gdb/language.h

   1 /* Source-language-related definitions for GDB.
   2
   3    Copyright (C) 1991-2014 Free Software Foundation, Inc.
   4
   5    Contributed by the Department of Computer Science at the State University
   6    of New York at Buffalo.
   7
   8    This file is part of GDB.
   9
  10    This program is free software; you can redistribute it and/or modify
  11    it under the terms of the GNU General Public License as published by
  12    the Free Software Foundation; either version 3 of the License, or
  13    (at your option) any later version.
  14
  15    This program is distributed in the hope that it will be useful,
  16    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18    GNU General Public License for more details.
  19
  20    You should have received a copy of the GNU General Public License
  21    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  22
  23 #if !defined (LANGUAGE_H)
  24 #define LANGUAGE_H 1
  25
  26 #include "symtab.h"
  27
  28 /* Forward decls for prototypes.  */
  29 struct value;
  30 struct objfile;
  31 struct frame_info;
  32 struct expression;
  33 struct ui_file;
  34 struct value_print_options;
  35 struct type_print_options;
  36 struct lang_varobj_ops;
  37 struct parser_state;
  38 struct compile_instance;
  39
  40 #define MAX_FORTRAN_DIMS  7     /* Maximum number of F77 array dims.  */
  41
  42 /* range_mode ==
  43    range_mode_auto:   range_check set automatically to default of language.
  44    range_mode_manual: range_check set manually by user.  */
  45
  46 extern enum range_mode
  47   {
  48     range_mode_auto, range_mode_manual
  49   }
  50 range_mode;
  51
  52 /* range_check ==
  53    range_check_on:    Ranges are checked in GDB expressions, producing errors.
  54    range_check_warn:  Ranges are checked, producing warnings.
  55    range_check_off:   Ranges are not checked in GDB expressions.  */
  56
  57 extern enum range_check
  58   {
  59     range_check_off, range_check_warn, range_check_on
  60   }
  61 range_check;
  62
  63 /* case_mode ==
  64    case_mode_auto:   case_sensitivity set upon selection of scope.
  65    case_mode_manual: case_sensitivity set only by user.  */
  66
  67 extern enum case_mode
  68   {
  69     case_mode_auto, case_mode_manual
  70   }
  71 case_mode;
  72
  73 /* array_ordering ==
  74    array_row_major:     Arrays are in row major order.
  75    array_column_major:  Arrays are in column major order.  */
  76
  77 extern enum array_ordering
  78   {
  79     array_row_major, array_column_major
  80   }
  81 array_ordering;
  82
  83
  84 /* case_sensitivity ==
  85    case_sensitive_on:   Case sensitivity in name matching is used.
  86    case_sensitive_off:  Case sensitivity in name matching is not used.  */
  87
  88 extern enum case_sensitivity
  89   {
  90     case_sensitive_on, case_sensitive_off
  91   }
  92 case_sensitivity;
  93
  94
  95 /* macro_expansion ==
  96    macro_expansion_no:  No macro expansion is available.
  97    macro_expansion_c:   C-like macro expansion is available.  */
  98
  99 enum macro_expansion
 100   {
 101     macro_expansion_no, macro_expansion_c
 102   };
 103
 104 \f
 105 /* Per architecture (OS/ABI) language information.  */
 106
 107 struct language_arch_info
 108 {
 109   /* Its primitive types.  This is a vector ended by a NULL pointer.
 110      These types can be specified by name in parsing types in
 111      expressions, regardless of whether the program being debugged
 112      actually defines such a type.  */
 113   struct type **primitive_type_vector;
 114   /* Type of elements of strings.  */
 115   struct type *string_char_type;
 116
 117   /* Symbol name of type to use as boolean type, if defined.  */
 118   const char *bool_type_symbol;
 119   /* Otherwise, this is the default boolean builtin type.  */
 120   struct type *bool_type_default;
 121 };
