 is_suffix (const char *str, const char *suffix) 535 { 536 int len1, len2; 537 538 if (str == NULL) 539 return 0; 540 len1 = strlen (str); 541 len2 = strlen (suffix); 542 return (len1 >= len2 && strcmp (str + len1 - len2, suffix) == 0); 543 } 544 545 /* The contents of value VAL, treated as a value of type TYPE. The 546 result is an lval in memory if VAL is. */ 547 548 static struct value * 549 coerce_unspec_val_to_type (struct value *val, struct type *type) 550 { 551 type = ada_check_typedef (type); 552 if (value_type (val) == type) 553 return val; 554 else 555 { 556 struct value *result; 557 558 if (value_optimized_out (val)) 559 result = allocate_optimized_out_value (type); 560 else if (value_lazy (val) 561 /* Be careful not to make a lazy not_lval value. */ 562 || (VALUE_LVAL (val) != not_lval 563 && type->length () > value_type (val)->length ())) 564 result = allocate_value_lazy (type); 565 else 566 { 567 result = allocate_value (type); 568 value_contents_copy (result, 0, val, 0, type->length ()); 569 } 570 set_value_component_location (result, val); 571 set_value_bitsize (result, value_bitsize (val)); 572 set_value_bitpos (result, value_bitpos (val)); 573 if (VALUE_LVAL (result) == lval_memory) 574 set_value_address (result, value_address (val)); 575 return result; 576 } 577 } 578 579 static const gdb_byte * 580 cond_offset_host (const gdb_byte *valaddr, long offset) 581 { 582 if (valaddr == NULL) 583 return NULL; 584 else 585 return valaddr + offset; 586 } 587 588 static CORE_ADDR 589 cond_offset_target (CORE_ADDR address, long offset) 590 { 591 if (address == 0) 592 return 0; 593 else 594 return address + offset; 595 } 596 597 /* Issue a warning (as for the definition of warning in utils.c, but 598 with exactly one argument rather than ...), unless the limit on the 599 number of warnings has passed during the evaluation of the current 600 expression. */ 601 602 /* FIXME: cagney/2004-10-10: This function is mimicking the behavior 603 provided by "complaint". */ 604 static void lim_warning (const char *format, ...) ATTRIBUTE_PRINTF (1, 2); 605 606 static void 607 lim_warning (const char *format, ...) 608 { 609 va_list args; 610 611 va_start (args, format); 612 warnings_issued += 1; 613 if (warnings_issued <= warning_limit) 614 vwarning (format, args); 615 616 va_end (args); 617 } 618 619 /* Maximum value of a SIZE-byte signed integer type. */ 620 static LONGEST 621 max_of_size (int size) 622 { 623 LONGEST top_bit = (LONGEST) 1 << (size * 8 - 2); 624 625 return top_bit | (top_bit - 1); 626 } 627 628 /* Minimum value of a SIZE-byte signed integer type. */ 629 static LONGEST 630 min_of_size (int size) 631 { 632 return -max_of_size (size) - 1; 633 } 634 635 /* Maximum value of a SIZE-byte unsigned integer type. */ 636 static ULONGEST 637 umax_of_size (int size) 638 { 639 ULONGEST top_bit = (ULONGEST) 1 << (size * 8 - 1); 640 641 return top_bit | (top_bit - 1); 642 } 643 644 /* Maximum value of integral type T, as a signed quantity. */ 645 static LONGEST 646 max_of_type (struct type *t) 647 { 648 if (t->is_unsigned ()) 649 return (LONGEST) umax_of_size (t->length ()); 650 else 651 return max_of_size (t->length ()); 652 } 653 654 /* Minimum value of integral type T, as a signed quantity. */ 655 static LONGEST 656 min_of_type (struct type *t) 657 { 658 if (t->is_unsigned ()) 659 return 0; 660 else 661 return min_of_size (t->length ()); 662 } 663 664 /* The largest value in the domain of TYPE, a discrete type, as an integer. */ 665 LONGEST 666 ada_discrete_type_high_bound (struct type *type) 667 { 668 type = resolve_dynamic_type (type, {}, 0); 669 switch (type->code ()) 670 { 671 case TYPE_CODE_RANGE: 672 { 673 const dynamic_prop &high = type->bounds ()->high; 674 675 if (high.kind () == PROP_CONST) 676 return high.const_val (); 677 else 678 { 679 gdb_assert (high.kind () == PROP_UNDEFINED); 680 681 /* This happens when trying to evaluate a type's dynamic bound 682 without a live target. There is nothing relevant for us to 683 return here, so return 0. */ 684 return 0; 685 } 686 } 687 case TYPE_CODE_ENUM: 688 return type->field (type->num_fields () - 1).loc_enumval (); 689 case TYPE_CODE_BOOL: 690 return 1; 691 case TYPE_CODE_CHAR: 692 case TYPE_CODE_INT: 693 return max_of_type (type); 694 default: 695 error (_("Unexpected type in ada_discrete_type_high_bound.")); 696 } 697 } 698 699 /* The smallest value in the domain of TYPE, a discrete type, as an integer. */ 700 LONGEST 701 ada_discrete_type_low_bound (struct type *type) 702 { 703 type = resolve_dynamic_type (type, {}, 0); 704 switch (type->code ()) 705 { 706 case TYPE_CODE_RANGE: 707 { 708 const dynamic_prop &low = type->bounds ()->low; 709 710 if (low.kind () == PROP_CONST) 711 return low.const_val (); 712 else 713 { 714 gdb_assert (low.kind () == PROP_UNDEFINED); 715 716 /* This happens when trying to evaluate a type's dynamic bound 717 without a live target. There is nothing relevant for us to 718 return here, so return 0. */ 719 return 0; 720 } 721 } 722 case TYPE_CODE_ENUM: 723 return type->field (0).loc_enumval (); 724 case TYPE_CODE_BOOL: 725 return 0; 726 case TYPE_CODE_CHAR: 727 case TYPE_CODE_INT: 728 return min_of_type (type); 729 default: 730 error (_("Unexpected type in ada_discrete_type_low_bound.")); 731 } 732 } 733 734 /* The identity on non-range types. For range types, the underlying 735 non-range scalar type. */ 736 737 static struct type * 738 get_base_type (struct type *type) 739 { 740 while (type != NULL && type->code () == TYPE_CODE_RANGE) 741 { 742 if (type == type->target_type () || type->target_type () == NULL) 743 return type; 744 type = type->target_type (); 745 } 746 return type; 747 } 748 749 /* Return a decoded version of the given VALUE. This means returning 750 a value whose type is obtained by applying all the GNAT-specific 751 encodings, making the resulting type a static but standard description 752 of the initial type. */ 753 754 struct value * 755 ada_get_decoded_value (struct value *value) 756 { 757 struct type *type = ada_check_typedef (value_type (value)); 758 759 if (ada_is_array_descriptor_type (type) 760 || (ada_is_constrained_packed_array_type (type) 761 && type->code () != TYPE_CODE_PTR)) 762 { 763 if (type->code () == TYPE_CODE_TYPEDEF) /* array access type. */ 764 value = ada_coerce_to_simple_array_ptr (value); 765 else 766 value = ada_coerce_to_simple_array (value); 767 } 768 else 769 value = ada_to_fixed_value (value); 770 771 return value; 772 } 773 774 /* Same as ada_get_decoded_value, but with the given TYPE. 775 Because there is no associated actual value for this type, 776 the resulting type might be a best-effort approximation in 777 the case of dynamic types. */ 778 779 struct type * 780 ada_get_decoded_type (struct type *type) 781 { 782 type = to_static_fixed_type (type); 783 if (ada_is_constrained_packed_array_type (type)) 784 type = ada_coerce_to_simple_array_type (type); 785 return type; 786 } 787 788 789 790 /* Language Selection */ 791 792 /* If the main program is in Ada, return language_ada, otherwise return LANG 793 (the main program is in Ada iif the adainit symbol is found). */ 794 795 static enum language 796 ada_update_initial_language (enum language lang) 797 { 798 if (lookup_minimal_symbol ("adainit", NULL, NULL).minsym != NULL) 799 return language_ada; 800 801 return lang; 802 } 803 804 /* If the main procedure is written in Ada, then return its name. 805 The result is good until the next call. Return NULL if the main 806 procedure doesn't appear to be in Ada. */ 807 808 char * 809 ada_main_name (void) 810 { 811 struct bound_minimal_symbol msym; 812 static gdb::unique_xmalloc_ptr<char> main_program_name; 813 814 /* For Ada, the name of the main procedure is stored in a specific 815 string constant, generated by the binder. Look for that symbol, 816 extract its address, and then read that string. If we didn't find 817 that string, then most probably the main procedure is not written 818 in Ada. */ 819 msym = lookup_minimal_symbol (ADA_MAIN_PROGRAM_SYMBOL_NAME, NULL, NULL); 820 821 if (msym.minsym != NULL) 822 { 823 CORE_ADDR main_program_name_addr = msym.value_address (); 824 if (main_program_name_addr == 0) 825 error (_("Invalid address for Ada main program name.")); 826 827 main_program_name = target_read_string (main_program_name_addr, 1024); 828 return main_program_name.get (); 829 } 830 831 /* The main procedure doesn't seem to be in Ada. */ 832 return NULL; 833 } 834 835 /* Symbols */ 836 837 /* Table of Ada operators and their GNAT-encoded names. Last entry is pair 838 of NULLs. */ 839 840 const struct ada_opname_map ada_opname_table[] = { 841 {"Oadd", "\"+\"", BINOP_ADD}, 842 {"Osubtract", "\"-\"", BINOP_SUB}, 843 {"Omultiply", "\"*\"", BINOP_MUL}, 844 {"Odivide", "\"/\"", BINOP_DIV}, 845 {"Omod", "\"mod\"", BINOP_MOD}, 846 {"Orem", "\"rem\"", BINOP_REM}, 847 {"Oexpon", "\"**\"", BINOP_EXP}, 848 {"Olt", "\"<\"", BINOP_LESS}, 849 {"Ole", "\"<=\"", BINOP_LEQ}, 850 {"Ogt", "\">\"", BINOP_GTR}, 851 {"Oge", "\">=\"", BINOP_GEQ}, 852 {"Oeq", "\"=\"", BINOP_EQUAL}, 853 {"One", "\"/=\"", BINOP_NOTEQUAL}, 854 {"Oand", "\"and\"", BINOP_BITWISE_AND}, 855 {"Oor", "\"or\"", BINOP_BITWISE_IOR}, 856 {"Oxor", "\"xor\"", BINOP_BITWISE_XOR}, 857 {"Oconcat", "\"&\"", BINOP_CONCAT}, 858 {"Oabs", "\"abs\"", UNOP_ABS}, 859 {"Onot", "\"not\"", UNOP_LOGICAL_NOT}, 860 {"Oadd", "\"+\"", UNOP_PLUS}, 861 {"Osubtract", "\"-\"", UNOP_NEG}, 862 {NULL, NULL} 863 }; 864 865 /* If STR is a decoded version of a compiler-provided suffix (like the 866 "[cold]" in "symbol[cold]"), return true. Otherwise, return 867 false. */ 868 869 static bool 870 is_compiler_suffix (const char *str) 871 { 872 gdb_assert (*str == '['); 873 ++str; 874 while (*str != '\0' && isalpha (*str)) 875 ++str; 876 /* We accept a missing "]" in order to support completion. */ 877 return *str == '\0' || (str[0] == ']' && str[1] == '\0'); 878 } 879 880 /* Append a non-ASCII character to RESULT. */ 881 static void 882 append_hex_encoded (std::string &result, uint32_t one_char) 883 { 884 if (one_char <= 0xff) 885 { 886 result.append ("U"); 887 result.append (phex (one_char, 1)); 888 } 889 else if (one_char <= 0xffff) 890 { 891 result.append ("W"); 892 result.append (phex (one_char, 2)); 893 } 894 else 895 { 896 result.append ("WW"); 897 result.append (phex (one_char, 4)); 898 } 899 } 900 901 /* Return a string that is a copy of the data in STORAGE, with 902 non-ASCII characters replaced by the appropriate hex encoding. A 903 template is used because, for UTF-8, we actually want to work with 904 UTF-32 codepoints. */ 905 template<typename T> 906 std::string 907 copy_and_hex_encode (struct obstack *storage) 908 { 909 const T *chars = (T *) obstack_base (storage); 910 int num_chars = obstack_object_size (storage) / sizeof (T); 911 std::string result; 912 for (int i = 0; i < num_chars; ++i) 913 { 914 if (chars[i] <= 0x7f) 915 { 916 /* The host character set has to be a superset of ASCII, as 917 are all the other character sets we can use. */ 918 result.push_back (chars[i]); 919 } 920 else 921 append_hex_encoded (result, chars[i]); 922 } 923 return result; 924 } 925 926 /* The "encoded" form of DECODED, according to GNAT conventions. If 927 THROW_ERRORS, throw an error if invalid operator name is found. 928 Otherwise, return the empty string in that case. */ 929 930 static std::string 931 ada_encode_1 (const char *decoded, bool throw_errors) 932 { 933 if (decoded == NULL) 934 return {}; 935 936 std::string encoding_buffer; 937 bool saw_non_ascii = false; 938 for (const char *p = decoded; *p != '\0'; p += 1) 939 { 940 if ((*p & 0x80) != 0) 941 saw_non_ascii = true; 942 943 if (*p == '.') 944 encoding_buffer.append ("__"); 945 else if (*p == '[' && is_compiler_suffix (p)) 946 { 947 encoding_buffer = encoding_buffer + "." + (p + 1); 948 if (encoding_buffer.back () == ']') 949 encoding_buffer.pop_back (); 950 break; 951 } 952 else if (*p == '"') 953 { 954 const struct ada_opname_map *mapping; 955 956 for (mapping = ada_opname_table; 957 mapping->encoded != NULL 958 && !startswith (p, mapping->decoded); mapping += 1) 959 ; 960 if (mapping->encoded == NULL) 961 { 962 if (throw_errors) 963 error (_("invalid Ada operator name: %s"), p); 964 else 965 return {}; 966 } 967 encoding_buffer.append (mapping->encoded); 968 break; 969 } 970 else 971 encoding_buffer.push_back (*p); 972 } 973 974 /* If a non-ASCII character is seen, we must convert it to the 975 appropriate hex form. As this is more expensive, we keep track 976 of whether it is even necessary. */ 977 if (saw_non_ascii) 978 { 979 auto_obstack storage; 980 bool is_utf8 = ada_source_charset == ada_utf8; 981 try 982 { 983 convert_between_encodings 984 (host_charset (), 985 is_utf8 ? HOST_UTF32 : ada_source_charset, 986 (const gdb_byte *) encoding_buffer.c_str (), 987 encoding_buffer.length (), 1, 988 &storage, translit_none); 989 } 990 catch (const gdb_exception &) 991 { 992 static bool warned = false; 993 994 /* Converting to UTF-32 shouldn't fail, so if it doesn't, we 995 might like to know why. */ 996 if (!warned) 997 { 998 warned = true; 999 warning (_("charset conversion failure for '%s'.\n" 1000 "You may have the wrong value for 'set ada source-charset'."), 1001 encoding_buffer.c_str ()); 1002 } 1003 1004 /* We don't try to recover from errors. */ 1005 return encoding_buffer; 1006 } 1007 1008 if (is_utf8) 1009 return copy_and_hex_encode<uint32_t> (&storage); 1010 return copy_and_hex_encode<gdb_byte> (&storage); 1011 } 1012 1013 return encoding_buffer; 1014 } 1015 1016 /* Find the entry for C in the case-folding table. Return nullptr if 1017 the entry does not cover C. */ 1018 static const utf8_entry * 1019 find_case_fold_entry (uint32_t c) 1020 { 1021 auto iter = std::lower_bound (std::begin (ada_case_fold), 1022 std::end (ada_case_fold), 1023 c); 1024 if (iter == std::end (ada_case_fold) 1025 || c < iter->start 1026 || c > iter->end) 1027 return nullptr; 1028 return &*iter; 1029 } 1030 1031 /* Return NAME folded to lower case, or, if surrounded by single 1032 quotes, unfolded, but with the quotes stripped away. If 1033 THROW_ON_ERROR is true, encoding failures will throw an exception 1034 rather than emitting a warning. Result good to next call. */ 1035 1036 static const char * 1037 ada_fold_name (gdb::string_view name, bool throw_on_error = false) 1038 { 1039 static std::string fold_storage; 1040 1041 if (!name.empty () && name[0] == '\'') 1042 fold_storage = gdb::to_string (name.substr (1, name.size () - 2)); 1043 else 1044 { 1045 /* Why convert to UTF-32 and implement our own case-folding, 1046 rather than convert to wchar_t and use the platform's 1047 functions? I'm glad you asked. 1048 1049 The main problem is that GNAT implements an unusual rule for 1050 case folding. For ASCII letters, letters in single-byte 1051 encodings (such as ISO-8859-*), and Unicode letters that fit 1052 in a single byte (i.e., code point is <= 0xff), the letter is 1053 folded to lower case. Other Unicode letters are folded to 1054 upper case. 1055 1056 This rule means that the code must be able to examine the 1057 value of the character. And, some hosts do not use Unicode 1058 for wchar_t, so examining the value of such characters is 1059 forbidden. */ 1060 auto_obstack storage; 1061 try 1062 { 1063 convert_between_encodings 1064 (host_charset (), HOST_UTF32, 1065 (const gdb_byte *) name.data (), 1066 name.length (), 1, 1067 &storage, translit_none); 1068 } 1069 catch (const gdb_exception &) 1070 { 1071 if (throw_on_error) 1072 throw; 1073 1074 static bool warned = false; 1075 1076 /* Converting to UTF-32 shouldn't fail, so if it doesn't, we 1077 might like to know why. */ 1078 if (!warned) 1079 { 1080 warned = true; 1081 warning (_("could not convert '%s' from the host encoding (%s) to UTF-32.\n" 1082 "This normally should not happen, please file a bug report."), 1083 gdb::to_string (name).c_str (), host_charset ()); 1084 } 1085 1086 /* We don't try to recover from errors; just return the 1087 original string. */ 1088 fold_storage = gdb::to_string (name); 1089 return fold_storage.c_str (); 1090 } 1091 1092 bool is_utf8 = ada_source_charset == ada_utf8; 1093 uint32_t *chars = (uint32_t *) obstack_base (&storage); 1094 int num_chars = obstack_object_size (&storage) / sizeof (uint32_t); 1095 for (int i = 0; i < num_chars; ++i) 1096 { 1097 const struct utf8_entry *entry = find_case_fold_entry (chars[i]); 1098 if (entry != nullptr) 1099 { 1100 uint32_t low = chars[i] + entry->lower_delta; 1101 if (!is_utf8 || low <= 0xff) 1102 chars[i] = low; 1103 else 1104 chars[i] = chars[i] + entry->upper_delta; 1105 } 1106 } 1107 1108 /* Now convert back to ordinary characters. */ 1109 auto_obstack reconverted; 1110 try 1111 { 1112 convert_between_encodings (HOST_UTF32, 1113 host_charset (), 1114 (const gdb_byte *) chars, 1115 num_chars * sizeof (uint32_t), 1116 sizeof (uint32_t), 1117 &reconverted, 1118 translit_none); 1119 obstack_1grow (&reconverted, '\0'); 1120 fold_storage = std::string ((const char *) obstack_base (&reconverted)); 1121 } 1122 catch (const gdb_exception &) 1123 { 1124 if (throw_on_error) 1125 throw; 1126 1127 static bool warned = false; 1128 1129 /* Converting back from UTF-32 shouldn't normally fail, but 1130 there are some host encodings without upper/lower 1131 equivalence. */ 1132 if (!warned) 1133 { 1134 warned = true; 1135 warning (_("could not convert the lower-cased variant of '%s'\n" 1136 "from UTF-32 to the host encoding (%s)."), 1137 gdb::to_string (name).c_str (), host_charset ()); 1138 } 1139 1140 /* We don't try to recover from errors; just return the 1141 original string. */ 1142 fold_storage = gdb::to_string (name); 1143 } 1144 } 1145 1146 return fold_storage.c_str (); 1147 } 1148 1149 /* The "encoded" form of DECODED, according to GNAT conventions. If 1150 FOLD is true (the default), case-fold any ordinary symbol. Symbols 1151 with <...> quoting are not folded in any case. */ 1152 1153 std::string 1154 ada_encode (const char *decoded, bool fold) 1155 { 1156 if (fold && decoded[0] != '<') 1157 decoded = ada_fold_name (decoded); 1158 return ada_encode_1 (decoded, true); 1159 } 1160 1161 /* Return nonzero if C is either a digit or a lowercase alphabet character. */ 1162 1163 static int 1164 is_lower_alphanum (const char c) 1165 { 1166 return (isdigit (c) || (isalpha (c) && islower (c))); 1167 } 1168 1169 /* ENCODED is the linkage name of a symbol and LEN contains its length. 1170 This function saves in LEN the length of that same symbol name but 1171 without either of these suffixes: 1172 . .{DIGIT}+ 1173 . ${DIGIT}+ 1174 . ___{DIGIT}+ 1175 . __{DIGIT}+. 1176 1177 These are suffixes introduced by the compiler for entities such as 1178 nested subprogram for instance, in order to avoid name clashes. 1179 They do not serve any purpose for the debugger. */ 1180 1181 static void 1182 ada_remove_trailing_digits (const char *encoded, int *len) 1183 { 1184 if (*len > 1 && isdigit (encoded[*len - 1])) 1185 { 1186 int i = *len - 2; 1187 1188 while (i > 0 && isdigit (encoded[i])) 1189 i--; 1190 if (i >= 0 && encoded[i] == '.') 1191 *len = i; 1192 else if (i >= 0 && encoded[i] == '$') 1193 *len = i; 1194 else if (i >= 2 && startswith (encoded + i - 2, "___")) 1195 *len = i - 2; 1196 else if (i >= 1 && startswith (encoded + i - 1, "__")) 1197 *len = i - 1; 1198 } 1199 } 1200 1201 /* Remove the suffix introduced by the compiler for protected object 1202 subprograms. */ 1203 1204 static void 1205 ada_remove_po_subprogram_suffix (const char *encoded, int *len) 1206 { 1207 /* Remove trailing N. */ 1208 1209 /* Protected entry subprograms are broken into two 1210 separate subprograms: The first one is unprotected, and has 1211 a 'N' suffix; the second is the protected version, and has 1212 the 'P' suffix. The second calls the first one after handling 1213 the protection. Since the P subprograms are internally generated, 1214 we leave these names undecoded, giving the user a clue that this 1215 entity is internal. */ 1216 1217 if (*len > 1 1218 && encoded[*len - 1] == 'N' 1219 && (isdigit (encoded[*len - 2]) || islower (encoded[*len - 2]))) 1220 *len = *len - 1; 1221 } 1222 1223 /* If ENCODED ends with a compiler-provided suffix (like ".cold"), 1224 then update *LEN to remove the suffix and return the offset of the 1225 character just past the ".". Otherwise, return -1. */ 1226 1227 static int 1228 remove_compiler_suffix (const char *encoded, int *len) 1229 { 1230 int offset = *len - 1; 1231 while (offset > 0 && isalpha (encoded[offset])) 1232 --offset; 1233 if (offset > 0 && encoded[offset] == '.') 1234 { 1235 *len = offset; 1236 return offset + 1; 1237 } 1238 return -1; 1239 } 1240 1241 /* Convert an ASCII hex string to a number. Reads exactly N 1242 characters from STR. Returns true on success, false if one of the 1243 digits was not a hex digit. */ 1244 static bool 1245 convert_hex (const char *str, int n, uint32_t *out) 1246 { 1247 uint32_t result = 0; 1248 1249 for (int i = 0; i < n; ++i) 1250 { 1251 if (!isxdigit (str[i])) 1252 return false; 1253 result <<= 4; 1254 result |= fromhex (str[i]); 1255 } 1256 1257 *out = result; 1258 return true; 1259 } 1260 1261 /* Convert a wide character from its ASCII hex representation in STR 1262 (consisting of exactly N characters) to the host encoding, 1263 appending the resulting bytes to OUT. If N==2 and the Ada source 1264 charset is not UTF-8, then hex refers to an encoding in the 1265 ADA_SOURCE_CHARSET; otherwise, use UTF-32. Return true on success. 1266 Return false and do not modify OUT on conversion failure. */ 1267 static bool 1268 convert_from_hex_encoded (std::string &out, const char *str, int n) 1269 { 1270 uint32_t value; 1271 1272 if (!convert_hex (str, n, &value)) 1273 return false; 1274 try 1275 { 1276 auto_obstack bytes; 1277 /* In the 'U' case, the hex digits encode the character in the 1278 Ada source charset. However, if the source charset is UTF-8, 1279 this really means it is a single-byte UTF-32 character. */ 1280 if (n == 2 && ada_source_charset != ada_utf8) 1281 { 1282 gdb_byte one_char = (gdb_byte) value; 1283 1284 convert_between_encodings (ada_source_charset, host_charset (), 1285 &one_char, 1286 sizeof (one_char), sizeof (one_char), 1287 &bytes, translit_none); 1288 } 1289 else 1290 convert_between_encodings (HOST_UTF32, host_charset (), 1291 (const gdb_byte *) &value, 1292 sizeof (value), sizeof (value), 1293 &bytes, translit_none); 1294 obstack_1grow (&bytes, '\0'); 1295 out.append ((const char *) obstack_base (&bytes)); 1296 } 1297 catch (const gdb_exception &) 1298 { 1299 /* On failure, the caller will just let the encoded form 1300 through, which seems basically reasonable. */ 1301 return false; 1302 } 1303 1304 return true; 1305 } 1306 1307 /* See ada-lang.h. */ 1308 1309 std::string 1310 ada_decode (const char *encoded, bool wrap, bool operators) 1311 { 1312 int i; 1313 int len0; 1314 const char *p; 1315 int at_start_name; 1316 std::string decoded; 1317 int suffix = -1; 1318 1319 /* With function descriptors on PPC64, the value of a symbol named 1320 ".FN", if it exists, is the entry point of the function "FN". */ 1321 if (encoded[0] == '.') 1322 encoded += 1; 1323 1324 /* The name of the Ada main procedure starts with "_ada_". 1325 This prefix is not part of the decoded name, so skip this part 1326 if we see this prefix. */ 1327 if (startswith (encoded, "_ada_")) 1328 encoded += 5; 1329 /* The "___ghost_" prefix is used for ghost entities. Normally 1330 these aren't preserved but when they are, it's useful to see 1331 them. */ 1332 if (startswith (encoded, "___ghost_")) 1333 encoded += 9; 1334 1335 /* If the name starts with '_', then it is not a properly encoded 1336 name, so do not attempt to decode it. Similarly, if the name 1337 starts with '<', the name should not be decoded. */ 1338 if (encoded[0] == '_' || encoded[0] == '<') 1339 goto Suppress; 1340 1341 len0 = strlen (encoded); 1342 1343 suffix = remove_compiler_suffix (encoded, &len0); 1344 1345 ada_remove_trailing_digits (encoded, &len0); 1346 ada_remove_po_subprogram_suffix (encoded, &len0); 1347 1348 /* Remove the ___X.* suffix if present. Do not forget to verify that 1349 the suffix is located before the current "end" of ENCODED. We want 1350 to avoid re-matching parts of ENCODED that have previously been 1351 marked as discarded (by decrementing LEN0). */ 1352 p = strstr (encoded, "___"); 1353 if (p != NULL && p - encoded < len0 - 3) 1354 { 1355 if (p[3] == 'X') 1356 len0 = p - encoded; 1357 else 1358 goto Suppress; 1359 } 1360 1361 /* Remove any trailing TKB suffix. It tells us that this symbol 1362 is for the body of a task, but that information does not actually 1363 appear in the decoded name. */ 1364 1365 if (len0 > 3 && startswith (encoded + len0 - 3, "TKB")) 1366 len0 -= 3; 1367 1368 /* Remove any trailing TB suffix. The TB suffix is slightly different 1369 from the TKB suffix because it is used for non-anonymous task 1370 bodies. */ 1371 1372 if (len0 > 2 && startswith (encoded + len0 - 2, "TB")) 1373 len0 -= 2; 1374 1375 /* Remove trailing "B" suffixes. */ 1376 /* FIXME: brobecker/2006-04-19: Not sure what this are used for... */ 1377 1378 if (len0 > 1 && startswith (encoded + len0 - 1, "B")) 1379 len0 -= 1; 1380 1381 /* Remove trailing __{digit}+ or trailing ${digit}+. */ 1382 1383 if (len0 > 1 && isdigit (encoded[len0 - 1])) 1384 { 1385 i = len0 - 2; 1386 while ((i >= 0 && isdigit (encoded[i])) 1387 || (i >= 1 && encoded[i] == '_' && isdigit (encoded[i - 1]))) 1388 i -= 1; 1389 if (i > 1 && encoded[i] == '_' && encoded[i - 1] == '_') 1390 len0 = i - 1; 1391 else if (encoded[i] == '$') 1392 len0 = i; 1393 } 1394 1395 /* The first few characters that are not alphabetic are not part 1396 of any encoding we use, so we can copy them over verbatim. */ 1397 1398 for (i = 0; i < len0 && !isalpha (encoded[i]); i += 1) 1399 decoded.push_back (encoded[i]); 1400 1401 at_start_name = 1; 1402 while (i < len0) 1403 { 1404 /* Is this a symbol function? */ 1405 if (operators && at_start_name && encoded[i] == 'O') 1406 { 1407 int k; 1408 1409 for (k = 0; ada_opname_table[k].encoded != NULL; k += 1) 1410 { 1411 int op_len = strlen (ada_opname_table[k].encoded); 1412 if ((strncmp (ada_opname_table[k].encoded + 1, encoded + i + 1, 1413 op_len - 1) == 0) 1414 && !isalnum (encoded[i + op_len])) 1415 { 1416 decoded.append (ada_opname_table[k].decoded); 1417 at_start_name = 0; 1418 i += op_len; 1419 break; 1420 } 1421 } 1422 if (ada_opname_table[k].encoded != NULL) 1423 continue; 1424 } 1425 at_start_name = 0; 1426 1427 /* Replace "TK__" with "__", which will eventually be translated 1428 into "." (just below). */ 1429 1430 if (i < len0 - 4 && startswith (encoded + i, "TK__")) 1431 i += 2; 1432 1433 /* Replace "__B_{DIGITS}+__" sequences by "__", which will eventually 1434 be translated into "." (just below). These are internal names 1435 generated for anonymous blocks inside which our symbol is nested. */ 1436 1437 if (len0 - i > 5 && encoded [i] == '_' && encoded [i+1] == '_' 1438 && encoded [i+2] == 'B' && encoded [i+3] == '_' 1439 && isdigit (encoded [i+4])) 1440 { 1441 int k = i + 5; 1442 1443 while (k < len0 && isdigit (encoded[k])) 1444 k++; /* Skip any extra digit. */ 1445 1446 /* Double-check that the "__B_{DIGITS}+" sequence we found 1447 is indeed followed by "__". */ 1448 if (len0 - k > 2 && encoded [k] == '_' && encoded [k+1] == '_') 1449 i = k; 1450 } 1451 1452 /* Remove _E{DIGITS}+[sb] */ 1453 1454 /* Just as for protected object subprograms, there are 2 categories 1455 of subprograms created by the compiler for each entry. The first 1456 one implements the actual entry code, and has a suffix following 1457 the convention above; the second one implements the barrier and 1458 uses the same convention as above, except that the 'E' is replaced 1459 by a 'B'. 1460 1461 Just as above, we do not decode the name of barrier functions 1462 to give the user a clue that the code he is debugging has been 1463 internally generated. */ 1464 1465 if (len0 - i > 3 && encoded [i] == '_' && encoded[i+1] == 'E' 1466 && isdigit (encoded[i+2])) 1467 { 1468 int k = i + 3; 1469 1470 while (k < len0 && isdigit (encoded[k])) 1471 k++; 1472 1473 if (k < len0 1474 && (encoded[k] == 'b' || encoded[k] == 's')) 1475 { 1476 k++; 1477 /* Just as an extra precaution, make sure that if this 1478 suffix is followed by anything else, it is a '_'. 1479 Otherwise, we matched this sequence by accident. */ 1480 if (k == len0 1481 || (k < len0 && encoded[k] == '_')) 1482 i = k; 1483 } 1484 } 1485 1486 /* Remove trailing "N" in [a-z0-9]+N__. The N is added by 1487 the GNAT front-end in protected object subprograms. */ 1488 1489 if (i < len0 + 3 1490 && encoded[i] == 'N' && encoded[i+1] == '_' && encoded[i+2] == '_') 1491 { 1492 /* Backtrack a bit up until we reach either the begining of 1493 the encoded name, or "__". Make sure that we only find 1494 digits or lowercase characters. */ 1495 const char *ptr = encoded + i - 1; 1496 1497 while (ptr >= encoded && is_lower_alphanum (ptr[0])) 1498 ptr--; 1499 if (ptr < encoded 1500 || (ptr > encoded && ptr[0] == '_' && ptr[-1] == '_')) 1501 i++; 1502 } 1503 1504 if (i < len0 + 3 && encoded[i] == 'U' && isxdigit (encoded[i + 1])) 1505 { 1506 if (convert_from_hex_encoded (decoded, &encoded[i + 1], 2)) 1507 { 1508 i += 3; 1509 continue; 1510 } 1511 } 1512 else if (i < len0 + 5 && encoded[i] == 'W' && isxdigit (encoded[i + 1])) 1513 { 1514 if (convert_from_hex_encoded (decoded, &encoded[i + 1], 4)) 1515 { 1516 i += 5; 1517 continue; 1518 } 1519 } 1520 else if (i < len0 + 10 && encoded[i] == 'W' && encoded[i + 1] == 'W' 1521 && isxdigit (encoded[i + 2])) 1522 { 1523 if (convert_from_hex_encoded (decoded, &encoded[i + 2], 8)) 1524 { 1525 i += 10; 1526 continue; 1527 } 1528 } 1529 1530 if (encoded[i] == 'X' && i != 0 && isalnum (encoded[i - 1])) 1531 { 1532 /* This is a X[bn]* sequence not separated from the previous 1533 part of the name with a non-alpha-numeric character (in other 1534 words, immediately following an alpha-numeric character), then 1535 verify that it is placed at the end of the encoded name. If 1536 not, then the encoding is not valid and we should abort the 1537 decoding. Otherwise, just skip it, it is used in body-nested 1538 package names. */ 1539 do 1540 i += 1; 1541 while (i < len0 && (encoded[i] == 'b' || encoded[i] == 'n')); 1542 if (i < len0) 1543 goto Suppress; 1544 } 1545 else if (i < len0 - 2 && encoded[i] == '_' && encoded[i + 1] == '_') 1546 { 1547 /* Replace '__' by '.'. */ 1548 decoded.push_back ('.'); 1549 at_start_name = 1; 1550 i += 2; 1551 } 1552 else 1553 { 1554 /* It's a character part of the decoded name, so just copy it 1555 over. */ 1556 decoded.push_back (encoded[i]); 1557 i += 1; 1558 } 1559 } 1560 1561 /* Decoded names should never contain any uppercase character. 1562 Double-check this, and abort the decoding if we find one. */ 1563 1564 if (operators) 1565 { 1566 for (i = 0; i < decoded.length(); ++i) 1567 if (isupper (decoded[i]) || decoded[i] == ' ') 1568 goto Suppress; 1569 } 1570 1571 /* If the compiler added a suffix, append it now. */ 1572 if (suffix >= 0) 1573 decoded = decoded + "[" + &encoded[suffix] + "]"; 1574 1575 return decoded; 1576 1577 Suppress: 1578 if (!wrap) 1579 return {}; 1580 1581 if (encoded[0] == '<') 1582 decoded = encoded; 1583 else 1584 decoded = '<' + std::string(encoded) + '>'; 1585 return decoded; 1586 } 1587 1588 /* Table for keeping permanent unique copies of decoded names. Once 1589 allocated, names in this table are never released. While this is a 1590 storage leak, it should not be significant unless there are massive 1591 changes in the set of decoded names in successive versions of a 1592 symbol table loaded during a single session. */ 1593 static struct htab *decoded_names_store; 1594 1595 /* Returns the decoded name of GSYMBOL, as for ada_decode, caching it 1596 in the language-specific part of GSYMBOL, if it has not been 1597 previously computed. Tries to save the decoded name in the same 1598 obstack as GSYMBOL, if possible, and otherwise on the heap (so that, 1599 in any case, the decoded symbol has a lifetime at least that of 1600 GSYMBOL). 1601 The GSYMBOL parameter is "mutable" in the C++ sense: logically 1602 const, but nevertheless modified to a semantically equivalent form 1603 when a decoded name is cached in it. */ 1604 1605 const char * 1606 ada_decode_symbol (const struct general_symbol_info *arg) 1607 { 1608 struct general_symbol_info *gsymbol = (struct general_symbol_info *) arg; 1609 const char **resultp = 1610 &gsymbol->language_specific.demangled_name; 1611 1612 if (!gsymbol->ada_mangled) 1613 { 1614 std::string decoded = ada_decode (gsymbol->linkage_name ()); 1615 struct obstack *obstack = gsymbol->language_specific.obstack; 1616 1617 gsymbol->ada_mangled = 1; 1618 1619 if (obstack != NULL) 1620 *resultp = obstack_strdup (obstack, decoded.c_str ()); 1621 else 1622 { 1623 /* Sometimes, we can't find a corresponding objfile, in 1624 which case, we put the result on the heap. Since we only 1625 decode when needed, we hope this usually does not cause a 1626 significant memory leak (FIXME). */ 1627 1628 char **slot = (char **) htab_find_slot (decoded_names_store, 1629 decoded.c_str (), INSERT); 1630 1631 if (*slot == NULL) 1632 *slot = xstrdup (decoded.c_str ()); 1633 *resultp = *slot; 1634 } 1635 } 1636 1637 return *resultp; 1638 } 1639 1640 1641 1642 /* Arrays */ 1643 1644 /* Assuming that INDEX_DESC_TYPE is an ___XA structure, a structure 1645 generated by the GNAT compiler to describe the index type used 1646 for each dimension of an array, check whether it follows the latest 1647 known encoding. If not, fix it up to conform to the latest encoding. 1648 Otherwise, do nothing. This function also does nothing if 1649 INDEX_DESC_TYPE is NULL. 1650 1651 The GNAT encoding used to describe the array index type evolved a bit. 1652 Initially, the information would be provided through the name of each 1653 field of the structure type only, while the type of these fields was 1654 described as unspecified and irrelevant. The debugger was then expected 1655 to perform a global type lookup using the name of that field in order 1656 to get access to the full index type description. Because these global 1657 lookups can be very expensive, the encoding was later enhanced to make 1658 the global lookup unnecessary by defining the field type as being 1659 the full index type description. 1660 1661 The purpose of this routine is to allow us to support older versions 1662 of the compiler by detecting the use of the older encoding, and by 1663 fixing up the INDEX_DESC_TYPE to follow the new one (at this point, 1664 we essentially replace each field's meaningless type by the associated 1665 index subtype). */ 1666 1667 void 1668 ada_fixup_array_indexes_type (struct type *index_desc_type) 1669 { 1670 int i; 1671 1672 if (index_desc_type == NULL) 1673 return; 1674 gdb_assert (index_desc_type->num_fields () > 0); 1675 1676 /* Check if INDEX_DESC_TYPE follows the older encoding (it is sufficient 1677 to check one field only, no need to check them all). If not, return 1678 now. 1679 1680 If our INDEX_DESC_TYPE was generated using the older encoding, 1681 the field type should be a meaningless integer type whose name 1682 is not equal to the field name. */ 1683 if (index_desc_type->field (0).type ()->name () != NULL 1684 && strcmp (index_desc_type->field (0).type ()->name (), 1685 index_desc_type->field (0).name ()) == 0) 1686 return; 1687 1688 /* Fixup each field of INDEX_DESC_TYPE. */ 1689 for (i = 0; i < index_desc_type->num_fields (); i++) 1690 { 1691 const char *name = index_desc_type->field (i).name (); 1692 struct type *raw_type = ada_check_typedef (ada_find_any_type (name)); 1693 1694 if (raw_type) 1695 index_desc_type->field (i).set_type (raw_type); 1696 } 1697 } 1698 1699 /* The desc_* routines return primitive portions of array descriptors 1700 (fat pointers). */ 1701 1702 /* The descriptor or array type, if any, indicated by TYPE; removes 1703 level of indirection, if needed. */ 1704 1705 static struct type * 1706 desc_base_type (struct type *type) 1707 { 1708 if (type == NULL) 1709 return NULL; 1710 type = ada_check_typedef (type); 1711 if (type->code () == TYPE_CODE_TYPEDEF) 1712 type = ada_typedef_target_type (type); 1713 1714 if (type != NULL 1715 && (type->code () == TYPE_CODE_PTR 1716 || type->code () == TYPE_CODE_REF)) 1717 return ada_check_typedef (type->target_type ()); 1718 else 1719 return type; 1720 } 1721 1722 /* True iff TYPE indicates a "thin" array pointer type. */ 1723 1724 static int 1725 is_thin_pntr (struct type *type) 1726 { 1727 return 1728 is_suffix (ada_type_name (desc_base_type (type)), "___XUT") 1729 || is_suffix (ada_type_name (desc_base_type (type)), "___XUT___XVE"); 1730 } 1731 1732 /* The descriptor type for thin pointer type TYPE. */ 1733 1734 static struct type * 1735 thin_descriptor_type (struct type *type) 1736 { 1737 struct type *base_type = desc_base_type (type); 1738 1739 if (base_type == NULL) 1740 return NULL; 1741 if (is_suffix (ada_type_name (base_type), "___XVE")) 1742 return base_type; 1743 else 1744 { 1745 struct type *alt_type = ada_find_parallel_type (base_type, "___XVE"); 1746 1747 if (alt_type == NULL) 1748 return base_type; 1749 else 1750 return alt_type; 1751 } 1752 } 1753 1754 /* A pointer to the array data for thin-pointer value VAL. */ 1755 1756 static struct value * 1757 thin_data_pntr (struct value *val) 1758 { 1759 struct type *type = ada_check_typedef (value_type (val)); 1760 struct type *data_type = desc_data_target_type (thin_descriptor_type (type)); 1761 1762 data_type = lookup_pointer_type (data_type); 1763 1764 if (type->code () == TYPE_CODE_PTR) 1765 return value_cast (data_type, value_copy (val)); 1766 else 1767 return value_from_longest (data_type, value_address (val)); 1768 } 1769 1770 /* True iff TYPE indicates a "thick" array pointer type. */ 1771 1772 static int 1773 is_thick_pntr (struct type *type) 1774 { 1775 type = desc_base_type (type); 1776 return (type != NULL && type->code () == TYPE_CODE_STRUCT 1777 && lookup_struct_elt_type (type, "P_BOUNDS", 1) != NULL); 1778 } 1779 1780 /* If TYPE is the type of an array descriptor (fat or thin pointer) or a 1781 pointer to one, the type of its bounds data; otherwise, NULL. */ 1782 1783 static struct type * 1784 desc_bounds_type (struct type *type) 1785 { 1786 struct type *r; 1787 1788 type = desc_base_type (type); 1789 1790 if (type == NULL) 1791 return NULL; 1792 else if (is_thin_pntr (type)) 1793 { 1794 type = thin_descriptor_type (type); 1795 if (type == NULL) 1796 return NULL; 1797 r = lookup_struct_elt_type (type, "BOUNDS", 1); 1798 if (r != NULL) 1799 return ada_check_typedef (r); 1800 } 1801 else if (type->code () == TYPE_CODE_STRUCT) 1802 { 1803 r = lookup_struct_elt_type (type, "P_BOUNDS", 1); 1804 if (r != NULL) 1805 return ada_check_typedef (ada_check_typedef (r)->target_type ()); 1806 } 1807 return NULL; 1808 } 1809 1810 /* If ARR is an array descriptor (fat or thin pointer), or pointer to 1811 one, a pointer to its bounds data. Otherwise NULL. */ 1812 1813 static struct value * 1814 desc_bounds (struct value *arr) 1815 { 1816 struct type *type = ada_check_typedef (value_type (arr)); 1817 1818 if (is_thin_pntr (type)) 1819 { 1820 struct type *bounds_type = 1821 desc_bounds_type (thin_descriptor_type (type)); 1822 LONGEST addr; 1823 1824 if (bounds_type == NULL) 1825 error (_("Bad GNAT array descriptor")); 1826 1827 /* NOTE: The following calculation is not really kosher, but 1828 since desc_type is an XVE-encoded type (and shouldn't be), 1829 the correct calculation is a real pain. FIXME (and fix GCC). */ 1830 if (type->code () == TYPE_CODE_PTR) 1831 addr = value_as_long (arr); 1832 else 1833 addr = value_address (arr); 1834 1835 return 1836 value_from_longest (lookup_pointer_type (bounds_type), 1837 addr - bounds_type->length ()); 1838 } 1839 1840 else if (is_thick_pntr (type)) 1841 { 1842 struct value *p_bounds = value_struct_elt (&arr, {}, "P_BOUNDS", NULL, 1843 _("Bad GNAT array descriptor")); 1844 struct type *p_bounds_type = value_type (p_bounds); 1845 1846 if (p_bounds_type 1847 && p_bounds_type->code () == TYPE_CODE_PTR) 1848 { 1849 struct type *target_type = p_bounds_type->target_type (); 1850 1851 if (target_type->is_stub ()) 1852 p_bounds = value_cast (lookup_pointer_type 1853 (ada_check_typedef (target_type)), 1854 p_bounds); 1855 } 1856 else 1857 error (_("Bad GNAT array descriptor")); 1858 1859 return p_bounds; 1860 } 1861 else 1862 return NULL; 1863 } 1864 1865 /* If TYPE is the type of an array-descriptor (fat pointer), the bit 1866 position of the field containing the address of the bounds data. */ 1867 1868 static int 1869 fat_pntr_bounds_bitpos (struct type *type) 1870 { 1871 return desc_base_type (type)->field (1).loc_bitpos (); 1872 } 1873 1874 /* If TYPE is the type of an array-descriptor (fat pointer), the bit 1875 size of the field containing the address of the bounds data. */ 1876 1877 static int 1878 fat_pntr_bounds_bitsize (struct type *type) 1879 { 1880 type = desc_base_type (type); 1881 1882 if (TYPE_FIELD_BITSIZE (type, 1) > 0) 1883 return TYPE_FIELD_BITSIZE (type, 1); 1884 else 1885 return 8 * ada_check_typedef (type->field (1).type ())->length (); 1886 } 1887 1888 /* If TYPE is the type of an array descriptor (fat or thin pointer) or a 1889 pointer to one, the type of its array data (a array-with-no-bounds type); 1890 otherwise, NULL. Use ada_type_of_array to get an array type with bounds 1891 data. */ 1892 1893 static struct type * 1894 desc_data_target_type (struct type *type) 1895 { 1896 type = desc_base_type (type); 1897 1898 /* NOTE: The following is bogus; see comment in desc_bounds. */ 1899 if (is_thin_pntr (type)) 1900 return desc_base_type (thin_descriptor_type (type)->field (1).type ()); 1901 else if (is_thick_pntr (type)) 1902 { 1903 struct type *data_type = lookup_struct_elt_type (type, "P_ARRAY", 1); 1904 1905 if (data_type 1906 && ada_check_typedef (data_type)->code () == TYPE_CODE_PTR) 1907 return ada_check_typedef (data_type->target_type ()); 1908 } 1909 1910 return NULL; 1911 } 1912 1913 /* If ARR is an array descriptor (fat or thin pointer), a pointer to 1914 its array data. */ 1915 1916 static struct value * 1917 desc_data (struct value *arr) 1918 { 1919 struct type *type = value_type (arr); 1920 1921 if (is_thin_pntr (type)) 1922 return thin_data_pntr (arr); 1923 else if (is_thick_pntr (type)) 1924 return value_struct_elt (&arr, {}, "P_ARRAY", NULL, 1925 _("Bad GNAT array descriptor")); 1926 else 1927 return NULL; 1928 } 1929 1930 1931 /* If TYPE is the type of an array-descriptor (fat pointer), the bit 1932 position of the field containing the address of the data. */ 1933 1934 static int 1935 fat_pntr_data_bitpos (struct type *type) 1936 { 1937 return desc_base_type (type)->field (0).loc_bitpos (); 1938 } 1939 1940 /* If TYPE is the type of an array-descriptor (fat pointer), the bit 1941 size of the field containing the address of the data. */ 1942 1943 static int 1944 fat_pntr_data_bitsize (struct type *type) 1945 { 1946 type = desc_base_type (type); 1947 1948 if (TYPE_FIELD_BITSIZE (type, 0) > 0) 1949 return TYPE_FIELD_BITSIZE (type, 0); 1950 else 1951 return TARGET_CHAR_BIT * type->field (0).type ()->length (); 1952 } 1953 1954 /* If BOUNDS is an array-bounds structure (or pointer to one), return 1955 the Ith lower bound stored in it, if WHICH is 0, and the Ith upper 1956 bound, if WHICH is 1. The first bound is I=1. */ 1957 1958 static struct value * 1959 desc_one_bound (struct value *bounds, int i, int which) 1960 { 1961 char bound_name[20]; 1962 xsnprintf (bound_name, sizeof (bound_name), "%cB%d", 1963 which ? 'U' : 'L', i - 1); 1964 return value_struct_elt (&bounds, {}, bound_name, NULL, 1965 _("Bad GNAT array descriptor bounds")); 1966 } 1967 1968 /* If BOUNDS is an array-bounds structure type, return the bit position 1969 of the Ith lower bound stored in it, if WHICH is 0, and the Ith upper 1970 bound, if WHICH is 1. The first bound is I=1. */ 1971 1972 static int 1973 desc_bound_bitpos (struct type *type, int i, int which) 1974 { 1975 return desc_base_type (type)->field (2 * i + which - 2).loc_bitpos (); 1976 } 1977 1978 /* If BOUNDS is an array-bounds structure type, return the bit field size 1979 of the Ith lower bound stored in it, if WHICH is 0, and the Ith upper 1980 bound, if WHICH is 1. The first bound is I=1. */ 1981 1982 static int 1983 desc_bound_bitsize (struct type *type, int i, int which) 1984 { 1985 type = desc_base_type (type); 1986 1987 if (TYPE_FIELD_BITSIZE (type, 2 * i + which - 2) > 0) 1988 return TYPE_FIELD_BITSIZE (type, 2 * i + which - 2); 1989 else 1990 return 8 * type->field (2 * i + which - 2).type ()->length (); 1991 } 1992 1993 /* If TYPE is the type of an array-bounds structure, the type of its 1994 Ith bound (numbering from 1). Otherwise, NULL. */ 1995 1996 static struct type * 1997 desc_index_type (struct type *type, int i) 1998 { 1999 type = desc_base_type (type); 2000 2001 if (type->code () == TYPE_CODE_STRUCT) 2002 { 2003 char bound_name[20]; 2004 xsnprintf (bound_name, sizeof (bound_name), "LB%d", i - 1); 2005 return lookup_struct_elt_type (type, bound_name, 1); 2006 } 2007 else 2008 return NULL; 2009 } 2010 2011 /* The number of index positions in the array-bounds type TYPE. 2012 Return 0 if TYPE is NULL. */ 2013 2014 static int 2015 desc_arity (struct type *type) 2016 { 2017 type = desc_base_type (type); 2018 2019 if (type != NULL) 2020 return type->num_fields () / 2; 2021 return 0; 2022 } 2023 2024 /* Non-zero iff TYPE is a simple array type (not a pointer to one) or 2025 an array descriptor type (representing an unconstrained array 2026 type). */ 2027 2028 static int 2029 ada_is_direct_array_type (struct type *type) 2030 { 2031 if (type == NULL) 2032 return 0; 2033 type = ada_check_typedef (type); 2034 return (type->code () == TYPE_CODE_ARRAY 2035 || ada_is_array_descriptor_type (type)); 2036 } 2037 2038 /* Non-zero iff TYPE represents any kind of array in Ada, or a pointer 2039 * to one. */ 2040 2041 static int 2042 ada_is_array_type (struct type *type) 2043 { 2044 while (type != NULL 2045 && (type->code () == TYPE_CODE_PTR 2046 || type->code () == TYPE_CODE_REF)) 2047 type = type->target_type (); 2048 return ada_is_direct_array_type (type); 2049 } 2050 2051 /* Non-zero iff TYPE is a simple array type or pointer to one. */ 2052 2053 int 2054 ada_is_simple_array_type (struct type *type) 2055 { 2056 if (type == NULL) 2057 return 0; 2058 type = ada_check_typedef (type); 2059 return (type->code () == TYPE_CODE_ARRAY 2060 || (type->code () == TYPE_CODE_PTR 2061 && (ada_check_typedef (type->target_type ())->code () 2062 == TYPE_CODE_ARRAY))); 2063 } 2064 2065 /* Non-zero iff TYPE belongs to a GNAT array descriptor. */ 2066 2067 int 2068 ada_is_array_descriptor_type (struct type *type) 2069 { 2070 struct type *data_type = desc_data_target_type (type); 2071 2072 if (type == NULL) 2073 return 0; 2074 type = ada_check_typedef (type); 2075 return (data_type != NULL 2076 && data_type->code () == TYPE_CODE_ARRAY 2077 && desc_arity (desc_bounds_type (type)) > 0); 2078 } 2079 2080 /* Non-zero iff type is a partially mal-formed GNAT array 2081 descriptor. FIXME: This is to compensate for some problems with 2082 debugging output from GNAT. Re-examine periodically to see if it 2083 is still needed. */ 2084 2085 int 2086 ada_is_bogus_array_descriptor (struct type *type) 2087 { 2088 return 2089 type != NULL 2090 && type->code () == TYPE_CODE_STRUCT 2091 && (lookup_struct_elt_type (type, "P_BOUNDS", 1) != NULL 2092 || lookup_struct_elt_type (type, "P_ARRAY", 1) != NULL) 2093 && !ada_is_array_descriptor_type (type); 2094 } 2095 2096 2097 /* If ARR has a record type in the form of a standard GNAT array descriptor, 2098 (fat pointer) returns the type of the array data described---specifically, 2099 a pointer-to-array type. If BOUNDS is non-zero, the bounds data are filled 2100 in from the descriptor; otherwise, they are left unspecified. If 2101 the ARR denotes a null array descriptor and BOUNDS is non-zero, 2102 returns NULL. The result is simply the type of ARR if ARR is not 2103 a descriptor. */ 2104 2105 static struct type * 2106 ada_type_of_array (struct value *arr, int bounds) 2107 { 2108 if (ada_is_constrained_packed_array_type (value_type (arr))) 2109 return decode_constrained_packed_array_type (value_type (arr)); 2110 2111 if (!ada_is_array_descriptor_type (value_type (arr))) 2112 return value_type (arr); 2113 2114 if (!bounds) 2115 { 2116 struct type *array_type = 2117 ada_check_typedef (desc_data_target_type (value_type (arr))); 2118 2119 if (ada_is_unconstrained_packed_array_type (value_type (arr))) 2120 TYPE_FIELD_BITSIZE (array_type, 0) = 2121 decode_packed_array_bitsize (value_type (arr)); 2122 2123 return array_type; 2124 } 2125 else 2126 { 2127 struct type *elt_type; 2128 int arity; 2129 struct value *descriptor; 2130 2131 elt_type = ada_array_element_type (value_type (arr), -1); 2132 arity = ada_array_arity (value_type (arr)); 2133 2134 if (elt_type == NULL || arity == 0) 2135 return ada_check_typedef (value_type (arr)); 2136 2137 descriptor = desc_bounds (arr); 2138 if (value_as_long (descriptor) == 0) 2139 return NULL; 2140 while (arity > 0) 2141 { 2142 struct type *range_type = alloc_type_copy (value_type (arr)); 2143 struct type *array_type = alloc_type_copy (value_type (arr)); 2144 struct value *low = desc_one_bound (descriptor, arity, 0); 2145 struct value *high = desc_one_bound (descriptor, arity, 1); 2146 2147 arity -= 1; 2148 create_static_range_type (range_type, value_type (low), 2149 longest_to_int (value_as_long (low)), 2150 longest_to_int (value_as_long (high))); 2151 elt_type = create_array_type (array_type, elt_type, range_type); 2152 2153 if (ada_is_unconstrained_packed_array_type (value_type (arr))) 2154 { 2155 /* We need to store the element packed bitsize, as well as 2156 recompute the array size, because it was previously 2157 computed based on the unpacked element size. */ 2158 LONGEST lo = value_as_long (low); 2159 LONGEST hi = value_as_long (high); 2160 2161 TYPE_FIELD_BITSIZE (elt_type, 0) = 2162 decode_packed_array_bitsize (value_type (arr)); 2163 /* If the array has no element, then the size is already 2164 zero, and does not need to be recomputed. */ 2165 if (lo < hi) 2166 { 2167 int array_bitsize = 2168 (hi - lo + 1) * TYPE_FIELD_BITSIZE (elt_type, 0); 2169 2170 array_type->set_length ((array_bitsize + 7) / 8); 2171 } 2172 } 2173 } 2174 2175 return lookup_pointer_type (elt_type); 2176 } 2177 } 2178 2179 /* If ARR does not represent an array, returns ARR unchanged. 2180 Otherwise, returns either a standard GDB array with bounds set 2181 appropriately or, if ARR is a non-null fat pointer, a pointer to a standard 2182 GDB array. Returns NULL if ARR is a null fat pointer. */ 2183 2184 struct value * 2185 ada_coerce_to_simple_array_ptr (struct value *arr) 2186 { 2187 if (ada_is_array_descriptor_type (value_type (arr))) 2188 { 2189 struct type *arrType = ada_type_of_array (arr, 1); 2190 2191 if (arrType == NULL) 2192 return NULL; 2193 return value_cast (arrType, value_copy (desc_data (arr))); 2194 } 2195 else if (ada_is_constrained_packed_array_type (value_type (arr))) 2196 return decode_constrained_packed_array (arr); 2197 else 2198 return arr; 2199 } 2200 2201 /* If ARR does not represent an array, returns ARR unchanged. 2202 Otherwise, returns a standard GDB array describing ARR (which may 2203 be ARR itself if it already is in the proper form). */ 2204 2205 struct value * 2206 ada_coerce_to_simple_array (struct value *arr) 2207 { 2208 if (ada_is_array_descriptor_type (value_type (arr))) 2209 { 2210 struct value *arrVal = ada_coerce_to_simple_array_ptr (arr); 2211 2212 if (arrVal == NULL) 2213 error (_("Bounds unavailable for null array pointer.")); 2214 return value_ind (arrVal); 2215 } 2216 else if (ada_is_constrained_packed_array_type (value_type (arr))) 2217 return decode_constrained_packed_array (arr); 2218 else 2219 return arr; 2220 } 2221 2222 /* If TYPE represents a GNAT array type, return it translated to an 2223 ordinary GDB array type (possibly with BITSIZE fields indicating 2224 packing). For other types, is the identity. */ 2225 2226 struct type * 2227 ada_coerce_to_simple_array_type (struct type *type) 2228 { 2229 if (ada_is_constrained_packed_array_type (type)) 2230 return decode_constrained_packed_array_type (type); 2231 2232 if (ada_is_array_descriptor_type (type)) 2233 return ada_check_typedef (desc_data_target_type (type)); 2234 2235 return type; 2236 } 2237 2238 /* Non-zero iff TYPE represents a standard GNAT packed-array type. */ 2239 2240 static int 2241 ada_is_gnat_encoded_packed_array_type (struct type *type) 2242 { 2243 if (type == NULL) 2244 return 0; 2245 type = desc_base_type (type); 2246 type = ada_check_typedef (type); 2247 return 2248 ada_type_name (type) != NULL 2249 && strstr (ada_type_name (type), "___XP") != NULL; 2250 } 2251 2252 /* Non-zero iff TYPE represents a standard GNAT constrained 2253 packed-array type. */ 2254 2255 int 2256 ada_is_constrained_packed_array_type (struct type *type) 2257 { 2258 return ada_is_gnat_encoded_packed_array_type (type) 2259 && !ada_is_array_descriptor_type (type); 2260 } 2261 2262 /* Non-zero iff TYPE represents an array descriptor for a 2263 unconstrained packed-array type. */ 2264 2265 static int 2266 ada_is_unconstrained_packed_array_type (struct type *type) 2267 { 2268 if (!ada_is_array_descriptor_type (type)) 2269 return 0; 2270 2271 if (ada_is_gnat_encoded_packed_array_type (type)) 2272 return 1; 2273 2274 /* If we saw GNAT encodings, then the above code is sufficient. 2275 However, with minimal encodings, we will just have a thick 2276 pointer instead. */ 2277 if (is_thick_pntr (type)) 2278 { 2279 type = desc_base_type (type); 2280 /* The structure's first field is a pointer to an array, so this 2281 fetches the array type. */ 2282 type = type->field (0).type ()->target_type (); 2283 if (type->code () == TYPE_CODE_TYPEDEF) 2284 type = ada_typedef_target_type (type); 2285 /* Now we can see if the array elements are packed. */ 2286 return TYPE_FIELD_BITSIZE (type, 0) > 0; 2287 } 2288 2289 return 0; 2290 } 2291 2292 /* Return true if TYPE is a (Gnat-encoded) constrained packed array 2293 type, or if it is an ordinary (non-Gnat-encoded) packed array. */ 2294 2295 static bool 2296 ada_is_any_packed_array_type (struct type *type) 2297 { 2298 return (ada_is_constrained_packed_array_type (type) 2299 || (type->code () == TYPE_CODE_ARRAY 2300 && TYPE_FIELD_BITSIZE (type, 0) % 8 != 0)); 2301 } 2302 2303 /* Given that TYPE encodes a packed array type (constrained or unconstrained), 2304 return the size of its elements in bits. */ 2305 2306 static long 2307 decode_packed_array_bitsize (struct type *type) 2308 { 2309 const char *raw_name; 2310 const char *tail; 2311 long bits; 2312 2313 /* Access to arrays implemented as fat pointers are encoded as a typedef 2314 of the fat pointer type. We need the name of the fat pointer type 2315 to do the decoding, so strip the typedef layer. */ 2316 if (type->code () == TYPE_CODE_TYPEDEF) 2317 type = ada_typedef_target_type (type); 2318 2319 raw_name = ada_type_name (ada_check_typedef (type)); 2320 if (!raw_name) 2321 raw_name = ada_type_name (desc_base_type (type)); 2322 2323 if (!raw_name) 2324 return 0; 2325 2326 tail = strstr (raw_name, "___XP"); 2327 if (tail == nullptr) 2328 { 2329 gdb_assert (is_thick_pntr (type)); 2330 /* The structure's first field is a pointer to an array, so this 2331 fetches the array type. */ 2332 type = type->field (0).type ()->target_type (); 2333 /* Now we can see if the array elements are packed. */ 2334 return TYPE_FIELD_BITSIZE (type, 0); 2335 } 2336 2337 if (sscanf (tail + sizeof ("___XP") - 1, "%ld", &bits) != 1) 2338 { 2339 lim_warning 2340 (_("could not understand bit size information on packed array")); 2341 return 0; 2342 } 2343 2344 return bits; 2345 } 2346 2347 /* Given that TYPE is a standard GDB array type with all bounds filled 2348 in, and that the element size of its ultimate scalar constituents 2349 (that is, either its elements, or, if it is an array of arrays, its 2350 elements' elements, etc.) is *ELT_BITS, return an identical type, 2351 but with the bit sizes of its elements (and those of any 2352 constituent arrays) recorded in the BITSIZE components of its 2353 TYPE_FIELD_BITSIZE values, and with *ELT_BITS set to its total size 2354 in bits. 2355 2356 Note that, for arrays whose index type has an XA encoding where 2357 a bound references a record discriminant, getting that discriminant, 2358 and therefore the actual value of that bound, is not possible 2359 because none of the given parameters gives us access to the record. 