src/symtab.h

   1 /* Definitions for symtab.c and callers, part of Bison.
   2
   3    Copyright (C) 1984, 1989, 1992, 2000-2002, 2004-2015, 2018-2021 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 <https://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   /** Undefined.  */
  42   unknown_sym,
  43   /** Declared with %type: same as Undefined, but triggered a Wyacc if
  44       applied to a terminal. */
  45   pct_type_sym,
  46   /** Terminal. */
  47   token_sym,
  48   /** Nonterminal. */
  49   nterm_sym
  50 } symbol_class;
  51
  52
  53 /** Internal token numbers. */
  54 typedef int symbol_number;
  55 # define SYMBOL_NUMBER_MAXIMUM INT_MAX
  56
  57
  58 typedef struct symbol symbol;
  59 typedef struct sym_content sym_content;
  60
  61 /* Declaration status of a symbol.
  62
  63    First, it is "undeclared".  Then, if "undeclared" and used in a
  64    %printer/%destructor, it is "used".  If not "declared" but used in
  65    a rule, it is "needed".  Finally, if declared (via a rule for
  66    nonterminals, or %token), it is "declared".
  67
  68    When status are checked at the end, "declared" symbols are fine,
  69    "used" symbols trigger warnings, otherwise it's an error.  */
  70
  71 typedef enum
  72   {
  73     /** Used in the input file for an unknown reason (error).  */
  74     undeclared,
  75     /** Used by %destructor/%printer but not defined (warning).  */
  76     used,
  77     /** Used in the grammar (rules) but not defined (error).  */
  78     needed,
  79     /** Defined with %type or %token (good).  */
  80     declared,
  81   } declaration_status;
  82
  83 enum code_props_type
  84   {
  85     destructor = 0,
  86     printer = 1,
  87   };
  88 typedef enum code_props_type code_props_type;
  89
  90 enum { CODE_PROPS_SIZE = 2 };
  91
  92 struct symbol
  93 {
  94   /** The key, name of the symbol.  */
  95   uniqstr tag;
  96
  97   /** The "defining" location.  */
  98   location location;
  99
 100   /** Whether this symbol is translatable. */
 101   bool translatable;
 102
 103   /** Whether \a location is about the first uses as left-hand side
 104       symbol of a rule (true), or simply the first occurrence (e.g.,
 105       in a %type, or as a rhs symbol of a rule).  The former type of
 106       location is more natural in error messages.  This Boolean helps
 107       moving from location of the first occurrence to first use as
 108       lhs. */
 109   bool location_of_lhs;
 110
 111   /** Points to the other in the symbol-string pair for an alias. */
 112   symbol *alias;
 113
 114   /** Whether this symbol is the alias of another or not. */
 115   bool is_alias;
 116
 117   /** All the info about the pointed symbol is there. */
 118   sym_content *content;
 119 };
 120
 121 struct sym_content
 122 {
 123   /** The main symbol that denotes this content (it contains the
 124       possible alias). */
 125   symbol *symbol;
 126
 127   /** Its \c \%type.
 128
 129       Beware that this is the type_name as was entered by the user,
 130       including silly things such as "]" if she entered "%token <]> t".
 131       Therefore, when outputting type_name to M4, be sure to escape it
 132       into "@}".  See quoted_output for instance.  */
 133   uniqstr type_name;
 134
 135   /** Its \c \%type's location.  */
 136   location type_loc;
 137
 138   /** Any \c \%destructor (resp. \%printer) declared specifically for this
 139       symbol.
 140
 141       Access this field only through <tt>symbol</tt>'s interface functions. For
 142       example, if <tt>symbol::destructor = NULL</tt> (resp. <tt>symbol::printer
 143       = NULL</tt>), a default \c \%destructor (resp. \%printer) or a per-type
 144       \c symbol_destructor_printer_get will compute the correct one. */
 145   code_props props[CODE_PROPS_SIZE];
 146
 147   symbol_number number;
 148   location prec_loc;
 149   int prec;
 150   assoc assoc;
 151
 152   /** Token code, possibly specified by the user (%token FOO 42).  */
 153   int code;
 154
 155   symbol_class class;
 156   declaration_status status;
 157 };
 158
 159 /** Fetch (or create) the symbol associated to KEY.  */
 160 symbol *symbol_from_uniqstr (const uniqstr key, location loc);
 161
 162 /** Fetch (or create) the symbol associated to KEY.  */
 163 symbol *symbol_get (const char *key, location loc);
 164
 165 /** Generate a dummy nonterminal.
 166
 167    Its name cannot conflict with the user's names.  */
 168 symbol *dummy_symbol_get (location loc);
 169
 170
 171 /*--------------------.
 172 | Methods on symbol.  |
 173 `--------------------*/
 174
 175 /** Print a symbol (for debugging). */
 176 void symbol_print (symbol const *s, FILE *f);
 177
 178 /** Is this a dummy nonterminal?  */
 179 bool symbol_is_dummy (symbol const *sym);
 180
 181 /** The name of the code_props type: "\%destructor" or "\%printer".  */
 182 char const *code_props_type_string (code_props_type kind);
 183
 184 /** The name of the symbol that can be used as an identifier.
 185  ** Consider the alias if needed.
 186  ** Return 0 if there is none (e.g., the symbol is only defined as
 187  ** a string). */
 188 uniqstr symbol_id_get (symbol const *sym);
 189
 190 /**
 191  * Make \c str the literal string alias of \c sym.  Copy token number,
 192  * symbol number, and type from \c sym to \c str.