 122
 123 /* A pointer to a function expected to return nonzero if
 124    SYMBOL_SEARCH_NAME matches the given LOOKUP_NAME.
 125
 126    SYMBOL_SEARCH_NAME should be a symbol's "search" name.
 127    LOOKUP_NAME should be the name of an entity after it has been
 128    transformed for lookup.  */
 129
 130 typedef int (*symbol_name_cmp_ftype) (const char *symbol_search_name,
 131                                           const char *lookup_name);
 132
 133 /* Structure tying together assorted information about a language.  */
 134
 135 struct language_defn
 136   {
 137     /* Name of the language.  */
 138
 139     const char *la_name;
 140
 141     /* Natural or official name of the language.  */
 142
 143     const char *la_natural_name;
 144
 145     /* its symtab language-enum (defs.h).  */
 146
 147     enum language la_language;
 148
 149     /* Default range checking.  */
 150
 151     enum range_check la_range_check;
 152
 153     /* Default case sensitivity.  */
 154     enum case_sensitivity la_case_sensitivity;
 155
 156     /* Multi-dimensional array ordering.  */
 157     enum array_ordering la_array_ordering;
 158
 159     /* Style of macro expansion, if any, supported by this language.  */
 160     enum macro_expansion la_macro_expansion;
 161
 162     /* Definitions related to expression printing, prefixifying, and
 163        dumping.  */
 164
 165     const struct exp_descriptor *la_exp_desc;
 166
 167     /* Parser function.  */
 168
 169     int (*la_parser) (struct parser_state *);
 170
 171     /* Parser error function.  */
 172
 173     void (*la_error) (char *);
 174
 175     /* Given an expression *EXPP created by prefixifying the result of
 176        la_parser, perform any remaining processing necessary to complete
 177        its translation.  *EXPP may change; la_post_parser is responsible
 178        for releasing its previous contents, if necessary.  If
 179        VOID_CONTEXT_P, then no value is expected from the expression.  */
 180
 181     void (*la_post_parser) (struct expression ** expp, int void_context_p);
 182
 183     void (*la_printchar) (int ch, struct type *chtype,
 184                           struct ui_file * stream);
 185
 186     void (*la_printstr) (struct ui_file * stream, struct type *elttype,
 187                          const gdb_byte *string, unsigned int length,
 188                          const char *encoding, int force_ellipses,
 189                          const struct value_print_options *);
 190
 191     void (*la_emitchar) (int ch, struct type *chtype,
 192                          struct ui_file * stream, int quoter);
 193
 194     /* Print a type using syntax appropriate for this language.  */
 195
 196     void (*la_print_type) (struct type *, const char *, struct ui_file *, int,
 197                            int, const struct type_print_options *);
 198
 199     /* Print a typedef using syntax appropriate for this language.
 200        TYPE is the underlying type.  NEW_SYMBOL is the symbol naming
 201        the type.  STREAM is the output stream on which to print.  */
 202
 203     void (*la_print_typedef) (struct type *type, struct symbol *new_symbol,
 204                               struct ui_file *stream);
 205
 206     /* Print a value using syntax appropriate for this language.
 207
 208        TYPE is the type of the sub-object to be printed.
 209
 210        CONTENTS holds the bits of the value.  This holds the entire
 211        enclosing object.
 212
 213        EMBEDDED_OFFSET is the offset into the outermost object of the
 214        sub-object represented by TYPE.  This is the object which this
 215        call should print.  Note that the enclosing type is not
 216        available.
 217
 218        ADDRESS is the address in the inferior of the enclosing object.
 219
 220        STREAM is the stream on which the value is to be printed.
 221
 222        RECURSE is the recursion depth.  It is zero-based.
 223
 224        OPTIONS are the formatting options to be used when
 225        printing.  */
 226
 227     void (*la_val_print) (struct type *type,
 228                           const gdb_byte *contents,
 229                           int embedded_offset, CORE_ADDR address,
 230                           struct ui_file *stream, int recurse,
 231                           const struct value *val,
 232                           const struct value_print_options *options);
 233
 234     /* Print a top-level value using syntax appropriate for this language.  */
 235
 236     void (*la_value_print) (struct value *, struct ui_file *,
 237                             const struct value_print_options *);
 238
 239     /* Given a symbol VAR, and a stack frame id FRAME, read the value
 240        of the variable an return (pointer to a) struct value containing
 241        the value.
 242
 243        Throw an error if the variable cannot be found.  */
 244
 245     struct value *(*la_read_var_value) (struct symbol *var,
 246                                         struct frame_info *frame);
 247
 248     /* PC is possibly an unknown languages trampoline.
