bpo-34636: Use fast path for more chars in SRE category macros. (GH-9170)

sir-sigurd · miss-islington · commit ec014a101a7f · 2018-09-11T15:47:59.000-07:00
When handling \s, \d, or \w (and their inverse) escapes in bytes regexes this a small but measurable performance improvement.  https://bugs.python.org/issue34636
diff --git a/Misc/NEWS.d/next/Library/2018-09-11-15-04-05.bpo-34636.capCmt.rst b/Misc/NEWS.d/next/Library/2018-09-11-15-04-05.bpo-34636.capCmt.rst
@@ -0,0 +1,2 @@
+Speed up re scanning of many non-matching characters for \s \w and \d within
+bytes objects. (microoptimization)
diff --git a/Modules/_sre.c b/Modules/_sre.c
@@ -87,13 +87,13 @@ static const char copyright[] =
 /* search engine state */
 
 #define SRE_IS_DIGIT(ch)\
-    ((ch) < 128 && Py_ISDIGIT(ch))
+    ((ch) <= '9' && Py_ISDIGIT(ch))
 #define SRE_IS_SPACE(ch)\
-    ((ch) < 128 && Py_ISSPACE(ch))
+    ((ch) <= ' ' && Py_ISSPACE(ch))
 #define SRE_IS_LINEBREAK(ch)\
     ((ch) == '\n')
 #define SRE_IS_WORD(ch)\
-    ((ch) < 128 && (Py_ISALNUM(ch) || (ch) == '_'))
+    ((ch) <= 'z' && (Py_ISALNUM(ch) || (ch) == '_'))
 
 static unsigned int sre_lower_ascii(unsigned int ch)
 {

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+Speed up re scanning of many non-matching characters for \s \w and \d within`
	`2`	`+bytes objects. (microoptimization)`