src/symtab.h

   1 /* Definitions for symtab.c and callers, part of Bison.
   2
   3    Copyright (C) 1984, 1989, 1992, 2000-2002, 2004-2015, 2018-2019 Free
   4    Software Foundation, Inc.
   5
   6    This file is part of Bison, the GNU Compiler Compiler.
   7
   8    This program is free software: you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation, either version 3 of the License, or
  11    (at your option) any later version.
  12
  13    This program is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  20
  21 /**
  22  * \file symtab.h
  23  * \brief Manipulating ::symbol.
  24  */
  25
  26 #ifndef SYMTAB_H_
  27 # define SYMTAB_H_
  28
  29 # include "assoc.h"
  30 # include "location.h"
  31 # include "scan-code.h"
  32 # include "uniqstr.h"
  33
  34 /*----------.
  35 | Symbols.  |
  36 `----------*/
  37
  38 /** Symbol classes.  */
  39 typedef enum
  40 {
  41   unknown_sym,          /**< Undefined.  */
  42   token_sym,            /**< Terminal. */
  43   nterm_sym             /**< Nonterminal. */
  44 } symbol_class;
  45
  46
  47 /** Internal token numbers. */
  48 typedef int symbol_number;
  49 # define SYMBOL_NUMBER_MAXIMUM INT_MAX
  50
  51
  52 typedef struct symbol symbol;
  53 typedef struct sym_content sym_content;
  54
  55 /* Declaration status of a symbol.
  56
  57    First, it is "undeclared".  Then, if "undeclared" and used in a
  58    %printer/%destructor, it is "used".  If not "declared" but used in
  59    a rule, it is "needed".  Finally, if declared (via a rule for
  60    nonterminals, or %token), it is "declared".
  61
  62    When status are checked at the end, "declared" symbols are fine,
  63    "used" symbols trigger warnings, otherwise it's an error.  */
  64
  65 typedef enum
  66   {
  67     /** Used in the input file for an unknown reason (error).  */
  68     undeclared,
  69     /** Used by %destructor/%printer but not defined (warning).  */
  70     used,
  71     /** Used in the grammar (rules) but not defined (error).  */
  72     needed,
  73     /** Defined with %type or %token (good).  */
  74     declared,
  75   } declaration_status;
  76
  77 enum code_props_type
  78   {
  79     destructor = 0,
  80     printer = 1,
  81   };
  82 typedef enum code_props_type code_props_type;
  83
  84 enum { CODE_PROPS_SIZE = 2 };
  85
  86 struct symbol
  87 {
  88   /** The key, name of the symbol.  */
  89   uniqstr tag;
  90
  91   /** The "defining" location.  */
  92   location location;
  93
  94   /** Whether \a location is about the first uses as left-hand side
  95       symbol of a rule (true), or simply the first occurrence (e.g.,
  96       in a %type, or as a rhs symbol of a rule).  The former type of
  97       location is more natural in error messages.  This Boolean helps
  98       moving from location of the first occurrence to first use as
  99       lhs. */
 100   bool location_of_lhs;
 101
 102   /** Points to the other in the symbol-string pair for an alias. */
 103   symbol *alias;
 104
 105   /** Whether this symbol is the alias of another or not. */
 106   bool is_alias;
 107
 108   /** All the info about the pointed symbol is there. */
 109   sym_content *content;
 110 };
 111
 112 struct sym_content
 113 {
 114   symbol *symbol;
 115
 116   /** Its \c \%type.
 117
 118       Beware that this is the type_name as was entered by the user,
 119       including silly things such as "]" if she entered "%token <]> t".
 120       Therefore, when outputting type_name to M4, be sure to escape it
 121       into "@}".  See quoted_output for instance.  */
 122   uniqstr type_name;
 123
 124   /** Its \c \%type's location.  */
 125   location type_loc;
 126
 127   /** Any \c \%destructor (resp. \%printer) declared specificially for this
 128       symbol.