 249        If that PC falls in a trampoline belonging to this language,
 250        return the address of the first pc in the real function, or 0
 251        if it isn't a language tramp for this language.  */
 252     CORE_ADDR (*skip_trampoline) (struct frame_info *, CORE_ADDR);
 253
 254     /* Now come some hooks for lookup_symbol.  */
 255
 256     /* If this is non-NULL, specifies the name that of the implicit
 257        local variable that refers to the current object instance.  */
 258
 259     char *la_name_of_this;
 260
 261     /* This is a function that lookup_symbol will call when it gets to
 262        the part of symbol lookup where C looks up static and global
 263        variables.  */
 264
 265     struct symbol *(*la_lookup_symbol_nonlocal) (const char *,
 266                                                  const struct block *,
 267                                                  const domain_enum);
 268
 269     /* Find the definition of the type with the given name.  */
 270     struct type *(*la_lookup_transparent_type) (const char *);
 271
 272     /* Return demangled language symbol, or NULL.  */
 273     char *(*la_demangle) (const char *mangled, int options);
 274
 275     /* Return class name of a mangled method name or NULL.  */
 276     char *(*la_class_name_from_physname) (const char *physname);
 277
 278     /* Table for printing expressions.  */
 279
 280     const struct op_print *la_op_print_tab;
 281
 282     /* Zero if the language has first-class arrays.  True if there are no
 283        array values, and array objects decay to pointers, as in C.  */
 284
 285     char c_style_arrays;
 286
 287     /* Index to use for extracting the first element of a string.  */
 288     char string_lower_bound;
 289
 290     /* The list of characters forming word boundaries.  */
 291     char *(*la_word_break_characters) (void);
 292
 293     /* Should return a vector of all symbols which are possible
 294        completions for TEXT.  WORD is the entire command on which the
 295        completion is being made.  If CODE is TYPE_CODE_UNDEF, then all
 296        symbols should be examined; otherwise, only STRUCT_DOMAIN
 297        symbols whose type has a code of CODE should be matched.  */
 298     VEC (char_ptr) *(*la_make_symbol_completion_list) (const char *text,
 299                                                        const char *word,
 300                                                        enum type_code code);
 301
 302     /* The per-architecture (OS/ABI) language information.  */
 303     void (*la_language_arch_info) (struct gdbarch *,
 304                                    struct language_arch_info *);
 305
 306     /* Print the index of an element of an array.  */
 307     void (*la_print_array_index) (struct value *index_value,
 308                                   struct ui_file *stream,
 309                                   const struct value_print_options *options);
 310
 311     /* Return non-zero if TYPE should be passed (and returned) by
 312        reference at the language level.  */
 313     int (*la_pass_by_reference) (struct type *type);
 314
 315     /* Obtain a string from the inferior, storing it in a newly allocated
 316        buffer in BUFFER, which should be freed by the caller.  If the
 317        in- and out-parameter *LENGTH is specified at -1, the string is
 318        read until a null character of the appropriate width is found -
 319        otherwise the string is read to the length of characters specified.
 320        On completion, *LENGTH will hold the size of the string in characters.
 321        If a *LENGTH of -1 was specified it will count only actual
 322        characters, excluding any eventual terminating null character.
 323        Otherwise *LENGTH will include all characters - including any nulls.
 324        CHARSET will hold the encoding used in the string.  */
 325     void (*la_get_string) (struct value *value, gdb_byte **buffer, int *length,
 326                            struct type **chartype, const char **charset);
 327
 328     /* Return a pointer to the function that should be used to match
 329        a symbol name against LOOKUP_NAME. This is mostly for languages
 330        such as Ada where the matching algorithm depends on LOOKUP_NAME.
 331
 332        This field may be NULL, in which case strcmp_iw will be used
 333        to perform the matching.  */
 334     symbol_name_cmp_ftype (*la_get_symbol_name_cmp) (const char *lookup_name);
 335
 336     /* Find all symbols in the current program space matching NAME in
 337        DOMAIN, according to this language's rules.
 338
 339        The search is done in BLOCK only.
 340        The caller is responsible for iterating up through superblocks
 341        if desired.
 342
 343        For each one, call CALLBACK with the symbol and the DATA
 344        argument.  If CALLBACK returns zero, the iteration ends at that
 345        point.
 346
 347        This field may not be NULL.  If the language does not need any
 348        special processing here, 'iterate_over_symbols' should be
 349        used as the definition.  */
 350     void (*la_iterate_over_symbols) (const struct block *block,
 351                                      const char *name,
 352                                      domain_enum domain,
 353                                      symbol_found_callback_ftype *callback,
 354                                      void *data);
 355
 356     /* Various operations on varobj.  */
 357     const struct lang_varobj_ops *la_varobj_ops;
 358
 359     /* If this language allows compilation from the gdb command line,
 360        this method should be non-NULL.  When called it should return
 361        an instance of struct gcc_context appropriate to the language.
