Mailing List Archive

https://github.com/python/cpython/commit/444156ede44204ef16c9d3cfcb03a637535fd5bf
commit: 444156ede44204ef16c9d3cfcb03a637535fd5bf
branch: main
author: Erlend E. Aasland <erlend@python.org>
committer: erlend-aasland <erlend.aasland@protonmail.com>
date: 2024-04-03T09:11:39+02:00
summary:

gh-117431: Adapt str.startswith and str.endswith to Argument Clinic (#117466)

This change gives a significant speedup, as the METH_FASTCALL calling
convention is now used.

files:
A Misc/NEWS.d/next/Core and Builtins/2024-04-02-17-37-35.gh-issue-117431.vDKAOn.rst
M Lib/test/string_tests.py
M Objects/clinic/unicodeobject.c.h
M Objects/unicodeobject.c

diff --git a/Lib/test/string_tests.py b/Lib/test/string_tests.py
index cecf309dca9194..5ade7013328d63 100644
--- a/Lib/test/string_tests.py
+++ b/Lib/test/string_tests.py
@@ -1513,9 +1513,9 @@ def test_find_etc_raise_correct_error_messages(self):
x, None, None, None)
self.assertRaisesRegex(TypeError, r'^count\(', s.count,
x, None, None, None)
- self.assertRaisesRegex(TypeError, r'^startswith\(', s.startswith,
+ self.assertRaisesRegex(TypeError, r'^startswith\b', s.startswith,
x, None, None, None)
- self.assertRaisesRegex(TypeError, r'^endswith\(', s.endswith,
+ self.assertRaisesRegex(TypeError, r'^endswith\b', s.endswith,
x, None, None, None)

# issue #15534
diff --git a/Misc/NEWS.d/next/Core and Builtins/2024-04-02-17-37-35.gh-issue-117431.vDKAOn.rst b/Misc/NEWS.d/next/Core and Builtins/2024-04-02-17-37-35.gh-issue-117431.vDKAOn.rst
new file mode 100644
index 00000000000000..96e14ea0c3b1bd
--- /dev/null
+++ b/Misc/NEWS.d/next/Core and Builtins/2024-04-02-17-37-35.gh-issue-117431.vDKAOn.rst
@@ -0,0 +1,2 @@
+Improve the performance of :meth:`str.startswith` and :meth:`str.endswith`
+by adapting them to the :c:macro:`METH_FASTCALL` calling convention.
diff --git a/Objects/clinic/unicodeobject.c.h b/Objects/clinic/unicodeobject.c.h
index 3e5167d9242fe4..c956bb1936776a 100644
--- a/Objects/clinic/unicodeobject.c.h
+++ b/Objects/clinic/unicodeobject.c.h
@@ -1369,6 +1369,108 @@ unicode_zfill(PyObject *self, PyObject *arg)
return return_value;
}

