cpython/Parser/parsetok.c

481 lines
14 KiB
C

/* Parser-tokenizer link implementation */
#include "Python.h"
#include "tokenizer.h"
#include "node.h"
#include "grammar.h"
#include "parser.h"
#include "parsetok.h"
#include "errcode.h"
#include "graminit.h"
/* Forward */
static node *parsetok(struct tok_state *, grammar *, int, perrdetail *, int *);
static int initerr(perrdetail *err_ret, PyObject * filename);
typedef struct {
struct {
int lineno;
char *comment;
} *items;
size_t size;
size_t num_items;
} growable_comment_array;
static int
growable_comment_array_init(growable_comment_array *arr, size_t initial_size) {
assert(initial_size > 0);
arr->items = malloc(initial_size * sizeof(*arr->items));
arr->size = initial_size;
arr->num_items = 0;
return arr->items != NULL;
}
static int
growable_comment_array_add(growable_comment_array *arr, int lineno, char *comment) {
if (arr->num_items >= arr->size) {
size_t new_size = arr->size * 2;
void *new_items_array = realloc(arr->items, new_size * sizeof(*arr->items));
if (!new_items_array) {
return 0;
}
arr->items = new_items_array;
arr->size = new_size;
}
arr->items[arr->num_items].lineno = lineno;
arr->items[arr->num_items].comment = comment;
arr->num_items++;
return 1;
}
static void
growable_comment_array_deallocate(growable_comment_array *arr) {
for (unsigned i = 0; i < arr->num_items; i++) {
PyObject_FREE(arr->items[i].comment);
}
free(arr->items);
}
/* Parse input coming from a string. Return error code, print some errors. */
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)
{
int iflags = flags;
return PyParser_ParseStringFlagsFilenameEx(s, filename, g, start,
err_ret, &iflags);
}
node *
PyParser_ParseStringObject(const char *s, PyObject *filename,
grammar *g, int start,
perrdetail *err_ret, int *flags)
{
struct tok_state *tok;
int exec_input = start == file_input;
if (initerr(err_ret, filename) < 0)
return NULL;
if (PySys_Audit("compile", "yO", s, err_ret->filename) < 0) {
err_ret->error = E_ERROR;
return NULL;
}
if (*flags & PyPARSE_IGNORE_COOKIE)
tok = PyTokenizer_FromUTF8(s, exec_input);
else
tok = PyTokenizer_FromString(s, exec_input);
if (tok == NULL) {
err_ret->error = PyErr_Occurred() ? E_DECODE : E_NOMEM;
return NULL;
}
if (*flags & PyPARSE_TYPE_COMMENTS) {
tok->type_comments = 1;
}
Py_INCREF(err_ret->filename);
tok->filename = err_ret->filename;
if (*flags & PyPARSE_ASYNC_HACKS)
tok->async_hacks = 1;
return parsetok(tok, g, start, err_ret, flags);
}
node *
PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename_str,
grammar *g, int start,
perrdetail *err_ret, int *flags)
{
node *n;
PyObject *filename = NULL;
if (filename_str != NULL) {
filename = PyUnicode_DecodeFSDefault(filename_str);
if (filename == NULL) {
err_ret->error = E_ERROR;
return NULL;
}
}
n = PyParser_ParseStringObject(s, filename, g, start, err_ret, flags);
Py_XDECREF(filename);
return n;
}
/* Parse input coming from a file. Return error code, print some errors. */
node *
PyParser_ParseFile(FILE *fp, const char *filename, grammar *g, int start,
const char *ps1, const char *ps2,
perrdetail *err_ret)
{
return PyParser_ParseFileFlags(fp, filename, NULL,
g, start, ps1, ps2, err_ret, 0);
}
node *
