bpo-30285: Optimize case-insensitive matching and searching (#1482)

of regular expressions.

2 files changed, 97 insertions, 1 deletions

diff --git a/Modules/_sre.c b/Modules/_sre.c
index a86c5f2..6873f1d 100644
--- a/Modules/_sre.c
+++ b/Modules/_sre.c
@@ -274,6 +274,38 @@ _sre_getcodesize_impl(PyObject *module)
 }
 
 /*[clinic input]
+_sre.ascii_iscased -> bool
+
+    character: int
+    /
+
+[clinic start generated code]*/
+
+static int
+_sre_ascii_iscased_impl(PyObject *module, int character)
+/*[clinic end generated code: output=4f454b630fbd19a2 input=9f0bd952812c7ed3]*/
+{
+    unsigned int ch = (unsigned int)character;
+    return ch != sre_lower(ch) || ch != sre_upper(ch);
+}
+
+/*[clinic input]
+_sre.unicode_iscased -> bool
+
+    character: int
+    /
+
+[clinic start generated code]*/
+
+static int
+_sre_unicode_iscased_impl(PyObject *module, int character)
+/*[clinic end generated code: output=9c5ddee0dc2bc258 input=51e42c3b8dddb78e]*/
+{
+    unsigned int ch = (unsigned int)character;
+    return ch != sre_lower_unicode(ch) || ch != sre_upper_unicode(ch);
+}
+
+/*[clinic input]
 _sre.ascii_tolower -> int
 
     character: int
@@ -2750,6 +2782,8 @@ static PyTypeObject Scanner_Type = {
 static PyMethodDef _functions[] = {
     _SRE_COMPILE_METHODDEF
     _SRE_GETCODESIZE_METHODDEF
+    _SRE_ASCII_ISCASED_METHODDEF
+    _SRE_UNICODE_ISCASED_METHODDEF
     _SRE_ASCII_TOLOWER_METHODDEF
     _SRE_UNICODE_TOLOWER_METHODDEF
     {NULL, NULL}
diff --git a/Modules/clinic/_sre.c.h b/Modules/clinic/_sre.c.h
index 8056eda..1e60686 100644
--- a/Modules/clinic/_sre.c.h
+++ b/Modules/clinic/_sre.c.h
@@ -29,6 +29,68 @@ exit:
     return return_value;
 }
 
+PyDoc_STRVAR(_sre_ascii_iscased__doc__,
+"ascii_iscased($module, character, /)\n"
+"--\n"
+"\n");
+
+#define _SRE_ASCII_ISCASED_METHODDEF    \
+    {"ascii_iscased", (PyCFunction)_sre_ascii_iscased, METH_O, _sre_ascii_iscased__doc__},
+
+static int
+_sre_ascii_iscased_impl(PyObject *module, int character);
+
+static PyObject *
+_sre_ascii_iscased(PyObject *module, PyObject *arg)
+{
+    PyObject *return_value = NULL;
+    int character;
+    int _return_value;
+
+    if (!PyArg_Parse(arg, "i:ascii_iscased", &character)) {
+        goto exit;
+    }
+    _return_value = _sre_ascii_iscased_impl(module, character);
+    if ((_return_value == -1) && PyErr_Occurred()) {
+        goto exit;
+    }
+    return_value = PyBool_FromLong((long)_return_value);
+
+exit:
+    return return_value;
+}
+
+PyDoc_STRVAR(_sre_unicode_iscased__doc__,
+"unicode_iscased($module, character, /)\n"
+"--\n"
+"\n");
+
+#define _SRE_UNICODE_ISCASED_METHODDEF    \
+    {"unicode_iscased", (PyCFunction)_sre_unicode_iscased, METH_O, _sre_unicode_iscased__doc__},
+
+static int
+_sre_unicode_iscased_impl(PyObject *module, int character);
+
+static PyObject *
+_sre_unicode_iscased(PyObject *module, PyObject *arg)
+{
+    PyObject *return_value = NULL;
+    int character;
+    int _return_value;
+
+    if (!PyArg_Parse(arg, "i:unicode_iscased", &character)) {
+        goto exit;
+    }
+    _return_value = _sre_unicode_iscased_impl(module, character);
+    if ((_return_value == -1) && PyErr_Occurred()) {
+        goto exit;
+    }
+    return_value = PyBool_FromLong((long)_return_value);
+
+exit:
+    return return_value;
+}
+
 PyDoc_STRVAR(_sre_ascii_tolower__doc__,
 "ascii_tolower($module, character, /)\n"
 "--\n"
@@ -715,4 +777,4 @@ _sre_SRE_Scanner_search(ScannerObject *self, PyObject *Py_UNUSED(ignored))
 {
     return _sre_SRE_Scanner_search_impl(self);
 }
-/*[clinic end generated code: output=811e67d7f8f5052e input=a9049054013a1b77]*/
+/*[clinic end generated code: output=5fe47c49e475cccb input=a9049054013a1b77]*/