Stowage/cpython
2
0
Fork 0
mirror of https://github.com/python/cpython.git synced 2025-11-04 07:31:38 +00:00
Code Issues Projects Releases Packages Wiki Activity

Optimize unicode_subtype_new(): don't encode to wchar_t and decode from wchar_t

Browse source 
Rewrite unicode_subtype_new(): allocate directly the right type.
This commit is contained in:
Victor Stinner 2011-10-01 16:16:43 +02:00
parent e90fe6a8f4
commit 07ac3ebd7b
2 changed files with 81 additions and 43 deletions
Download patch file Download diff file Expand all files Collapse all files

11
Lib/test/test_unicode.py
View file

@ -1010,10 +1010,13 @@ def test_constructor(self):
class UnicodeSubclass(str): class UnicodeSubclass(str):
pass pass
self.assertEqual( for text in ('ascii', '\xe9', '\u20ac', '\U0010FFFF'):
str(UnicodeSubclass('unicode subclass becomes unicode')), subclass = UnicodeSubclass(text)
'unicode subclass becomes unicode' self.assertEqual(str(subclass), text)
) self.assertEqual(len(subclass), len(text))
if text == 'ascii':
self.assertEqual(subclass.encode('ascii'), b'ascii')
self.assertEqual(subclass.encode('utf-8'), b'ascii')
self.assertEqual( self.assertEqual(
str('strings are converted to unicode'), str('strings are converted to unicode'),

111
Objects/unicodeobject.c
View file

@ -12410,56 +12410,91 @@ unicode_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
static PyObject * static PyObject *
unicode_subtype_new(PyTypeObject *type, PyObject *args, PyObject *kwds) unicode_subtype_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
{ {
PyUnicodeObject *tmp, *pnew; PyUnicodeObject *unicode, *self;
Py_ssize_t n; Py_ssize_t length, char_size;
PyObject *err = NULL; int share_wstr, share_utf8;
unsigned int kind;
void *data;
assert(PyType_IsSubtype(type, &PyUnicode_Type)); assert(PyType_IsSubtype(type, &PyUnicode_Type));
tmp = (PyUnicodeObject *)unicode_new(&PyUnicode_Type, args, kwds);
if (tmp == NULL) unicode = (PyUnicodeObject *)unicode_new(&PyUnicode_Type, args, kwds);
if (unicode == NULL)
return NULL; return NULL;
assert(PyUnicode_Check(tmp)); assert(PyUnicode_Check(unicode));
// TODO: Verify the PyUnicode_GET_SIZE does the right thing. if (PyUnicode_READY(unicode))
// it seems kind of strange that tp_alloc gets passed the size return NULL;
// of the unicode string because there will follow another
// malloc. self = (PyUnicodeObject *) type->tp_alloc(type, 0);
pnew = (PyUnicodeObject *) type->tp_alloc(type, if (self == NULL) {
n = PyUnicode_GET_SIZE(tmp)); Py_DECREF(unicode);
if (pnew == NULL) {
Py_DECREF(tmp);
return NULL; return NULL;
} }
_PyUnicode_WSTR(pnew) = (Py_UNICODE*) PyObject_MALLOC(sizeof(Py_UNICODE) * (n+1)); kind = PyUnicode_KIND(unicode);
if (_PyUnicode_WSTR(pnew) == NULL) { length = PyUnicode_GET_LENGTH(unicode);
err = PyErr_NoMemory();
_PyUnicode_LENGTH(self) = length;
_PyUnicode_HASH(self) = _PyUnicode_HASH(unicode);
_PyUnicode_STATE(self).interned = 0;
_PyUnicode_STATE(self).kind = kind;
_PyUnicode_STATE(self).compact = 0;
_PyUnicode_STATE(self).ascii = 0;
_PyUnicode_STATE(self).ready = 1;
_PyUnicode_WSTR(self) = NULL;
_PyUnicode_UTF8_LENGTH(self) = 0;
_PyUnicode_UTF8(self) = NULL;
_PyUnicode_WSTR_LENGTH(self) = 0;
self->data.any = NULL;
share_utf8 = 0;
share_wstr = 0;
if (kind == PyUnicode_1BYTE_KIND) {
char_size = 1;
if (PyUnicode_MAX_CHAR_VALUE(unicode) < 128)
share_utf8 = 1;
}
else if (kind == PyUnicode_2BYTE_KIND) {
char_size = 2;
if (sizeof(wchar_t) == 2)
share_wstr = 1;
}
else {
assert(kind == PyUnicode_4BYTE_KIND);
char_size = 4;
if (sizeof(wchar_t) == 4)
share_wstr = 1;
}
/* Ensure we won't overflow the length. */
if (length > (PY_SSIZE_T_MAX / char_size - 1)) {
PyErr_NoMemory();
goto onError; goto onError;
} }
Py_UNICODE_COPY(_PyUnicode_WSTR(pnew), PyUnicode_AS_UNICODE(tmp), n+1); data = PyObject_MALLOC((length + 1) * char_size);
_PyUnicode_WSTR_LENGTH(pnew) = n; if (data == NULL) {
_PyUnicode_HASH(pnew) = _PyUnicode_HASH(tmp); PyErr_NoMemory();
_PyUnicode_STATE(pnew).interned = 0;
_PyUnicode_STATE(pnew).kind = 0;
_PyUnicode_STATE(pnew).compact = 0;
_PyUnicode_STATE(pnew).ready = 0;
_PyUnicode_STATE(pnew).ascii = 0;
pnew->data.any = NULL;
_PyUnicode_LENGTH(pnew) = 0;
pnew->_base.utf8 = NULL;
pnew->_base.utf8_length = 0;
if (PyUnicode_READY(pnew) == -1) {
PyObject_FREE(_PyUnicode_WSTR(pnew));
goto onError; goto onError;
} }
Py_DECREF(tmp); self->data.any = data;
return (PyObject *)pnew; if (share_utf8) {
_PyUnicode_UTF8_LENGTH(self) = length;
_PyUnicode_UTF8(self) = data;
}
if (share_wstr) {
_PyUnicode_WSTR_LENGTH(self) = length;
_PyUnicode_WSTR(self) = (wchar_t *)data;
}
Py_MEMCPY(data, PyUnicode_DATA(unicode),
PyUnicode_KIND_SIZE(kind, length + 1));
Py_DECREF(unicode);
return (PyObject *)self;
onError: onError:
_Py_ForgetReference((PyObject *)pnew); Py_DECREF(unicode);
PyObject_Del(pnew); Py_DECREF(self);
Py_DECREF(tmp); return NULL;
return err;
} }
PyDoc_STRVAR(unicode_doc, PyDoc_STRVAR(unicode_doc,