Patch #1970 by Antoine Pitrou: Speedup unicode whitespace and linebreak detection. The speedup is about 25% for split() (571 / 457 usec) and 35% (175 / 127 usec )for splitlines()
diff --git a/Include/unicodeobject.h b/Include/unicodeobject.h
index 2c11bfd..8809b34 100644
--- a/Include/unicodeobject.h
+++ b/Include/unicodeobject.h
@@ -348,7 +348,14 @@
#else
-#define Py_UNICODE_ISSPACE(ch) _PyUnicode_IsWhitespace(ch)
+/* Since splitting on whitespace is an important use case, and whitespace
+ in most situations is solely ASCII whitespace, we optimize for the common
+ case by using a quick look-up table with an inlined check.
+ */
+extern const unsigned char _Py_ascii_whitespace[];
+
+#define Py_UNICODE_ISSPACE(ch) \
+ ((ch) < 128U ? _Py_ascii_whitespace[(ch)] : _PyUnicode_IsWhitespace(ch))
#define Py_UNICODE_ISLOWER(ch) _PyUnicode_IsLowercase(ch)
#define Py_UNICODE_ISUPPER(ch) _PyUnicode_IsUppercase(ch)