src/symtab.h

   1 /* Definitions for symtab.c and callers, part of Bison.
   2
   3    Copyright (C) 1984, 1989, 1992, 2000-2002, 2004-2015 Free Software
   4    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
  54 /* Declaration status of a symbol.
  55
  56    First, it is "undeclared".  Then, if "undeclared" and used in a
  57    %printer/%destructor, it is "used".  If not "declared" but used in
  58    a rule, it is "needed".  Finally, if declared (via a rule for
  59    nonterminals, or %token), it is "declared".
  60
  61    When status are checked at the end, "declared" symbols are fine,
  62    "used" symbols trigger warnings, otherwise it's an error.  */
  63
  64 typedef enum
  65   {
  66     /** Used in the input file for an unknown reason (error).  */
  67     undeclared,
  68     /** Used by %destructor/%printer but not defined (warning).  */
  69     used,
  70     /** Used in the gramar (rules) but not defined (error).  */
  71     needed,
  72     /** Defined with %type or %token (good).  */
  73     declared,
  74   } status;
  75
  76 typedef enum code_props_type code_props_type;
  77 enum code_props_type
  78   {
  79     destructor = 0,
  80     printer = 1,
  81   };
  82
  83 enum { CODE_PROPS_SIZE = 2 };
  84
  85 /* When extending this structure, be sure to complete
  86    symbol_check_alias_consistency.  */
  87 struct symbol
  88 {
  89   /** The key, name of the symbol.  */
  90   uniqstr tag;
  91   /** The location of its first occurrence.  */
  92   location location;
  93
  94   /** Its \c \%type.
  95
  96       Beware that this is the type_name as was entered by the user,
  97       including silly things such as "]" if she entered "%token <]> t".
  98       Therefore, when outputting type_name to M4, be sure to escape it
  99       into "@}".  See quoted_output for instance.  */
 100   uniqstr type_name;
 101
 102   /** Its \c \%type's location.  */
 103   location type_location;
 104
 105   /** Any \c \%destructor (resp. \%printer) declared specificially for this
 106       symbol.
 107
 108       Access this field only through <tt>symbol</tt>'s interface functions. For
 109       example, if <tt>symbol::destructor = NULL</tt> (resp. <tt>symbol::printer
 110       = NULL</tt>), a default \c \%destructor (resp. \%printer) or a per-type
 111       \c symbol_destructor_printer_get will compute the correct one. */
 112   code_props props[CODE_PROPS_SIZE];
 113
 114   symbol_number number;
 115   location prec_location;
 116   int prec;
 117   assoc assoc;
 118   int user_token_number;
 119
 120   /* Points to the other in the symbol-string pair for an alias.
 121      Special value USER_NUMBER_HAS_STRING_ALIAS in the symbol half of the
 122      symbol-string pair for an alias.  */
 123   symbol *alias;
 124   symbol_class class;
 125   status status;
 126 };
 127
 128 /** Undefined user number.  */
 129 # define USER_NUMBER_UNDEFINED -1
 130
 131 /* 'symbol->user_token_number == USER_NUMBER_HAS_STRING_ALIAS' means
 132    this symbol has a literal string alias.  For instance, '%token foo
 133    "foo"' has '"foo"' numbered regularly, and 'foo' numbered as
 134    USER_NUMBER_HAS_STRING_ALIAS.  */
 135 # define USER_NUMBER_HAS_STRING_ALIAS -9991
 136
 137 /* Undefined internal token number.  */
 138 # define NUMBER_UNDEFINED (-1)
 139
 140 /** Fetch (or create) the symbol associated to KEY.  */
 141 symbol *symbol_from_uniqstr (const uniqstr key, location loc);
 142
 143 /** Fetch (or create) the symbol associated to KEY.  */
 144 symbol *symbol_get (const char *key, location loc);
 145
 146 /** Generate a dummy nonterminal.
 147
 148    Its name cannot conflict with the user's names.  */
 149 symbol *dummy_symbol_get (location loc);
 150
 151
 152 /*--------------------.
 153 | Methods on symbol.  |
 154 `--------------------*/
 155
 156 /** Print a symbol (for debugging). */
 157 void symbol_print (symbol const *s, FILE *f);
 158
 159 /** Is this a dummy nonterminal?  */
 160 bool symbol_is_dummy (const symbol *sym);
 161
 162 /** The name of the code_props type: "\%destructor" or "\%printer".  */
 163 char const *code_props_type_string (code_props_type kind);
 164
 165 /** The name of the symbol that can be used as an identifier.
 166  ** Consider the alias if needed.
 167  ** Return 0 if there is none (e.g., the symbol is only defined as
 168  ** a string). */
 169 uniqstr symbol_id_get (symbol const *sym);
 170
 171 /**
 172  * Make \c str the literal string alias of \c sym.  Copy token number,
 173  * symbol number, and type from \c sym to \c str.
 174  */
 175 void symbol_make_alias (symbol *sym, symbol *str, location loc);
 176
 177 /** Set the \c type_name associated with \c sym.
