[Webkit-unassigned] [Bug 16548] REGRESSION(r28810): Font style and sizes are weird for Japanese text

[bugzilla-daemon at webkit.org]

http://bugs.webkit.org/show_bug.cgi?id=16548





------- Comment #13 from mitz at webkit.org  2007-12-21 22:45 PDT -------
(In reply to comment #11)
> The quickest *temporary* hack is exclude Japanese installation, simply
> checkable:
>         if (GetACP()==932){ /* is Japanese */}
> so hacks like:
>         if (/*TEMPHACK*/GetACP()!=932 && actualCodePages && ...

I believe the above will still break if you are on a Japanese installation once
you go to a Chinese page. We will then ask MLang to map from codepage 936
alone, so it will return a Chinese font, and subsequently it will return the
same font for all characters, even those that are both in 932 and in 936. This
inconsistent behavior and dependency on the order of operations is what r28810
was trying to prevent.

The system code page should probably factor into the font linking process, but
hopefully there is a way for that to happen that is internal to MLang or
another Windows API without having to query it explicitly in WebKit.

Perhaps hiding information from MLang is the way to go: when a character
belongs to multiple code pages, if one of them is the system code page ask only
about that code page. Then figure out what to do if none of the multiple code
pages is the system code page (as is the case on an English installation). I
think Mac OS X uses font traits to pick the fallback font.


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