2360 This function assumes that it is OK in the context where it is being 2361 used to return an array whose bounds are still dynamic and where 2362 the length is arbitrary. */ 2363 2364 static struct type * 2365 constrained_packed_array_type (struct type *type, long *elt_bits) 2366 { 2367 struct type *new_elt_type; 2368 struct type *new_type; 2369 struct type *index_type_desc; 2370 struct type *index_type; 2371 LONGEST low_bound, high_bound; 2372 2373 type = ada_check_typedef (type); 2374 if (type->code () != TYPE_CODE_ARRAY) 2375 return type; 2376 2377 index_type_desc = ada_find_parallel_type (type, "___XA"); 2378 if (index_type_desc) 2379 index_type = to_fixed_range_type (index_type_desc->field (0).type (), 2380 NULL); 2381 else 2382 index_type = type->index_type (); 2383 2384 new_type = alloc_type_copy (type); 2385 new_elt_type = 2386 constrained_packed_array_type (ada_check_typedef (type->target_type ()), 2387 elt_bits); 2388 create_array_type (new_type, new_elt_type, index_type); 2389 TYPE_FIELD_BITSIZE (new_type, 0) = *elt_bits; 2390 new_type->set_name (ada_type_name (type)); 2391 2392 if ((check_typedef (index_type)->code () == TYPE_CODE_RANGE 2393 && is_dynamic_type (check_typedef (index_type))) 2394 || !get_discrete_bounds (index_type, &low_bound, &high_bound)) 2395 low_bound = high_bound = 0; 2396 if (high_bound < low_bound) 2397 { 2398 *elt_bits = 0; 2399 new_type->set_length (0); 2400 } 2401 else 2402 { 2403 *elt_bits *= (high_bound - low_bound + 1); 2404 new_type->set_length ((*elt_bits + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT); 2405 } 2406 2407 new_type->set_is_fixed_instance (true); 2408 return new_type; 2409 } 2410 2411 /* The array type encoded by TYPE, where 2412 ada_is_constrained_packed_array_type (TYPE). */ 2413 2414 static struct type * 2415 decode_constrained_packed_array_type (struct type *type) 2416 { 2417 const char *raw_name = ada_type_name (ada_check_typedef (type)); 2418 char *name; 2419 const char *tail; 2420 struct type *shadow_type; 2421 long bits; 2422 2423 if (!raw_name) 2424 raw_name = ada_type_name (desc_base_type (type)); 2425 2426 if (!raw_name) 2427 return NULL; 2428 2429 name = (char *) alloca (strlen (raw_name) + 1); 2430 tail = strstr (raw_name, "___XP"); 2431 type = desc_base_type (type); 2432 2433 memcpy (name, raw_name, tail - raw_name); 2434 name[tail - raw_name] = '\000'; 2435 2436 shadow_type = ada_find_parallel_type_with_name (type, name); 2437 2438 if (shadow_type == NULL) 2439 { 2440 lim_warning (_("could not find bounds information on packed array")); 2441 return NULL; 2442 } 2443 shadow_type = check_typedef (shadow_type); 2444 2445 if (shadow_type->code () != TYPE_CODE_ARRAY) 2446 { 2447 lim_warning (_("could not understand bounds " 2448 "information on packed array")); 2449 return NULL; 2450 } 2451 2452 bits = decode_packed_array_bitsize (type); 2453 return constrained_packed_array_type (shadow_type, &bits); 2454 } 2455 2456 /* Helper function for decode_constrained_packed_array. Set the field 2457 bitsize on a series of packed arrays. Returns the number of 2458 elements in TYPE. */ 2459 2460 static LONGEST 2461 recursively_update_array_bitsize (struct type *type) 2462 { 2463 gdb_assert (type->code () == TYPE_CODE_ARRAY); 2464 2465 LONGEST low, high; 2466 if (!get_discrete_bounds (type->index_type (), &low, &high) 2467 || low > high) 2468 return 0; 2469 LONGEST our_len = high - low + 1; 2470 2471 struct type *elt_type = type->target_type (); 2472 if (elt_type->code () == TYPE_CODE_ARRAY) 2473 { 2474 LONGEST elt_len = recursively_update_array_bitsize (elt_type); 2475 LONGEST elt_bitsize = elt_len * TYPE_FIELD_BITSIZE (elt_type, 0); 2476 TYPE_FIELD_BITSIZE (type, 0) = elt_bitsize; 2477 2478 type->set_length (((our_len * elt_bitsize + HOST_CHAR_BIT - 1) 2479 / HOST_CHAR_BIT)); 2480 } 2481 2482 return our_len; 2483 } 2484 2485 /* Given that ARR is a struct value *indicating a GNAT constrained packed 2486 array, returns a simple array that denotes that array. Its type is a 2487 standard GDB array type except that the BITSIZEs of the array 2488 target types are set to the number of bits in each element, and the 2489 type length is set appropriately. */ 2490 2491 static struct value * 2492 decode_constrained_packed_array (struct value *arr) 2493 { 2494 struct type *type; 2495 2496 /* If our value is a pointer, then dereference it. Likewise if 2497 the value is a reference. Make sure that this operation does not 2498 cause the target type to be fixed, as this would indirectly cause 2499 this array to be decoded. The rest of the routine assumes that 2500 the array hasn't been decoded yet, so we use the basic "coerce_ref" 2501 and "value_ind" routines to perform the dereferencing, as opposed 2502 to using "ada_coerce_ref" or "ada_value_ind". */ 2503 arr = coerce_ref (arr); 2504 if (ada_check_typedef (value_type (arr))->code () == TYPE_CODE_PTR) 2505 arr = value_ind (arr); 2506 2507 type = decode_constrained_packed_array_type (value_type (arr)); 2508 if (type == NULL) 2509 { 2510 error (_("can't unpack array")); 2511 return NULL; 2512 } 2513 2514 /* Decoding the packed array type could not correctly set the field 2515 bitsizes for any dimension except the innermost, because the 2516 bounds may be variable and were not passed to that function. So, 2517 we further resolve the array bounds here and then update the 2518 sizes. */ 2519 const gdb_byte *valaddr = value_contents_for_printing (arr).data (); 2520 CORE_ADDR address = value_address (arr); 2521 gdb::array_view<const gdb_byte> view 2522 = gdb::make_array_view (valaddr, type->length ()); 2523 type = resolve_dynamic_type (type, view, address); 2524 recursively_update_array_bitsize (type); 2525 2526 if (type_byte_order (value_type (arr)) == BFD_ENDIAN_BIG 2527 && ada_is_modular_type (value_type (arr))) 2528 { 2529 /* This is a (right-justified) modular type representing a packed 2530 array with no wrapper. In order to interpret the value through 2531 the (left-justified) packed array type we just built, we must 2532 first left-justify it. */ 2533 int bit_size, bit_pos; 2534 ULONGEST mod; 2535 2536 mod = ada_modulus (value_type (arr)) - 1; 2537 bit_size = 0; 2538 while (mod > 0) 2539 { 2540 bit_size += 1; 2541 mod >>= 1; 2542 } 2543 bit_pos = HOST_CHAR_BIT * value_type (arr)->length () - bit_size; 2544 arr = ada_value_primitive_packed_val (arr, NULL, 2545 bit_pos / HOST_CHAR_BIT, 2546 bit_pos % HOST_CHAR_BIT, 2547 bit_size, 2548 type); 2549 } 2550 2551 return coerce_unspec_val_to_type (arr, type); 2552 } 2553 2554 2555 /* The value of the element of packed array ARR at the ARITY indices 2556 given in IND. ARR must be a simple array. */ 2557 2558 static struct value * 2559 value_subscript_packed (struct value *arr, int arity, struct value **ind) 2560 { 2561 int i; 2562 int bits, elt_off, bit_off; 2563 long elt_total_bit_offset; 2564 struct type *elt_type; 2565 struct value *v; 2566 2567 bits = 0; 2568 elt_total_bit_offset = 0; 2569 elt_type = ada_check_typedef (value_type (arr)); 2570 for (i = 0; i < arity; i += 1) 2571 { 2572 if (elt_type->code () != TYPE_CODE_ARRAY 2573 || TYPE_FIELD_BITSIZE (elt_type, 0) == 0) 2574 error 2575 (_("attempt to do packed indexing of " 2576 "something other than a packed array")); 2577 else 2578 { 2579 struct type *range_type = elt_type->index_type (); 2580 LONGEST lowerbound, upperbound; 2581 LONGEST idx; 2582 2583 if (!get_discrete_bounds (range_type, &lowerbound, &upperbound)) 2584 { 2585 lim_warning (_("don't know bounds of array")); 2586 lowerbound = upperbound = 0; 2587 } 2588 2589 idx = pos_atr (ind[i]); 2590 if (idx < lowerbound || idx > upperbound) 2591 lim_warning (_("packed array index %ld out of bounds"), 2592 (long) idx); 2593 bits = TYPE_FIELD_BITSIZE (elt_type, 0); 2594 elt_total_bit_offset += (idx - lowerbound) * bits; 2595 elt_type = ada_check_typedef (elt_type->target_type ()); 2596 } 2597 } 2598 elt_off = elt_total_bit_offset / HOST_CHAR_BIT; 2599 bit_off = elt_total_bit_offset % HOST_CHAR_BIT; 2600 2601 v = ada_value_primitive_packed_val (arr, NULL, elt_off, bit_off, 2602 bits, elt_type); 2603 return v; 2604 } 2605 2606 /* Non-zero iff TYPE includes negative integer values. */ 2607 2608 static int 2609 has_negatives (struct type *type) 2610 { 2611 switch (type->code ()) 2612 { 2613 default: 2614 return 0; 2615 case TYPE_CODE_INT: 2616 return !type->is_unsigned (); 2617 case TYPE_CODE_RANGE: 2618 return type->bounds ()->low.const_val () - type->bounds ()->bias < 0; 2619 } 2620 } 2621 2622 /* With SRC being a buffer containing BIT_SIZE bits of data at BIT_OFFSET, 2623 unpack that data into UNPACKED. UNPACKED_LEN is the size in bytes of 2624 the unpacked buffer. 2625 2626 The size of the unpacked buffer (UNPACKED_LEN) is expected to be large 2627 enough to contain at least BIT_OFFSET bits. If not, an error is raised. 2628 2629 IS_BIG_ENDIAN is nonzero if the data is stored in big endian mode, 2630 zero otherwise. 2631 2632 IS_SIGNED_TYPE is nonzero if the data corresponds to a signed type. 2633 2634 IS_SCALAR is nonzero if the data corresponds to a signed type. */ 2635 2636 static void 2637 ada_unpack_from_contents (const gdb_byte *src, int bit_offset, int bit_size, 2638 gdb_byte *unpacked, int unpacked_len, 2639 int is_big_endian, int is_signed_type, 2640 int is_scalar) 2641 { 2642 int src_len = (bit_size + bit_offset + HOST_CHAR_BIT - 1) / 8; 2643 int src_idx; /* Index into the source area */ 2644 int src_bytes_left; /* Number of source bytes left to process. */ 2645 int srcBitsLeft; /* Number of source bits left to move */ 2646 int unusedLS; /* Number of bits in next significant 2647 byte of source that are unused */ 2648 2649 int unpacked_idx; /* Index into the unpacked buffer */ 2650 int unpacked_bytes_left; /* Number of bytes left to set in unpacked. */ 2651 2652 unsigned long accum; /* Staging area for bits being transferred */ 2653 int accumSize; /* Number of meaningful bits in accum */ 2654 unsigned char sign; 2655 2656 /* Transmit bytes from least to most significant; delta is the direction 2657 the indices move. */ 2658 int delta = is_big_endian ? -1 : 1; 2659 2660 /* Make sure that unpacked is large enough to receive the BIT_SIZE 2661 bits from SRC. .*/ 2662 if ((bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT > unpacked_len) 2663 error (_("Cannot unpack %d bits into buffer of %d bytes"), 2664 bit_size, unpacked_len); 2665 2666 srcBitsLeft = bit_size; 2667 src_bytes_left = src_len; 2668 unpacked_bytes_left = unpacked_len; 2669 sign = 0; 2670 2671 if (is_big_endian) 2672 { 2673 src_idx = src_len - 1; 2674 if (is_signed_type 2675 && ((src[0] << bit_offset) & (1 << (HOST_CHAR_BIT - 1)))) 2676 sign = ~0; 2677 2678 unusedLS = 2679 (HOST_CHAR_BIT - (bit_size + bit_offset) % HOST_CHAR_BIT) 2680 % HOST_CHAR_BIT; 2681 2682 if (is_scalar) 2683 { 2684 accumSize = 0; 2685 unpacked_idx = unpacked_len - 1; 2686 } 2687 else 2688 { 2689 /* Non-scalar values must be aligned at a byte boundary... */ 2690 accumSize = 2691 (HOST_CHAR_BIT - bit_size % HOST_CHAR_BIT) % HOST_CHAR_BIT; 2692 /* ... And are placed at the beginning (most-significant) bytes 2693 of the target. */ 2694 unpacked_idx = (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT - 1; 2695 unpacked_bytes_left = unpacked_idx + 1; 2696 } 2697 } 2698 else 2699 { 2700 int sign_bit_offset = (bit_size + bit_offset - 1) % 8; 2701 2702 src_idx = unpacked_idx = 0; 2703 unusedLS = bit_offset; 2704 accumSize = 0; 2705 2706 if (is_signed_type && (src[src_len - 1] & (1 << sign_bit_offset))) 2707 sign = ~0; 2708 } 2709 2710 accum = 0; 2711 while (src_bytes_left > 0) 2712 { 2713 /* Mask for removing bits of the next source byte that are not 2714 part of the value. */ 2715 unsigned int unusedMSMask = 2716 (1 << (srcBitsLeft >= HOST_CHAR_BIT ? HOST_CHAR_BIT : srcBitsLeft)) - 2717 1; 2718 /* Sign-extend bits for this byte. */ 2719 unsigned int signMask = sign & ~unusedMSMask; 2720 2721 accum |= 2722 (((src[src_idx] >> unusedLS) & unusedMSMask) | signMask) << accumSize; 2723 accumSize += HOST_CHAR_BIT - unusedLS; 2724 if (accumSize >= HOST_CHAR_BIT) 2725 { 2726 unpacked[unpacked_idx] = accum & ~(~0UL << HOST_CHAR_BIT); 2727 accumSize -= HOST_CHAR_BIT; 2728 accum >>= HOST_CHAR_BIT; 2729 unpacked_bytes_left -= 1; 2730 unpacked_idx += delta; 2731 } 2732 srcBitsLeft -= HOST_CHAR_BIT - unusedLS; 2733 unusedLS = 0; 2734 src_bytes_left -= 1; 2735 src_idx += delta; 2736 } 2737 while (unpacked_bytes_left > 0) 2738 { 2739 accum |= sign << accumSize; 2740 unpacked[unpacked_idx] = accum & ~(~0UL << HOST_CHAR_BIT); 2741 accumSize -= HOST_CHAR_BIT; 2742 if (accumSize < 0) 2743 accumSize = 0; 2744 accum >>= HOST_CHAR_BIT; 2745 unpacked_bytes_left -= 1; 2746 unpacked_idx += delta; 2747 } 2748 } 2749 2750 /* Create a new value of type TYPE from the contents of OBJ starting 2751 at byte OFFSET, and bit offset BIT_OFFSET within that byte, 2752 proceeding for BIT_SIZE bits. If OBJ is an lval in memory, then 2753 assigning through the result will set the field fetched from. 2754 VALADDR is ignored unless OBJ is NULL, in which case, 2755 VALADDR+OFFSET must address the start of storage containing the 2756 packed value. The value returned in this case is never an lval. 2757 Assumes 0 <= BIT_OFFSET < HOST_CHAR_BIT. */ 2758 2759 struct value * 2760 ada_value_primitive_packed_val (struct value *obj, const gdb_byte *valaddr, 2761 long offset, int bit_offset, int bit_size, 2762 struct type *type) 2763 { 2764 struct value *v; 2765 const gdb_byte *src; /* First byte containing data to unpack */ 2766 gdb_byte *unpacked; 2767 const int is_scalar = is_scalar_type (type); 2768 const int is_big_endian = type_byte_order (type) == BFD_ENDIAN_BIG; 2769 gdb::byte_vector staging; 2770 2771 type = ada_check_typedef (type); 2772 2773 if (obj == NULL) 2774 src = valaddr + offset; 2775 else 2776 src = value_contents (obj).data () + offset; 2777 2778 if (is_dynamic_type (type)) 2779 { 2780 /* The length of TYPE might by dynamic, so we need to resolve 2781 TYPE in order to know its actual size, which we then use 2782 to create the contents buffer of the value we return. 2783 The difficulty is that the data containing our object is 2784 packed, and therefore maybe not at a byte boundary. So, what 2785 we do, is unpack the data into a byte-aligned buffer, and then 2786 use that buffer as our object's value for resolving the type. */ 2787 int staging_len = (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT; 2788 staging.resize (staging_len); 2789 2790 ada_unpack_from_contents (src, bit_offset, bit_size, 2791 staging.data (), staging.size (), 2792 is_big_endian, has_negatives (type), 2793 is_scalar); 2794 type = resolve_dynamic_type (type, staging, 0); 2795 if (type->length () < (bit_size + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT) 2796 { 2797 /* This happens when the length of the object is dynamic, 2798 and is actually smaller than the space reserved for it. 2799 For instance, in an array of variant records, the bit_size 2800 we're given is the array stride, which is constant and 2801 normally equal to the maximum size of its element. 2802 But, in reality, each element only actually spans a portion 2803 of that stride. */ 2804 bit_size = type->length () * HOST_CHAR_BIT; 2805 } 2806 } 2807 2808 if (obj == NULL) 2809 { 2810 v = allocate_value (type); 2811 src = valaddr + offset; 2812 } 2813 else if (VALUE_LVAL (obj) == lval_memory && value_lazy (obj)) 2814 { 2815 int src_len = (bit_size + bit_offset + HOST_CHAR_BIT - 1) / 8; 2816 gdb_byte *buf; 2817 2818 v = value_at (type, value_address (obj) + offset); 2819 buf = (gdb_byte *) alloca (src_len); 2820 read_memory (value_address (v), buf, src_len); 2821 src = buf; 2822 } 2823 else 2824 { 2825 v = allocate_value (type); 2826 src = value_contents (obj).data () + offset; 2827 } 2828 2829 if (obj != NULL) 2830 { 2831 long new_offset = offset; 2832 2833 set_value_component_location (v, obj); 2834 set_value_bitpos (v, bit_offset + value_bitpos (obj)); 2835 set_value_bitsize (v, bit_size); 2836 if (value_bitpos (v) >= HOST_CHAR_BIT) 2837 { 2838 ++new_offset; 2839 set_value_bitpos (v, value_bitpos (v) - HOST_CHAR_BIT); 2840 } 2841 set_value_offset (v, new_offset); 2842 2843 /* Also set the parent value. This is needed when trying to 2844 assign a new value (in inferior memory). */ 2845 set_value_parent (v, obj); 2846 } 2847 else 2848 set_value_bitsize (v, bit_size); 2849 unpacked = value_contents_writeable (v).data (); 2850 2851 if (bit_size == 0) 2852 { 2853 memset (unpacked, 0, type->length ()); 2854 return v; 2855 } 2856 2857 if (staging.size () == type->length ()) 2858 { 2859 /* Small short-cut: If we've unpacked the data into a buffer 2860 of the same size as TYPE's length, then we can reuse that, 2861 instead of doing the unpacking again. */ 2862 memcpy (unpacked, staging.data (), staging.size ()); 2863 } 2864 else 2865 ada_unpack_from_contents (src, bit_offset, bit_size, 2866 unpacked, type->length (), 2867 is_big_endian, has_negatives (type), is_scalar); 2868 2869 return v; 2870 } 2871 2872 /* Store the contents of FROMVAL into the location of TOVAL. 2873 Return a new value with the location of TOVAL and contents of 2874 FROMVAL. Handles assignment into packed fields that have 2875 floating-point or non-scalar types. */ 2876 2877 static struct value * 2878 ada_value_assign (struct value *toval, struct value *fromval) 2879 { 2880 struct type *type = value_type (toval); 2881 int bits = value_bitsize (toval); 2882 2883 toval = ada_coerce_ref (toval); 2884 fromval = ada_coerce_ref (fromval); 2885 2886 if (ada_is_direct_array_type (value_type (toval))) 2887 toval = ada_coerce_to_simple_array (toval); 2888 if (ada_is_direct_array_type (value_type (fromval))) 2889 fromval = ada_coerce_to_simple_array (fromval); 2890 2891 if (!deprecated_value_modifiable (toval)) 2892 error (_("Left operand of assignment is not a modifiable lvalue.")); 2893 2894 if (VALUE_LVAL (toval) == lval_memory 2895 && bits > 0 2896 && (type->code () == TYPE_CODE_FLT 2897 || type->code () == TYPE_CODE_STRUCT)) 2898 { 2899 int len = (value_bitpos (toval) 2900 + bits + HOST_CHAR_BIT - 1) / HOST_CHAR_BIT; 2901 int from_size; 2902 gdb_byte *buffer = (gdb_byte *) alloca (len); 2903 struct value *val; 2904 CORE_ADDR to_addr = value_address (toval); 2905 2906 if (type->code () == TYPE_CODE_FLT) 2907 fromval = value_cast (type, fromval); 2908 2909 read_memory (to_addr, buffer, len); 2910 from_size = value_bitsize (fromval); 2911 if (from_size == 0) 2912 from_size = value_type (fromval)->length () * TARGET_CHAR_BIT; 2913 2914 const int is_big_endian = type_byte_order (type) == BFD_ENDIAN_BIG; 2915 ULONGEST from_offset = 0; 2916 if (is_big_endian && is_scalar_type (value_type (fromval))) 2917 from_offset = from_size - bits; 2918 copy_bitwise (buffer, value_bitpos (toval), 2919 value_contents (fromval).data (), from_offset, 2920 bits, is_big_endian); 2921 write_memory_with_notification (to_addr, buffer, len); 2922 2923 val = value_copy (toval); 2924 memcpy (value_contents_raw (val).data (), 2925 value_contents (fromval).data (), 2926 type->length ()); 2927 deprecated_set_value_type (val, type); 2928 2929 return val; 2930 } 2931 2932 return value_assign (toval, fromval); 2933 } 2934 2935 2936 /* Given that COMPONENT is a memory lvalue that is part of the lvalue 2937 CONTAINER, assign the contents of VAL to COMPONENTS's place in 2938 CONTAINER. Modifies the VALUE_CONTENTS of CONTAINER only, not 2939 COMPONENT, and not the inferior's memory. The current contents 2940 of COMPONENT are ignored. 2941 2942 Although not part of the initial design, this function also works 2943 when CONTAINER and COMPONENT are not_lval's: it works as if CONTAINER 2944 had a null address, and COMPONENT had an address which is equal to 2945 its offset inside CONTAINER. */ 2946 2947 static void 2948 value_assign_to_component (struct value *container, struct value *component, 2949 struct value *val) 2950 { 2951 LONGEST offset_in_container = 2952 (LONGEST) (value_address (component) - value_address (container)); 2953 int bit_offset_in_container = 2954 value_bitpos (component) - value_bitpos (container); 2955 int bits; 2956 2957 val = value_cast (value_type (component), val); 2958 2959 if (value_bitsize (component) == 0) 2960 bits = TARGET_CHAR_BIT * value_type (component)->length (); 2961 else 2962 bits = value_bitsize (component); 2963 2964 if (type_byte_order (value_type (container)) == BFD_ENDIAN_BIG) 2965 { 2966 int src_offset; 2967 2968 if (is_scalar_type (check_typedef (value_type (component)))) 2969 src_offset 2970 = value_type (component)->length () * TARGET_CHAR_BIT - bits; 2971 else 2972 src_offset = 0; 2973 copy_bitwise ((value_contents_writeable (container).data () 2974 + offset_in_container), 2975 value_bitpos (container) + bit_offset_in_container, 2976 value_contents (val).data (), src_offset, bits, 1); 2977 } 2978 else 2979 copy_bitwise ((value_contents_writeable (container).data () 2980 + offset_in_container), 2981 value_bitpos (container) + bit_offset_in_container, 2982 value_contents (val).data (), 0, bits, 0); 2983 } 2984 2985 /* Determine if TYPE is an access to an unconstrained array. */ 2986 2987 bool 2988 ada_is_access_to_unconstrained_array (struct type *type) 2989 { 2990 return (type->code () == TYPE_CODE_TYPEDEF 2991 && is_thick_pntr (ada_typedef_target_type (type))); 2992 } 2993 2994 /* The value of the element of array ARR at the ARITY indices given in IND. 2995 ARR may be either a simple array, GNAT array descriptor, or pointer 2996 thereto. */ 2997 2998 struct value * 2999 ada_value_subscript (struct value *arr, int arity, struct value **ind) 3000 { 3001 int k; 3002 struct value *elt; 3003 struct type *elt_type; 3004 3005 elt = ada_coerce_to_simple_array (arr); 3006 3007 elt_type = ada_check_typedef (value_type (elt)); 3008 if (elt_type->code () == TYPE_CODE_ARRAY 3009 && TYPE_FIELD_BITSIZE (elt_type, 0) > 0) 3010 return value_subscript_packed (elt, arity, ind); 3011 3012 for (k = 0; k < arity; k += 1) 3013 { 3014 struct type *saved_elt_type = elt_type->target_type (); 3015 3016 if (elt_type->code () != TYPE_CODE_ARRAY) 3017 error (_("too many subscripts (%d expected)"), k); 3018 3019 elt = value_subscript (elt, pos_atr (ind[k])); 3020 3021 if (ada_is_access_to_unconstrained_array (saved_elt_type) 3022 && value_type (elt)->code () != TYPE_CODE_TYPEDEF) 3023 { 3024 /* The element is a typedef to an unconstrained array, 3025 except that the value_subscript call stripped the 3026 typedef layer. The typedef layer is GNAT's way to 3027 specify that the element is, at the source level, an 3028 access to the unconstrained array, rather than the 3029 unconstrained array. So, we need to restore that 3030 typedef layer, which we can do by forcing the element's 3031 type back to its original type. Otherwise, the returned 3032 value is going to be printed as the array, rather 3033 than as an access. Another symptom of the same issue 3034 would be that an expression trying to dereference the 3035 element would also be improperly rejected. */ 3036 deprecated_set_value_type (elt, saved_elt_type); 3037 } 3038 3039 elt_type = ada_check_typedef (value_type (elt)); 3040 } 3041 3042 return elt; 3043 } 3044 3045 /* Assuming ARR is a pointer to a GDB array, the value of the element 3046 of *ARR at the ARITY indices given in IND. 3047 Does not read the entire array into memory. 3048 3049 Note: Unlike what one would expect, this function is used instead of 3050 ada_value_subscript for basically all non-packed array types. The reason 3051 for this is that a side effect of doing our own pointer arithmetics instead 3052 of relying on value_subscript is that there is no implicit typedef peeling. 3053 This is important for arrays of array accesses, where it allows us to 3054 preserve the fact that the array's element is an array access, where the 3055 access part os encoded in a typedef layer. */ 3056 3057 static struct value * 3058 ada_value_ptr_subscript (struct value *arr, int arity, struct value **ind) 3059 { 3060 int k; 3061 struct value *array_ind = ada_value_ind (arr); 3062 struct type *type 3063 = check_typedef (value_enclosing_type (array_ind)); 3064 3065 if (type->code () == TYPE_CODE_ARRAY 3066 && TYPE_FIELD_BITSIZE (type, 0) > 0) 3067 return value_subscript_packed (array_ind, arity, ind); 3068 3069 for (k = 0; k < arity; k += 1) 3070 { 3071 LONGEST lwb, upb; 3072 3073 if (type->code () != TYPE_CODE_ARRAY) 3074 error (_("too many subscripts (%d expected)"), k); 3075 arr = value_cast (lookup_pointer_type (type->target_type ()), 3076 value_copy (arr)); 3077 get_discrete_bounds (type->index_type (), &lwb, &upb); 3078 arr = value_ptradd (arr, pos_atr (ind[k]) - lwb); 3079 type = type->target_type (); 3080 } 3081 3082 return value_ind (arr); 3083 } 3084 3085 /* Given that ARRAY_PTR is a pointer or reference to an array of type TYPE (the 3086 actual type of ARRAY_PTR is ignored), returns the Ada slice of 3087 HIGH'Pos-LOW'Pos+1 elements starting at index LOW. The lower bound of 3088 this array is LOW, as per Ada rules. */ 3089 static struct value * 3090 ada_value_slice_from_ptr (struct value *array_ptr, struct type *type, 3091 int low, int high) 3092 { 3093 struct type *type0 = ada_check_typedef (type); 3094 struct type *base_index_type = type0->index_type ()->target_type (); 3095 struct type *index_type 3096 = create_static_range_type (NULL, base_index_type, low, high); 3097 struct type *slice_type = create_array_type_with_stride 3098 (NULL, type0->target_type (), index_type, 3099 type0->dyn_prop (DYN_PROP_BYTE_STRIDE), 3100 TYPE_FIELD_BITSIZE (type0, 0)); 3101 int base_low = ada_discrete_type_low_bound (type0->index_type ()); 3102 gdb::optional<LONGEST> base_low_pos, low_pos; 3103 CORE_ADDR base; 3104 3105 low_pos = discrete_position (base_index_type, low); 3106 base_low_pos = discrete_position (base_index_type, base_low); 3107 3108 if (!low_pos.has_value () || !base_low_pos.has_value ()) 3109 { 3110 warning (_("unable to get positions in slice, use bounds instead")); 3111 low_pos = low; 3112 base_low_pos = base_low; 3113 } 3114 3115 ULONGEST stride = TYPE_FIELD_BITSIZE (slice_type, 0) / 8; 3116 if (stride == 0) 3117 stride = type0->target_type ()->length (); 3118 3119 base = value_as_address (array_ptr) + (*low_pos - *base_low_pos) * stride; 3120 return value_at_lazy (slice_type, base); 3121 } 3122 3123 3124 static struct value * 3125 ada_value_slice (struct value *array, int low, int high) 3126 { 3127 struct type *type = ada_check_typedef (value_type (array)); 3128 struct type *base_index_type = type->index_type ()->target_type (); 3129 struct type *index_type 3130 = create_static_range_type (NULL, type->index_type (), low, high); 3131 struct type *slice_type = create_array_type_with_stride 3132 (NULL, type->target_type (), index_type, 3133 type->dyn_prop (DYN_PROP_BYTE_STRIDE), 3134 TYPE_FIELD_BITSIZE (type, 0)); 3135 gdb::optional<LONGEST> low_pos, high_pos; 3136 3137 3138 low_pos = discrete_position (base_index_type, low); 3139 high_pos = discrete_position (base_index_type, high); 3140 3141 if (!low_pos.has_value () || !high_pos.has_value ()) 3142 { 3143 warning (_("unable to get positions in slice, use bounds instead")); 3144 low_pos = low; 3145 high_pos = high; 3146 } 3147 3148 return value_cast (slice_type, 3149 value_slice (array, low, *high_pos - *low_pos + 1)); 3150 } 3151 3152 /* If type is a record type in the form of a standard GNAT array 3153 descriptor, returns the number of dimensions for type. If arr is a 3154 simple array, returns the number of "array of"s that prefix its 3155 type designation. Otherwise, returns 0. */ 3156 3157 int 3158 ada_array_arity (struct type *type) 3159 { 3160 int arity; 3161 3162 if (type == NULL) 3163 return 0; 3164 3165 type = desc_base_type (type); 3166 3167 arity = 0; 3168 if (type->code () == TYPE_CODE_STRUCT) 3169 return desc_arity (desc_bounds_type (type)); 3170 else 3171 while (type->code () == TYPE_CODE_ARRAY) 3172 { 3173 arity += 1; 3174 type = ada_check_typedef (type->target_type ()); 3175 } 3176 3177 return arity; 3178 } 3179 3180 /* If TYPE is a record type in the form of a standard GNAT array 3181 descriptor or a simple array type, returns the element type for 3182 TYPE after indexing by NINDICES indices, or by all indices if 3183 NINDICES is -1. Otherwise, returns NULL. */ 3184 3185 struct type * 3186 ada_array_element_type (struct type *type, int nindices) 3187 { 3188 type = desc_base_type (type); 3189 3190 if (type->code () == TYPE_CODE_STRUCT) 3191 { 3192 int k; 3193 struct type *p_array_type; 3194 3195 p_array_type = desc_data_target_type (type); 3196 3197 k = ada_array_arity (type); 3198 if (k == 0) 3199 return NULL; 3200 3201 /* Initially p_array_type = elt_type(*)[]...(k times)...[]. */ 3202 if (nindices >= 0 && k > nindices) 3203 k = nindices; 3204 while (k > 0 && p_array_type != NULL) 3205 { 3206 p_array_type = ada_check_typedef (p_array_type->target_type ()); 3207 k -= 1; 3208 } 3209 return p_array_type; 3210 } 3211 else if (type->code () == TYPE_CODE_ARRAY) 3212 { 3213 while (nindices != 0 && type->code () == TYPE_CODE_ARRAY) 3214 { 3215 type = type->target_type (); 3216 /* A multi-dimensional array is represented using a sequence 3217 of array types. If one of these types has a name, then 3218 it is not another dimension of the outer array, but 3219 rather the element type of the outermost array. */ 3220 if (type->name () != nullptr) 3221 break; 3222 nindices -= 1; 3223 } 3224 return type; 3225 } 3226 3227 return NULL; 3228 } 3229 3230 /* See ada-lang.h. */ 3231 3232 struct type * 3233 ada_index_type (struct type *type, int n, const char *name) 3234 { 3235 struct type *result_type; 3236 3237 type = desc_base_type (type); 3238 3239 if (n < 0 || n > ada_array_arity (type)) 3240 error (_("invalid dimension number to '%s"), name); 3241 3242 if (ada_is_simple_array_type (type)) 3243 { 3244 int i; 3245 3246 for (i = 1; i < n; i += 1) 3247 { 3248 type = ada_check_typedef (type); 3249 type = type->target_type (); 3250 } 3251 result_type = ada_check_typedef (type)->index_type ()->target_type (); 3252 /* FIXME: The stabs type r(0,0);bound;bound in an array type 3253 has a target type of TYPE_CODE_UNDEF. We compensate here, but 3254 perhaps stabsread.c would make more sense. */ 3255 if (result_type && result_type->code () == TYPE_CODE_UNDEF) 3256 result_type = NULL; 3257 } 3258 else 3259 { 3260 result_type = desc_index_type (desc_bounds_type (type), n); 3261 if (result_type == NULL) 3262 error (_("attempt to take bound of something that is not an array")); 3263 } 3264 3265 return result_type; 3266 } 3267 3268 /* Given that arr is an array type, returns the lower bound of the 3269 Nth index (numbering from 1) if WHICH is 0, and the upper bound if 3270 WHICH is 1. This returns bounds 0 .. -1 if ARR_TYPE is an 3271 array-descriptor type. It works for other arrays with bounds supplied 3272 by run-time quantities other than discriminants. */ 3273 3274 static LONGEST 3275 ada_array_bound_from_type (struct type *arr_type, int n, int which) 3276 { 3277 struct type *type, *index_type_desc, *index_type; 3278 int i; 3279 3280 gdb_assert (which == 0 || which == 1); 3281 3282 if (ada_is_constrained_packed_array_type (arr_type)) 3283 arr_type = decode_constrained_packed_array_type (arr_type); 3284 3285 if (arr_type == NULL || !ada_is_simple_array_type (arr_type)) 3286 return (LONGEST) - which; 3287 3288 if (arr_type->code () == TYPE_CODE_PTR) 3289 type = arr_type->target_type (); 3290 else 3291 type = arr_type; 3292 3293 if (type->is_fixed_instance ()) 3294 { 3295 /* The array has already been fixed, so we do not need to 3296 check the parallel ___XA type again. That encoding has 3297 already been applied, so ignore it now. */ 3298 index_type_desc = NULL; 3299 } 3300 else 3301 { 3302 index_type_desc = ada_find_parallel_type (type, "___XA"); 3303 ada_fixup_array_indexes_type (index_type_desc); 3304 } 3305 3306 if (index_type_desc != NULL) 3307 index_type = to_fixed_range_type (index_type_desc->field (n - 1).type (), 3308 NULL); 3309 else 3310 { 3311 struct type *elt_type = check_typedef (type); 3312 3313 for (i = 1; i < n; i++) 3314 elt_type = check_typedef (elt_type->target_type ()); 3315 3316 index_type = elt_type->index_type (); 3317 } 3318 3319 return 3320 (LONGEST) (which == 0 3321 ? ada_discrete_type_low_bound (index_type) 3322 : ada_discrete_type_high_bound (index_type)); 3323 } 3324 3325 /* Given that arr is an array value, returns the lower bound of the 3326 nth index (numbering from 1) if WHICH is 0, and the upper bound if 3327 WHICH is 1. This routine will also work for arrays with bounds 3328 supplied by run-time quantities other than discriminants. */ 3329 3330 static LONGEST 3331 ada_array_bound (struct value *arr, int n, int which) 3332 { 3333 struct type *arr_type; 3334 3335 if (check_typedef (value_type (arr))->code () == TYPE_CODE_PTR) 3336 arr = value_ind (arr); 3337 arr_type = value_enclosing_type (arr); 3338 3339 if (ada_is_constrained_packed_array_type (arr_type)) 3340 return ada_array_bound (decode_constrained_packed_array (arr), n, which); 3341 else if (ada_is_simple_array_type (arr_type)) 3342 return ada_array_bound_from_type (arr_type, n, which); 3343 else 3344 return value_as_long (desc_one_bound (desc_bounds (arr), n, which)); 3345 } 3346 3347 /* Given that arr is an array value, returns the length of the 3348 nth index. This routine will also work for arrays with bounds 3349 supplied by run-time quantities other than discriminants. 3350 Does not work for arrays indexed by enumeration types with representation 3351 clauses at the moment. */ 3352 3353 static LONGEST 3354 ada_array_length (struct value *arr, int n) 3355 { 3356 struct type *arr_type, *index_type; 3357 int low, high; 3358 3359 if (check_typedef (value_type (arr))->code () == TYPE_CODE_PTR) 3360 arr = value_ind (arr); 3361 arr_type = value_enclosing_type (arr); 3362 3363 if (ada_is_constrained_packed_array_type (arr_type)) 3364 return ada_array_length (decode_constrained_packed_array (arr), n); 3365 3366 if (ada_is_simple_array_type (arr_type)) 3367 { 3368 low = ada_array_bound_from_type (arr_type, n, 0); 3369 high = ada_array_bound_from_type (arr_type, n, 1); 3370 } 3371 else 3372 { 3373 low = value_as_long (desc_one_bound (desc_bounds (arr), n, 0)); 3374 high = value_as_long (desc_one_bound (desc_bounds (arr), n, 1)); 3375 } 3376 3377 arr_type = check_typedef (arr_type); 3378 index_type = ada_index_type (arr_type, n, "length"); 3379 if (index_type != NULL) 3380 { 3381 struct type *base_type; 3382 if (index_type->code () == TYPE_CODE_RANGE) 3383 base_type = index_type->target_type (); 3384 else 3385 base_type = index_type; 3386 3387 low = pos_atr (value_from_longest (base_type, low)); 3388 high = pos_atr (value_from_longest (base_type, high)); 3389 } 3390 return high - low + 1; 3391 } 3392 3393 /* An array whose type is that of ARR_TYPE (an array type), with 3394 bounds LOW to HIGH, but whose contents are unimportant. If HIGH is 3395 less than LOW, then LOW-1 is used. */ 3396 3397 static struct value * 3398 empty_array (struct type *arr_type, int low, int high) 3399 { 3400 struct type *arr_type0 = ada_check_typedef (arr_type); 3401 struct type *index_type 3402 = create_static_range_type 3403 (NULL, arr_type0->index_type ()->target_type (), low, 3404 high < low ? low - 1 : high); 3405 struct type *elt_type = ada_array_element_type (arr_type0, 1); 3406 3407 return allocate_value (create_array_type (NULL, elt_type, index_type)); 3408 } 3409 3410 3411 /* Name resolution */ 3412 3413 /* The "decoded" name for the user-definable Ada operator corresponding 3414 to OP. */ 3415 3416 static const char * 3417 ada_decoded_op_name (enum exp_opcode op) 3418 { 3419 int i; 3420 3421 for (i = 0; ada_opname_table[i].encoded != NULL; i += 1) 3422 { 3423 if (ada_opname_table[i].op == op) 3424 return ada_opname_table[i].decoded; 3425 } 3426 error (_("Could not find operator name for opcode")); 3427 } 3428 3429 /* Returns true (non-zero) iff decoded name N0 should appear before N1 3430 in a listing of choices during disambiguation (see sort_choices, below). 3431 The idea is that overloadings of a subprogram name from the 3432 same package should sort in their source order. We settle for ordering 3433 such symbols by their trailing number (__N or $N). */ 3434 3435 static int 3436 encoded_ordered_before (const char *N0, const char *N1) 3437 { 3438 if (N1 == NULL) 3439 return 0; 3440 else if (N0 == NULL) 3441 return 1; 3442 else 3443 { 3444 int k0, k1; 3445 3446 for (k0 = strlen (N0) - 1; k0 > 0 && isdigit (N0[k0]); k0 -= 1) 3447 ; 3448 for (k1 = strlen (N1) - 1; k1 > 0 && isdigit (N1[k1]); k1 -= 1) 3449 ; 3450 if ((N0[k0] == '_' || N0[k0] == '$') && N0[k0 + 1] != '\000' 3451 && (N1[k1] == '_' || N1[k1] == '$') && N1[k1 + 1] != '\000') 3452 { 3453 int n0, n1; 3454 3455 n0 = k0; 3456 while (N0[n0] == '_' && n0 > 0 && N0[n0 - 1] == '_') 3457 n0 -= 1; 3458 n1 = k1; 3459 while (N1[n1] == '_' && n1 > 0 && N1[n1 - 1] == '_') 3460 n1 -= 1; 3461 if (n0 == n1 && strncmp (N0, N1, n0) == 0) 3462 return (atoi (N0 + k0 + 1) < atoi (N1 + k1 + 1)); 3463 } 3464 return (strcmp (N0, N1) < 0); 3465 } 3466 } 3467 3468 /* Sort SYMS[0..NSYMS-1] to put the choices in a canonical order by the 3469 encoded names. */ 3470 3471 static void 3472 sort_choices (struct block_symbol syms[], int nsyms) 3473 { 3474 int i; 3475 3476 for (i = 1; i < nsyms; i += 1) 3477 { 3478 struct block_symbol sym = syms[i]; 3479 int j; 3480 3481 for (j = i - 1; j >= 0; j -= 1) 3482 { 3483 if (encoded_ordered_before (syms[j].symbol->linkage_name (), 3484 sym.symbol->linkage_name ())) 3485 break; 3486 syms[j + 1] = syms[j]; 3487 } 3488 syms[j + 1] = sym; 3489 } 3490 } 3491 3492 /* Whether GDB should display formals and return types for functions in the 3493 overloads selection menu. */ 3494 static bool print_signatures = true; 3495 3496 /* Print the signature for SYM on STREAM according to the FLAGS options. For 3497 all but functions, the signature is just the name of the symbol. For 3498 functions, this is the name of the function, the list of types for formals 3499 and the return type (if any). */ 3500 3501 static void 3502 ada_print_symbol_signature (struct ui_file *stream, struct symbol *sym, 3503 const struct type_print_options *flags) 3504 { 3505 struct type *type = sym->type (); 3506 3507 gdb_printf (stream, "%s", sym->print_name ()); 3508 if (!print_signatures 3509 || type == NULL 3510 || type->code () != TYPE_CODE_FUNC) 3511 return; 3512 3513 if (type->num_fields () > 0) 3514 { 3515 int i; 3516 3517 gdb_printf (stream, " ("); 3518 for (i = 0; i < type->num_fields (); ++i) 3519 { 3520 if (i > 0) 3521 gdb_printf (stream, "; "); 3522 ada_print_type (type->field (i).type (), NULL, stream, -1, 0, 3523 flags); 3524 } 3525 gdb_printf (stream, ")"); 3526 } 3527 if (type->target_type () != NULL 3528 && type->target_type ()->code () != TYPE_CODE_VOID) 3529 { 3530 gdb_printf (stream, " return "); 3531 ada_print_type (type->target_type (), NULL, stream, -1, 0, flags); 3532 } 3533 } 3534 3535 /* Read and validate a set of numeric choices from the user in the 3536 range 0 .. N_CHOICES-1. Place the results in increasing 3537 order in CHOICES[0 .. N-1], and return N. 3538 3539 The user types choices as a sequence of numbers on one line 3540 separated by blanks, encoding them as follows: 3541 3542 + A choice of 0 means to cancel the selection, throwing an error. 3543 + If IS_ALL_CHOICE, a choice of 1 selects the entire set 0 .. N_CHOICES-1. 3544 + The user chooses k by typing k+IS_ALL_CHOICE+1. 3545 3546 The user is not allowed to choose more than MAX_RESULTS values. 3547 3548 ANNOTATION_SUFFIX, if present, is used to annotate the input 3549 prompts (for use with the -f switch). */ 3550 3551 static int 3552 get_selections (int *choices, int n_choices, int max_results, 3553 int is_all_choice, const char *annotation_suffix) 3554 { 3555 const char *args; 3556 const char *prompt; 3557 int n_chosen; 3558 int first_choice = is_all_choice ? 2 : 1; 3559 3560 prompt = getenv ("PS2"); 3561 if (prompt == NULL) 3562 prompt = "> "; 3563 3564 std::string buffer; 3565 args = command_line_input (buffer, prompt, annotation_suffix); 3566 3567 if (args == NULL) 3568 error_no_arg (_("one or more choice numbers")); 3569 3570 n_chosen = 0; 3571 3572 /* Set choices[0 .. n_chosen-1] to the users' choices in ascending 3573 order, as given in args. Choices are validated. */ 3574 while (1) 3575 { 3576 char *args2; 3577 int choice, j; 3578 3579 args = skip_spaces (args); 3580 if (*args == '\0' && n_chosen == 0) 3581 error_no_arg (_("one or more choice numbers")); 3582 else if (*args == '\0') 3583 break; 3584 3585 choice = strtol (args, &args2, 10); 3586 if (args == args2 || choice < 0 3587 || choice > n_choices + first_choice - 1) 3588 error (_("Argument must be choice number")); 3589 args = args2; 3590 3591 if (choice == 0) 3592 error (_("cancelled")); 3593 3594 if (choice < first_choice) 3595 { 3596 n_chosen = n_choices; 3597 for (j = 0; j < n_choices; j += 1) 3598 choices[j] = j; 3599 break; 3600 } 3601 choice -= first_choice; 3602 3603 for (j = n_chosen - 1; j >= 0 && choice < choices[j]; j -= 1) 3604 { 3605 } 3606 3607 if (j < 0 || choice != choices[j]) 3608 { 3609 int k; 3610 3611 for (k = n_chosen - 1; k > j; k -= 1) 3612 choices[k + 1] = choices[k]; 3613 choices[j + 1] = choice; 3614 n_chosen += 1; 3615 } 3616 } 3617 3618 if (n_chosen > max_results) 3619 error (_("Select no more than %d of the above"), max_results); 3620 3621 return n_chosen; 3622 } 3623 3624 /* Given a list of NSYMS symbols in SYMS, select up to MAX_RESULTS>0 3625 by asking the user (if necessary), returning the number selected, 3626 and setting the first elements of SYMS items. Error if no symbols 3627 selected. */ 3628 3629 /* NOTE: Adapted from decode_line_2 in symtab.c, with which it ought 3630 to be re-integrated one of these days. */ 3631 3632 static int 3633 user_select_syms (struct block_symbol *syms, int nsyms, int max_results) 3634 { 3635 int i; 3636 int *chosen = XALLOCAVEC (int , nsyms); 3637 int n_chosen; 3638 int first_choice = (max_results == 1) ? 1 : 2; 3639 const char *select_mode = multiple_symbols_select_mode (); 3640 3641 if (max_results < 1) 3642 error (_("Request to select 0 symbols!")); 3643 if (nsyms <= 1) 3644 return nsyms; 3645 3646 if (select_mode == multiple_symbols_cancel) 3647 error (_("\ 3648 canceled because the command is ambiguous\n\ 3649 See set/show multiple-symbol.")); 3650 3651 /* If select_mode is "all", then return all possible symbols. 3652 Only do that if more than one symbol can be selected, of course. 3653 Otherwise, display the menu as usual. */ 3654 if (select_mode == multiple_symbols_all && max_results > 1) 3655 return nsyms; 3656 3657 gdb_printf (_("[0] cancel\n")); 3658 if (max_results > 1) 3659 gdb_printf (_("[1] all\n")); 3660 3661 sort_choices (syms, nsyms); 3662 3663 for (i = 0; i < nsyms; i += 1) 3664 { 3665 if (syms[i].symbol == NULL) 3666 continue; 3667 3668 if (syms[i].symbol->aclass () == LOC_BLOCK) 3669 { 3670 struct symtab_and_line sal = 3671 find_function_start_sal (syms[i].symbol, 1); 3672 3673 gdb_printf ("[%d] ", i + first_choice); 3674 ada_print_symbol_signature (gdb_stdout, syms[i].symbol, 3675 &type_print_raw_options); 3676 if (sal.symtab == NULL) 3677 gdb_printf (_(" at %p[<no source file available>%p]:%d\n"), 3678 metadata_style.style ().ptr (), nullptr, sal.line); 3679 else 3680 gdb_printf 3681 (_(" at %ps:%d\n"), 3682 styled_string (file_name_style.style (), 3683 symtab_to_filename_for_display (sal.symtab)), 3684 sal.line); 3685 continue; 3686 } 3687 else 3688 { 3689 int is_enumeral = 3690 (syms[i].symbol->aclass () == LOC_CONST 3691 && syms[i].symbol->type () != NULL 3692 && syms[i].symbol->type ()->code () == TYPE_CODE_ENUM); 3693 struct symtab *symtab = NULL; 3694 3695 if (syms[i].symbol->is_objfile_owned ()) 3696 symtab = syms[i].symbol->symtab (); 3697 3698 if (syms[i].symbol->line () != 0 && symtab != NULL) 3699 { 3700 gdb_printf ("[%d] ", i + first_choice); 3701 ada_print_symbol_signature (gdb_stdout, syms[i].symbol, 3702 &type_print_raw_options); 3703 gdb_printf (_(" at %s:%d\n"), 3704 symtab_to_filename_for_display (symtab), 3705 syms[i].symbol->line ()); 3706 } 3707 else if (is_enumeral 3708 && syms[i].symbol->type ()->name () != NULL) 3709 { 3710 gdb_printf (("[%d] "), i + first_choice); 3711 ada_print_type (syms[i].symbol->type (), NULL, 3712 gdb_stdout, -1, 0, &type_print_raw_options); 3713 gdb_printf (_("'(%s) (enumeral)\n"), 3714 syms[i].symbol->print_name ()); 3715 } 3716 else 3717 { 3718 gdb_printf ("[%d] ", i + first_choice); 3719 ada_print_symbol_signature (gdb_stdout, syms[i].symbol, 3720 &type_print_raw_options); 3721 3722 if (symtab != NULL) 3723 gdb_printf (is_enumeral 3724 ? _(" in %s (enumeral)\n") 3725 : _(" at %s:?\n"), 3726 symtab_to_filename_for_display (symtab)); 3727 else 3728 gdb_printf (is_enumeral 3729 ? _(" (enumeral)\n") 3730 : _(" at ?\n")); 3731 } 3732 } 3733 } 3734 3735 n_chosen = get_selections (chosen, nsyms, max_results, max_results > 1, 3736 "overload-choice"); 3737 3738 for (i = 0; i < n_chosen; i += 1) 3739 syms[i] = syms[chosen[i]]; 3740 3741 return n_chosen; 3742 } 3743 3744 /* See ada-lang.h. */ 3745 3746 block_symbol 3747 ada_find_operator_symbol (enum exp_opcode op, bool parse_completion, 3748 int nargs, value *argvec[]) 3749 { 3750 if (possible_user_operator_p (op, argvec)) 3751 { 3752 std::vector<struct block_symbol> candidates 3753 = ada_lookup_symbol_list (ada_decoded_op_name (op), 3754 NULL, VAR_DOMAIN); 3755 3756 int i = ada_resolve_function (candidates, argvec, 3757 nargs, ada_decoded_op_name (op), NULL, 3758 parse_completion); 3759 if (i >= 0) 3760 return candidates[i]; 3761 } 3762 return {}; 3763 } 3764 3765 /* See ada-lang.h. */ 3766 3767 block_symbol 3768 ada_resolve_funcall (struct symbol *sym, const struct block *block, 3769 struct type *context_type, 3770 bool parse_completion, 3771 int nargs, value *argvec[], 3772 innermost_block_tracker *tracker) 3773 { 3774 std::vector<struct block_symbol> candidates 3775 = ada_lookup_symbol_list (sym->linkage_name (), block, VAR_DOMAIN); 3776 3777 int i; 3778 if (candidates.size () == 1) 3779 i = 0; 3780 else 3781 { 3782 i = ada_resolve_function 3783 (candidates, 3784 argvec, nargs, 3785 sym->linkage_name (), 3786 context_type, parse_completion); 3787 if (i < 0) 3788 error (_("Could not find a match for %s"), sym->print_name ()); 3789 } 3790 3791 tracker->update (candidates[i]); 3792 return candidates[i]; 3793 } 3794 3795 /* Resolve a mention of a name where the context type is an 3796 enumeration type. */ 3797 3798 static int 3799 ada_resolve_enum (std::vector<struct block_symbol> &syms, 3800 const char *name, struct type *context_type, 3801 bool parse_completion) 3802 { 3803 gdb_assert (context_type->code () == TYPE_CODE_ENUM); 3804 context_type = ada_check_typedef (context_type); 3805 3806 for (int i = 0; i < syms.size (); ++i) 3807 { 3808 /* We already know the name matches, so we're just looking for 3809 an element of the correct enum type. */ 3810 if (ada_check_typedef (syms[i].symbol->type ()) == context_type) 3811 return i; 3812 } 3813 3814 error (_("No name '%s' in enumeration type '%s'"), name, 3815 ada_type_name (context_type)); 3816 } 3817 3818 /* See ada-lang.h. */ 3819 3820 block_symbol 3821 ada_resolve_variable (struct symbol *sym, const struct block *block, 3822 struct type *context_type, 3823 bool parse_completion, 3824 int deprocedure_p, 3825 innermost_block_tracker *tracker) 3826 { 3827 std::vector<struct block_symbol> candidates 3828 = ada_lookup_symbol_list (sym->linkage_name (), block, VAR_DOMAIN); 3829 3830 if (std::any_of (candidates.begin (), 3831 candidates.end (), 3832 [] (block_symbol &bsym) 3833 { 3834 switch (bsym.symbol->aclass ()) 3835 { 3836 case LOC_REGISTER: 3837 case LOC_ARG: 3838 case LOC_REF_ARG: 3839 case LOC_REGPARM_ADDR: 3840 case LOC_LOCAL: 3841 case LOC_COMPUTED: 3842 return true; 3843 default: 3844 return false; 3845 } 3846 })) 3847 { 3848 /* Types tend to get re-introduced locally, so if there 3849 are any local symbols that are not types, first filter 3850 out all types. */ 3851 candidates.erase 3852 (std::remove_if 3853 (candidates.begin (), 3854 candidates.end (), 3855 [] (block_symbol &bsym) 3856 { 3857 return bsym.symbol->aclass () == LOC_TYPEDEF; 3858 }), 3859 candidates.end ()); 3860 } 3861 3862 /* Filter out artificial symbols. */ 3863 candidates.erase 3864 (std::remove_if 3865 (candidates.begin (), 3866 candidates.end (), 3867 [] (block_symbol &bsym) 3868 { 3869 return bsym.symbol->is_artificial (); 3870 }), 3871 candidates.end ()); 3872 3873 int i; 3874 if (candidates.empty ()) 3875 error (_("No definition found for %s"), sym->print_name ()); 3876 else if (candidates.size () == 1) 3877 i = 0; 3878 else if (context_type != nullptr 3879 && context_type->code () == TYPE_CODE_ENUM) 3880 i = ada_resolve_enum (candidates, sym->linkage_name (), context_type, 3881 parse_completion); 3882 else if (deprocedure_p && !is_nonfunction (candidates)) 3883 { 3884 i = ada_resolve_function 3885 (candidates, NULL, 0, 3886 sym->linkage_name (), 3887 context_type, parse_completion); 3888 if (i < 0) 3889 error (_("Could not find a match for %s"), sym->print_name ()); 3890 } 3891 else 3892 { 3893 gdb_printf (_("Multiple matches for %s\n"), sym->print_name ()); 3894 user_select_syms (candidates.data (), candidates.size (), 1); 3895 i = 0; 3896 } 3897 3898 tracker->update (candidates[i]); 3899 return candidates[i]; 3900 } 3901 3902 /* Return non-zero if formal type FTYPE matches actual type ATYPE. */ 3903 /* The term "match" here is rather loose. The match is heuristic and 3904 liberal. */ 3905 3906 static int 3907 ada_type_match (struct type *ftype, struct type *atype) 3908 { 3909 ftype = ada_check_typedef (ftype); 3910 atype = ada_check_typedef (atype); 3911 3912 if (ftype->code () == TYPE_CODE_REF) 3913 ftype = ftype->target_type (); 3914 if (atype->code () == TYPE_CODE_REF) 3915 atype = atype->target_type (); 3916 3917 switch (ftype->code ()) 3918 { 3919 default: 3920 return ftype->code () == atype->code (); 3921 case TYPE_CODE_PTR: 3922 if (atype->code () != TYPE_CODE_PTR) 3923 return 0; 3924 atype = atype->target_type (); 3925 /* This can only happen if the actual argument is 'null'. */ 3926 if (atype->code () == TYPE_CODE_INT && atype->length () == 0) 3927 return 1; 3928 return ada_type_match (ftype->target_type (), atype); 3929 case TYPE_CODE_INT: 3930 case TYPE_CODE_ENUM: 3931 case TYPE_CODE_RANGE: 3932 switch (atype->code ()) 3933 { 3934 case TYPE_CODE_INT: 3935 case TYPE_CODE_ENUM: 3936 case TYPE_CODE_RANGE: 3937 return 1; 3938 default: 3939 return 0; 3940 } 3941 3942 case TYPE_CODE_ARRAY: 3943 return (atype->code () == TYPE_CODE_ARRAY 3944 || ada_is_array_descriptor_type (atype)); 3945 3946 case TYPE_CODE_STRUCT: 3947 if (ada_is_array_descriptor_type (ftype)) 3948 return (atype->code () == TYPE_CODE_ARRAY 3949 || ada_is_array_descriptor_type (atype)); 3950 else 3951 return (atype->code () == TYPE_CODE_STRUCT 3952 && !ada_is_array_descriptor_type (atype)); 3953 3954 case TYPE_CODE_UNION: 3955 case TYPE_CODE_FLT: 3956 return (atype->code () == ftype->code ()); 3957 } 3958 } 3959 3960 /* Return non-zero if the formals of FUNC "sufficiently match" the 3961 vector of actual argument types ACTUALS of size N_ACTUALS. FUNC 3962 may also be an enumeral, in which case it is treated as a 0- 3963 argument function. */ 3964 3965 static int 3966 ada_args_match (struct symbol *func, struct value **actuals, int n_actuals) 3967 { 3968 int i; 3969 struct type *func_type = func->type (); 3970 3971 if (func->aclass () == LOC_CONST 3972 && func_type->code () == TYPE_CODE_ENUM) 3973 return (n_actuals == 0); 3974 else if (func_type == NULL || func_type->code () != TYPE_CODE_FUNC) 3975 return 0; 3976 3977 if (func_type->num_fields () != n_actuals) 3978 return 0; 3979 3980 for (i = 0; i < n_actuals; i += 1) 3981 { 3982 if (actuals[i] == NULL) 3983 return 0; 3984 else 3985 { 3986 struct type *ftype = ada_check_typedef (func_type->field (i).type ()); 3987 struct type *atype = ada_check_typedef (value_type (actuals[i])); 3988 3989 if (!ada_type_match (ftype, atype)) 3990 return 0; 3991 } 3992 } 3993 return 1; 3994 } 3995 3996 /* False iff function type FUNC_TYPE definitely does not produce a value 3997 compatible with type CONTEXT_TYPE. Conservatively returns 1 if 3998 FUNC_TYPE is not a valid function type with a non-null return type 3999 or an enumerated type. A null CONTEXT_TYPE indicates any non-void type. */ 4000 4001 static int 4002 return_match (struct type *func_type, struct type *context_type) 4003 { 4004 struct type *return_type; 4005 4006 if (func_type == NULL) 4007 return 1; 4008 4009 if (func_type->code () == TYPE_CODE_FUNC) 4010 return_type = get_base_type (func_type->target_type ()); 4011 else 4012 return_type = get_base_type (func_type); 4013 if (return_type == NULL) 4014 return 1; 4015 4016 context_type = get_base_type (context_type); 4017 4018 if (return_type->code () == TYPE_CODE_ENUM) 4019 return context_type == NULL || return_type == context_type; 4020 else if (context_type == NULL) 4021 return return_type->code () != TYPE_CODE_VOID; 4022 else 4023 return return_type->code () == context_type->code (); 4024 } 4025 4026 4027 /* Returns the index in SYMS that contains the symbol for the 4028 function (if any) that matches the types of the NARGS arguments in 4029 ARGS. If CONTEXT_TYPE is non-null and there is at least one match 4030 that returns that type, then eliminate matches that don't. If 4031 CONTEXT_TYPE is void and there is at least one match that does not 4032 return void, eliminate all matches that do. 4033 4034 Asks the user if there is more than one match remaining. Returns -1 4035 if there is no such symbol or none is selected. NAME is used 4036 solely for messages. May re-arrange and modify SYMS in 4037 the process; the index returned is for the modified vector. */ 4038 4039 static int 4040 ada_resolve_function (std::vector<struct block_symbol> &syms, 4041 struct value **args, int nargs, 4042 const char *name, struct type *context_type, 4043 bool parse_completion) 4044 { 4045 int fallback; 4046 int k; 4047 int m; /* Number of hits */ 4048 4049 m = 0; 4050 /* In the first pass of the loop, we only accept functions matching 4051 context_type. If none are found, we add a second pass of the loop 4052 where every function is accepted. */ 4053 for (fallback = 0; m == 0 && fallback < 2; fallback++) 4054 { 4055 for (k = 0; k < syms.size (); k += 1) 4056 { 4057 struct type *type = ada_check_typedef (syms[k].symbol->type ()); 4058 4059 if (ada_args_match (syms[k].symbol, args, nargs) 4060 && (fallback || return_match (type, context_type))) 4061 { 4062 syms[m] = syms[k]; 4063 m += 1; 4064 } 4065 } 4066 } 4067 4068 /* If we got multiple matches, ask the user which one to use. Don't do this 4069 interactive thing during completion, though, as the purpose of the 4070 completion is providing a list of all possible matches. Prompting the 4071 user to filter it down would be completely unexpected in this case. */ 4072 if (m == 0) 4073 return -1; 4074 else if (m > 1 && !parse_completion) 4075 { 4076 gdb_printf (_("Multiple matches for %s\n"), name); 4077 user_select_syms (syms.data (), m, 1); 4078 return 0; 4079 } 4080 return 0; 4081 } 4082 4083 /* Type-class predicates */ 4084 4085 /* True iff TYPE is numeric (i.e., an INT, RANGE (of numeric type), 4086 or FLOAT). */ 4087 4088 static int 4089 numeric_type_p (struct type *type) 4090 { 4091 if (type == NULL) 4092 return 0; 4093 else 4094 { 4095 switch (type->code ()) 4096 { 4097 case TYPE_CODE_INT: 4098 case TYPE_CODE_FLT: 4099 case TYPE_CODE_FIXED_POINT: 4100 return 1; 4101 case TYPE_CODE_RANGE: 4102 return (type == type->target_type () 4103 || numeric_type_p (type->target_type ())); 4104 default: 4105 return 0; 4106 } 4107 } 4108 } 4109 4110 /* True iff TYPE is integral (an INT or RANGE of INTs). */ 4111 4112 static int 4113 integer_type_p (struct type *type) 4114 { 4115 if (type == NULL) 4116 return 0; 4117 else 4118 { 4119 switch (type->code ()) 4120 { 4121 case TYPE_CODE_INT: 4122 return 1; 4123 case TYPE_CODE_RANGE: 4124 return (type == type->target_type () 4125 || integer_type_p (type->target_type ())); 4126 default: 4127 return 0; 4128 } 4129 } 4130 } 4131 4132 /* True iff TYPE is scalar (INT, RANGE, FLOAT, ENUM). */ 4133 4134 static int 4135 scalar_type_p (struct type *type) 4136 { 4137 if (type == NULL) 4138 return 0; 4139 else 4140 { 4141 switch (type->code ()) 4142 { 4143 case TYPE_CODE_INT: 4144 case TYPE_CODE_RANGE: 4145 case TYPE_CODE_ENUM: 4146 case TYPE_CODE_FLT: 4147 case TYPE_CODE_FIXED_POINT: 4148 return 1; 4149 default: 4150 return 0; 4151 } 4152 } 4153 } 4154 4155 /* True iff TYPE is discrete, as defined in the Ada Reference Manual. 