Renamed /shadow-all/ to /shadow/.

Source/core/css/CSSParserValues.h

Index: Source/core/css/CSSParserValues.h
diff --git a/Source/core/css/CSSParserValues.h b/Source/core/css/CSSParserValues.h
index 37d56aac75202f6815ea7697921650889b626eb1..3ef91b4391bb69216c9c5ea71f5b6dd66ef534fc 100644
--- a/Source/core/css/CSSParserValues.h
+++ b/Source/core/css/CSSParserValues.h
@@ -98,9 +98,8 @@ struct CSSParserString {
 
     bool equalIgnoringCase(const char* str) const
     {
-        if (is8Bit())
-            return WTF::equalIgnoringCase(str, characters8(), length());
-        return WTF::equalIgnoringCase(str, characters16(), length());
+        bool match = is8Bit() ? WTF::equalIgnoringCase(str, characters8(), length()) : WTF::equalIgnoringCase(str, characters16(), length());
+        return match ? str[length()] == '\0' : false;

[dglazkov, 2014/03/06 18:35:03]

We shouldn't wallpaper of what looks like a bug in WTF::equalIgnoringCase here.
Can you look into why this occurs?

[adamk, 2014/03/06 22:42:18]

Seems like you this function should just start with:

if (strlen(str) != length())
    return false;

This isn't a bug in WTF::equalIgnoringCase(), it's a big in this function
(CSSParserString::equalIgnoringCase).

[tasak, 2014/03/07 05:08:18]

Yeah. So is it ok to use strlen?

I think, strlen(str) is slower than str[length()] == '\0'.
Yeah. This is CSSParserString's bug.
It misuses WTF::equalIgnoringCase.

     }
 
     template <size_t strLength>