+PyDoc_STRVAR(unicode_startswith__doc__,
+"startswith($self, prefix[, start[, end]], /)\n"
+"--\n"
+"\n"
+"Return True if the string starts with the specified prefix, False otherwise.\n"
+"\n"
+" prefix\n"
+" A string or a tuple of strings to try.\n"
+" start\n"
+" Optional start position. Default: start of the string.\n"
+" end\n"
+" Optional stop position. Default: end of the string.");
+
+#define UNICODE_STARTSWITH_METHODDEF \
+ {"startswith", _PyCFunction_CAST(unicode_startswith), METH_FASTCALL, unicode_startswith__doc__},
+
+static PyObject *
+unicode_startswith_impl(PyObject *self, PyObject *subobj, Py_ssize_t start,
+ Py_ssize_t end);
+
+static PyObject *
+unicode_startswith(PyObject *self, PyObject *const *args, Py_ssize_t nargs)
+{
+ PyObject *return_value = NULL;
+ PyObject *subobj;
+ Py_ssize_t start = 0;
+ Py_ssize_t end = PY_SSIZE_T_MAX;
+
+ if (!_PyArg_CheckPositional("startswith", nargs, 1, 3)) {
+ goto exit;
+ }
+ subobj = args[0];
+ if (nargs < 2) {
+ goto skip_optional;
+ }
+ if (!_PyEval_SliceIndex(args[1], &start)) {
+ goto exit;
+ }
+ if (nargs < 3) {
+ goto skip_optional;
+ }
+ if (!_PyEval_SliceIndex(args[2], &end)) {
+ goto exit;
+ }
+skip_optional:
+ return_value = unicode_startswith_impl(self, subobj, start, end);
+
+exit:
+ return return_value;
+}
+
+PyDoc_STRVAR(unicode_endswith__doc__,
+"endswith($self, prefix[, start[, end]], /)\n"
+"--\n"
+"\n"
+"Return True if the string ends with the specified prefix, False otherwise.\n"
+"\n"
+" prefix\n"
+" A string or a tuple of strings to try.\n"
+" start\n"
+" Optional start position. Default: start of the string.\n"
+" end\n"
+" Optional stop position. Default: end of the string.");
+
+#define UNICODE_ENDSWITH_METHODDEF \
+ {"endswith", _PyCFunction_CAST(unicode_endswith), METH_FASTCALL, unicode_endswith__doc__},
+
+static PyObject *
+unicode_endswith_impl(PyObject *self, PyObject *subobj, Py_ssize_t start,
+ Py_ssize_t end);
+
+static PyObject *
+unicode_endswith(PyObject *self, PyObject *const *args, Py_ssize_t nargs)
+{
+ PyObject *return_value = NULL;
+ PyObject *subobj;
+ Py_ssize_t start = 0;
+ Py_ssize_t end = PY_SSIZE_T_MAX;
+
+ if (!_PyArg_CheckPositional("endswith", nargs, 1, 3)) {
+ goto exit;
+ }
+ subobj = args[0];
+ if (nargs < 2) {
+ goto skip_optional;
+ }
+ if (!_PyEval_SliceIndex(args[1], &start)) {
+ goto exit;
+ }
+ if (nargs < 3) {
+ goto skip_optional;
+ }
+ if (!_PyEval_SliceIndex(args[2], &end)) {
+ goto exit;
+ }
+skip_optional:
+ return_value = unicode_endswith_impl(self, subobj, start, end);
+
+exit:
+ return return_value;
+}
+
PyDoc_STRVAR(unicode___format____doc__,
"__format__($self, format_spec, /)\n"
"--\n"
@@ -1507,4 +1609,4 @@ unicode_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
exit:
return return_value;
}
-/*[clinic end generated code: output=1aab29bab5201c78 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=e495e878d8283217 input=a9049054013a1b77]*/
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index e412af5f797e7a..ac59419b3c7a50 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -13021,30 +13021,30 @@ unicode_zfill_impl(PyObject *self, Py_ssize_t width)
return u;
}

-PyDoc_STRVAR(startswith__doc__,
- "S.startswith(prefix[, start[, end]]) -> bool\n\
-\n\
-Return True if S starts with the specified prefix, False otherwise.\n\
-With optional start, test S beginning at that position.\n\
-With optional end, stop comparing S at that position.\n\
-prefix can also be a tuple of strings to try.");
+/*[clinic input]
+@text_signature "($self, prefix[, start[, end]], /)"
+str.startswith as unicode_startswith
+
+ prefix as subobj: object
+ A string or a tuple of strings to try.
+ start: slice_index(accept={int, NoneType}, c_default='0') = None
+ Optional start position. Default: start of the string.
+ end: slice_index(accept={int, NoneType}, c_default='PY_SSIZE_T_MAX') = None
+ Optional stop position. Default: end of the string.
+ /
+
+Return True if the string starts with the specified prefix, False otherwise.
+[clinic start generated code]*/

static PyObject *
-unicode_startswith(PyObject *self,
- PyObject *args)
+unicode_startswith_impl(PyObject *self, PyObject *subobj, Py_ssize_t start,
+ Py_ssize_t end)
+/*[clinic end generated code: output=4bd7cfd0803051d4 input=5f918b5f5f89d856]*/
{
- PyObject *subobj;
- PyObject *substring;
- Py_ssize_t start = 0;
- Py_ssize_t end = PY_SSIZE_T_MAX;
- int result;
-
- if (!asciilib_parse_args_finds("startswith", args, &subobj, &start, &end))
- return NULL;
if (PyTuple_Check(subobj)) {
Py_ssize_t i;
for (i = 0; i < PyTuple_GET_SIZE(subobj); i++) {
- substring = PyTuple_GET_ITEM(subobj, i);
+ PyObject *substring = PyTuple_GET_ITEM(subobj, i);
if (!PyUnicode_Check(substring)) {
PyErr_Format(PyExc_TypeError,
"tuple for startswith must only contain str, "
@@ -13052,9 +13052,10 @@ unicode_startswith(PyObject *self,
Py_TYPE(substring)->tp_name);
return NULL;
}
- result = tailmatch(self, substring, start, end, -1);
- if (result == -1)
+ int result = tailmatch(self, substring, start, end, -1);
+ if (result < 0) {
return NULL;
+ }
if (result) {
Py_RETURN_TRUE;
}
@@ -13068,37 +13069,30 @@ unicode_startswith(PyObject *self,
"a tuple of str, not %.100s", Py_TYPE(subobj)->tp_name);
return NULL;
}
- result = tailmatch(self, subobj, start, end, -1);
- if (result == -1)
+ int result = tailmatch(self, subobj, start, end, -1);
+ if (result < 0) {
return NULL;
+ }
return PyBool_FromLong(result);
}


