[Webkit-unassigned] [Bug 25119] IME modifies the DOM before keydown

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=25119





------- Comment #18 from ojan at chromium.org  2009-05-17 23:42 PDT -------
(In reply to comment #16)
Eric, all those sound reasonable. Will post a new patch soon.

(In reply to comment #17)
> +    // IE sends VK_PROCESSKEY which has value 229;
> Weird semicolon.
> +        // Currently, only Mac Safari uses this. Eventually GTK, and Chromium
> Linux/Win
> 
> If it affects linux at the IME level, it'll likely affect all linux webkit
> ports (importantly Qt).  Might be worth rephrasing this so you don't need to
> enumerate them.  (This comment is duplicated across two locations.)

I don't actually know if it needs to affect linux. Currently, Qt goes through
the same codepath as Mac (at least, according to the comments). That said, the
Mac code path is a side effect of the requirements of Mac IME. It's not clear
to me that linux needs to go through all the same contortions. Mainly, it might
be possible to cancel IME keydowns on Linux as it is on Windows (Safari Windows
will ignore an IME keypress if the keydown event is canceled). Specifically,
the Mac issue is that you need to tell the IME a keypress happened before you
know if it will handle the keypress. But you need to know if it will handle the
keypress in order to set the 229 keycode on the keydown event, at which point,
it's too late to cancel the keydown as the IME has no undo (basically what
Alexey said above). Does Linux have the same issue?


-- 
Configure bugmail: https://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.