 129
 130       Access this field only through <tt>symbol</tt>'s interface functions. For
 131       example, if <tt>symbol::destructor = NULL</tt> (resp. <tt>symbol::printer
 132       = NULL</tt>), a default \c \%destructor (resp. \%printer) or a per-type
 133       \c symbol_destructor_printer_get will compute the correct one. */
 134   code_props props[CODE_PROPS_SIZE];
 135
 136   symbol_number number;
 137   location prec_loc;
 138   int prec;
 139   assoc assoc;
 140
 141   /** The user specified token number.
 142
 143       E.g., %token FOO 42.*/
 144   int user_token_number;
 145
 146   symbol_class class;
 147   declaration_status status;
 148 };
 149
 150 /** Undefined user number.  */
 151 # define USER_NUMBER_UNDEFINED -1
 152
 153 /* Undefined internal token number.  */
 154 # define NUMBER_UNDEFINED (-1)
 155
 156 /** Fetch (or create) the symbol associated to KEY.  */
 157 symbol *symbol_from_uniqstr (const uniqstr key, location loc);
 158
 159 /** Fetch (or create) the symbol associated to KEY.  */
 160 symbol *symbol_get (const char *key, location loc);
 161
 162 /** Generate a dummy nonterminal.
 163
 164    Its name cannot conflict with the user's names.  */
 165 symbol *dummy_symbol_get (location loc);
 166
 167
 168 /*--------------------.
 169 | Methods on symbol.  |
 170 `--------------------*/
 171
 172 /** Print a symbol (for debugging). */
 173 void symbol_print (symbol const *s, FILE *f);
 174
 175 /** Is this a dummy nonterminal?  */
 176 bool symbol_is_dummy (const symbol *sym);
 177
 178 /** The name of the code_props type: "\%destructor" or "\%printer".  */
 179 char const *code_props_type_string (code_props_type kind);
 180
 181 /** The name of the symbol that can be used as an identifier.
 182  ** Consider the alias if needed.
 183  ** Return 0 if there is none (e.g., the symbol is only defined as
 184  ** a string). */
 185 uniqstr symbol_id_get (symbol const *sym);
 186
 187 /**
 188  * Make \c str the literal string alias of \c sym.  Copy token number,
 189  * symbol number, and type from \c sym to \c str.
 190  */
 191 void symbol_make_alias (symbol *sym, symbol *str, location loc);
 192
 193 /**
 194  * This symbol is used as the lhs of a rule.  Record this location
 195  * as definition point, if not already done.
 196  */
 197 void symbol_location_as_lhs_set (symbol *sym, location loc);
 198
 199 /** Set the \c type_name associated with \c sym.
 200
 201     Do nothing if passed 0 as \c type_name.  */
 202 void symbol_type_set (symbol *sym, uniqstr type_name, location loc);
 203
 204 /** Set the \c \%destructor or \c \%printer associated with \c sym.  */
 205 void symbol_code_props_set (symbol *sym, code_props_type kind,
 206                             code_props const *destructor);
 207
 208 /** Get the computed \c \%destructor or \c %printer for \c sym, which was
 209     initialized with \c code_props_none_init if there's no \c \%destructor or
 210     \c %printer.  */
 211 code_props *symbol_code_props_get (symbol *sym, code_props_type kind);
 212
 213 /** Set the \c precedence associated with \c sym.
 214
 215     Ensure that \a symbol is a terminal.
 216     Do nothing if invoked with \c undef_assoc as \c assoc.  */
 217 void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 218
 219 /** Set the \c class associated with \c sym.
 220
 221     Whether \c declaring means whether this class definition comes
 222     from %nterm or %token.  */
 223 void symbol_class_set (symbol *sym, symbol_class class, location loc,
 224                        bool declaring);
 225
 226 /** Set the \c user_token_number associated with \c sym.  */
 227 void symbol_user_token_number_set (symbol *sym, int user_number, location loc);
 228
 229
 230
 231 /*------------------.
 232 | Special symbols.  |
 233 `------------------*/
 234
 235 /** The error token. */
 236 extern symbol *errtoken;
 237 /** The token for unknown tokens.  */
 238 extern symbol *undeftoken;
 239 /** The end of input token.  */
 240 extern symbol *endtoken;
 241 /** The genuine start symbol.
