Issue #14687: Cleanup unicode_writer_prepare()
"Inline" PyUnicode_Resize(): call directly resize_compact()
This commit is contained in:
parent
f2c76aa6cb
commit
7989157e49
|
@ -654,6 +654,7 @@ resize_compact(PyObject *unicode, Py_ssize_t length)
|
|||
Py_ssize_t new_size;
|
||||
int share_wstr;
|
||||
PyObject *new_unicode;
|
||||
assert(unicode_modifiable(unicode));
|
||||
assert(PyUnicode_IS_READY(unicode));
|
||||
assert(PyUnicode_IS_COMPACT(unicode));
|
||||
|
||||
|
@ -690,6 +691,7 @@ resize_compact(PyObject *unicode, Py_ssize_t length)
|
|||
}
|
||||
PyUnicode_WRITE(PyUnicode_KIND(unicode), PyUnicode_DATA(unicode),
|
||||
length, 0);
|
||||
assert(_PyUnicode_CheckConsistency(unicode, 0));
|
||||
return unicode;
|
||||
}
|
||||
|
||||
|
@ -1603,7 +1605,6 @@ unicode_resize(PyObject **p_unicode, Py_ssize_t length)
|
|||
if (new_unicode == NULL)
|
||||
return -1;
|
||||
*p_unicode = new_unicode;
|
||||
assert(_PyUnicode_CheckConsistency(*p_unicode, 0));
|
||||
return 0;
|
||||
}
|
||||
return resize_inplace(unicode, length);
|
||||
|
@ -13690,6 +13691,7 @@ unicode_writer_prepare(struct unicode_writer_t *writer,
|
|||
Py_ssize_t length, Py_UCS4 maxchar)
|
||||
{
|
||||
Py_ssize_t newlen;
|
||||
PyObject *newbuffer;
|
||||
|
||||
if (length > PY_SSIZE_T_MAX - writer->pos) {
|
||||
PyErr_NoMemory();
|
||||
|
@ -13697,15 +13699,14 @@ unicode_writer_prepare(struct unicode_writer_t *writer,
|
|||
}
|
||||
newlen = writer->pos + length;
|
||||
|
||||
if (newlen > writer->length && maxchar > writer->maxchar) {
|
||||
PyObject *newbuffer;
|
||||
|
||||
if (newlen > writer->length) {
|
||||
/* overallocate 25% to limit the number of resize */
|
||||
if (newlen > PY_SSIZE_T_MAX - newlen / 4)
|
||||
writer->length = newlen;
|
||||
else
|
||||
writer->length = newlen + newlen / 4;
|
||||
|
||||
if (maxchar > writer->maxchar) {
|
||||
/* resize + widen */
|
||||
newbuffer = PyUnicode_New(writer->length, maxchar);
|
||||
if (newbuffer == NULL)
|
||||
|
@ -13713,21 +13714,16 @@ unicode_writer_prepare(struct unicode_writer_t *writer,
|
|||
PyUnicode_CopyCharacters(newbuffer, 0,
|
||||
writer->buffer, 0, writer->pos);
|
||||
Py_DECREF(writer->buffer);
|
||||
}
|
||||
else {
|
||||
newbuffer = resize_compact(writer->buffer, writer->length);
|
||||
if (newbuffer == NULL)
|
||||
return -1;
|
||||
}
|
||||
writer->buffer = newbuffer;
|
||||
unicode_writer_update(writer);
|
||||
return 0;
|
||||
}
|
||||
if (newlen > writer->length) {
|
||||
/* overallocate 25% to limit the number of resize */
|
||||
if (newlen > PY_SSIZE_T_MAX - newlen / 4)
|
||||
writer->length = newlen;
|
||||
else
|
||||
writer->length = newlen + newlen / 4;
|
||||
if (PyUnicode_Resize(&writer->buffer, writer->length) < 0)
|
||||
return -1;
|
||||
unicode_writer_update(writer);
|
||||
}
|
||||
if (maxchar > writer->maxchar) {
|
||||
else if (maxchar > writer->maxchar) {
|
||||
if (unicode_widen(&writer->buffer, writer->pos, maxchar) < 0)
|
||||
return -1;
|
||||
unicode_writer_update(writer);
|
||||
|
|
Loading…
Reference in New Issue