4156 This essentially means one of (INT, RANGE, ENUM) -- but note that 4157 "enum" includes character and boolean as well. */ 4158 4159 static int 4160 discrete_type_p (struct type *type) 4161 { 4162 if (type == NULL) 4163 return 0; 4164 else 4165 { 4166 switch (type->code ()) 4167 { 4168 case TYPE_CODE_INT: 4169 case TYPE_CODE_RANGE: 4170 case TYPE_CODE_ENUM: 4171 case TYPE_CODE_BOOL: 4172 case TYPE_CODE_CHAR: 4173 return 1; 4174 default: 4175 return 0; 4176 } 4177 } 4178 } 4179 4180 /* Returns non-zero if OP with operands in the vector ARGS could be 4181 a user-defined function. Errs on the side of pre-defined operators 4182 (i.e., result 0). */ 4183 4184 static int 4185 possible_user_operator_p (enum exp_opcode op, struct value *args[]) 4186 { 4187 struct type *type0 = 4188 (args[0] == NULL) ? NULL : ada_check_typedef (value_type (args[0])); 4189 struct type *type1 = 4190 (args[1] == NULL) ? NULL : ada_check_typedef (value_type (args[1])); 4191 4192 if (type0 == NULL) 4193 return 0; 4194 4195 switch (op) 4196 { 4197 default: 4198 return 0; 4199 4200 case BINOP_ADD: 4201 case BINOP_SUB: 4202 case BINOP_MUL: 4203 case BINOP_DIV: 4204 return (!(numeric_type_p (type0) && numeric_type_p (type1))); 4205 4206 case BINOP_REM: 4207 case BINOP_MOD: 4208 case BINOP_BITWISE_AND: 4209 case BINOP_BITWISE_IOR: 4210 case BINOP_BITWISE_XOR: 4211 return (!(integer_type_p (type0) && integer_type_p (type1))); 4212 4213 case BINOP_EQUAL: 4214 case BINOP_NOTEQUAL: 4215 case BINOP_LESS: 4216 case BINOP_GTR: 4217 case BINOP_LEQ: 4218 case BINOP_GEQ: 4219 return (!(scalar_type_p (type0) && scalar_type_p (type1))); 4220 4221 case BINOP_CONCAT: 4222 return !ada_is_array_type (type0) || !ada_is_array_type (type1); 4223 4224 case BINOP_EXP: 4225 return (!(numeric_type_p (type0) && integer_type_p (type1))); 4226 4227 case UNOP_NEG: 4228 case UNOP_PLUS: 4229 case UNOP_LOGICAL_NOT: 4230 case UNOP_ABS: 4231 return (!numeric_type_p (type0)); 4232 4233 } 4234 } 4235 4236 /* Renaming */ 4237 4238 /* NOTES: 4239 4240 1. In the following, we assume that a renaming type's name may 4241 have an ___XD suffix. It would be nice if this went away at some 4242 point. 4243 2. We handle both the (old) purely type-based representation of 4244 renamings and the (new) variable-based encoding. At some point, 4245 it is devoutly to be hoped that the former goes away 4246 (FIXME: hilfinger-2007-07-09). 4247 3. Subprogram renamings are not implemented, although the XRS 4248 suffix is recognized (FIXME: hilfinger-2007-07-09). */ 4249 4250 /* If SYM encodes a renaming, 4251 4252 <renaming> renames <renamed entity>, 4253 4254 sets *LEN to the length of the renamed entity's name, 4255 *RENAMED_ENTITY to that name (not null-terminated), and *RENAMING_EXPR to 4256 the string describing the subcomponent selected from the renamed 4257 entity. Returns ADA_NOT_RENAMING if SYM does not encode a renaming 4258 (in which case, the values of *RENAMED_ENTITY, *LEN, and *RENAMING_EXPR 4259 are undefined). Otherwise, returns a value indicating the category 4260 of entity renamed: an object (ADA_OBJECT_RENAMING), exception 4261 (ADA_EXCEPTION_RENAMING), package (ADA_PACKAGE_RENAMING), or 4262 subprogram (ADA_SUBPROGRAM_RENAMING). Does no allocation; the 4263 strings returned in *RENAMED_ENTITY and *RENAMING_EXPR should not be 4264 deallocated. The values of RENAMED_ENTITY, LEN, or RENAMING_EXPR 4265 may be NULL, in which case they are not assigned. 4266 4267 [Currently, however, GCC does not generate subprogram renamings.] */ 4268 4269 enum ada_renaming_category 4270 ada_parse_renaming (struct symbol *sym, 4271 const char **renamed_entity, int *len, 4272 const char **renaming_expr) 4273 { 4274 enum ada_renaming_category kind; 4275 const char *info; 4276 const char *suffix; 4277 4278 if (sym == NULL) 4279 return ADA_NOT_RENAMING; 4280 switch (sym->aclass ()) 4281 { 4282 default: 4283 return ADA_NOT_RENAMING; 4284 case LOC_LOCAL: 4285 case LOC_STATIC: 4286 case LOC_COMPUTED: 4287 case LOC_OPTIMIZED_OUT: 4288 info = strstr (sym->linkage_name (), "___XR"); 4289 if (info == NULL) 4290 return ADA_NOT_RENAMING; 4291 switch (info[5]) 4292 { 4293 case '_': 4294 kind = ADA_OBJECT_RENAMING; 4295 info += 6; 4296 break; 4297 case 'E': 4298 kind = ADA_EXCEPTION_RENAMING; 4299 info += 7; 4300 break; 4301 case 'P': 4302 kind = ADA_PACKAGE_RENAMING; 4303 info += 7; 4304 break; 4305 case 'S': 4306 kind = ADA_SUBPROGRAM_RENAMING; 4307 info += 7; 4308 break; 4309 default: 4310 return ADA_NOT_RENAMING; 4311 } 4312 } 4313 4314 if (renamed_entity != NULL) 4315 *renamed_entity = info; 4316 suffix = strstr (info, "___XE"); 4317 if (suffix == NULL || suffix == info) 4318 return ADA_NOT_RENAMING; 4319 if (len != NULL) 4320 *len = strlen (info) - strlen (suffix); 4321 suffix += 5; 4322 if (renaming_expr != NULL) 4323 *renaming_expr = suffix; 4324 return kind; 4325 } 4326 4327 /* Compute the value of the given RENAMING_SYM, which is expected to 4328 be a symbol encoding a renaming expression. BLOCK is the block 4329 used to evaluate the renaming. */ 4330 4331 static struct value * 4332 ada_read_renaming_var_value (struct symbol *renaming_sym, 4333 const struct block *block) 4334 { 4335 const char *sym_name; 4336 4337 sym_name = renaming_sym->linkage_name (); 4338 expression_up expr = parse_exp_1 (&sym_name, 0, block, 0); 4339 return evaluate_expression (expr.get ()); 4340 } 4341 4342 4343 /* Evaluation: Function Calls */ 4344 4345 /* Return an lvalue containing the value VAL. This is the identity on 4346 lvalues, and otherwise has the side-effect of allocating memory 4347 in the inferior where a copy of the value contents is copied. */ 4348 4349 static struct value * 4350 ensure_lval (struct value *val) 4351 { 4352 if (VALUE_LVAL (val) == not_lval 4353 || VALUE_LVAL (val) == lval_internalvar) 4354 { 4355 int len = ada_check_typedef (value_type (val))->length (); 4356 const CORE_ADDR addr = 4357 value_as_long (value_allocate_space_in_inferior (len)); 4358 4359 VALUE_LVAL (val) = lval_memory; 4360 set_value_address (val, addr); 4361 write_memory (addr, value_contents (val).data (), len); 4362 } 4363 4364 return val; 4365 } 4366 4367 /* Given ARG, a value of type (pointer or reference to a)* 4368 structure/union, extract the component named NAME from the ultimate 4369 target structure/union and return it as a value with its 4370 appropriate type. 4371 4372 The routine searches for NAME among all members of the structure itself 4373 and (recursively) among all members of any wrapper members 4374 (e.g., '_parent'). 4375 4376 If NO_ERR, then simply return NULL in case of error, rather than 4377 calling error. */ 4378 4379 static struct value * 4380 ada_value_struct_elt (struct value *arg, const char *name, int no_err) 4381 { 4382 struct type *t, *t1; 4383 struct value *v; 4384 int check_tag; 4385 4386 v = NULL; 4387 t1 = t = ada_check_typedef (value_type (arg)); 4388 if (t->code () == TYPE_CODE_REF) 4389 { 4390 t1 = t->target_type (); 4391 if (t1 == NULL) 4392 goto BadValue; 4393 t1 = ada_check_typedef (t1); 4394 if (t1->code () == TYPE_CODE_PTR) 4395 { 4396 arg = coerce_ref (arg); 4397 t = t1; 4398 } 4399 } 4400 4401 while (t->code () == TYPE_CODE_PTR) 4402 { 4403 t1 = t->target_type (); 4404 if (t1 == NULL) 4405 goto BadValue; 4406 t1 = ada_check_typedef (t1); 4407 if (t1->code () == TYPE_CODE_PTR) 4408 { 4409 arg = value_ind (arg); 4410 t = t1; 4411 } 4412 else 4413 break; 4414 } 4415 4416 if (t1->code () != TYPE_CODE_STRUCT && t1->code () != TYPE_CODE_UNION) 4417 goto BadValue; 4418 4419 if (t1 == t) 4420 v = ada_search_struct_field (name, arg, 0, t); 4421 else 4422 { 4423 int bit_offset, bit_size, byte_offset; 4424 struct type *field_type; 4425 CORE_ADDR address; 4426 4427 if (t->code () == TYPE_CODE_PTR) 4428 address = value_address (ada_value_ind (arg)); 4429 else 4430 address = value_address (ada_coerce_ref (arg)); 4431 4432 /* Check to see if this is a tagged type. We also need to handle 4433 the case where the type is a reference to a tagged type, but 4434 we have to be careful to exclude pointers to tagged types. 4435 The latter should be shown as usual (as a pointer), whereas 4436 a reference should mostly be transparent to the user. */ 4437 4438 if (ada_is_tagged_type (t1, 0) 4439 || (t1->code () == TYPE_CODE_REF 4440 && ada_is_tagged_type (t1->target_type (), 0))) 4441 { 4442 /* We first try to find the searched field in the current type. 4443 If not found then let's look in the fixed type. */ 4444 4445 if (!find_struct_field (name, t1, 0, 4446 nullptr, nullptr, nullptr, 4447 nullptr, nullptr)) 4448 check_tag = 1; 4449 else 4450 check_tag = 0; 4451 } 4452 else 4453 check_tag = 0; 4454 4455 /* Convert to fixed type in all cases, so that we have proper 4456 offsets to each field in unconstrained record types. */ 4457 t1 = ada_to_fixed_type (ada_get_base_type (t1), NULL, 4458 address, NULL, check_tag); 4459 4460 /* Resolve the dynamic type as well. */ 4461 arg = value_from_contents_and_address (t1, nullptr, address); 4462 t1 = value_type (arg); 4463 4464 if (find_struct_field (name, t1, 0, 4465 &field_type, &byte_offset, &bit_offset, 4466 &bit_size, NULL)) 4467 { 4468 if (bit_size != 0) 4469 { 4470 if (t->code () == TYPE_CODE_REF) 4471 arg = ada_coerce_ref (arg); 4472 else 4473 arg = ada_value_ind (arg); 4474 v = ada_value_primitive_packed_val (arg, NULL, byte_offset, 4475 bit_offset, bit_size, 4476 field_type); 4477 } 4478 else 4479 v = value_at_lazy (field_type, address + byte_offset); 4480 } 4481 } 4482 4483 if (v != NULL || no_err) 4484 return v; 4485 else 4486 error (_("There is no member named %s."), name); 4487 4488 BadValue: 4489 if (no_err) 4490 return NULL; 4491 else 4492 error (_("Attempt to extract a component of " 4493 "a value that is not a record.")); 4494 } 4495 4496 /* Return the value ACTUAL, converted to be an appropriate value for a 4497 formal of type FORMAL_TYPE. Use *SP as a stack pointer for 4498 allocating any necessary descriptors (fat pointers), or copies of 4499 values not residing in memory, updating it as needed. */ 4500 4501 struct value * 4502 ada_convert_actual (struct value *actual, struct type *formal_type0) 4503 { 4504 struct type *actual_type = ada_check_typedef (value_type (actual)); 4505 struct type *formal_type = ada_check_typedef (formal_type0); 4506 struct type *formal_target = 4507 formal_type->code () == TYPE_CODE_PTR 4508 ? ada_check_typedef (formal_type->target_type ()) : formal_type; 4509 struct type *actual_target = 4510 actual_type->code () == TYPE_CODE_PTR 4511 ? ada_check_typedef (actual_type->target_type ()) : actual_type; 4512 4513 if (ada_is_array_descriptor_type (formal_target) 4514 && actual_target->code () == TYPE_CODE_ARRAY) 4515 return make_array_descriptor (formal_type, actual); 4516 else if (formal_type->code () == TYPE_CODE_PTR 4517 || formal_type->code () == TYPE_CODE_REF) 4518 { 4519 struct value *result; 4520 4521 if (formal_target->code () == TYPE_CODE_ARRAY 4522 && ada_is_array_descriptor_type (actual_target)) 4523 result = desc_data (actual); 4524 else if (formal_type->code () != TYPE_CODE_PTR) 4525 { 4526 if (VALUE_LVAL (actual) != lval_memory) 4527 { 4528 struct value *val; 4529 4530 actual_type = ada_check_typedef (value_type (actual)); 4531 val = allocate_value (actual_type); 4532 copy (value_contents (actual), value_contents_raw (val)); 4533 actual = ensure_lval (val); 4534 } 4535 result = value_addr (actual); 4536 } 4537 else 4538 return actual; 4539 return value_cast_pointers (formal_type, result, 0); 4540 } 4541 else if (actual_type->code () == TYPE_CODE_PTR) 4542 return ada_value_ind (actual); 4543 else if (ada_is_aligner_type (formal_type)) 4544 { 4545 /* We need to turn this parameter into an aligner type 4546 as well. */ 4547 struct value *aligner = allocate_value (formal_type); 4548 struct value *component = ada_value_struct_elt (aligner, "F", 0); 4549 4550 value_assign_to_component (aligner, component, actual); 4551 return aligner; 4552 } 4553 4554 return actual; 4555 } 4556 4557 /* Convert VALUE (which must be an address) to a CORE_ADDR that is a pointer of 4558 type TYPE. This is usually an inefficient no-op except on some targets 4559 (such as AVR) where the representation of a pointer and an address 4560 differs. */ 4561 4562 static CORE_ADDR 4563 value_pointer (struct value *value, struct type *type) 4564 { 4565 unsigned len = type->length (); 4566 gdb_byte *buf = (gdb_byte *) alloca (len); 4567 CORE_ADDR addr; 4568 4569 addr = value_address (value); 4570 gdbarch_address_to_pointer (type->arch (), type, buf, addr); 4571 addr = extract_unsigned_integer (buf, len, type_byte_order (type)); 4572 return addr; 4573 } 4574 4575 4576 /* Push a descriptor of type TYPE for array value ARR on the stack at 4577 *SP, updating *SP to reflect the new descriptor. Return either 4578 an lvalue representing the new descriptor, or (if TYPE is a pointer- 4579 to-descriptor type rather than a descriptor type), a struct value * 4580 representing a pointer to this descriptor. */ 4581 4582 static struct value * 4583 make_array_descriptor (struct type *type, struct value *arr) 4584 { 4585 struct type *bounds_type = desc_bounds_type (type); 4586 struct type *desc_type = desc_base_type (type); 4587 struct value *descriptor = allocate_value (desc_type); 4588 struct value *bounds = allocate_value (bounds_type); 4589 int i; 4590 4591 for (i = ada_array_arity (ada_check_typedef (value_type (arr))); 4592 i > 0; i -= 1) 4593 { 4594 modify_field (value_type (bounds), 4595 value_contents_writeable (bounds).data (), 4596 ada_array_bound (arr, i, 0), 4597 desc_bound_bitpos (bounds_type, i, 0), 4598 desc_bound_bitsize (bounds_type, i, 0)); 4599 modify_field (value_type (bounds), 4600 value_contents_writeable (bounds).data (), 4601 ada_array_bound (arr, i, 1), 4602 desc_bound_bitpos (bounds_type, i, 1), 4603 desc_bound_bitsize (bounds_type, i, 1)); 4604 } 4605 4606 bounds = ensure_lval (bounds); 4607 4608 modify_field (value_type (descriptor), 4609 value_contents_writeable (descriptor).data (), 4610 value_pointer (ensure_lval (arr), 4611 desc_type->field (0).type ()), 4612 fat_pntr_data_bitpos (desc_type), 4613 fat_pntr_data_bitsize (desc_type)); 4614 4615 modify_field (value_type (descriptor), 4616 value_contents_writeable (descriptor).data (), 4617 value_pointer (bounds, 4618 desc_type->field (1).type ()), 4619 fat_pntr_bounds_bitpos (desc_type), 4620 fat_pntr_bounds_bitsize (desc_type)); 4621 4622 descriptor = ensure_lval (descriptor); 4623 4624 if (type->code () == TYPE_CODE_PTR) 4625 return value_addr (descriptor); 4626 else 4627 return descriptor; 4628 } 4629 4630 /* Symbol Cache Module */ 4631 4632 /* Performance measurements made as of 2010-01-15 indicate that 4633 this cache does bring some noticeable improvements. Depending 4634 on the type of entity being printed, the cache can make it as much 4635 as an order of magnitude faster than without it. 4636 4637 The descriptive type DWARF extension has significantly reduced 4638 the need for this cache, at least when DWARF is being used. However, 4639 even in this case, some expensive name-based symbol searches are still 4640 sometimes necessary - to find an XVZ variable, mostly. */ 4641 4642 /* Return the symbol cache associated to the given program space PSPACE. 4643 If not allocated for this PSPACE yet, allocate and initialize one. */ 4644 4645 static struct ada_symbol_cache * 4646 ada_get_symbol_cache (struct program_space *pspace) 4647 { 4648 struct ada_pspace_data *pspace_data = get_ada_pspace_data (pspace); 4649 4650 if (pspace_data->sym_cache == nullptr) 4651 pspace_data->sym_cache.reset (new ada_symbol_cache); 4652 4653 return pspace_data->sym_cache.get (); 4654 } 4655 4656 /* Clear all entries from the symbol cache. */ 4657 4658 static void 4659 ada_clear_symbol_cache () 4660 { 4661 struct ada_pspace_data *pspace_data 4662 = get_ada_pspace_data (current_program_space); 4663 4664 if (pspace_data->sym_cache != nullptr) 4665 pspace_data->sym_cache.reset (); 4666 } 4667 4668 /* Search our cache for an entry matching NAME and DOMAIN. 4669 Return it if found, or NULL otherwise. */ 4670 4671 static struct cache_entry ** 4672 find_entry (const char *name, domain_enum domain) 4673 { 4674 struct ada_symbol_cache *sym_cache 4675 = ada_get_symbol_cache (current_program_space); 4676 int h = msymbol_hash (name) % HASH_SIZE; 4677 struct cache_entry **e; 4678 4679 for (e = &sym_cache->root[h]; *e != NULL; e = &(*e)->next) 4680 { 4681 if (domain == (*e)->domain && strcmp (name, (*e)->name) == 0) 4682 return e; 4683 } 4684 return NULL; 4685 } 4686 4687 /* Search the symbol cache for an entry matching NAME and DOMAIN. 4688 Return 1 if found, 0 otherwise. 4689 4690 If an entry was found and SYM is not NULL, set *SYM to the entry's 4691 SYM. Same principle for BLOCK if not NULL. */ 4692 4693 static int 4694 lookup_cached_symbol (const char *name, domain_enum domain, 4695 struct symbol **sym, const struct block **block) 4696 { 4697 struct cache_entry **e = find_entry (name, domain); 4698 4699 if (e == NULL) 4700 return 0; 4701 if (sym != NULL) 4702 *sym = (*e)->sym; 4703 if (block != NULL) 4704 *block = (*e)->block; 4705 return 1; 4706 } 4707 4708 /* Assuming that (SYM, BLOCK) is the result of the lookup of NAME 4709 in domain DOMAIN, save this result in our symbol cache. */ 4710 4711 static void 4712 cache_symbol (const char *name, domain_enum domain, struct symbol *sym, 4713 const struct block *block) 4714 { 4715 struct ada_symbol_cache *sym_cache 4716 = ada_get_symbol_cache (current_program_space); 4717 int h; 4718 struct cache_entry *e; 4719 4720 /* Symbols for builtin types don't have a block. 4721 For now don't cache such symbols. */ 4722 if (sym != NULL && !sym->is_objfile_owned ()) 4723 return; 4724 4725 /* If the symbol is a local symbol, then do not cache it, as a search 4726 for that symbol depends on the context. To determine whether 4727 the symbol is local or not, we check the block where we found it 4728 against the global and static blocks of its associated symtab. */ 4729 if (sym != nullptr) 4730 { 4731 const blockvector &bv = *sym->symtab ()->compunit ()->blockvector (); 4732 4733 if (bv.global_block () != block && bv.static_block () != block) 4734 return; 4735 } 4736 4737 h = msymbol_hash (name) % HASH_SIZE; 4738 e = XOBNEW (&sym_cache->cache_space, cache_entry); 4739 e->next = sym_cache->root[h]; 4740 sym_cache->root[h] = e; 4741 e->name = obstack_strdup (&sym_cache->cache_space, name); 4742 e->sym = sym; 4743 e->domain = domain; 4744 e->block = block; 4745 } 4746 4747 /* Symbol Lookup */ 4748 4749 /* Return the symbol name match type that should be used used when 4750 searching for all symbols matching LOOKUP_NAME. 4751 4752 LOOKUP_NAME is expected to be a symbol name after transformation 4753 for Ada lookups. */ 4754 4755 static symbol_name_match_type 4756 name_match_type_from_name (const char *lookup_name) 4757 { 4758 return (strstr (lookup_name, "__") == NULL 4759 ? symbol_name_match_type::WILD 4760 : symbol_name_match_type::FULL); 4761 } 4762 4763 /* Return the result of a standard (literal, C-like) lookup of NAME in 4764 given DOMAIN, visible from lexical block BLOCK. */ 4765 4766 static struct symbol * 4767 standard_lookup (const char *name, const struct block *block, 4768 domain_enum domain) 4769 { 4770 /* Initialize it just to avoid a GCC false warning. */ 4771 struct block_symbol sym = {}; 4772 4773 if (lookup_cached_symbol (name, domain, &sym.symbol, NULL)) 4774 return sym.symbol; 4775 ada_lookup_encoded_symbol (name, block, domain, &sym); 4776 cache_symbol (name, domain, sym.symbol, sym.block); 4777 return sym.symbol; 4778 } 4779 4780 4781 /* Non-zero iff there is at least one non-function/non-enumeral symbol 4782 in the symbol fields of SYMS. We treat enumerals as functions, 4783 since they contend in overloading in the same way. */ 4784 static int 4785 is_nonfunction (const std::vector<struct block_symbol> &syms) 4786 { 4787 for (const block_symbol &sym : syms) 4788 if (sym.symbol->type ()->code () != TYPE_CODE_FUNC 4789 && (sym.symbol->type ()->code () != TYPE_CODE_ENUM 4790 || sym.symbol->aclass () != LOC_CONST)) 4791 return 1; 4792 4793 return 0; 4794 } 4795 4796 /* If true (non-zero), then TYPE0 and TYPE1 represent equivalent 4797 struct types. Otherwise, they may not. */ 4798 4799 static int 4800 equiv_types (struct type *type0, struct type *type1) 4801 { 4802 if (type0 == type1) 4803 return 1; 4804 if (type0 == NULL || type1 == NULL 4805 || type0->code () != type1->code ()) 4806 return 0; 4807 if ((type0->code () == TYPE_CODE_STRUCT 4808 || type0->code () == TYPE_CODE_ENUM) 4809 && ada_type_name (type0) != NULL && ada_type_name (type1) != NULL 4810 && strcmp (ada_type_name (type0), ada_type_name (type1)) == 0) 4811 return 1; 4812 4813 return 0; 4814 } 4815 4816 /* True iff SYM0 represents the same entity as SYM1, or one that is 4817 no more defined than that of SYM1. */ 4818 4819 static int 4820 lesseq_defined_than (struct symbol *sym0, struct symbol *sym1) 4821 { 4822 if (sym0 == sym1) 4823 return 1; 4824 if (sym0->domain () != sym1->domain () 4825 || sym0->aclass () != sym1->aclass ()) 4826 return 0; 4827 4828 switch (sym0->aclass ()) 4829 { 4830 case LOC_UNDEF: 4831 return 1; 4832 case LOC_TYPEDEF: 4833 { 4834 struct type *type0 = sym0->type (); 4835 struct type *type1 = sym1->type (); 4836 const char *name0 = sym0->linkage_name (); 4837 const char *name1 = sym1->linkage_name (); 4838 int len0 = strlen (name0); 4839 4840 return 4841 type0->code () == type1->code () 4842 && (equiv_types (type0, type1) 4843 || (len0 < strlen (name1) && strncmp (name0, name1, len0) == 0 4844 && startswith (name1 + len0, "___XV"))); 4845 } 4846 case LOC_CONST: 4847 return sym0->value_longest () == sym1->value_longest () 4848 && equiv_types (sym0->type (), sym1->type ()); 4849 4850 case LOC_STATIC: 4851 { 4852 const char *name0 = sym0->linkage_name (); 4853 const char *name1 = sym1->linkage_name (); 4854 return (strcmp (name0, name1) == 0 4855 && sym0->value_address () == sym1->value_address ()); 4856 } 4857 4858 default: 4859 return 0; 4860 } 4861 } 4862 4863 /* Append (SYM,BLOCK) to the end of the array of struct block_symbol 4864 records in RESULT. Do nothing if SYM is a duplicate. */ 4865 4866 static void 4867 add_defn_to_vec (std::vector<struct block_symbol> &result, 4868 struct symbol *sym, 4869 const struct block *block) 4870 { 4871 /* Do not try to complete stub types, as the debugger is probably 4872 already scanning all symbols matching a certain name at the 4873 time when this function is called. Trying to replace the stub 4874 type by its associated full type will cause us to restart a scan 4875 which may lead to an infinite recursion. Instead, the client 4876 collecting the matching symbols will end up collecting several 4877 matches, with at least one of them complete. It can then filter 4878 out the stub ones if needed. */ 4879 4880 for (int i = result.size () - 1; i >= 0; i -= 1) 4881 { 4882 if (lesseq_defined_than (sym, result[i].symbol)) 4883 return; 4884 else if (lesseq_defined_than (result[i].symbol, sym)) 4885 { 4886 result[i].symbol = sym; 4887 result[i].block = block; 4888 return; 4889 } 4890 } 4891 4892 struct block_symbol info; 4893 info.symbol = sym; 4894 info.block = block; 4895 result.push_back (info); 4896 } 4897 4898 /* Return a bound minimal symbol matching NAME according to Ada 4899 decoding rules. Returns an invalid symbol if there is no such 4900 minimal symbol. Names prefixed with "standard__" are handled 4901 specially: "standard__" is first stripped off, and only static and 4902 global symbols are searched. */ 4903 4904 struct bound_minimal_symbol 4905 ada_lookup_simple_minsym (const char *name, struct objfile *objfile) 4906 { 4907 struct bound_minimal_symbol result; 4908 4909 symbol_name_match_type match_type = name_match_type_from_name (name); 4910 lookup_name_info lookup_name (name, match_type); 4911 4912 symbol_name_matcher_ftype *match_name 4913 = ada_get_symbol_name_matcher (lookup_name); 4914 4915 gdbarch_iterate_over_objfiles_in_search_order 4916 (objfile != NULL ? objfile->arch () : target_gdbarch (), 4917 [&result, lookup_name, match_name] (struct objfile *obj) 4918 { 4919 for (minimal_symbol *msymbol : obj->msymbols ()) 4920 { 4921 if (match_name (msymbol->linkage_name (), lookup_name, nullptr) 4922 && msymbol->type () != mst_solib_trampoline) 4923 { 4924 result.minsym = msymbol; 4925 result.objfile = obj; 4926 return 1; 4927 } 4928 } 4929 4930 return 0; 4931 }, objfile); 4932 4933 return result; 4934 } 4935 4936 /* True if TYPE is definitely an artificial type supplied to a symbol 4937 for which no debugging information was given in the symbol file. */ 4938 4939 static int 4940 is_nondebugging_type (struct type *type) 4941 { 4942 const char *name = ada_type_name (type); 4943 4944 return (name != NULL && strcmp (name, "<variable, no debug info>") == 0); 4945 } 4946 4947 /* Return nonzero if TYPE1 and TYPE2 are two enumeration types 4948 that are deemed "identical" for practical purposes. 4949 4950 This function assumes that TYPE1 and TYPE2 are both TYPE_CODE_ENUM 4951 types and that their number of enumerals is identical (in other 4952 words, type1->num_fields () == type2->num_fields ()). */ 4953 4954 static int 4955 ada_identical_enum_types_p (struct type *type1, struct type *type2) 4956 { 4957 int i; 4958 4959 /* The heuristic we use here is fairly conservative. We consider 4960 that 2 enumerate types are identical if they have the same 4961 number of enumerals and that all enumerals have the same 4962 underlying value and name. */ 4963 4964 /* All enums in the type should have an identical underlying value. */ 4965 for (i = 0; i < type1->num_fields (); i++) 4966 if (type1->field (i).loc_enumval () != type2->field (i).loc_enumval ()) 4967 return 0; 4968 4969 /* All enumerals should also have the same name (modulo any numerical 4970 suffix). */ 4971 for (i = 0; i < type1->num_fields (); i++) 4972 { 4973 const char *name_1 = type1->field (i).name (); 4974 const char *name_2 = type2->field (i).name (); 4975 int len_1 = strlen (name_1); 4976 int len_2 = strlen (name_2); 4977 4978 ada_remove_trailing_digits (type1->field (i).name (), &len_1); 4979 ada_remove_trailing_digits (type2->field (i).name (), &len_2); 4980 if (len_1 != len_2 4981 || strncmp (type1->field (i).name (), 4982 type2->field (i).name (), 4983 len_1) != 0) 4984 return 0; 4985 } 4986 4987 return 1; 4988 } 4989 4990 /* Return nonzero if all the symbols in SYMS are all enumeral symbols 4991 that are deemed "identical" for practical purposes. Sometimes, 4992 enumerals are not strictly identical, but their types are so similar 4993 that they can be considered identical. 4994 4995 For instance, consider the following code: 4996 4997 type Color is (Black, Red, Green, Blue, White); 4998 type RGB_Color is new Color range Red .. Blue; 4999 5000 Type RGB_Color is a subrange of an implicit type which is a copy 5001 of type Color. If we call that implicit type RGB_ColorB ("B" is 5002 for "Base Type"), then type RGB_ColorB is a copy of type Color. 5003 As a result, when an expression references any of the enumeral 5004 by name (Eg. "print green"), the expression is technically 5005 ambiguous and the user should be asked to disambiguate. But 5006 doing so would only hinder the user, since it wouldn't matter 5007 what choice he makes, the outcome would always be the same. 5008 So, for practical purposes, we consider them as the same. */ 5009 5010 static int 5011 symbols_are_identical_enums (const std::vector<struct block_symbol> &syms) 5012 { 5013 int i; 5014 5015 /* Before performing a thorough comparison check of each type, 5016 we perform a series of inexpensive checks. We expect that these 5017 checks will quickly fail in the vast majority of cases, and thus 5018 help prevent the unnecessary use of a more expensive comparison. 5019 Said comparison also expects us to make some of these checks 5020 (see ada_identical_enum_types_p). */ 5021 5022 /* Quick check: All symbols should have an enum type. */ 5023 for (i = 0; i < syms.size (); i++) 5024 if (syms[i].symbol->type ()->code () != TYPE_CODE_ENUM) 5025 return 0; 5026 5027 /* Quick check: They should all have the same value. */ 5028 for (i = 1; i < syms.size (); i++) 5029 if (syms[i].symbol->value_longest () != syms[0].symbol->value_longest ()) 5030 return 0; 5031 5032 /* Quick check: They should all have the same number of enumerals. */ 5033 for (i = 1; i < syms.size (); i++) 5034 if (syms[i].symbol->type ()->num_fields () 5035 != syms[0].symbol->type ()->num_fields ()) 5036 return 0; 5037 5038 /* All the sanity checks passed, so we might have a set of 5039 identical enumeration types. Perform a more complete 5040 comparison of the type of each symbol. */ 5041 for (i = 1; i < syms.size (); i++) 5042 if (!ada_identical_enum_types_p (syms[i].symbol->type (), 5043 syms[0].symbol->type ())) 5044 return 0; 5045 5046 return 1; 5047 } 5048 5049 /* Remove any non-debugging symbols in SYMS that definitely 5050 duplicate other symbols in the list (The only case I know of where 5051 this happens is when object files containing stabs-in-ecoff are 5052 linked with files containing ordinary ecoff debugging symbols (or no 5053 debugging symbols)). Modifies SYMS to squeeze out deleted entries. */ 5054 5055 static void 5056 remove_extra_symbols (std::vector<struct block_symbol> *syms) 5057 { 5058 int i, j; 5059 5060 /* We should never be called with less than 2 symbols, as there 5061 cannot be any extra symbol in that case. But it's easy to 5062 handle, since we have nothing to do in that case. */ 5063 if (syms->size () < 2) 5064 return; 5065 5066 i = 0; 5067 while (i < syms->size ()) 5068 { 5069 int remove_p = 0; 5070 5071 /* If two symbols have the same name and one of them is a stub type, 5072 the get rid of the stub. */ 5073 5074 if ((*syms)[i].symbol->type ()->is_stub () 5075 && (*syms)[i].symbol->linkage_name () != NULL) 5076 { 5077 for (j = 0; j < syms->size (); j++) 5078 { 5079 if (j != i 5080 && !(*syms)[j].symbol->type ()->is_stub () 5081 && (*syms)[j].symbol->linkage_name () != NULL 5082 && strcmp ((*syms)[i].symbol->linkage_name (), 5083 (*syms)[j].symbol->linkage_name ()) == 0) 5084 remove_p = 1; 5085 } 5086 } 5087 5088 /* Two symbols with the same name, same class and same address 5089 should be identical. */ 5090 5091 else if ((*syms)[i].symbol->linkage_name () != NULL 5092 && (*syms)[i].symbol->aclass () == LOC_STATIC 5093 && is_nondebugging_type ((*syms)[i].symbol->type ())) 5094 { 5095 for (j = 0; j < syms->size (); j += 1) 5096 { 5097 if (i != j 5098 && (*syms)[j].symbol->linkage_name () != NULL 5099 && strcmp ((*syms)[i].symbol->linkage_name (), 5100 (*syms)[j].symbol->linkage_name ()) == 0 5101 && ((*syms)[i].symbol->aclass () 5102 == (*syms)[j].symbol->aclass ()) 5103 && (*syms)[i].symbol->value_address () 5104 == (*syms)[j].symbol->value_address ()) 5105 remove_p = 1; 5106 } 5107 } 5108 5109 if (remove_p) 5110 syms->erase (syms->begin () + i); 5111 else 5112 i += 1; 5113 } 5114 5115 /* If all the remaining symbols are identical enumerals, then 5116 just keep the first one and discard the rest. 5117 5118 Unlike what we did previously, we do not discard any entry 5119 unless they are ALL identical. This is because the symbol 5120 comparison is not a strict comparison, but rather a practical 5121 comparison. If all symbols are considered identical, then 5122 we can just go ahead and use the first one and discard the rest. 5123 But if we cannot reduce the list to a single element, we have 5124 to ask the user to disambiguate anyways. And if we have to 5125 present a multiple-choice menu, it's less confusing if the list 5126 isn't missing some choices that were identical and yet distinct. */ 5127 if (symbols_are_identical_enums (*syms)) 5128 syms->resize (1); 5129 } 5130 5131 /* Given a type that corresponds to a renaming entity, use the type name 5132 to extract the scope (package name or function name, fully qualified, 5133 and following the GNAT encoding convention) where this renaming has been 5134 defined. */ 5135 5136 static std::string 5137 xget_renaming_scope (struct type *renaming_type) 5138 { 5139 /* The renaming types adhere to the following convention: 5140 <scope>__<rename>___<XR extension>. 5141 So, to extract the scope, we search for the "___XR" extension, 5142 and then backtrack until we find the first "__". */ 5143 5144 const char *name = renaming_type->name (); 5145 const char *suffix = strstr (name, "___XR"); 5146 const char *last; 5147 5148 /* Now, backtrack a bit until we find the first "__". Start looking 5149 at suffix - 3, as the <rename> part is at least one character long. */ 5150 5151 for (last = suffix - 3; last > name; last--) 5152 if (last[0] == '_' && last[1] == '_') 5153 break; 5154 5155 /* Make a copy of scope and return it. */ 5156 return std::string (name, last); 5157 } 5158 5159 /* Return nonzero if NAME corresponds to a package name. */ 5160 5161 static int 5162 is_package_name (const char *name) 5163 { 5164 /* Here, We take advantage of the fact that no symbols are generated 5165 for packages, while symbols are generated for each function. 5166 So the condition for NAME represent a package becomes equivalent 5167 to NAME not existing in our list of symbols. There is only one 5168 small complication with library-level functions (see below). */ 5169 5170 /* If it is a function that has not been defined at library level, 5171 then we should be able to look it up in the symbols. */ 5172 if (standard_lookup (name, NULL, VAR_DOMAIN) != NULL) 5173 return 0; 5174 5175 /* Library-level function names start with "_ada_". See if function 5176 "_ada_" followed by NAME can be found. */ 5177 5178 /* Do a quick check that NAME does not contain "__", since library-level 5179 functions names cannot contain "__" in them. */ 5180 if (strstr (name, "__") != NULL) 5181 return 0; 5182 5183 std::string fun_name = string_printf ("_ada_%s", name); 5184 5185 return (standard_lookup (fun_name.c_str (), NULL, VAR_DOMAIN) == NULL); 5186 } 5187 5188 /* Return nonzero if SYM corresponds to a renaming entity that is 5189 not visible from FUNCTION_NAME. */ 5190 5191 static int 5192 old_renaming_is_invisible (const struct symbol *sym, const char *function_name) 5193 { 5194 if (sym->aclass () != LOC_TYPEDEF) 5195 return 0; 5196 5197 std::string scope = xget_renaming_scope (sym->type ()); 5198 5199 /* If the rename has been defined in a package, then it is visible. */ 5200 if (is_package_name (scope.c_str ())) 5201 return 0; 5202 5203 /* Check that the rename is in the current function scope by checking 5204 that its name starts with SCOPE. */ 5205 5206 /* If the function name starts with "_ada_", it means that it is 5207 a library-level function. Strip this prefix before doing the 5208 comparison, as the encoding for the renaming does not contain 5209 this prefix. */ 5210 if (startswith (function_name, "_ada_")) 5211 function_name += 5; 5212 5213 return !startswith (function_name, scope.c_str ()); 5214 } 5215 5216 /* Remove entries from SYMS that corresponds to a renaming entity that 5217 is not visible from the function associated with CURRENT_BLOCK or 5218 that is superfluous due to the presence of more specific renaming 5219 information. Places surviving symbols in the initial entries of 5220 SYMS. 5221 5222 Rationale: 5223 First, in cases where an object renaming is implemented as a 5224 reference variable, GNAT may produce both the actual reference 5225 variable and the renaming encoding. In this case, we discard the 5226 latter. 5227 5228 Second, GNAT emits a type following a specified encoding for each renaming 5229 entity. Unfortunately, STABS currently does not support the definition 5230 of types that are local to a given lexical block, so all renamings types 5231 are emitted at library level. As a consequence, if an application 5232 contains two renaming entities using the same name, and a user tries to 5233 print the value of one of these entities, the result of the ada symbol 5234 lookup will also contain the wrong renaming type. 5235 5236 This function partially covers for this limitation by attempting to 5237 remove from the SYMS list renaming symbols that should be visible 5238 from CURRENT_BLOCK. However, there does not seem be a 100% reliable 5239 method with the current information available. The implementation 5240 below has a couple of limitations (FIXME: brobecker-2003-05-12): 5241 5242 - When the user tries to print a rename in a function while there 5243 is another rename entity defined in a package: Normally, the 5244 rename in the function has precedence over the rename in the 5245 package, so the latter should be removed from the list. This is 5246 currently not the case. 5247 5248 - This function will incorrectly remove valid renames if 5249 the CURRENT_BLOCK corresponds to a function which symbol name 5250 has been changed by an "Export" pragma. As a consequence, 5251 the user will be unable to print such rename entities. */ 5252 5253 static void 5254 remove_irrelevant_renamings (std::vector<struct block_symbol> *syms, 5255 const struct block *current_block) 5256 { 5257 struct symbol *current_function; 5258 const char *current_function_name; 5259 int i; 5260 int is_new_style_renaming; 5261 5262 /* If there is both a renaming foo___XR... encoded as a variable and 5263 a simple variable foo in the same block, discard the latter. 5264 First, zero out such symbols, then compress. */ 5265 is_new_style_renaming = 0; 5266 for (i = 0; i < syms->size (); i += 1) 5267 { 5268 struct symbol *sym = (*syms)[i].symbol; 5269 const struct block *block = (*syms)[i].block; 5270 const char *name; 5271 const char *suffix; 5272 5273 if (sym == NULL || sym->aclass () == LOC_TYPEDEF) 5274 continue; 5275 name = sym->linkage_name (); 5276 suffix = strstr (name, "___XR"); 5277 5278 if (suffix != NULL) 5279 { 5280 int name_len = suffix - name; 5281 int j; 5282 5283 is_new_style_renaming = 1; 5284 for (j = 0; j < syms->size (); j += 1) 5285 if (i != j && (*syms)[j].symbol != NULL 5286 && strncmp (name, (*syms)[j].symbol->linkage_name (), 5287 name_len) == 0 5288 && block == (*syms)[j].block) 5289 (*syms)[j].symbol = NULL; 5290 } 5291 } 5292 if (is_new_style_renaming) 5293 { 5294 int j, k; 5295 5296 for (j = k = 0; j < syms->size (); j += 1) 5297 if ((*syms)[j].symbol != NULL) 5298 { 5299 (*syms)[k] = (*syms)[j]; 5300 k += 1; 5301 } 5302 syms->resize (k); 5303 return; 5304 } 5305 5306 /* Extract the function name associated to CURRENT_BLOCK. 5307 Abort if unable to do so. */ 5308 5309 if (current_block == NULL) 5310 return; 5311 5312 current_function = block_linkage_function (current_block); 5313 if (current_function == NULL) 5314 return; 5315 5316 current_function_name = current_function->linkage_name (); 5317 if (current_function_name == NULL) 5318 return; 5319 5320 /* Check each of the symbols, and remove it from the list if it is 5321 a type corresponding to a renaming that is out of the scope of 5322 the current block. */ 5323 5324 i = 0; 5325 while (i < syms->size ()) 5326 { 5327 if (ada_parse_renaming ((*syms)[i].symbol, NULL, NULL, NULL) 5328 == ADA_OBJECT_RENAMING 5329 && old_renaming_is_invisible ((*syms)[i].symbol, 5330 current_function_name)) 5331 syms->erase (syms->begin () + i); 5332 else 5333 i += 1; 5334 } 5335 } 5336 5337 /* Add to RESULT all symbols from BLOCK (and its super-blocks) 5338 whose name and domain match LOOKUP_NAME and DOMAIN respectively. 5339 5340 Note: This function assumes that RESULT is empty. */ 5341 5342 static void 5343 ada_add_local_symbols (std::vector<struct block_symbol> &result, 5344 const lookup_name_info &lookup_name, 5345 const struct block *block, domain_enum domain) 5346 { 5347 while (block != NULL) 5348 { 5349 ada_add_block_symbols (result, block, lookup_name, domain, NULL); 5350 5351 /* If we found a non-function match, assume that's the one. We 5352 only check this when finding a function boundary, so that we 5353 can accumulate all results from intervening blocks first. */ 5354 if (block->function () != nullptr && is_nonfunction (result)) 5355 return; 5356 5357 block = block->superblock (); 5358 } 5359 } 5360 5361 /* An object of this type is used as the callback argument when 5362 calling the map_matching_symbols method. */ 5363 5364 struct match_data 5365 { 5366 explicit match_data (std::vector<struct block_symbol> *rp) 5367 : resultp (rp) 5368 { 5369 } 5370 DISABLE_COPY_AND_ASSIGN (match_data); 5371 5372 bool operator() (struct block_symbol *bsym); 5373 5374 struct objfile *objfile = nullptr; 5375 std::vector<struct block_symbol> *resultp; 5376 struct symbol *arg_sym = nullptr; 5377 bool found_sym = false; 5378 }; 5379 5380 /* A callback for add_nonlocal_symbols that adds symbol, found in 5381 BSYM, to a list of symbols. */ 5382 5383 bool 5384 match_data::operator() (struct block_symbol *bsym) 5385 { 5386 const struct block *block = bsym->block; 5387 struct symbol *sym = bsym->symbol; 5388 5389 if (sym == NULL) 5390 { 5391 if (!found_sym && arg_sym != NULL) 5392 add_defn_to_vec (*resultp, 5393 fixup_symbol_section (arg_sym, objfile), 5394 block); 5395 found_sym = false; 5396 arg_sym = NULL; 5397 } 5398 else 5399 { 5400 if (sym->aclass () == LOC_UNRESOLVED) 5401 return true; 5402 else if (sym->is_argument ()) 5403 arg_sym = sym; 5404 else 5405 { 5406 found_sym = true; 5407 add_defn_to_vec (*resultp, 5408 fixup_symbol_section (sym, objfile), 5409 block); 5410 } 5411 } 5412 return true; 5413 } 5414 5415 /* Helper for add_nonlocal_symbols. Find symbols in DOMAIN which are 5416 targeted by renamings matching LOOKUP_NAME in BLOCK. Add these 5417 symbols to RESULT. Return whether we found such symbols. */ 5418 5419 static int 5420 ada_add_block_renamings (std::vector<struct block_symbol> &result, 5421 const struct block *block, 5422 const lookup_name_info &lookup_name, 5423 domain_enum domain) 5424 { 5425 struct using_direct *renaming; 5426 int defns_mark = result.size (); 5427 5428 symbol_name_matcher_ftype *name_match 5429 = ada_get_symbol_name_matcher (lookup_name); 5430 5431 for (renaming = block_using (block); 5432 renaming != NULL; 5433 renaming = renaming->next) 5434 { 5435 const char *r_name; 5436 5437 /* Avoid infinite recursions: skip this renaming if we are actually 5438 already traversing it. 5439 5440 Currently, symbol lookup in Ada don't use the namespace machinery from 5441 C++/Fortran support: skip namespace imports that use them. */ 5442 if (renaming->searched 5443 || (renaming->import_src != NULL 5444 && renaming->import_src[0] != '\0') 5445 || (renaming->import_dest != NULL 5446 && renaming->import_dest[0] != '\0')) 5447 continue; 5448 renaming->searched = 1; 5449 5450 /* TODO: here, we perform another name-based symbol lookup, which can 5451 pull its own multiple overloads. In theory, we should be able to do 5452 better in this case since, in DWARF, DW_AT_import is a DIE reference, 5453 not a simple name. But in order to do this, we would need to enhance 5454 the DWARF reader to associate a symbol to this renaming, instead of a 5455 name. So, for now, we do something simpler: re-use the C++/Fortran 5456 namespace machinery. */ 5457 r_name = (renaming->alias != NULL 5458 ? renaming->alias 5459 : renaming->declaration); 5460 if (name_match (r_name, lookup_name, NULL)) 5461 { 5462 lookup_name_info decl_lookup_name (renaming->declaration, 5463 lookup_name.match_type ()); 5464 ada_add_all_symbols (result, block, decl_lookup_name, domain, 5465 1, NULL); 5466 } 5467 renaming->searched = 0; 5468 } 5469 return result.size () != defns_mark; 5470 } 5471 5472 /* Implements compare_names, but only applying the comparision using 5473 the given CASING. */ 5474 5475 static int 5476 compare_names_with_case (const char *string1, const char *string2, 5477 enum case_sensitivity casing) 5478 { 5479 while (*string1 != '\0' && *string2 != '\0') 5480 { 5481 char c1, c2; 5482 5483 if (isspace (*string1) || isspace (*string2)) 5484 return strcmp_iw_ordered (string1, string2); 5485 5486 if (casing == case_sensitive_off) 5487 { 5488 c1 = tolower (*string1); 5489 c2 = tolower (*string2); 5490 } 5491 else 5492 { 5493 c1 = *string1; 5494 c2 = *string2; 5495 } 5496 if (c1 != c2) 5497 break; 5498 5499 string1 += 1; 5500 string2 += 1; 5501 } 5502 5503 switch (*string1) 5504 { 5505 case '(': 5506 return strcmp_iw_ordered (string1, string2); 5507 case '_': 5508 if (*string2 == '\0') 5509 { 5510 if (is_name_suffix (string1)) 5511 return 0; 5512 else 5513 return 1; 5514 } 5515 /* FALLTHROUGH */ 5516 default: 5517 if (*string2 == '(') 5518 return strcmp_iw_ordered (string1, string2); 5519 else 5520 { 5521 if (casing == case_sensitive_off) 5522 return tolower (*string1) - tolower (*string2); 5523 else 5524 return *string1 - *string2; 5525 } 5526 } 5527 } 5528 5529 /* Compare STRING1 to STRING2, with results as for strcmp. 5530 Compatible with strcmp_iw_ordered in that... 5531 5532 strcmp_iw_ordered (STRING1, STRING2) <= 0 5533 5534 ... implies... 5535 5536 compare_names (STRING1, STRING2) <= 0 5537 5538 (they may differ as to what symbols compare equal). */ 5539 5540 static int 5541 compare_names (const char *string1, const char *string2) 5542 { 5543 int result; 5544 5545 /* Similar to what strcmp_iw_ordered does, we need to perform 5546 a case-insensitive comparison first, and only resort to 5547 a second, case-sensitive, comparison if the first one was 5548 not sufficient to differentiate the two strings. */ 5549 5550 result = compare_names_with_case (string1, string2, case_sensitive_off); 5551 if (result == 0) 5552 result = compare_names_with_case (string1, string2, case_sensitive_on); 5553 5554 return result; 5555 } 5556 5557 /* Convenience function to get at the Ada encoded lookup name for 5558 LOOKUP_NAME, as a C string. */ 5559 5560 static const char * 5561 ada_lookup_name (const lookup_name_info &lookup_name) 5562 { 5563 return lookup_name.ada ().lookup_name ().c_str (); 5564 } 5565 5566 /* A helper for add_nonlocal_symbols. Call expand_matching_symbols 5567 for OBJFILE, then walk the objfile's symtabs and update the 5568 results. */ 5569 5570 static void 5571 map_matching_symbols (struct objfile *objfile, 5572 const lookup_name_info &lookup_name, 5573 bool is_wild_match, 5574 domain_enum domain, 5575 int global, 5576 match_data &data) 5577 { 5578 data.objfile = objfile; 5579 objfile->expand_matching_symbols (lookup_name, domain, global, 5580 is_wild_match ? nullptr : compare_names); 5581 5582 const int block_kind = global ? GLOBAL_BLOCK : STATIC_BLOCK; 5583 for (compunit_symtab *symtab : objfile->compunits ()) 5584 { 5585 const struct block *block 5586 = symtab->blockvector ()->block (block_kind); 5587 if (!iterate_over_symbols_terminated (block, lookup_name, 5588 domain, data)) 5589 break; 5590 } 5591 } 5592 5593 /* Add to RESULT all non-local symbols whose name and domain match 5594 LOOKUP_NAME and DOMAIN respectively. The search is performed on 5595 GLOBAL_BLOCK symbols if GLOBAL is non-zero, or on STATIC_BLOCK 5596 symbols otherwise. */ 5597 5598 static void 5599 add_nonlocal_symbols (std::vector<struct block_symbol> &result, 5600 const lookup_name_info &lookup_name, 5601 domain_enum domain, int global) 5602 { 5603 struct match_data data (&result); 5604 5605 bool is_wild_match = lookup_name.ada ().wild_match_p (); 5606 5607 for (objfile *objfile : current_program_space->objfiles ()) 5608 { 5609 map_matching_symbols (objfile, lookup_name, is_wild_match, domain, 5610 global, data); 5611 5612 for (compunit_symtab *cu : objfile->compunits ()) 5613 { 5614 const struct block *global_block 5615 = cu->blockvector ()->global_block (); 5616 5617 if (ada_add_block_renamings (result, global_block, lookup_name, 5618 domain)) 5619 data.found_sym = true; 5620 } 5621 } 5622 5623 if (result.empty () && global && !is_wild_match) 5624 { 5625 const char *name = ada_lookup_name (lookup_name); 5626 std::string bracket_name = std::string ("<_ada_") + name + '>'; 5627 lookup_name_info name1 (bracket_name, symbol_name_match_type::FULL); 5628 5629 for (objfile *objfile : current_program_space->objfiles ()) 5630 map_matching_symbols (objfile, name1, false, domain, global, data); 5631 } 5632 } 5633 5634 /* Find symbols in DOMAIN matching LOOKUP_NAME, in BLOCK and, if 5635 FULL_SEARCH is non-zero, enclosing scope and in global scopes, 5636 returning the number of matches. Add these to RESULT. 5637 5638 When FULL_SEARCH is non-zero, any non-function/non-enumeral 5639 symbol match within the nest of blocks whose innermost member is BLOCK, 5640 is the one match returned (no other matches in that or 5641 enclosing blocks is returned). If there are any matches in or 5642 surrounding BLOCK, then these alone are returned. 5643 5644 Names prefixed with "standard__" are handled specially: 5645 "standard__" is first stripped off (by the lookup_name 5646 constructor), and only static and global symbols are searched. 5647 5648 If MADE_GLOBAL_LOOKUP_P is non-null, set it before return to whether we had 5649 to lookup global symbols. */ 5650 5651 static void 5652 ada_add_all_symbols (std::vector<struct block_symbol> &result, 5653 const struct block *block, 5654 const lookup_name_info &lookup_name, 5655 domain_enum domain, 5656 int full_search, 5657 int *made_global_lookup_p) 5658 { 5659 struct symbol *sym; 5660 5661 if (made_global_lookup_p) 5662 *made_global_lookup_p = 0; 5663 5664 /* Special case: If the user specifies a symbol name inside package 5665 Standard, do a non-wild matching of the symbol name without 5666 the "standard__" prefix. This was primarily introduced in order 5667 to allow the user to specifically access the standard exceptions 5668 using, for instance, Standard.Constraint_Error when Constraint_Error 5669 is ambiguous (due to the user defining its own Constraint_Error 5670 entity inside its program). */ 5671 if (lookup_name.ada ().standard_p ()) 5672 block = NULL; 5673 5674 /* Check the non-global symbols. If we have ANY match, then we're done. */ 5675 5676 if (block != NULL) 5677 { 5678 if (full_search) 5679 ada_add_local_symbols (result, lookup_name, block, domain); 5680 else 5681 { 5682 /* In the !full_search case we're are being called by 5683 iterate_over_symbols, and we don't want to search 5684 superblocks. */ 5685 ada_add_block_symbols (result, block, lookup_name, domain, NULL); 5686 } 5687 if (!result.empty () || !full_search) 5688 return; 5689 } 5690 5691 /* No non-global symbols found. Check our cache to see if we have 5692 already performed this search before. If we have, then return 5693 the same result. */ 5694 5695 if (lookup_cached_symbol (ada_lookup_name (lookup_name), 5696 domain, &sym, &block)) 5697 { 5698 if (sym != NULL) 5699 add_defn_to_vec (result, sym, block); 5700 return; 5701 } 5702 5703 if (made_global_lookup_p) 5704 *made_global_lookup_p = 1; 5705 5706 /* Search symbols from all global blocks. */ 5707 5708 add_nonlocal_symbols (result, lookup_name, domain, 1); 5709 5710 /* Now add symbols from all per-file blocks if we've gotten no hits 5711 (not strictly correct, but perhaps better than an error). */ 5712 5713 if (result.empty ()) 5714 add_nonlocal_symbols (result, lookup_name, domain, 0); 5715 } 5716 5717 /* Find symbols in DOMAIN matching LOOKUP_NAME, in BLOCK and, if FULL_SEARCH 5718 is non-zero, enclosing scope and in global scopes. 5719 5720 Returns (SYM,BLOCK) tuples, indicating the symbols found and the 5721 blocks and symbol tables (if any) in which they were found. 5722 5723 When full_search is non-zero, any non-function/non-enumeral 5724 symbol match within the nest of blocks whose innermost member is BLOCK, 5725 is the one match returned (no other matches in that or 5726 enclosing blocks is returned). If there are any matches in or 5727 surrounding BLOCK, then these alone are returned. 