mirror of
https://github.com/python/cpython.git
synced 2024-11-30 05:15:14 +08:00
261 lines
5.9 KiB
C
261 lines
5.9 KiB
C
#include "Python.h"
|
|
#include "node.h"
|
|
#include "token.h"
|
|
#include "graminit.h"
|
|
#include "compile.h"
|
|
#include "symtable.h"
|
|
|
|
#define UNDEFINED_FUTURE_FEATURE "future feature %.100s is not defined"
|
|
#define FUTURE_IMPORT_STAR "future statement does not support import *"
|
|
|
|
/* FUTURE_POSSIBLE() is provided to accomodate doc strings, which is
|
|
the only statement that can occur before a future statement.
|
|
*/
|
|
#define FUTURE_POSSIBLE(FF) ((FF)->ff_last_lineno == -1)
|
|
|
|
static int
|
|
future_check_features(PyFutureFeatures *ff, node *n, char *filename)
|
|
{
|
|
int i;
|
|
char *feature;
|
|
node *ch;
|
|
|
|
REQ(n, import_stmt); /* must by from __future__ import ... */
|
|
|
|
for (i = 3; i < NCH(n); i += 2) {
|
|
ch = CHILD(n, i);
|
|
if (TYPE(ch) == STAR) {
|
|
PyErr_SetString(PyExc_SyntaxError,
|
|
FUTURE_IMPORT_STAR);
|
|
PyErr_SyntaxLocation(filename, ch->n_lineno);
|
|
return -1;
|
|
}
|
|
REQ(ch, import_as_name);
|
|
feature = STR(CHILD(ch, 0));
|
|
if (strcmp(feature, FUTURE_NESTED_SCOPES) == 0) {
|
|
continue;
|
|
} else if (strcmp(feature, FUTURE_GENERATORS) == 0) {
|
|
continue;
|
|
} else if (strcmp(feature, FUTURE_DIVISION) == 0) {
|
|
ff->ff_features |= CO_FUTURE_DIVISION;
|
|
} else if (strcmp(feature, "braces") == 0) {
|
|
PyErr_SetString(PyExc_SyntaxError,
|
|
"not a chance");
|
|
PyErr_SyntaxLocation(filename, CHILD(ch, 0)->n_lineno);
|
|
return -1;
|
|
} else {
|
|
PyErr_Format(PyExc_SyntaxError,
|
|
UNDEFINED_FUTURE_FEATURE, feature);
|
|
PyErr_SyntaxLocation(filename, CHILD(ch, 0)->n_lineno);
|
|
return -1;
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static void
|
|
future_error(node *n, char *filename)
|
|
{
|
|
PyErr_SetString(PyExc_SyntaxError,
|
|
"from __future__ imports must occur at the "
|
|
"beginning of the file");
|
|
PyErr_SyntaxLocation(filename, n->n_lineno);
|
|
}
|
|
|
|
/* Relevant portions of the grammar:
|
|
|
|
single_input: NEWLINE | simple_stmt | compound_stmt NEWLINE
|
|
file_input: (NEWLINE | stmt)* ENDMARKER
|
|
stmt: simple_stmt | compound_stmt
|
|
simple_stmt: small_stmt (';' small_stmt)* [';'] NEWLINE
|
|
small_stmt: expr_stmt | print_stmt | del_stmt | pass_stmt | flow_stmt
|
|
| import_stmt | global_stmt | exec_stmt | assert_stmt
|
|
import_stmt: 'import' dotted_as_name (',' dotted_as_name)*
|
|
| 'from' dotted_name 'import' ('*' | import_as_name (',' import_as_name)*)
|
|
import_as_name: NAME [NAME NAME]
|
|
dotted_as_name: dotted_name [NAME NAME]
|
|
dotted_name: NAME ('.' NAME)*
|
|
*/
|
|
|
|
/* future_parse() finds future statements at the beginnning of a
|
|
module. The function calls itself recursively, rather than
|
|
factoring out logic for different kinds of statements into
|
|
different routines.
|
|
|
|
Return values:
|
|
-1 indicates an error occurred, e.g. unknown feature name
|
|
0 indicates no feature was found
|
|
1 indicates a feature was found
|
|
*/
|
|
|
|
static int
|
|
future_parse(PyFutureFeatures *ff, node *n, char *filename)
|
|
{
|
|
int i, r;
|
|
loop:
|
|
|
|
switch (TYPE(n)) {
|
|
|
|
case single_input:
|
|
if (TYPE(CHILD(n, 0)) == simple_stmt) {
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
}
|
|
return 0;
|
|
|
|
case file_input:
|
|
/* Check each statement in the file, starting with the
|
|
first, and continuing until the first statement
|
|
that isn't a future statement.
