cpython/Parser/parsetok.c

269 lines
6.4 KiB
C
Raw Normal View History

1991-02-19 08:39:46 -04:00
1990-10-14 09:07:46 -03:00
/* Parser-tokenizer link implementation */
1990-12-20 11:06:42 -04:00
#include "pgenheaders.h"
1990-10-14 09:07:46 -03:00
#include "tokenizer.h"
#include "node.h"
#include "grammar.h"
#include "parser.h"
1990-12-20 11:06:42 -04:00
#include "parsetok.h"
1990-10-14 09:07:46 -03:00
#include "errcode.h"
#include "graminit.h"
1990-10-14 09:07:46 -03:00
1998-04-10 16:35:06 -03:00
int Py_TabcheckFlag;
1990-10-14 09:07:46 -03:00
1990-12-20 11:06:42 -04:00
/* Forward */
static node *parsetok(struct tok_state *, grammar *, int, perrdetail *, int);
static void initerr(perrdetail *err_ret, const char* filename);
1990-10-14 09:07:46 -03:00
1990-12-20 11:06:42 -04:00
/* Parse input coming from a string. Return error code, print some errors. */
1994-08-29 09:25:45 -03:00
node *
PyParser_ParseString(const char *s, grammar *g, int start, perrdetail *err_ret)
{
return PyParser_ParseStringFlagsFilename(s, NULL, g, start, err_ret, 0);
}
node *
PyParser_ParseStringFlags(const char *s, grammar *g, int start,
perrdetail *err_ret, int flags)
{
return PyParser_ParseStringFlagsFilename(s, NULL,
g, start, err_ret, flags);
}
node *
PyParser_ParseStringFlagsFilename(const char *s, const char *filename,
grammar *g, int start,
perrdetail *err_ret, int flags)
1990-10-14 09:07:46 -03:00
{
1994-08-29 09:25:45 -03:00
struct tok_state *tok;
initerr(err_ret, filename);
1994-08-29 09:25:45 -03:00
1997-04-29 18:03:06 -03:00
if ((tok = PyTokenizer_FromString(s)) == NULL) {
err_ret->error = PyErr_Occurred() ? E_DECODE : E_NOMEM;
1994-08-29 09:25:45 -03:00
return NULL;
1990-10-14 09:07:46 -03:00
}
1994-08-29 09:25:45 -03:00
tok->filename = filename ? filename : "<string>";
if (Py_TabcheckFlag || Py_VerboseFlag) {
tok->altwarning = (tok->filename != NULL);
if (Py_TabcheckFlag >= 2)
tok->alterror++;
}
return parsetok(tok, g, start, err_ret, flags);
1990-10-14 09:07:46 -03:00
}
1990-12-20 11:06:42 -04:00
/* Parse input coming from a file. Return error code, print some errors. */
1990-10-14 09:07:46 -03:00
1994-08-29 09:25:45 -03:00
node *
PyParser_ParseFile(FILE *fp, const char *filename, grammar *g, int start,
char *ps1, char *ps2, perrdetail *err_ret)
{
return PyParser_ParseFileFlags(fp, filename, g, start, ps1, ps2,
err_ret, 0);
}
node *
PyParser_ParseFileFlags(FILE *fp, const char *filename, grammar *g, int start,
char *ps1, char *ps2, perrdetail *err_ret, int flags)
1990-10-14 09:07:46 -03:00
{
1994-08-29 09:25:45 -03:00
struct tok_state *tok;
initerr(err_ret, filename);
1994-08-29 09:25:45 -03:00
1997-04-29 18:03:06 -03:00
if ((tok = PyTokenizer_FromFile(fp, ps1, ps2)) == NULL) {
1994-08-29 09:25:45 -03:00
err_ret->error = E_NOMEM;
return NULL;
1990-10-14 09:07:46 -03:00
}
tok->filename = filename;
1998-04-10 16:35:06 -03:00
if (Py_TabcheckFlag || Py_VerboseFlag) {
tok->altwarning = (filename != NULL);
if (Py_TabcheckFlag >= 2)
tok->alterror++;
}
1994-08-29 09:25:45 -03:00
1990-12-20 11:06:42 -04:00
return parsetok(tok, g, start, err_ret, flags);
1994-08-29 09:25:45 -03:00
}
1990-12-20 11:06:42 -04:00
/* Parse input coming from the given tokenizer structure.
