Send key equivalents to renderer first, sendEvent-based.

Send key equivalents to renderer first, sendEvent-based.

Idea by suzhe@.

See also http://codereview.chromium.org/242069 .

Known issues:
* Uses a different code path than Safari and Firefox.

(* Suppressing keypress() doesn't prevent default action (suppressing keydown() works though. Both Firefox and Safari prevent the default action when either is suppressed). This seems to be broken on linux and windows too, though  – it seems to match IE8, but not Firefox or Safari. I guess this is the question for the web council.)


BUG=15090
TEST=Cmd-f now sucks when used on google docs.

[Avi (use Gerrit), 2009-10-12 16:10:42]

Hmmm...

I'm liking this one a little bit less. Yes, fixing the application's sendEvent
to send commandkeys in a sane way is nice, but it's a weird special-casing that
doesn't make me happy (and the dependency is odd). The other patch plays within
the standard rules.

This does LGTM but I'm feeling a bit better about the other one.

http://codereview.chromium.org/271054/diff/1/2
File chrome/browser/chrome_application_mac.mm (right):

http://codereview.chromium.org/271054/diff/1/2#newcode90
Line 90: // dispatching them to |keyDown:|. Since we want to send keys do the
renderer
TO the renderer...

[Nico, 2009-10-12 19:44:28]

For the records: We decided to not go with this patch but with the other version
for now.