Consider IME is associated with the top level widget rather than child widget

Consider IME is associated with the top level widget rather than child widget.

An important thing is that [NativeWidget A] owns Win32 input focus even
when [View X] is logically focused by FocusManager. As a result, an Win32
IME may want to interact with the native view of [NativeWidget A] rather
than that of [NativeWidget B]. This is why we need to call
GetTopLevelWidget() here.

Previous implementation treated [NativeWidget B] as associated with
the IME but it was wrong. This CL fixes the this confusion.

BUG=246534
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=208018

[Seigo Nonaka, 2013-06-19 08:45:32]

lgtm

[Yohei Yukawa, 2013-06-19 08:50:59]

+msw@ for the owner review.

The motivation of my change is described in the description.
Could you take a look?

[msw, 2013-06-19 16:15:50]

LGTM, but I'm CC'ing Oshima and Peter in case they have any comments.

[oshima, 2013-06-19 16:49:43]

+yusukes who knows IME stuff better. One comment about the
bug description/comment in the code.

The description isn't correct. NativeWidget B can get focus
and input in Aura. It is true that only toplevel widget has an InputMethod
object, so this change is probably
correct though.

[Yohei Yukawa, 2013-06-19 19:31:54]

On 2013/06/19 16:49:43, oshima wrote:
> +yusukes who knows IME stuff better. One comment about the
> bug description/comment in the code.
> 
> The description isn't correct. NativeWidget B can get focus
> and input in Aura. It is true that only toplevel widget has an InputMethod
> object, so this change is probably
> correct though.

Thank you for the info. Updated the description and comment.

yusukes: Could you take a look?

[Yohei Yukawa, 2013-06-20 22:14:22]

On 2013/06/19 19:31:54, Yohei Yukawa wrote:
> On 2013/06/19 16:49:43, oshima wrote:
> > +yusukes who knows IME stuff better. One comment about the
> > bug description/comment in the code.
> > 
> > The description isn't correct. NativeWidget B can get focus
> > and input in Aura. It is true that only toplevel widget has an InputMethod
> > object, so this change is probably
> > correct though.
> 
> Thank you for the info. Updated the description and comment.
> 
> yusukes: Could you take a look?

yusukes: ping?
I'd like to commit this CL into M29. So it would be helpful if you can review
this CL by its branch cut.
Thanks!

[Yusuke Sato, 2013-06-21 04:36:34]

On 2013/06/20 22:14:22, Yohei Yukawa wrote:
> On 2013/06/19 19:31:54, Yohei Yukawa wrote:
> > On 2013/06/19 16:49:43, oshima wrote:
> > > +yusukes who knows IME stuff better. One comment about the
> > > bug description/comment in the code.
> > > 
> > > The description isn't correct. NativeWidget B can get focus
> > > and input in Aura. It is true that only toplevel widget has an InputMethod
> > > object, so this change is probably
> > > correct though.
> > 
> > Thank you for the info. Updated the description and comment.
> > 
> > yusukes: Could you take a look?
> 
> yusukes: ping?
> I'd like to commit this CL into M29. So it would be helpful if you can review
> this CL by its branch cut.
> Thanks!

lgtm

[Yohei Yukawa, 2013-06-21 04:47:38]

On 2013/06/21 04:36:34, Yusuke Sato wrote:
> On 2013/06/20 22:14:22, Yohei Yukawa wrote:
> > On 2013/06/19 19:31:54, Yohei Yukawa wrote:
> > > On 2013/06/19 16:49:43, oshima wrote:
> > > > +yusukes who knows IME stuff better. One comment about the
> > > > bug description/comment in the code.
> > > > 
> > > > The description isn't correct. NativeWidget B can get focus
> > > > and input in Aura. It is true that only toplevel widget has an
InputMethod
> > > > object, so this change is probably
> > > > correct though.
> > > 
> > > Thank you for the info. Updated the description and comment.
> > > 
> > > yusukes: Could you take a look?
> > 
> > yusukes: ping?
> > I'd like to commit this CL into M29. So it would be helpful if you can
review
> > this CL by its branch cut.
> > Thanks!
> 
> lgtm

Thank you for the review.

Oshima-san, please let me know if the code comment/CL description are still not
correct. I'll commit this CL in 17 hours. Thanks!

[commit-bot: I haz the power, 2013-06-22 10:36:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/Yukawa@chromium.org/17101019/8002

[commit-bot: I haz the power, 2013-06-22 13:37:59]

Change committed as 208018