db: move type_size to use raw SQL
[smatch.git] / token.h
blobfbeacb8eadf9e4eedff725da5cf47688284ba2a2
1 #ifndef TOKEN_H
2 #define TOKEN_H
3 /*
4 * Basic tokenization structures. NOTE! Those tokens had better
5 * be pretty small, since we're going to keep them all in memory
6 * indefinitely.
8 * Copyright (C) 2003 Transmeta Corp.
9 * 2003 Linus Torvalds
11 * Licensed under the Open Software License version 1.1
14 #include <sys/types.h>
15 #include "lib.h"
18 * This describes the pure lexical elements (tokens), with
19 * no semantic meaning. In other words, an identifier doesn't
20 * have a type or meaning, it is only a specific string in
21 * the input stream.
23 * Semantic meaning is handled elsewhere.
26 enum constantfile {
27 CONSTANT_FILE_MAYBE, // To be determined, not inside any #ifs in this file
28 CONSTANT_FILE_IFNDEF, // To be determined, currently inside #ifndef
29 CONSTANT_FILE_NOPE, // No
30 CONSTANT_FILE_YES // Yes
33 extern const char *includepath[];
35 struct stream {
36 int fd;
37 const char *name;
38 const char *path; // input-file path - see set_stream_include_path()
39 const char **next_path;
41 /* Use these to check for "already parsed" */
42 enum constantfile constant;
43 int dirty, next_stream;
44 struct ident *protect;
45 struct token *ifndef;
46 struct token *top_if;
49 extern int input_stream_nr;
50 extern struct stream *input_streams;
51 extern unsigned int tabstop;
52 extern int no_lineno;
53 extern int *hash_stream(const char *name);
55 struct ident {
56 struct ident *next; /* Hash chain of identifiers */
57 struct symbol *symbols; /* Pointer to semantic meaning list */
58 unsigned char len; /* Length of identifier name */
59 unsigned char tainted:1,
60 reserved:1,
61 keyword:1;
62 char name[]; /* Actual identifier */
65 enum token_type {
66 TOKEN_EOF,
67 TOKEN_ERROR,
68 TOKEN_IDENT,
69 TOKEN_ZERO_IDENT,
70 TOKEN_NUMBER,
71 TOKEN_CHAR,
72 TOKEN_WIDE_CHAR,
73 TOKEN_STRING,
74 TOKEN_WIDE_STRING,
75 TOKEN_SPECIAL,
76 TOKEN_STREAMBEGIN,
77 TOKEN_STREAMEND,
78 TOKEN_MACRO_ARGUMENT,
79 TOKEN_STR_ARGUMENT,
80 TOKEN_QUOTED_ARGUMENT,
81 TOKEN_CONCAT,
82 TOKEN_GNU_KLUDGE,
83 TOKEN_UNTAINT,
84 TOKEN_ARG_COUNT,
85 TOKEN_IF,
86 TOKEN_SKIP_GROUPS,
87 TOKEN_ELSE,
90 /* Combination tokens */
91 #define COMBINATION_STRINGS { \
92 "+=", "++", \
93 "-=", "--", "->", \
94 "*=", \
95 "/=", \
96 "%=", \
97 "<=", ">=", \
98 "==", "!=", \
99 "&&", "&=", \
100 "||", "|=", \
101 "^=", "##", \
102 "<<", ">>", "..", \
103 "<<=", ">>=", "...", \
104 "", \
105 "<", ">", "<=", ">=" \
108 extern unsigned char combinations[][4];
110 enum special_token {
111 SPECIAL_BASE = 256,
112 SPECIAL_ADD_ASSIGN = SPECIAL_BASE,
113 SPECIAL_INCREMENT,
114 SPECIAL_SUB_ASSIGN,
115 SPECIAL_DECREMENT,
116 SPECIAL_DEREFERENCE,
117 SPECIAL_MUL_ASSIGN,
118 SPECIAL_DIV_ASSIGN,
119 SPECIAL_MOD_ASSIGN,
120 SPECIAL_LTE,
121 SPECIAL_GTE,
122 SPECIAL_EQUAL,
123 SPECIAL_NOTEQUAL,
124 SPECIAL_LOGICAL_AND,
125 SPECIAL_AND_ASSIGN,
126 SPECIAL_LOGICAL_OR,
127 SPECIAL_OR_ASSIGN,
128 SPECIAL_XOR_ASSIGN,
129 SPECIAL_HASHHASH,
130 SPECIAL_LEFTSHIFT,
131 SPECIAL_RIGHTSHIFT,
132 SPECIAL_DOTDOT,
133 SPECIAL_SHL_ASSIGN,
134 SPECIAL_SHR_ASSIGN,
135 SPECIAL_ELLIPSIS,
136 SPECIAL_ARG_SEPARATOR,
137 SPECIAL_UNSIGNED_LT,
138 SPECIAL_UNSIGNED_GT,
139 SPECIAL_UNSIGNED_LTE,
140 SPECIAL_UNSIGNED_GTE,
143 struct string {
144 unsigned int length;
145 char data[];
148 /* will fit into 32 bits */
149 struct argcount {
150 unsigned normal:10;
151 unsigned quoted:10;
152 unsigned str:10;
153 unsigned vararg:1;
157 * This is a very common data structure, it should be kept
158 * as small as humanly possible. Big (rare) types go as
159 * pointers.
161 struct token {
162 struct position pos;
163 struct token *next;
164 union {
165 const char *number;
166 struct ident *ident;
167 unsigned int special;
168 struct string *string;
169 int character;
170 int argnum;
171 struct argcount count;
175 #define MAX_STRING 4095
177 static inline struct token *containing_token(struct token **p)
179 void *addr = (char *)p - ((char *)&((struct token *)0)->next - (char *)0);
180 return addr;
183 #define token_type(x) ((x)->pos.type)
186 * Last token in the stream - points to itself.
187 * This allows us to not test for NULL pointers
188 * when following the token->next chain..
190 extern struct token eof_token_entry;
191 #define eof_token(x) ((x) == &eof_token_entry)
193 extern int init_stream(const char *, int fd, const char **next_path);
194 extern const char *stream_name(int stream);
195 extern struct ident *hash_ident(struct ident *);
196 extern struct ident *built_in_ident(const char *);
197 extern struct token *built_in_token(int, const char *);
198 extern const char *show_special(int);
199 extern const char *show_ident(const struct ident *);
200 extern const char *show_string(const struct string *string);
201 extern const char *show_token(const struct token *);
202 extern struct token * tokenize(const char *, int, struct token *, const char **next_path);
203 extern struct token * tokenize_buffer(void *, unsigned long, struct token **);
205 extern void show_identifier_stats(void);
206 extern struct token *preprocess(struct token *);
208 extern void store_all_tokens(struct token *token);
209 extern struct token *pos_get_token(struct position pos);
210 extern char *pos_ident(struct position pos);
212 extern void store_macro_pos(struct token *);
213 extern char *get_macro_name(struct position pos);
215 static inline int match_op(struct token *token, int op)
217 return token->pos.type == TOKEN_SPECIAL && token->special == op;
220 static inline int match_ident(struct token *token, struct ident *id)
222 return token->pos.type == TOKEN_IDENT && token->ident == id;
225 #endif