blob: 9782666fd82450dfcd2f9e197f7cd9cd2c9c20cd [file] [log] [blame]
Guido van Rossuma3309961993-07-28 09:05:47 +00001#ifndef Py_TOKENIZER_H
2#define Py_TOKENIZER_H
3#ifdef __cplusplus
4extern "C" {
5#endif
6
Martin v. Löwis00f1e3f2002-08-04 17:29:52 +00007#include "object.h"
Guido van Rossumf70e43a1991-02-19 12:39:46 +00008
Guido van Rossum85a5fbb1990-10-14 12:07:46 +00009/* Tokenizer interface */
10
11#include "token.h" /* For token types */
12
13#define MAXINDENT 100 /* Max indentation level */
14
15/* Tokenizer state */
16struct tok_state {
17 /* Input state; buf <= cur <= inp <= end */
Guido van Rossum6ac258d1993-05-12 08:24:20 +000018 /* NB an entire line is held in the buffer */
19 char *buf; /* Input buffer, or NULL; malloc'ed if fp != NULL */
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000020 char *cur; /* Next character in buffer */
21 char *inp; /* End of data in buffer */
Guido van Rossum6ac258d1993-05-12 08:24:20 +000022 char *end; /* End of input buffer if buf != NULL */
Guido van Rossum1d5735e1994-08-30 08:27:36 +000023 char *start; /* Start of current token if not NULL */
Guido van Rossum6ac258d1993-05-12 08:24:20 +000024 int done; /* E_OK normally, E_EOF at EOF, otherwise error code */
25 /* NB If done != E_OK, cur must be == inp!!! */
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000026 FILE *fp; /* Rest of input; NULL if tokenizing a string */
27 int tabsize; /* Tab spacing */
28 int indent; /* Current indentation index */
29 int indstack[MAXINDENT]; /* Stack of indents */
30 int atbol; /* Nonzero if at begin of new line */
31 int pendin; /* Pending indents (if > 0) or dedents (if < 0) */
32 char *prompt, *nextprompt; /* For interactive prompting */
33 int lineno; /* Current line number */
Guido van Rossuma849b831993-05-12 11:35:44 +000034 int level; /* () [] {} Parentheses nesting level */
35 /* Used to allow free continuations inside them */
Guido van Rossum926f13a1998-04-09 21:38:06 +000036 /* Stuff for checking on different tab sizes */
37 char *filename; /* For error messages */
38 int altwarning; /* Issue warning if alternate tabs don't match */
39 int alterror; /* Issue error if alternate tabs don't match */
40 int alttabsize; /* Alternate tab spacing */
41 int altindstack[MAXINDENT]; /* Stack of alternate indents */
Martin v. Löwis00f1e3f2002-08-04 17:29:52 +000042 /* Stuff for PEP 0263 */
43 int decoding_state; /* -1:decoding, 0:init, 1:raw */
44 int decoding_erred; /* whether erred in decoding */
45 int read_coding_spec; /* whether 'coding:...' has been read */
46 int issued_encoding_warning; /* whether non-ASCII warning was issued */
47 char *encoding;
48 PyObject *decoding_readline; /* codecs.open(...).readline */
49 PyObject *decoding_buffer;
50 const char* enc;
51 const char* str;
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000052};
53
Tim Petersdbd9ba62000-07-09 03:09:57 +000054extern struct tok_state *PyTokenizer_FromString(char *);
55extern struct tok_state *PyTokenizer_FromFile(FILE *, char *, char *);
56extern void PyTokenizer_Free(struct tok_state *);
57extern int PyTokenizer_Get(struct tok_state *, char **, char **);
Guido van Rossuma3309961993-07-28 09:05:47 +000058
59#ifdef __cplusplus
60}
61#endif
62#endif /* !Py_TOKENIZER_H */