Pass unhandled keyup messages coming in from the renderer to the underlying v...

Pass unhandled keyup messages coming in from the renderer to the underlying view. This is needed
for system keyboard messages which are generated when the Alt key is pressed in addition to other
keys to function correctly in ChromeFrame.

The TabContentsView implementation for windows does the right thing by specifically checking for keydown
events before processing them as accelerators. The TabContentsView implementation for Linux also treated
keyup messages as accelerators, which is now fixed.

This fixes parts of http://code.google.com/p/chromium/issues/detail?id=22996

Bug=22996
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=27926

[jam, 2009-09-30 20:46:58]

I'm actually removing this whole if statement in
http://codereview.chromium.org/224023, which I'm about to check in now.  The if
isn't necessary since those 4 checks are all that exist.

[jcampan, 2009-10-02 18:55:08]

LGTM

[jam, 2009-10-02 19:08:15]

http://codereview.chromium.org/255028/diff/6001/6003
File chrome/browser/tab_contents/tab_contents_view_gtk.cc (right):

http://codereview.chromium.org/255028/diff/6001/6003#newcode575
Line 575: event.type == WebKit::WebInputEvent::KeyUp)
nit: should the comment be updated?  also what about mac?

[ananta, 2009-10-02 19:19:41]

Adding Avi so he can opine on whether there are issues with keyUp messages being
processed as accelerators on the Mac.

http://codereview.chromium.org/255028/diff/6001/6003
File chrome/browser/tab_contents/tab_contents_view_gtk.cc (right):

http://codereview.chromium.org/255028/diff/6001/6003#newcode575
Line 575: event.type == WebKit::WebInputEvent::KeyUp)
On 2009/10/02 19:08:15, John Abd-El-Malek wrote:
> nit: should the comment be updated?  also what about mac?

Done. Added avi to the list of reviewers for his opinion on the mac.