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 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             /**< Non-terminal. */
  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 gramar (rules) but not defined (error).  */
  72     needed,
  73     /** Defined with %type or %token (good).  */
  74     declared,
  75   } 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_location;
 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_location;
 138   int prec;
 139   assoc assoc;
 140   int user_token_number;
 141
 142   symbol_class class;
 143   status status;
 144 };
 145
 146 /** Undefined user number.  */
 147 # define USER_NUMBER_UNDEFINED -1
 148
 149 /* Undefined internal token number.  */
 150 # define NUMBER_UNDEFINED (-1)
 151
 152 /** Fetch (or create) the symbol associated to KEY.  */
 153 symbol *symbol_from_uniqstr (const uniqstr key, location loc);
 154
 155 /** Fetch (or create) the symbol associated to KEY.  */
 156 symbol *symbol_get (const char *key, location loc);
 157
 158 /** Generate a dummy nonterminal.
 159
 160    Its name cannot conflict with the user's names.  */
 161 symbol *dummy_symbol_get (location loc);
 162
 163
 164 /*--------------------.
 165 | Methods on symbol.  |
 166 `--------------------*/
 167
 168 /** Print a symbol (for debugging). */
 169 void symbol_print (symbol const *s, FILE *f);
 170
 171 /** Is this a dummy nonterminal?  */
 172 bool symbol_is_dummy (const symbol *sym);
 173
 174 /** The name of the code_props type: "\%destructor" or "\%printer".  */
 175 char const *code_props_type_string (code_props_type kind);
 176
 177 /** The name of the symbol that can be used as an identifier.
 178  ** Consider the alias if needed.
 179  ** Return 0 if there is none (e.g., the symbol is only defined as
 180  ** a string). */
 181 uniqstr symbol_id_get (symbol const *sym);
 182
 183 /**
 184  * Make \c str the literal string alias of \c sym.  Copy token number,
 185  * symbol number, and type from \c sym to \c str.
 186  */
 187 void symbol_make_alias (symbol *sym, symbol *str, location loc);
 188
 189 /**
 190  * This symbol is used as the lhs of a rule.  Record this location
 191  * as definition point, if not already done.
 192  */
 193 void symbol_location_as_lhs_set (symbol *sym, location loc);
 194
 195 /** Set the \c type_name associated with \c sym.
 196
 197     Do nothing if passed 0 as \c type_name.  */
 198 void symbol_type_set (symbol *sym, uniqstr type_name, location loc);
 199
 200 /** Set the \c \%destructor or \c \%printer associated with \c sym.  */
 201 void symbol_code_props_set (symbol *sym, code_props_type kind,
 202                             code_props const *destructor);
 203
 204 /** Get the computed \c \%destructor or \c %printer for \c sym, which was
 205     initialized with \c code_props_none_init if there's no \c \%destructor or
 206     \c %printer.  */
 207 code_props *symbol_code_props_get (symbol *sym, code_props_type kind);
 208
 209 /** Set the \c precedence associated with \c sym.
 210
 211     Ensure that \a symbol is a terminal.
 212     Do nothing if invoked with \c undef_assoc as \c assoc.  */
 213 void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 214
 215 /** Set the \c class associated with \c sym.  */
 216 void symbol_class_set (symbol *sym, symbol_class class, location loc,
 217                        bool declaring);
 218
 219 /** Set the \c user_token_number associated with \c sym.  */
 220 void symbol_user_token_number_set (symbol *sym, int user_number, location loc);
 221
 222
 223
 224 /*------------------.
 225 | Special symbols.  |
 226 `------------------*/
 227
 228 /** The error token. */
 229 extern symbol *errtoken;
 230 /** The token for unknown tokens.  */
 231 extern symbol *undeftoken;
 232 /** The end of input token.  */
 233 extern symbol *endtoken;
 234 /** The genuine start symbol.
 235
 236    $accept: start-symbol $end */
 237 extern symbol *accept;
 238
 239 /** The user start symbol. */
 240 extern symbol *startsymbol;
 241 /** The location of the \c \%start declaration.  */
 242 extern location startsymbol_location;
 243
 244
 245
 246 /*-------------------.
 247 | Symbol Relations.  |
 248 `-------------------*/
 249
 250 /* The symbol relations are represented by a directed graph. */
 251
 252 /* The id of a node */
 253 typedef int graphid;
 254
 255 typedef struct symgraphlink symgraphlink;
 256
 257 struct symgraphlink
 258 {
 259   /** The second \c symbol or group of a precedence relation.
 260    * See \c symgraph. */
 261   graphid id;
 262
 263   symgraphlink *next;
 264 };
 265
 266 /* Symbol precedence graph, to store the used precedence relations between
 267  * symbols. */
 268
 269 typedef struct symgraph symgraph;
 270
 271 struct symgraph
 272 {
 273   /** Identifier for the node: equal to the number of the symbol. */
 274   graphid id;
 275
 276   /** The list of related symbols that have a smaller precedence. */
 277   symgraphlink *succ;
 278
 279   /** The list of related symbols that have a greater precedence. */
 280   symgraphlink *pred;
 281 };
 282
 283 /** Register a new precedence relation as used. */
 284
 285 void register_precedence (graphid first, graphid snd);
 286
 287 /** Print a warning for each symbol whose precedence and/or associativity
 288  * is useless. */
 289
 290 void print_precedence_warnings (void);
 291
 292 /*----------------------.
 293 | Symbol associativity  |
 294 `----------------------*/
 295
 296 void register_assoc (graphid i, graphid j);
 297
 298 /*-----------------.
 299 | Semantic types.  |
 300 `-----------------*/
 301
 302 /** A semantic type and its associated \c \%destructor and \c \%printer.
 303
 304    Access the fields of this struct only through the interface functions in
 305    this file.  \sa symbol::destructor  */
 306 typedef struct {
 307   /** The key, name of the semantic type.  */
 308   uniqstr tag;
 309
 310   /** The location of its first occurence.  */
 311   location location;
 312
 313   /** Its status : "undeclared", "used" or "declared".
 314       It cannot be "needed".  */
 315   status status;
 316
 317   /** Any \c %destructor and %printer declared for this
 318       semantic type.  */
 319   code_props props[CODE_PROPS_SIZE];
 320
 321 } semantic_type;
 322
 323 /** Fetch (or create) the semantic type associated to KEY.  */
 324 semantic_type *semantic_type_from_uniqstr (const uniqstr key,
 325                                            const location *loc);
 326
 327 /** Fetch (or create) the semantic type associated to KEY.  */
 328 semantic_type *semantic_type_get (const char *key, const location *loc);
 329
 330 /** Set the \c destructor or \c printer associated with \c type.  */
 331 void semantic_type_code_props_set (semantic_type *type,
 332                                    code_props_type kind,
 333                                    code_props const *code);
 334
 335 /*----------------------------------.
 336 | Symbol and semantic type tables.  |
 337 `----------------------------------*/
 338
 339 /** Create the symbol and semantic type tables.  */
 340 void symbols_new (void);
 341
 342 /** Free all the memory allocated for symbols and semantic types.  */
 343 void symbols_free (void);
 344
 345 /** Check that all the symbols are defined.
 346
 347     Report any undefined symbols and consider them nonterminals.  */
 348 void symbols_check_defined (void);
 349
 350 /** Sanity checks and #token_translations construction.
 351
 352    Perform various sanity checks, assign symbol numbers, and set up
 353    #token_translations.  */
 354 void symbols_pack (void);
 355
 356 #endif /* !SYMTAB_H_ */