 193  */
 194 void symbol_make_alias (symbol *sym, symbol *str, location loc);
 195
 196 /**
 197  * This symbol is used as the lhs of a rule.  Record this location
 198  * as definition point, if not already done.
 199  */
 200 void symbol_location_as_lhs_set (symbol *sym, location loc);
 201
 202 /** Set the \c type_name associated with \c sym.
 203
 204     Do nothing if passed 0 as \c type_name.  */
 205 void symbol_type_set (symbol *sym, uniqstr type_name, location loc);
 206
 207 /** Set the \c \%destructor or \c \%printer associated with \c sym.  */
 208 void symbol_code_props_set (symbol *sym, code_props_type kind,
 209                             code_props const *destructor);
 210
 211 /** Get the computed \c \%destructor or \c %printer for \c sym, which was
 212     initialized with \c code_props_none_init if there's no \c \%destructor or
 213     \c %printer.  */
 214 code_props *symbol_code_props_get (symbol *sym, code_props_type kind);
 215
 216 /** Set the \c precedence associated with \c sym.
 217
 218     Ensure that \a symbol is a terminal.
 219     Do nothing if invoked with \c undef_assoc as \c assoc.  */
 220 void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 221
 222 /** Set the \c class associated with \c sym.
 223
 224     Whether \c declaring means whether this class definition comes
 225     from %nterm or %token (but not %type, prec/assoc, etc.).  A symbol
 226     can have "declaring" set only at most once.  */
 227 void symbol_class_set (symbol *sym, symbol_class class, location loc,
 228                        bool declaring);
 229
 230 /** Set the token \c code of \c sym, specified by the user at \c loc.  */
 231 void symbol_code_set (symbol *sym, int code, location loc);
 232
 233
 234
 235 /*------------------.
 236 | Special symbols.  |
 237 `------------------*/
 238
 239 /** The error token. */
 240 extern symbol *errtoken;
 241 /** The token for unknown tokens.  */
 242 extern symbol *undeftoken;
 243 /** The end of input token.  */
 244 extern symbol *eoftoken;
 245 /** The genuine start symbol.
 246
 247    $accept: start-symbol $end */
 248 extern symbol *acceptsymbol;
 249
 250 /** Whether a symbol declared with a type tag.  */
 251 extern bool tag_seen;
 252
 253
 254 /*-------------------.
 255 | Symbol Relations.  |
 256 `-------------------*/
 257
 258 /* The symbol relations are represented by a directed graph. */
 259
 260 /* The id of a node */
 261 typedef int graphid;
 262
 263 typedef struct symgraphlink symgraphlink;
 264
 265 struct symgraphlink
 266 {
 267   /** The second \c symbol or group of a precedence relation.
 268    * See \c symgraph. */
 269   graphid id;
 270
 271   symgraphlink *next;
 272 };
 273
 274 /* Symbol precedence graph, to store the used precedence relations between
 275  * symbols. */
 276
 277 typedef struct symgraph symgraph;
 278
 279 struct symgraph
 280 {
 281   /** Identifier for the node: equal to the number of the symbol. */
 282   graphid id;
 283
 284   /** The list of related symbols that have a smaller precedence. */
 285   symgraphlink *succ;
 286
 287   /** The list of related symbols that have a greater precedence. */
 288   symgraphlink *pred;
 289 };
 290
 291 /** Register a new precedence relation as used. */
 292
 293 void register_precedence (graphid first, graphid snd);
 294
 295 /** Print a warning for each symbol whose precedence and/or associativity
 296  * is useless. */
 297
 298 void print_precedence_warnings (void);
 299
 300 /*----------------------.
 301 | Symbol associativity  |
 302 `----------------------*/
 303
 304 void register_assoc (graphid i, graphid j);
 305
 306 /*-----------------.
 307 | Semantic types.  |
 308 `-----------------*/
 309
 310 /** A semantic type and its associated \c \%destructor and \c \%printer.
 311
 312    Access the fields of this struct only through the interface functions in
 313    this file.  \sa symbol::destructor  */
 314 typedef struct {
 315   /** The key, name of the semantic type.  */
 316   uniqstr tag;
 317
 318   /** The location of its first occurrence.  */
 319   location location;
 320
 321   /** Its status : "undeclared", "used" or "declared".
 322       It cannot be "needed".  */
 323   declaration_status status;
 324
 325   /** Any \c %destructor and %printer declared for this
 326       semantic type.  */
 327   code_props props[CODE_PROPS_SIZE];
 328
 329 } semantic_type;
 330
 331 /** Fetch (or create) the semantic type associated to KEY.  */
 332 semantic_type *semantic_type_from_uniqstr (const uniqstr key,
 333                                            const location *loc);
 334
 335 /** Fetch (or create) the semantic type associated to KEY.  */
 336 semantic_type *semantic_type_get (const char *key, const location *loc);
 337
 338 /** Set the \c destructor or \c printer associated with \c type.  */
 339 void semantic_type_code_props_set (semantic_type *type,
 340                                    code_props_type kind,
 341                                    code_props const *code);
 342
 343 /*----------------------------------.
 344 | Symbol and semantic type tables.  |
 345 `----------------------------------*/
 346
 347 /** Create the symbol and semantic type tables, and the built-in
 348     symbols.  */
 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_ */