Aura: Add Window::MoveCursorTo() taking relative location to the window.

Aura: Add Window::MoveCursorTo() taking relative location to the window.

Previously, although RootWindow::MoveCursorTo() takes a ponit coordinate relative to Root, but OnMouseEvent/LockMouse in RWHVA are using coordinate relative to parent window (which may not be root). UnlockMouse is using global X/Y from webkit. 

This CL added Window::MoveCursorTo() taking relative location to the window, and made RWHVA use it.

BUG=none
TEST=manual

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=142888

[oshima, 2012-06-15 08:53:23]

thanks. can you add unit tests for MoveCursorTo?

[yoshiki, 2012-06-15 13:49:52]

oshima: Added a test. PTAL.

[oshima, 2012-06-15 14:06:17]

https://chromiumcodereview.appspot.com/10543174/diff/10001/ui/aura/window_uni...
File ui/aura/window_unittest.cc (right):

https://chromiumcodereview.appspot.com/10543174/diff/10001/ui/aura/window_uni...
ui/aura/window_unittest.cc:325: EXPECT_EQ(gfx::Point(35, 35),
root->last_mouse_location());
can you also add cases that applies transforms (rotation, scaling etc) to
windows?

[yoshiki, 2012-06-15 15:10:39]

Added tests of transform. And changed RootWindow::MoveCursorTo() to treat
root_window's transform as http://codereview.chromium.org/10388141/.

https://chromiumcodereview.appspot.com/10543174/diff/10001/ui/aura/window_uni...
File ui/aura/window_unittest.cc (right):

https://chromiumcodereview.appspot.com/10543174/diff/10001/ui/aura/window_uni...
ui/aura/window_unittest.cc:325: EXPECT_EQ(gfx::Point(35, 35),
root->last_mouse_location());
On 2012/06/15 14:06:17, oshima wrote:
> can you also add cases that applies transforms (rotation, scaling etc) to
> windows?

Done.

[oshima, 2012-06-15 16:57:01]

https://chromiumcodereview.appspot.com/10543174/diff/18001/ui/aura/window_uni...
File ui/aura/window_unittest.cc (right):

https://chromiumcodereview.appspot.com/10543174/diff/18001/ui/aura/window_uni...
ui/aura/window_unittest.cc:318: EXPECT_EQ(gfx::Point(10, 10),
root->last_mouse_location());
could you please use the style
EXPECT_EQ("10,10", root->last_mouse_location().ToString());
?
This makes it easier to read error messages.

https://chromiumcodereview.appspot.com/10543174/diff/18001/ui/aura/window_uni...
ui/aura/window_unittest.cc:332: transform0.ConcatRotate(-45.0f);
non orthogonal angles makes it difficult to verify that this test is correct.
Just use -90. same for others.

https://chromiumcodereview.appspot.com/10543174/diff/18001/ui/aura/window_uni...
ui/aura/window_unittest.cc:363: EXPECT_EQ(gfx::Point(15, 33),
root->last_mouse_location());
It's hard to tell these tests are correct. It'd be better to keep test for
complex transforms and test for multiple transforms separated.

[yoshiki, 2012-06-18 06:55:36]

oshima: Thanks, PTAL

http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc
File ui/aura/window_unittest.cc (right):

http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc...
ui/aura/window_unittest.cc:318: EXPECT_EQ(gfx::Point(10, 10),
root->last_mouse_location());
On 2012/06/15 16:57:01, oshima wrote:
> could you please use the style
> EXPECT_EQ("10,10", root->last_mouse_location().ToString());
> ?
> This makes it easier to read error messages.

Done.

http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc...
ui/aura/window_unittest.cc:332: transform0.ConcatRotate(-45.0f);
On 2012/06/15 16:57:01, oshima wrote:
> non orthogonal angles makes it difficult to verify that this test is correct.
> Just use -90. same for others.

Done.

http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc...
ui/aura/window_unittest.cc:363: EXPECT_EQ(gfx::Point(15, 33),
root->last_mouse_location());
On 2012/06/15 16:57:01, oshima wrote:
> It's hard to tell these tests are correct. It'd be better to keep test for
> complex transforms and test for multiple transforms separated.

