Adds a porting layer so aura can be made to work with mus

Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org

Committed: https://crrev.com/631094cdb13348d8946c753c10a8f8bd9fe353ef
Cr-Commit-Position: refs/heads/master@{#427750}

[sky, 2016-10-25 15:46:56]

Description was changed from

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org
==========

to

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org
==========

[sky, 2016-10-25 15:46:56]

sky@chromium.org changed reviewers:
+ sadrul@chromium.org

[sky, 2016-10-25 15:47:16]

The CQ bit was checked by sky@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-25 15:47:42]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-25 16:41:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-25 16:41:06]

Dry run: This issue passed the CQ dry run.

[sadrul, 2016-10-26 04:25:55]

I have some comments, which are really just nits. Also comment re naming: does
WindowPortLocal vs. WindowPortRemote (where WindowPortRemote is the mus-version)
work? aura::WindowPortAura feels a bit awkward. As with all naming, I don't feel
strongly about it.

lgtm

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.cc
File ui/aura/window.cc (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.cc#newco...
ui/aura/window.cc:70: // See comemnt in header as to why this is done.
*comment

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.h
File ui/aura/window.h (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.h#newcod...
ui/aura/window.h:475: WindowPort* port_;
Should we still use a unique_ptr<> here, and std::move() that out in the dtor
with the explanation there?

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h
File ui/aura/window_port.h (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:40: // If a Window is created without an explicit
WindwPort then
*WindowPort

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:49: virtual std::unique_ptr<WindowPortInitData>
Init(Window* window) = 0;
Alternatively, this could be OnPreInit() and OnPostInit(), and since each
WindowPort services only a single window, it can manage the init-data by itself.

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:84: std::unique_ptr<WindowPortPropertyData> data) = 0;
Same comment here re: WindowPortPropertyData.

[sky, 2016-10-26 16:25:21]

I renamed WindowPortAura to WindowPortLocal, as I agree that name is better. I'm
going to keep the name WindowPortMus though as I think that better conveys the
backend vs WindowPortRemote.

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.cc
File ui/aura/window.cc (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.cc#newco...
ui/aura/window.cc:70: // See comemnt in header as to why this is done.
On 2016/10/26 04:25:55, sadrul wrote:
> *comment

Done.

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.h
File ui/aura/window.h (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window.h#newcod...
ui/aura/window.h:475: WindowPort* port_;
On 2016/10/26 04:25:55, sadrul wrote:
> Should we still use a unique_ptr<> here, and std::move() that out in the dtor
> with the explanation there?

I still need the ability to get the port_ during destruction. I changed this to
a unique_ptr and add a WindowPort* member that references the port_.get().

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h
File ui/aura/window_port.h (right):

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:40: // If a Window is created without an explicit
WindwPort then
On 2016/10/26 04:25:55, sadrul wrote:
> *WindowPort

Done.

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:49: virtual std::unique_ptr<WindowPortInitData>
Init(Window* window) = 0;
On 2016/10/26 04:25:55, sadrul wrote:
> Alternatively, this could be OnPreInit() and OnPostInit(), and since each
> WindowPort services only a single window, it can manage the init-data by
itself.

The reason I created the structure is so that WindowPortMus doesn't need to
track state that is only used during init. That is, WindowPortMus needs
calculate some values that are only needed in PostInit. If I didn't have the
data type that is returned from OnPreInit and fed into OnPostInit then
WindowPortMus would need to keep around data that is only used during init.

I renamed the members.

https://codereview.chromium.org/2446893005/diff/20001/ui/aura/window_port.h#n...
ui/aura/window_port.h:84: std::unique_ptr<WindowPortPropertyData> data) = 0;
On 2016/10/26 04:25:55, sadrul wrote:
> Same comment here re: WindowPortPropertyData.

Same comment as above. Allowing this to pass around data makes it easy for
windowportmus.

[sky, 2016-10-26 16:25:29]

The CQ bit was checked by sky@chromium.org

[sky, 2016-10-26 16:25:30]

The patchset sent to the CQ was uploaded after l-g-t-m from sadrul@chromium.org
Link to the patchset: https://codereview.chromium.org/2446893005/#ps40001
(title: "feedback")

[commit-bot: I haz the power, 2016-10-26 16:25:50]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sky, 2016-10-26 16:51:28]

The CQ bit was unchecked by sky@chromium.org

[sky, 2016-10-26 17:00:05]

The CQ bit was checked by sky@chromium.org

[sky, 2016-10-26 17:00:06]

The patchset sent to the CQ was uploaded after l-g-t-m from sadrul@chromium.org
Link to the patchset: https://codereview.chromium.org/2446893005/#ps80001
(title: "better")

[commit-bot: I haz the power, 2016-10-26 17:00:45]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-26 18:21:34]

Description was changed from

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org
==========

to

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org
==========

[commit-bot: I haz the power, 2016-10-26 18:21:35]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-10-26 18:53:12]

Description was changed from

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org
==========

to

==========
Adds a porting layer so aura can be made to work with mus

This introduces WindowPort and the single concrete implementation of
it WindowPortAura. WindowPort contains the set of functions needed to
get aura to work with mus. A new constructor is added to Window and
WindowTreeNhost that can be used to specify an explicit
WindowPort. This is not the common case though, the common case is to
use a constructor that does not take a WindowPort, in which case a
WindowPort is created using a callback specified at the time of
creating Env.

BUG=659155
TEST=covered by existing tests
sadrul@chromium.org

Committed: https://crrev.com/631094cdb13348d8946c753c10a8f8bd9fe353ef
Cr-Commit-Position: refs/heads/master@{#427750}
==========

[commit-bot: I haz the power, 2016-10-26 18:53:14]

Patchset 5 (id:??) landed as
https://crrev.com/631094cdb13348d8946c753c10a8f8bd9fe353ef
Cr-Commit-Position: refs/heads/master@{#427750}

[fwang, 2016-10-26 19:08:06]

fwang@igalia.com changed reviewers:
+ fwang@igalia.com