Allow the ZWJ and ZWNJ characters in IdentifierPart as per ES5.

src/char-predicates.h

 // Copyright 2011 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 38 matching lines @@
       case '$': case '_': case '\\': return true;
       default: return unibrow::Letter::Is(c);
     }
   }
 };
 
 
 struct IdentifierPart {
   static inline bool Is(uc32 c) {
     return IdentifierStart::Is(c)
+        || c == 0x200C
+        || c == 0x200D

[mathias, 2012/02/22 09:49:37]

As an `IdentifierStart` character is the most common case for an
`IdentifierPart`, I figured it would be best to perform the U+200C and U+200D
checks right after, but before the more expensive Unicode category lookups.

[Lasse Reichstein Nielsen, 2012/02/22 14:03:28]

Good thought.
I'd prefer moving the number (digit, really) test before the ZWNJ/ZWJ test,
since digits are probably more likely to be in actual identifiers. 

Perhaps add a comment after the constants (like "  // U+200C is
ZeroWidthNonJoiner.").

         || unibrow::Number::Is(c)
         || unibrow::CombiningMark::Is(c)
         || unibrow::ConnectorPunctuation::Is(c);
   }
 };
 
 } }  // namespace v8::internal
 
 #endif  // V8_CHAR_PREDICATES_H_