 362        When defined this method must never return NULL; instead it
 363        should throw an exception on failure.  The returned compiler
 364        instance is owned by its caller and must be deallocated by
 365        calling its 'destroy' method.  */
 366
 367     struct compile_instance *(*la_get_compile_instance) (void);
 368
 369     /* This method must be defined if 'la_get_gcc_context' is defined.
 370        If 'la_get_gcc_context' is not defined, then this method is
 371        ignored.
 372
 373        This takes the user-supplied text and returns a newly malloc'd
 374        bit of code to compile.  The caller owns the result.
 375
 376        INST is the compiler instance being used.
 377        INPUT is the user's input text.
 378        GDBARCH is the architecture to use.
 379        EXPR_BLOCK is the block in which the expression is being
 380        parsed.
 381        EXPR_PC is the PC at which the expression is being parsed.  */
 382
 383     char *(*la_compute_program) (struct compile_instance *inst,
 384                                  const char *input,
 385                                  struct gdbarch *gdbarch,
 386                                  const struct block *expr_block,
 387                                  CORE_ADDR expr_pc);
 388
 389     /* Add fields above this point, so the magic number is always last.  */
 390     /* Magic number for compat checking.  */
 391
 392     long la_magic;
 393
 394   };
 395
 396 #define LANG_MAGIC      910823L
 397
 398 /* Pointer to the language_defn for our current language.  This pointer
 399    always points to *some* valid struct; it can be used without checking
 400    it for validity.
 401
 402    The current language affects expression parsing and evaluation
 403    (FIXME: it might be cleaner to make the evaluation-related stuff
 404    separate exp_opcodes for each different set of semantics.  We
 405    should at least think this through more clearly with respect to
 406    what happens if the language is changed between parsing and
 407    evaluation) and printing of things like types and arrays.  It does
 408    *not* affect symbol-reading-- each source file in a symbol-file has
 409    its own language and we should keep track of that regardless of the
 410    language when symbols are read.  If we want some manual setting for
 411    the language of symbol files (e.g. detecting when ".c" files are
 412    C++), it should be a separate setting from the current_language.  */
 413
 414 extern const struct language_defn *current_language;
 415
 416 /* Pointer to the language_defn expected by the user, e.g. the language
 417    of main(), or the language we last mentioned in a message, or C.  */
 418
 419 extern const struct language_defn *expected_language;
 420
 421 /* language_mode ==
 422    language_mode_auto:   current_language automatically set upon selection
 423    of scope (e.g. stack frame)
 424    language_mode_manual: current_language set only by user.  */
 425
 426 extern enum language_mode
 427   {
 428     language_mode_auto, language_mode_manual
 429   }
 430 language_mode;
 431
 432 struct type *language_bool_type (const struct language_defn *l,
 433                                  struct gdbarch *gdbarch);
 434
 435 struct type *language_string_char_type (const struct language_defn *l,
 436                                         struct gdbarch *gdbarch);
 437
 438 struct type *language_lookup_primitive_type_by_name (const struct language_defn *l,
 439                                                      struct gdbarch *gdbarch,
 440                                                      const char *name);
 441
 442 \f
 443 /* These macros define the behaviour of the expression
 444    evaluator.  */
 445
 446 /* Should we range check values against the domain of their type?  */
 447 #define RANGE_CHECK (range_check != range_check_off)
 448
 449 /* "cast" really means conversion.  */
 450 /* FIXME -- should be a setting in language_defn.  */
 451 #define CAST_IS_CONVERSION(LANG) ((LANG)->la_language == language_c  || \
 452                                   (LANG)->la_language == language_cplus || \
 453                                   (LANG)->la_language == language_objc)
 454
 455 extern void language_info (int);
 456
 457 extern enum language set_language (enum language);
 458 \f
 459
 460 /* This page contains functions that return things that are
 461    specific to languages.  Each of these functions is based on
 462    the current setting of working_lang, which the user sets
 463    with the "set language" command.  */
 464
 465 #define LA_PRINT_TYPE(type,varstring,stream,show,level,flags)           \
 466   (current_language->la_print_type(type,varstring,stream,show,level,flags))
 467
