src/gram.h

   1 /* Data definitions for internal representation of bison's input,
   2    Copyright (C) 1984, 1986, 1989, 1992 Free Software Foundation, Inc.
   3
   4 This file is part of Bison, the GNU Compiler Compiler.
   5
   6 Bison is free software; you can redistribute it and/or modify
   7 it under the terms of the GNU General Public License as published by
   8 the Free Software Foundation; either version 2, or (at your option)
   9 any later version.
  10
  11 Bison is distributed in the hope that it will be useful,
  12 but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 GNU General Public License for more details.
  15
  16 You should have received a copy of the GNU General Public License
  17 along with Bison; see the file COPYING.  If not, write to
  18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19 Boston, MA 02111-1307, USA.  */
  20
  21
  22 /* representation of the grammar rules:
  23
  24 ntokens is the number of tokens, and nvars is the number of variables
  25 (nonterminals).  nsyms is the total number, ntokens + nvars.
  26
  27         (the true number of token values assigned is ntokens
  28         reduced by one for each alias declaration)
  29
  30 Each symbol (either token or variable) receives a symbol number.
  31 Numbers 0 to ntokens-1 are for tokens, and ntokens to nsyms-1 are for
  32 variables.  Symbol number zero is the end-of-input token.  This token
  33 is counted in ntokens.
  34
  35 The rules receive rule numbers 1 to nrules in the order they are written.
  36 Actions and guards are accessed via the rule number.
  37
  38 The rules themselves are described by three arrays: rrhs, rlhs and
  39 ritem.  rlhs[R] is the symbol number of the left hand side of rule R.
  40 The right hand side is stored as symbol numbers in a portion of
  41 ritem.  rrhs[R] contains the index in ritem of the beginning of the
  42 portion for rule R.
  43
  44 If rlhs[R] is -1, the rule has been thrown out by reduce.c
  45 and should be ignored.
  46
  47 The length of the portion is one greater
  48  than the number of symbols in the rule's right hand side.
  49 The last element in the portion contains minus R, which
  50 identifies it as the end of a portion and says which rule it is for.
  51
  52 The portions of ritem come in order of increasing rule number and are
  53 followed by an element which is zero to mark the end.  nitems is the
  54 total length of ritem, not counting the final zero.  Each element of
  55 ritem is called an "item" and its index in ritem is an item number.
  56
  57 Item numbers are used in the finite state machine to represent
  58 places that parsing can get to.
  59
  60 Precedence levels are recorded in the vectors sprec and rprec.
  61 sprec records the precedence level of each symbol,
  62 rprec the precedence level of each rule.
  63 rprecsym is the symbol-number of the symbol in %prec for this rule (if any).
  64
  65 Precedence levels are assigned in increasing order starting with 1 so
  66 that numerically higher precedence values mean tighter binding as they
  67 ought to.  Zero as a symbol or rule's precedence means none is
  68 assigned.
  69
  70 Associativities are recorded similarly in rassoc and sassoc.  */
  71
  72
  73 #define ISTOKEN(s)      ((s) < ntokens)
  74 #define ISVAR(s)        ((s) >= ntokens)
  75
  76
  77 extern int nitems;
  78 extern int nrules;
  79 extern int nsyms;
  80 extern int ntokens;
  81 extern int nvars;
  82
  83 extern short *ritem;
  84 extern short *rlhs;
  85 extern short *rrhs;
  86 extern short *rprec;
  87 extern short *rprecsym;
  88 extern short *sprec;
  89 extern short *rassoc;
  90 extern short *sassoc;
  91 extern short *rline;            /* Source line number of each rule */
  92
  93 extern int start_symbol;
  94
  95
  96 /* associativity values in elements of rassoc, sassoc.  */
  97
  98 #define RIGHT_ASSOC 1
  99 #define LEFT_ASSOC 2
 100 #define NON_ASSOC 3
 101
 102 /* token translation table:
 103 indexed by a token number as returned by the user's yylex routine,
 104 it yields the internal token number used by the parser and throughout bison.
 105 If translations is zero, the translation table is not used because
 106 the two kinds of token numbers are the same.
 107 (It is noted in reader.c that "Nowadays translations is always set to 1...")
 108 */
 109
 110 extern short *token_translations;
 111 extern int translations;
 112 extern int max_user_token_number;
 113
 114 /* semantic_parser is nonzero if the input file says to use the hairy parser
 115 that provides for semantic error recovery.  If it is zero, the yacc-compatible
 116 simplified parser is used.  */
 117
 118 extern int semantic_parser;
 119
 120 /* pure_parser is nonzero if should generate a parser that is all pure and reentrant. */
 121
 122 extern int pure_parser;
 123
 124 /* error_token_number is the token number of the error token.  */
 125
 126 extern int error_token_number;