[Mac] Confirm-to-Quit: Wait for a KeyUp event if the user held Cmd+Q on the second tap.

[Mac] Confirm-to-Quit: Wait for a KeyUp event if the user held Cmd+Q on the second tap.

This also refactors the logic that was in AppController to be in the
ConfirmQuitPanelController

BUG=74976
TEST=Press and hold Cmd+Q. Windows fade out and Chromium quits when you release the keys.
TEST=Tap Cmd+Q and get the floaty window. Tap it again quickly and Chromium quits.
TEST=Tap Cmd+Q to get the floaty window. Press and hold again and the windows hide and Chromium quits when the keys are released.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=77157

[Nico, 2011-03-07 17:40:19]

LG

I'm fine with you landing this. If any of my comments need addressing, you can
do so in a follow-up (but chances are that it's just me being confused again :-)
).

http://codereview.chromium.org/6625025/diff/1/chrome/browser/app_controller_m...
File chrome/browser/app_controller_mac.mm (right):

http://codereview.chromium.org/6625025/diff/1/chrome/browser/app_controller_m...
chrome/browser/app_controller_mac.mm:295: return NSTerminateNow;
nit: I'd probably push this block into the controller as well.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
File chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm (right):

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:171: } while
(!nextEvent);
Wait, this busy-polls, right? Why not call nextEventMatchingMask just once, but
with [NSDate distantFuture] as date?

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:189: nextEvent = [self
pumpEventQueueForKeyUp:app];
I guess you could pass in |targetDate| to nEMM as well here?

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:296:
//base::mac::ScopedNSAutoreleasePool pool;  // Do we need this?
do we? my guess is 'no'.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:308: for (NSWindow*
aWindow in [app windows]) {
nit: just |window|

[Nico, 2011-03-07 17:40:50]

(oh, and it would've been nice to do the moving stuff around in a separate CL;
would've been easier to see what changed in that case.)

[Robert Sesek, 2011-03-07 18:07:43]

On 2011/03/07 17:40:50, Nico wrote:
> (oh, and it would've been nice to do the moving stuff around in a separate CL;
> would've been easier to see what changed in that case.)

Yea, sorry. The changes here necessitated the move for code reuse reasons.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/app_controller_m...
File chrome/browser/app_controller_mac.mm (right):

http://codereview.chromium.org/6625025/diff/1/chrome/browser/app_controller_m...
chrome/browser/app_controller_mac.mm:295: return NSTerminateNow;
On 2011/03/07 17:40:19, Nico wrote:
> nit: I'd probably push this block into the controller as well.

Done.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
File chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm (right):

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:171: } while
(!nextEvent);
On 2011/03/07 17:40:19, Nico wrote:
> Wait, this busy-polls, right? Why not call nextEventMatchingMask just once,
but
> with [NSDate distantFuture] as date?

Done.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:189: nextEvent = [self
pumpEventQueueForKeyUp:app];
On 2011/03/07 17:40:19, Nico wrote:
> I guess you could pass in |targetDate| to nEMM as well here?

I looked at doing this, but the FuzzFactor also must be considered. I'm going to
leave this one with untilDate:nil until I have time to explore more.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:296:
//base::mac::ScopedNSAutoreleasePool pool;  // Do we need this?
On 2011/03/07 17:40:19, Nico wrote:
> do we? my guess is 'no'.

Nope.

http://codereview.chromium.org/6625025/diff/1/chrome/browser/ui/cocoa/confirm...
chrome/browser/ui/cocoa/confirm_quit_panel_controller.mm:308: for (NSWindow*
aWindow in [app windows]) {
On 2011/03/07 17:40:19, Nico wrote:
> nit: just |window|

Done.

[Nico, 2011-03-07 18:14:56]

LG