 242
 243    $accept: start-symbol $end */
 244 extern symbol *accept;
 245
 246 /** The user start symbol. */
 247 extern symbol *startsymbol;
 248 /** The location of the \c \%start declaration.  */
 249 extern location startsymbol_loc;
 250
 251 /** Whether a symbol declared with a type tag.  */
 252 extern bool tag_seen;
 253
 254
 255 /*-------------------.
 256 | Symbol Relations.  |
 257 `-------------------*/
 258
 259 /* The symbol relations are represented by a directed graph. */
 260
 261 /* The id of a node */
 262 typedef int graphid;
 263
 264 typedef struct symgraphlink symgraphlink;
 265
 266 struct symgraphlink
 267 {
 268   /** The second \c symbol or group of a precedence relation.
 269    * See \c symgraph. */
 270   graphid id;
 271
 272   symgraphlink *next;
 273 };
 274
 275 /* Symbol precedence graph, to store the used precedence relations between
 276  * symbols. */
 277
 278 typedef struct symgraph symgraph;
 279
 280 struct symgraph
 281 {
 282   /** Identifier for the node: equal to the number of the symbol. */
 283   graphid id;
 284
 285   /** The list of related symbols that have a smaller precedence. */
 286   symgraphlink *succ;
 287
 288   /** The list of related symbols that have a greater precedence. */
 289   symgraphlink *pred;
 290 };
 291
 292 /** Register a new precedence relation as used. */
 293
 294 void register_precedence (graphid first, graphid snd);
 295
 296 /** Print a warning for each symbol whose precedence and/or associativity
 297  * is useless. */
 298
 299 void print_precedence_warnings (void);
 300
 301 /*----------------------.
 302 | Symbol associativity  |
 303 `----------------------*/
 304
 305 void register_assoc (graphid i, graphid j);
 306
 307 /*-----------------.
 308 | Semantic types.  |
 309 `-----------------*/
 310
 311 /** A semantic type and its associated \c \%destructor and \c \%printer.
 312
 313    Access the fields of this struct only through the interface functions in
 314    this file.  \sa symbol::destructor  */
 315 typedef struct {
 316   /** The key, name of the semantic type.  */
 317   uniqstr tag;
 318
 319   /** The location of its first occurrence.  */
 320   location location;
 321
 322   /** Its status : "undeclared", "used" or "declared".
 323       It cannot be "needed".  */
 324   declaration_status status;
 325
 326   /** Any \c %destructor and %printer declared for this
 327       semantic type.  */
 328   code_props props[CODE_PROPS_SIZE];
 329
 330 } semantic_type;
 331
 332 /** Fetch (or create) the semantic type associated to KEY.  */
 333 semantic_type *semantic_type_from_uniqstr (const uniqstr key,
 334                                            const location *loc);
 335
 336 /** Fetch (or create) the semantic type associated to KEY.  */
 337 semantic_type *semantic_type_get (const char *key, const location *loc);
 338
 339 /** Set the \c destructor or \c printer associated with \c type.  */
 340 void semantic_type_code_props_set (semantic_type *type,
 341                                    code_props_type kind,
 342                                    code_props const *code);
 343
 344 /*----------------------------------.
 345 | Symbol and semantic type tables.  |
 346 `----------------------------------*/
 347
 348 /** Create the symbol and semantic type tables.  */
 349 void symbols_new (void);
 350
 351 /** Free all the memory allocated for symbols and semantic types.  */
 352 void symbols_free (void);
 353
 354 /** Check that all the symbols are defined.
 355
 356     Report any undefined symbols and consider them nonterminals.  */
 357 void symbols_check_defined (void);
 358
 359 /** Sanity checks and #token_translations construction.
 360
 361    Perform various sanity checks, assign symbol numbers, and set up
 362    #token_translations.  */
 363 void symbols_pack (void);
 364
 365 #endif /* !SYMTAB_H_ */