]> git.saurik.com Git - bison.git/blob - src/gram.h
* src/files.c (open_files): Yipee! We no longer need all the code
[bison.git] / src / gram.h
1 /* Data definitions for internal representation of bison's input,
2 Copyright 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 #ifndef GRAM_H_
22 # define GRAM_H_
23
24 /* representation of the grammar rules:
25
26 ntokens is the number of tokens, and nvars is the number of
27 variables (nonterminals). nsyms is the total number, ntokens +
28 nvars.
29
30 (the true number of token values assigned is ntokens reduced by one
31 for each alias declaration)
32
33 Each symbol (either token or variable) receives a symbol number.
34 Numbers 0 to ntokens-1 are for tokens, and ntokens to nsyms-1 are
35 for variables. Symbol number zero is the end-of-input token. This
36 token is counted in ntokens.
37
38 The rules receive rule numbers 1 to nrules in the order they are
39 written. Actions and guards are accessed via the rule number.
40
41 The rules themselves are described by three arrays: rrhs, rlhs and
42 ritem. rlhs[R] is the symbol number of the left hand side of rule
43 R. The right hand side is stored as symbol numbers in a portion of
44 ritem. rrhs[R] contains the index in ritem of the beginning of the
45 portion for rule R.
46
47 If rlhs[R] is -1, the rule has been thrown out by reduce.c and
48 should be ignored.
49
50 The length of the portion is one greater than the number of symbols
51 in the rule's right hand side. The last element in the portion
52 contains minus R, which identifies it as the end of a portion and
53 says which rule it is for.
54
55 The portions of ritem come in order of increasing rule number and
56 are followed by an element which is zero to mark the end. nitems
57 is the total length of ritem, not counting the final zero. Each
58 element of ritem is called an "item" and its index in ritem is an
59 item number.
60
61 Item numbers are used in the finite state machine to represent
62 places that parsing can get to.
63
64 Precedence levels are recorded in the vectors sprec and rprec.
65 sprec records the precedence level of each symbol, rprec the
66 precedence level of each rule. rprecsym is the symbol-number of
67 the symbol in %prec for this rule (if any).
68
69 Precedence levels are assigned in increasing order starting with 1
70 so that numerically higher precedence values mean tighter binding
71 as they ought to. Zero as a symbol or rule's precedence means none
72 is assigned.
73
74 Associativities are recorded similarly in rassoc and sassoc. */
75
76
77 #define ISTOKEN(s) ((s) < ntokens)
78 #define ISVAR(s) ((s) >= ntokens)
79
80 extern int nitems;
81 extern int nrules;
82 extern int nsyms;
83 extern int ntokens;
84 extern int nvars;
85
86 extern short *ritem;
87 extern short *rlhs;
88 extern short *rrhs;
89 extern short *rprec;
90 extern short *rprecsym;
91 extern short *sprec;
92 extern short *rassoc;
93 extern short *sassoc;
94 extern short *rline; /* Source line number of each rule */
95
96 extern int start_symbol;
97
98
99 /* associativity values in elements of rassoc, sassoc. */
100 typedef enum
101 {
102 right_assoc,
103 left_assoc,
104 non_assoc
105 } associativity;
106
107
108 /* token translation table: indexed by a token number as returned by
109 the user's yylex routine, it yields the internal token number used
110 by the parser and throughout bison. If translations is zero, the
111 translation table is not used because the two kinds of token
112 numbers are the same. (It is noted in reader.c that "Nowadays
113 translations is always set to 1...") */
114
115 extern short *token_translations;
116 extern int translations;
117 extern int max_user_token_number;
118
119 /* SEMANTIC_PARSER is nonzero if the input file says to use the hairy
120 parser that provides for semantic error recovery. If it is zero,
121 the yacc-compatible simplified parser is used. */
122
123 extern int semantic_parser;
124
125 /* PURE_PARSER is nonzero if should generate a parser that is all pure
126 and reentrant. */
127
128 extern int pure_parser;
129
130 /* ERROR_TOKEN_NUMBER is the token number of the error token. */
131
132 extern int error_token_number;
133 #endif /* !GRAM_H_ */