 178
 179     Do nothing if passed 0 as \c type_name.  */
 180 void symbol_type_set (symbol *sym, uniqstr type_name, location loc);
 181
 182 /** Set the \c \%destructor or \c \%printer associated with \c sym.  */
 183 void symbol_code_props_set (symbol *sym, code_props_type kind,
 184                             code_props const *destructor);
 185
 186 /** Get the computed \c \%destructor or \c %printer for \c sym, which was
 187     initialized with \c code_props_none_init if there's no \c \%destructor or
 188     \c %printer.  */
 189 code_props *symbol_code_props_get (symbol *sym, code_props_type kind);
 190
 191 /** Set the \c precedence associated with \c sym.
 192
 193     Ensure that \a symbol is a terminal.
 194     Do nothing if invoked with \c undef_assoc as \c assoc.  */
 195 void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 196
 197 /** Set the \c class associated with \c sym.  */
 198 void symbol_class_set (symbol *sym, symbol_class class, location loc,
 199                        bool declaring);
 200
 201 /** Set the \c user_token_number associated with \c sym.  */
 202 void symbol_user_token_number_set (symbol *sym, int user_number, location loc);
 203
 204
 205
 206 /*------------------.
 207 | Special symbols.  |
 208 `------------------*/
 209
 210 /** The error token. */
 211 extern symbol *errtoken;
 212 /** The token for unknown tokens.  */
 213 extern symbol *undeftoken;
 214 /** The end of input token.  */
 215 extern symbol *endtoken;
 216 /** The genuine start symbol.
 217
 218    $accept: start-symbol $end */
 219 extern symbol *accept;
 220
 221 /** The user start symbol. */
 222 extern symbol *startsymbol;
 223 /** The location of the \c \%start declaration.  */
 224 extern location startsymbol_location;
 225
 226
 227
 228 /*-------------------.
 229 | Symbol Relations.  |
 230 `-------------------*/
 231
 232 /* The symbol relations are represented by a directed graph. */
 233
 234 /* The id of a node */
 235 typedef int graphid;
 236
 237 typedef struct symgraphlink symgraphlink;
 238
 239 struct symgraphlink
 240 {
 241   /** The second \c symbol or group of a precedence relation.
 242    * See \c symgraph. */
 243   graphid id;
 244
 245   symgraphlink *next;
 246 };
 247
 248 /* Symbol precedence graph, to store the used precedence relations between
 249  * symbols. */
 250
 251 typedef struct symgraph symgraph;
 252
 253 struct symgraph
 254 {
 255   /** Identifier for the node: equal to the number of the symbol. */
 256   graphid id;
 257
 258   /** The list of related symbols that have a smaller precedence. */
 259   symgraphlink *succ;
 260
 261   /** The list of related symbols that have a greater precedence. */
 262   symgraphlink *pred;
 263 };
 264
 265 /** Register a new precedence relation as used. */
 266
 267 void register_precedence (graphid first, graphid snd);
 268
 269 /** Print a warning for each symbol whose precedence and/or associativity
 270  * is useless. */
 271
 272 void print_precedence_warnings (void);
 273
 274 /*----------------------.
 275 | Symbol associativity  |
 276 `----------------------*/
 277
 278 void register_assoc (graphid i, graphid j);
 279
 280 /*-----------------.
 281 | Semantic types.  |
 282 `-----------------*/
 283
 284 /** A semantic type and its associated \c \%destructor and \c \%printer.
 285
 286    Access the fields of this struct only through the interface functions in
 287    this file.  \sa symbol::destructor  */
 288 typedef struct {
 289   /** The key, name of the semantic type.  */
 290   uniqstr tag;
 291
 292   /** The location of its first occurence.  */
 293   location location;
 294
 295   /** Its status : "undeclared", "used" or "declared".
 296       It cannot be "needed".  */
 297   status status;
 298
 299   /** Any \c %destructor and %printer declared for this
 300       semantic type.  */
 301   code_props props[CODE_PROPS_SIZE];
 302
 303 } semantic_type;
 304
 305 /** Fetch (or create) the semantic type associated to KEY.  */
 306 semantic_type *semantic_type_from_uniqstr (const uniqstr key,
 307                                            const location *loc);
 308
 309 /** Fetch (or create) the semantic type associated to KEY.  */
 310 semantic_type *semantic_type_get (const char *key, const location *loc);
 311
 312 /** Set the \c destructor or \c printer associated with \c type.  */
 313 void semantic_type_code_props_set (semantic_type *type,
 314                                    code_props_type kind,
 315                                    code_props const *code);
 316
 317 /*----------------------------------.
 318 | Symbol and semantic type tables.  |
 319 `----------------------------------*/
 320
 321 /** Create the symbol and semantic type tables.  */
 322 void symbols_new (void);
 323
 324 /** Free all the memory allocated for symbols and semantic types.  */
 325 void symbols_free (void);
 326
 327 /** Check that all the symbols are defined.
 328
 329     Report any undefined symbols and consider them nonterminals.  */
 330 void symbols_check_defined (void);
 331
 332 /** Sanity checks and #token_translations construction.
 333
 334    Perform various sanity checks, assign symbol numbers, and set up
 335    #token_translations.  */
 336 void symbols_pack (void);
 337
 338 #endif /* !SYMTAB_H_ */