|
|
*/
|
|
for (i = 0; i < NCH(n); i++) {
|
|
node *ch = CHILD(n, i);
|
|
if (TYPE(ch) == stmt) {
|
|
r = future_parse(ff, ch, filename);
|
|
/* Need to check both conditions below
|
|
to accomodate doc strings, which
|
|
causes r < 0.
|
|
*/
|
|
if (r < 1 && !FUTURE_POSSIBLE(ff))
|
|
return r;
|
|
}
|
|
}
|
|
return 0;
|
|
|
|
case simple_stmt:
|
|
if (NCH(n) == 2) {
|
|
REQ(CHILD(n, 0), small_stmt);
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
} else {
|
|
/* Deal with the special case of a series of
|
|
small statements on a single line. If a
|
|
future statement follows some other
|
|
statement, the SyntaxError is raised here.
|
|
In all other cases, the symtable pass
|
|
raises the exception.
|
|
*/
|
|
int found = 0, end_of_future = 0;
|
|
|
|
for (i = 0; i < NCH(n); i += 2) {
|
|
if (TYPE(CHILD(n, i)) == small_stmt) {
|
|
r = future_parse(ff, CHILD(n, i),
|
|
filename);
|
|
if (r < 1)
|
|
end_of_future = 1;
|
|
else {
|
|
found = 1;
|
|
if (end_of_future) {
|
|
future_error(n,
|
|
filename);
|
|
return -1;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
/* If we found one and only one, then the
|
|
current lineno is legal.
|
|
*/
|
|
if (found)
|
|
ff->ff_last_lineno = n->n_lineno + 1;
|
|
else
|
|
ff->ff_last_lineno = n->n_lineno;
|
|
|
|
if (end_of_future && found)
|
|
return 1;
|
|
else
|
|
return 0;
|
|
}
|
|
|
|
case stmt:
|
|
if (TYPE(CHILD(n, 0)) == simple_stmt) {
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
} else if (TYPE(CHILD(n, 0)) == expr_stmt) {
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
} else {
|
|
REQ(CHILD(n, 0), compound_stmt);
|
|
ff->ff_last_lineno = n->n_lineno;
|
|
return 0;
|
|
}
|
|
|
|
case small_stmt:
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
|
|
case import_stmt: {
|
|
node *name;
|
|
|
|
if (STR(CHILD(n, 0))[0] != 'f') { /* from */
|
|
ff->ff_last_lineno = n->n_lineno;
|
|
return 0;
|
|
}
|
|
name = CHILD(n, 1);
|
|
if (strcmp(STR(CHILD(name, 0)), "__future__") != 0)
|
|
return 0;
|
|
if (future_check_features(ff, n, filename) < 0)
|
|
return -1;
|
|
ff->ff_last_lineno = n->n_lineno + 1;
|
|
return 1;
|
|
}
|
|
|
|
/* The cases below -- all of them! -- are necessary to find
|
|
and skip doc strings. */
|
|
case expr_stmt:
|
|
case testlist:
|
|
case test:
|
|
case and_test:
|
|
case not_test:
|
|
case comparison:
|
|
case expr:
|
|
case xor_expr:
|
|
case and_expr:
|
|
case shift_expr:
|
|
case arith_expr:
|
|
case term:
|
|
case factor:
|
|
case power:
|
|
if (NCH(n) == 1) {
|
|
n = CHILD(n, 0);
|
|
goto loop;
|
|
}
|
|
break;
|
|
|
|
case atom:
|
|
if (TYPE(CHILD(n, 0)) == STRING
|
|
&& ff->ff_found_docstring == 0) {
|
|
ff->ff_found_docstring = 1;
|
|
return 0;
|
|
}
|
|
ff->ff_last_lineno = n->n_lineno;
|
|
return 0;
|
|
|
|
default:
|
|
ff->ff_last_lineno = n->n_lineno;
|
|
return 0;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
PyFutureFeatures *
|
|
PyNode_Future(node *n, char *filename)
|
|
{
|
|
PyFutureFeatures *ff;
|
|
|
|
ff = (PyFutureFeatures *)PyMem_Malloc(sizeof(PyFutureFeatures));
|
|
if (ff == NULL)
|
|
return NULL;
|
|
ff->ff_found_docstring = 0;
|
|
ff->ff_last_lineno = -1;
|
|
ff->ff_features = 0;
|
|
|
|
if (future_parse(ff, n, filename) < 0) {
|
|
PyMem_Free((void *)ff);
|
|
return NULL;
|
|
}
|
|
return ff;
|
|
}
|
|
|