[Python-checkins] CVS: python/dist/src/Modules _sre.c,2.67,2.68
Fredrik Lundh
effbot@users.sourceforge.net
Sat, 20 Oct 2001 10:48:49 -0700
Update of /cvsroot/python/python/dist/src/Modules
In directory usw-pr-cvs1:/tmp/cvs-serv8184/Modules
Modified Files:
_sre.c
Log Message:
rewrote the pattern.split method in C
also restored SRE Unicode support for 1.6/2.0/2.1
Index: _sre.c
===================================================================
RCS file: /cvsroot/python/python/dist/src/Modules/_sre.c,v
retrieving revision 2.67
retrieving revision 2.68
diff -C2 -d -r2.67 -r2.68
*** _sre.c 2001/10/18 19:30:11 2.67
--- _sre.c 2001/10/20 17:48:46 2.68
***************
*** 34,37 ****
--- 34,38 ----
* 2001-09-18 fl added _getliteral helper
* 2001-10-18 fl fixed group reset issue (from Matthew Mueller)
+ * 2001-10-20 fl added split primitive; reenable unicode for 1.6/2.0/2.1
*
* Copyright (c) 1997-2001 by Secret Labs AB. All rights reserved.
***************
*** 66,77 ****
#undef VERBOSE
! #if PY_VERSION_HEX >= 0x01060000 && defined(Py_USING_UNICODE)
/* defining this enables unicode support (default under 1.6a1 and later) */
#define HAVE_UNICODE
#endif
/* -------------------------------------------------------------------- */
/* optional features */
/* prevent run-away recursion (bad patterns on long strings) */
--- 67,83 ----
#undef VERBOSE
! #if PY_VERSION_HEX >= 0x01060000
! #if PY_VERSION_HEX < 0x02020000 || defined(Py_USING_UNICODE)
/* defining this enables unicode support (default under 1.6a1 and later) */
#define HAVE_UNICODE
#endif
+ #endif
/* -------------------------------------------------------------------- */
/* optional features */
+ /* test: define to use sre._split helper instead of C code */
+ #undef USE_PYTHON_SPLIT
+
/* prevent run-away recursion (bad patterns on long strings) */
***************
*** 1489,1493 ****
LOCAL(PyObject*)
! state_getslice(SRE_STATE* state, int index, PyObject* string)
{
int i, j;
--- 1495,1499 ----
LOCAL(PyObject*)
! state_getslice(SRE_STATE* state, int index, PyObject* string, int empty)
{
int i, j;
***************
*** 1496,1500 ****
if (string == Py_None || !state->mark[index] || !state->mark[index+1]) {
! i = j = 0;
} else {
i = ((char*)state->mark[index] - (char*)state->beginning) /
--- 1502,1512 ----
if (string == Py_None || !state->mark[index] || !state->mark[index+1]) {
! if (empty)
! /* want empty string */
! i = j = 0;
! else {
! Py_INCREF(Py_None);
! return Py_None;
! }
} else {
i = ((char*)state->mark[index] - (char*)state->beginning) /
***************
*** 1783,1786 ****
--- 1795,1799 ----
}
+ #if defined(USE_PYTHON_SPLIT)
static PyObject*
pattern_split(PatternObject* self, PyObject* args, PyObject* kw)
***************
*** 1799,1802 ****
--- 1812,1816 ----
);
}
+ #endif
static PyObject*
***************
*** 1806,1810 ****
PyObject* list;
int status;
! int i;
PyObject* string;
--- 1820,1824 ----
PyObject* list;
int status;
! int i, b, e;
PyObject* string;
***************
*** 1843,1857 ****
switch (self->groups) {
case 0:
! item = PySequence_GetSlice(
! string,
! ((char*) state.start - (char*) state.beginning) /
! state.charsize,
! ((char*) state.ptr - (char*) state.beginning) /
! state.charsize);
if (!item)
goto error;
break;
case 1:
! item = state_getslice(&state, 1, string);
if (!item)
goto error;
--- 1857,1870 ----
switch (self->groups) {
case 0:
! b = ((char*) state.start - (char*) state.beginning) /
! state.charsize;
! e = ((char*) state.ptr - (char*) state.beginning) /
! state.charsize;
! item = PySequence_GetSlice(string, b, e);
if (!item)
goto error;
break;
case 1:
! item = state_getslice(&state, 1, string, 1);
if (!item)
goto error;
***************
*** 1862,1866 ****
goto error;
for (i = 0; i < self->groups; i++) {
! PyObject* o = state_getslice(&state, i+1, string);
if (!o) {
Py_DECREF(item);
--- 1875,1879 ----
goto error;
for (i = 0; i < self->groups; i++) {
! PyObject* o = state_getslice(&state, i+1, string, 1);
if (!o) {
Py_DECREF(item);
***************
*** 1903,1906 ****
--- 1916,2030 ----
}
+
+ #if !defined(USE_PYTHON_SPLIT)
+ static PyObject*
+ pattern_split(PatternObject* self, PyObject* args, PyObject* kw)
+ {
+ SRE_STATE state;
+ PyObject* list;
+ PyObject* item;
+ int status;
+ int n;
+ int i, b, e;
+ int g;
+
+ PyObject* string;
+ int maxsplit = 0;
+ static char* kwlist[] = { "source", "maxsplit", NULL };
+ if (!PyArg_ParseTupleAndKeywords(args, kw, "O|i:split", kwlist,
+ &string, &maxsplit))
+ return NULL;
+
+ string = state_init(&state, self, string, 0, INT_MAX);
+ if (!string)
+ return NULL;
+
+ list = PyList_New(0);
+
+ i = n = 0;
+
+ while (maxsplit == 0 || n < maxsplit) {
+
+ state_reset(&state);
+
+ state.ptr = state.start;
+
+ if (state.charsize == 1) {
+ status = sre_search(&state, PatternObject_GetCode(self));
+ } else {
+ #if defined(HAVE_UNICODE)
+ status = sre_usearch(&state, PatternObject_GetCode(self));
+ #endif
+ }
+
+ if (status > 0) {
+
+ if (state.start == state.ptr) {
+ if (i >= state.endpos)
+ break;
+ /* skip one character */
+ state.start = (void*) ((char*) state.ptr + state.charsize);
+ continue;
+ }
+
+ b = ((char*) state.start - (char*) state.beginning) /
+ state.charsize;
+ e = ((char*) state.ptr - (char*) state.beginning) /
+ state.charsize;
+
+ /* get segment before this match */
+ item = PySequence_GetSlice(string, i, b);
+ if (!item)
+ goto error;
+ status = PyList_Append(list, item);
+ Py_DECREF(item);
+ if (status < 0)
+ goto error;
+
+ for (g = 0; g < self->groups; g++) {
+ item = state_getslice(&state, g+1, string, 0);
+ if (!item)
+ goto error;
+ status = PyList_Append(list, item);
+ Py_DECREF(item);
+ if (status < 0)
+ goto error;
+ }
+
+ i = e;
+ n = n + 1;
+
+ state.start = state.ptr;
+
+ } else {
+
+ if (status == 0)
+ break;
+
+ pattern_error(status);
+ goto error;
+
+ }
+ }
+
+ /* get segment following last match */
+ item = PySequence_GetSlice(string, i, state.endpos);
+ if (!item)
+ goto error;
+ status = PyList_Append(list, item);
+ Py_DECREF(item);
+ if (status < 0)
+ goto error;
+
+ state_fini(&state);
+ return list;
+
+ error:
+ Py_DECREF(list);
+ state_fini(&state);
+ return NULL;
+
+ }
+ #endif
static PyObject*