Enable CursorManager::LockCursor to lock cursor visibility.

Enable CursorManager::LockCursor to lock cursor visibility.

This CL changes mainly three parts:
- Add LockCursor and UnlockCursor to CursorClient.
- Move the code to change cursor visibility from aura::RootWindow and 
   aura::RootWindowHostLinux to ash::CursorManager.
- Remove CursorDelegate, which was too much simple delegation and unnecessary.

BUG=153702
TEST=CursorManagerTest.ShowCursor


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=161454

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=161629

[mazda, 2012-10-09 18:03:27]

ping?

[oshima, 2012-10-09 18:59:44]

On 2012/10/09 18:03:27, mazda wrote:
> ping?

Sorry I somehow missed it. Will review it shortly.

[oshima, 2012-10-09 23:08:38]

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc
File ash/wm/cursor_manager.cc (right):

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc#ne...
ash/wm/cursor_manager.cc:15: void SetRootWindowsCursor(gfx::NativeCursor cursor)
{
SetCursorOnAllRootWindows would be better

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc#ne...
ash/wm/cursor_manager.cc:120: image_cursors_->GetDeviceScaleFactor());
can you explain why we need to set the scale factor to invisible ones? It's not
obvious as you can't see it anyway.

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.h
File ash/wm/cursor_manager.h (right):

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.h#new...
ash/wm/cursor_manager.h:33: void UnlockCursor();
Can you define these API in CursorClient? RWHVA needs this feature.

[mazda, 2012-10-10 00:59:12]

Please take another look.

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc
File ash/wm/cursor_manager.cc (right):

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc#ne...
ash/wm/cursor_manager.cc:15: void SetRootWindowsCursor(gfx::NativeCursor cursor)
{
On 2012/10/09 23:08:38, oshima wrote:
> SetCursorOnAllRootWindows would be better

Done.

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.cc#ne...
ash/wm/cursor_manager.cc:120: image_cursors_->GetDeviceScaleFactor());
On 2012/10/09 23:08:38, oshima wrote:
> can you explain why we need to set the scale factor to invisible ones? It's
not
> obvious as you can't see it anyway.

This isn't necessary in practice. I just wanted to make sure device scale factor
value is correct regardless of visibility. Removed.

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.h
File ash/wm/cursor_manager.h (right):

http://codereview.chromium.org/11035050/diff/9002/ash/wm/cursor_manager.h#new...
ash/wm/cursor_manager.h:33: void UnlockCursor();
On 2012/10/09 23:08:38, oshima wrote:
> Can you define these API in CursorClient? RWHVA needs this feature.

Done.
I'd like to add the implementations other than CursorManager after I refactor
CursorManager code furthermore for code sharing.

[oshima, 2012-10-10 01:24:41]

lgtm

http://codereview.chromium.org/11035050/diff/23002/ui/aura/client/cursor_clie...
File ui/aura/client/cursor_client.h (right):

http://codereview.chromium.org/11035050/diff/23002/ui/aura/client/cursor_clie...
ui/aura/client/cursor_client.h:30: // Locks the cursor change.
can you document how this interact with Show/Set Cursor?

[mazda, 2012-10-11 20:58:50]

+sky
Could you do an OWNERS review for ui/views, ash, and ui/aura?
As we talked offline, current lock/unlock behavior is not ideal. I'll file a bug
to improve it and work on it later.

http://codereview.chromium.org/11035050/diff/23002/ui/aura/client/cursor_clie...
File ui/aura/client/cursor_client.h (right):

http://codereview.chromium.org/11035050/diff/23002/ui/aura/client/cursor_clie...
ui/aura/client/cursor_client.h:30: // Locks the cursor change.
On 2012/10/10 01:24:41, oshima wrote:
> can you document how this interact with Show/Set Cursor?

Done.

[sky, 2012-10-11 21:42:34]

LGTM - I tend to think some of this code needs to be refactored for better
sharing, but that can be done later.

[mazda, 2012-10-11 21:46:22]

On 2012/10/11 21:42:34, sky wrote:
> LGTM - I tend to think some of this code needs to be refactored for better
> sharing, but that can be done later.

Yes, I'll refactor CursorManager's cursor lock logic and share it with other
platforms.

[commit-bot: I haz the power, 2012-10-11 21:58:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mazda@chromium.org/11035050/10008

[commit-bot: I haz the power, 2012-10-12 00:08:18]

Change committed as 161454

[commit-bot: I haz the power, 2012-10-12 16:49:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mazda@chromium.org/11035050/42001

[commit-bot: I haz the power, 2012-10-12 18:55:26]

Change committed as 161629