PyParser_ParseFileFlags(FILE *fp, const char *filename, const char *enc,
grammar *g, int start,
const char *ps1, const char *ps2,
perrdetail *err_ret, int flags)
{
int iflags = flags;
return PyParser_ParseFileFlagsEx(fp, filename, enc, g, start, ps1,
ps2, err_ret, &iflags);
}
node *
PyParser_ParseFileObject(FILE *fp, PyObject *filename,
const char *enc, grammar *g, int start,
const char *ps1, const char *ps2,
perrdetail *err_ret, int *flags)
{
struct tok_state *tok;
if (initerr(err_ret, filename) < 0)
return NULL;
if (PySys_Audit("compile", "OO", Py_None, err_ret->filename) < 0) {
return NULL;
}
if ((tok = PyTokenizer_FromFile(fp, enc, ps1, ps2)) == NULL) {
err_ret->error = E_NOMEM;
return NULL;
}
if (*flags & PyPARSE_TYPE_COMMENTS) {
tok->type_comments = 1;
}
Py_INCREF(err_ret->filename);
tok->filename = err_ret->filename;
return parsetok(tok, g, start, err_ret, flags);
}
node *
PyParser_ParseFileFlagsEx(FILE *fp, const char *filename,
const char *enc, grammar *g, int start,
const char *ps1, const char *ps2,
perrdetail *err_ret, int *flags)
{
node *n;
PyObject *fileobj = NULL;
if (filename != NULL) {
fileobj = PyUnicode_DecodeFSDefault(filename);
if (fileobj == NULL) {
err_ret->error = E_ERROR;
return NULL;
}
}
n = PyParser_ParseFileObject(fp, fileobj, enc, g,
start, ps1, ps2, err_ret, flags);
Py_XDECREF(fileobj);
return n;
}
/* Parse input coming from the given tokenizer structure.
Return error code. */
static node *
parsetok(struct tok_state *tok, grammar *g, int start, perrdetail *err_ret,
int *flags)
{
parser_state *ps;
node *n;
int started = 0;
int col_offset, end_col_offset;
growable_comment_array type_ignores;
if (!growable_comment_array_init(&type_ignores, 10)) {
err_ret->error = E_NOMEM;
PyTokenizer_Free(tok);
return NULL;
}
if ((ps = PyParser_New(g, start)) == NULL) {
err_ret->error = E_NOMEM;
growable_comment_array_deallocate(&type_ignores);
PyTokenizer_Free(tok);
return NULL;
}
#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
if (*flags & PyPARSE_BARRY_AS_BDFL)
ps->p_flags |= CO_FUTURE_BARRY_AS_BDFL;
if (*flags & PyPARSE_TYPE_COMMENTS)
ps->p_flags |= PyCF_TYPE_COMMENTS;
#endif
for (;;) {
const char *a, *b;
int type;
size_t len;
char *str;
col_offset = -1;
int lineno;
const char *line_start;
type = PyTokenizer_Get(tok, &a, &b);
if (type == ERRORTOKEN) {
err_ret->error = tok->done;
break;
}
if (type == ENDMARKER && started) {
type = NEWLINE; /* Add an extra newline */
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;
len = (a != NULL && b != NULL) ? b - a : 0;
str = (char *) PyObject_MALLOC(len + 1);
if (str == NULL) {
err_ret->error = E_NOMEM;
break;
}
if (len > 0)
strncpy(str, a, len);
str[len] = '\0';
#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
if (type == NOTEQUAL) {
if (!(ps->p_flags & CO_FUTURE_BARRY_AS_BDFL) &&
strcmp(str, "!=")) {
PyObject_FREE(str);
err_ret->error = E_SYNTAX;
break;
}
else if ((ps->p_flags & CO_FUTURE_BARRY_AS_BDFL) &&
strcmp(str, "<>")) {
PyObject_FREE(str);
err_ret->expected = NOTEQUAL;
err_ret->error = E_SYNTAX;
break;
}
}
#endif
/* Nodes of type STRING, especially multi line strings
must be handled differently in order to get both
the starting line number and the column offset right.