5728 5729 Names prefixed with "standard__" are handled specially: "standard__" 5730 is first stripped off, and only static and global symbols are searched. */ 5731 5732 static std::vector<struct block_symbol> 5733 ada_lookup_symbol_list_worker (const lookup_name_info &lookup_name, 5734 const struct block *block, 5735 domain_enum domain, 5736 int full_search) 5737 { 5738 int syms_from_global_search; 5739 std::vector<struct block_symbol> results; 5740 5741 ada_add_all_symbols (results, block, lookup_name, 5742 domain, full_search, &syms_from_global_search); 5743 5744 remove_extra_symbols (&results); 5745 5746 if (results.empty () && full_search && syms_from_global_search) 5747 cache_symbol (ada_lookup_name (lookup_name), domain, NULL, NULL); 5748 5749 if (results.size () == 1 && full_search && syms_from_global_search) 5750 cache_symbol (ada_lookup_name (lookup_name), domain, 5751 results[0].symbol, results[0].block); 5752 5753 remove_irrelevant_renamings (&results, block); 5754 return results; 5755 } 5756 5757 /* Find symbols in DOMAIN matching NAME, in BLOCK and enclosing scope and 5758 in global scopes, returning (SYM,BLOCK) tuples. 5759 5760 See ada_lookup_symbol_list_worker for further details. */ 5761 5762 std::vector<struct block_symbol> 5763 ada_lookup_symbol_list (const char *name, const struct block *block, 5764 domain_enum domain) 5765 { 5766 symbol_name_match_type name_match_type = name_match_type_from_name (name); 5767 lookup_name_info lookup_name (name, name_match_type); 5768 5769 return ada_lookup_symbol_list_worker (lookup_name, block, domain, 1); 5770 } 5771 5772 /* The result is as for ada_lookup_symbol_list with FULL_SEARCH set 5773 to 1, but choosing the first symbol found if there are multiple 5774 choices. 5775 5776 The result is stored in *INFO, which must be non-NULL. 5777 If no match is found, INFO->SYM is set to NULL. */ 5778 5779 void 5780 ada_lookup_encoded_symbol (const char *name, const struct block *block, 5781 domain_enum domain, 5782 struct block_symbol *info) 5783 { 5784 /* Since we already have an encoded name, wrap it in '<>' to force a 5785 verbatim match. Otherwise, if the name happens to not look like 5786 an encoded name (because it doesn't include a "__"), 5787 ada_lookup_name_info would re-encode/fold it again, and that 5788 would e.g., incorrectly lowercase object renaming names like 5789 "R28b" -> "r28b". */ 5790 std::string verbatim = add_angle_brackets (name); 5791 5792 gdb_assert (info != NULL); 5793 *info = ada_lookup_symbol (verbatim.c_str (), block, domain); 5794 } 5795 5796 /* Return a symbol in DOMAIN matching NAME, in BLOCK0 and enclosing 5797 scope and in global scopes, or NULL if none. NAME is folded and 5798 encoded first. Otherwise, the result is as for ada_lookup_symbol_list, 5799 choosing the first symbol if there are multiple choices. */ 5800 5801 struct block_symbol 5802 ada_lookup_symbol (const char *name, const struct block *block0, 5803 domain_enum domain) 5804 { 5805 std::vector<struct block_symbol> candidates 5806 = ada_lookup_symbol_list (name, block0, domain); 5807 5808 if (candidates.empty ()) 5809 return {}; 5810 5811 block_symbol info = candidates[0]; 5812 info.symbol = fixup_symbol_section (info.symbol, NULL); 5813 return info; 5814 } 5815 5816 5817 /* True iff STR is a possible encoded suffix of a normal Ada name 5818 that is to be ignored for matching purposes. Suffixes of parallel 5819 names (e.g., XVE) are not included here. Currently, the possible suffixes 5820 are given by any of the regular expressions: 5821 5822 [.$][0-9]+ [nested subprogram suffix, on platforms such as GNU/Linux] 5823 ___[0-9]+ [nested subprogram suffix, on platforms such as HP/UX] 5824 TKB [subprogram suffix for task bodies] 5825 _E[0-9]+[bs]$ [protected object entry suffixes] 5826 (X[nb]*)?((\$|__)[0-9](_?[0-9]+)|___(JM|LJM|X([FDBUP].*|R[^T]?)))?$ 5827 5828 Also, any leading "__[0-9]+" sequence is skipped before the suffix 5829 match is performed. This sequence is used to differentiate homonyms, 5830 is an optional part of a valid name suffix. */ 5831 5832 static int 5833 is_name_suffix (const char *str) 5834 { 5835 int k; 5836 const char *matching; 5837 const int len = strlen (str); 5838 5839 /* Skip optional leading __[0-9]+. */ 5840 5841 if (len > 3 && str[0] == '_' && str[1] == '_' && isdigit (str[2])) 5842 { 5843 str += 3; 5844 while (isdigit (str[0])) 5845 str += 1; 5846 } 5847 5848 /* [.$][0-9]+ */ 5849 5850 if (str[0] == '.' || str[0] == '$') 5851 { 5852 matching = str + 1; 5853 while (isdigit (matching[0])) 5854 matching += 1; 5855 if (matching[0] == '\0') 5856 return 1; 5857 } 5858 5859 /* ___[0-9]+ */ 5860 5861 if (len > 3 && str[0] == '_' && str[1] == '_' && str[2] == '_') 5862 { 5863 matching = str + 3; 5864 while (isdigit (matching[0])) 5865 matching += 1; 5866 if (matching[0] == '\0') 5867 return 1; 5868 } 5869 5870 /* "TKB" suffixes are used for subprograms implementing task bodies. */ 5871 5872 if (strcmp (str, "TKB") == 0) 5873 return 1; 5874 5875 #if 0 5876 /* FIXME: brobecker/2005-09-23: Protected Object subprograms end 5877 with a N at the end. Unfortunately, the compiler uses the same 5878 convention for other internal types it creates. So treating 5879 all entity names that end with an "N" as a name suffix causes 5880 some regressions. For instance, consider the case of an enumerated 5881 type. To support the 'Image attribute, it creates an array whose 5882 name ends with N. 5883 Having a single character like this as a suffix carrying some 5884 information is a bit risky. Perhaps we should change the encoding 5885 to be something like "_N" instead. In the meantime, do not do 5886 the following check. */ 5887 /* Protected Object Subprograms */ 5888 if (len == 1 && str [0] == 'N') 5889 return 1; 5890 #endif 5891 5892 /* _E[0-9]+[bs]$ */ 5893 if (len > 3 && str[0] == '_' && str [1] == 'E' && isdigit (str[2])) 5894 { 5895 matching = str + 3; 5896 while (isdigit (matching[0])) 5897 matching += 1; 5898 if ((matching[0] == 'b' || matching[0] == 's') 5899 && matching [1] == '\0') 5900 return 1; 5901 } 5902 5903 /* ??? We should not modify STR directly, as we are doing below. This 5904 is fine in this case, but may become problematic later if we find 5905 that this alternative did not work, and want to try matching 5906 another one from the begining of STR. Since we modified it, we 5907 won't be able to find the begining of the string anymore! */ 5908 if (str[0] == 'X') 5909 { 5910 str += 1; 5911 while (str[0] != '_' && str[0] != '\0') 5912 { 5913 if (str[0] != 'n' && str[0] != 'b') 5914 return 0; 5915 str += 1; 5916 } 5917 } 5918 5919 if (str[0] == '\000') 5920 return 1; 5921 5922 if (str[0] == '_') 5923 { 5924 if (str[1] != '_' || str[2] == '\000') 5925 return 0; 5926 if (str[2] == '_') 5927 { 5928 if (strcmp (str + 3, "JM") == 0) 5929 return 1; 5930 /* FIXME: brobecker/2004-09-30: GNAT will soon stop using 5931 the LJM suffix in favor of the JM one. But we will 5932 still accept LJM as a valid suffix for a reasonable 5933 amount of time, just to allow ourselves to debug programs 5934 compiled using an older version of GNAT. */ 5935 if (strcmp (str + 3, "LJM") == 0) 5936 return 1; 5937 if (str[3] != 'X') 5938 return 0; 5939 if (str[4] == 'F' || str[4] == 'D' || str[4] == 'B' 5940 || str[4] == 'U' || str[4] == 'P') 5941 return 1; 5942 if (str[4] == 'R' && str[5] != 'T') 5943 return 1; 5944 return 0; 5945 } 5946 if (!isdigit (str[2])) 5947 return 0; 5948 for (k = 3; str[k] != '\0'; k += 1) 5949 if (!isdigit (str[k]) && str[k] != '_') 5950 return 0; 5951 return 1; 5952 } 5953 if (str[0] == '$' && isdigit (str[1])) 5954 { 5955 for (k = 2; str[k] != '\0'; k += 1) 5956 if (!isdigit (str[k]) && str[k] != '_') 5957 return 0; 5958 return 1; 5959 } 5960 return 0; 5961 } 5962 5963 /* Return non-zero if the string starting at NAME and ending before 5964 NAME_END contains no capital letters. */ 5965 5966 static int 5967 is_valid_name_for_wild_match (const char *name0) 5968 { 5969 std::string decoded_name = ada_decode (name0); 5970 int i; 5971 5972 /* If the decoded name starts with an angle bracket, it means that 5973 NAME0 does not follow the GNAT encoding format. It should then 5974 not be allowed as a possible wild match. */ 5975 if (decoded_name[0] == '<') 5976 return 0; 5977 5978 for (i=0; decoded_name[i] != '\0'; i++) 5979 if (isalpha (decoded_name[i]) && !islower (decoded_name[i])) 5980 return 0; 5981 5982 return 1; 5983 } 5984 5985 /* Advance *NAMEP to next occurrence in the string NAME0 of the TARGET0 5986 character which could start a simple name. Assumes that *NAMEP points 5987 somewhere inside the string beginning at NAME0. */ 5988 5989 static int 5990 advance_wild_match (const char **namep, const char *name0, char target0) 5991 { 5992 const char *name = *namep; 5993 5994 while (1) 5995 { 5996 char t0, t1; 5997 5998 t0 = *name; 5999 if (t0 == '_') 6000 { 6001 t1 = name[1]; 6002 if ((t1 >= 'a' && t1 <= 'z') || (t1 >= '0' && t1 <= '9')) 6003 { 6004 name += 1; 6005 if (name == name0 + 5 && startswith (name0, "_ada")) 6006 break; 6007 else 6008 name += 1; 6009 } 6010 else if (t1 == '_' && ((name[2] >= 'a' && name[2] <= 'z') 6011 || name[2] == target0)) 6012 { 6013 name += 2; 6014 break; 6015 } 6016 else if (t1 == '_' && name[2] == 'B' && name[3] == '_') 6017 { 6018 /* Names like "pkg__B_N__name", where N is a number, are 6019 block-local. We can handle these by simply skipping 6020 the "B_" here. */ 6021 name += 4; 6022 } 6023 else 6024 return 0; 6025 } 6026 else if ((t0 >= 'a' && t0 <= 'z') || (t0 >= '0' && t0 <= '9')) 6027 name += 1; 6028 else 6029 return 0; 6030 } 6031 6032 *namep = name; 6033 return 1; 6034 } 6035 6036 /* Return true iff NAME encodes a name of the form prefix.PATN. 6037 Ignores any informational suffixes of NAME (i.e., for which 6038 is_name_suffix is true). Assumes that PATN is a lower-cased Ada 6039 simple name. */ 6040 6041 static bool 6042 wild_match (const char *name, const char *patn) 6043 { 6044 const char *p; 6045 const char *name0 = name; 6046 6047 if (startswith (name, "___ghost_")) 6048 name += 9; 6049 6050 while (1) 6051 { 6052 const char *match = name; 6053 6054 if (*name == *patn) 6055 { 6056 for (name += 1, p = patn + 1; *p != '\0'; name += 1, p += 1) 6057 if (*p != *name) 6058 break; 6059 if (*p == '\0' && is_name_suffix (name)) 6060 return match == name0 || is_valid_name_for_wild_match (name0); 6061 6062 if (name[-1] == '_') 6063 name -= 1; 6064 } 6065 if (!advance_wild_match (&name, name0, *patn)) 6066 return false; 6067 } 6068 } 6069 6070 /* Add symbols from BLOCK matching LOOKUP_NAME in DOMAIN to RESULT (if 6071 necessary). OBJFILE is the section containing BLOCK. */ 6072 6073 static void 6074 ada_add_block_symbols (std::vector<struct block_symbol> &result, 6075 const struct block *block, 6076 const lookup_name_info &lookup_name, 6077 domain_enum domain, struct objfile *objfile) 6078 { 6079 struct block_iterator iter; 6080 /* A matching argument symbol, if any. */ 6081 struct symbol *arg_sym; 6082 /* Set true when we find a matching non-argument symbol. */ 6083 bool found_sym; 6084 struct symbol *sym; 6085 6086 arg_sym = NULL; 6087 found_sym = false; 6088 for (sym = block_iter_match_first (block, lookup_name, &iter); 6089 sym != NULL; 6090 sym = block_iter_match_next (lookup_name, &iter)) 6091 { 6092 if (symbol_matches_domain (sym->language (), sym->domain (), domain)) 6093 { 6094 if (sym->aclass () != LOC_UNRESOLVED) 6095 { 6096 if (sym->is_argument ()) 6097 arg_sym = sym; 6098 else 6099 { 6100 found_sym = true; 6101 add_defn_to_vec (result, 6102 fixup_symbol_section (sym, objfile), 6103 block); 6104 } 6105 } 6106 } 6107 } 6108 6109 /* Handle renamings. */ 6110 6111 if (ada_add_block_renamings (result, block, lookup_name, domain)) 6112 found_sym = true; 6113 6114 if (!found_sym && arg_sym != NULL) 6115 { 6116 add_defn_to_vec (result, 6117 fixup_symbol_section (arg_sym, objfile), 6118 block); 6119 } 6120 6121 if (!lookup_name.ada ().wild_match_p ()) 6122 { 6123 arg_sym = NULL; 6124 found_sym = false; 6125 const std::string &ada_lookup_name = lookup_name.ada ().lookup_name (); 6126 const char *name = ada_lookup_name.c_str (); 6127 size_t name_len = ada_lookup_name.size (); 6128 6129 ALL_BLOCK_SYMBOLS (block, iter, sym) 6130 { 6131 if (symbol_matches_domain (sym->language (), 6132 sym->domain (), domain)) 6133 { 6134 int cmp; 6135 6136 cmp = (int) '_' - (int) sym->linkage_name ()[0]; 6137 if (cmp == 0) 6138 { 6139 cmp = !startswith (sym->linkage_name (), "_ada_"); 6140 if (cmp == 0) 6141 cmp = strncmp (name, sym->linkage_name () + 5, 6142 name_len); 6143 } 6144 6145 if (cmp == 0 6146 && is_name_suffix (sym->linkage_name () + name_len + 5)) 6147 { 6148 if (sym->aclass () != LOC_UNRESOLVED) 6149 { 6150 if (sym->is_argument ()) 6151 arg_sym = sym; 6152 else 6153 { 6154 found_sym = true; 6155 add_defn_to_vec (result, 6156 fixup_symbol_section (sym, objfile), 6157 block); 6158 } 6159 } 6160 } 6161 } 6162 } 6163 6164 /* NOTE: This really shouldn't be needed for _ada_ symbols. 6165 They aren't parameters, right? */ 6166 if (!found_sym && arg_sym != NULL) 6167 { 6168 add_defn_to_vec (result, 6169 fixup_symbol_section (arg_sym, objfile), 6170 block); 6171 } 6172 } 6173 } 6174 6175 6176 /* Symbol Completion */ 6177 6178 /* See symtab.h. */ 6179 6180 bool 6181 ada_lookup_name_info::matches 6182 (const char *sym_name, 6183 symbol_name_match_type match_type, 6184 completion_match_result *comp_match_res) const 6185 { 6186 bool match = false; 6187 const char *text = m_encoded_name.c_str (); 6188 size_t text_len = m_encoded_name.size (); 6189 6190 /* First, test against the fully qualified name of the symbol. */ 6191 6192 if (strncmp (sym_name, text, text_len) == 0) 6193 match = true; 6194 6195 std::string decoded_name = ada_decode (sym_name); 6196 if (match && !m_encoded_p) 6197 { 6198 /* One needed check before declaring a positive match is to verify 6199 that iff we are doing a verbatim match, the decoded version 6200 of the symbol name starts with '<'. Otherwise, this symbol name 6201 is not a suitable completion. */ 6202 6203 bool has_angle_bracket = (decoded_name[0] == '<'); 6204 match = (has_angle_bracket == m_verbatim_p); 6205 } 6206 6207 if (match && !m_verbatim_p) 6208 { 6209 /* When doing non-verbatim match, another check that needs to 6210 be done is to verify that the potentially matching symbol name 6211 does not include capital letters, because the ada-mode would 6212 not be able to understand these symbol names without the 6213 angle bracket notation. */ 6214 const char *tmp; 6215 6216 for (tmp = sym_name; *tmp != '\0' && !isupper (*tmp); tmp++); 6217 if (*tmp != '\0') 6218 match = false; 6219 } 6220 6221 /* Second: Try wild matching... */ 6222 6223 if (!match && m_wild_match_p) 6224 { 6225 /* Since we are doing wild matching, this means that TEXT 6226 may represent an unqualified symbol name. We therefore must 6227 also compare TEXT against the unqualified name of the symbol. */ 6228 sym_name = ada_unqualified_name (decoded_name.c_str ()); 6229 6230 if (strncmp (sym_name, text, text_len) == 0) 6231 match = true; 6232 } 6233 6234 /* Finally: If we found a match, prepare the result to return. */ 6235 6236 if (!match) 6237 return false; 6238 6239 if (comp_match_res != NULL) 6240 { 6241 std::string &match_str = comp_match_res->match.storage (); 6242 6243 if (!m_encoded_p) 6244 match_str = ada_decode (sym_name); 6245 else 6246 { 6247 if (m_verbatim_p) 6248 match_str = add_angle_brackets (sym_name); 6249 else 6250 match_str = sym_name; 6251 6252 } 6253 6254 comp_match_res->set_match (match_str.c_str ()); 6255 } 6256 6257 return true; 6258 } 6259 6260 /* Field Access */ 6261 6262 /* Return non-zero if TYPE is a pointer to the GNAT dispatch table used 6263 for tagged types. */ 6264 6265 static int 6266 ada_is_dispatch_table_ptr_type (struct type *type) 6267 { 6268 const char *name; 6269 6270 if (type->code () != TYPE_CODE_PTR) 6271 return 0; 6272 6273 name = type->target_type ()->name (); 6274 if (name == NULL) 6275 return 0; 6276 6277 return (strcmp (name, "ada__tags__dispatch_table") == 0); 6278 } 6279 6280 /* Return non-zero if TYPE is an interface tag. */ 6281 6282 static int 6283 ada_is_interface_tag (struct type *type) 6284 { 6285 const char *name = type->name (); 6286 6287 if (name == NULL) 6288 return 0; 6289 6290 return (strcmp (name, "ada__tags__interface_tag") == 0); 6291 } 6292 6293 /* True if field number FIELD_NUM in struct or union type TYPE is supposed 6294 to be invisible to users. */ 6295 6296 int 6297 ada_is_ignored_field (struct type *type, int field_num) 6298 { 6299 if (field_num < 0 || field_num > type->num_fields ()) 6300 return 1; 6301 6302 /* Check the name of that field. */ 6303 { 6304 const char *name = type->field (field_num).name (); 6305 6306 /* Anonymous field names should not be printed. 6307 brobecker/2007-02-20: I don't think this can actually happen 6308 but we don't want to print the value of anonymous fields anyway. */ 6309 if (name == NULL) 6310 return 1; 6311 6312 /* Normally, fields whose name start with an underscore ("_") 6313 are fields that have been internally generated by the compiler, 6314 and thus should not be printed. The "_parent" field is special, 6315 however: This is a field internally generated by the compiler 6316 for tagged types, and it contains the components inherited from 6317 the parent type. This field should not be printed as is, but 6318 should not be ignored either. */ 6319 if (name[0] == '_' && !startswith (name, "_parent")) 6320 return 1; 6321 6322 /* The compiler doesn't document this, but sometimes it emits 6323 a field whose name starts with a capital letter, like 'V148s'. 6324 These aren't marked as artificial in any way, but we know they 6325 should be ignored. However, wrapper fields should not be 6326 ignored. */ 6327 if (name[0] == 'S' || name[0] == 'R' || name[0] == 'O') 6328 { 6329 /* Wrapper field. */ 6330 } 6331 else if (isupper (name[0])) 6332 return 1; 6333 } 6334 6335 /* If this is the dispatch table of a tagged type or an interface tag, 6336 then ignore. */ 6337 if (ada_is_tagged_type (type, 1) 6338 && (ada_is_dispatch_table_ptr_type (type->field (field_num).type ()) 6339 || ada_is_interface_tag (type->field (field_num).type ()))) 6340 return 1; 6341 6342 /* Not a special field, so it should not be ignored. */ 6343 return 0; 6344 } 6345 6346 /* True iff TYPE has a tag field. If REFOK, then TYPE may also be a 6347 pointer or reference type whose ultimate target has a tag field. */ 6348 6349 int 6350 ada_is_tagged_type (struct type *type, int refok) 6351 { 6352 return (ada_lookup_struct_elt_type (type, "_tag", refok, 1) != NULL); 6353 } 6354 6355 /* True iff TYPE represents the type of X'Tag */ 6356 6357 int 6358 ada_is_tag_type (struct type *type) 6359 { 6360 type = ada_check_typedef (type); 6361 6362 if (type == NULL || type->code () != TYPE_CODE_PTR) 6363 return 0; 6364 else 6365 { 6366 const char *name = ada_type_name (type->target_type ()); 6367 6368 return (name != NULL 6369 && strcmp (name, "ada__tags__dispatch_table") == 0); 6370 } 6371 } 6372 6373 /* The type of the tag on VAL. */ 6374 6375 static struct type * 6376 ada_tag_type (struct value *val) 6377 { 6378 return ada_lookup_struct_elt_type (value_type (val), "_tag", 1, 0); 6379 } 6380 6381 /* Return 1 if TAG follows the old scheme for Ada tags (used for Ada 95, 6382 retired at Ada 05). */ 6383 6384 static int 6385 is_ada95_tag (struct value *tag) 6386 { 6387 return ada_value_struct_elt (tag, "tsd", 1) != NULL; 6388 } 6389 6390 /* The value of the tag on VAL. */ 6391 6392 static struct value * 6393 ada_value_tag (struct value *val) 6394 { 6395 return ada_value_struct_elt (val, "_tag", 0); 6396 } 6397 6398 /* The value of the tag on the object of type TYPE whose contents are 6399 saved at VALADDR, if it is non-null, or is at memory address 6400 ADDRESS. */ 6401 6402 static struct value * 6403 value_tag_from_contents_and_address (struct type *type, 6404 const gdb_byte *valaddr, 6405 CORE_ADDR address) 6406 { 6407 int tag_byte_offset; 6408 struct type *tag_type; 6409 6410 gdb::array_view<const gdb_byte> contents; 6411 if (valaddr != nullptr) 6412 contents = gdb::make_array_view (valaddr, type->length ()); 6413 struct type *resolved_type = resolve_dynamic_type (type, contents, address); 6414 if (find_struct_field ("_tag", resolved_type, 0, &tag_type, &tag_byte_offset, 6415 NULL, NULL, NULL)) 6416 { 6417 const gdb_byte *valaddr1 = ((valaddr == NULL) 6418 ? NULL 6419 : valaddr + tag_byte_offset); 6420 CORE_ADDR address1 = (address == 0) ? 0 : address + tag_byte_offset; 6421 6422 return value_from_contents_and_address (tag_type, valaddr1, address1); 6423 } 6424 return NULL; 6425 } 6426 6427 static struct type * 6428 type_from_tag (struct value *tag) 6429 { 6430 gdb::unique_xmalloc_ptr<char> type_name = ada_tag_name (tag); 6431 6432 if (type_name != NULL) 6433 return ada_find_any_type (ada_encode (type_name.get ()).c_str ()); 6434 return NULL; 6435 } 6436 6437 /* Given a value OBJ of a tagged type, return a value of this 6438 type at the base address of the object. The base address, as 6439 defined in Ada.Tags, it is the address of the primary tag of 6440 the object, and therefore where the field values of its full 6441 view can be fetched. */ 6442 6443 struct value * 6444 ada_tag_value_at_base_address (struct value *obj) 6445 { 6446 struct value *val; 6447 LONGEST offset_to_top = 0; 6448 struct type *ptr_type, *obj_type; 6449 struct value *tag; 6450 CORE_ADDR base_address; 6451 6452 obj_type = value_type (obj); 6453 6454 /* It is the responsability of the caller to deref pointers. */ 6455 6456 if (obj_type->code () == TYPE_CODE_PTR || obj_type->code () == TYPE_CODE_REF) 6457 return obj; 6458 6459 tag = ada_value_tag (obj); 6460 if (!tag) 6461 return obj; 6462 6463 /* Base addresses only appeared with Ada 05 and multiple inheritance. */ 6464 6465 if (is_ada95_tag (tag)) 6466 return obj; 6467 6468 struct type *offset_type 6469 = language_lookup_primitive_type (language_def (language_ada), 6470 target_gdbarch(), "storage_offset"); 6471 ptr_type = lookup_pointer_type (offset_type); 6472 val = value_cast (ptr_type, tag); 6473 if (!val) 6474 return obj; 6475 6476 /* It is perfectly possible that an exception be raised while 6477 trying to determine the base address, just like for the tag; 6478 see ada_tag_name for more details. We do not print the error 6479 message for the same reason. */ 6480 6481 try 6482 { 6483 offset_to_top = value_as_long (value_ind (value_ptradd (val, -2))); 6484 } 6485 6486 catch (const gdb_exception_error &e) 6487 { 6488 return obj; 6489 } 6490 6491 /* If offset is null, nothing to do. */ 6492 6493 if (offset_to_top == 0) 6494 return obj; 6495 6496 /* -1 is a special case in Ada.Tags; however, what should be done 6497 is not quite clear from the documentation. So do nothing for 6498 now. */ 6499 6500 if (offset_to_top == -1) 6501 return obj; 6502 6503 /* Storage_Offset'Last is used to indicate that a dynamic offset to 6504 top is used. In this situation the offset is stored just after 6505 the tag, in the object itself. */ 6506 ULONGEST last = (((ULONGEST) 1) << (8 * offset_type->length () - 1)) - 1; 6507 if (offset_to_top == last) 6508 { 6509 struct value *tem = value_addr (tag); 6510 tem = value_ptradd (tem, 1); 6511 tem = value_cast (ptr_type, tem); 6512 offset_to_top = value_as_long (value_ind (tem)); 6513 } 6514 6515 if (offset_to_top > 0) 6516 { 6517 /* OFFSET_TO_TOP used to be a positive value to be subtracted 6518 from the base address. This was however incompatible with 6519 C++ dispatch table: C++ uses a *negative* value to *add* 6520 to the base address. Ada's convention has therefore been 6521 changed in GNAT 19.0w 20171023: since then, C++ and Ada 6522 use the same convention. Here, we support both cases by 6523 checking the sign of OFFSET_TO_TOP. */ 6524 offset_to_top = -offset_to_top; 6525 } 6526 6527 base_address = value_address (obj) + offset_to_top; 6528 tag = value_tag_from_contents_and_address (obj_type, NULL, base_address); 6529 6530 /* Make sure that we have a proper tag at the new address. 6531 Otherwise, offset_to_top is bogus (which can happen when 6532 the object is not initialized yet). */ 6533 6534 if (!tag) 6535 return obj; 6536 6537 obj_type = type_from_tag (tag); 6538 6539 if (!obj_type) 6540 return obj; 6541 6542 return value_from_contents_and_address (obj_type, NULL, base_address); 6543 } 6544 6545 /* Return the "ada__tags__type_specific_data" type. */ 6546 6547 static struct type * 6548 ada_get_tsd_type (struct inferior *inf) 6549 { 6550 struct ada_inferior_data *data = get_ada_inferior_data (inf); 6551 6552 if (data->tsd_type == 0) 6553 data->tsd_type = ada_find_any_type ("ada__tags__type_specific_data"); 6554 return data->tsd_type; 6555 } 6556 6557 /* Return the TSD (type-specific data) associated to the given TAG. 6558 TAG is assumed to be the tag of a tagged-type entity. 6559 6560 May return NULL if we are unable to get the TSD. */ 6561 6562 static struct value * 6563 ada_get_tsd_from_tag (struct value *tag) 6564 { 6565 struct value *val; 6566 struct type *type; 6567 6568 /* First option: The TSD is simply stored as a field of our TAG. 6569 Only older versions of GNAT would use this format, but we have 6570 to test it first, because there are no visible markers for 6571 the current approach except the absence of that field. */ 6572 6573 val = ada_value_struct_elt (tag, "tsd", 1); 6574 if (val) 6575 return val; 6576 6577 /* Try the second representation for the dispatch table (in which 6578 there is no explicit 'tsd' field in the referent of the tag pointer, 6579 and instead the tsd pointer is stored just before the dispatch 6580 table. */ 6581 6582 type = ada_get_tsd_type (current_inferior()); 6583 if (type == NULL) 6584 return NULL; 6585 type = lookup_pointer_type (lookup_pointer_type (type)); 6586 val = value_cast (type, tag); 6587 if (val == NULL) 6588 return NULL; 6589 return value_ind (value_ptradd (val, -1)); 6590 } 6591 6592 /* Given the TSD of a tag (type-specific data), return a string 6593 containing the name of the associated type. 6594 6595 May return NULL if we are unable to determine the tag name. */ 6596 6597 static gdb::unique_xmalloc_ptr<char> 6598 ada_tag_name_from_tsd (struct value *tsd) 6599 { 6600 struct value *val; 6601 6602 val = ada_value_struct_elt (tsd, "expanded_name", 1); 6603 if (val == NULL) 6604 return NULL; 6605 gdb::unique_xmalloc_ptr<char> buffer 6606 = target_read_string (value_as_address (val), INT_MAX); 6607 if (buffer == nullptr) 6608 return nullptr; 6609 6610 try 6611 { 6612 /* Let this throw an exception on error. If the data is 6613 uninitialized, we'd rather not have the user see a 6614 warning. */ 6615 const char *folded = ada_fold_name (buffer.get (), true); 6616 return make_unique_xstrdup (folded); 6617 } 6618 catch (const gdb_exception &) 6619 { 6620 return nullptr; 6621 } 6622 } 6623 6624 /* The type name of the dynamic type denoted by the 'tag value TAG, as 6625 a C string. 6626 6627 Return NULL if the TAG is not an Ada tag, or if we were unable to 6628 determine the name of that tag. */ 6629 6630 gdb::unique_xmalloc_ptr<char> 6631 ada_tag_name (struct value *tag) 6632 { 6633 gdb::unique_xmalloc_ptr<char> name; 6634 6635 if (!ada_is_tag_type (value_type (tag))) 6636 return NULL; 6637 6638 /* It is perfectly possible that an exception be raised while trying 6639 to determine the TAG's name, even under normal circumstances: 6640 The associated variable may be uninitialized or corrupted, for 6641 instance. We do not let any exception propagate past this point. 6642 instead we return NULL. 6643 6644 We also do not print the error message either (which often is very 6645 low-level (Eg: "Cannot read memory at 0x[...]"), but instead let 6646 the caller print a more meaningful message if necessary. */ 6647 try 6648 { 6649 struct value *tsd = ada_get_tsd_from_tag (tag); 6650 6651 if (tsd != NULL) 6652 name = ada_tag_name_from_tsd (tsd); 6653 } 6654 catch (const gdb_exception_error &e) 6655 { 6656 } 6657 6658 return name; 6659 } 6660 6661 /* The parent type of TYPE, or NULL if none. */ 6662 6663 struct type * 6664 ada_parent_type (struct type *type) 6665 { 6666 int i; 6667 6668 type = ada_check_typedef (type); 6669 6670 if (type == NULL || type->code () != TYPE_CODE_STRUCT) 6671 return NULL; 6672 6673 for (i = 0; i < type->num_fields (); i += 1) 6674 if (ada_is_parent_field (type, i)) 6675 { 6676 struct type *parent_type = type->field (i).type (); 6677 6678 /* If the _parent field is a pointer, then dereference it. */ 6679 if (parent_type->code () == TYPE_CODE_PTR) 6680 parent_type = parent_type->target_type (); 6681 /* If there is a parallel XVS type, get the actual base type. */ 6682 parent_type = ada_get_base_type (parent_type); 6683 6684 return ada_check_typedef (parent_type); 6685 } 6686 6687 return NULL; 6688 } 6689 6690 /* True iff field number FIELD_NUM of structure type TYPE contains the 6691 parent-type (inherited) fields of a derived type. Assumes TYPE is 6692 a structure type with at least FIELD_NUM+1 fields. */ 6693 6694 int 6695 ada_is_parent_field (struct type *type, int field_num) 6696 { 6697 const char *name = ada_check_typedef (type)->field (field_num).name (); 6698 6699 return (name != NULL 6700 && (startswith (name, "PARENT") 6701 || startswith (name, "_parent"))); 6702 } 6703 6704 /* True iff field number FIELD_NUM of structure type TYPE is a 6705 transparent wrapper field (which should be silently traversed when doing 6706 field selection and flattened when printing). Assumes TYPE is a 6707 structure type with at least FIELD_NUM+1 fields. Such fields are always 6708 structures. */ 6709 6710 int 6711 ada_is_wrapper_field (struct type *type, int field_num) 6712 { 6713 const char *name = type->field (field_num).name (); 6714 6715 if (name != NULL && strcmp (name, "RETVAL") == 0) 6716 { 6717 /* This happens in functions with "out" or "in out" parameters 6718 which are passed by copy. For such functions, GNAT describes 6719 the function's return type as being a struct where the return 6720 value is in a field called RETVAL, and where the other "out" 6721 or "in out" parameters are fields of that struct. This is not 6722 a wrapper. */ 6723 return 0; 6724 } 6725 6726 return (name != NULL 6727 && (startswith (name, "PARENT") 6728 || strcmp (name, "REP") == 0 6729 || startswith (name, "_parent") 6730 || name[0] == 'S' || name[0] == 'R' || name[0] == 'O')); 6731 } 6732 6733 /* True iff field number FIELD_NUM of structure or union type TYPE 6734 is a variant wrapper. Assumes TYPE is a structure type with at least 6735 FIELD_NUM+1 fields. */ 6736 6737 int 6738 ada_is_variant_part (struct type *type, int field_num) 6739 { 6740 /* Only Ada types are eligible. */ 6741 if (!ADA_TYPE_P (type)) 6742 return 0; 6743 6744 struct type *field_type = type->field (field_num).type (); 6745 6746 return (field_type->code () == TYPE_CODE_UNION 6747 || (is_dynamic_field (type, field_num) 6748 && (field_type->target_type ()->code () 6749 == TYPE_CODE_UNION))); 6750 } 6751 6752 /* Assuming that VAR_TYPE is a variant wrapper (type of the variant part) 6753 whose discriminants are contained in the record type OUTER_TYPE, 6754 returns the type of the controlling discriminant for the variant. 6755 May return NULL if the type could not be found. */ 6756 6757 struct type * 6758 ada_variant_discrim_type (struct type *var_type, struct type *outer_type) 6759 { 6760 const char *name = ada_variant_discrim_name (var_type); 6761 6762 return ada_lookup_struct_elt_type (outer_type, name, 1, 1); 6763 } 6764 6765 /* Assuming that TYPE is the type of a variant wrapper, and FIELD_NUM is a 6766 valid field number within it, returns 1 iff field FIELD_NUM of TYPE 6767 represents a 'when others' clause; otherwise 0. */ 6768 6769 static int 6770 ada_is_others_clause (struct type *type, int field_num) 6771 { 6772 const char *name = type->field (field_num).name (); 6773 6774 return (name != NULL && name[0] == 'O'); 6775 } 6776 6777 /* Assuming that TYPE0 is the type of the variant part of a record, 6778 returns the name of the discriminant controlling the variant. 6779 The value is valid until the next call to ada_variant_discrim_name. */ 6780 6781 const char * 6782 ada_variant_discrim_name (struct type *type0) 6783 { 6784 static std::string result; 6785 struct type *type; 6786 const char *name; 6787 const char *discrim_end; 6788 const char *discrim_start; 6789 6790 if (type0->code () == TYPE_CODE_PTR) 6791 type = type0->target_type (); 6792 else 6793 type = type0; 6794 6795 name = ada_type_name (type); 6796 6797 if (name == NULL || name[0] == '\000') 6798 return ""; 6799 6800 for (discrim_end = name + strlen (name) - 6; discrim_end != name; 6801 discrim_end -= 1) 6802 { 6803 if (startswith (discrim_end, "___XVN")) 6804 break; 6805 } 6806 if (discrim_end == name) 6807 return ""; 6808 6809 for (discrim_start = discrim_end; discrim_start != name + 3; 6810 discrim_start -= 1) 6811 { 6812 if (discrim_start == name + 1) 6813 return ""; 6814 if ((discrim_start > name + 3 6815 && startswith (discrim_start - 3, "___")) 6816 || discrim_start[-1] == '.') 6817 break; 6818 } 6819 6820 result = std::string (discrim_start, discrim_end - discrim_start); 6821 return result.c_str (); 6822 } 6823 6824 /* Scan STR for a subtype-encoded number, beginning at position K. 6825 Put the position of the character just past the number scanned in 6826 *NEW_K, if NEW_K!=NULL. Put the scanned number in *R, if R!=NULL. 6827 Return 1 if there was a valid number at the given position, and 0 6828 otherwise. A "subtype-encoded" number consists of the absolute value 6829 in decimal, followed by the letter 'm' to indicate a negative number. 6830 Assumes 0m does not occur. */ 6831 6832 int 6833 ada_scan_number (const char str[], int k, LONGEST * R, int *new_k) 6834 { 6835 ULONGEST RU; 6836 6837 if (!isdigit (str[k])) 6838 return 0; 6839 6840 /* Do it the hard way so as not to make any assumption about 6841 the relationship of unsigned long (%lu scan format code) and 6842 LONGEST. */ 6843 RU = 0; 6844 while (isdigit (str[k])) 6845 { 6846 RU = RU * 10 + (str[k] - '0'); 6847 k += 1; 6848 } 6849 6850 if (str[k] == 'm') 6851 { 6852 if (R != NULL) 6853 *R = (-(LONGEST) (RU - 1)) - 1; 6854 k += 1; 6855 } 6856 else if (R != NULL) 6857 *R = (LONGEST) RU; 6858 6859 /* NOTE on the above: Technically, C does not say what the results of 6860 - (LONGEST) RU or (LONGEST) -RU are for RU == largest positive 6861 number representable as a LONGEST (although either would probably work 6862 in most implementations). When RU>0, the locution in the then branch 6863 above is always equivalent to the negative of RU. */ 6864 6865 if (new_k != NULL) 6866 *new_k = k; 6867 return 1; 6868 } 6869 6870 /* Assuming that TYPE is a variant part wrapper type (a VARIANTS field), 6871 and FIELD_NUM is a valid field number within it, returns 1 iff VAL is 6872 in the range encoded by field FIELD_NUM of TYPE; otherwise 0. */ 6873 6874 static int 6875 ada_in_variant (LONGEST val, struct type *type, int field_num) 6876 { 6877 const char *name = type->field (field_num).name (); 6878 int p; 6879 6880 p = 0; 6881 while (1) 6882 { 6883 switch (name[p]) 6884 { 6885 case '\0': 6886 return 0; 6887 case 'S': 6888 { 6889 LONGEST W; 6890 6891 if (!ada_scan_number (name, p + 1, &W, &p)) 6892 return 0; 6893 if (val == W) 6894 return 1; 6895 break; 6896 } 6897 case 'R': 6898 { 6899 LONGEST L, U; 6900 6901 if (!ada_scan_number (name, p + 1, &L, &p) 6902 || name[p] != 'T' || !ada_scan_number (name, p + 1, &U, &p)) 6903 return 0; 6904 if (val >= L && val <= U) 6905 return 1; 6906 break; 6907 } 6908 case 'O': 6909 return 1; 6910 default: 6911 return 0; 6912 } 6913 } 6914 } 6915 6916 /* FIXME: Lots of redundancy below. Try to consolidate. */ 6917 6918 /* Given a value ARG1 (offset by OFFSET bytes) of a struct or union type 6919 ARG_TYPE, extract and return the value of one of its (non-static) 6920 fields. FIELDNO says which field. Differs from value_primitive_field 6921 only in that it can handle packed values of arbitrary type. */ 6922 6923 struct value * 6924 ada_value_primitive_field (struct value *arg1, int offset, int fieldno, 6925 struct type *arg_type) 6926 { 6927 struct type *type; 6928 6929 arg_type = ada_check_typedef (arg_type); 6930 type = arg_type->field (fieldno).type (); 6931 6932 /* Handle packed fields. It might be that the field is not packed 6933 relative to its containing structure, but the structure itself is 6934 packed; in this case we must take the bit-field path. */ 6935 if (TYPE_FIELD_BITSIZE (arg_type, fieldno) != 0 || value_bitpos (arg1) != 0) 6936 { 6937 int bit_pos = arg_type->field (fieldno).loc_bitpos (); 6938 int bit_size = TYPE_FIELD_BITSIZE (arg_type, fieldno); 6939 6940 return ada_value_primitive_packed_val (arg1, 6941 value_contents (arg1).data (), 6942 offset + bit_pos / 8, 6943 bit_pos % 8, bit_size, type); 6944 } 6945 else 6946 return value_primitive_field (arg1, offset, fieldno, arg_type); 6947 } 6948 6949 /* Find field with name NAME in object of type TYPE. If found, 6950 set the following for each argument that is non-null: 6951 - *FIELD_TYPE_P to the field's type; 6952 - *BYTE_OFFSET_P to OFFSET + the byte offset of the field within 6953 an object of that type; 6954 - *BIT_OFFSET_P to the bit offset modulo byte size of the field; 6955 - *BIT_SIZE_P to its size in bits if the field is packed, and 6956 0 otherwise; 6957 If INDEX_P is non-null, increment *INDEX_P by the number of source-visible 6958 fields up to but not including the desired field, or by the total 6959 number of fields if not found. A NULL value of NAME never 6960 matches; the function just counts visible fields in this case. 6961 6962 Notice that we need to handle when a tagged record hierarchy 6963 has some components with the same name, like in this scenario: 6964 6965 type Top_T is tagged record 6966 N : Integer := 1; 6967 U : Integer := 974; 6968 A : Integer := 48; 6969 end record; 6970 6971 type Middle_T is new Top.Top_T with record 6972 N : Character := 'a'; 6973 C : Integer := 3; 6974 end record; 6975 6976 type Bottom_T is new Middle.Middle_T with record 6977 N : Float := 4.0; 6978 C : Character := '5'; 6979 X : Integer := 6; 6980 A : Character := 'J'; 6981 end record; 6982 6983 Let's say we now have a variable declared and initialized as follow: 6984 6985 TC : Top_A := new Bottom_T; 6986 6987 And then we use this variable to call this function 6988 6989 procedure Assign (Obj: in out Top_T; TV : Integer); 6990 6991 as follow: 6992 6993 Assign (Top_T (B), 12); 6994 6995 Now, we're in the debugger, and we're inside that procedure 6996 then and we want to print the value of obj.c: 6997 6998 Usually, the tagged record or one of the parent type owns the 6999 component to print and there's no issue but in this particular 7000 case, what does it mean to ask for Obj.C? Since the actual 7001 type for object is type Bottom_T, it could mean two things: type 7002 component C from the Middle_T view, but also component C from 7003 Bottom_T. So in that "undefined" case, when the component is 7004 not found in the non-resolved type (which includes all the 7005 components of the parent type), then resolve it and see if we 7006 get better luck once expanded. 7007 7008 In the case of homonyms in the derived tagged type, we don't 7009 guaranty anything, and pick the one that's easiest for us 7010 to program. 7011 7012 Returns 1 if found, 0 otherwise. */ 7013 7014 static int 7015 find_struct_field (const char *name, struct type *type, int offset, 7016 struct type **field_type_p, 7017 int *byte_offset_p, int *bit_offset_p, int *bit_size_p, 7018 int *index_p) 7019 { 7020 int i; 7021 int parent_offset = -1; 7022 7023 type = ada_check_typedef (type); 7024 7025 if (field_type_p != NULL) 7026 *field_type_p = NULL; 7027 if (byte_offset_p != NULL) 7028 *byte_offset_p = 0; 7029 if (bit_offset_p != NULL) 7030 *bit_offset_p = 0; 7031 if (bit_size_p != NULL) 7032 *bit_size_p = 0; 7033 7034 for (i = 0; i < type->num_fields (); i += 1) 7035 { 7036 /* These can't be computed using TYPE_FIELD_BITPOS for a dynamic 7037 type. However, we only need the values to be correct when 7038 the caller asks for them. */ 7039 int bit_pos = 0, fld_offset = 0; 7040 if (byte_offset_p != nullptr || bit_offset_p != nullptr) 7041 { 7042 bit_pos = type->field (i).loc_bitpos (); 7043 fld_offset = offset + bit_pos / 8; 7044 } 7045 7046 const char *t_field_name = type->field (i).name (); 7047 7048 if (t_field_name == NULL) 7049 continue; 7050 7051 else if (ada_is_parent_field (type, i)) 7052 { 7053 /* This is a field pointing us to the parent type of a tagged 7054 type. As hinted in this function's documentation, we give 7055 preference to fields in the current record first, so what 7056 we do here is just record the index of this field before 7057 we skip it. If it turns out we couldn't find our field 7058 in the current record, then we'll get back to it and search 7059 inside it whether the field might exist in the parent. */ 7060 7061 parent_offset = i; 7062 continue; 7063 } 7064 7065 else if (name != NULL && field_name_match (t_field_name, name)) 7066 { 7067 int bit_size = TYPE_FIELD_BITSIZE (type, i); 7068 7069 if (field_type_p != NULL) 7070 *field_type_p = type->field (i).type (); 7071 if (byte_offset_p != NULL) 7072 *byte_offset_p = fld_offset; 7073 if (bit_offset_p != NULL) 7074 *bit_offset_p = bit_pos % 8; 7075 if (bit_size_p != NULL) 7076 *bit_size_p = bit_size; 7077 return 1; 7078 } 7079 else if (ada_is_wrapper_field (type, i)) 7080 { 7081 if (find_struct_field (name, type->field (i).type (), fld_offset, 7082 field_type_p, byte_offset_p, bit_offset_p, 7083 bit_size_p, index_p)) 7084 return 1; 7085 } 7086 else if (ada_is_variant_part (type, i)) 7087 { 7088 /* PNH: Wait. Do we ever execute this section, or is ARG always of 7089 fixed type?? */ 7090 int j; 7091 struct type *field_type 7092 = ada_check_typedef (type->field (i).type ()); 7093 7094 for (j = 0; j < field_type->num_fields (); j += 1) 7095 { 7096 if (find_struct_field (name, field_type->field (j).type (), 7097 fld_offset 7098 + field_type->field (j).loc_bitpos () / 8, 7099 field_type_p, byte_offset_p, 7100 bit_offset_p, bit_size_p, index_p)) 7101 return 1; 7102 } 7103 } 7104 else if (index_p != NULL) 7105 *index_p += 1; 7106 } 7107 7108 /* Field not found so far. If this is a tagged type which 7109 has a parent, try finding that field in the parent now. */ 7110 7111 if (parent_offset != -1) 7112 { 7113 /* As above, only compute the offset when truly needed. */ 7114 int fld_offset = offset; 7115 if (byte_offset_p != nullptr || bit_offset_p != nullptr) 7116 { 7117 int bit_pos = type->field (parent_offset).loc_bitpos (); 7118 fld_offset += bit_pos / 8; 7119 } 7120 7121 if (find_struct_field (name, type->field (parent_offset).type (), 7122 fld_offset, field_type_p, byte_offset_p, 7123 bit_offset_p, bit_size_p, index_p)) 7124 return 1; 7125 } 7126 7127 return 0; 7128 } 7129 7130 /* Number of user-visible fields in record type TYPE. */ 7131 7132 static int 7133 num_visible_fields (struct type *type) 7134 { 7135 int n; 7136 7137 n = 0; 7138 find_struct_field (NULL, type, 0, NULL, NULL, NULL, NULL, &n); 7139 return n; 7140 } 7141 7142 /* Look for a field NAME in ARG. Adjust the address of ARG by OFFSET bytes, 7143 and search in it assuming it has (class) type TYPE. 7144 If found, return value, else return NULL. 7145 7146 Searches recursively through wrapper fields (e.g., '_parent'). 7147 7148 In the case of homonyms in the tagged types, please refer to the 7149 long explanation in find_struct_field's function documentation. */ 7150 7151 static struct value * 7152 ada_search_struct_field (const char *name, struct value *arg, int offset, 7153 struct type *type) 7154 { 7155 int i; 7156 int parent_offset = -1; 7157 7158 type = ada_check_typedef (type); 7159 for (i = 0; i < type->num_fields (); i += 1) 7160 { 7161 const char *t_field_name = type->field (i).name (); 7162 7163 if (t_field_name == NULL) 7164 continue; 7165 7166 else if (ada_is_parent_field (type, i)) 7167 { 7168 /* This is a field pointing us to the parent type of a tagged 7169 type. As hinted in this function's documentation, we give 7170 preference to fields in the current record first, so what 7171 we do here is just record the index of this field before 7172 we skip it. If it turns out we couldn't find our field 7173 in the current record, then we'll get back to it and search 7174 inside it whether the field might exist in the parent. */ 7175 7176 parent_offset = i; 7177 continue; 7178 } 7179 7180 else if (field_name_match (t_field_name, name)) 7181 return ada_value_primitive_field (arg, offset, i, type); 7182 7183 else if (ada_is_wrapper_field (type, i)) 7184 { 7185 struct value *v = /* Do not let indent join lines here. */ 7186 ada_search_struct_field (name, arg, 7187 offset + type->field (i).loc_bitpos () / 8, 7188 type->field (i).type ()); 7189 7190 if (v != NULL) 7191 return v; 7192 } 7193 7194 else if (ada_is_variant_part (type, i)) 7195 { 7196 /* PNH: Do we ever get here? See find_struct_field. */ 7197 int j; 7198 struct type *field_type = ada_check_typedef (type->field (i).type ()); 7199 int var_offset = offset + type->field (i).loc_bitpos () / 8; 7200 7201 for (j = 0; j < field_type->num_fields (); j += 1) 7202 { 7203 struct value *v = ada_search_struct_field /* Force line 7204 break. */ 7205 (name, arg, 7206 var_offset + field_type->field (j).loc_bitpos () / 8, 7207 field_type->field (j).type ()); 7208 7209 if (v != NULL) 7210 return v; 7211 } 7212 } 7213 } 7214 7215 /* Field not found so far. If this is a tagged type which 7216 has a parent, try finding that field in the parent now. */ 7217 7218 if (parent_offset != -1) 7219 { 7220 struct value *v = ada_search_struct_field ( 7221 name, arg, offset + type->field (parent_offset).loc_bitpos () / 8, 7222 type->field (parent_offset).type ()); 7223 7224 if (v != NULL) 7225 return v; 7226 } 7227 7228 return NULL; 7229 } 7230 7231 static struct value *ada_index_struct_field_1 (int *, struct value *, 7232 int, struct type *); 7233 7234 7235 /* Return field #INDEX in ARG, where the index is that returned by 7236 * find_struct_field through its INDEX_P argument. Adjust the address 7237 * of ARG by OFFSET bytes, and search in it assuming it has (class) type TYPE. 7238 * If found, return value, else return NULL. */ 7239 7240 static struct value * 7241 ada_index_struct_field (int index, struct value *arg, int offset, 7242 struct type *type) 7243 { 7244 return ada_index_struct_field_1 (&index, arg, offset, type); 7245 } 7246 7247 7248 /* Auxiliary function for ada_index_struct_field. Like 7249 * ada_index_struct_field, but takes index from *INDEX_P and modifies 7250 * *INDEX_P. */ 7251 7252 static struct value * 7253 ada_index_struct_field_1 (int *index_p, struct value *arg, int offset, 7254 struct type *type) 7255 { 7256 int i; 7257 type = ada_check_typedef (type); 7258 7259 for (i = 0; i < type->num_fields (); i += 1) 7260 { 7261 if (type->field (i).name () == NULL) 7262 continue; 7263 else if (ada_is_wrapper_field (type, i)) 7264 { 7265 struct value *v = /* Do not let indent join lines here. */ 7266 ada_index_struct_field_1 (index_p, arg, 7267 offset + type->field (i).loc_bitpos () / 8, 7268 type->field (i).type ()); 7269 7270 if (v != NULL) 7271 return v; 7272 } 7273 7274 else if (ada_is_variant_part (type, i)) 7275 { 7276 /* PNH: Do we ever get here? See ada_search_struct_field, 7277 find_struct_field. */ 7278 error (_("Cannot assign this kind of variant record")); 7279 } 7280 else if (*index_p == 0) 7281 return ada_value_primitive_field (arg, offset, i, type); 7282 else 7283 *index_p -= 1; 7284 } 7285 return NULL; 7286 } 7287 7288 /* Return a string representation of type TYPE. */ 7289 7290 static std::string 7291 type_as_string (struct type *type) 7292 { 7293 string_file tmp_stream; 7294 7295 type_print (type, "", &tmp_stream, -1); 7296 7297 return tmp_stream.release (); 7298 } 7299 7300 /* Given a type TYPE, look up the type of the component of type named NAME. 7301 If DISPP is non-null, add its byte displacement from the beginning of a 7302 structure (pointed to by a value) of type TYPE to *DISPP (does not 7303 work for packed fields). 7304 7305 Matches any field whose name has NAME as a prefix, possibly 7306 followed by "___". 7307 7308 TYPE can be either a struct or union. If REFOK, TYPE may also 7309 be a (pointer or reference)+ to a struct or union, and the 7310 ultimate target type will be searched. 7311 7312 Looks recursively into variant clauses and parent types. 7313 7314 In the case of homonyms in the tagged types, please refer to the 7315 long explanation in find_struct_field's function documentation. 7316 7317 If NOERR is nonzero, return NULL if NAME is not suitably defined or 7318 TYPE is not a type of the right kind. */ 7319 7320 static struct type * 7321 ada_lookup_struct_elt_type (struct type *type, const char *name, int refok, 7322 int noerr) 7323 { 7324 int i; 7325 int parent_offset = -1; 7326 7327 if (name == NULL) 7328 goto BadName; 7329 7330 if (refok && type != NULL) 7331 while (1) 7332 { 7333 type = ada_check_typedef (type); 7334 if (type->code () != TYPE_CODE_PTR && type->code () != TYPE_CODE_REF) 7335 break; 7336 type = type->target_type (); 7337 } 7338 7339 if (type == NULL 7340 || (type->code () != TYPE_CODE_STRUCT 7341 && type->code () != TYPE_CODE_UNION)) 7342 { 7343 if (noerr) 7344 return NULL; 7345 7346 error (_("Type %s is not a structure or union type"), 7347 type != NULL ? type_as_string (type).c_str () : _("(null)")); 7348 } 7349 7350 type = to_static_fixed_type (type); 7351 7352 for (i = 0; i < type->num_fields (); i += 1) 7353 { 7354 const char *t_field_name = type->field (i).name (); 7355 struct type *t; 7356 7357 if (t_field_name == NULL) 7358 continue; 7359 7360 else if (ada_is_parent_field (type, i)) 7361 { 7362 /* This is a field pointing us to the parent type of a tagged 7363 type. As hinted in this function's documentation, we give 7364 preference to fields in the current record first, so what 7365 we do here is just record the index of this field before 7366 we skip it. If it turns out we couldn't find our field 7367 in the current record, then we'll get back to it and search 7368 inside it whether the field might exist in the parent. */ 7369 7370 parent_offset = i; 7371 continue; 7372 } 7373 7374 else if (field_name_match (t_field_name, name)) 7375 return type->field (i).type (); 7376 7377 else if (ada_is_wrapper_field (type, i)) 7378 { 7379 t = ada_lookup_struct_elt_type (type->field (i).type (), name, 7380 0, 1); 7381 if (t != NULL) 7382 return t; 7383 } 7384 7385 else if (ada_is_variant_part (type, i)) 7386 { 7387 int j; 7388 struct type *field_type = ada_check_typedef (type->field (i).type ()); 7389 7390 for (j = field_type->num_fields () - 1; j >= 0; j -= 1) 7391 { 7392 /* FIXME pnh 2008/01/26: We check for a field that is 7393 NOT wrapped in a struct, since the compiler sometimes 7394 generates these for unchecked variant types. Revisit 7395 if the compiler changes this practice. */ 7396 const char *v_field_name = field_type->field (j).name (); 7397 7398 if (v_field_name != NULL 7399 && field_name_match (v_field_name, name)) 7400 t = field_type->field (j).type (); 7401 else 7402 t = ada_lookup_struct_elt_type (field_type->field (j).type (), 7403 name, 0, 1); 7404 7405 if (t != NULL) 7406 return t; 7407 } 7408 } 7409 7410 } 7411 7412 /* Field not found so far. If this is a tagged type which 7413 has a parent, try finding that field in the parent now. */ 7414 7415 if (parent_offset != -1) 7416 { 7417 struct type *t; 7418 7419 t = ada_lookup_struct_elt_type (type->field (parent_offset).type (), 7420 name, 0, 1); 7421 if (t != NULL) 7422 return t; 7423 } 7424 7425 BadName: 7426 if (!noerr) 7427 { 7428 const char *name_str = name != NULL ? name : _("<null>"); 7429 7430 error (_("Type %s has no component named %s"), 7431 type_as_string (type).c_str (), name_str); 7432 } 7433 7434 return NULL; 7435 } 7436 7437 /* Assuming that VAR_TYPE is the type of a variant part of a record (a union), 7438 within a value of type OUTER_TYPE, return true iff VAR_TYPE 7439 represents an unchecked union (that is, the variant part of a 7440 record that is named in an Unchecked_Union pragma). */ 7441 7442 static int 7443 is_unchecked_variant (struct type *var_type, struct type *outer_type) 7444 { 7445 const char *discrim_name = ada_variant_discrim_name (var_type); 7446 7447 return (ada_lookup_struct_elt_type (outer_type, discrim_name, 0, 1) == NULL); 7448 } 7449 7450 7451 /* Assuming that VAR_TYPE is the type of a variant part of a record (a union), 7452 within OUTER, determine which variant clause (field number in VAR_TYPE, 7453 numbering from 0) is applicable. Returns -1 if none are. */ 7454 7455 int 7456 ada_which_variant_applies (struct type *var_type, struct value *outer) 7457 { 7458 int others_clause; 7459 int i; 7460 const char *discrim_name = ada_variant_discrim_name (var_type); 7461 struct value *discrim; 7462 LONGEST discrim_val; 7463 7464 /* Using plain value_from_contents_and_address here causes problems 7465 because we will end up trying to resolve a type that is currently 7466 being constructed. */ 7467 discrim = ada_value_struct_elt (outer, discrim_name, 1); 7468 if (discrim == NULL) 7469 return -1; 7470 discrim_val = value_as_long (discrim); 7471 7472 others_clause = -1; 7473 for (i = 0; i < var_type->num_fields (); i += 1) 7474 { 7475 if (ada_is_others_clause (var_type, i)) 7476 others_clause = i; 7477 else if (ada_in_variant (discrim_val, var_type, i)) 7478 return i; 7479 } 7480 7481 return others_clause; 7482 } 7483 7484 7485 7486 /* Dynamic-Sized Records */ 7487 7488 /* Strategy: The type ostensibly attached to a value with dynamic size 7489 (i.e., a size that is not statically recorded in the debugging 7490 data) does not accurately reflect the size or layout of the value. 7491 Our strategy is to convert these values to values with accurate, 7492 conventional types that are constructed on the fly. */ 7493 7494 /* There is a subtle and tricky problem here. In general, we cannot 7495 determine the size of dynamic records without its data. However, 7496 the 'struct value' data structure, which GDB uses to represent 7497 quantities in the inferior process (the target), requires the size 7498 of the type at the time of its allocation in order to reserve space 7499 for GDB's internal copy of the data. That's why the 7500 'to_fixed_xxx_type' routines take (target) addresses as parameters, 7501 rather than struct value*s. 7502 7503 However, GDB's internal history variables ($1, $2, etc.) are 7504 struct value*s containing internal copies of the data that are not, in 7505 general, the same as the data at their corresponding addresses in 7506 the target. Fortunately, the types we give to these values are all 7507 conventional, fixed-size types (as per the strategy described 7508 above), so that we don't usually have to perform the 7509 'to_fixed_xxx_type' conversions to look at their values. 7510 Unfortunately, there is one exception: if one of the internal 7511 history variables is an array whose elements are unconstrained 7512 records, then we will need to create distinct fixed types for each 7513 element selected. */ 7514 7515 /* The upshot of all of this is that many routines take a (type, host 7516 address, target address) triple as arguments to represent a value. 