Work in progress: Adds MessagePumpDispatcher::ShouldExit(). This is used to handle the

Adds MessagePumpDispatcher::ShouldExit(). This is used to handle the
case of needing to exit a nested message loop before processing a
Task. With Aura we get far fewer real X events, which means it's
entirely possible we want to exit the nested message loop but can't
because the exit won't happen until the next X event.

BUG=124123
TEST=none
R=darin@chromium.org

[sky, 2012-04-19 21:33:37]

This isn't done, I wanted to make sure you're ok with it before I clean up other
code and get all places working.

The core of the change is the new method in MessagePumpDispatcher, the call to
it in MessagePumpGlib and the override in MenuController.

  -Scott

[sky, 2012-04-24 03:31:55]

Ping

[darin (slow to review), 2012-04-24 03:36:49]

MessageLoop::QuitNow() does not suffice?

[sky, 2012-04-24 04:25:49]

It too has the same problem I'm addressing here. Also, QuitNow is
risky here since I can't tell if I happened to have scheduled
multiple.

  -Scott

On Mon, Apr 23, 2012 at 8:36 PM,  <darin@chromium.org> wrote:
> MessageLoop::QuitNow() does not suffice?
>
> http://codereview.chromium.org/10143001/

[sky, 2012-04-26 00:14:28]

As you said, QuitNow does what I need. I'm closing this patch so it isn't
needed.

Thanks!

[darin (slow to review), 2012-04-26 02:47:59]

That's great to hear!

On Wed, Apr 25, 2012 at 5:14 PM, <sky@chromium.org> wrote:

> As you said, QuitNow does what I need. I'm closing this patch so it isn't
> needed.
>
> Thanks!
>
>
http://codereview.chromium.**org/10143001/<http://codereview.chromium.org/101...
>

[sky, 2012-05-10 01:10:13]

So, of course I ran into a bug where my quit was quitting someone
elses nested message loop. I can work around this particular loop, but
oy. Arbitrary quitting is error prone.
Ugh!

  -Scott

On Wed, Apr 25, 2012 at 7:47 PM, Darin Fisher <darin@chromium.org> wrote:
> That's great to hear!
>
>
> On Wed, Apr 25, 2012 at 5:14 PM, <sky@chromium.org> wrote:
>>
>> As you said, QuitNow does what I need. I'm closing this patch so it isn't
>> needed.
>>
>> Thanks!
>>
>> http://codereview.chromium.org/10143001/
>
>

[darin (slow to review), 2012-05-10 07:00:22]

Maybe we need a mechanism to refer to a particular invocation of
MessageLoop::Run() so that Quit / QuitNow will only quit the intended
invocation.  Hmm...

-Darin


On Wed, May 9, 2012 at 6:10 PM, Scott Violet <sky@chromium.org> wrote:

> So, of course I ran into a bug where my quit was quitting someone
> elses nested message loop. I can work around this particular loop, but
> oy. Arbitrary quitting is error prone.
> Ugh!
>
>  -Scott
>
> On Wed, Apr 25, 2012 at 7:47 PM, Darin Fisher <darin@chromium.org> wrote:
> > That's great to hear!
> >
> >
> > On Wed, Apr 25, 2012 at 5:14 PM, <sky@chromium.org> wrote:
> >>
> >> As you said, QuitNow does what I need. I'm closing this patch so it
> isn't
> >> needed.
> >>
> >> Thanks!
> >>
> >> http://codereview.chromium.org/10143001/
> >
> >
>

[sky, 2012-05-10 16:08:57]

Better yet, I should redo menus so they don't need the nested message
loop. Other than changing all existing code the only thing I really
need is the ability to get in the way of native message dispatching.
Maybe we should add the ability to set a Dispatcher without running a
nested message loop.

  -Scott

On Thu, May 10, 2012 at 12:00 AM, Darin Fisher <darin@chromium.org> wrote:
> Maybe we need a mechanism to refer to a particular invocation of
> MessageLoop::Run() so that Quit / QuitNow will only quit the intended
> invocation.  Hmm...
>
> -Darin
>
>
> On Wed, May 9, 2012 at 6:10 PM, Scott Violet <sky@chromium.org> wrote:
>>
>> So, of course I ran into a bug where my quit was quitting someone
>> elses nested message loop. I can work around this particular loop, but
>> oy. Arbitrary quitting is error prone.
>> Ugh!
>>
>>  -Scott
>>
>> On Wed, Apr 25, 2012 at 7:47 PM, Darin Fisher <darin@chromium.org> wrote:
>> > That's great to hear!
>> >
>> >
>> > On Wed, Apr 25, 2012 at 5:14 PM, <sky@chromium.org> wrote:
>> >>
>> >> As you said, QuitNow does what I need. I'm closing this patch so it
>> >> isn't
>> >> needed.
>> >>
>> >> Thanks!
>> >>
>> >> http://codereview.chromium.org/10143001/
>> >
>> >
>
>