[Webkit-unassigned] [Bug 4920] Non-BMP characters in JavaScript identifiers

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=4920





--- Comment #16 from Gavin Barraclough <barraclough at apple.com>  2011-06-14 12:15:52 PST ---
> > The Unicode spec defines types for Unicode Code Points
> 
> As you said, they are defined for code points, and not for code units. These two-bytes sequences of UTF-16 are code units.

Sure, and the code unit to code point mapping is outside the definition of the Unicode spec.  ECMAScript used to specifically defined that the source text encoding was UCS-2, which is a direct mapping from code units to code points.  They no longer call out the name of this encoding, but the spec defined lexing behaviour still matches UCS-2 decoding.

> I didn't reopen this bug, because I don't think that there is enough practical difference for us to care. But tracking a obvious JS spec mistake with a WebKit bug is not unthinkable, even when we're in compliance.

That's an understandable and reasonable goal.  I'm just worried that this bug report as it stands is likely to be misconstrued as a bug in our code that should be fixed, and that a contributor might work towards making changes that we would not accept.  Maybe we could try to get a bug filed against the ECMA spec (I believe their bug track is only open to ES members :-( ), and then you could reopen and bug & mark this bug as blocked on the ES one?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.