[webkit-reviews] review denied: [Bug 14453] REGRESSION: www.nzherald.co.nz almost all the formating is gone : [Attachment 15416] Improved patch and testcase

[bugzilla-daemon at webkit.org]

Darin Adler <darin at apple.com> has denied Rob Buis <rwlbuis at gmail.com>'s request
for review:
Bug 14453: REGRESSION: www.nzherald.co.nz almost all the formating is gone
http://bugs.webkit.org/show_bug.cgi?id=14453

Attachment 15416: Improved patch and testcase
http://bugs.webkit.org/attachment.cgi?id=15416&action=edit

------- Additional Comments from Darin Adler <darin at apple.com>
It's not portable to call isdigit on a 16-bit character and count on it
returning false for non-ASCII values. In particular I think it either doesn't
compile at all or warns on Windows. Instead the code should just compare the
UChar to check if it's >= '0' and <= '9'. We should probably add some helpful
ASCII versions of ctype.h functions too, because the ctype.h functions change
their behavior based on the current locale and we really don't want that. I
think Sam Weinig is working on that.

Sorry to keep nitpicking this, but I still don't think that HASH is a clear
name for one of two different tokens that both start with "#". My original
comment still stands -- every code path for IDSEL and HASH is identical and
it's not helpful that the lexer gives two different tokens for these two
different cases. They should be combined into a single token. I think HASH
*would* be a reasonable name for that single token.