7517 The host address, if non-null, is supposed to contain an internal 7518 copy of the relevant data; otherwise, the program is to consult the 7519 target at the target address. */ 7520 7521 /* Assuming that VAL0 represents a pointer value, the result of 7522 dereferencing it. Differs from value_ind in its treatment of 7523 dynamic-sized types. */ 7524 7525 struct value * 7526 ada_value_ind (struct value *val0) 7527 { 7528 struct value *val = value_ind (val0); 7529 7530 if (ada_is_tagged_type (value_type (val), 0)) 7531 val = ada_tag_value_at_base_address (val); 7532 7533 return ada_to_fixed_value (val); 7534 } 7535 7536 /* The value resulting from dereferencing any "reference to" 7537 qualifiers on VAL0. */ 7538 7539 static struct value * 7540 ada_coerce_ref (struct value *val0) 7541 { 7542 if (value_type (val0)->code () == TYPE_CODE_REF) 7543 { 7544 struct value *val = val0; 7545 7546 val = coerce_ref (val); 7547 7548 if (ada_is_tagged_type (value_type (val), 0)) 7549 val = ada_tag_value_at_base_address (val); 7550 7551 return ada_to_fixed_value (val); 7552 } 7553 else 7554 return val0; 7555 } 7556 7557 /* Return the bit alignment required for field #F of template type TYPE. */ 7558 7559 static unsigned int 7560 field_alignment (struct type *type, int f) 7561 { 7562 const char *name = type->field (f).name (); 7563 int len; 7564 int align_offset; 7565 7566 /* The field name should never be null, unless the debugging information 7567 is somehow malformed. In this case, we assume the field does not 7568 require any alignment. */ 7569 if (name == NULL) 7570 return 1; 7571 7572 len = strlen (name); 7573 7574 if (!isdigit (name[len - 1])) 7575 return 1; 7576 7577 if (isdigit (name[len - 2])) 7578 align_offset = len - 2; 7579 else 7580 align_offset = len - 1; 7581 7582 if (align_offset < 7 || !startswith (name + align_offset - 6, "___XV")) 7583 return TARGET_CHAR_BIT; 7584 7585 return atoi (name + align_offset) * TARGET_CHAR_BIT; 7586 } 7587 7588 /* Find a typedef or tag symbol named NAME. Ignores ambiguity. */ 7589 7590 static struct symbol * 7591 ada_find_any_type_symbol (const char *name) 7592 { 7593 struct symbol *sym; 7594 7595 sym = standard_lookup (name, get_selected_block (NULL), VAR_DOMAIN); 7596 if (sym != NULL && sym->aclass () == LOC_TYPEDEF) 7597 return sym; 7598 7599 sym = standard_lookup (name, NULL, STRUCT_DOMAIN); 7600 return sym; 7601 } 7602 7603 /* Find a type named NAME. Ignores ambiguity. This routine will look 7604 solely for types defined by debug info, it will not search the GDB 7605 primitive types. */ 7606 7607 static struct type * 7608 ada_find_any_type (const char *name) 7609 { 7610 struct symbol *sym = ada_find_any_type_symbol (name); 7611 7612 if (sym != NULL) 7613 return sym->type (); 7614 7615 return NULL; 7616 } 7617 7618 /* Given NAME_SYM and an associated BLOCK, find a "renaming" symbol 7619 associated with NAME_SYM's name. NAME_SYM may itself be a renaming 7620 symbol, in which case it is returned. Otherwise, this looks for 7621 symbols whose name is that of NAME_SYM suffixed with "___XR". 7622 Return symbol if found, and NULL otherwise. */ 7623 7624 static bool 7625 ada_is_renaming_symbol (struct symbol *name_sym) 7626 { 7627 const char *name = name_sym->linkage_name (); 7628 return strstr (name, "___XR") != NULL; 7629 } 7630 7631 /* Because of GNAT encoding conventions, several GDB symbols may match a 7632 given type name. If the type denoted by TYPE0 is to be preferred to 7633 that of TYPE1 for purposes of type printing, return non-zero; 7634 otherwise return 0. */ 7635 7636 int 7637 ada_prefer_type (struct type *type0, struct type *type1) 7638 { 7639 if (type1 == NULL) 7640 return 1; 7641 else if (type0 == NULL) 7642 return 0; 7643 else if (type1->code () == TYPE_CODE_VOID) 7644 return 1; 7645 else if (type0->code () == TYPE_CODE_VOID) 7646 return 0; 7647 else if (type1->name () == NULL && type0->name () != NULL) 7648 return 1; 7649 else if (ada_is_constrained_packed_array_type (type0)) 7650 return 1; 7651 else if (ada_is_array_descriptor_type (type0) 7652 && !ada_is_array_descriptor_type (type1)) 7653 return 1; 7654 else 7655 { 7656 const char *type0_name = type0->name (); 7657 const char *type1_name = type1->name (); 7658 7659 if (type0_name != NULL && strstr (type0_name, "___XR") != NULL 7660 && (type1_name == NULL || strstr (type1_name, "___XR") == NULL)) 7661 return 1; 7662 } 7663 return 0; 7664 } 7665 7666 /* The name of TYPE, which is its TYPE_NAME. Null if TYPE is 7667 null. */ 7668 7669 const char * 7670 ada_type_name (struct type *type) 7671 { 7672 if (type == NULL) 7673 return NULL; 7674 return type->name (); 7675 } 7676 7677 /* Search the list of "descriptive" types associated to TYPE for a type 7678 whose name is NAME. */ 7679 7680 static struct type * 7681 find_parallel_type_by_descriptive_type (struct type *type, const char *name) 7682 { 7683 struct type *result, *tmp; 7684 7685 if (ada_ignore_descriptive_types_p) 7686 return NULL; 7687 7688 /* If there no descriptive-type info, then there is no parallel type 7689 to be found. */ 7690 if (!HAVE_GNAT_AUX_INFO (type)) 7691 return NULL; 7692 7693 result = TYPE_DESCRIPTIVE_TYPE (type); 7694 while (result != NULL) 7695 { 7696 const char *result_name = ada_type_name (result); 7697 7698 if (result_name == NULL) 7699 { 7700 warning (_("unexpected null name on descriptive type")); 7701 return NULL; 7702 } 7703 7704 /* If the names match, stop. */ 7705 if (strcmp (result_name, name) == 0) 7706 break; 7707 7708 /* Otherwise, look at the next item on the list, if any. */ 7709 if (HAVE_GNAT_AUX_INFO (result)) 7710 tmp = TYPE_DESCRIPTIVE_TYPE (result); 7711 else 7712 tmp = NULL; 7713 7714 /* If not found either, try after having resolved the typedef. */ 7715 if (tmp != NULL) 7716 result = tmp; 7717 else 7718 { 7719 result = check_typedef (result); 7720 if (HAVE_GNAT_AUX_INFO (result)) 7721 result = TYPE_DESCRIPTIVE_TYPE (result); 7722 else 7723 result = NULL; 7724 } 7725 } 7726 7727 /* If we didn't find a match, see whether this is a packed array. With 7728 older compilers, the descriptive type information is either absent or 7729 irrelevant when it comes to packed arrays so the above lookup fails. 7730 Fall back to using a parallel lookup by name in this case. */ 7731 if (result == NULL && ada_is_constrained_packed_array_type (type)) 7732 return ada_find_any_type (name); 7733 7734 return result; 7735 } 7736 7737 /* Find a parallel type to TYPE with the specified NAME, using the 7738 descriptive type taken from the debugging information, if available, 7739 and otherwise using the (slower) name-based method. */ 7740 7741 static struct type * 7742 ada_find_parallel_type_with_name (struct type *type, const char *name) 7743 { 7744 struct type *result = NULL; 7745 7746 if (HAVE_GNAT_AUX_INFO (type)) 7747 result = find_parallel_type_by_descriptive_type (type, name); 7748 else 7749 result = ada_find_any_type (name); 7750 7751 return result; 7752 } 7753 7754 /* Same as above, but specify the name of the parallel type by appending 7755 SUFFIX to the name of TYPE. */ 7756 7757 struct type * 7758 ada_find_parallel_type (struct type *type, const char *suffix) 7759 { 7760 char *name; 7761 const char *type_name = ada_type_name (type); 7762 int len; 7763 7764 if (type_name == NULL) 7765 return NULL; 7766 7767 len = strlen (type_name); 7768 7769 name = (char *) alloca (len + strlen (suffix) + 1); 7770 7771 strcpy (name, type_name); 7772 strcpy (name + len, suffix); 7773 7774 return ada_find_parallel_type_with_name (type, name); 7775 } 7776 7777 /* If TYPE is a variable-size record type, return the corresponding template 7778 type describing its fields. Otherwise, return NULL. */ 7779 7780 static struct type * 7781 dynamic_template_type (struct type *type) 7782 { 7783 type = ada_check_typedef (type); 7784 7785 if (type == NULL || type->code () != TYPE_CODE_STRUCT 7786 || ada_type_name (type) == NULL) 7787 return NULL; 7788 else 7789 { 7790 int len = strlen (ada_type_name (type)); 7791 7792 if (len > 6 && strcmp (ada_type_name (type) + len - 6, "___XVE") == 0) 7793 return type; 7794 else 7795 return ada_find_parallel_type (type, "___XVE"); 7796 } 7797 } 7798 7799 /* Assuming that TEMPL_TYPE is a union or struct type, returns 7800 non-zero iff field FIELD_NUM of TEMPL_TYPE has dynamic size. */ 7801 7802 static int 7803 is_dynamic_field (struct type *templ_type, int field_num) 7804 { 7805 const char *name = templ_type->field (field_num).name (); 7806 7807 return name != NULL 7808 && templ_type->field (field_num).type ()->code () == TYPE_CODE_PTR 7809 && strstr (name, "___XVL") != NULL; 7810 } 7811 7812 /* The index of the variant field of TYPE, or -1 if TYPE does not 7813 represent a variant record type. */ 7814 7815 static int 7816 variant_field_index (struct type *type) 7817 { 7818 int f; 7819 7820 if (type == NULL || type->code () != TYPE_CODE_STRUCT) 7821 return -1; 7822 7823 for (f = 0; f < type->num_fields (); f += 1) 7824 { 7825 if (ada_is_variant_part (type, f)) 7826 return f; 7827 } 7828 return -1; 7829 } 7830 7831 /* A record type with no fields. */ 7832 7833 static struct type * 7834 empty_record (struct type *templ) 7835 { 7836 struct type *type = alloc_type_copy (templ); 7837 7838 type->set_code (TYPE_CODE_STRUCT); 7839 INIT_NONE_SPECIFIC (type); 7840 type->set_name ("<empty>"); 7841 type->set_length (0); 7842 return type; 7843 } 7844 7845 /* An ordinary record type (with fixed-length fields) that describes 7846 the value of type TYPE at VALADDR or ADDRESS (see comments at 7847 the beginning of this section) VAL according to GNAT conventions. 7848 DVAL0 should describe the (portion of a) record that contains any 7849 necessary discriminants. It should be NULL if value_type (VAL) is 7850 an outer-level type (i.e., as opposed to a branch of a variant.) A 7851 variant field (unless unchecked) is replaced by a particular branch 7852 of the variant. 7853 7854 If not KEEP_DYNAMIC_FIELDS, then all fields whose position or 7855 length are not statically known are discarded. As a consequence, 7856 VALADDR, ADDRESS and DVAL0 are ignored. 7857 7858 NOTE: Limitations: For now, we assume that dynamic fields and 7859 variants occupy whole numbers of bytes. However, they need not be 7860 byte-aligned. */ 7861 7862 struct type * 7863 ada_template_to_fixed_record_type_1 (struct type *type, 7864 const gdb_byte *valaddr, 7865 CORE_ADDR address, struct value *dval0, 7866 int keep_dynamic_fields) 7867 { 7868 struct value *dval; 7869 struct type *rtype; 7870 int nfields, bit_len; 7871 int variant_field; 7872 long off; 7873 int fld_bit_len; 7874 int f; 7875 7876 scoped_value_mark mark; 7877 7878 /* Compute the number of fields in this record type that are going 7879 to be processed: unless keep_dynamic_fields, this includes only 7880 fields whose position and length are static will be processed. */ 7881 if (keep_dynamic_fields) 7882 nfields = type->num_fields (); 7883 else 7884 { 7885 nfields = 0; 7886 while (nfields < type->num_fields () 7887 && !ada_is_variant_part (type, nfields) 7888 && !is_dynamic_field (type, nfields)) 7889 nfields++; 7890 } 7891 7892 rtype = alloc_type_copy (type); 7893 rtype->set_code (TYPE_CODE_STRUCT); 7894 INIT_NONE_SPECIFIC (rtype); 7895 rtype->set_num_fields (nfields); 7896 rtype->set_fields 7897 ((struct field *) TYPE_ZALLOC (rtype, nfields * sizeof (struct field))); 7898 rtype->set_name (ada_type_name (type)); 7899 rtype->set_is_fixed_instance (true); 7900 7901 off = 0; 7902 bit_len = 0; 7903 variant_field = -1; 7904 7905 for (f = 0; f < nfields; f += 1) 7906 { 7907 off = align_up (off, field_alignment (type, f)) 7908 + type->field (f).loc_bitpos (); 7909 rtype->field (f).set_loc_bitpos (off); 7910 TYPE_FIELD_BITSIZE (rtype, f) = 0; 7911 7912 if (ada_is_variant_part (type, f)) 7913 { 7914 variant_field = f; 7915 fld_bit_len = 0; 7916 } 7917 else if (is_dynamic_field (type, f)) 7918 { 7919 const gdb_byte *field_valaddr = valaddr; 7920 CORE_ADDR field_address = address; 7921 struct type *field_type = type->field (f).type ()->target_type (); 7922 7923 if (dval0 == NULL) 7924 { 7925 /* Using plain value_from_contents_and_address here 7926 causes problems because we will end up trying to 7927 resolve a type that is currently being 7928 constructed. */ 7929 dval = value_from_contents_and_address_unresolved (rtype, 7930 valaddr, 7931 address); 7932 rtype = value_type (dval); 7933 } 7934 else 7935 dval = dval0; 7936 7937 /* If the type referenced by this field is an aligner type, we need 7938 to unwrap that aligner type, because its size might not be set. 7939 Keeping the aligner type would cause us to compute the wrong 7940 size for this field, impacting the offset of the all the fields 7941 that follow this one. */ 7942 if (ada_is_aligner_type (field_type)) 7943 { 7944 long field_offset = type->field (f).loc_bitpos (); 7945 7946 field_valaddr = cond_offset_host (field_valaddr, field_offset); 7947 field_address = cond_offset_target (field_address, field_offset); 7948 field_type = ada_aligned_type (field_type); 7949 } 7950 7951 field_valaddr = cond_offset_host (field_valaddr, 7952 off / TARGET_CHAR_BIT); 7953 field_address = cond_offset_target (field_address, 7954 off / TARGET_CHAR_BIT); 7955 7956 /* Get the fixed type of the field. Note that, in this case, 7957 we do not want to get the real type out of the tag: if 7958 the current field is the parent part of a tagged record, 7959 we will get the tag of the object. Clearly wrong: the real 7960 type of the parent is not the real type of the child. We 7961 would end up in an infinite loop. */ 7962 field_type = ada_get_base_type (field_type); 7963 field_type = ada_to_fixed_type (field_type, field_valaddr, 7964 field_address, dval, 0); 7965 7966 rtype->field (f).set_type (field_type); 7967 rtype->field (f).set_name (type->field (f).name ()); 7968 /* The multiplication can potentially overflow. But because 7969 the field length has been size-checked just above, and 7970 assuming that the maximum size is a reasonable value, 7971 an overflow should not happen in practice. So rather than 7972 adding overflow recovery code to this already complex code, 7973 we just assume that it's not going to happen. */ 7974 fld_bit_len = rtype->field (f).type ()->length () * TARGET_CHAR_BIT; 7975 } 7976 else 7977 { 7978 /* Note: If this field's type is a typedef, it is important 7979 to preserve the typedef layer. 7980 7981 Otherwise, we might be transforming a typedef to a fat 7982 pointer (encoding a pointer to an unconstrained array), 7983 into a basic fat pointer (encoding an unconstrained 7984 array). As both types are implemented using the same 7985 structure, the typedef is the only clue which allows us 7986 to distinguish between the two options. Stripping it 7987 would prevent us from printing this field appropriately. */ 7988 rtype->field (f).set_type (type->field (f).type ()); 7989 rtype->field (f).set_name (type->field (f).name ()); 7990 if (TYPE_FIELD_BITSIZE (type, f) > 0) 7991 fld_bit_len = 7992 TYPE_FIELD_BITSIZE (rtype, f) = TYPE_FIELD_BITSIZE (type, f); 7993 else 7994 { 7995 struct type *field_type = type->field (f).type (); 7996 7997 /* We need to be careful of typedefs when computing 7998 the length of our field. If this is a typedef, 7999 get the length of the target type, not the length 8000 of the typedef. */ 8001 if (field_type->code () == TYPE_CODE_TYPEDEF) 8002 field_type = ada_typedef_target_type (field_type); 8003 8004 fld_bit_len = 8005 ada_check_typedef (field_type)->length () * TARGET_CHAR_BIT; 8006 } 8007 } 8008 if (off + fld_bit_len > bit_len) 8009 bit_len = off + fld_bit_len; 8010 off += fld_bit_len; 8011 rtype->set_length (align_up (bit_len, TARGET_CHAR_BIT) / TARGET_CHAR_BIT); 8012 } 8013 8014 /* We handle the variant part, if any, at the end because of certain 8015 odd cases in which it is re-ordered so as NOT to be the last field of 8016 the record. This can happen in the presence of representation 8017 clauses. */ 8018 if (variant_field >= 0) 8019 { 8020 struct type *branch_type; 8021 8022 off = rtype->field (variant_field).loc_bitpos (); 8023 8024 if (dval0 == NULL) 8025 { 8026 /* Using plain value_from_contents_and_address here causes 8027 problems because we will end up trying to resolve a type 8028 that is currently being constructed. */ 8029 dval = value_from_contents_and_address_unresolved (rtype, valaddr, 8030 address); 8031 rtype = value_type (dval); 8032 } 8033 else 8034 dval = dval0; 8035 8036 branch_type = 8037 to_fixed_variant_branch_type 8038 (type->field (variant_field).type (), 8039 cond_offset_host (valaddr, off / TARGET_CHAR_BIT), 8040 cond_offset_target (address, off / TARGET_CHAR_BIT), dval); 8041 if (branch_type == NULL) 8042 { 8043 for (f = variant_field + 1; f < rtype->num_fields (); f += 1) 8044 rtype->field (f - 1) = rtype->field (f); 8045 rtype->set_num_fields (rtype->num_fields () - 1); 8046 } 8047 else 8048 { 8049 rtype->field (variant_field).set_type (branch_type); 8050 rtype->field (variant_field).set_name ("S"); 8051 fld_bit_len = 8052 rtype->field (variant_field).type ()->length () * TARGET_CHAR_BIT; 8053 if (off + fld_bit_len > bit_len) 8054 bit_len = off + fld_bit_len; 8055 8056 rtype->set_length 8057 (align_up (bit_len, TARGET_CHAR_BIT) / TARGET_CHAR_BIT); 8058 } 8059 } 8060 8061 /* According to exp_dbug.ads, the size of TYPE for variable-size records 8062 should contain the alignment of that record, which should be a strictly 8063 positive value. If null or negative, then something is wrong, most 8064 probably in the debug info. In that case, we don't round up the size 8065 of the resulting type. If this record is not part of another structure, 8066 the current RTYPE length might be good enough for our purposes. */ 8067 if (type->length () <= 0) 8068 { 8069 if (rtype->name ()) 8070 warning (_("Invalid type size for `%s' detected: %s."), 8071 rtype->name (), pulongest (type->length ())); 8072 else 8073 warning (_("Invalid type size for <unnamed> detected: %s."), 8074 pulongest (type->length ())); 8075 } 8076 else 8077 rtype->set_length (align_up (rtype->length (), type->length ())); 8078 8079 return rtype; 8080 } 8081 8082 /* As for ada_template_to_fixed_record_type_1 with KEEP_DYNAMIC_FIELDS 8083 of 1. */ 8084 8085 static struct type * 8086 template_to_fixed_record_type (struct type *type, const gdb_byte *valaddr, 8087 CORE_ADDR address, struct value *dval0) 8088 { 8089 return ada_template_to_fixed_record_type_1 (type, valaddr, 8090 address, dval0, 1); 8091 } 8092 8093 /* An ordinary record type in which ___XVL-convention fields and 8094 ___XVU- and ___XVN-convention field types in TYPE0 are replaced with 8095 static approximations, containing all possible fields. Uses 8096 no runtime values. Useless for use in values, but that's OK, 8097 since the results are used only for type determinations. Works on both 8098 structs and unions. Representation note: to save space, we memorize 8099 the result of this function in the type::target_type of the 8100 template type. */ 8101 8102 static struct type * 8103 template_to_static_fixed_type (struct type *type0) 8104 { 8105 struct type *type; 8106 int nfields; 8107 int f; 8108 8109 /* No need no do anything if the input type is already fixed. */ 8110 if (type0->is_fixed_instance ()) 8111 return type0; 8112 8113 /* Likewise if we already have computed the static approximation. */ 8114 if (type0->target_type () != NULL) 8115 return type0->target_type (); 8116 8117 /* Don't clone TYPE0 until we are sure we are going to need a copy. */ 8118 type = type0; 8119 nfields = type0->num_fields (); 8120 8121 /* Whether or not we cloned TYPE0, cache the result so that we don't do 8122 recompute all over next time. */ 8123 type0->set_target_type (type); 8124 8125 for (f = 0; f < nfields; f += 1) 8126 { 8127 struct type *field_type = type0->field (f).type (); 8128 struct type *new_type; 8129 8130 if (is_dynamic_field (type0, f)) 8131 { 8132 field_type = ada_check_typedef (field_type); 8133 new_type = to_static_fixed_type (field_type->target_type ()); 8134 } 8135 else 8136 new_type = static_unwrap_type (field_type); 8137 8138 if (new_type != field_type) 8139 { 8140 /* Clone TYPE0 only the first time we get a new field type. */ 8141 if (type == type0) 8142 { 8143 type = alloc_type_copy (type0); 8144 type0->set_target_type (type); 8145 type->set_code (type0->code ()); 8146 INIT_NONE_SPECIFIC (type); 8147 type->set_num_fields (nfields); 8148 8149 field *fields = 8150 ((struct field *) 8151 TYPE_ALLOC (type, nfields * sizeof (struct field))); 8152 memcpy (fields, type0->fields (), 8153 sizeof (struct field) * nfields); 8154 type->set_fields (fields); 8155 8156 type->set_name (ada_type_name (type0)); 8157 type->set_is_fixed_instance (true); 8158 type->set_length (0); 8159 } 8160 type->field (f).set_type (new_type); 8161 type->field (f).set_name (type0->field (f).name ()); 8162 } 8163 } 8164 8165 return type; 8166 } 8167 8168 /* Given an object of type TYPE whose contents are at VALADDR and 8169 whose address in memory is ADDRESS, returns a revision of TYPE, 8170 which should be a non-dynamic-sized record, in which the variant 8171 part, if any, is replaced with the appropriate branch. Looks 8172 for discriminant values in DVAL0, which can be NULL if the record 8173 contains the necessary discriminant values. */ 8174 8175 static struct type * 8176 to_record_with_fixed_variant_part (struct type *type, const gdb_byte *valaddr, 8177 CORE_ADDR address, struct value *dval0) 8178 { 8179 struct value *dval; 8180 struct type *rtype; 8181 struct type *branch_type; 8182 int nfields = type->num_fields (); 8183 int variant_field = variant_field_index (type); 8184 8185 if (variant_field == -1) 8186 return type; 8187 8188 scoped_value_mark mark; 8189 if (dval0 == NULL) 8190 { 8191 dval = value_from_contents_and_address (type, valaddr, address); 8192 type = value_type (dval); 8193 } 8194 else 8195 dval = dval0; 8196 8197 rtype = alloc_type_copy (type); 8198 rtype->set_code (TYPE_CODE_STRUCT); 8199 INIT_NONE_SPECIFIC (rtype); 8200 rtype->set_num_fields (nfields); 8201 8202 field *fields = 8203 (struct field *) TYPE_ALLOC (rtype, nfields * sizeof (struct field)); 8204 memcpy (fields, type->fields (), sizeof (struct field) * nfields); 8205 rtype->set_fields (fields); 8206 8207 rtype->set_name (ada_type_name (type)); 8208 rtype->set_is_fixed_instance (true); 8209 rtype->set_length (type->length ()); 8210 8211 branch_type = to_fixed_variant_branch_type 8212 (type->field (variant_field).type (), 8213 cond_offset_host (valaddr, 8214 type->field (variant_field).loc_bitpos () 8215 / TARGET_CHAR_BIT), 8216 cond_offset_target (address, 8217 type->field (variant_field).loc_bitpos () 8218 / TARGET_CHAR_BIT), dval); 8219 if (branch_type == NULL) 8220 { 8221 int f; 8222 8223 for (f = variant_field + 1; f < nfields; f += 1) 8224 rtype->field (f - 1) = rtype->field (f); 8225 rtype->set_num_fields (rtype->num_fields () - 1); 8226 } 8227 else 8228 { 8229 rtype->field (variant_field).set_type (branch_type); 8230 rtype->field (variant_field).set_name ("S"); 8231 TYPE_FIELD_BITSIZE (rtype, variant_field) = 0; 8232 rtype->set_length (rtype->length () + branch_type->length ()); 8233 } 8234 8235 rtype->set_length (rtype->length () 8236 - type->field (variant_field).type ()->length ()); 8237 8238 return rtype; 8239 } 8240 8241 /* An ordinary record type (with fixed-length fields) that describes 8242 the value at (TYPE0, VALADDR, ADDRESS) [see explanation at 8243 beginning of this section]. Any necessary discriminants' values 8244 should be in DVAL, a record value; it may be NULL if the object 8245 at ADDR itself contains any necessary discriminant values. 8246 Additionally, VALADDR and ADDRESS may also be NULL if no discriminant 8247 values from the record are needed. Except in the case that DVAL, 8248 VALADDR, and ADDRESS are all 0 or NULL, a variant field (unless 8249 unchecked) is replaced by a particular branch of the variant. 8250 8251 NOTE: the case in which DVAL and VALADDR are NULL and ADDRESS is 0 8252 is questionable and may be removed. It can arise during the 8253 processing of an unconstrained-array-of-record type where all the 8254 variant branches have exactly the same size. This is because in 8255 such cases, the compiler does not bother to use the XVS convention 8256 when encoding the record. I am currently dubious of this 8257 shortcut and suspect the compiler should be altered. FIXME. */ 8258 8259 static struct type * 8260 to_fixed_record_type (struct type *type0, const gdb_byte *valaddr, 8261 CORE_ADDR address, struct value *dval) 8262 { 8263 struct type *templ_type; 8264 8265 if (type0->is_fixed_instance ()) 8266 return type0; 8267 8268 templ_type = dynamic_template_type (type0); 8269 8270 if (templ_type != NULL) 8271 return template_to_fixed_record_type (templ_type, valaddr, address, dval); 8272 else if (variant_field_index (type0) >= 0) 8273 { 8274 if (dval == NULL && valaddr == NULL && address == 0) 8275 return type0; 8276 return to_record_with_fixed_variant_part (type0, valaddr, address, 8277 dval); 8278 } 8279 else 8280 { 8281 type0->set_is_fixed_instance (true); 8282 return type0; 8283 } 8284 8285 } 8286 8287 /* An ordinary record type (with fixed-length fields) that describes 8288 the value at (VAR_TYPE0, VALADDR, ADDRESS), where VAR_TYPE0 is a 8289 union type. Any necessary discriminants' values should be in DVAL, 8290 a record value. That is, this routine selects the appropriate 8291 branch of the union at ADDR according to the discriminant value 8292 indicated in the union's type name. Returns VAR_TYPE0 itself if 8293 it represents a variant subject to a pragma Unchecked_Union. */ 8294 8295 static struct type * 8296 to_fixed_variant_branch_type (struct type *var_type0, const gdb_byte *valaddr, 8297 CORE_ADDR address, struct value *dval) 8298 { 8299 int which; 8300 struct type *templ_type; 8301 struct type *var_type; 8302 8303 if (var_type0->code () == TYPE_CODE_PTR) 8304 var_type = var_type0->target_type (); 8305 else 8306 var_type = var_type0; 8307 8308 templ_type = ada_find_parallel_type (var_type, "___XVU"); 8309 8310 if (templ_type != NULL) 8311 var_type = templ_type; 8312 8313 if (is_unchecked_variant (var_type, value_type (dval))) 8314 return var_type0; 8315 which = ada_which_variant_applies (var_type, dval); 8316 8317 if (which < 0) 8318 return empty_record (var_type); 8319 else if (is_dynamic_field (var_type, which)) 8320 return to_fixed_record_type 8321 (var_type->field (which).type ()->target_type(), valaddr, address, dval); 8322 else if (variant_field_index (var_type->field (which).type ()) >= 0) 8323 return 8324 to_fixed_record_type 8325 (var_type->field (which).type (), valaddr, address, dval); 8326 else 8327 return var_type->field (which).type (); 8328 } 8329 8330 /* Assuming RANGE_TYPE is a TYPE_CODE_RANGE, return nonzero if 8331 ENCODING_TYPE, a type following the GNAT conventions for discrete 8332 type encodings, only carries redundant information. */ 8333 8334 static int 8335 ada_is_redundant_range_encoding (struct type *range_type, 8336 struct type *encoding_type) 8337 { 8338 const char *bounds_str; 8339 int n; 8340 LONGEST lo, hi; 8341 8342 gdb_assert (range_type->code () == TYPE_CODE_RANGE); 8343 8344 if (get_base_type (range_type)->code () 8345 != get_base_type (encoding_type)->code ()) 8346 { 8347 /* The compiler probably used a simple base type to describe 8348 the range type instead of the range's actual base type, 8349 expecting us to get the real base type from the encoding 8350 anyway. In this situation, the encoding cannot be ignored 8351 as redundant. */ 8352 return 0; 8353 } 8354 8355 if (is_dynamic_type (range_type)) 8356 return 0; 8357 8358 if (encoding_type->name () == NULL) 8359 return 0; 8360 8361 bounds_str = strstr (encoding_type->name (), "___XDLU_"); 8362 if (bounds_str == NULL) 8363 return 0; 8364 8365 n = 8; /* Skip "___XDLU_". */ 8366 if (!ada_scan_number (bounds_str, n, &lo, &n)) 8367 return 0; 8368 if (range_type->bounds ()->low.const_val () != lo) 8369 return 0; 8370 8371 n += 2; /* Skip the "__" separator between the two bounds. */ 8372 if (!ada_scan_number (bounds_str, n, &hi, &n)) 8373 return 0; 8374 if (range_type->bounds ()->high.const_val () != hi) 8375 return 0; 8376 8377 return 1; 8378 } 8379 8380 /* Given the array type ARRAY_TYPE, return nonzero if DESC_TYPE, 8381 a type following the GNAT encoding for describing array type 8382 indices, only carries redundant information. */ 8383 8384 static int 8385 ada_is_redundant_index_type_desc (struct type *array_type, 8386 struct type *desc_type) 8387 { 8388 struct type *this_layer = check_typedef (array_type); 8389 int i; 8390 8391 for (i = 0; i < desc_type->num_fields (); i++) 8392 { 8393 if (!ada_is_redundant_range_encoding (this_layer->index_type (), 8394 desc_type->field (i).type ())) 8395 return 0; 8396 this_layer = check_typedef (this_layer->target_type ()); 8397 } 8398 8399 return 1; 8400 } 8401 8402 /* Assuming that TYPE0 is an array type describing the type of a value 8403 at ADDR, and that DVAL describes a record containing any 8404 discriminants used in TYPE0, returns a type for the value that 8405 contains no dynamic components (that is, no components whose sizes 8406 are determined by run-time quantities). Unless IGNORE_TOO_BIG is 8407 true, gives an error message if the resulting type's size is over 8408 varsize_limit. */ 8409 8410 static struct type * 8411 to_fixed_array_type (struct type *type0, struct value *dval, 8412 int ignore_too_big) 8413 { 8414 struct type *index_type_desc; 8415 struct type *result; 8416 int constrained_packed_array_p; 8417 static const char *xa_suffix = "___XA"; 8418 8419 type0 = ada_check_typedef (type0); 8420 if (type0->is_fixed_instance ()) 8421 return type0; 8422 8423 constrained_packed_array_p = ada_is_constrained_packed_array_type (type0); 8424 if (constrained_packed_array_p) 8425 { 8426 type0 = decode_constrained_packed_array_type (type0); 8427 if (type0 == nullptr) 8428 error (_("could not decode constrained packed array type")); 8429 } 8430 8431 index_type_desc = ada_find_parallel_type (type0, xa_suffix); 8432 8433 /* As mentioned in exp_dbug.ads, for non bit-packed arrays an 8434 encoding suffixed with 'P' may still be generated. If so, 8435 it should be used to find the XA type. */ 8436 8437 if (index_type_desc == NULL) 8438 { 8439 const char *type_name = ada_type_name (type0); 8440 8441 if (type_name != NULL) 8442 { 8443 const int len = strlen (type_name); 8444 char *name = (char *) alloca (len + strlen (xa_suffix)); 8445 8446 if (type_name[len - 1] == 'P') 8447 { 8448 strcpy (name, type_name); 8449 strcpy (name + len - 1, xa_suffix); 8450 index_type_desc = ada_find_parallel_type_with_name (type0, name); 8451 } 8452 } 8453 } 8454 8455 ada_fixup_array_indexes_type (index_type_desc); 8456 if (index_type_desc != NULL 8457 && ada_is_redundant_index_type_desc (type0, index_type_desc)) 8458 { 8459 /* Ignore this ___XA parallel type, as it does not bring any 8460 useful information. This allows us to avoid creating fixed 8461 versions of the array's index types, which would be identical 8462 to the original ones. This, in turn, can also help avoid 8463 the creation of fixed versions of the array itself. */ 8464 index_type_desc = NULL; 8465 } 8466 8467 if (index_type_desc == NULL) 8468 { 8469 struct type *elt_type0 = ada_check_typedef (type0->target_type ()); 8470 8471 /* NOTE: elt_type---the fixed version of elt_type0---should never 8472 depend on the contents of the array in properly constructed 8473 debugging data. */ 8474 /* Create a fixed version of the array element type. 8475 We're not providing the address of an element here, 8476 and thus the actual object value cannot be inspected to do 8477 the conversion. This should not be a problem, since arrays of 8478 unconstrained objects are not allowed. In particular, all 8479 the elements of an array of a tagged type should all be of 8480 the same type specified in the debugging info. No need to 8481 consult the object tag. */ 8482 struct type *elt_type = ada_to_fixed_type (elt_type0, 0, 0, dval, 1); 8483 8484 /* Make sure we always create a new array type when dealing with 8485 packed array types, since we're going to fix-up the array 8486 type length and element bitsize a little further down. */ 8487 if (elt_type0 == elt_type && !constrained_packed_array_p) 8488 result = type0; 8489 else 8490 result = create_array_type (alloc_type_copy (type0), 8491 elt_type, type0->index_type ()); 8492 } 8493 else 8494 { 8495 int i; 8496 struct type *elt_type0; 8497 8498 elt_type0 = type0; 8499 for (i = index_type_desc->num_fields (); i > 0; i -= 1) 8500 elt_type0 = elt_type0->target_type (); 8501 8502 /* NOTE: result---the fixed version of elt_type0---should never 8503 depend on the contents of the array in properly constructed 8504 debugging data. */ 8505 /* Create a fixed version of the array element type. 8506 We're not providing the address of an element here, 8507 and thus the actual object value cannot be inspected to do 8508 the conversion. This should not be a problem, since arrays of 8509 unconstrained objects are not allowed. In particular, all 8510 the elements of an array of a tagged type should all be of 8511 the same type specified in the debugging info. No need to 8512 consult the object tag. */ 8513 result = 8514 ada_to_fixed_type (ada_check_typedef (elt_type0), 0, 0, dval, 1); 8515 8516 elt_type0 = type0; 8517 for (i = index_type_desc->num_fields () - 1; i >= 0; i -= 1) 8518 { 8519 struct type *range_type = 8520 to_fixed_range_type (index_type_desc->field (i).type (), dval); 8521 8522 result = create_array_type (alloc_type_copy (elt_type0), 8523 result, range_type); 8524 elt_type0 = elt_type0->target_type (); 8525 } 8526 } 8527 8528 /* We want to preserve the type name. This can be useful when 8529 trying to get the type name of a value that has already been 8530 printed (for instance, if the user did "print VAR; whatis $". */ 8531 result->set_name (type0->name ()); 8532 8533 if (constrained_packed_array_p) 8534 { 8535 /* So far, the resulting type has been created as if the original 8536 type was a regular (non-packed) array type. As a result, the 8537 bitsize of the array elements needs to be set again, and the array 8538 length needs to be recomputed based on that bitsize. */ 8539 int len = result->length () / result->target_type ()->length (); 8540 int elt_bitsize = TYPE_FIELD_BITSIZE (type0, 0); 8541 8542 TYPE_FIELD_BITSIZE (result, 0) = TYPE_FIELD_BITSIZE (type0, 0); 8543 result->set_length (len * elt_bitsize / HOST_CHAR_BIT); 8544 if (result->length () * HOST_CHAR_BIT < len * elt_bitsize) 8545 result->set_length (result->length () + 1); 8546 } 8547 8548 result->set_is_fixed_instance (true); 8549 return result; 8550 } 8551 8552 8553 /* A standard type (containing no dynamically sized components) 8554 corresponding to TYPE for the value (TYPE, VALADDR, ADDRESS) 8555 DVAL describes a record containing any discriminants used in TYPE0, 8556 and may be NULL if there are none, or if the object of type TYPE at 8557 ADDRESS or in VALADDR contains these discriminants. 8558 8559 If CHECK_TAG is not null, in the case of tagged types, this function 8560 attempts to locate the object's tag and use it to compute the actual 8561 type. However, when ADDRESS is null, we cannot use it to determine the 8562 location of the tag, and therefore compute the tagged type's actual type. 8563 So we return the tagged type without consulting the tag. */ 8564 8565 static struct type * 8566 ada_to_fixed_type_1 (struct type *type, const gdb_byte *valaddr, 8567 CORE_ADDR address, struct value *dval, int check_tag) 8568 { 8569 type = ada_check_typedef (type); 8570 8571 /* Only un-fixed types need to be handled here. */ 8572 if (!HAVE_GNAT_AUX_INFO (type)) 8573 return type; 8574 8575 switch (type->code ()) 8576 { 8577 default: 8578 return type; 8579 case TYPE_CODE_STRUCT: 8580 { 8581 struct type *static_type = to_static_fixed_type (type); 8582 struct type *fixed_record_type = 8583 to_fixed_record_type (type, valaddr, address, NULL); 8584 8585 /* If STATIC_TYPE is a tagged type and we know the object's address, 8586 then we can determine its tag, and compute the object's actual 8587 type from there. Note that we have to use the fixed record 8588 type (the parent part of the record may have dynamic fields 8589 and the way the location of _tag is expressed may depend on 8590 them). */ 8591 8592 if (check_tag && address != 0 && ada_is_tagged_type (static_type, 0)) 8593 { 8594 struct value *tag = 8595 value_tag_from_contents_and_address 8596 (fixed_record_type, 8597 valaddr, 8598 address); 8599 struct type *real_type = type_from_tag (tag); 8600 struct value *obj = 8601 value_from_contents_and_address (fixed_record_type, 8602 valaddr, 8603 address); 8604 fixed_record_type = value_type (obj); 8605 if (real_type != NULL) 8606 return to_fixed_record_type 8607 (real_type, NULL, 8608 value_address (ada_tag_value_at_base_address (obj)), NULL); 8609 } 8610 8611 /* Check to see if there is a parallel ___XVZ variable. 8612 If there is, then it provides the actual size of our type. */ 8613 else if (ada_type_name (fixed_record_type) != NULL) 8614 { 8615 const char *name = ada_type_name (fixed_record_type); 8616 char *xvz_name 8617 = (char *) alloca (strlen (name) + 7 /* "___XVZ\0" */); 8618 bool xvz_found = false; 8619 LONGEST size; 8620 8621 xsnprintf (xvz_name, strlen (name) + 7, "%s___XVZ", name); 8622 try 8623 { 8624 xvz_found = get_int_var_value (xvz_name, size); 8625 } 8626 catch (const gdb_exception_error &except) 8627 { 8628 /* We found the variable, but somehow failed to read 8629 its value. Rethrow the same error, but with a little 8630 bit more information, to help the user understand 8631 what went wrong (Eg: the variable might have been 8632 optimized out). */ 8633 throw_error (except.error, 8634 _("unable to read value of %s (%s)"), 8635 xvz_name, except.what ()); 8636 } 8637 8638 if (xvz_found && fixed_record_type->length () != size) 8639 { 8640 fixed_record_type = copy_type (fixed_record_type); 8641 fixed_record_type->set_length (size); 8642 8643 /* The FIXED_RECORD_TYPE may have be a stub. We have 8644 observed this when the debugging info is STABS, and 8645 apparently it is something that is hard to fix. 8646 8647 In practice, we don't need the actual type definition 8648 at all, because the presence of the XVZ variable allows us 8649 to assume that there must be a XVS type as well, which we 8650 should be able to use later, when we need the actual type 8651 definition. 8652 8653 In the meantime, pretend that the "fixed" type we are 8654 returning is NOT a stub, because this can cause trouble 8655 when using this type to create new types targeting it. 8656 Indeed, the associated creation routines often check 8657 whether the target type is a stub and will try to replace 8658 it, thus using a type with the wrong size. This, in turn, 8659 might cause the new type to have the wrong size too. 8660 Consider the case of an array, for instance, where the size 8661 of the array is computed from the number of elements in 8662 our array multiplied by the size of its element. */ 8663 fixed_record_type->set_is_stub (false); 8664 } 8665 } 8666 return fixed_record_type; 8667 } 8668 case TYPE_CODE_ARRAY: 8669 return to_fixed_array_type (type, dval, 1); 8670 case TYPE_CODE_UNION: 8671 if (dval == NULL) 8672 return type; 8673 else 8674 return to_fixed_variant_branch_type (type, valaddr, address, dval); 8675 } 8676 } 8677 8678 /* The same as ada_to_fixed_type_1, except that it preserves the type 8679 if it is a TYPE_CODE_TYPEDEF of a type that is already fixed. 8680 8681 The typedef layer needs be preserved in order to differentiate between 8682 arrays and array pointers when both types are implemented using the same 8683 fat pointer. In the array pointer case, the pointer is encoded as 8684 a typedef of the pointer type. For instance, considering: 8685 8686 type String_Access is access String; 8687 S1 : String_Access := null; 8688 8689 To the debugger, S1 is defined as a typedef of type String. But 8690 to the user, it is a pointer. So if the user tries to print S1, 8691 we should not dereference the array, but print the array address 8692 instead. 8693 8694 If we didn't preserve the typedef layer, we would lose the fact that 8695 the type is to be presented as a pointer (needs de-reference before 8696 being printed). And we would also use the source-level type name. */ 8697 8698 struct type * 8699 ada_to_fixed_type (struct type *type, const gdb_byte *valaddr, 8700 CORE_ADDR address, struct value *dval, int check_tag) 8701 8702 { 8703 struct type *fixed_type = 8704 ada_to_fixed_type_1 (type, valaddr, address, dval, check_tag); 8705 8706 /* If TYPE is a typedef and its target type is the same as the FIXED_TYPE, 8707 then preserve the typedef layer. 8708 8709 Implementation note: We can only check the main-type portion of 8710 the TYPE and FIXED_TYPE, because eliminating the typedef layer 8711 from TYPE now returns a type that has the same instance flags 8712 as TYPE. For instance, if TYPE is a "typedef const", and its 8713 target type is a "struct", then the typedef elimination will return 8714 a "const" version of the target type. See check_typedef for more 8715 details about how the typedef layer elimination is done. 8716 8717 brobecker/2010-11-19: It seems to me that the only case where it is 8718 useful to preserve the typedef layer is when dealing with fat pointers. 8719 Perhaps, we could add a check for that and preserve the typedef layer 8720 only in that situation. But this seems unnecessary so far, probably 8721 because we call check_typedef/ada_check_typedef pretty much everywhere. 8722 */ 8723 if (type->code () == TYPE_CODE_TYPEDEF 8724 && (TYPE_MAIN_TYPE (ada_typedef_target_type (type)) 8725 == TYPE_MAIN_TYPE (fixed_type))) 8726 return type; 8727 8728 return fixed_type; 8729 } 8730 8731 /* A standard (static-sized) type corresponding as well as possible to 8732 TYPE0, but based on no runtime data. */ 8733 8734 static struct type * 8735 to_static_fixed_type (struct type *type0) 8736 { 8737 struct type *type; 8738 8739 if (type0 == NULL) 8740 return NULL; 8741 8742 if (type0->is_fixed_instance ()) 8743 return type0; 8744 8745 type0 = ada_check_typedef (type0); 8746 8747 switch (type0->code ()) 8748 { 8749 default: 8750 return type0; 8751 case TYPE_CODE_STRUCT: 8752 type = dynamic_template_type (type0); 8753 if (type != NULL) 8754 return template_to_static_fixed_type (type); 8755 else 8756 return template_to_static_fixed_type (type0); 8757 case TYPE_CODE_UNION: 8758 type = ada_find_parallel_type (type0, "___XVU"); 8759 if (type != NULL) 8760 return template_to_static_fixed_type (type); 8761 else 8762 return template_to_static_fixed_type (type0); 8763 } 8764 } 8765 8766 /* A static approximation of TYPE with all type wrappers removed. */ 8767 8768 static struct type * 8769 static_unwrap_type (struct type *type) 8770 { 8771 if (ada_is_aligner_type (type)) 8772 { 8773 struct type *type1 = ada_check_typedef (type)->field (0).type (); 8774 if (ada_type_name (type1) == NULL) 8775 type1->set_name (ada_type_name (type)); 8776 8777 return static_unwrap_type (type1); 8778 } 8779 else 8780 { 8781 struct type *raw_real_type = ada_get_base_type (type); 8782 8783 if (raw_real_type == type) 8784 return type; 8785 else 8786 return to_static_fixed_type (raw_real_type); 8787 } 8788 } 8789 8790 /* In some cases, incomplete and private types require 8791 cross-references that are not resolved as records (for example, 8792 type Foo; 8793 type FooP is access Foo; 8794 V: FooP; 8795 type Foo is array ...; 8796 ). In these cases, since there is no mechanism for producing 8797 cross-references to such types, we instead substitute for FooP a 8798 stub enumeration type that is nowhere resolved, and whose tag is 8799 the name of the actual type. Call these types "non-record stubs". */ 8800 8801 /* A type equivalent to TYPE that is not a non-record stub, if one 8802 exists, otherwise TYPE. */ 8803 8804 struct type * 8805 ada_check_typedef (struct type *type) 8806 { 8807 if (type == NULL) 8808 return NULL; 8809 8810 /* If our type is an access to an unconstrained array, which is encoded 8811 as a TYPE_CODE_TYPEDEF of a fat pointer, then we're done. 8812 We don't want to strip the TYPE_CODE_TYPDEF layer, because this is 8813 what allows us to distinguish between fat pointers that represent 8814 array types, and fat pointers that represent array access types 8815 (in both cases, the compiler implements them as fat pointers). */ 8816 if (ada_is_access_to_unconstrained_array (type)) 8817 return type; 8818 8819 type = check_typedef (type); 8820 if (type == NULL || type->code () != TYPE_CODE_ENUM 8821 || !type->is_stub () 8822 || type->name () == NULL) 8823 return type; 8824 else 8825 { 8826 const char *name = type->name (); 8827 struct type *type1 = ada_find_any_type (name); 8828 8829 if (type1 == NULL) 8830 return type; 8831 8832 /* TYPE1 might itself be a TYPE_CODE_TYPEDEF (this can happen with 8833 stubs pointing to arrays, as we don't create symbols for array 8834 types, only for the typedef-to-array types). If that's the case, 8835 strip the typedef layer. */ 8836 if (type1->code () == TYPE_CODE_TYPEDEF) 8837 type1 = ada_check_typedef (type1); 8838 8839 return type1; 8840 } 8841 } 8842 8843 /* A value representing the data at VALADDR/ADDRESS as described by 8844 type TYPE0, but with a standard (static-sized) type that correctly 8845 describes it. If VAL0 is not NULL and TYPE0 already is a standard 8846 type, then return VAL0 [this feature is simply to avoid redundant 8847 creation of struct values]. */ 8848 8849 static struct value * 8850 ada_to_fixed_value_create (struct type *type0, CORE_ADDR address, 8851 struct value *val0) 8852 { 8853 struct type *type = ada_to_fixed_type (type0, 0, address, NULL, 1); 8854 8855 if (type == type0 && val0 != NULL) 8856 return val0; 8857 8858 if (VALUE_LVAL (val0) != lval_memory) 8859 { 8860 /* Our value does not live in memory; it could be a convenience 8861 variable, for instance. Create a not_lval value using val0's 8862 contents. */ 8863 return value_from_contents (type, value_contents (val0).data ()); 8864 } 8865 8866 return value_from_contents_and_address (type, 0, address); 8867 } 8868 8869 /* A value representing VAL, but with a standard (static-sized) type 8870 that correctly describes it. Does not necessarily create a new 8871 value. */ 8872 8873 struct value * 8874 ada_to_fixed_value (struct value *val) 8875 { 8876 val = unwrap_value (val); 8877 val = ada_to_fixed_value_create (value_type (val), value_address (val), val); 8878 return val; 8879 } 8880 8881 8882 /* Attributes */ 8883 8884 /* Table mapping attribute numbers to names. 8885 NOTE: Keep up to date with enum ada_attribute definition in ada-lang.h. */ 8886 8887 static const char * const attribute_names[] = { 8888 "<?>", 8889 8890 "first", 8891 "last", 8892 "length", 8893 "image", 8894 "max", 8895 "min", 8896 "modulus", 8897 "pos", 8898 "size", 8899 "tag", 8900 "val", 8901 0 8902 }; 8903 8904 static const char * 8905 ada_attribute_name (enum exp_opcode n) 8906 { 8907 if (n >= OP_ATR_FIRST && n <= (int) OP_ATR_VAL) 8908 return attribute_names[n - OP_ATR_FIRST + 1]; 8909 else 8910 return attribute_names[0]; 8911 } 8912 8913 /* Evaluate the 'POS attribute applied to ARG. */ 8914 8915 static LONGEST 8916 pos_atr (struct value *arg) 8917 { 8918 struct value *val = coerce_ref (arg); 8919 struct type *type = value_type (val); 8920 8921 if (!discrete_type_p (type)) 8922 error (_("'POS only defined on discrete types")); 8923 8924 gdb::optional<LONGEST> result = discrete_position (type, value_as_long (val)); 8925 if (!result.has_value ()) 8926 error (_("enumeration value is invalid: can't find 'POS")); 8927 8928 return *result; 8929 } 8930 8931 struct value * 8932 ada_pos_atr (struct type *expect_type, 8933 struct expression *exp, 8934 enum noside noside, enum exp_opcode op, 8935 struct value *arg) 8936 { 8937 struct type *type = builtin_type (exp->gdbarch)->builtin_int; 8938 if (noside == EVAL_AVOID_SIDE_EFFECTS) 8939 return value_zero (type, not_lval); 8940 return value_from_longest (type, pos_atr (arg)); 8941 } 8942 8943 /* Evaluate the TYPE'VAL attribute applied to ARG. */ 8944 8945 static struct value * 8946 val_atr (struct type *type, LONGEST val) 8947 { 8948 gdb_assert (discrete_type_p (type)); 8949 if (type->code () == TYPE_CODE_RANGE) 8950 type = type->target_type (); 8951 if (type->code () == TYPE_CODE_ENUM) 8952 { 8953 if (val < 0 || val >= type->num_fields ()) 8954 error (_("argument to 'VAL out of range")); 8955 val = type->field (val).loc_enumval (); 8956 } 8957 return value_from_longest (type, val); 8958 } 8959 8960 struct value * 8961 ada_val_atr (enum noside noside, struct type *type, struct value *arg) 8962 { 8963 if (noside == EVAL_AVOID_SIDE_EFFECTS) 8964 return value_zero (type, not_lval); 8965 8966 if (!discrete_type_p (type)) 8967 error (_("'VAL only defined on discrete types")); 8968 if (!integer_type_p (value_type (arg))) 8969 error (_("'VAL requires integral argument")); 8970 8971 return val_atr (type, value_as_long (arg)); 8972 } 8973 8974 8975 /* Evaluation */ 8976 8977 /* True if TYPE appears to be an Ada character type. 8978 [At the moment, this is true only for Character and Wide_Character; 8979 It is a heuristic test that could stand improvement]. */ 8980 8981 bool 8982 ada_is_character_type (struct type *type) 8983 { 8984 const char *name; 8985 8986 /* If the type code says it's a character, then assume it really is, 8987 and don't check any further. */ 8988 if (type->code () == TYPE_CODE_CHAR) 8989 return true; 8990 8991 /* Otherwise, assume it's a character type iff it is a discrete type 8992 with a known character type name. */ 8993 name = ada_type_name (type); 8994 return (name != NULL 8995 && (type->code () == TYPE_CODE_INT 8996 || type->code () == TYPE_CODE_RANGE) 8997 && (strcmp (name, "character") == 0 8998 || strcmp (name, "wide_character") == 0 8999 || strcmp (name, "wide_wide_character") == 0 9000 || strcmp (name, "unsigned char") == 0)); 9001 } 9002 9003 /* True if TYPE appears to be an Ada string type. */ 9004 9005 bool 9006 ada_is_string_type (struct type *type) 9007 { 9008 type = ada_check_typedef (type); 9009 if (type != NULL 9010 && type->code () != TYPE_CODE_PTR 9011 && (ada_is_simple_array_type (type) 9012 || ada_is_array_descriptor_type (type)) 9013 && ada_array_arity (type) == 1) 9014 { 9015 struct type *elttype = ada_array_element_type (type, 1); 9016 9017 return ada_is_character_type (elttype); 9018 } 9019 else 9020 return false; 9021 } 9022 9023 /* The compiler sometimes provides a parallel XVS type for a given 9024 PAD type. Normally, it is safe to follow the PAD type directly, 9025 but older versions of the compiler have a bug that causes the offset 9026 of its "F" field to be wrong. Following that field in that case 9027 would lead to incorrect results, but this can be worked around 9028 by ignoring the PAD type and using the associated XVS type instead. 9029 9030 Set to True if the debugger should trust the contents of PAD types. 