ozone: Remove the explicit call to start listening for events.

ozone: Remove the explicit call to start listening for events.

It is not necessary for WindowTreeHostOzone to call into the event-factory
to explicitly request it to start receiving events. The evdev implementation
can start listening to the events immediately, and the canca implementation
can start listening for events when the dispatcher list changes (this is the
same we do for the libevent based x11 event-source).

BUG=361137
R=spang@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=271036

[sadrul, 2014-05-15 16:30:04]

https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
File ui/events/ozone/evdev/event_factory_evdev.cc (right):

https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
ui/events/ozone/evdev/event_factory_evdev.cc:155:
device_manager_->ScanDevices(this);
I am not sure if this depends on other parts of the system being initialized
first? If so, we can move this from here into OnDispatcherListChanged()
override.

[spang, 2014-05-15 17:04:39]

On 2014/05/15 16:30:04, sadrul wrote:
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> File ui/events/ozone/evdev/event_factory_evdev.cc (right):
> 
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> ui/events/ozone/evdev/event_factory_evdev.cc:155:
> device_manager_->ScanDevices(this);
> I am not sure if this depends on other parts of the system being initialized
> first? If so, we can move this from here into OnDispatcherListChanged()
> override.

Before landing this, we need to fix the platform to build the EventFactoryOzone
in the browser process only.

Maybe we can have

  ui::OzonePlatform::GetInstance()->CreateEventFactoryOzone()

inside aura::Env::Init()?

Otherwise, GPU process will break.

[sadrul, 2014-05-15 17:12:24]

On 2014/05/15 17:04:39, spang wrote:
> On 2014/05/15 16:30:04, sadrul wrote:
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > File ui/events/ozone/evdev/event_factory_evdev.cc (right):
> > 
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > ui/events/ozone/evdev/event_factory_evdev.cc:155:
> > device_manager_->ScanDevices(this);
> > I am not sure if this depends on other parts of the system being initialized
> > first? If so, we can move this from here into OnDispatcherListChanged()
> > override.
> 
> Before landing this, we need to fix the platform to build the
EventFactoryOzone
> in the browser process only.
> 
> Maybe we can have
> 
>   ui::OzonePlatform::GetInstance()->CreateEventFactoryOzone()
> 
> inside aura::Env::Init()?
> 
> Otherwise, GPU process will break.

Ah yes. I remember the issue now. I was thinking we could have explicit
OzonePlatform::InitializeForUI() and InitializeForGPU(), and have the platform
take care of initializing the right set of subsystems. We can then have the
InitializeForUI() done in Env::Init, and InitializeForGPU() in <insert
appropriate place ... GpuMain?>

[spang, 2014-05-15 17:17:25]

On 2014/05/15 17:12:24, sadrul wrote:
> On 2014/05/15 17:04:39, spang wrote:
> > On 2014/05/15 16:30:04, sadrul wrote:
> > >
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > > File ui/events/ozone/evdev/event_factory_evdev.cc (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > > ui/events/ozone/evdev/event_factory_evdev.cc:155:
> > > device_manager_->ScanDevices(this);
> > > I am not sure if this depends on other parts of the system being
initialized
> > > first? If so, we can move this from here into OnDispatcherListChanged()
> > > override.
> > 
> > Before landing this, we need to fix the platform to build the
> EventFactoryOzone
> > in the browser process only.
> > 
> > Maybe we can have
> > 
> >   ui::OzonePlatform::GetInstance()->CreateEventFactoryOzone()
> > 
> > inside aura::Env::Init()?
> > 
> > Otherwise, GPU process will break.
> 
> Ah yes. I remember the issue now. I was thinking we could have explicit
> OzonePlatform::InitializeForUI() and InitializeForGPU(), and have the platform
> take care of initializing the right set of subsystems. We can then have the
> InitializeForUI() done in Env::Init, and InitializeForGPU() in <insert
> appropriate place ... GpuMain?>

Ok.

[sadrul, 2014-05-16 06:01:41]

On 2014/05/15 17:17:25, spang wrote:
> On 2014/05/15 17:12:24, sadrul wrote:
> > On 2014/05/15 17:04:39, spang wrote:
> > > On 2014/05/15 16:30:04, sadrul wrote:
> > > >
> > >
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > > > File ui/events/ozone/evdev/event_factory_evdev.cc (right):
> > > > 
> > > >
> > >
> >
>
https://codereview.chromium.org/285303004/diff/1/ui/events/ozone/evdev/event_...
> > > > ui/events/ozone/evdev/event_factory_evdev.cc:155:
> > > > device_manager_->ScanDevices(this);
> > > > I am not sure if this depends on other parts of the system being
> initialized
> > > > first? If so, we can move this from here into OnDispatcherListChanged()
> > > > override.
> > > 
> > > Before landing this, we need to fix the platform to build the
> > EventFactoryOzone
> > > in the browser process only.
> > > 
> > > Maybe we can have
> > > 
> > >   ui::OzonePlatform::GetInstance()->CreateEventFactoryOzone()
> > > 
> > > inside aura::Env::Init()?
> > > 
> > > Otherwise, GPU process will break.
> > 
> > Ah yes. I remember the issue now. I was thinking we could have explicit
> > OzonePlatform::InitializeForUI() and InitializeForGPU(), and have the
platform
> > take care of initializing the right set of subsystems. We can then have the
> > InitializeForUI() done in Env::Init, and InitializeForGPU() in <insert
> > appropriate place ... GpuMain?>
> 
> Ok.

This is now done. Does this CL now seem ready to land?

[spang, 2014-05-16 15:01:16]

lgtm

https://codereview.chromium.org/285303004/diff/20001/ui/events/ozone/evdev/ev...
File ui/events/ozone/evdev/event_factory_evdev.cc (left):

https://codereview.chromium.org/285303004/diff/20001/ui/events/ozone/evdev/ev...
ui/events/ozone/evdev/event_factory_evdev.cc:138:
ui_task_runner_(base::MessageLoopProxy::current()),
I think you need to remove this initializer.

[sadrul, 2014-05-16 16:34:14]

https://codereview.chromium.org/285303004/diff/20001/ui/events/ozone/evdev/ev...
File ui/events/ozone/evdev/event_factory_evdev.cc (left):

https://codereview.chromium.org/285303004/diff/20001/ui/events/ozone/evdev/ev...
ui/events/ozone/evdev/event_factory_evdev.cc:138:
ui_task_runner_(base::MessageLoopProxy::current()),
On 2014/05/16 15:01:17, spang wrote:
> I think you need to remove this initializer.

Done.

[sadrul, 2014-05-16 16:57:12]

Committed patchset #3 manually as r271036 (presubmit successful).