Return error code. */
static char with_msg[] =
"%s:%d: Warning: 'with' will become a reserved keyword in Python 2.6\n";
static char as_msg[] =
"%s:%d: Warning: 'as' will become a reserved keyword in Python 2.6\n";
static void
warn(const char *msg, const char *filename, int lineno)
{
if (filename == NULL)
filename = "<string>";
PySys_WriteStderr(msg, filename, lineno);
}
1994-08-29 09:25:45 -03:00
static node *
parsetok(struct tok_state *tok, grammar *g, int start, perrdetail *err_ret,
int flags)
1990-12-20 11:06:42 -04:00
{
parser_state *ps;
1994-08-29 09:25:45 -03:00
node *n;
int started = 0, handling_import = 0, handling_with = 0;
1994-08-29 09:25:45 -03:00
1997-04-29 18:03:06 -03:00
if ((ps = PyParser_New(g, start)) == NULL) {
1990-12-20 11:06:42 -04:00
fprintf(stderr, "no mem for new parser\n");
1994-08-29 09:25:45 -03:00
err_ret->error = E_NOMEM;
PyTokenizer_Free(tok);
1994-08-29 09:25:45 -03:00
return NULL;
1990-12-20 11:06:42 -04:00
}
#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
if (flags & PyPARSE_WITH_IS_KEYWORD)
ps->p_flags |= CO_FUTURE_WITH_STATEMENT;
#endif
1994-08-29 09:25:45 -03:00
1990-12-20 11:06:42 -04:00
for (;;) {
char *a, *b;
int type;
size_t len;
1990-12-20 11:06:42 -04:00
char *str;
int col_offset;
1994-08-29 09:25:45 -03:00
1997-04-29 18:03:06 -03:00
type = PyTokenizer_Get(tok, &a, &b);
1990-12-20 11:06:42 -04:00
if (type == ERRORTOKEN) {
1994-08-29 09:25:45 -03:00
err_ret->error = tok->done;
1990-12-20 11:06:42 -04:00
break;
}
if (type == ENDMARKER && started) {
type = NEWLINE; /* Add an extra newline */
handling_with = handling_import = 0;
started = 0;
/* Add the right number of dedent tokens,
except if a certain flag is given --
codeop.py uses this. */
if (tok->indent &&
!(flags & PyPARSE_DONT_IMPLY_DEDENT))
{
tok->pendin = -tok->indent;
tok->indent = 0;
}
}
else
started = 1;
1994-08-29 09:25:45 -03:00
len = b - a; /* XXX this may compute NULL - NULL */
str = (char *) PyObject_MALLOC(len + 1);
1990-12-20 11:06:42 -04:00
if (str == NULL) {
fprintf(stderr, "no mem for next token\n");
1994-08-29 09:25:45 -03:00
err_ret->error = E_NOMEM;
1990-12-20 11:06:42 -04:00
break;
}
1994-08-29 09:25:45 -03:00
if (len > 0)
strncpy(str, a, len);
1990-12-20 11:06:42 -04:00
str[len] = '\0';
#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
/* This is only necessary to support the "as" warning, but
we don't want to warn about "as" in import statements. */
if (type == NAME &&
len == 6 && str[0] == 'i' && strcmp(str, "import") == 0)
handling_import = 1;
/* Warn about with as NAME */
if (type == NAME &&
!(ps->p_flags & CO_FUTURE_WITH_STATEMENT)) {
if (len == 4 && str[0] == 'w' && strcmp(str, "with") == 0)
warn(with_msg, err_ret->filename, tok->lineno);
else if (!(handling_import || handling_with) &&
len == 2 && str[0] == 'a' &&
strcmp(str, "as") == 0)
warn(as_msg, err_ret->filename, tok->lineno);
}
else if (type == NAME &&
(ps->p_flags & CO_FUTURE_WITH_STATEMENT) &&
len == 4 && str[0] == 'w' && strcmp(str, "with") == 0)
handling_with = 1;
#endif
if (a >= tok->line_start)
col_offset = a - tok->line_start;
else
col_offset = -1;
1994-08-29 09:25:45 -03:00
if ((err_ret->error =
PyParser_AddToken(ps, (int)type, str, tok->lineno, col_offset,
&(err_ret->expected))) != E_OK) {
if (err_ret->error != E_DONE) {
PyObject_FREE(str);
err_ret->token = type;
}
1990-12-20 11:06:42 -04:00
break;
}
1990-12-20 11:06:42 -04:00
}
1994-08-29 09:25:45 -03:00
if (err_ret->error == E_DONE) {
n = ps->p_tree;
ps->p_tree = NULL;
}
else
n = NULL;
1997-04-29 18:03:06 -03:00
PyParser_Delete(ps);
1994-08-29 09:25:45 -03:00
if (n == NULL) {
if (tok->lineno <= 1 && tok->done == E_EOF)
err_ret->error = E_EOF;
err_ret->lineno = tok->lineno;
if (tok->buf != NULL) {
char *text = NULL;
2006-02-15 15:56:18 -04:00
size_t len;
2006-02-15 13:27:45 -04:00
assert(tok->cur - tok->buf < INT_MAX);
err_ret->offset = (int)(tok->cur - tok->buf);
2006-02-15 15:56:18 -04:00
len = tok->inp - tok->buf;
#ifdef Py_USING_UNICODE
text = PyTokenizer_RestoreEncoding(tok, len, &err_ret->offset);
#endif
if (text == NULL) {
text = (char *) PyObject_MALLOC(len + 1);
if (text != NULL) {
if (len > 0)
strncpy(text, tok->buf, len);
text[len] = '\0';
}
1994-08-29 09:25:45 -03:00
}
err_ret->text = text;
1994-08-29 09:25:45 -03:00
}
} else if (tok->encoding != NULL) {
node* r = PyNode_New(encoding_decl);
if (!r) {
err_ret->error = E_NOMEM;
n = NULL;
goto done;
}
r->n_str = tok->encoding;
r->n_nchildren = 1;
r->n_child = n;
tok->encoding = NULL;
n = r;
1994-08-29 09:25:45 -03:00
}
done:
1997-04-29 18:03:06 -03:00
PyTokenizer_Free(tok);
1994-08-29 09:25:45 -03:00
return n;
1990-12-20 11:06:42 -04:00
}
static void
initerr(perrdetail *err_ret, const char *filename)
{
err_ret->error = E_OK;
err_ret->filename = filename;
err_ret->lineno = 0;
err_ret->offset = 0;
err_ret->text = NULL;
err_ret->token = -1;
err_ret->expected = -1;
}