9031 Otherwise, ignore the PAD type if there is a parallel XVS type. */ 9032 static bool trust_pad_over_xvs = true; 9033 9034 /* True if TYPE is a struct type introduced by the compiler to force the 9035 alignment of a value. Such types have a single field with a 9036 distinctive name. */ 9037 9038 int 9039 ada_is_aligner_type (struct type *type) 9040 { 9041 type = ada_check_typedef (type); 9042 9043 if (!trust_pad_over_xvs && ada_find_parallel_type (type, "___XVS") != NULL) 9044 return 0; 9045 9046 return (type->code () == TYPE_CODE_STRUCT 9047 && type->num_fields () == 1 9048 && strcmp (type->field (0).name (), "F") == 0); 9049 } 9050 9051 /* If there is an ___XVS-convention type parallel to SUBTYPE, return 9052 the parallel type. */ 9053 9054 struct type * 9055 ada_get_base_type (struct type *raw_type) 9056 { 9057 struct type *real_type_namer; 9058 struct type *raw_real_type; 9059 9060 if (raw_type == NULL || raw_type->code () != TYPE_CODE_STRUCT) 9061 return raw_type; 9062 9063 if (ada_is_aligner_type (raw_type)) 9064 /* The encoding specifies that we should always use the aligner type. 9065 So, even if this aligner type has an associated XVS type, we should 9066 simply ignore it. 9067 9068 According to the compiler gurus, an XVS type parallel to an aligner 9069 type may exist because of a stabs limitation. In stabs, aligner 9070 types are empty because the field has a variable-sized type, and 9071 thus cannot actually be used as an aligner type. As a result, 9072 we need the associated parallel XVS type to decode the type. 9073 Since the policy in the compiler is to not change the internal 9074 representation based on the debugging info format, we sometimes 9075 end up having a redundant XVS type parallel to the aligner type. */ 9076 return raw_type; 9077 9078 real_type_namer = ada_find_parallel_type (raw_type, "___XVS"); 9079 if (real_type_namer == NULL 9080 || real_type_namer->code () != TYPE_CODE_STRUCT 9081 || real_type_namer->num_fields () != 1) 9082 return raw_type; 9083 9084 if (real_type_namer->field (0).type ()->code () != TYPE_CODE_REF) 9085 { 9086 /* This is an older encoding form where the base type needs to be 9087 looked up by name. We prefer the newer encoding because it is 9088 more efficient. */ 9089 raw_real_type = ada_find_any_type (real_type_namer->field (0).name ()); 9090 if (raw_real_type == NULL) 9091 return raw_type; 9092 else 9093 return raw_real_type; 9094 } 9095 9096 /* The field in our XVS type is a reference to the base type. */ 9097 return real_type_namer->field (0).type ()->target_type (); 9098 } 9099 9100 /* The type of value designated by TYPE, with all aligners removed. */ 9101 9102 struct type * 9103 ada_aligned_type (struct type *type) 9104 { 9105 if (ada_is_aligner_type (type)) 9106 return ada_aligned_type (type->field (0).type ()); 9107 else 9108 return ada_get_base_type (type); 9109 } 9110 9111 9112 /* The address of the aligned value in an object at address VALADDR 9113 having type TYPE. Assumes ada_is_aligner_type (TYPE). */ 9114 9115 const gdb_byte * 9116 ada_aligned_value_addr (struct type *type, const gdb_byte *valaddr) 9117 { 9118 if (ada_is_aligner_type (type)) 9119 return ada_aligned_value_addr 9120 (type->field (0).type (), 9121 valaddr + type->field (0).loc_bitpos () / TARGET_CHAR_BIT); 9122 else 9123 return valaddr; 9124 } 9125 9126 9127 9128 /* The printed representation of an enumeration literal with encoded 9129 name NAME. The value is good to the next call of ada_enum_name. */ 9130 const char * 9131 ada_enum_name (const char *name) 9132 { 9133 static std::string storage; 9134 const char *tmp; 9135 9136 /* First, unqualify the enumeration name: 9137 1. Search for the last '.' character. If we find one, then skip 9138 all the preceding characters, the unqualified name starts 9139 right after that dot. 9140 2. Otherwise, we may be debugging on a target where the compiler 9141 translates dots into "__". Search forward for double underscores, 9142 but stop searching when we hit an overloading suffix, which is 9143 of the form "__" followed by digits. */ 9144 9145 tmp = strrchr (name, '.'); 9146 if (tmp != NULL) 9147 name = tmp + 1; 9148 else 9149 { 9150 while ((tmp = strstr (name, "__")) != NULL) 9151 { 9152 if (isdigit (tmp[2])) 9153 break; 9154 else 9155 name = tmp + 2; 9156 } 9157 } 9158 9159 if (name[0] == 'Q') 9160 { 9161 int v; 9162 9163 if (name[1] == 'U' || name[1] == 'W') 9164 { 9165 int offset = 2; 9166 if (name[1] == 'W' && name[2] == 'W') 9167 { 9168 /* Also handle the QWW case. */ 9169 ++offset; 9170 } 9171 if (sscanf (name + offset, "%x", &v) != 1) 9172 return name; 9173 } 9174 else if (((name[1] >= '0' && name[1] <= '9') 9175 || (name[1] >= 'a' && name[1] <= 'z')) 9176 && name[2] == '\0') 9177 { 9178 storage = string_printf ("'%c'", name[1]); 9179 return storage.c_str (); 9180 } 9181 else 9182 return name; 9183 9184 if (isascii (v) && isprint (v)) 9185 storage = string_printf ("'%c'", v); 9186 else if (name[1] == 'U') 9187 storage = string_printf ("'[\"%02x\"]'", v); 9188 else if (name[2] != 'W') 9189 storage = string_printf ("'[\"%04x\"]'", v); 9190 else 9191 storage = string_printf ("'[\"%06x\"]'", v); 9192 9193 return storage.c_str (); 9194 } 9195 else 9196 { 9197 tmp = strstr (name, "__"); 9198 if (tmp == NULL) 9199 tmp = strstr (name, "$"); 9200 if (tmp != NULL) 9201 { 9202 storage = std::string (name, tmp - name); 9203 return storage.c_str (); 9204 } 9205 9206 return name; 9207 } 9208 } 9209 9210 /* If VAL is wrapped in an aligner or subtype wrapper, return the 9211 value it wraps. */ 9212 9213 static struct value * 9214 unwrap_value (struct value *val) 9215 { 9216 struct type *type = ada_check_typedef (value_type (val)); 9217 9218 if (ada_is_aligner_type (type)) 9219 { 9220 struct value *v = ada_value_struct_elt (val, "F", 0); 9221 struct type *val_type = ada_check_typedef (value_type (v)); 9222 9223 if (ada_type_name (val_type) == NULL) 9224 val_type->set_name (ada_type_name (type)); 9225 9226 return unwrap_value (v); 9227 } 9228 else 9229 { 9230 struct type *raw_real_type = 9231 ada_check_typedef (ada_get_base_type (type)); 9232 9233 /* If there is no parallel XVS or XVE type, then the value is 9234 already unwrapped. Return it without further modification. */ 9235 if ((type == raw_real_type) 9236 && ada_find_parallel_type (type, "___XVE") == NULL) 9237 return val; 9238 9239 return 9240 coerce_unspec_val_to_type 9241 (val, ada_to_fixed_type (raw_real_type, 0, 9242 value_address (val), 9243 NULL, 1)); 9244 } 9245 } 9246 9247 /* Given two array types T1 and T2, return nonzero iff both arrays 9248 contain the same number of elements. */ 9249 9250 static int 9251 ada_same_array_size_p (struct type *t1, struct type *t2) 9252 { 9253 LONGEST lo1, hi1, lo2, hi2; 9254 9255 /* Get the array bounds in order to verify that the size of 9256 the two arrays match. */ 9257 if (!get_array_bounds (t1, &lo1, &hi1) 9258 || !get_array_bounds (t2, &lo2, &hi2)) 9259 error (_("unable to determine array bounds")); 9260 9261 /* To make things easier for size comparison, normalize a bit 9262 the case of empty arrays by making sure that the difference 9263 between upper bound and lower bound is always -1. */ 9264 if (lo1 > hi1) 9265 hi1 = lo1 - 1; 9266 if (lo2 > hi2) 9267 hi2 = lo2 - 1; 9268 9269 return (hi1 - lo1 == hi2 - lo2); 9270 } 9271 9272 /* Assuming that VAL is an array of integrals, and TYPE represents 9273 an array with the same number of elements, but with wider integral 9274 elements, return an array "casted" to TYPE. In practice, this 9275 means that the returned array is built by casting each element 9276 of the original array into TYPE's (wider) element type. */ 9277 9278 static struct value * 9279 ada_promote_array_of_integrals (struct type *type, struct value *val) 9280 { 9281 struct type *elt_type = type->target_type (); 9282 LONGEST lo, hi; 9283 LONGEST i; 9284 9285 /* Verify that both val and type are arrays of scalars, and 9286 that the size of val's elements is smaller than the size 9287 of type's element. */ 9288 gdb_assert (type->code () == TYPE_CODE_ARRAY); 9289 gdb_assert (is_integral_type (type->target_type ())); 9290 gdb_assert (value_type (val)->code () == TYPE_CODE_ARRAY); 9291 gdb_assert (is_integral_type (value_type (val)->target_type ())); 9292 gdb_assert (type->target_type ()->length () 9293 > value_type (val)->target_type ()->length ()); 9294 9295 if (!get_array_bounds (type, &lo, &hi)) 9296 error (_("unable to determine array bounds")); 9297 9298 value *res = allocate_value (type); 9299 gdb::array_view<gdb_byte> res_contents = value_contents_writeable (res); 9300 9301 /* Promote each array element. */ 9302 for (i = 0; i < hi - lo + 1; i++) 9303 { 9304 struct value *elt = value_cast (elt_type, value_subscript (val, lo + i)); 9305 int elt_len = elt_type->length (); 9306 9307 copy (value_contents_all (elt), res_contents.slice (elt_len * i, elt_len)); 9308 } 9309 9310 return res; 9311 } 9312 9313 /* Coerce VAL as necessary for assignment to an lval of type TYPE, and 9314 return the converted value. */ 9315 9316 static struct value * 9317 coerce_for_assign (struct type *type, struct value *val) 9318 { 9319 struct type *type2 = value_type (val); 9320 9321 if (type == type2) 9322 return val; 9323 9324 type2 = ada_check_typedef (type2); 9325 type = ada_check_typedef (type); 9326 9327 if (type2->code () == TYPE_CODE_PTR 9328 && type->code () == TYPE_CODE_ARRAY) 9329 { 9330 val = ada_value_ind (val); 9331 type2 = value_type (val); 9332 } 9333 9334 if (type2->code () == TYPE_CODE_ARRAY 9335 && type->code () == TYPE_CODE_ARRAY) 9336 { 9337 if (!ada_same_array_size_p (type, type2)) 9338 error (_("cannot assign arrays of different length")); 9339 9340 if (is_integral_type (type->target_type ()) 9341 && is_integral_type (type2->target_type ()) 9342 && type2->target_type ()->length () < type->target_type ()->length ()) 9343 { 9344 /* Allow implicit promotion of the array elements to 9345 a wider type. */ 9346 return ada_promote_array_of_integrals (type, val); 9347 } 9348 9349 if (type2->target_type ()->length () != type->target_type ()->length ()) 9350 error (_("Incompatible types in assignment")); 9351 deprecated_set_value_type (val, type); 9352 } 9353 return val; 9354 } 9355 9356 static struct value * 9357 ada_value_binop (struct value *arg1, struct value *arg2, enum exp_opcode op) 9358 { 9359 struct value *val; 9360 struct type *type1, *type2; 9361 LONGEST v, v1, v2; 9362 9363 arg1 = coerce_ref (arg1); 9364 arg2 = coerce_ref (arg2); 9365 type1 = get_base_type (ada_check_typedef (value_type (arg1))); 9366 type2 = get_base_type (ada_check_typedef (value_type (arg2))); 9367 9368 if (type1->code () != TYPE_CODE_INT 9369 || type2->code () != TYPE_CODE_INT) 9370 return value_binop (arg1, arg2, op); 9371 9372 switch (op) 9373 { 9374 case BINOP_MOD: 9375 case BINOP_DIV: 9376 case BINOP_REM: 9377 break; 9378 default: 9379 return value_binop (arg1, arg2, op); 9380 } 9381 9382 v2 = value_as_long (arg2); 9383 if (v2 == 0) 9384 { 9385 const char *name; 9386 if (op == BINOP_MOD) 9387 name = "mod"; 9388 else if (op == BINOP_DIV) 9389 name = "/"; 9390 else 9391 { 9392 gdb_assert (op == BINOP_REM); 9393 name = "rem"; 9394 } 9395 9396 error (_("second operand of %s must not be zero."), name); 9397 } 9398 9399 if (type1->is_unsigned () || op == BINOP_MOD) 9400 return value_binop (arg1, arg2, op); 9401 9402 v1 = value_as_long (arg1); 9403 switch (op) 9404 { 9405 case BINOP_DIV: 9406 v = v1 / v2; 9407 if (!TRUNCATION_TOWARDS_ZERO && v1 * (v1 % v2) < 0) 9408 v += v > 0 ? -1 : 1; 9409 break; 9410 case BINOP_REM: 9411 v = v1 % v2; 9412 if (v * v1 < 0) 9413 v -= v2; 9414 break; 9415 default: 9416 /* Should not reach this point. */ 9417 v = 0; 9418 } 9419 9420 val = allocate_value (type1); 9421 store_unsigned_integer (value_contents_raw (val).data (), 9422 value_type (val)->length (), 9423 type_byte_order (type1), v); 9424 return val; 9425 } 9426 9427 static int 9428 ada_value_equal (struct value *arg1, struct value *arg2) 9429 { 9430 if (ada_is_direct_array_type (value_type (arg1)) 9431 || ada_is_direct_array_type (value_type (arg2))) 9432 { 9433 struct type *arg1_type, *arg2_type; 9434 9435 /* Automatically dereference any array reference before 9436 we attempt to perform the comparison. */ 9437 arg1 = ada_coerce_ref (arg1); 9438 arg2 = ada_coerce_ref (arg2); 9439 9440 arg1 = ada_coerce_to_simple_array (arg1); 9441 arg2 = ada_coerce_to_simple_array (arg2); 9442 9443 arg1_type = ada_check_typedef (value_type (arg1)); 9444 arg2_type = ada_check_typedef (value_type (arg2)); 9445 9446 if (arg1_type->code () != TYPE_CODE_ARRAY 9447 || arg2_type->code () != TYPE_CODE_ARRAY) 9448 error (_("Attempt to compare array with non-array")); 9449 /* FIXME: The following works only for types whose 9450 representations use all bits (no padding or undefined bits) 9451 and do not have user-defined equality. */ 9452 return (arg1_type->length () == arg2_type->length () 9453 && memcmp (value_contents (arg1).data (), 9454 value_contents (arg2).data (), 9455 arg1_type->length ()) == 0); 9456 } 9457 return value_equal (arg1, arg2); 9458 } 9459 9460 namespace expr 9461 { 9462 9463 bool 9464 check_objfile (const std::unique_ptr<ada_component> &comp, 9465 struct objfile *objfile) 9466 { 9467 return comp->uses_objfile (objfile); 9468 } 9469 9470 /* Assign the result of evaluating ARG starting at *POS to the INDEXth 9471 component of LHS (a simple array or a record). Does not modify the 9472 inferior's memory, nor does it modify LHS (unless LHS == 9473 CONTAINER). */ 9474 9475 static void 9476 assign_component (struct value *container, struct value *lhs, LONGEST index, 9477 struct expression *exp, operation_up &arg) 9478 { 9479 scoped_value_mark mark; 9480 9481 struct value *elt; 9482 struct type *lhs_type = check_typedef (value_type (lhs)); 9483 9484 if (lhs_type->code () == TYPE_CODE_ARRAY) 9485 { 9486 struct type *index_type = builtin_type (exp->gdbarch)->builtin_int; 9487 struct value *index_val = value_from_longest (index_type, index); 9488 9489 elt = unwrap_value (ada_value_subscript (lhs, 1, &index_val)); 9490 } 9491 else 9492 { 9493 elt = ada_index_struct_field (index, lhs, 0, value_type (lhs)); 9494 elt = ada_to_fixed_value (elt); 9495 } 9496 9497 ada_aggregate_operation *ag_op 9498 = dynamic_cast<ada_aggregate_operation *> (arg.get ()); 9499 if (ag_op != nullptr) 9500 ag_op->assign_aggregate (container, elt, exp); 9501 else 9502 value_assign_to_component (container, elt, 9503 arg->evaluate (nullptr, exp, 9504 EVAL_NORMAL)); 9505 } 9506 9507 bool 9508 ada_aggregate_component::uses_objfile (struct objfile *objfile) 9509 { 9510 for (const auto &item : m_components) 9511 if (item->uses_objfile (objfile)) 9512 return true; 9513 return false; 9514 } 9515 9516 void 9517 ada_aggregate_component::dump (ui_file *stream, int depth) 9518 { 9519 gdb_printf (stream, _("%*sAggregate\n"), depth, ""); 9520 for (const auto &item : m_components) 9521 item->dump (stream, depth + 1); 9522 } 9523 9524 void 9525 ada_aggregate_component::assign (struct value *container, 9526 struct value *lhs, struct expression *exp, 9527 std::vector<LONGEST> &indices, 9528 LONGEST low, LONGEST high) 9529 { 9530 for (auto &item : m_components) 9531 item->assign (container, lhs, exp, indices, low, high); 9532 } 9533 9534 /* See ada-exp.h. */ 9535 9536 value * 9537 ada_aggregate_operation::assign_aggregate (struct value *container, 9538 struct value *lhs, 9539 struct expression *exp) 9540 { 9541 struct type *lhs_type; 9542 LONGEST low_index, high_index; 9543 9544 container = ada_coerce_ref (container); 9545 if (ada_is_direct_array_type (value_type (container))) 9546 container = ada_coerce_to_simple_array (container); 9547 lhs = ada_coerce_ref (lhs); 9548 if (!deprecated_value_modifiable (lhs)) 9549 error (_("Left operand of assignment is not a modifiable lvalue.")); 9550 9551 lhs_type = check_typedef (value_type (lhs)); 9552 if (ada_is_direct_array_type (lhs_type)) 9553 { 9554 lhs = ada_coerce_to_simple_array (lhs); 9555 lhs_type = check_typedef (value_type (lhs)); 9556 low_index = lhs_type->bounds ()->low.const_val (); 9557 high_index = lhs_type->bounds ()->high.const_val (); 9558 } 9559 else if (lhs_type->code () == TYPE_CODE_STRUCT) 9560 { 9561 low_index = 0; 9562 high_index = num_visible_fields (lhs_type) - 1; 9563 } 9564 else 9565 error (_("Left-hand side must be array or record.")); 9566 9567 std::vector<LONGEST> indices (4); 9568 indices[0] = indices[1] = low_index - 1; 9569 indices[2] = indices[3] = high_index + 1; 9570 9571 std::get<0> (m_storage)->assign (container, lhs, exp, indices, 9572 low_index, high_index); 9573 9574 return container; 9575 } 9576 9577 bool 9578 ada_positional_component::uses_objfile (struct objfile *objfile) 9579 { 9580 return m_op->uses_objfile (objfile); 9581 } 9582 9583 void 9584 ada_positional_component::dump (ui_file *stream, int depth) 9585 { 9586 gdb_printf (stream, _("%*sPositional, index = %d\n"), 9587 depth, "", m_index); 9588 m_op->dump (stream, depth + 1); 9589 } 9590 9591 /* Assign into the component of LHS indexed by the OP_POSITIONAL 9592 construct, given that the positions are relative to lower bound 9593 LOW, where HIGH is the upper bound. Record the position in 9594 INDICES. CONTAINER is as for assign_aggregate. */ 9595 void 9596 ada_positional_component::assign (struct value *container, 9597 struct value *lhs, struct expression *exp, 9598 std::vector<LONGEST> &indices, 9599 LONGEST low, LONGEST high) 9600 { 9601 LONGEST ind = m_index + low; 9602 9603 if (ind - 1 == high) 9604 warning (_("Extra components in aggregate ignored.")); 9605 if (ind <= high) 9606 { 9607 add_component_interval (ind, ind, indices); 9608 assign_component (container, lhs, ind, exp, m_op); 9609 } 9610 } 9611 9612 bool 9613 ada_discrete_range_association::uses_objfile (struct objfile *objfile) 9614 { 9615 return m_low->uses_objfile (objfile) || m_high->uses_objfile (objfile); 9616 } 9617 9618 void 9619 ada_discrete_range_association::dump (ui_file *stream, int depth) 9620 { 9621 gdb_printf (stream, _("%*sDiscrete range:\n"), depth, ""); 9622 m_low->dump (stream, depth + 1); 9623 m_high->dump (stream, depth + 1); 9624 } 9625 9626 void 9627 ada_discrete_range_association::assign (struct value *container, 9628 struct value *lhs, 9629 struct expression *exp, 9630 std::vector<LONGEST> &indices, 9631 LONGEST low, LONGEST high, 9632 operation_up &op) 9633 { 9634 LONGEST lower = value_as_long (m_low->evaluate (nullptr, exp, EVAL_NORMAL)); 9635 LONGEST upper = value_as_long (m_high->evaluate (nullptr, exp, EVAL_NORMAL)); 9636 9637 if (lower <= upper && (lower < low || upper > high)) 9638 error (_("Index in component association out of bounds.")); 9639 9640 add_component_interval (lower, upper, indices); 9641 while (lower <= upper) 9642 { 9643 assign_component (container, lhs, lower, exp, op); 9644 lower += 1; 9645 } 9646 } 9647 9648 bool 9649 ada_name_association::uses_objfile (struct objfile *objfile) 9650 { 9651 return m_val->uses_objfile (objfile); 9652 } 9653 9654 void 9655 ada_name_association::dump (ui_file *stream, int depth) 9656 { 9657 gdb_printf (stream, _("%*sName:\n"), depth, ""); 9658 m_val->dump (stream, depth + 1); 9659 } 9660 9661 void 9662 ada_name_association::assign (struct value *container, 9663 struct value *lhs, 9664 struct expression *exp, 9665 std::vector<LONGEST> &indices, 9666 LONGEST low, LONGEST high, 9667 operation_up &op) 9668 { 9669 int index; 9670 9671 if (ada_is_direct_array_type (value_type (lhs))) 9672 index = longest_to_int (value_as_long (m_val->evaluate (nullptr, exp, 9673 EVAL_NORMAL))); 9674 else 9675 { 9676 ada_string_operation *strop 9677 = dynamic_cast<ada_string_operation *> (m_val.get ()); 9678 9679 const char *name; 9680 if (strop != nullptr) 9681 name = strop->get_name (); 9682 else 9683 { 9684 ada_var_value_operation *vvo 9685 = dynamic_cast<ada_var_value_operation *> (m_val.get ()); 9686 if (vvo != nullptr) 9687 error (_("Invalid record component association.")); 9688 name = vvo->get_symbol ()->natural_name (); 9689 } 9690 9691 index = 0; 9692 if (! find_struct_field (name, value_type (lhs), 0, 9693 NULL, NULL, NULL, NULL, &index)) 9694 error (_("Unknown component name: %s."), name); 9695 } 9696 9697 add_component_interval (index, index, indices); 9698 assign_component (container, lhs, index, exp, op); 9699 } 9700 9701 bool 9702 ada_choices_component::uses_objfile (struct objfile *objfile) 9703 { 9704 if (m_op->uses_objfile (objfile)) 9705 return true; 9706 for (const auto &item : m_assocs) 9707 if (item->uses_objfile (objfile)) 9708 return true; 9709 return false; 9710 } 9711 9712 void 9713 ada_choices_component::dump (ui_file *stream, int depth) 9714 { 9715 gdb_printf (stream, _("%*sChoices:\n"), depth, ""); 9716 m_op->dump (stream, depth + 1); 9717 for (const auto &item : m_assocs) 9718 item->dump (stream, depth + 1); 9719 } 9720 9721 /* Assign into the components of LHS indexed by the OP_CHOICES 9722 construct at *POS, updating *POS past the construct, given that 9723 the allowable indices are LOW..HIGH. Record the indices assigned 9724 to in INDICES. CONTAINER is as for assign_aggregate. */ 9725 void 9726 ada_choices_component::assign (struct value *container, 9727 struct value *lhs, struct expression *exp, 9728 std::vector<LONGEST> &indices, 9729 LONGEST low, LONGEST high) 9730 { 9731 for (auto &item : m_assocs) 9732 item->assign (container, lhs, exp, indices, low, high, m_op); 9733 } 9734 9735 bool 9736 ada_others_component::uses_objfile (struct objfile *objfile) 9737 { 9738 return m_op->uses_objfile (objfile); 9739 } 9740 9741 void 9742 ada_others_component::dump (ui_file *stream, int depth) 9743 { 9744 gdb_printf (stream, _("%*sOthers:\n"), depth, ""); 9745 m_op->dump (stream, depth + 1); 9746 } 9747 9748 /* Assign the value of the expression in the OP_OTHERS construct in 9749 EXP at *POS into the components of LHS indexed from LOW .. HIGH that 9750 have not been previously assigned. The index intervals already assigned 9751 are in INDICES. CONTAINER is as for assign_aggregate. */ 9752 void 9753 ada_others_component::assign (struct value *container, 9754 struct value *lhs, struct expression *exp, 9755 std::vector<LONGEST> &indices, 9756 LONGEST low, LONGEST high) 9757 { 9758 int num_indices = indices.size (); 9759 for (int i = 0; i < num_indices - 2; i += 2) 9760 { 9761 for (LONGEST ind = indices[i + 1] + 1; ind < indices[i + 2]; ind += 1) 9762 assign_component (container, lhs, ind, exp, m_op); 9763 } 9764 } 9765 9766 struct value * 9767 ada_assign_operation::evaluate (struct type *expect_type, 9768 struct expression *exp, 9769 enum noside noside) 9770 { 9771 value *arg1 = std::get<0> (m_storage)->evaluate (nullptr, exp, noside); 9772 9773 ada_aggregate_operation *ag_op 9774 = dynamic_cast<ada_aggregate_operation *> (std::get<1> (m_storage).get ()); 9775 if (ag_op != nullptr) 9776 { 9777 if (noside != EVAL_NORMAL) 9778 return arg1; 9779 9780 arg1 = ag_op->assign_aggregate (arg1, arg1, exp); 9781 return ada_value_assign (arg1, arg1); 9782 } 9783 /* Force the evaluation of the rhs ARG2 to the type of the lhs ARG1, 9784 except if the lhs of our assignment is a convenience variable. 9785 In the case of assigning to a convenience variable, the lhs 9786 should be exactly the result of the evaluation of the rhs. */ 9787 struct type *type = value_type (arg1); 9788 if (VALUE_LVAL (arg1) == lval_internalvar) 9789 type = NULL; 9790 value *arg2 = std::get<1> (m_storage)->evaluate (type, exp, noside); 9791 if (noside == EVAL_AVOID_SIDE_EFFECTS) 9792 return arg1; 9793 if (VALUE_LVAL (arg1) == lval_internalvar) 9794 { 9795 /* Nothing. */ 9796 } 9797 else 9798 arg2 = coerce_for_assign (value_type (arg1), arg2); 9799 return ada_value_assign (arg1, arg2); 9800 } 9801 9802 } /* namespace expr */ 9803 9804 /* Add the interval [LOW .. HIGH] to the sorted set of intervals 9805 [ INDICES[0] .. INDICES[1] ],... The resulting intervals do not 9806 overlap. */ 9807 static void 9808 add_component_interval (LONGEST low, LONGEST high, 9809 std::vector<LONGEST> &indices) 9810 { 9811 int i, j; 9812 9813 int size = indices.size (); 9814 for (i = 0; i < size; i += 2) { 9815 if (high >= indices[i] && low <= indices[i + 1]) 9816 { 9817 int kh; 9818 9819 for (kh = i + 2; kh < size; kh += 2) 9820 if (high < indices[kh]) 9821 break; 9822 if (low < indices[i]) 9823 indices[i] = low; 9824 indices[i + 1] = indices[kh - 1]; 9825 if (high > indices[i + 1]) 9826 indices[i + 1] = high; 9827 memcpy (indices.data () + i + 2, indices.data () + kh, size - kh); 9828 indices.resize (kh - i - 2); 9829 return; 9830 } 9831 else if (high < indices[i]) 9832 break; 9833 } 9834 9835 indices.resize (indices.size () + 2); 9836 for (j = indices.size () - 1; j >= i + 2; j -= 1) 9837 indices[j] = indices[j - 2]; 9838 indices[i] = low; 9839 indices[i + 1] = high; 9840 } 9841 9842 /* Perform and Ada cast of ARG2 to type TYPE if the type of ARG2 9843 is different. */ 9844 9845 static struct value * 9846 ada_value_cast (struct type *type, struct value *arg2) 9847 { 9848 if (type == ada_check_typedef (value_type (arg2))) 9849 return arg2; 9850 9851 return value_cast (type, arg2); 9852 } 9853 9854 /* Evaluating Ada expressions, and printing their result. 9855 ------------------------------------------------------ 9856 9857 1. Introduction: 9858 ---------------- 9859 9860 We usually evaluate an Ada expression in order to print its value. 9861 We also evaluate an expression in order to print its type, which 9862 happens during the EVAL_AVOID_SIDE_EFFECTS phase of the evaluation, 9863 but we'll focus mostly on the EVAL_NORMAL phase. In practice, the 9864 EVAL_AVOID_SIDE_EFFECTS phase allows us to simplify certain aspects of 9865 the evaluation compared to the EVAL_NORMAL, but is otherwise very 9866 similar. 9867 9868 Evaluating expressions is a little more complicated for Ada entities 9869 than it is for entities in languages such as C. The main reason for 9870 this is that Ada provides types whose definition might be dynamic. 9871 One example of such types is variant records. Or another example 9872 would be an array whose bounds can only be known at run time. 9873 9874 The following description is a general guide as to what should be 9875 done (and what should NOT be done) in order to evaluate an expression 9876 involving such types, and when. This does not cover how the semantic 9877 information is encoded by GNAT as this is covered separatly. For the 9878 document used as the reference for the GNAT encoding, see exp_dbug.ads 9879 in the GNAT sources. 9880 9881 Ideally, we should embed each part of this description next to its 9882 associated code. Unfortunately, the amount of code is so vast right 9883 now that it's hard to see whether the code handling a particular 9884 situation might be duplicated or not. One day, when the code is 9885 cleaned up, this guide might become redundant with the comments 9886 inserted in the code, and we might want to remove it. 9887 9888 2. ``Fixing'' an Entity, the Simple Case: 9889 ----------------------------------------- 9890 9891 When evaluating Ada expressions, the tricky issue is that they may 9892 reference entities whose type contents and size are not statically 9893 known. Consider for instance a variant record: 9894 9895 type Rec (Empty : Boolean := True) is record 9896 case Empty is 9897 when True => null; 9898 when False => Value : Integer; 9899 end case; 9900 end record; 9901 Yes : Rec := (Empty => False, Value => 1); 9902 No : Rec := (empty => True); 9903 9904 The size and contents of that record depends on the value of the 9905 descriminant (Rec.Empty). At this point, neither the debugging 9906 information nor the associated type structure in GDB are able to 9907 express such dynamic types. So what the debugger does is to create 9908 "fixed" versions of the type that applies to the specific object. 9909 We also informally refer to this operation as "fixing" an object, 9910 which means creating its associated fixed type. 9911 9912 Example: when printing the value of variable "Yes" above, its fixed 9913 type would look like this: 9914 9915 type Rec is record 9916 Empty : Boolean; 9917 Value : Integer; 9918 end record; 9919 9920 On the other hand, if we printed the value of "No", its fixed type 9921 would become: 9922 9923 type Rec is record 9924 Empty : Boolean; 9925 end record; 9926 9927 Things become a little more complicated when trying to fix an entity 9928 with a dynamic type that directly contains another dynamic type, 9929 such as an array of variant records, for instance. There are 9930 two possible cases: Arrays, and records. 9931 9932 3. ``Fixing'' Arrays: 9933 --------------------- 9934 9935 The type structure in GDB describes an array in terms of its bounds, 9936 and the type of its elements. By design, all elements in the array 9937 have the same type and we cannot represent an array of variant elements 9938 using the current type structure in GDB. When fixing an array, 9939 we cannot fix the array element, as we would potentially need one 9940 fixed type per element of the array. As a result, the best we can do 9941 when fixing an array is to produce an array whose bounds and size 9942 are correct (allowing us to read it from memory), but without having 9943 touched its element type. Fixing each element will be done later, 9944 when (if) necessary. 9945 9946 Arrays are a little simpler to handle than records, because the same 9947 amount of memory is allocated for each element of the array, even if 9948 the amount of space actually used by each element differs from element 9949 to element. Consider for instance the following array of type Rec: 9950 9951 type Rec_Array is array (1 .. 2) of Rec; 9952 9953 The actual amount of memory occupied by each element might be different 9954 from element to element, depending on the value of their discriminant. 9955 But the amount of space reserved for each element in the array remains 9956 fixed regardless. So we simply need to compute that size using 9957 the debugging information available, from which we can then determine 9958 the array size (we multiply the number of elements of the array by 9959 the size of each element). 9960 9961 The simplest case is when we have an array of a constrained element 9962 type. For instance, consider the following type declarations: 9963 9964 type Bounded_String (Max_Size : Integer) is 9965 Length : Integer; 9966 Buffer : String (1 .. Max_Size); 9967 end record; 9968 type Bounded_String_Array is array (1 ..2) of Bounded_String (80); 9969 9970 In this case, the compiler describes the array as an array of 9971 variable-size elements (identified by its XVS suffix) for which 9972 the size can be read in the parallel XVZ variable. 9973 9974 In the case of an array of an unconstrained element type, the compiler 9975 wraps the array element inside a private PAD type. This type should not 9976 be shown to the user, and must be "unwrap"'ed before printing. Note 9977 that we also use the adjective "aligner" in our code to designate 9978 these wrapper types. 9979 9980 In some cases, the size allocated for each element is statically 9981 known. In that case, the PAD type already has the correct size, 9982 and the array element should remain unfixed. 9983 9984 But there are cases when this size is not statically known. 9985 For instance, assuming that "Five" is an integer variable: 9986 9987 type Dynamic is array (1 .. Five) of Integer; 9988 type Wrapper (Has_Length : Boolean := False) is record 9989 Data : Dynamic; 9990 case Has_Length is 9991 when True => Length : Integer; 9992 when False => null; 9993 end case; 9994 end record; 9995 type Wrapper_Array is array (1 .. 2) of Wrapper; 9996 9997 Hello : Wrapper_Array := (others => (Has_Length => True, 9998 Data => (others => 17), 9999 Length => 1)); 10000 10001 10002 The debugging info would describe variable Hello as being an 10003 array of a PAD type. The size of that PAD type is not statically 10004 known, but can be determined using a parallel XVZ variable. 10005 In that case, a copy of the PAD type with the correct size should 10006 be used for the fixed array. 10007 10008 3. ``Fixing'' record type objects: 10009 ---------------------------------- 10010 10011 Things are slightly different from arrays in the case of dynamic 10012 record types. In this case, in order to compute the associated 10013 fixed type, we need to determine the size and offset of each of 10014 its components. This, in turn, requires us to compute the fixed 10015 type of each of these components. 10016 10017 Consider for instance the example: 10018 10019 type Bounded_String (Max_Size : Natural) is record 10020 Str : String (1 .. Max_Size); 10021 Length : Natural; 10022 end record; 10023 My_String : Bounded_String (Max_Size => 10); 10024 10025 In that case, the position of field "Length" depends on the size 10026 of field Str, which itself depends on the value of the Max_Size 10027 discriminant. In order to fix the type of variable My_String, 10028 we need to fix the type of field Str. Therefore, fixing a variant 10029 record requires us to fix each of its components. 10030 10031 However, if a component does not have a dynamic size, the component 10032 should not be fixed. In particular, fields that use a PAD type 10033 should not fixed. Here is an example where this might happen 10034 (assuming type Rec above): 10035 10036 type Container (Big : Boolean) is record 10037 First : Rec; 10038 After : Integer; 10039 case Big is 10040 when True => Another : Integer; 10041 when False => null; 10042 end case; 10043 end record; 10044 My_Container : Container := (Big => False, 10045 First => (Empty => True), 10046 After => 42); 10047 10048 In that example, the compiler creates a PAD type for component First, 10049 whose size is constant, and then positions the component After just 10050 right after it. The offset of component After is therefore constant 10051 in this case. 10052 10053 The debugger computes the position of each field based on an algorithm 10054 that uses, among other things, the actual position and size of the field 10055 preceding it. Let's now imagine that the user is trying to print 10056 the value of My_Container. If the type fixing was recursive, we would 10057 end up computing the offset of field After based on the size of the 10058 fixed version of field First. And since in our example First has 10059 only one actual field, the size of the fixed type is actually smaller 10060 than the amount of space allocated to that field, and thus we would 10061 compute the wrong offset of field After. 10062 10063 To make things more complicated, we need to watch out for dynamic 10064 components of variant records (identified by the ___XVL suffix in 10065 the component name). Even if the target type is a PAD type, the size 10066 of that type might not be statically known. So the PAD type needs 10067 to be unwrapped and the resulting type needs to be fixed. Otherwise, 10068 we might end up with the wrong size for our component. This can be 10069 observed with the following type declarations: 10070 10071 type Octal is new Integer range 0 .. 7; 10072 type Octal_Array is array (Positive range <>) of Octal; 10073 pragma Pack (Octal_Array); 10074 10075 type Octal_Buffer (Size : Positive) is record 10076 Buffer : Octal_Array (1 .. Size); 10077 Length : Integer; 10078 end record; 10079 10080 In that case, Buffer is a PAD type whose size is unset and needs 10081 to be computed by fixing the unwrapped type. 10082 10083 4. When to ``Fix'' un-``Fixed'' sub-elements of an entity: 10084 ---------------------------------------------------------- 10085 10086 Lastly, when should the sub-elements of an entity that remained unfixed 10087 thus far, be actually fixed? 10088 10089 The answer is: Only when referencing that element. For instance 10090 when selecting one component of a record, this specific component 10091 should be fixed at that point in time. Or when printing the value 10092 of a record, each component should be fixed before its value gets 10093 printed. Similarly for arrays, the element of the array should be 10094 fixed when printing each element of the array, or when extracting 10095 one element out of that array. On the other hand, fixing should 10096 not be performed on the elements when taking a slice of an array! 10097 10098 Note that one of the side effects of miscomputing the offset and 10099 size of each field is that we end up also miscomputing the size 10100 of the containing type. This can have adverse results when computing 10101 the value of an entity. GDB fetches the value of an entity based 10102 on the size of its type, and thus a wrong size causes GDB to fetch 10103 the wrong amount of memory. In the case where the computed size is 10104 too small, GDB fetches too little data to print the value of our 10105 entity. Results in this case are unpredictable, as we usually read 10106 past the buffer containing the data =:-o. */ 10107 10108 /* A helper function for TERNOP_IN_RANGE. */ 10109 10110 static value * 10111 eval_ternop_in_range (struct type *expect_type, struct expression *exp, 10112 enum noside noside, 10113 value *arg1, value *arg2, value *arg3) 10114 { 10115 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10116 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3); 10117 struct type *type = language_bool_type (exp->language_defn, exp->gdbarch); 10118 return 10119 value_from_longest (type, 10120 (value_less (arg1, arg3) 10121 || value_equal (arg1, arg3)) 10122 && (value_less (arg2, arg1) 10123 || value_equal (arg2, arg1))); 10124 } 10125 10126 /* A helper function for UNOP_NEG. */ 10127 10128 value * 10129 ada_unop_neg (struct type *expect_type, 10130 struct expression *exp, 10131 enum noside noside, enum exp_opcode op, 10132 struct value *arg1) 10133 { 10134 unop_promote (exp->language_defn, exp->gdbarch, &arg1); 10135 return value_neg (arg1); 10136 } 10137 10138 /* A helper function for UNOP_IN_RANGE. */ 10139 10140 value * 10141 ada_unop_in_range (struct type *expect_type, 10142 struct expression *exp, 10143 enum noside noside, enum exp_opcode op, 10144 struct value *arg1, struct type *type) 10145 { 10146 struct value *arg2, *arg3; 10147 switch (type->code ()) 10148 { 10149 default: 10150 lim_warning (_("Membership test incompletely implemented; " 10151 "always returns true")); 10152 type = language_bool_type (exp->language_defn, exp->gdbarch); 10153 return value_from_longest (type, (LONGEST) 1); 10154 10155 case TYPE_CODE_RANGE: 10156 arg2 = value_from_longest (type, 10157 type->bounds ()->low.const_val ()); 10158 arg3 = value_from_longest (type, 10159 type->bounds ()->high.const_val ()); 10160 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10161 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3); 10162 type = language_bool_type (exp->language_defn, exp->gdbarch); 10163 return 10164 value_from_longest (type, 10165 (value_less (arg1, arg3) 10166 || value_equal (arg1, arg3)) 10167 && (value_less (arg2, arg1) 10168 || value_equal (arg2, arg1))); 10169 } 10170 } 10171 10172 /* A helper function for OP_ATR_TAG. */ 10173 10174 value * 10175 ada_atr_tag (struct type *expect_type, 10176 struct expression *exp, 10177 enum noside noside, enum exp_opcode op, 10178 struct value *arg1) 10179 { 10180 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10181 return value_zero (ada_tag_type (arg1), not_lval); 10182 10183 return ada_value_tag (arg1); 10184 } 10185 10186 /* A helper function for OP_ATR_SIZE. */ 10187 10188 value * 10189 ada_atr_size (struct type *expect_type, 10190 struct expression *exp, 10191 enum noside noside, enum exp_opcode op, 10192 struct value *arg1) 10193 { 10194 struct type *type = value_type (arg1); 10195 10196 /* If the argument is a reference, then dereference its type, since 10197 the user is really asking for the size of the actual object, 10198 not the size of the pointer. */ 10199 if (type->code () == TYPE_CODE_REF) 10200 type = type->target_type (); 10201 10202 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10203 return value_zero (builtin_type (exp->gdbarch)->builtin_int, not_lval); 10204 else 10205 return value_from_longest (builtin_type (exp->gdbarch)->builtin_int, 10206 TARGET_CHAR_BIT * type->length ()); 10207 } 10208 10209 /* A helper function for UNOP_ABS. */ 10210 10211 value * 10212 ada_abs (struct type *expect_type, 10213 struct expression *exp, 10214 enum noside noside, enum exp_opcode op, 10215 struct value *arg1) 10216 { 10217 unop_promote (exp->language_defn, exp->gdbarch, &arg1); 10218 if (value_less (arg1, value_zero (value_type (arg1), not_lval))) 10219 return value_neg (arg1); 10220 else 10221 return arg1; 10222 } 10223 10224 /* A helper function for BINOP_MUL. */ 10225 10226 value * 10227 ada_mult_binop (struct type *expect_type, 10228 struct expression *exp, 10229 enum noside noside, enum exp_opcode op, 10230 struct value *arg1, struct value *arg2) 10231 { 10232 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10233 { 10234 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10235 return value_zero (value_type (arg1), not_lval); 10236 } 10237 else 10238 { 10239 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10240 return ada_value_binop (arg1, arg2, op); 10241 } 10242 } 10243 10244 /* A helper function for BINOP_EQUAL and BINOP_NOTEQUAL. */ 10245 10246 value * 10247 ada_equal_binop (struct type *expect_type, 10248 struct expression *exp, 10249 enum noside noside, enum exp_opcode op, 10250 struct value *arg1, struct value *arg2) 10251 { 10252 int tem; 10253 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10254 tem = 0; 10255 else 10256 { 10257 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10258 tem = ada_value_equal (arg1, arg2); 10259 } 10260 if (op == BINOP_NOTEQUAL) 10261 tem = !tem; 10262 struct type *type = language_bool_type (exp->language_defn, exp->gdbarch); 10263 return value_from_longest (type, (LONGEST) tem); 10264 } 10265 10266 /* A helper function for TERNOP_SLICE. */ 10267 10268 value * 10269 ada_ternop_slice (struct expression *exp, 10270 enum noside noside, 10271 struct value *array, struct value *low_bound_val, 10272 struct value *high_bound_val) 10273 { 10274 LONGEST low_bound; 10275 LONGEST high_bound; 10276 10277 low_bound_val = coerce_ref (low_bound_val); 10278 high_bound_val = coerce_ref (high_bound_val); 10279 low_bound = value_as_long (low_bound_val); 10280 high_bound = value_as_long (high_bound_val); 10281 10282 /* If this is a reference to an aligner type, then remove all 10283 the aligners. */ 10284 if (value_type (array)->code () == TYPE_CODE_REF 10285 && ada_is_aligner_type (value_type (array)->target_type ())) 10286 value_type (array)->set_target_type 10287 (ada_aligned_type (value_type (array)->target_type ())); 10288 10289 if (ada_is_any_packed_array_type (value_type (array))) 10290 error (_("cannot slice a packed array")); 10291 10292 /* If this is a reference to an array or an array lvalue, 10293 convert to a pointer. */ 10294 if (value_type (array)->code () == TYPE_CODE_REF 10295 || (value_type (array)->code () == TYPE_CODE_ARRAY 10296 && VALUE_LVAL (array) == lval_memory)) 10297 array = value_addr (array); 10298 10299 if (noside == EVAL_AVOID_SIDE_EFFECTS 10300 && ada_is_array_descriptor_type (ada_check_typedef 10301 (value_type (array)))) 10302 return empty_array (ada_type_of_array (array, 0), low_bound, 10303 high_bound); 10304 10305 array = ada_coerce_to_simple_array_ptr (array); 10306 10307 /* If we have more than one level of pointer indirection, 10308 dereference the value until we get only one level. */ 10309 while (value_type (array)->code () == TYPE_CODE_PTR 10310 && (value_type (array)->target_type ()->code () 10311 == TYPE_CODE_PTR)) 10312 array = value_ind (array); 10313 10314 /* Make sure we really do have an array type before going further, 10315 to avoid a SEGV when trying to get the index type or the target 10316 type later down the road if the debug info generated by 10317 the compiler is incorrect or incomplete. */ 10318 if (!ada_is_simple_array_type (value_type (array))) 10319 error (_("cannot take slice of non-array")); 10320 10321 if (ada_check_typedef (value_type (array))->code () 10322 == TYPE_CODE_PTR) 10323 { 10324 struct type *type0 = ada_check_typedef (value_type (array)); 10325 10326 if (high_bound < low_bound || noside == EVAL_AVOID_SIDE_EFFECTS) 10327 return empty_array (type0->target_type (), low_bound, high_bound); 10328 else 10329 { 10330 struct type *arr_type0 = 10331 to_fixed_array_type (type0->target_type (), NULL, 1); 10332 10333 return ada_value_slice_from_ptr (array, arr_type0, 10334 longest_to_int (low_bound), 10335 longest_to_int (high_bound)); 10336 } 10337 } 10338 else if (noside == EVAL_AVOID_SIDE_EFFECTS) 10339 return array; 10340 else if (high_bound < low_bound) 10341 return empty_array (value_type (array), low_bound, high_bound); 10342 else 10343 return ada_value_slice (array, longest_to_int (low_bound), 10344 longest_to_int (high_bound)); 10345 } 10346 10347 /* A helper function for BINOP_IN_BOUNDS. */ 10348 10349 value * 10350 ada_binop_in_bounds (struct expression *exp, enum noside noside, 10351 struct value *arg1, struct value *arg2, int n) 10352 { 10353 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10354 { 10355 struct type *type = language_bool_type (exp->language_defn, 10356 exp->gdbarch); 10357 return value_zero (type, not_lval); 10358 } 10359 10360 struct type *type = ada_index_type (value_type (arg2), n, "range"); 10361 if (!type) 10362 type = value_type (arg1); 10363 10364 value *arg3 = value_from_longest (type, ada_array_bound (arg2, n, 1)); 10365 arg2 = value_from_longest (type, ada_array_bound (arg2, n, 0)); 10366 10367 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10368 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg3); 10369 type = language_bool_type (exp->language_defn, exp->gdbarch); 10370 return value_from_longest (type, 10371 (value_less (arg1, arg3) 10372 || value_equal (arg1, arg3)) 10373 && (value_less (arg2, arg1) 10374 || value_equal (arg2, arg1))); 10375 } 10376 10377 /* A helper function for some attribute operations. */ 10378 10379 static value * 10380 ada_unop_atr (struct expression *exp, enum noside noside, enum exp_opcode op, 10381 struct value *arg1, struct type *type_arg, int tem) 10382 { 10383 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10384 { 10385 if (type_arg == NULL) 10386 type_arg = value_type (arg1); 10387 10388 if (ada_is_constrained_packed_array_type (type_arg)) 10389 type_arg = decode_constrained_packed_array_type (type_arg); 10390 10391 if (!discrete_type_p (type_arg)) 10392 { 10393 switch (op) 10394 { 10395 default: /* Should never happen. */ 10396 error (_("unexpected attribute encountered")); 10397 case OP_ATR_FIRST: 10398 case OP_ATR_LAST: 10399 type_arg = ada_index_type (type_arg, tem, 10400 ada_attribute_name (op)); 10401 break; 10402 case OP_ATR_LENGTH: 10403 type_arg = builtin_type (exp->gdbarch)->builtin_int; 10404 break; 10405 } 10406 } 10407 10408 return value_zero (type_arg, not_lval); 10409 } 10410 else if (type_arg == NULL) 10411 { 10412 arg1 = ada_coerce_ref (arg1); 10413 10414 if (ada_is_constrained_packed_array_type (value_type (arg1))) 10415 arg1 = ada_coerce_to_simple_array (arg1); 10416 10417 struct type *type; 10418 if (op == OP_ATR_LENGTH) 10419 type = builtin_type (exp->gdbarch)->builtin_int; 10420 else 10421 { 10422 type = ada_index_type (value_type (arg1), tem, 10423 ada_attribute_name (op)); 10424 if (type == NULL) 10425 type = builtin_type (exp->gdbarch)->builtin_int; 10426 } 10427 10428 switch (op) 10429 { 10430 default: /* Should never happen. */ 10431 error (_("unexpected attribute encountered")); 10432 case OP_ATR_FIRST: 10433 return value_from_longest 10434 (type, ada_array_bound (arg1, tem, 0)); 10435 case OP_ATR_LAST: 10436 return value_from_longest 10437 (type, ada_array_bound (arg1, tem, 1)); 10438 case OP_ATR_LENGTH: 10439 return value_from_longest 10440 (type, ada_array_length (arg1, tem)); 10441 } 10442 } 10443 else if (discrete_type_p (type_arg)) 10444 { 10445 struct type *range_type; 10446 const char *name = ada_type_name (type_arg); 10447 10448 range_type = NULL; 10449 if (name != NULL && type_arg->code () != TYPE_CODE_ENUM) 10450 range_type = to_fixed_range_type (type_arg, NULL); 10451 if (range_type == NULL) 10452 range_type = type_arg; 10453 switch (op) 10454 { 10455 default: 10456 error (_("unexpected attribute encountered")); 10457 case OP_ATR_FIRST: 10458 return value_from_longest 10459 (range_type, ada_discrete_type_low_bound (range_type)); 10460 case OP_ATR_LAST: 10461 return value_from_longest 10462 (range_type, ada_discrete_type_high_bound (range_type)); 10463 case OP_ATR_LENGTH: 10464 error (_("the 'length attribute applies only to array types")); 10465 } 10466 } 10467 else if (type_arg->code () == TYPE_CODE_FLT) 10468 error (_("unimplemented type attribute")); 10469 else 10470 { 10471 LONGEST low, high; 10472 10473 if (ada_is_constrained_packed_array_type (type_arg)) 10474 type_arg = decode_constrained_packed_array_type (type_arg); 10475 10476 struct type *type; 10477 if (op == OP_ATR_LENGTH) 10478 type = builtin_type (exp->gdbarch)->builtin_int; 10479 else 10480 { 10481 type = ada_index_type (type_arg, tem, ada_attribute_name (op)); 10482 if (type == NULL) 10483 type = builtin_type (exp->gdbarch)->builtin_int; 10484 } 10485 10486 switch (op) 10487 { 10488 default: 10489 error (_("unexpected attribute encountered")); 10490 case OP_ATR_FIRST: 10491 low = ada_array_bound_from_type (type_arg, tem, 0); 10492 return value_from_longest (type, low); 10493 case OP_ATR_LAST: 10494 high = ada_array_bound_from_type (type_arg, tem, 1); 10495 return value_from_longest (type, high); 10496 case OP_ATR_LENGTH: 10497 low = ada_array_bound_from_type (type_arg, tem, 0); 10498 high = ada_array_bound_from_type (type_arg, tem, 1); 10499 return value_from_longest (type, high - low + 1); 10500 } 10501 } 10502 } 10503 10504 /* A helper function for OP_ATR_MIN and OP_ATR_MAX. */ 10505 10506 struct value * 10507 ada_binop_minmax (struct type *expect_type, 10508 struct expression *exp, 10509 enum noside noside, enum exp_opcode op, 10510 struct value *arg1, struct value *arg2) 10511 { 10512 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10513 return value_zero (value_type (arg1), not_lval); 10514 else 10515 { 10516 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10517 return value_binop (arg1, arg2, op); 10518 } 10519 } 10520 10521 /* A helper function for BINOP_EXP. */ 10522 10523 struct value * 10524 ada_binop_exp (struct type *expect_type, 10525 struct expression *exp, 10526 enum noside noside, enum exp_opcode op, 10527 struct value *arg1, struct value *arg2) 10528 { 10529 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10530 return value_zero (value_type (arg1), not_lval); 10531 else 10532 { 10533 /* For integer exponentiation operations, 10534 only promote the first argument. */ 10535 if (is_integral_type (value_type (arg2))) 10536 unop_promote (exp->language_defn, exp->gdbarch, &arg1); 10537 else 10538 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10539 10540 return value_binop (arg1, arg2, op); 10541 } 10542 } 10543 10544 namespace expr 10545 { 10546 10547 /* See ada-exp.h. */ 10548 10549 operation_up 10550 ada_resolvable::replace (operation_up &&owner, 10551 struct expression *exp, 10552 bool deprocedure_p, 10553 bool parse_completion, 10554 innermost_block_tracker *tracker, 10555 struct type *context_type) 10556 { 10557 if (resolve (exp, deprocedure_p, parse_completion, tracker, context_type)) 10558 return (make_operation<ada_funcall_operation> 10559 (std::move (owner), 10560 std::vector<operation_up> ())); 10561 return std::move (owner); 10562 } 10563 10564 /* Convert the character literal whose value would be VAL to the 10565 appropriate value of type TYPE, if there is a translation. 10566 Otherwise return VAL. Hence, in an enumeration type ('A', 'B'), 10567 the literal 'A' (VAL == 65), returns 0. */ 10568 10569 static LONGEST 10570 convert_char_literal (struct type *type, LONGEST val) 10571 { 10572 char name[12]; 10573 int f; 10574 10575 if (type == NULL) 10576 return val; 10577 type = check_typedef (type); 10578 if (type->code () != TYPE_CODE_ENUM) 10579 return val; 10580 10581 if ((val >= 'a' && val <= 'z') || (val >= '0' && val <= '9')) 10582 xsnprintf (name, sizeof (name), "Q%c", (int) val); 10583 else if (val >= 0 && val < 256) 10584 xsnprintf (name, sizeof (name), "QU%02x", (unsigned) val); 10585 else if (val >= 0 && val < 0x10000) 10586 xsnprintf (name, sizeof (name), "QW%04x", (unsigned) val); 10587 else 10588 xsnprintf (name, sizeof (name), "QWW%08lx", (unsigned long) val); 10589 size_t len = strlen (name); 10590 for (f = 0; f < type->num_fields (); f += 1) 10591 { 10592 /* Check the suffix because an enum constant in a package will 10593 have a name like "pkg__QUxx". This is safe enough because we 10594 already have the correct type, and because mangling means 10595 there can't be clashes. */ 10596 const char *ename = type->field (f).name (); 10597 size_t elen = strlen (ename); 10598 10599 if (elen >= len && strcmp (name, ename + elen - len) == 0) 10600 return type->field (f).loc_enumval (); 10601 } 10602 return val; 10603 } 10604 10605 value * 10606 ada_char_operation::evaluate (struct type *expect_type, 10607 struct expression *exp, 10608 enum noside noside) 10609 { 10610 value *result = long_const_operation::evaluate (expect_type, exp, noside); 10611 if (expect_type != nullptr) 10612 result = ada_value_cast (expect_type, result); 10613 return result; 10614 } 10615 10616 /* See ada-exp.h. */ 10617 10618 operation_up 10619 ada_char_operation::replace (operation_up &&owner, 10620 struct expression *exp, 10621 bool deprocedure_p, 10622 bool parse_completion, 10623 innermost_block_tracker *tracker, 10624 struct type *context_type) 10625 { 10626 operation_up result = std::move (owner); 10627 10628 if (context_type != nullptr && context_type->code () == TYPE_CODE_ENUM) 10629 { 10630 gdb_assert (result.get () == this); 10631 std::get<0> (m_storage) = context_type; 10632 std::get<1> (m_storage) 10633 = convert_char_literal (context_type, std::get<1> (m_storage)); 10634 } 10635 10636 return result; 10637 } 10638 10639 value * 10640 ada_wrapped_operation::evaluate (struct type *expect_type, 10641 struct expression *exp, 10642 enum noside noside) 10643 { 10644 value *result = std::get<0> (m_storage)->evaluate (expect_type, exp, noside); 10645 if (noside == EVAL_NORMAL) 10646 result = unwrap_value (result); 10647 10648 /* If evaluating an OP_FLOAT and an EXPECT_TYPE was provided, 10649 then we need to perform the conversion manually, because 10650 evaluate_subexp_standard doesn't do it. This conversion is 10651 necessary in Ada because the different kinds of float/fixed 10652 types in Ada have different representations. 10653 10654 Similarly, we need to perform the conversion from OP_LONG 10655 ourselves. */ 10656 if ((opcode () == OP_FLOAT || opcode () == OP_LONG) && expect_type != NULL) 10657 result = ada_value_cast (expect_type, result); 10658 10659 return result; 10660 } 10661 10662 value * 10663 ada_string_operation::evaluate (struct type *expect_type, 10664 struct expression *exp, 10665 enum noside noside) 10666 { 10667 struct type *char_type; 10668 if (expect_type != nullptr && ada_is_string_type (expect_type)) 10669 char_type = ada_array_element_type (expect_type, 1); 10670 else 10671 char_type = language_string_char_type (exp->language_defn, exp->gdbarch); 10672 10673 const std::string &str = std::get<0> (m_storage); 10674 const char *encoding; 10675 switch (char_type->length ()) 10676 { 10677 case 1: 10678 { 10679 /* Simply copy over the data -- this isn't perhaps strictly 10680 correct according to the encodings, but it is gdb's 10681 historical behavior. */ 10682 struct type *stringtype 10683 = lookup_array_range_type (char_type, 1, str.length ()); 10684 struct value *val = allocate_value (stringtype); 10685 memcpy (value_contents_raw (val).data (), str.c_str (), 10686 str.length ()); 10687 return val; 10688 } 10689 10690 case 2: 10691 if (gdbarch_byte_order (exp->gdbarch) == BFD_ENDIAN_BIG) 10692 encoding = "UTF-16BE"; 10693 else 10694 encoding = "UTF-16LE"; 10695 break; 10696 10697 case 4: 10698 if (gdbarch_byte_order (exp->gdbarch) == BFD_ENDIAN_BIG) 10699 encoding = "UTF-32BE"; 10700 else 10701 encoding = "UTF-32LE"; 10702 break; 10703 10704 default: 10705 error (_("unexpected character type size %s"), 10706 pulongest (char_type->length ())); 10707 } 10708 10709 auto_obstack converted; 10710 convert_between_encodings (host_charset (), encoding, 10711 (const gdb_byte *) str.c_str (), 10712 str.length (), 1, 10713 &converted, translit_none); 10714 10715 struct type *stringtype 10716 = lookup_array_range_type (char_type, 1, 10717 obstack_object_size (&converted) 10718 / char_type->length ()); 10719 struct value *val = allocate_value (stringtype); 10720 memcpy (value_contents_raw (val).data (), 10721 obstack_base (&converted), 10722 obstack_object_size (&converted)); 10723 return val; 10724 } 10725 10726 value * 10727 ada_concat_operation::evaluate (struct type *expect_type, 10728 struct expression *exp, 10729 enum noside noside) 10730 { 10731 /* If one side is a literal, evaluate the other side first so that 10732 the expected type can be set properly. */ 10733 const operation_up &lhs_expr = std::get<0> (m_storage); 10734 const operation_up &rhs_expr = std::get<1> (m_storage); 10735 10736 value *lhs, *rhs; 10737 if (dynamic_cast<ada_string_operation *> (lhs_expr.get ()) != nullptr) 10738 { 10739 rhs = rhs_expr->evaluate (nullptr, exp, noside); 10740 lhs = lhs_expr->evaluate (value_type (rhs), exp, noside); 10741 } 10742 else if (dynamic_cast<ada_char_operation *> (lhs_expr.get ()) != nullptr) 10743 { 10744 rhs = rhs_expr->evaluate (nullptr, exp, noside); 10745 struct type *rhs_type = check_typedef (value_type (rhs)); 10746 struct type *elt_type = nullptr; 10747 if (rhs_type->code () == TYPE_CODE_ARRAY) 10748 elt_type = rhs_type->target_type (); 10749 lhs = lhs_expr->evaluate (elt_type, exp, noside); 10750 } 10751 else if (dynamic_cast<ada_string_operation *> (rhs_expr.get ()) != nullptr) 10752 { 10753 lhs = lhs_expr->evaluate (nullptr, exp, noside); 10754 rhs = rhs_expr->evaluate (value_type (lhs), exp, noside); 10755 } 10756 else if (dynamic_cast<ada_char_operation *> (rhs_expr.get ()) != nullptr) 10757 { 10758 lhs = lhs_expr->evaluate (nullptr, exp, noside); 10759 struct type *lhs_type = check_typedef (value_type (lhs)); 10760 struct type *elt_type = nullptr; 10761 if (lhs_type->code () == TYPE_CODE_ARRAY) 10762 elt_type = lhs_type->target_type (); 10763 rhs = rhs_expr->evaluate (elt_type, exp, noside); 10764 } 10765 else 10766 return concat_operation::evaluate (expect_type, exp, noside); 10767 10768 return value_concat (lhs, rhs); 10769 } 10770 10771 value * 10772 ada_qual_operation::evaluate (struct type *expect_type, 10773 struct expression *exp, 10774 enum noside noside) 10775 { 10776 struct type *type = std::get<1> (m_storage); 10777 return std::get<0> (m_storage)->evaluate (type, exp, noside); 10778 } 10779 10780 value * 10781 ada_ternop_range_operation::evaluate (struct type *expect_type, 10782 struct expression *exp, 10783 enum noside noside) 10784 { 10785 value *arg0 = std::get<0> (m_storage)->evaluate (nullptr, exp, noside); 10786 value *arg1 = std::get<1> (m_storage)->evaluate (nullptr, exp, noside); 10787 value *arg2 = std::get<2> (m_storage)->evaluate (nullptr, exp, noside); 10788 return eval_ternop_in_range (expect_type, exp, noside, arg0, arg1, arg2); 10789 } 10790 10791 value * 10792 ada_binop_addsub_operation::evaluate (struct type *expect_type, 10793 struct expression *exp, 10794 enum noside noside) 10795 { 10796 value *arg1 = std::get<1> (m_storage)->evaluate_with_coercion (exp, noside); 10797 value *arg2 = std::get<2> (m_storage)->evaluate_with_coercion (exp, noside); 10798 10799 auto do_op = [=] (LONGEST x, LONGEST y) 10800 { 10801 if (std::get<0> (m_storage) == BINOP_ADD) 10802 return x + y; 10803 return x - y; 10804 }; 10805 10806 if (value_type (arg1)->code () == TYPE_CODE_PTR) 10807 return (value_from_longest 10808 (value_type (arg1), 10809 do_op (value_as_long (arg1), value_as_long (arg2)))); 10810 if (value_type (arg2)->code () == TYPE_CODE_PTR) 10811 return (value_from_longest 10812 (value_type (arg2), 10813 do_op (value_as_long (arg1), value_as_long (arg2)))); 10814 /* Preserve the original type for use by the range case below. 