 468 #define LA_PRINT_TYPEDEF(type,new_symbol,stream) \
 469   (current_language->la_print_typedef(type,new_symbol,stream))
 470
 471 #define LA_VAL_PRINT(type,valaddr,offset,addr,stream,val,recurse,options) \
 472   (current_language->la_val_print(type,valaddr,offset,addr,stream, \
 473                                   val,recurse,options))
 474 #define LA_VALUE_PRINT(val,stream,options) \
 475   (current_language->la_value_print(val,stream,options))
 476
 477 #define LA_PRINT_CHAR(ch, type, stream) \
 478   (current_language->la_printchar(ch, type, stream))
 479 #define LA_PRINT_STRING(stream, elttype, string, length, encoding, force_ellipses, options) \
 480   (current_language->la_printstr(stream, elttype, string, length, \
 481                                  encoding, force_ellipses,options))
 482 #define LA_EMIT_CHAR(ch, type, stream, quoter) \
 483   (current_language->la_emitchar(ch, type, stream, quoter))
 484 #define LA_GET_STRING(value, buffer, length, chartype, encoding) \
 485   (current_language->la_get_string(value, buffer, length, chartype, encoding))
 486
 487 #define LA_PRINT_ARRAY_INDEX(index_value, stream, options) \
 488   (current_language->la_print_array_index(index_value, stream, options))
 489
 490 #define LA_ITERATE_OVER_SYMBOLS(BLOCK, NAME, DOMAIN, CALLBACK, DATA) \
 491   (current_language->la_iterate_over_symbols (BLOCK, NAME, DOMAIN, CALLBACK, \
 492                                               DATA))
 493
 494 /* Test a character to decide whether it can be printed in literal form
 495    or needs to be printed in another representation.  For example,
 496    in C the literal form of the character with octal value 141 is 'a'
 497    and the "other representation" is '\141'.  The "other representation"
 498    is program language dependent.  */
 499
 500 #define PRINT_LITERAL_FORM(c)           \
 501   ((c) >= 0x20                          \
 502    && ((c) < 0x7F || (c) >= 0xA0)       \
 503    && (!sevenbit_strings || (c) < 0x80))
 504
 505 /* Type predicates */
 506
 507 extern int pointer_type (struct type *);
 508
 509 /* Error messages */
 510
 511 extern void range_error (const char *, ...) ATTRIBUTE_PRINTF (1, 2);
 512
 513 /* Data:  Does this value represent "truth" to the current language?  */
 514
 515 extern int value_true (struct value *);
 516
 517 /* Misc:  The string representing a particular enum language.  */
 518
 519 extern enum language language_enum (char *str);
 520
 521 extern const struct language_defn *language_def (enum language);
 522
 523 extern const char *language_str (enum language);
 524
 525 /* Add a language to the set known by GDB (at initialization time).  */
 526
 527 extern void add_language (const struct language_defn *);
 528
 529 extern enum language get_frame_language (void); /* In stack.c */
 530
 531 /* Check for a language-specific trampoline.  */
 532
 533 extern CORE_ADDR skip_language_trampoline (struct frame_info *, CORE_ADDR pc);
 534
 535 /* Return demangled language symbol, or NULL.  */
 536 extern char *language_demangle (const struct language_defn *current_language,
 537                                 const char *mangled, int options);
 538
 539 /* Return class name from physname, or NULL.  */
 540 extern char *language_class_name_from_physname (const struct language_defn *,
 541                                                 const char *physname);
 542
 543 /* Splitting strings into words.  */
 544 extern char *default_word_break_characters (void);
 545
 546 /* Print the index of an array element using the C99 syntax.  */
 547 extern void default_print_array_index (struct value *index_value,
 548                                        struct ui_file *stream,
 549                                        const struct value_print_options *options);
 550
 551 /* Return non-zero if TYPE should be passed (and returned) by
 552    reference at the language level.  */
 553 int language_pass_by_reference (struct type *type);
 554
 555 /* Return zero; by default, types are passed by value at the language
 556    level.  The target ABI may pass or return some structs by reference
 557    independent of this.  */
 558 int default_pass_by_reference (struct type *type);
 559
 560 /* The default implementation of la_print_typedef.  */
 561 void default_print_typedef (struct type *type, struct symbol *new_symbol,
 562                             struct ui_file *stream);
 563
 564 void default_get_string (struct value *value, gdb_byte **buffer, int *length,
 565                          struct type **char_type, const char **charset);
 566
 567 void c_get_string (struct value *value, gdb_byte **buffer, int *length,
 568                    struct type **char_type, const char **charset);
 569
 570 #endif /* defined (LANGUAGE_H) */