Done.

[oshima, 2012-06-18 15:41:13]

lgtm

On Sun, Jun 17, 2012 at 11:55 PM, <yoshiki@chromium.org> wrote:

> oshima: Thanks, PTAL
>
>
> http://codereview.chromium.**org/10543174/diff/18001/ui/**
>
aura/window_unittest.cc<http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc>
> File ui/aura/window_unittest.cc (right):
>
> http://codereview.chromium.**org/10543174/diff/18001/ui/**
>
aura/window_unittest.cc#**newcode318<http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc#newcode318>
> ui/aura/window_unittest.cc:**318: EXPECT_EQ(gfx::Point(10, 10),
> root->last_mouse_location());
> On 2012/06/15 16:57:01, oshima wrote:
>
>> could you please use the style
>> EXPECT_EQ("10,10", root->last_mouse_location().**ToString());
>> ?
>> This makes it easier to read error messages.
>>
>
> Done.
>
> http://codereview.chromium.**org/10543174/diff/18001/ui/**
>
aura/window_unittest.cc#**newcode332<http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc#newcode332>
> ui/aura/window_unittest.cc:**332: transform0.ConcatRotate(-45.**0f);
> On 2012/06/15 16:57:01, oshima wrote:
>
>> non orthogonal angles makes it difficult to verify that this test is
>>
> correct.
>
>> Just use -90. same for others.
>>
>
> Done.
>
> http://codereview.chromium.**org/10543174/diff/18001/ui/**
>
aura/window_unittest.cc#**newcode363<http://codereview.chromium.org/10543174/diff/18001/ui/aura/window_unittest.cc#newcode363>
> ui/aura/window_unittest.cc:**363: EXPECT_EQ(gfx::Point(15, 33),
> root->last_mouse_location());
> On 2012/06/15 16:57:01, oshima wrote:
>
>> It's hard to tell these tests are correct. It'd be better to keep test
>>
> for
>
>> complex transforms and test for multiple transforms separated.
>>
>
> Done.
>
>
http://codereview.chromium.**org/10543174/<http://codereview.chromium.org/105...
>

[yoshiki, 2012-06-18 15:49:06]

ben: Could you take a look and approve of check-in?

[Ben Goodger (Google), 2012-06-18 16:50:21]

http://codereview.chromium.org/10543174/diff/25001/ui/aura/root_window.h
File ui/aura/root_window.h (right):

http://codereview.chromium.org/10543174/diff/25001/ui/aura/root_window.h#newc...
ui/aura/root_window.h:104: RootWindowHost* host() { return host_.get(); }
can we expose QueryMouseLocation() on RootWindow rather than exposing the whole
interface?

[yoshiki, 2012-06-18 18:49:39]

ben: PTAL

http://codereview.chromium.org/10543174/diff/25001/ui/aura/root_window.h
File ui/aura/root_window.h (right):

http://codereview.chromium.org/10543174/diff/25001/ui/aura/root_window.h#newc...
ui/aura/root_window.h:104: RootWindowHost* host() { return host_.get(); }
On 2012/06/18 16:50:21, Ben Goodger (Google) wrote:
> can we expose QueryMouseLocation() on RootWindow rather than exposing the
whole
> interface?

Done.

[Ben Goodger (Google), 2012-06-18 20:30:52]

LGTM

http://codereview.chromium.org/10543174/diff/17008/ui/aura/root_window.cc
File ui/aura/root_window.cc (right):

http://codereview.chromium.org/10543174/diff/17008/ui/aura/root_window.cc#new...
ui/aura/root_window.cc:657:
////////////////////////////////////////////////////////////////////////////////
delete these two lines

http://codereview.chromium.org/10543174/diff/17008/ui/aura/root_window.h
File ui/aura/root_window.h (right):

http://codereview.chromium.org/10543174/diff/17008/ui/aura/root_window.h#newc...
ui/aura/root_window.h:267: // Exposes RootWindowHost::QueryMouseLocation() for
test purpose.
purposes