10815 We cannot cast the result to a reference type, so if ARG1 is 10816 a reference type, find its underlying type. */ 10817 struct type *type = value_type (arg1); 10818 while (type->code () == TYPE_CODE_REF) 10819 type = type->target_type (); 10820 binop_promote (exp->language_defn, exp->gdbarch, &arg1, &arg2); 10821 arg1 = value_binop (arg1, arg2, std::get<0> (m_storage)); 10822 /* We need to special-case the result with a range. 10823 This is done for the benefit of "ptype". gdb's Ada support 10824 historically used the LHS to set the result type here, so 10825 preserve this behavior. */ 10826 if (type->code () == TYPE_CODE_RANGE) 10827 arg1 = value_cast (type, arg1); 10828 return arg1; 10829 } 10830 10831 value * 10832 ada_unop_atr_operation::evaluate (struct type *expect_type, 10833 struct expression *exp, 10834 enum noside noside) 10835 { 10836 struct type *type_arg = nullptr; 10837 value *val = nullptr; 10838 10839 if (std::get<0> (m_storage)->opcode () == OP_TYPE) 10840 { 10841 value *tem = std::get<0> (m_storage)->evaluate (nullptr, exp, 10842 EVAL_AVOID_SIDE_EFFECTS); 10843 type_arg = value_type (tem); 10844 } 10845 else 10846 val = std::get<0> (m_storage)->evaluate (nullptr, exp, noside); 10847 10848 return ada_unop_atr (exp, noside, std::get<1> (m_storage), 10849 val, type_arg, std::get<2> (m_storage)); 10850 } 10851 10852 value * 10853 ada_var_msym_value_operation::evaluate_for_cast (struct type *expect_type, 10854 struct expression *exp, 10855 enum noside noside) 10856 { 10857 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10858 return value_zero (expect_type, not_lval); 10859 10860 const bound_minimal_symbol &b = std::get<0> (m_storage); 10861 value *val = evaluate_var_msym_value (noside, b.objfile, b.minsym); 10862 10863 val = ada_value_cast (expect_type, val); 10864 10865 /* Follow the Ada language semantics that do not allow taking 10866 an address of the result of a cast (view conversion in Ada). */ 10867 if (VALUE_LVAL (val) == lval_memory) 10868 { 10869 if (value_lazy (val)) 10870 value_fetch_lazy (val); 10871 VALUE_LVAL (val) = not_lval; 10872 } 10873 return val; 10874 } 10875 10876 value * 10877 ada_var_value_operation::evaluate_for_cast (struct type *expect_type, 10878 struct expression *exp, 10879 enum noside noside) 10880 { 10881 value *val = evaluate_var_value (noside, 10882 std::get<0> (m_storage).block, 10883 std::get<0> (m_storage).symbol); 10884 10885 val = ada_value_cast (expect_type, val); 10886 10887 /* Follow the Ada language semantics that do not allow taking 10888 an address of the result of a cast (view conversion in Ada). */ 10889 if (VALUE_LVAL (val) == lval_memory) 10890 { 10891 if (value_lazy (val)) 10892 value_fetch_lazy (val); 10893 VALUE_LVAL (val) = not_lval; 10894 } 10895 return val; 10896 } 10897 10898 value * 10899 ada_var_value_operation::evaluate (struct type *expect_type, 10900 struct expression *exp, 10901 enum noside noside) 10902 { 10903 symbol *sym = std::get<0> (m_storage).symbol; 10904 10905 if (sym->domain () == UNDEF_DOMAIN) 10906 /* Only encountered when an unresolved symbol occurs in a 10907 context other than a function call, in which case, it is 10908 invalid. */ 10909 error (_("Unexpected unresolved symbol, %s, during evaluation"), 10910 sym->print_name ()); 10911 10912 if (noside == EVAL_AVOID_SIDE_EFFECTS) 10913 { 10914 struct type *type = static_unwrap_type (sym->type ()); 10915 /* Check to see if this is a tagged type. We also need to handle 10916 the case where the type is a reference to a tagged type, but 10917 we have to be careful to exclude pointers to tagged types. 10918 The latter should be shown as usual (as a pointer), whereas 10919 a reference should mostly be transparent to the user. */ 10920 if (ada_is_tagged_type (type, 0) 10921 || (type->code () == TYPE_CODE_REF 10922 && ada_is_tagged_type (type->target_type (), 0))) 10923 { 10924 /* Tagged types are a little special in the fact that the real 10925 type is dynamic and can only be determined by inspecting the 10926 object's tag. This means that we need to get the object's 10927 value first (EVAL_NORMAL) and then extract the actual object 10928 type from its tag. 10929 10930 Note that we cannot skip the final step where we extract 10931 the object type from its tag, because the EVAL_NORMAL phase 10932 results in dynamic components being resolved into fixed ones. 10933 This can cause problems when trying to print the type 10934 description of tagged types whose parent has a dynamic size: 10935 We use the type name of the "_parent" component in order 10936 to print the name of the ancestor type in the type description. 10937 If that component had a dynamic size, the resolution into 10938 a fixed type would result in the loss of that type name, 10939 thus preventing us from printing the name of the ancestor 10940 type in the type description. */ 10941 value *arg1 = evaluate (nullptr, exp, EVAL_NORMAL); 10942 10943 if (type->code () != TYPE_CODE_REF) 10944 { 10945 struct type *actual_type; 10946 10947 actual_type = type_from_tag (ada_value_tag (arg1)); 10948 if (actual_type == NULL) 10949 /* If, for some reason, we were unable to determine 10950 the actual type from the tag, then use the static 10951 approximation that we just computed as a fallback. 10952 This can happen if the debugging information is 10953 incomplete, for instance. */ 10954 actual_type = type; 10955 return value_zero (actual_type, not_lval); 10956 } 10957 else 10958 { 10959 /* In the case of a ref, ada_coerce_ref takes care 10960 of determining the actual type. But the evaluation 10961 should return a ref as it should be valid to ask 10962 for its address; so rebuild a ref after coerce. */ 10963 arg1 = ada_coerce_ref (arg1); 10964 return value_ref (arg1, TYPE_CODE_REF); 10965 } 10966 } 10967 10968 /* Records and unions for which GNAT encodings have been 10969 generated need to be statically fixed as well. 10970 Otherwise, non-static fixing produces a type where 10971 all dynamic properties are removed, which prevents "ptype" 10972 from being able to completely describe the type. 10973 For instance, a case statement in a variant record would be 10974 replaced by the relevant components based on the actual 10975 value of the discriminants. */ 10976 if ((type->code () == TYPE_CODE_STRUCT 10977 && dynamic_template_type (type) != NULL) 10978 || (type->code () == TYPE_CODE_UNION 10979 && ada_find_parallel_type (type, "___XVU") != NULL)) 10980 return value_zero (to_static_fixed_type (type), not_lval); 10981 } 10982 10983 value *arg1 = var_value_operation::evaluate (expect_type, exp, noside); 10984 return ada_to_fixed_value (arg1); 10985 } 10986 10987 bool 10988 ada_var_value_operation::resolve (struct expression *exp, 10989 bool deprocedure_p, 10990 bool parse_completion, 10991 innermost_block_tracker *tracker, 10992 struct type *context_type) 10993 { 10994 symbol *sym = std::get<0> (m_storage).symbol; 10995 if (sym->domain () == UNDEF_DOMAIN) 10996 { 10997 block_symbol resolved 10998 = ada_resolve_variable (sym, std::get<0> (m_storage).block, 10999 context_type, parse_completion, 11000 deprocedure_p, tracker); 11001 std::get<0> (m_storage) = resolved; 11002 } 11003 11004 if (deprocedure_p 11005 && (std::get<0> (m_storage).symbol->type ()->code () 11006 == TYPE_CODE_FUNC)) 11007 return true; 11008 11009 return false; 11010 } 11011 11012 value * 11013 ada_atr_val_operation::evaluate (struct type *expect_type, 11014 struct expression *exp, 11015 enum noside noside) 11016 { 11017 value *arg = std::get<1> (m_storage)->evaluate (nullptr, exp, noside); 11018 return ada_val_atr (noside, std::get<0> (m_storage), arg); 11019 } 11020 11021 value * 11022 ada_unop_ind_operation::evaluate (struct type *expect_type, 11023 struct expression *exp, 11024 enum noside noside) 11025 { 11026 value *arg1 = std::get<0> (m_storage)->evaluate (expect_type, exp, noside); 11027 11028 struct type *type = ada_check_typedef (value_type (arg1)); 11029 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11030 { 11031 if (ada_is_array_descriptor_type (type)) 11032 /* GDB allows dereferencing GNAT array descriptors. */ 11033 { 11034 struct type *arrType = ada_type_of_array (arg1, 0); 11035 11036 if (arrType == NULL) 11037 error (_("Attempt to dereference null array pointer.")); 11038 return value_at_lazy (arrType, 0); 11039 } 11040 else if (type->code () == TYPE_CODE_PTR 11041 || type->code () == TYPE_CODE_REF 11042 /* In C you can dereference an array to get the 1st elt. */ 11043 || type->code () == TYPE_CODE_ARRAY) 11044 { 11045 /* As mentioned in the OP_VAR_VALUE case, tagged types can 11046 only be determined by inspecting the object's tag. 11047 This means that we need to evaluate completely the 11048 expression in order to get its type. */ 11049 11050 if ((type->code () == TYPE_CODE_REF 11051 || type->code () == TYPE_CODE_PTR) 11052 && ada_is_tagged_type (type->target_type (), 0)) 11053 { 11054 arg1 = std::get<0> (m_storage)->evaluate (nullptr, exp, 11055 EVAL_NORMAL); 11056 type = value_type (ada_value_ind (arg1)); 11057 } 11058 else 11059 { 11060 type = to_static_fixed_type 11061 (ada_aligned_type 11062 (ada_check_typedef (type->target_type ()))); 11063 } 11064 return value_zero (type, lval_memory); 11065 } 11066 else if (type->code () == TYPE_CODE_INT) 11067 { 11068 /* GDB allows dereferencing an int. */ 11069 if (expect_type == NULL) 11070 return value_zero (builtin_type (exp->gdbarch)->builtin_int, 11071 lval_memory); 11072 else 11073 { 11074 expect_type = 11075 to_static_fixed_type (ada_aligned_type (expect_type)); 11076 return value_zero (expect_type, lval_memory); 11077 } 11078 } 11079 else 11080 error (_("Attempt to take contents of a non-pointer value.")); 11081 } 11082 arg1 = ada_coerce_ref (arg1); /* FIXME: What is this for?? */ 11083 type = ada_check_typedef (value_type (arg1)); 11084 11085 if (type->code () == TYPE_CODE_INT) 11086 /* GDB allows dereferencing an int. If we were given 11087 the expect_type, then use that as the target type. 11088 Otherwise, assume that the target type is an int. */ 11089 { 11090 if (expect_type != NULL) 11091 return ada_value_ind (value_cast (lookup_pointer_type (expect_type), 11092 arg1)); 11093 else 11094 return value_at_lazy (builtin_type (exp->gdbarch)->builtin_int, 11095 (CORE_ADDR) value_as_address (arg1)); 11096 } 11097 11098 if (ada_is_array_descriptor_type (type)) 11099 /* GDB allows dereferencing GNAT array descriptors. */ 11100 return ada_coerce_to_simple_array (arg1); 11101 else 11102 return ada_value_ind (arg1); 11103 } 11104 11105 value * 11106 ada_structop_operation::evaluate (struct type *expect_type, 11107 struct expression *exp, 11108 enum noside noside) 11109 { 11110 value *arg1 = std::get<0> (m_storage)->evaluate (nullptr, exp, noside); 11111 const char *str = std::get<1> (m_storage).c_str (); 11112 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11113 { 11114 struct type *type; 11115 struct type *type1 = value_type (arg1); 11116 11117 if (ada_is_tagged_type (type1, 1)) 11118 { 11119 type = ada_lookup_struct_elt_type (type1, str, 1, 1); 11120 11121 /* If the field is not found, check if it exists in the 11122 extension of this object's type. This means that we 11123 need to evaluate completely the expression. */ 11124 11125 if (type == NULL) 11126 { 11127 arg1 = std::get<0> (m_storage)->evaluate (nullptr, exp, 11128 EVAL_NORMAL); 11129 arg1 = ada_value_struct_elt (arg1, str, 0); 11130 arg1 = unwrap_value (arg1); 11131 type = value_type (ada_to_fixed_value (arg1)); 11132 } 11133 } 11134 else 11135 type = ada_lookup_struct_elt_type (type1, str, 1, 0); 11136 11137 return value_zero (ada_aligned_type (type), lval_memory); 11138 } 11139 else 11140 { 11141 arg1 = ada_value_struct_elt (arg1, str, 0); 11142 arg1 = unwrap_value (arg1); 11143 return ada_to_fixed_value (arg1); 11144 } 11145 } 11146 11147 value * 11148 ada_funcall_operation::evaluate (struct type *expect_type, 11149 struct expression *exp, 11150 enum noside noside) 11151 { 11152 const std::vector<operation_up> &args_up = std::get<1> (m_storage); 11153 int nargs = args_up.size (); 11154 std::vector<value *> argvec (nargs); 11155 operation_up &callee_op = std::get<0> (m_storage); 11156 11157 ada_var_value_operation *avv 11158 = dynamic_cast<ada_var_value_operation *> (callee_op.get ()); 11159 if (avv != nullptr 11160 && avv->get_symbol ()->domain () == UNDEF_DOMAIN) 11161 error (_("Unexpected unresolved symbol, %s, during evaluation"), 11162 avv->get_symbol ()->print_name ()); 11163 11164 value *callee = callee_op->evaluate (nullptr, exp, noside); 11165 for (int i = 0; i < args_up.size (); ++i) 11166 argvec[i] = args_up[i]->evaluate (nullptr, exp, noside); 11167 11168 if (ada_is_constrained_packed_array_type 11169 (desc_base_type (value_type (callee)))) 11170 callee = ada_coerce_to_simple_array (callee); 11171 else if (value_type (callee)->code () == TYPE_CODE_ARRAY 11172 && TYPE_FIELD_BITSIZE (value_type (callee), 0) != 0) 11173 /* This is a packed array that has already been fixed, and 11174 therefore already coerced to a simple array. Nothing further 11175 to do. */ 11176 ; 11177 else if (value_type (callee)->code () == TYPE_CODE_REF) 11178 { 11179 /* Make sure we dereference references so that all the code below 11180 feels like it's really handling the referenced value. Wrapping 11181 types (for alignment) may be there, so make sure we strip them as 11182 well. */ 11183 callee = ada_to_fixed_value (coerce_ref (callee)); 11184 } 11185 else if (value_type (callee)->code () == TYPE_CODE_ARRAY 11186 && VALUE_LVAL (callee) == lval_memory) 11187 callee = value_addr (callee); 11188 11189 struct type *type = ada_check_typedef (value_type (callee)); 11190 11191 /* Ada allows us to implicitly dereference arrays when subscripting 11192 them. So, if this is an array typedef (encoding use for array 11193 access types encoded as fat pointers), strip it now. */ 11194 if (type->code () == TYPE_CODE_TYPEDEF) 11195 type = ada_typedef_target_type (type); 11196 11197 if (type->code () == TYPE_CODE_PTR) 11198 { 11199 switch (ada_check_typedef (type->target_type ())->code ()) 11200 { 11201 case TYPE_CODE_FUNC: 11202 type = ada_check_typedef (type->target_type ()); 11203 break; 11204 case TYPE_CODE_ARRAY: 11205 break; 11206 case TYPE_CODE_STRUCT: 11207 if (noside != EVAL_AVOID_SIDE_EFFECTS) 11208 callee = ada_value_ind (callee); 11209 type = ada_check_typedef (type->target_type ()); 11210 break; 11211 default: 11212 error (_("cannot subscript or call something of type `%s'"), 11213 ada_type_name (value_type (callee))); 11214 break; 11215 } 11216 } 11217 11218 switch (type->code ()) 11219 { 11220 case TYPE_CODE_FUNC: 11221 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11222 { 11223 if (type->target_type () == NULL) 11224 error_call_unknown_return_type (NULL); 11225 return allocate_value (type->target_type ()); 11226 } 11227 return call_function_by_hand (callee, NULL, argvec); 11228 case TYPE_CODE_INTERNAL_FUNCTION: 11229 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11230 /* We don't know anything about what the internal 11231 function might return, but we have to return 11232 something. */ 11233 return value_zero (builtin_type (exp->gdbarch)->builtin_int, 11234 not_lval); 11235 else 11236 return call_internal_function (exp->gdbarch, exp->language_defn, 11237 callee, nargs, 11238 argvec.data ()); 11239 11240 case TYPE_CODE_STRUCT: 11241 { 11242 int arity; 11243 11244 arity = ada_array_arity (type); 11245 type = ada_array_element_type (type, nargs); 11246 if (type == NULL) 11247 error (_("cannot subscript or call a record")); 11248 if (arity != nargs) 11249 error (_("wrong number of subscripts; expecting %d"), arity); 11250 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11251 return value_zero (ada_aligned_type (type), lval_memory); 11252 return 11253 unwrap_value (ada_value_subscript 11254 (callee, nargs, argvec.data ())); 11255 } 11256 case TYPE_CODE_ARRAY: 11257 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11258 { 11259 type = ada_array_element_type (type, nargs); 11260 if (type == NULL) 11261 error (_("element type of array unknown")); 11262 else 11263 return value_zero (ada_aligned_type (type), lval_memory); 11264 } 11265 return 11266 unwrap_value (ada_value_subscript 11267 (ada_coerce_to_simple_array (callee), 11268 nargs, argvec.data ())); 11269 case TYPE_CODE_PTR: /* Pointer to array */ 11270 if (noside == EVAL_AVOID_SIDE_EFFECTS) 11271 { 11272 type = to_fixed_array_type (type->target_type (), NULL, 1); 11273 type = ada_array_element_type (type, nargs); 11274 if (type == NULL) 11275 error (_("element type of array unknown")); 11276 else 11277 return value_zero (ada_aligned_type (type), lval_memory); 11278 } 11279 return 11280 unwrap_value (ada_value_ptr_subscript (callee, nargs, 11281 argvec.data ())); 11282 11283 default: 11284 error (_("Attempt to index or call something other than an " 11285 "array or function")); 11286 } 11287 } 11288 11289 bool 11290 ada_funcall_operation::resolve (struct expression *exp, 11291 bool deprocedure_p, 11292 bool parse_completion, 11293 innermost_block_tracker *tracker, 11294 struct type *context_type) 11295 { 11296 operation_up &callee_op = std::get<0> (m_storage); 11297 11298 ada_var_value_operation *avv 11299 = dynamic_cast<ada_var_value_operation *> (callee_op.get ()); 11300 if (avv == nullptr) 11301 return false; 11302 11303 symbol *sym = avv->get_symbol (); 11304 if (sym->domain () != UNDEF_DOMAIN) 11305 return false; 11306 11307 const std::vector<operation_up> &args_up = std::get<1> (m_storage); 11308 int nargs = args_up.size (); 11309 std::vector<value *> argvec (nargs); 11310 11311 for (int i = 0; i < args_up.size (); ++i) 11312 argvec[i] = args_up[i]->evaluate (nullptr, exp, EVAL_AVOID_SIDE_EFFECTS); 11313 11314 const block *block = avv->get_block (); 11315 block_symbol resolved 11316 = ada_resolve_funcall (sym, block, 11317 context_type, parse_completion, 11318 nargs, argvec.data (), 11319 tracker); 11320 11321 std::get<0> (m_storage) 11322 = make_operation<ada_var_value_operation> (resolved); 11323 return false; 11324 } 11325 11326 bool 11327 ada_ternop_slice_operation::resolve (struct expression *exp, 11328 bool deprocedure_p, 11329 bool parse_completion, 11330 innermost_block_tracker *tracker, 11331 struct type *context_type) 11332 { 11333 /* Historically this check was done during resolution, so we 11334 continue that here. */ 11335 value *v = std::get<0> (m_storage)->evaluate (context_type, exp, 11336 EVAL_AVOID_SIDE_EFFECTS); 11337 if (ada_is_any_packed_array_type (value_type (v))) 11338 error (_("cannot slice a packed array")); 11339 return false; 11340 } 11341 11342 } 11343 11344 11345 11346 /* Return non-zero iff TYPE represents a System.Address type. */ 11347 11348 int 11349 ada_is_system_address_type (struct type *type) 11350 { 11351 return (type->name () && strcmp (type->name (), "system__address") == 0); 11352 } 11353 11354 11355 11356 /* Range types */ 11357 11358 /* Scan STR beginning at position K for a discriminant name, and 11359 return the value of that discriminant field of DVAL in *PX. If 11360 PNEW_K is not null, put the position of the character beyond the 11361 name scanned in *PNEW_K. Return 1 if successful; return 0 and do 11362 not alter *PX and *PNEW_K if unsuccessful. */ 11363 11364 static int 11365 scan_discrim_bound (const char *str, int k, struct value *dval, LONGEST * px, 11366 int *pnew_k) 11367 { 11368 static std::string storage; 11369 const char *pstart, *pend, *bound; 11370 struct value *bound_val; 11371 11372 if (dval == NULL || str == NULL || str[k] == '\0') 11373 return 0; 11374 11375 pstart = str + k; 11376 pend = strstr (pstart, "__"); 11377 if (pend == NULL) 11378 { 11379 bound = pstart; 11380 k += strlen (bound); 11381 } 11382 else 11383 { 11384 int len = pend - pstart; 11385 11386 /* Strip __ and beyond. */ 11387 storage = std::string (pstart, len); 11388 bound = storage.c_str (); 11389 k = pend - str; 11390 } 11391 11392 bound_val = ada_search_struct_field (bound, dval, 0, value_type (dval)); 11393 if (bound_val == NULL) 11394 return 0; 11395 11396 *px = value_as_long (bound_val); 11397 if (pnew_k != NULL) 11398 *pnew_k = k; 11399 return 1; 11400 } 11401 11402 /* Value of variable named NAME. Only exact matches are considered. 11403 If no such variable found, then if ERR_MSG is null, returns 0, and 11404 otherwise causes an error with message ERR_MSG. */ 11405 11406 static struct value * 11407 get_var_value (const char *name, const char *err_msg) 11408 { 11409 std::string quoted_name = add_angle_brackets (name); 11410 11411 lookup_name_info lookup_name (quoted_name, symbol_name_match_type::FULL); 11412 11413 std::vector<struct block_symbol> syms 11414 = ada_lookup_symbol_list_worker (lookup_name, 11415 get_selected_block (0), 11416 VAR_DOMAIN, 1); 11417 11418 if (syms.size () != 1) 11419 { 11420 if (err_msg == NULL) 11421 return 0; 11422 else 11423 error (("%s"), err_msg); 11424 } 11425 11426 return value_of_variable (syms[0].symbol, syms[0].block); 11427 } 11428 11429 /* Value of integer variable named NAME in the current environment. 11430 If no such variable is found, returns false. Otherwise, sets VALUE 11431 to the variable's value and returns true. */ 11432 11433 bool 11434 get_int_var_value (const char *name, LONGEST &value) 11435 { 11436 struct value *var_val = get_var_value (name, 0); 11437 11438 if (var_val == 0) 11439 return false; 11440 11441 value = value_as_long (var_val); 11442 return true; 11443 } 11444 11445 11446 /* Return a range type whose base type is that of the range type named 11447 NAME in the current environment, and whose bounds are calculated 11448 from NAME according to the GNAT range encoding conventions. 11449 Extract discriminant values, if needed, from DVAL. ORIG_TYPE is the 11450 corresponding range type from debug information; fall back to using it 11451 if symbol lookup fails. If a new type must be created, allocate it 11452 like ORIG_TYPE was. The bounds information, in general, is encoded 11453 in NAME, the base type given in the named range type. */ 11454 11455 static struct type * 11456 to_fixed_range_type (struct type *raw_type, struct value *dval) 11457 { 11458 const char *name; 11459 struct type *base_type; 11460 const char *subtype_info; 11461 11462 gdb_assert (raw_type != NULL); 11463 gdb_assert (raw_type->name () != NULL); 11464 11465 if (raw_type->code () == TYPE_CODE_RANGE) 11466 base_type = raw_type->target_type (); 11467 else 11468 base_type = raw_type; 11469 11470 name = raw_type->name (); 11471 subtype_info = strstr (name, "___XD"); 11472 if (subtype_info == NULL) 11473 { 11474 LONGEST L = ada_discrete_type_low_bound (raw_type); 11475 LONGEST U = ada_discrete_type_high_bound (raw_type); 11476 11477 if (L < INT_MIN || U > INT_MAX) 11478 return raw_type; 11479 else 11480 return create_static_range_type (alloc_type_copy (raw_type), raw_type, 11481 L, U); 11482 } 11483 else 11484 { 11485 int prefix_len = subtype_info - name; 11486 LONGEST L, U; 11487 struct type *type; 11488 const char *bounds_str; 11489 int n; 11490 11491 subtype_info += 5; 11492 bounds_str = strchr (subtype_info, '_'); 11493 n = 1; 11494 11495 if (*subtype_info == 'L') 11496 { 11497 if (!ada_scan_number (bounds_str, n, &L, &n) 11498 && !scan_discrim_bound (bounds_str, n, dval, &L, &n)) 11499 return raw_type; 11500 if (bounds_str[n] == '_') 11501 n += 2; 11502 else if (bounds_str[n] == '.') /* FIXME? SGI Workshop kludge. */ 11503 n += 1; 11504 subtype_info += 1; 11505 } 11506 else 11507 { 11508 std::string name_buf = std::string (name, prefix_len) + "___L"; 11509 if (!get_int_var_value (name_buf.c_str (), L)) 11510 { 11511 lim_warning (_("Unknown lower bound, using 1.")); 11512 L = 1; 11513 } 11514 } 11515 11516 if (*subtype_info == 'U') 11517 { 11518 if (!ada_scan_number (bounds_str, n, &U, &n) 11519 && !scan_discrim_bound (bounds_str, n, dval, &U, &n)) 11520 return raw_type; 11521 } 11522 else 11523 { 11524 std::string name_buf = std::string (name, prefix_len) + "___U"; 11525 if (!get_int_var_value (name_buf.c_str (), U)) 11526 { 11527 lim_warning (_("Unknown upper bound, using %ld."), (long) L); 11528 U = L; 11529 } 11530 } 11531 11532 type = create_static_range_type (alloc_type_copy (raw_type), 11533 base_type, L, U); 11534 /* create_static_range_type alters the resulting type's length 11535 to match the size of the base_type, which is not what we want. 11536 Set it back to the original range type's length. */ 11537 type->set_length (raw_type->length ()); 11538 type->set_name (name); 11539 return type; 11540 } 11541 } 11542 11543 /* True iff NAME is the name of a range type. */ 11544 11545 int 11546 ada_is_range_type_name (const char *name) 11547 { 11548 return (name != NULL && strstr (name, "___XD")); 11549 } 11550 11551 11552 /* Modular types */ 11553 11554 /* True iff TYPE is an Ada modular type. */ 11555 11556 int 11557 ada_is_modular_type (struct type *type) 11558 { 11559 struct type *subranged_type = get_base_type (type); 11560 11561 return (subranged_type != NULL && type->code () == TYPE_CODE_RANGE 11562 && subranged_type->code () == TYPE_CODE_INT 11563 && subranged_type->is_unsigned ()); 11564 } 11565 11566 /* Assuming ada_is_modular_type (TYPE), the modulus of TYPE. */ 11567 11568 ULONGEST 11569 ada_modulus (struct type *type) 11570 { 11571 const dynamic_prop &high = type->bounds ()->high; 11572 11573 if (high.kind () == PROP_CONST) 11574 return (ULONGEST) high.const_val () + 1; 11575 11576 /* If TYPE is unresolved, the high bound might be a location list. Return 11577 0, for lack of a better value to return. */ 11578 return 0; 11579 } 11580 11581 11582 /* Ada exception catchpoint support: 11583 --------------------------------- 11584 11585 We support 3 kinds of exception catchpoints: 11586 . catchpoints on Ada exceptions 11587 . catchpoints on unhandled Ada exceptions 11588 . catchpoints on failed assertions 11589 11590 Exceptions raised during failed assertions, or unhandled exceptions 11591 could perfectly be caught with the general catchpoint on Ada exceptions. 11592 However, we can easily differentiate these two special cases, and having 11593 the option to distinguish these two cases from the rest can be useful 11594 to zero-in on certain situations. 11595 11596 Exception catchpoints are a specialized form of breakpoint, 11597 since they rely on inserting breakpoints inside known routines 11598 of the GNAT runtime. The implementation therefore uses a standard 11599 breakpoint structure of the BP_BREAKPOINT type, but with its own set 11600 of breakpoint_ops. 11601 11602 Support in the runtime for exception catchpoints have been changed 11603 a few times already, and these changes affect the implementation 11604 of these catchpoints. In order to be able to support several 11605 variants of the runtime, we use a sniffer that will determine 11606 the runtime variant used by the program being debugged. */ 11607 11608 /* Ada's standard exceptions. 11609 11610 The Ada 83 standard also defined Numeric_Error. But there so many 11611 situations where it was unclear from the Ada 83 Reference Manual 11612 (RM) whether Constraint_Error or Numeric_Error should be raised, 11613 that the ARG (Ada Rapporteur Group) eventually issued a Binding 11614 Interpretation saying that anytime the RM says that Numeric_Error 11615 should be raised, the implementation may raise Constraint_Error. 11616 Ada 95 went one step further and pretty much removed Numeric_Error 11617 from the list of standard exceptions (it made it a renaming of 11618 Constraint_Error, to help preserve compatibility when compiling 11619 an Ada83 compiler). As such, we do not include Numeric_Error from 11620 this list of standard exceptions. */ 11621 11622 static const char * const standard_exc[] = { 11623 "constraint_error", 11624 "program_error", 11625 "storage_error", 11626 "tasking_error" 11627 }; 11628 11629 typedef CORE_ADDR (ada_unhandled_exception_name_addr_ftype) (void); 11630 11631 /* A structure that describes how to support exception catchpoints 11632 for a given executable. */ 11633 11634 struct exception_support_info 11635 { 11636 /* The name of the symbol to break on in order to insert 11637 a catchpoint on exceptions. */ 11638 const char *catch_exception_sym; 11639 11640 /* The name of the symbol to break on in order to insert 11641 a catchpoint on unhandled exceptions. */ 11642 const char *catch_exception_unhandled_sym; 11643 11644 /* The name of the symbol to break on in order to insert 11645 a catchpoint on failed assertions. */ 11646 const char *catch_assert_sym; 11647 11648 /* The name of the symbol to break on in order to insert 11649 a catchpoint on exception handling. */ 11650 const char *catch_handlers_sym; 11651 11652 /* Assuming that the inferior just triggered an unhandled exception 11653 catchpoint, this function is responsible for returning the address 11654 in inferior memory where the name of that exception is stored. 11655 Return zero if the address could not be computed. */ 11656 ada_unhandled_exception_name_addr_ftype *unhandled_exception_name_addr; 11657 }; 11658 11659 static CORE_ADDR ada_unhandled_exception_name_addr (void); 11660 static CORE_ADDR ada_unhandled_exception_name_addr_from_raise (void); 11661 11662 /* The following exception support info structure describes how to 11663 implement exception catchpoints with the latest version of the 11664 Ada runtime (as of 2019-08-??). */ 11665 11666 static const struct exception_support_info default_exception_support_info = 11667 { 11668 "__gnat_debug_raise_exception", /* catch_exception_sym */ 11669 "__gnat_unhandled_exception", /* catch_exception_unhandled_sym */ 11670 "__gnat_debug_raise_assert_failure", /* catch_assert_sym */ 11671 "__gnat_begin_handler_v1", /* catch_handlers_sym */ 11672 ada_unhandled_exception_name_addr 11673 }; 11674 11675 /* The following exception support info structure describes how to 11676 implement exception catchpoints with an earlier version of the 11677 Ada runtime (as of 2007-03-06) using v0 of the EH ABI. */ 11678 11679 static const struct exception_support_info exception_support_info_v0 = 11680 { 11681 "__gnat_debug_raise_exception", /* catch_exception_sym */ 11682 "__gnat_unhandled_exception", /* catch_exception_unhandled_sym */ 11683 "__gnat_debug_raise_assert_failure", /* catch_assert_sym */ 11684 "__gnat_begin_handler", /* catch_handlers_sym */ 11685 ada_unhandled_exception_name_addr 11686 }; 11687 11688 /* The following exception support info structure describes how to 11689 implement exception catchpoints with a slightly older version 11690 of the Ada runtime. */ 11691 11692 static const struct exception_support_info exception_support_info_fallback = 11693 { 11694 "__gnat_raise_nodefer_with_msg", /* catch_exception_sym */ 11695 "__gnat_unhandled_exception", /* catch_exception_unhandled_sym */ 11696 "system__assertions__raise_assert_failure", /* catch_assert_sym */ 11697 "__gnat_begin_handler", /* catch_handlers_sym */ 11698 ada_unhandled_exception_name_addr_from_raise 11699 }; 11700 11701 /* Return nonzero if we can detect the exception support routines 11702 described in EINFO. 11703 11704 This function errors out if an abnormal situation is detected 11705 (for instance, if we find the exception support routines, but 11706 that support is found to be incomplete). */ 11707 11708 static int 11709 ada_has_this_exception_support (const struct exception_support_info *einfo) 11710 { 11711 struct symbol *sym; 11712 11713 /* The symbol we're looking up is provided by a unit in the GNAT runtime 11714 that should be compiled with debugging information. As a result, we 11715 expect to find that symbol in the symtabs. */ 11716 11717 sym = standard_lookup (einfo->catch_exception_sym, NULL, VAR_DOMAIN); 11718 if (sym == NULL) 11719 { 11720 /* Perhaps we did not find our symbol because the Ada runtime was 11721 compiled without debugging info, or simply stripped of it. 11722 It happens on some GNU/Linux distributions for instance, where 11723 users have to install a separate debug package in order to get 11724 the runtime's debugging info. In that situation, let the user 11725 know why we cannot insert an Ada exception catchpoint. 11726 11727 Note: Just for the purpose of inserting our Ada exception 11728 catchpoint, we could rely purely on the associated minimal symbol. 11729 But we would be operating in degraded mode anyway, since we are 11730 still lacking the debugging info needed later on to extract 11731 the name of the exception being raised (this name is printed in 11732 the catchpoint message, and is also used when trying to catch 11733 a specific exception). We do not handle this case for now. */ 11734 struct bound_minimal_symbol msym 11735 = lookup_minimal_symbol (einfo->catch_exception_sym, NULL, NULL); 11736 11737 if (msym.minsym && msym.minsym->type () != mst_solib_trampoline) 11738 error (_("Your Ada runtime appears to be missing some debugging " 11739 "information.\nCannot insert Ada exception catchpoint " 11740 "in this configuration.")); 11741 11742 return 0; 11743 } 11744 11745 /* Make sure that the symbol we found corresponds to a function. */ 11746 11747 if (sym->aclass () != LOC_BLOCK) 11748 { 11749 error (_("Symbol \"%s\" is not a function (class = %d)"), 11750 sym->linkage_name (), sym->aclass ()); 11751 return 0; 11752 } 11753 11754 sym = standard_lookup (einfo->catch_handlers_sym, NULL, VAR_DOMAIN); 11755 if (sym == NULL) 11756 { 11757 struct bound_minimal_symbol msym 11758 = lookup_minimal_symbol (einfo->catch_handlers_sym, NULL, NULL); 11759 11760 if (msym.minsym && msym.minsym->type () != mst_solib_trampoline) 11761 error (_("Your Ada runtime appears to be missing some debugging " 11762 "information.\nCannot insert Ada exception catchpoint " 11763 "in this configuration.")); 11764 11765 return 0; 11766 } 11767 11768 /* Make sure that the symbol we found corresponds to a function. */ 11769 11770 if (sym->aclass () != LOC_BLOCK) 11771 { 11772 error (_("Symbol \"%s\" is not a function (class = %d)"), 11773 sym->linkage_name (), sym->aclass ()); 11774 return 0; 11775 } 11776 11777 return 1; 11778 } 11779 11780 /* Inspect the Ada runtime and determine which exception info structure 11781 should be used to provide support for exception catchpoints. 11782 11783 This function will always set the per-inferior exception_info, 11784 or raise an error. */ 11785 11786 static void 11787 ada_exception_support_info_sniffer (void) 11788 { 11789 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ()); 11790 11791 /* If the exception info is already known, then no need to recompute it. */ 11792 if (data->exception_info != NULL) 11793 return; 11794 11795 /* Check the latest (default) exception support info. */ 11796 if (ada_has_this_exception_support (&default_exception_support_info)) 11797 { 11798 data->exception_info = &default_exception_support_info; 11799 return; 11800 } 11801 11802 /* Try the v0 exception suport info. */ 11803 if (ada_has_this_exception_support (&exception_support_info_v0)) 11804 { 11805 data->exception_info = &exception_support_info_v0; 11806 return; 11807 } 11808 11809 /* Try our fallback exception suport info. */ 11810 if (ada_has_this_exception_support (&exception_support_info_fallback)) 11811 { 11812 data->exception_info = &exception_support_info_fallback; 11813 return; 11814 } 11815 11816 /* Sometimes, it is normal for us to not be able to find the routine 11817 we are looking for. This happens when the program is linked with 11818 the shared version of the GNAT runtime, and the program has not been 11819 started yet. Inform the user of these two possible causes if 11820 applicable. */ 11821 11822 if (ada_update_initial_language (language_unknown) != language_ada) 11823 error (_("Unable to insert catchpoint. Is this an Ada main program?")); 11824 11825 /* If the symbol does not exist, then check that the program is 11826 already started, to make sure that shared libraries have been 11827 loaded. If it is not started, this may mean that the symbol is 11828 in a shared library. */ 11829 11830 if (inferior_ptid.pid () == 0) 11831 error (_("Unable to insert catchpoint. Try to start the program first.")); 11832 11833 /* At this point, we know that we are debugging an Ada program and 11834 that the inferior has been started, but we still are not able to 11835 find the run-time symbols. That can mean that we are in 11836 configurable run time mode, or that a-except as been optimized 11837 out by the linker... In any case, at this point it is not worth 11838 supporting this feature. */ 11839 11840 error (_("Cannot insert Ada exception catchpoints in this configuration.")); 11841 } 11842 11843 /* True iff FRAME is very likely to be that of a function that is 11844 part of the runtime system. This is all very heuristic, but is 11845 intended to be used as advice as to what frames are uninteresting 11846 to most users. */ 11847 11848 static int 11849 is_known_support_routine (frame_info_ptr frame) 11850 { 11851 enum language func_lang; 11852 int i; 11853 const char *fullname; 11854 11855 /* If this code does not have any debugging information (no symtab), 11856 This cannot be any user code. */ 11857 11858 symtab_and_line sal = find_frame_sal (frame); 11859 if (sal.symtab == NULL) 11860 return 1; 11861 11862 /* If there is a symtab, but the associated source file cannot be 11863 located, then assume this is not user code: Selecting a frame 11864 for which we cannot display the code would not be very helpful 11865 for the user. This should also take care of case such as VxWorks 11866 where the kernel has some debugging info provided for a few units. */ 11867 11868 fullname = symtab_to_fullname (sal.symtab); 11869 if (access (fullname, R_OK) != 0) 11870 return 1; 11871 11872 /* Check the unit filename against the Ada runtime file naming. 11873 We also check the name of the objfile against the name of some 11874 known system libraries that sometimes come with debugging info 11875 too. */ 11876 11877 for (i = 0; known_runtime_file_name_patterns[i] != NULL; i += 1) 11878 { 11879 re_comp (known_runtime_file_name_patterns[i]); 11880 if (re_exec (lbasename (sal.symtab->filename))) 11881 return 1; 11882 if (sal.symtab->compunit ()->objfile () != NULL 11883 && re_exec (objfile_name (sal.symtab->compunit ()->objfile ()))) 11884 return 1; 11885 } 11886 11887 /* Check whether the function is a GNAT-generated entity. */ 11888 11889 gdb::unique_xmalloc_ptr<char> func_name 11890 = find_frame_funname (frame, &func_lang, NULL); 11891 if (func_name == NULL) 11892 return 1; 11893 11894 for (i = 0; known_auxiliary_function_name_patterns[i] != NULL; i += 1) 11895 { 11896 re_comp (known_auxiliary_function_name_patterns[i]); 11897 if (re_exec (func_name.get ())) 11898 return 1; 11899 } 11900 11901 return 0; 11902 } 11903 11904 /* Find the first frame that contains debugging information and that is not 11905 part of the Ada run-time, starting from FI and moving upward. */ 11906 11907 void 11908 ada_find_printable_frame (frame_info_ptr fi) 11909 { 11910 for (; fi != NULL; fi = get_prev_frame (fi)) 11911 { 11912 if (!is_known_support_routine (fi)) 11913 { 11914 select_frame (fi); 11915 break; 11916 } 11917 } 11918 11919 } 11920 11921 /* Assuming that the inferior just triggered an unhandled exception 11922 catchpoint, return the address in inferior memory where the name 11923 of the exception is stored. 11924 11925 Return zero if the address could not be computed. */ 11926 11927 static CORE_ADDR 11928 ada_unhandled_exception_name_addr (void) 11929 { 11930 return parse_and_eval_address ("e.full_name"); 11931 } 11932 11933 /* Same as ada_unhandled_exception_name_addr, except that this function 11934 should be used when the inferior uses an older version of the runtime, 11935 where the exception name needs to be extracted from a specific frame 11936 several frames up in the callstack. */ 11937 11938 static CORE_ADDR 11939 ada_unhandled_exception_name_addr_from_raise (void) 11940 { 11941 int frame_level; 11942 frame_info_ptr fi; 11943 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ()); 11944 11945 /* To determine the name of this exception, we need to select 11946 the frame corresponding to RAISE_SYM_NAME. This frame is 11947 at least 3 levels up, so we simply skip the first 3 frames 11948 without checking the name of their associated function. */ 11949 fi = get_current_frame (); 11950 for (frame_level = 0; frame_level < 3; frame_level += 1) 11951 if (fi != NULL) 11952 fi = get_prev_frame (fi); 11953 11954 while (fi != NULL) 11955 { 11956 enum language func_lang; 11957 11958 gdb::unique_xmalloc_ptr<char> func_name 11959 = find_frame_funname (fi, &func_lang, NULL); 11960 if (func_name != NULL) 11961 { 11962 if (strcmp (func_name.get (), 11963 data->exception_info->catch_exception_sym) == 0) 11964 break; /* We found the frame we were looking for... */ 11965 } 11966 fi = get_prev_frame (fi); 11967 } 11968 11969 if (fi == NULL) 11970 return 0; 11971 11972 select_frame (fi); 11973 return parse_and_eval_address ("id.full_name"); 11974 } 11975 11976 /* Assuming the inferior just triggered an Ada exception catchpoint 11977 (of any type), return the address in inferior memory where the name 11978 of the exception is stored, if applicable. 11979 11980 Assumes the selected frame is the current frame. 11981 11982 Return zero if the address could not be computed, or if not relevant. */ 11983 11984 static CORE_ADDR 11985 ada_exception_name_addr_1 (enum ada_exception_catchpoint_kind ex) 11986 { 11987 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ()); 11988 11989 switch (ex) 11990 { 11991 case ada_catch_exception: 11992 return (parse_and_eval_address ("e.full_name")); 11993 break; 11994 11995 case ada_catch_exception_unhandled: 11996 return data->exception_info->unhandled_exception_name_addr (); 11997 break; 11998 11999 case ada_catch_handlers: 12000 return 0; /* The runtimes does not provide access to the exception 12001 name. */ 12002 break; 12003 12004 case ada_catch_assert: 12005 return 0; /* Exception name is not relevant in this case. */ 12006 break; 12007 12008 default: 12009 internal_error (_("unexpected catchpoint type")); 12010 break; 12011 } 12012 12013 return 0; /* Should never be reached. */ 12014 } 12015 12016 /* Assuming the inferior is stopped at an exception catchpoint, 12017 return the message which was associated to the exception, if 12018 available. Return NULL if the message could not be retrieved. 12019 12020 Note: The exception message can be associated to an exception 12021 either through the use of the Raise_Exception function, or 12022 more simply (Ada 2005 and later), via: 12023 12024 raise Exception_Name with "exception message"; 12025 12026 */ 12027 12028 static gdb::unique_xmalloc_ptr<char> 12029 ada_exception_message_1 (void) 12030 { 12031 struct value *e_msg_val; 12032 int e_msg_len; 12033 12034 /* For runtimes that support this feature, the exception message 12035 is passed as an unbounded string argument called "message". */ 12036 e_msg_val = parse_and_eval ("message"); 12037 if (e_msg_val == NULL) 12038 return NULL; /* Exception message not supported. */ 12039 12040 e_msg_val = ada_coerce_to_simple_array (e_msg_val); 12041 gdb_assert (e_msg_val != NULL); 12042 e_msg_len = value_type (e_msg_val)->length (); 12043 12044 /* If the message string is empty, then treat it as if there was 12045 no exception message. */ 12046 if (e_msg_len <= 0) 12047 return NULL; 12048 12049 gdb::unique_xmalloc_ptr<char> e_msg ((char *) xmalloc (e_msg_len + 1)); 12050 read_memory (value_address (e_msg_val), (gdb_byte *) e_msg.get (), 12051 e_msg_len); 12052 e_msg.get ()[e_msg_len] = '\0'; 12053 12054 return e_msg; 12055 } 12056 12057 /* Same as ada_exception_message_1, except that all exceptions are 12058 contained here (returning NULL instead). */ 12059 12060 static gdb::unique_xmalloc_ptr<char> 12061 ada_exception_message (void) 12062 { 12063 gdb::unique_xmalloc_ptr<char> e_msg; 12064 12065 try 12066 { 12067 e_msg = ada_exception_message_1 (); 12068 } 12069 catch (const gdb_exception_error &e) 12070 { 12071 e_msg.reset (nullptr); 12072 } 12073 12074 return e_msg; 12075 } 12076 12077 /* Same as ada_exception_name_addr_1, except that it intercepts and contains 12078 any error that ada_exception_name_addr_1 might cause to be thrown. 12079 When an error is intercepted, a warning with the error message is printed, 12080 and zero is returned. */ 12081 12082 static CORE_ADDR 12083 ada_exception_name_addr (enum ada_exception_catchpoint_kind ex) 12084 { 12085 CORE_ADDR result = 0; 12086 12087 try 12088 { 12089 result = ada_exception_name_addr_1 (ex); 12090 } 12091 12092 catch (const gdb_exception_error &e) 12093 { 12094 warning (_("failed to get exception name: %s"), e.what ()); 12095 return 0; 12096 } 12097 12098 return result; 12099 } 12100 12101 static std::string ada_exception_catchpoint_cond_string 12102 (const char *excep_string, 12103 enum ada_exception_catchpoint_kind ex); 12104 12105 /* Ada catchpoints. 12106 12107 In the case of catchpoints on Ada exceptions, the catchpoint will 12108 stop the target on every exception the program throws. When a user 12109 specifies the name of a specific exception, we translate this 12110 request into a condition expression (in text form), and then parse 12111 it into an expression stored in each of the catchpoint's locations. 12112 We then use this condition to check whether the exception that was 12113 raised is the one the user is interested in. If not, then the 12114 target is resumed again. We store the name of the requested 12115 exception, in order to be able to re-set the condition expression 12116 when symbols change. */ 12117 12118 /* An instance of this type is used to represent an Ada catchpoint. */ 12119 12120 struct ada_catchpoint : public code_breakpoint 12121 { 12122 ada_catchpoint (struct gdbarch *gdbarch_, 12123 enum ada_exception_catchpoint_kind kind, 12124 struct symtab_and_line sal, 12125 const char *addr_string_, 12126 bool tempflag, 12127 bool enabled, 12128 bool from_tty) 12129 : code_breakpoint (gdbarch_, bp_catchpoint), 12130 m_kind (kind) 12131 { 12132 add_location (sal); 12133 12134 /* Unlike most code_breakpoint types, Ada catchpoints are 12135 pspace-specific. */ 12136 gdb_assert (sal.pspace != nullptr); 12137 this->pspace = sal.pspace; 12138 12139 if (from_tty) 12140 { 12141 struct gdbarch *loc_gdbarch = get_sal_arch (sal); 12142 if (!loc_gdbarch) 12143 loc_gdbarch = gdbarch; 12144 12145 describe_other_breakpoints (loc_gdbarch, 12146 sal.pspace, sal.pc, sal.section, -1); 12147 /* FIXME: brobecker/2006-12-28: Actually, re-implement a special 12148 version for exception catchpoints, because two catchpoints 12149 used for different exception names will use the same address. 12150 In this case, a "breakpoint ... also set at..." warning is 12151 unproductive. Besides, the warning phrasing is also a bit 12152 inappropriate, we should use the word catchpoint, and tell 12153 the user what type of catchpoint it is. The above is good 12154 enough for now, though. */ 12155 } 12156 12157 enable_state = enabled ? bp_enabled : bp_disabled; 12158 disposition = tempflag ? disp_del : disp_donttouch; 12159 locspec = string_to_location_spec (&addr_string_, 12160 language_def (language_ada)); 12161 language = language_ada; 12162 } 12163 12164 struct bp_location *allocate_location () override; 12165 void re_set () override; 12166 void check_status (struct bpstat *bs) override; 12167 enum print_stop_action print_it (const bpstat *bs) const override; 12168 bool print_one (bp_location **) const override; 12169 void print_mention () const override; 12170 void print_recreate (struct ui_file *fp) const override; 12171 12172 /* The name of the specific exception the user specified. */ 12173 std::string excep_string; 12174 12175 /* What kind of catchpoint this is. */ 12176 enum ada_exception_catchpoint_kind m_kind; 12177 }; 12178 12179 /* An instance of this type is used to represent an Ada catchpoint 12180 breakpoint location. */ 12181 12182 class ada_catchpoint_location : public bp_location 12183 { 12184 public: 12185 explicit ada_catchpoint_location (ada_catchpoint *owner) 12186 : bp_location (owner, bp_loc_software_breakpoint) 12187 {} 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 expression_up excep_cond_expr; 12193 }; 12194 12195 /* Parse the exception condition string in the context of each of the 12196 catchpoint's locations, and store them for later evaluation. */ 12197 12198 static void 12199 create_excep_cond_exprs (struct ada_catchpoint *c, 12200 enum ada_exception_catchpoint_kind ex) 12201 { 12202 /* Nothing to do if there's no specific exception to catch. */ 12203 if (c->excep_string.empty ()) 12204 return; 12205 12206 /* Same if there are no locations... */ 12207 if (c->loc == NULL) 12208 return; 12209 12210 /* Compute the condition expression in text form, from the specific 12211 expection we want to catch. */ 12212 std::string cond_string 12213 = ada_exception_catchpoint_cond_string (c->excep_string.c_str (), ex); 12214 12215 /* Iterate over all the catchpoint's locations, and parse an 12216 expression for each. */ 12217 for (bp_location *bl : c->locations ()) 12218 { 12219 struct ada_catchpoint_location *ada_loc 12220 = (struct ada_catchpoint_location *) bl; 12221 expression_up exp; 12222 12223 if (!bl->shlib_disabled) 12224 { 12225 const char *s; 12226 12227 s = cond_string.c_str (); 12228 try 12229 { 12230 exp = parse_exp_1 (&s, bl->address, 12231 block_for_pc (bl->address), 12232 0); 12233 } 12234 catch (const gdb_exception_error &e) 12235 { 12236 warning (_("failed to reevaluate internal exception condition " 12237 "for catchpoint %d: %s"), 12238 c->number, e.what ()); 12239 } 12240 } 12241 12242 ada_loc->excep_cond_expr = std::move (exp); 12243 } 12244 } 12245 12246 /* Implement the ALLOCATE_LOCATION method in the structure for all 12247 exception catchpoint kinds. */ 12248 12249 struct bp_location * 12250 ada_catchpoint::allocate_location () 12251 { 12252 return new ada_catchpoint_location (this); 12253 } 12254 12255 /* Implement the RE_SET method in the structure for all exception 12256 catchpoint kinds. */ 12257 12258 void 12259 ada_catchpoint::re_set () 12260 { 12261 /* Call the base class's method. This updates the catchpoint's 12262 locations. */ 12263 this->code_breakpoint::re_set (); 12264 12265 /* Reparse the exception conditional expressions. One for each 12266 location. */ 12267 create_excep_cond_exprs (this, m_kind); 12268 } 12269 12270 /* Returns true if we should stop for this breakpoint hit. If the 12271 user specified a specific exception, we only want to cause a stop 12272 if the program thrown that exception. */ 12273 12274 static bool 12275 should_stop_exception (const struct bp_location *bl) 12276 { 12277 struct ada_catchpoint *c = (struct ada_catchpoint *) bl->owner; 12278 const struct ada_catchpoint_location *ada_loc 12279 = (const struct ada_catchpoint_location *) bl; 12280 bool stop; 12281 12282 struct internalvar *var = lookup_internalvar ("_ada_exception"); 12283 if (c->m_kind == ada_catch_assert) 12284 clear_internalvar (var); 12285 else 12286 { 12287 try 12288 { 12289 const char *expr; 12290 12291 if (c->m_kind == ada_catch_handlers) 12292 expr = ("GNAT_GCC_exception_Access(gcc_exception)" 12293 ".all.occurrence.id"); 12294 else 12295 expr = "e"; 12296 12297 struct value *exc = parse_and_eval (expr); 12298 set_internalvar (var, exc); 12299 } 12300 catch (const gdb_exception_error &ex) 12301 { 12302 clear_internalvar (var); 12303 } 12304 } 12305 12306 /* With no specific exception, should always stop. */ 12307 if (c->excep_string.empty ()) 12308 return true; 12309 12310 if (ada_loc->excep_cond_expr == NULL) 12311 { 12312 /* We will have a NULL expression if back when we were creating 12313 the expressions, this location's had failed to parse. */ 12314 return true; 12315 } 12316 12317 stop = true; 12318 try 12319 { 12320 scoped_value_mark mark; 12321 stop = value_true (evaluate_expression (ada_loc->excep_cond_expr.get ())); 12322 } 12323 catch (const gdb_exception &ex) 12324 { 12325 exception_fprintf (gdb_stderr, ex, 12326 _("Error in testing exception condition:\n")); 12327 } 12328 12329 return stop; 12330 } 12331 12332 /* Implement the CHECK_STATUS method in the structure for all 12333 exception catchpoint kinds. */ 12334 12335 void 12336 ada_catchpoint::check_status (bpstat *bs) 12337 { 12338 bs->stop = should_stop_exception (bs->bp_location_at.get ()); 12339 } 12340 12341 /* Implement the PRINT_IT method in the structure for all exception 12342 catchpoint kinds. */ 12343 12344 enum print_stop_action 12345 ada_catchpoint::print_it (const bpstat *bs) const 12346 { 12347 struct ui_out *uiout = current_uiout; 12348 12349 annotate_catchpoint (number); 12350 12351 if (uiout->is_mi_like_p ()) 12352 { 12353 uiout->field_string ("reason", 12354 async_reason_lookup (EXEC_ASYNC_BREAKPOINT_HIT)); 12355 uiout->field_string ("disp", bpdisp_text (disposition)); 12356 } 12357 12358 uiout->text (disposition == disp_del 12359 ? "\nTemporary catchpoint " : "\nCatchpoint "); 12360 print_num_locno (bs, uiout); 12361 uiout->text (", "); 12362 12363 /* ada_exception_name_addr relies on the selected frame being the 12364 current frame. Need to do this here because this function may be 12365 called more than once when printing a stop, and below, we'll 12366 select the first frame past the Ada run-time (see 12367 ada_find_printable_frame). */ 12368 select_frame (get_current_frame ()); 12369 12370 switch (m_kind) 12371 { 12372 case ada_catch_exception: 12373 case ada_catch_exception_unhandled: 12374 case ada_catch_handlers: 12375 { 12376 const CORE_ADDR addr = ada_exception_name_addr (m_kind); 12377 char exception_name[256]; 12378 12379 if (addr != 0) 12380 { 12381 read_memory (addr, (gdb_byte *) exception_name, 12382 sizeof (exception_name) - 1); 12383 exception_name [sizeof (exception_name) - 1] = '\0'; 12384 } 12385 else 12386 { 12387 /* For some reason, we were unable to read the exception 12388 name. This could happen if the Runtime was compiled 12389 without debugging info, for instance. In that case, 12390 just replace the exception name by the generic string 12391 "exception" - it will read as "an exception" in the 12392 notification we are about to print. */ 12393 memcpy (exception_name, "exception", sizeof ("exception")); 12394 } 12395 /* In the case of unhandled exception breakpoints, we print 12396 the exception name as "unhandled EXCEPTION_NAME", to make 12397 it clearer to the user which kind of catchpoint just got 12398 hit. We used ui_out_text to make sure that this extra 12399 info does not pollute the exception name in the MI case. */ 12400 if (m_kind == ada_catch_exception_unhandled) 12401 uiout->text ("unhandled "); 12402 uiout->field_string ("exception-name", exception_name); 12403 } 12404 break; 12405 case ada_catch_assert: 12406 /* In this case, the name of the exception is not really 12407 important. Just print "failed assertion" to make it clearer 12408 that his program just hit an assertion-failure catchpoint. 