(cf. issue 16806) */
lineno = type == STRING ? tok->first_lineno : tok->lineno;
line_start = type == STRING ? tok->multi_line_start : tok->line_start;
if (a != NULL && a >= line_start) {
col_offset = Py_SAFE_DOWNCAST(a - line_start,
intptr_t, int);
}
else {
col_offset = -1;
}
if (b != NULL && b >= tok->line_start) {
end_col_offset = Py_SAFE_DOWNCAST(b - tok->line_start,
intptr_t, int);
}
else {
end_col_offset = -1;
}
if (type == TYPE_IGNORE) {
if (!growable_comment_array_add(&type_ignores, tok->lineno, str)) {
err_ret->error = E_NOMEM;
break;
}
continue;
}
if ((err_ret->error =
PyParser_AddToken(ps, (int)type, str,
lineno, col_offset, tok->lineno, end_col_offset,
&(err_ret->expected))) != E_OK) {
if (err_ret->error != E_DONE) {
PyObject_FREE(str);
err_ret->token = type;
}
break;
}
}
if (err_ret->error == E_DONE) {
n = ps->p_tree;
ps->p_tree = NULL;
if (n->n_type == file_input) {
/* Put type_ignore nodes in the ENDMARKER of file_input. */
int num;
node *ch;
size_t i;
num = NCH(n);
ch = CHILD(n, num - 1);
REQ(ch, ENDMARKER);
for (i = 0; i < type_ignores.num_items; i++) {
int res = PyNode_AddChild(ch, TYPE_IGNORE, type_ignores.items[i].comment,
type_ignores.items[i].lineno, 0,
type_ignores.items[i].lineno, 0);
if (res != 0) {
err_ret->error = res;
PyNode_Free(n);
n = NULL;
break;
}
type_ignores.items[i].comment = NULL;
}
}
/* Check that the source for a single input statement really
is a single statement by looking at what is left in the
buffer after parsing. Trailing whitespace and comments
are OK. */
if (err_ret->error == E_DONE && start == single_input) {
const char *cur = tok->cur;
char c = *tok->cur;
for (;;) {
while (c == ' ' || c == '\t' || c == '\n' || c == '\014')
c = *++cur;
if (!c)
break;
if (c != '#') {
err_ret->error = E_BADSINGLE;
PyNode_Free(n);
n = NULL;
break;
}
/* Suck up comment. */
while (c && c != '\n')
c = *++cur;
}
}
}
else
n = NULL;
growable_comment_array_deallocate(&type_ignores);
#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
*flags = ps->p_flags;
#endif
PyParser_Delete(ps);
if (n == NULL) {
if (tok->done == E_EOF)
err_ret->error = E_EOF;
err_ret->lineno = tok->lineno;
if (tok->buf != NULL) {
size_t len;
assert(tok->cur - tok->buf < INT_MAX);
/* if we've managed to parse a token, point the offset to its start,
* else use the current reading position of the tokenizer
*/
err_ret->offset = col_offset != -1 ? col_offset + 1 : ((int)(tok->cur - tok->buf));
len = tok->inp - tok->buf;
err_ret->text = (char *) PyObject_MALLOC(len + 1);
if (err_ret->text != NULL) {
if (len > 0)
strncpy(err_ret->text, tok->buf, len);
err_ret->text[len] = '\0';
}
}
} else if (tok->encoding != NULL) {
/* 'nodes->n_str' uses PyObject_*, while 'tok->encoding' was
* allocated using PyMem_
*/
node* r = PyNode_New(encoding_decl);
if (r)
r->n_str = PyObject_MALLOC(strlen(tok->encoding)+1);
if (!r || !r->n_str) {
err_ret->error = E_NOMEM;
if (r)
PyObject_FREE(r);
n = NULL;
goto done;
}
strcpy(r->n_str, tok->encoding);
PyMem_FREE(tok->encoding);
tok->encoding = NULL;
r->n_nchildren = 1;
r->n_child = n;
n = r;
}
done:
PyTokenizer_Free(tok);
if (n != NULL) {
_PyNode_FinalizeEndPos(n);
}
return n;
}
static int
initerr(perrdetail *err_ret, PyObject *filename)
{
err_ret->error = E_OK;
err_ret->lineno = 0;
err_ret->offset = 0;
err_ret->text = NULL;
err_ret->token = -1;
err_ret->expected = -1;
if (filename) {
Py_INCREF(filename);
err_ret->filename = filename;
}
else {
err_ret->filename = PyUnicode_FromString("<string>");
if (err_ret->filename == NULL) {
err_ret->error = E_ERROR;
return -1;
}
}
return 0;
}