isspace: reduce need for category_code call, and inline

jiahao · May 28, 2015 · f04f10d · f04f10d
1 parent bb848b8
commit f04f10d
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/base/utf8proc.jl b/base/utf8proc.jl
@@ -155,8 +155,9 @@ iscntrl(c::Char) = (c <= Char(0x1f) || Char(0x7f) <= c <= Char(0x9f))
 
 ispunct(c::Char) = (UTF8PROC_CATEGORY_PC <= category_code(c) <= UTF8PROC_CATEGORY_PO)
 
-# 0x85 is the Unicode Next Line (NEL) character
-isspace(c::Char) = c == ' ' || '\t' <= c <='\r' || c == Char(0x85) || category_code(c)==UTF8PROC_CATEGORY_ZS
+# \u85 is the Unicode Next Line (NEL) character
+# the check for \ufffd allows for branch removal on ASCIIStrings
+@inline isspace(c::Char) = c == ' ' || '\t' <= c <='\r' || c == '\u85' || '\ua0' <= c && c != '\ufffd' && category_code(c)==UTF8PROC_CATEGORY_ZS
 
 isprint(c::Char) = (UTF8PROC_CATEGORY_LU <= category_code(c) <= UTF8PROC_CATEGORY_ZS)