12409 We used ui_out_text because this info does not belong in 12410 the MI output. */ 12411 uiout->text ("failed assertion"); 12412 break; 12413 } 12414 12415 gdb::unique_xmalloc_ptr<char> exception_message = ada_exception_message (); 12416 if (exception_message != NULL) 12417 { 12418 uiout->text (" ("); 12419 uiout->field_string ("exception-message", exception_message.get ()); 12420 uiout->text (")"); 12421 } 12422 12423 uiout->text (" at "); 12424 ada_find_printable_frame (get_current_frame ()); 12425 12426 return PRINT_SRC_AND_LOC; 12427 } 12428 12429 /* Implement the PRINT_ONE method in the structure for all exception 12430 catchpoint kinds. */ 12431 12432 bool 12433 ada_catchpoint::print_one (bp_location **last_loc) const 12434 { 12435 struct ui_out *uiout = current_uiout; 12436 struct value_print_options opts; 12437 12438 get_user_print_options (&opts); 12439 12440 if (opts.addressprint) 12441 uiout->field_skip ("addr"); 12442 12443 annotate_field (5); 12444 switch (m_kind) 12445 { 12446 case ada_catch_exception: 12447 if (!excep_string.empty ()) 12448 { 12449 std::string msg = string_printf (_("`%s' Ada exception"), 12450 excep_string.c_str ()); 12451 12452 uiout->field_string ("what", msg); 12453 } 12454 else 12455 uiout->field_string ("what", "all Ada exceptions"); 12456 12457 break; 12458 12459 case ada_catch_exception_unhandled: 12460 uiout->field_string ("what", "unhandled Ada exceptions"); 12461 break; 12462 12463 case ada_catch_handlers: 12464 if (!excep_string.empty ()) 12465 { 12466 uiout->field_fmt ("what", 12467 _("`%s' Ada exception handlers"), 12468 excep_string.c_str ()); 12469 } 12470 else 12471 uiout->field_string ("what", "all Ada exceptions handlers"); 12472 break; 12473 12474 case ada_catch_assert: 12475 uiout->field_string ("what", "failed Ada assertions"); 12476 break; 12477 12478 default: 12479 internal_error (_("unexpected catchpoint type")); 12480 break; 12481 } 12482 12483 return true; 12484 } 12485 12486 /* Implement the PRINT_MENTION method in the breakpoint_ops structure 12487 for all exception catchpoint kinds. */ 12488 12489 void 12490 ada_catchpoint::print_mention () const 12491 { 12492 struct ui_out *uiout = current_uiout; 12493 12494 uiout->text (disposition == disp_del ? _("Temporary catchpoint ") 12495 : _("Catchpoint ")); 12496 uiout->field_signed ("bkptno", number); 12497 uiout->text (": "); 12498 12499 switch (m_kind) 12500 { 12501 case ada_catch_exception: 12502 if (!excep_string.empty ()) 12503 { 12504 std::string info = string_printf (_("`%s' Ada exception"), 12505 excep_string.c_str ()); 12506 uiout->text (info); 12507 } 12508 else 12509 uiout->text (_("all Ada exceptions")); 12510 break; 12511 12512 case ada_catch_exception_unhandled: 12513 uiout->text (_("unhandled Ada exceptions")); 12514 break; 12515 12516 case ada_catch_handlers: 12517 if (!excep_string.empty ()) 12518 { 12519 std::string info 12520 = string_printf (_("`%s' Ada exception handlers"), 12521 excep_string.c_str ()); 12522 uiout->text (info); 12523 } 12524 else 12525 uiout->text (_("all Ada exceptions handlers")); 12526 break; 12527 12528 case ada_catch_assert: 12529 uiout->text (_("failed Ada assertions")); 12530 break; 12531 12532 default: 12533 internal_error (_("unexpected catchpoint type")); 12534 break; 12535 } 12536 } 12537 12538 /* Implement the PRINT_RECREATE method in the structure for all 12539 exception catchpoint kinds. */ 12540 12541 void 12542 ada_catchpoint::print_recreate (struct ui_file *fp) const 12543 { 12544 switch (m_kind) 12545 { 12546 case ada_catch_exception: 12547 gdb_printf (fp, "catch exception"); 12548 if (!excep_string.empty ()) 12549 gdb_printf (fp, " %s", excep_string.c_str ()); 12550 break; 12551 12552 case ada_catch_exception_unhandled: 12553 gdb_printf (fp, "catch exception unhandled"); 12554 break; 12555 12556 case ada_catch_handlers: 12557 gdb_printf (fp, "catch handlers"); 12558 break; 12559 12560 case ada_catch_assert: 12561 gdb_printf (fp, "catch assert"); 12562 break; 12563 12564 default: 12565 internal_error (_("unexpected catchpoint type")); 12566 } 12567 print_recreate_thread (fp); 12568 } 12569 12570 /* See ada-lang.h. */ 12571 12572 bool 12573 is_ada_exception_catchpoint (breakpoint *bp) 12574 { 12575 return dynamic_cast<ada_catchpoint *> (bp) != nullptr; 12576 } 12577 12578 /* Split the arguments specified in a "catch exception" command. 12579 Set EX to the appropriate catchpoint type. 12580 Set EXCEP_STRING to the name of the specific exception if 12581 specified by the user. 12582 IS_CATCH_HANDLERS_CMD: True if the arguments are for a 12583 "catch handlers" command. False otherwise. 12584 If a condition is found at the end of the arguments, the condition 12585 expression is stored in COND_STRING (memory must be deallocated 12586 after use). Otherwise COND_STRING is set to NULL. */ 12587 12588 static void 12589 catch_ada_exception_command_split (const char *args, 12590 bool is_catch_handlers_cmd, 12591 enum ada_exception_catchpoint_kind *ex, 12592 std::string *excep_string, 12593 std::string *cond_string) 12594 { 12595 std::string exception_name; 12596 12597 exception_name = extract_arg (&args); 12598 if (exception_name == "if") 12599 { 12600 /* This is not an exception name; this is the start of a condition 12601 expression for a catchpoint on all exceptions. So, "un-get" 12602 this token, and set exception_name to NULL. */ 12603 exception_name.clear (); 12604 args -= 2; 12605 } 12606 12607 /* Check to see if we have a condition. */ 12608 12609 args = skip_spaces (args); 12610 if (startswith (args, "if") 12611 && (isspace (args[2]) || args[2] == '\0')) 12612 { 12613 args += 2; 12614 args = skip_spaces (args); 12615 12616 if (args[0] == '\0') 12617 error (_("Condition missing after `if' keyword")); 12618 *cond_string = args; 12619 12620 args += strlen (args); 12621 } 12622 12623 /* Check that we do not have any more arguments. Anything else 12624 is unexpected. */ 12625 12626 if (args[0] != '\0') 12627 error (_("Junk at end of expression")); 12628 12629 if (is_catch_handlers_cmd) 12630 { 12631 /* Catch handling of exceptions. */ 12632 *ex = ada_catch_handlers; 12633 *excep_string = exception_name; 12634 } 12635 else if (exception_name.empty ()) 12636 { 12637 /* Catch all exceptions. */ 12638 *ex = ada_catch_exception; 12639 excep_string->clear (); 12640 } 12641 else if (exception_name == "unhandled") 12642 { 12643 /* Catch unhandled exceptions. */ 12644 *ex = ada_catch_exception_unhandled; 12645 excep_string->clear (); 12646 } 12647 else 12648 { 12649 /* Catch a specific exception. */ 12650 *ex = ada_catch_exception; 12651 *excep_string = exception_name; 12652 } 12653 } 12654 12655 /* Return the name of the symbol on which we should break in order to 12656 implement a catchpoint of the EX kind. */ 12657 12658 static const char * 12659 ada_exception_sym_name (enum ada_exception_catchpoint_kind ex) 12660 { 12661 struct ada_inferior_data *data = get_ada_inferior_data (current_inferior ()); 12662 12663 gdb_assert (data->exception_info != NULL); 12664 12665 switch (ex) 12666 { 12667 case ada_catch_exception: 12668 return (data->exception_info->catch_exception_sym); 12669 break; 12670 case ada_catch_exception_unhandled: 12671 return (data->exception_info->catch_exception_unhandled_sym); 12672 break; 12673 case ada_catch_assert: 12674 return (data->exception_info->catch_assert_sym); 12675 break; 12676 case ada_catch_handlers: 12677 return (data->exception_info->catch_handlers_sym); 12678 break; 12679 default: 12680 internal_error (_("unexpected catchpoint kind (%d)"), ex); 12681 } 12682 } 12683 12684 /* Return the condition that will be used to match the current exception 12685 being raised with the exception that the user wants to catch. This 12686 assumes that this condition is used when the inferior just triggered 12687 an exception catchpoint. 12688 EX: the type of catchpoints used for catching Ada exceptions. */ 12689 12690 static std::string 12691 ada_exception_catchpoint_cond_string (const char *excep_string, 12692 enum ada_exception_catchpoint_kind ex) 12693 { 12694 bool is_standard_exc = false; 12695 std::string result; 12696 12697 if (ex == ada_catch_handlers) 12698 { 12699 /* For exception handlers catchpoints, the condition string does 12700 not use the same parameter as for the other exceptions. */ 12701 result = ("long_integer (GNAT_GCC_exception_Access" 12702 "(gcc_exception).all.occurrence.id)"); 12703 } 12704 else 12705 result = "long_integer (e)"; 12706 12707 /* The standard exceptions are a special case. They are defined in 12708 runtime units that have been compiled without debugging info; if 12709 EXCEP_STRING is the not-fully-qualified name of a standard 12710 exception (e.g. "constraint_error") then, during the evaluation 12711 of the condition expression, the symbol lookup on this name would 12712 *not* return this standard exception. The catchpoint condition 12713 may then be set only on user-defined exceptions which have the 12714 same not-fully-qualified name (e.g. my_package.constraint_error). 12715 12716 To avoid this unexcepted behavior, these standard exceptions are 12717 systematically prefixed by "standard". This means that "catch 12718 exception constraint_error" is rewritten into "catch exception 12719 standard.constraint_error". 12720 12721 If an exception named constraint_error is defined in another package of 12722 the inferior program, then the only way to specify this exception as a 12723 breakpoint condition is to use its fully-qualified named: 12724 e.g. my_package.constraint_error. */ 12725 12726 for (const char *name : standard_exc) 12727 { 12728 if (strcmp (name, excep_string) == 0) 12729 { 12730 is_standard_exc = true; 12731 break; 12732 } 12733 } 12734 12735 result += " = "; 12736 12737 if (is_standard_exc) 12738 string_appendf (result, "long_integer (&standard.%s)", excep_string); 12739 else 12740 string_appendf (result, "long_integer (&%s)", excep_string); 12741 12742 return result; 12743 } 12744 12745 /* Return the symtab_and_line that should be used to insert an exception 12746 catchpoint of the TYPE kind. 12747 12748 ADDR_STRING returns the name of the function where the real 12749 breakpoint that implements the catchpoints is set, depending on the 12750 type of catchpoint we need to create. */ 12751 12752 static struct symtab_and_line 12753 ada_exception_sal (enum ada_exception_catchpoint_kind ex, 12754 std::string *addr_string) 12755 { 12756 const char *sym_name; 12757 struct symbol *sym; 12758 12759 /* First, find out which exception support info to use. */ 12760 ada_exception_support_info_sniffer (); 12761 12762 /* Then lookup the function on which we will break in order to catch 12763 the Ada exceptions requested by the user. */ 12764 sym_name = ada_exception_sym_name (ex); 12765 sym = standard_lookup (sym_name, NULL, VAR_DOMAIN); 12766 12767 if (sym == NULL) 12768 error (_("Catchpoint symbol not found: %s"), sym_name); 12769 12770 if (sym->aclass () != LOC_BLOCK) 12771 error (_("Unable to insert catchpoint. %s is not a function."), sym_name); 12772 12773 /* Set ADDR_STRING. */ 12774 *addr_string = sym_name; 12775 12776 return find_function_start_sal (sym, 1); 12777 } 12778 12779 /* Create an Ada exception catchpoint. 12780 12781 EX_KIND is the kind of exception catchpoint to be created. 12782 12783 If EXCEPT_STRING is empty, this catchpoint is expected to trigger 12784 for all exceptions. Otherwise, EXCEPT_STRING indicates the name 12785 of the exception to which this catchpoint applies. 12786 12787 COND_STRING, if not empty, is the catchpoint condition. 12788 12789 TEMPFLAG, if nonzero, means that the underlying breakpoint 12790 should be temporary. 12791 12792 FROM_TTY is the usual argument passed to all commands implementations. */ 12793 12794 void 12795 create_ada_exception_catchpoint (struct gdbarch *gdbarch, 12796 enum ada_exception_catchpoint_kind ex_kind, 12797 const std::string &excep_string, 12798 const std::string &cond_string, 12799 int tempflag, 12800 int disabled, 12801 int from_tty) 12802 { 12803 std::string addr_string; 12804 struct symtab_and_line sal = ada_exception_sal (ex_kind, &addr_string); 12805 12806 std::unique_ptr<ada_catchpoint> c 12807 (new ada_catchpoint (gdbarch, ex_kind, sal, addr_string.c_str (), 12808 tempflag, disabled, from_tty)); 12809 c->excep_string = excep_string; 12810 create_excep_cond_exprs (c.get (), ex_kind); 12811 if (!cond_string.empty ()) 12812 set_breakpoint_condition (c.get (), cond_string.c_str (), from_tty, false); 12813 install_breakpoint (0, std::move (c), 1); 12814 } 12815 12816 /* Implement the "catch exception" command. */ 12817 12818 static void 12819 catch_ada_exception_command (const char *arg_entry, int from_tty, 12820 struct cmd_list_element *command) 12821 { 12822 const char *arg = arg_entry; 12823 struct gdbarch *gdbarch = get_current_arch (); 12824 int tempflag; 12825 enum ada_exception_catchpoint_kind ex_kind; 12826 std::string excep_string; 12827 std::string cond_string; 12828 12829 tempflag = command->context () == CATCH_TEMPORARY; 12830 12831 if (!arg) 12832 arg = ""; 12833 catch_ada_exception_command_split (arg, false, &ex_kind, &excep_string, 12834 &cond_string); 12835 create_ada_exception_catchpoint (gdbarch, ex_kind, 12836 excep_string, cond_string, 12837 tempflag, 1 /* enabled */, 12838 from_tty); 12839 } 12840 12841 /* Implement the "catch handlers" command. */ 12842 12843 static void 12844 catch_ada_handlers_command (const char *arg_entry, int from_tty, 12845 struct cmd_list_element *command) 12846 { 12847 const char *arg = arg_entry; 12848 struct gdbarch *gdbarch = get_current_arch (); 12849 int tempflag; 12850 enum ada_exception_catchpoint_kind ex_kind; 12851 std::string excep_string; 12852 std::string cond_string; 12853 12854 tempflag = command->context () == CATCH_TEMPORARY; 12855 12856 if (!arg) 12857 arg = ""; 12858 catch_ada_exception_command_split (arg, true, &ex_kind, &excep_string, 12859 &cond_string); 12860 create_ada_exception_catchpoint (gdbarch, ex_kind, 12861 excep_string, cond_string, 12862 tempflag, 1 /* enabled */, 12863 from_tty); 12864 } 12865 12866 /* Completion function for the Ada "catch" commands. */ 12867 12868 static void 12869 catch_ada_completer (struct cmd_list_element *cmd, completion_tracker &tracker, 12870 const char *text, const char *word) 12871 { 12872 std::vector<ada_exc_info> exceptions = ada_exceptions_list (NULL); 12873 12874 for (const ada_exc_info &info : exceptions) 12875 { 12876 if (startswith (info.name, word)) 12877 tracker.add_completion (make_unique_xstrdup (info.name)); 12878 } 12879 } 12880 12881 /* Split the arguments specified in a "catch assert" command. 12882 12883 ARGS contains the command's arguments (or the empty string if 12884 no arguments were passed). 12885 12886 If ARGS contains a condition, set COND_STRING to that condition 12887 (the memory needs to be deallocated after use). */ 12888 12889 static void 12890 catch_ada_assert_command_split (const char *args, std::string &cond_string) 12891 { 12892 args = skip_spaces (args); 12893 12894 /* Check whether a condition was provided. */ 12895 if (startswith (args, "if") 12896 && (isspace (args[2]) || args[2] == '\0')) 12897 { 12898 args += 2; 12899 args = skip_spaces (args); 12900 if (args[0] == '\0') 12901 error (_("condition missing after `if' keyword")); 12902 cond_string.assign (args); 12903 } 12904 12905 /* Otherwise, there should be no other argument at the end of 12906 the command. */ 12907 else if (args[0] != '\0') 12908 error (_("Junk at end of arguments.")); 12909 } 12910 12911 /* Implement the "catch assert" command. */ 12912 12913 static void 12914 catch_assert_command (const char *arg_entry, int from_tty, 12915 struct cmd_list_element *command) 12916 { 12917 const char *arg = arg_entry; 12918 struct gdbarch *gdbarch = get_current_arch (); 12919 int tempflag; 12920 std::string cond_string; 12921 12922 tempflag = command->context () == CATCH_TEMPORARY; 12923 12924 if (!arg) 12925 arg = ""; 12926 catch_ada_assert_command_split (arg, cond_string); 12927 create_ada_exception_catchpoint (gdbarch, ada_catch_assert, 12928 "", cond_string, 12929 tempflag, 1 /* enabled */, 12930 from_tty); 12931 } 12932 12933 /* Return non-zero if the symbol SYM is an Ada exception object. */ 12934 12935 static int 12936 ada_is_exception_sym (struct symbol *sym) 12937 { 12938 const char *type_name = sym->type ()->name (); 12939 12940 return (sym->aclass () != LOC_TYPEDEF 12941 && sym->aclass () != LOC_BLOCK 12942 && sym->aclass () != LOC_CONST 12943 && sym->aclass () != LOC_UNRESOLVED 12944 && type_name != NULL && strcmp (type_name, "exception") == 0); 12945 } 12946 12947 /* Given a global symbol SYM, return non-zero iff SYM is a non-standard 12948 Ada exception object. This matches all exceptions except the ones 12949 defined by the Ada language. */ 12950 12951 static int 12952 ada_is_non_standard_exception_sym (struct symbol *sym) 12953 { 12954 if (!ada_is_exception_sym (sym)) 12955 return 0; 12956 12957 for (const char *name : standard_exc) 12958 if (strcmp (sym->linkage_name (), name) == 0) 12959 return 0; /* A standard exception. */ 12960 12961 /* Numeric_Error is also a standard exception, so exclude it. 12962 See the STANDARD_EXC description for more details as to why 12963 this exception is not listed in that array. */ 12964 if (strcmp (sym->linkage_name (), "numeric_error") == 0) 12965 return 0; 12966 12967 return 1; 12968 } 12969 12970 /* A helper function for std::sort, comparing two struct ada_exc_info 12971 objects. 12972 12973 The comparison is determined first by exception name, and then 12974 by exception address. */ 12975 12976 bool 12977 ada_exc_info::operator< (const ada_exc_info &other) const 12978 { 12979 int result; 12980 12981 result = strcmp (name, other.name); 12982 if (result < 0) 12983 return true; 12984 if (result == 0 && addr < other.addr) 12985 return true; 12986 return false; 12987 } 12988 12989 bool 12990 ada_exc_info::operator== (const ada_exc_info &other) const 12991 { 12992 return addr == other.addr && strcmp (name, other.name) == 0; 12993 } 12994 12995 /* Sort EXCEPTIONS using compare_ada_exception_info as the comparison 12996 routine, but keeping the first SKIP elements untouched. 12997 12998 All duplicates are also removed. */ 12999 13000 static void 13001 sort_remove_dups_ada_exceptions_list (std::vector<ada_exc_info> *exceptions, 13002 int skip) 13003 { 13004 std::sort (exceptions->begin () + skip, exceptions->end ()); 13005 exceptions->erase (std::unique (exceptions->begin () + skip, exceptions->end ()), 13006 exceptions->end ()); 13007 } 13008 13009 /* Add all exceptions defined by the Ada standard whose name match 13010 a regular expression. 13011 13012 If PREG is not NULL, then this regexp_t object is used to 13013 perform the symbol name matching. Otherwise, no name-based 13014 filtering is performed. 13015 13016 EXCEPTIONS is a vector of exceptions to which matching exceptions 13017 gets pushed. */ 13018 13019 static void 13020 ada_add_standard_exceptions (compiled_regex *preg, 13021 std::vector<ada_exc_info> *exceptions) 13022 { 13023 for (const char *name : standard_exc) 13024 { 13025 if (preg == NULL || preg->exec (name, 0, NULL, 0) == 0) 13026 { 13027 symbol_name_match_type match_type = name_match_type_from_name (name); 13028 lookup_name_info lookup_name (name, match_type); 13029 13030 symbol_name_matcher_ftype *match_name 13031 = ada_get_symbol_name_matcher (lookup_name); 13032 13033 /* Iterate over all objfiles irrespective of scope or linker 13034 namespaces so we get all exceptions anywhere in the 13035 progspace. */ 13036 for (objfile *objfile : current_program_space->objfiles ()) 13037 { 13038 for (minimal_symbol *msymbol : objfile->msymbols ()) 13039 { 13040 if (match_name (msymbol->linkage_name (), lookup_name, 13041 nullptr) 13042 && msymbol->type () != mst_solib_trampoline) 13043 { 13044 ada_exc_info info 13045 = {name, msymbol->value_address (objfile)}; 13046 13047 exceptions->push_back (info); 13048 } 13049 } 13050 } 13051 } 13052 } 13053 } 13054 13055 /* Add all Ada exceptions defined locally and accessible from the given 13056 FRAME. 13057 13058 If PREG is not NULL, then this regexp_t object is used to 13059 perform the symbol name matching. Otherwise, no name-based 13060 filtering is performed. 13061 13062 EXCEPTIONS is a vector of exceptions to which matching exceptions 13063 gets pushed. */ 13064 13065 static void 13066 ada_add_exceptions_from_frame (compiled_regex *preg, 13067 frame_info_ptr frame, 13068 std::vector<ada_exc_info> *exceptions) 13069 { 13070 const struct block *block = get_frame_block (frame, 0); 13071 13072 while (block != 0) 13073 { 13074 struct block_iterator iter; 13075 struct symbol *sym; 13076 13077 ALL_BLOCK_SYMBOLS (block, iter, sym) 13078 { 13079 switch (sym->aclass ()) 13080 { 13081 case LOC_TYPEDEF: 13082 case LOC_BLOCK: 13083 case LOC_CONST: 13084 break; 13085 default: 13086 if (ada_is_exception_sym (sym)) 13087 { 13088 struct ada_exc_info info = {sym->print_name (), 13089 sym->value_address ()}; 13090 13091 exceptions->push_back (info); 13092 } 13093 } 13094 } 13095 if (block->function () != NULL) 13096 break; 13097 block = block->superblock (); 13098 } 13099 } 13100 13101 /* Return true if NAME matches PREG or if PREG is NULL. */ 13102 13103 static bool 13104 name_matches_regex (const char *name, compiled_regex *preg) 13105 { 13106 return (preg == NULL 13107 || preg->exec (ada_decode (name).c_str (), 0, NULL, 0) == 0); 13108 } 13109 13110 /* Add all exceptions defined globally whose name name match 13111 a regular expression, excluding standard exceptions. 13112 13113 The reason we exclude standard exceptions is that they need 13114 to be handled separately: Standard exceptions are defined inside 13115 a runtime unit which is normally not compiled with debugging info, 13116 and thus usually do not show up in our symbol search. However, 13117 if the unit was in fact built with debugging info, we need to 13118 exclude them because they would duplicate the entry we found 13119 during the special loop that specifically searches for those 13120 standard exceptions. 13121 13122 If PREG is not NULL, then this regexp_t object is used to 13123 perform the symbol name matching. Otherwise, no name-based 13124 filtering is performed. 13125 13126 EXCEPTIONS is a vector of exceptions to which matching exceptions 13127 gets pushed. */ 13128 13129 static void 13130 ada_add_global_exceptions (compiled_regex *preg, 13131 std::vector<ada_exc_info> *exceptions) 13132 { 13133 /* In Ada, the symbol "search name" is a linkage name, whereas the 13134 regular expression used to do the matching refers to the natural 13135 name. So match against the decoded name. */ 13136 expand_symtabs_matching (NULL, 13137 lookup_name_info::match_any (), 13138 [&] (const char *search_name) 13139 { 13140 std::string decoded = ada_decode (search_name); 13141 return name_matches_regex (decoded.c_str (), preg); 13142 }, 13143 NULL, 13144 SEARCH_GLOBAL_BLOCK | SEARCH_STATIC_BLOCK, 13145 VARIABLES_DOMAIN); 13146 13147 /* Iterate over all objfiles irrespective of scope or linker namespaces 13148 so we get all exceptions anywhere in the progspace. */ 13149 for (objfile *objfile : current_program_space->objfiles ()) 13150 { 13151 for (compunit_symtab *s : objfile->compunits ()) 13152 { 13153 const struct blockvector *bv = s->blockvector (); 13154 int i; 13155 13156 for (i = GLOBAL_BLOCK; i <= STATIC_BLOCK; i++) 13157 { 13158 const struct block *b = bv->block (i); 13159 struct block_iterator iter; 13160 struct symbol *sym; 13161 13162 ALL_BLOCK_SYMBOLS (b, iter, sym) 13163 if (ada_is_non_standard_exception_sym (sym) 13164 && name_matches_regex (sym->natural_name (), preg)) 13165 { 13166 struct ada_exc_info info 13167 = {sym->print_name (), sym->value_address ()}; 13168 13169 exceptions->push_back (info); 13170 } 13171 } 13172 } 13173 } 13174 } 13175 13176 /* Implements ada_exceptions_list with the regular expression passed 13177 as a regex_t, rather than a string. 13178 13179 If not NULL, PREG is used to filter out exceptions whose names 13180 do not match. Otherwise, all exceptions are listed. */ 13181 13182 static std::vector<ada_exc_info> 13183 ada_exceptions_list_1 (compiled_regex *preg) 13184 { 13185 std::vector<ada_exc_info> result; 13186 int prev_len; 13187 13188 /* First, list the known standard exceptions. These exceptions 13189 need to be handled separately, as they are usually defined in 13190 runtime units that have been compiled without debugging info. */ 13191 13192 ada_add_standard_exceptions (preg, &result); 13193 13194 /* Next, find all exceptions whose scope is local and accessible 13195 from the currently selected frame. */ 13196 13197 if (has_stack_frames ()) 13198 { 13199 prev_len = result.size (); 13200 ada_add_exceptions_from_frame (preg, get_selected_frame (NULL), 13201 &result); 13202 if (result.size () > prev_len) 13203 sort_remove_dups_ada_exceptions_list (&result, prev_len); 13204 } 13205 13206 /* Add all exceptions whose scope is global. */ 13207 13208 prev_len = result.size (); 13209 ada_add_global_exceptions (preg, &result); 13210 if (result.size () > prev_len) 13211 sort_remove_dups_ada_exceptions_list (&result, prev_len); 13212 13213 return result; 13214 } 13215 13216 /* Return a vector of ada_exc_info. 13217 13218 If REGEXP is NULL, all exceptions are included in the result. 13219 Otherwise, it should contain a valid regular expression, 13220 and only the exceptions whose names match that regular expression 13221 are included in the result. 13222 13223 The exceptions are sorted in the following order: 13224 - Standard exceptions (defined by the Ada language), in 13225 alphabetical order; 13226 - Exceptions only visible from the current frame, in 13227 alphabetical order; 13228 - Exceptions whose scope is global, in alphabetical order. */ 13229 13230 std::vector<ada_exc_info> 13231 ada_exceptions_list (const char *regexp) 13232 { 13233 if (regexp == NULL) 13234 return ada_exceptions_list_1 (NULL); 13235 13236 compiled_regex reg (regexp, REG_NOSUB, _("invalid regular expression")); 13237 return ada_exceptions_list_1 (®); 13238 } 13239 13240 /* Implement the "info exceptions" command. */ 13241 13242 static void 13243 info_exceptions_command (const char *regexp, int from_tty) 13244 { 13245 struct gdbarch *gdbarch = get_current_arch (); 13246 13247 std::vector<ada_exc_info> exceptions = ada_exceptions_list (regexp); 13248 13249 if (regexp != NULL) 13250 gdb_printf 13251 (_("All Ada exceptions matching regular expression \"%s\":\n"), regexp); 13252 else 13253 gdb_printf (_("All defined Ada exceptions:\n")); 13254 13255 for (const ada_exc_info &info : exceptions) 13256 gdb_printf ("%s: %s\n", info.name, paddress (gdbarch, info.addr)); 13257 } 13258 13259 13260 /* Language vector */ 13261 13262 /* symbol_name_matcher_ftype adapter for wild_match. */ 13263 13264 static bool 13265 do_wild_match (const char *symbol_search_name, 13266 const lookup_name_info &lookup_name, 13267 completion_match_result *comp_match_res) 13268 { 13269 return wild_match (symbol_search_name, ada_lookup_name (lookup_name)); 13270 } 13271 13272 /* symbol_name_matcher_ftype adapter for full_match. */ 13273 13274 static bool 13275 do_full_match (const char *symbol_search_name, 13276 const lookup_name_info &lookup_name, 13277 completion_match_result *comp_match_res) 13278 { 13279 const char *lname = lookup_name.ada ().lookup_name ().c_str (); 13280 13281 /* If both symbols start with "_ada_", just let the loop below 13282 handle the comparison. However, if only the symbol name starts 13283 with "_ada_", skip the prefix and let the match proceed as 13284 usual. */ 13285 if (startswith (symbol_search_name, "_ada_") 13286 && !startswith (lname, "_ada")) 13287 symbol_search_name += 5; 13288 /* Likewise for ghost entities. */ 13289 if (startswith (symbol_search_name, "___ghost_") 13290 && !startswith (lname, "___ghost_")) 13291 symbol_search_name += 9; 13292 13293 int uscore_count = 0; 13294 while (*lname != '\0') 13295 { 13296 if (*symbol_search_name != *lname) 13297 { 13298 if (*symbol_search_name == 'B' && uscore_count == 2 13299 && symbol_search_name[1] == '_') 13300 { 13301 symbol_search_name += 2; 13302 while (isdigit (*symbol_search_name)) 13303 ++symbol_search_name; 13304 if (symbol_search_name[0] == '_' 13305 && symbol_search_name[1] == '_') 13306 { 13307 symbol_search_name += 2; 13308 continue; 13309 } 13310 } 13311 return false; 13312 } 13313 13314 if (*symbol_search_name == '_') 13315 ++uscore_count; 13316 else 13317 uscore_count = 0; 13318 13319 ++symbol_search_name; 13320 ++lname; 13321 } 13322 13323 return is_name_suffix (symbol_search_name); 13324 } 13325 13326 /* symbol_name_matcher_ftype for exact (verbatim) matches. */ 13327 13328 static bool 13329 do_exact_match (const char *symbol_search_name, 13330 const lookup_name_info &lookup_name, 13331 completion_match_result *comp_match_res) 13332 { 13333 return strcmp (symbol_search_name, ada_lookup_name (lookup_name)) == 0; 13334 } 13335 13336 /* Build the Ada lookup name for LOOKUP_NAME. */ 13337 13338 ada_lookup_name_info::ada_lookup_name_info (const lookup_name_info &lookup_name) 13339 { 13340 gdb::string_view user_name = lookup_name.name (); 13341 13342 if (!user_name.empty () && user_name[0] == '<') 13343 { 13344 if (user_name.back () == '>') 13345 m_encoded_name 13346 = gdb::to_string (user_name.substr (1, user_name.size () - 2)); 13347 else 13348 m_encoded_name 13349 = gdb::to_string (user_name.substr (1, user_name.size () - 1)); 13350 m_encoded_p = true; 13351 m_verbatim_p = true; 13352 m_wild_match_p = false; 13353 m_standard_p = false; 13354 } 13355 else 13356 { 13357 m_verbatim_p = false; 13358 13359 m_encoded_p = user_name.find ("__") != gdb::string_view::npos; 13360 13361 if (!m_encoded_p) 13362 { 13363 const char *folded = ada_fold_name (user_name); 13364 m_encoded_name = ada_encode_1 (folded, false); 13365 if (m_encoded_name.empty ()) 13366 m_encoded_name = gdb::to_string (user_name); 13367 } 13368 else 13369 m_encoded_name = gdb::to_string (user_name); 13370 13371 /* Handle the 'package Standard' special case. See description 13372 of m_standard_p. */ 13373 if (startswith (m_encoded_name.c_str (), "standard__")) 13374 { 13375 m_encoded_name = m_encoded_name.substr (sizeof ("standard__") - 1); 13376 m_standard_p = true; 13377 } 13378 else 13379 m_standard_p = false; 13380 13381 /* If the name contains a ".", then the user is entering a fully 13382 qualified entity name, and the match must not be done in wild 13383 mode. Similarly, if the user wants to complete what looks 13384 like an encoded name, the match must not be done in wild 13385 mode. Also, in the standard__ special case always do 13386 non-wild matching. */ 13387 m_wild_match_p 13388 = (lookup_name.match_type () != symbol_name_match_type::FULL 13389 && !m_encoded_p 13390 && !m_standard_p 13391 && user_name.find ('.') == std::string::npos); 13392 } 13393 } 13394 13395 /* symbol_name_matcher_ftype method for Ada. This only handles 13396 completion mode. */ 13397 13398 static bool 13399 ada_symbol_name_matches (const char *symbol_search_name, 13400 const lookup_name_info &lookup_name, 13401 completion_match_result *comp_match_res) 13402 { 13403 return lookup_name.ada ().matches (symbol_search_name, 13404 lookup_name.match_type (), 13405 comp_match_res); 13406 } 13407 13408 /* A name matcher that matches the symbol name exactly, with 13409 strcmp. */ 13410 13411 static bool 13412 literal_symbol_name_matcher (const char *symbol_search_name, 13413 const lookup_name_info &lookup_name, 13414 completion_match_result *comp_match_res) 13415 { 13416 gdb::string_view name_view = lookup_name.name (); 13417 13418 if (lookup_name.completion_mode () 13419 ? (strncmp (symbol_search_name, name_view.data (), 13420 name_view.size ()) == 0) 13421 : symbol_search_name == name_view) 13422 { 13423 if (comp_match_res != NULL) 13424 comp_match_res->set_match (symbol_search_name); 13425 return true; 13426 } 13427 else 13428 return false; 13429 } 13430 13431 /* Implement the "get_symbol_name_matcher" language_defn method for 13432 Ada. */ 13433 13434 static symbol_name_matcher_ftype * 13435 ada_get_symbol_name_matcher (const lookup_name_info &lookup_name) 13436 { 13437 if (lookup_name.match_type () == symbol_name_match_type::SEARCH_NAME) 13438 return literal_symbol_name_matcher; 13439 13440 if (lookup_name.completion_mode ()) 13441 return ada_symbol_name_matches; 13442 else 13443 { 13444 if (lookup_name.ada ().wild_match_p ()) 13445 return do_wild_match; 13446 else if (lookup_name.ada ().verbatim_p ()) 13447 return do_exact_match; 13448 else 13449 return do_full_match; 13450 } 13451 } 13452 13453 /* Class representing the Ada language. */ 13454 13455 class ada_language : public language_defn 13456 { 13457 public: 13458 ada_language () 13459 : language_defn (language_ada) 13460 { /* Nothing. */ } 13461 13462 /* See language.h. */ 13463 13464 const char *name () const override 13465 { return "ada"; } 13466 13467 /* See language.h. */ 13468 13469 const char *natural_name () const override 13470 { return "Ada"; } 13471 13472 /* See language.h. */ 13473 13474 const std::vector<const char *> &filename_extensions () const override 13475 { 13476 static const std::vector<const char *> extensions 13477 = { ".adb", ".ads", ".a", ".ada", ".dg" }; 13478 return extensions; 13479 } 13480 13481 /* Print an array element index using the Ada syntax. */ 13482 13483 void print_array_index (struct type *index_type, 13484 LONGEST index, 13485 struct ui_file *stream, 13486 const value_print_options *options) const override 13487 { 13488 struct value *index_value = val_atr (index_type, index); 13489 13490 value_print (index_value, stream, options); 13491 gdb_printf (stream, " => "); 13492 } 13493 13494 /* Implement the "read_var_value" language_defn method for Ada. */ 13495 13496 struct value *read_var_value (struct symbol *var, 13497 const struct block *var_block, 13498 frame_info_ptr frame) const override 13499 { 13500 /* The only case where default_read_var_value is not sufficient 13501 is when VAR is a renaming... */ 13502 if (frame != nullptr) 13503 { 13504 const struct block *frame_block = get_frame_block (frame, NULL); 13505 if (frame_block != nullptr && ada_is_renaming_symbol (var)) 13506 return ada_read_renaming_var_value (var, frame_block); 13507 } 13508 13509 /* This is a typical case where we expect the default_read_var_value 13510 function to work. */ 13511 return language_defn::read_var_value (var, var_block, frame); 13512 } 13513 13514 /* See language.h. */ 13515 bool symbol_printing_suppressed (struct symbol *symbol) const override 13516 { 13517 return symbol->is_artificial (); 13518 } 13519 13520 /* See language.h. */ 13521 void language_arch_info (struct gdbarch *gdbarch, 13522 struct language_arch_info *lai) const override 13523 { 13524 const struct builtin_type *builtin = builtin_type (gdbarch); 13525 13526 /* Helper function to allow shorter lines below. */ 13527 auto add = [&] (struct type *t) 13528 { 13529 lai->add_primitive_type (t); 13530 }; 13531 13532 add (arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch), 13533 0, "integer")); 13534 add (arch_integer_type (gdbarch, gdbarch_long_bit (gdbarch), 13535 0, "long_integer")); 13536 add (arch_integer_type (gdbarch, gdbarch_short_bit (gdbarch), 13537 0, "short_integer")); 13538 struct type *char_type = arch_character_type (gdbarch, TARGET_CHAR_BIT, 13539 1, "character"); 13540 lai->set_string_char_type (char_type); 13541 add (char_type); 13542 add (arch_character_type (gdbarch, 16, 1, "wide_character")); 13543 add (arch_character_type (gdbarch, 32, 1, "wide_wide_character")); 13544 add (arch_float_type (gdbarch, gdbarch_float_bit (gdbarch), 13545 "float", gdbarch_float_format (gdbarch))); 13546 add (arch_float_type (gdbarch, gdbarch_double_bit (gdbarch), 13547 "long_float", gdbarch_double_format (gdbarch))); 13548 add (arch_integer_type (gdbarch, gdbarch_long_long_bit (gdbarch), 13549 0, "long_long_integer")); 13550 add (arch_float_type (gdbarch, gdbarch_long_double_bit (gdbarch), 13551 "long_long_float", 13552 gdbarch_long_double_format (gdbarch))); 13553 add (arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch), 13554 0, "natural")); 13555 add (arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch), 13556 0, "positive")); 13557 add (builtin->builtin_void); 13558 13559 struct type *system_addr_ptr 13560 = lookup_pointer_type (arch_type (gdbarch, TYPE_CODE_VOID, TARGET_CHAR_BIT, 13561 "void")); 13562 system_addr_ptr->set_name ("system__address"); 13563 add (system_addr_ptr); 13564 13565 /* Create the equivalent of the System.Storage_Elements.Storage_Offset 13566 type. This is a signed integral type whose size is the same as 13567 the size of addresses. */ 13568 unsigned int addr_length = system_addr_ptr->length (); 13569 add (arch_integer_type (gdbarch, addr_length * HOST_CHAR_BIT, 0, 13570 "storage_offset")); 13571 13572 lai->set_bool_type (builtin->builtin_bool); 13573 } 13574 13575 /* See language.h. */ 13576 13577 bool iterate_over_symbols 13578 (const struct block *block, const lookup_name_info &name, 13579 domain_enum domain, 13580 gdb::function_view<symbol_found_callback_ftype> callback) const override 13581 { 13582 std::vector<struct block_symbol> results 13583 = ada_lookup_symbol_list_worker (name, block, domain, 0); 13584 for (block_symbol &sym : results) 13585 { 13586 if (!callback (&sym)) 13587 return false; 13588 } 13589 13590 return true; 13591 } 13592 13593 /* See language.h. */ 13594 bool sniff_from_mangled_name 13595 (const char *mangled, 13596 gdb::unique_xmalloc_ptr<char> *out) const override 13597 { 13598 std::string demangled = ada_decode (mangled); 13599 13600 *out = NULL; 13601 13602 if (demangled != mangled && demangled[0] != '<') 13603 { 13604 /* Set the gsymbol language to Ada, but still return 0. 13605 Two reasons for that: 13606 13607 1. For Ada, we prefer computing the symbol's decoded name 13608 on the fly rather than pre-compute it, in order to save 13609 memory (Ada projects are typically very large). 13610 13611 2. There are some areas in the definition of the GNAT 13612 encoding where, with a bit of bad luck, we might be able 13613 to decode a non-Ada symbol, generating an incorrect 13614 demangled name (Eg: names ending with "TB" for instance 13615 are identified as task bodies and so stripped from 13616 the decoded name returned). 13617 13618 Returning true, here, but not setting *DEMANGLED, helps us get 13619 a little bit of the best of both worlds. Because we're last, 13620 we should not affect any of the other languages that were 13621 able to demangle the symbol before us; we get to correctly 13622 tag Ada symbols as such; and even if we incorrectly tagged a 13623 non-Ada symbol, which should be rare, any routing through the 13624 Ada language should be transparent (Ada tries to behave much 13625 like C/C++ with non-Ada symbols). */ 13626 return true; 13627 } 13628 13629 return false; 13630 } 13631 13632 /* See language.h. */ 13633 13634 gdb::unique_xmalloc_ptr<char> demangle_symbol (const char *mangled, 13635 int options) const override 13636 { 13637 return make_unique_xstrdup (ada_decode (mangled).c_str ()); 13638 } 13639 13640 /* See language.h. */ 13641 13642 void print_type (struct type *type, const char *varstring, 13643 struct ui_file *stream, int show, int level, 13644 const struct type_print_options *flags) const override 13645 { 13646 ada_print_type (type, varstring, stream, show, level, flags); 13647 } 13648 13649 /* See language.h. */ 13650 13651 const char *word_break_characters (void) const override 13652 { 13653 return ada_completer_word_break_characters; 13654 } 13655 13656 /* See language.h. */ 13657 13658 void collect_symbol_completion_matches (completion_tracker &tracker, 13659 complete_symbol_mode mode, 13660 symbol_name_match_type name_match_type, 13661 const char *text, const char *word, 13662 enum type_code code) const override 13663 { 13664 struct symbol *sym; 13665 const struct block *b, *surrounding_static_block = 0; 13666 struct block_iterator iter; 13667 13668 gdb_assert (code == TYPE_CODE_UNDEF); 13669 13670 lookup_name_info lookup_name (text, name_match_type, true); 13671 13672 /* First, look at the partial symtab symbols. */ 13673 expand_symtabs_matching (NULL, 13674 lookup_name, 13675 NULL, 13676 NULL, 13677 SEARCH_GLOBAL_BLOCK | SEARCH_STATIC_BLOCK, 13678 ALL_DOMAIN); 13679 13680 /* At this point scan through the misc symbol vectors and add each 13681 symbol you find to the list. Eventually we want to ignore 13682 anything that isn't a text symbol (everything else will be 13683 handled by the psymtab code above). */ 13684 13685 for (objfile *objfile : current_program_space->objfiles ()) 13686 { 13687 for (minimal_symbol *msymbol : objfile->msymbols ()) 13688 { 13689 QUIT; 13690 13691 if (completion_skip_symbol (mode, msymbol)) 13692 continue; 13693 13694 language symbol_language = msymbol->language (); 13695 13696 /* Ada minimal symbols won't have their language set to Ada. If 13697 we let completion_list_add_name compare using the 13698 default/C-like matcher, then when completing e.g., symbols in a 13699 package named "pck", we'd match internal Ada symbols like 13700 "pckS", which are invalid in an Ada expression, unless you wrap 13701 them in '<' '>' to request a verbatim match. 13702 13703 Unfortunately, some Ada encoded names successfully demangle as 13704 C++ symbols (using an old mangling scheme), such as "name__2Xn" 13705 -> "Xn::name(void)" and thus some Ada minimal symbols end up 13706 with the wrong language set. Paper over that issue here. */ 13707 if (symbol_language == language_auto 13708 || symbol_language == language_cplus) 13709 symbol_language = language_ada; 13710 13711 completion_list_add_name (tracker, 13712 symbol_language, 13713 msymbol->linkage_name (), 13714 lookup_name, text, word); 13715 } 13716 } 13717 13718 /* Search upwards from currently selected frame (so that we can 13719 complete on local vars. */ 13720 13721 for (b = get_selected_block (0); b != NULL; b = b->superblock ()) 13722 { 13723 if (!b->superblock ()) 13724 surrounding_static_block = b; /* For elmin of dups */ 13725 13726 ALL_BLOCK_SYMBOLS (b, iter, sym) 13727 { 13728 if (completion_skip_symbol (mode, sym)) 13729 continue; 13730 13731 completion_list_add_name (tracker, 13732 sym->language (), 13733 sym->linkage_name (), 13734 lookup_name, text, word); 13735 } 13736 } 13737 13738 /* Go through the symtabs and check the externs and statics for 13739 symbols which match. */ 13740 13741 for (objfile *objfile : current_program_space->objfiles ()) 13742 { 13743 for (compunit_symtab *s : objfile->compunits ()) 13744 { 13745 QUIT; 13746 b = s->blockvector ()->global_block (); 13747 ALL_BLOCK_SYMBOLS (b, iter, sym) 13748 { 13749 if (completion_skip_symbol (mode, sym)) 13750 continue; 13751 13752 completion_list_add_name (tracker, 13753 sym->language (), 13754 sym->linkage_name (), 13755 lookup_name, text, word); 13756 } 13757 } 13758 } 13759 13760 for (objfile *objfile : current_program_space->objfiles ()) 13761 { 13762 for (compunit_symtab *s : objfile->compunits ()) 13763 { 13764 QUIT; 13765 b = s->blockvector ()->static_block (); 13766 /* Don't do this block twice. */ 13767 if (b == surrounding_static_block) 13768 continue; 13769 ALL_BLOCK_SYMBOLS (b, iter, sym) 13770 { 13771 if (completion_skip_symbol (mode, sym)) 13772 continue; 13773 13774 completion_list_add_name (tracker, 13775 sym->language (), 13776 sym->linkage_name (), 13777 lookup_name, text, word); 13778 } 13779 } 13780 } 13781 } 13782 13783 /* See language.h. */ 13784 13785 gdb::unique_xmalloc_ptr<char> watch_location_expression 13786 (struct type *type, CORE_ADDR addr) const override 13787 { 13788 type = check_typedef (check_typedef (type)->target_type ()); 13789 std::string name = type_to_string (type); 13790 return xstrprintf ("{%s} %s", name.c_str (), core_addr_to_string (addr)); 13791 } 13792 13793 /* See language.h. */ 13794 13795 void value_print (struct value *val, struct ui_file *stream, 13796 const struct value_print_options *options) const override 13797 { 13798 return ada_value_print (val, stream, options); 13799 } 13800 13801 /* See language.h. */ 13802 13803 void value_print_inner 13804 (struct value *val, struct ui_file *stream, int recurse, 13805 const struct value_print_options *options) const override 13806 { 13807 return ada_value_print_inner (val, stream, recurse, options); 13808 } 13809 13810 /* See language.h. */ 13811 13812 struct block_symbol lookup_symbol_nonlocal 13813 (const char *name, const struct block *block, 13814 const domain_enum domain) const override 13815 { 13816 struct block_symbol sym; 13817 13818 sym = ada_lookup_symbol (name, block_static_block (block), domain); 13819 if (sym.symbol != NULL) 13820 return sym; 13821 13822 /* If we haven't found a match at this point, try the primitive 13823 types. In other languages, this search is performed before 13824 searching for global symbols in order to short-circuit that 13825 global-symbol search if it happens that the name corresponds 13826 to a primitive type. But we cannot do the same in Ada, because 13827 it is perfectly legitimate for a program to declare a type which 13828 has the same name as a standard type. If looking up a type in 13829 that situation, we have traditionally ignored the primitive type 13830 in favor of user-defined types. This is why, unlike most other 13831 languages, we search the primitive types this late and only after 13832 having searched the global symbols without success. */ 13833 13834 if (domain == VAR_DOMAIN) 13835 { 13836 struct gdbarch *gdbarch; 13837 13838 if (block == NULL) 13839 gdbarch = target_gdbarch (); 13840 else 13841 gdbarch = block_gdbarch (block); 13842 sym.symbol 13843 = language_lookup_primitive_type_as_symbol (this, gdbarch, name); 13844 if (sym.symbol != NULL) 13845 return sym; 13846 } 13847 13848 return {}; 13849 } 13850 13851 /* See language.h. */ 13852 13853 int parser (struct parser_state *ps) const override 13854 { 13855 warnings_issued = 0; 13856 return ada_parse (ps); 13857 } 13858 13859 /* See language.h. */ 13860 13861 void emitchar (int ch, struct type *chtype, 13862 struct ui_file *stream, int quoter) const override 13863 { 13864 ada_emit_char (ch, chtype, stream, quoter, 1); 13865 } 13866 13867 /* See language.h. */ 13868 13869 void printchar (int ch, struct type *chtype, 13870 struct ui_file *stream) const override 13871 { 13872 ada_printchar (ch, chtype, stream); 13873 } 13874 13875 /* See language.h. */ 13876 13877 void printstr (struct ui_file *stream, struct type *elttype, 13878 const gdb_byte *string, unsigned int length, 13879 const char *encoding, int force_ellipses, 13880 const struct value_print_options *options) const override 13881 { 13882 ada_printstr (stream, elttype, string, length, encoding, 13883 force_ellipses, options); 13884 } 13885 13886 /* See language.h. */ 13887 13888 void print_typedef (struct type *type, struct symbol *new_symbol, 13889 struct ui_file *stream) const override 13890 { 13891 ada_print_typedef (type, new_symbol, stream); 13892 } 13893 13894 /* See language.h. */ 13895 13896 bool is_string_type_p (struct type *type) const override 13897 { 13898 return ada_is_string_type (type); 13899 } 13900 13901 /* See language.h. */ 13902 13903 const char *struct_too_deep_ellipsis () const override 13904 { return "(...)"; } 13905 13906 /* See language.h. */ 13907 13908 bool c_style_arrays_p () const override 13909 { return false; } 13910 13911 /* See language.h. */ 13912 13913 bool store_sym_names_in_linkage_form_p () const override 13914 { return true; } 13915 13916 /* See language.h. */ 13917 13918 const struct lang_varobj_ops *varobj_ops () const override 13919 { return &ada_varobj_ops; } 13920 13921 protected: 13922 /* See language.h. */ 13923 13924 symbol_name_matcher_ftype *get_symbol_name_matcher_inner 13925 (const lookup_name_info &lookup_name) const override 13926 { 13927 return ada_get_symbol_name_matcher (lookup_name); 13928 } 13929 }; 13930 13931 /* Single instance of the Ada language class. */ 13932 13933 static ada_language ada_language_defn; 13934 13935 /* Command-list for the "set/show ada" prefix command. */ 13936 static struct cmd_list_element *set_ada_list; 13937 static struct cmd_list_element *show_ada_list; 13938 13939 /* This module's 'new_objfile' observer. */ 13940 13941 static void 13942 ada_new_objfile_observer (struct objfile *objfile) 13943 { 13944 ada_clear_symbol_cache (); 13945 } 13946 13947 /* This module's 'free_objfile' observer. */ 13948 13949 static void 13950 ada_free_objfile_observer (struct objfile *objfile) 13951 { 13952 ada_clear_symbol_cache (); 13953 } 13954 13955 /* Charsets known to GNAT. */ 13956 static const char * const gnat_source_charsets[] = 13957 { 13958 /* Note that code below assumes that the default comes first. 13959 Latin-1 is the default here, because that is also GNAT's 13960 default. */ 13961 "ISO-8859-1", 13962 "ISO-8859-2", 13963 "ISO-8859-3", 13964 "ISO-8859-4", 13965 "ISO-8859-5", 13966 "ISO-8859-15", 13967 "CP437", 13968 "CP850", 13969 /* Note that this value is special-cased in the encoder and 13970 decoder. */ 13971 ada_utf8, 13972 nullptr 13973 }; 13974 13975 void _initialize_ada_language (); 13976 void 13977 _initialize_ada_language () 13978 { 13979 add_setshow_prefix_cmd 13980 ("ada", no_class, 13981 _("Prefix command for changing Ada-specific settings."), 13982 _("Generic command for showing Ada-specific settings."), 13983 &set_ada_list, &show_ada_list, 13984 &setlist, &showlist); 13985 13986 add_setshow_boolean_cmd ("trust-PAD-over-XVS", class_obscure, 13987 &trust_pad_over_xvs, _("\ 13988 Enable or disable an optimization trusting PAD types over XVS types."), _("\ 13989 Show whether an optimization trusting PAD types over XVS types is activated."), 13990 _("\ 13991 This is related to the encoding used by the GNAT compiler. The debugger\n\ 13992 should normally trust the contents of PAD types, but certain older versions\n\ 13993 of GNAT have a bug that sometimes causes the information in the PAD type\n\ 13994 to be incorrect. Turning this setting \"off\" allows the debugger to\n\ 13995 work around this bug. It is always safe to turn this option \"off\", but\n\ 13996 this incurs a slight performance penalty, so it is recommended to NOT change\n\ 13997 this option to \"off\" unless necessary."), 13998 NULL, NULL, &set_ada_list, &show_ada_list); 13999 14000 add_setshow_boolean_cmd ("print-signatures", class_vars, 14001 &print_signatures, _("\ 14002 Enable or disable the output of formal and return types for functions in the \ 14003 overloads selection menu."), _("\ 14004 Show whether the output of formal and return types for functions in the \ 14005 overloads selection menu is activated."), 14006 NULL, NULL, NULL, &set_ada_list, &show_ada_list); 14007 14008 ada_source_charset = gnat_source_charsets[0]; 14009 add_setshow_enum_cmd ("source-charset", class_files, 14010 gnat_source_charsets, 14011 &ada_source_charset, _("\ 14012 Set the Ada source character set."), _("\ 14013 Show the Ada source character set."), _("\ 14014 The character set used for Ada source files.\n\ 14015 This must correspond to the '-gnati' or '-gnatW' option passed to GNAT."), 14016 nullptr, nullptr, 14017 &set_ada_list, &show_ada_list); 14018 14019 add_catch_command ("exception", _("\ 14020 Catch Ada exceptions, when raised.\n\ 14021 Usage: catch exception [ARG] [if CONDITION]\n\ 14022 Without any argument, stop when any Ada exception is raised.\n\ 14023 If ARG is \"unhandled\" (without the quotes), only stop when the exception\n\ 14024 being raised does not have a handler (and will therefore lead to the task's\n\ 14025 termination).\n\ 14026 Otherwise, the catchpoint only stops when the name of the exception being\n\ 14027 raised is the same as ARG.\n\ 14028 CONDITION is a boolean expression that is evaluated to see whether the\n\ 14029 exception should cause a stop."), 14030 catch_ada_exception_command, 14031 catch_ada_completer, 14032 CATCH_PERMANENT, 14033 CATCH_TEMPORARY); 14034 14035 add_catch_command ("handlers", _("\ 14036 Catch Ada exceptions, when handled.\n\ 14037 Usage: catch handlers [ARG] [if CONDITION]\n\ 14038 Without any argument, stop when any Ada exception is handled.\n\ 14039 With an argument, catch only exceptions with the given name.\n\ 14040 CONDITION is a boolean expression that is evaluated to see whether the\n\ 14041 exception should cause a stop."), 14042 catch_ada_handlers_command, 14043 catch_ada_completer, 14044 CATCH_PERMANENT, 14045 CATCH_TEMPORARY); 14046 add_catch_command ("assert", _("\ 14047 Catch failed Ada assertions, when raised.\n\ 14048 Usage: catch assert [if CONDITION]\n\ 14049 CONDITION is a boolean expression that is evaluated to see whether the\n\ 14050 exception should cause a stop."), 14051 catch_assert_command, 14052 NULL, 14053 CATCH_PERMANENT, 14054 CATCH_TEMPORARY); 14055 14056 add_info ("exceptions", info_exceptions_command, 14057 _("\ 14058 List all Ada exception names.\n\ 14059 Usage: info exceptions [REGEXP]\n\ 14060 If a regular expression is passed as an argument, only those matching\n\ 14061 the regular expression are listed.")); 14062 14063 add_setshow_prefix_cmd ("ada", class_maintenance, 14064 _("Set Ada maintenance-related variables."), 14065 _("Show Ada maintenance-related variables."), 14066 &maint_set_ada_cmdlist, &maint_show_ada_cmdlist, 14067 &maintenance_set_cmdlist, &maintenance_show_cmdlist); 14068 14069 add_setshow_boolean_cmd 14070 ("ignore-descriptive-types", class_maintenance, 14071 &ada_ignore_descriptive_types_p, 14072 _("Set whether descriptive types generated by GNAT should be ignored."), 14073 _("Show whether descriptive types generated by GNAT should be ignored."), 14074 _("\ 14075 When enabled, the debugger will stop using the DW_AT_GNAT_descriptive_type\n\ 14076 DWARF attribute."), 14077 NULL, NULL, &maint_set_ada_cmdlist, &maint_show_ada_cmdlist); 14078 14079 decoded_names_store = htab_create_alloc (256, htab_hash_string, 14080 htab_eq_string, 14081 NULL, xcalloc, xfree); 14082 14083 /* The ada-lang observers. */ 14084 gdb::observers::new_objfile.attach (ada_new_objfile_observer, "ada-lang"); 14085 gdb::observers::free_objfile.attach (ada_free_objfile_observer, "ada-lang"); 14086 gdb::observers::inferior_exit.attach (ada_inferior_exit, "ada-lang"); 14087 } 14088