-PyDoc_STRVAR(endswith__doc__,
- "S.endswith(suffix[, start[, end]]) -> bool\n\
-\n\
-Return True if S ends with the specified suffix, False otherwise.\n\
-With optional start, test S beginning at that position.\n\
-With optional end, stop comparing S at that position.\n\
-suffix can also be a tuple of strings to try.");
+/*[clinic input]
+@text_signature "($self, prefix[, start[, end]], /)"
+str.endswith as unicode_endswith = str.startswith
+
+Return True if the string ends with the specified prefix, False otherwise.
+[clinic start generated code]*/

static PyObject *
-unicode_endswith(PyObject *self,
- PyObject *args)
+unicode_endswith_impl(PyObject *self, PyObject *subobj, Py_ssize_t start,
+ Py_ssize_t end)
+/*[clinic end generated code: output=cce6f8ceb0102ca9 input=82cd5ce9e7623646]*/
{
- PyObject *subobj;
- PyObject *substring;
- Py_ssize_t start = 0;
- Py_ssize_t end = PY_SSIZE_T_MAX;
- int result;
-
- if (!asciilib_parse_args_finds("endswith", args, &subobj, &start, &end))
- return NULL;
if (PyTuple_Check(subobj)) {
Py_ssize_t i;
for (i = 0; i < PyTuple_GET_SIZE(subobj); i++) {
- substring = PyTuple_GET_ITEM(subobj, i);
+ PyObject *substring = PyTuple_GET_ITEM(subobj, i);
if (!PyUnicode_Check(substring)) {
PyErr_Format(PyExc_TypeError,
"tuple for endswith must only contain str, "
@@ -13106,9 +13100,10 @@ unicode_endswith(PyObject *self,
Py_TYPE(substring)->tp_name);
return NULL;
}
- result = tailmatch(self, substring, start, end, +1);
- if (result == -1)
+ int result = tailmatch(self, substring, start, end, +1);
+ if (result < 0) {
return NULL;
+ }
if (result) {
Py_RETURN_TRUE;
}
@@ -13121,9 +13116,10 @@ unicode_endswith(PyObject *self,
"a tuple of str, not %.100s", Py_TYPE(subobj)->tp_name);
return NULL;
}
- result = tailmatch(self, subobj, start, end, +1);
- if (result == -1)
+ int result = tailmatch(self, subobj, start, end, +1);
+ if (result < 0) {
return NULL;
+ }
return PyBool_FromLong(result);
}

@@ -13576,8 +13572,8 @@ static PyMethodDef unicode_methods[] = {
UNICODE_SWAPCASE_METHODDEF
UNICODE_TRANSLATE_METHODDEF
UNICODE_UPPER_METHODDEF
- {"startswith", (PyCFunction) unicode_startswith, METH_VARARGS, startswith__doc__},
- {"endswith", (PyCFunction) unicode_endswith, METH_VARARGS, endswith__doc__},
+ UNICODE_STARTSWITH_METHODDEF
+ UNICODE_ENDSWITH_METHODDEF
UNICODE_REMOVEPREFIX_METHODDEF
UNICODE_REMOVESUFFIX_METHODDEF
UNICODE_ISASCII_METHODDEF

_______________________________________________
Python-checkins mailing list -- python-checkins@python.org
To unsubscribe send an email to python-checkins-leave@python.org
https://mail.python.org/mailman3/lists/python-checkins.python.org/
Member address: list-python-checkins@lists.gossamer-threads.com