ozone: Interpret absolute events for graphics tablets

ozone: Interpret absolute events for graphics tablets

This change adds a TabletEventConverterEvdev that handles absolute evdev events
generated by graphics tablets (like a Wacom.) It maps absolute co-ordinate from
the input range into the screen that currently has the cursor, thus providing
multiple desktop support.

In order to handle the absolute position relative to the current screen a new
API call (GetCursorDisplayBounds) is added to EvdevCursorDelegate and its
implementation in DriCursor to provide the bounds of the curren display.
MoveCursorTo is also overloaded to add a version that reuses the existing
cursor window.

TEST=On link_freon with a Wacom graphics tablet behaves as per X11, stylus
mapped to active screen where cursor is present, tap emulates left mouse
button, stylus button emulates right mouse click. All other input devices
(keyboard, touchpad, touchscreen and mouse) all function as expected. New tests
added to event_unittests.
BUG=432547
Committed: https://crrev.com/d5e789eaa18748652ffb4211a73485fb59d2d886
Cr-Commit-Position: refs/heads/master@{#305013}

[robert.bradford, 2014-11-18 19:28:22]

robert.bradford@intel.com changed reviewers:
+ dnicoara@chromium.org, spang@chromium.org

[robert.bradford, 2014-11-18 19:28:23]

Hi PTAL at this patch to add support for Wacom tablets to ozone. 

I know the newly added method in EvdevCursorDelegate/DriCursor is a bit "weird"
and would love to get feedback on an alternative name.

Thanks in advance.

[spang, 2014-11-18 20:24:37]

On 2014/11/18 19:28:23, robert.bradford wrote:
> Hi PTAL at this patch to add support for Wacom tablets to ozone. 
> 
> I know the newly added method in EvdevCursorDelegate/DriCursor is a bit
"weird"
> and would love to get feedback on an alternative name.
> 
> Thanks in advance.

I think we can sidestep the naming issue by adding a different function.

// ui/events/ozone/evdev/cursor_delegate_evdev.h
// Get bounds of the display the cursor is under.
virtual gfx::RectF GetCursorDisplayBounds() = 0;

[spang, 2014-11-18 20:26:14]

On 2014/11/18 20:24:37, spang wrote:
> On 2014/11/18 19:28:23, robert.bradford wrote:
> > Hi PTAL at this patch to add support for Wacom tablets to ozone. 
> > 
> > I know the newly added method in EvdevCursorDelegate/DriCursor is a bit
> "weird"
> > and would love to get feedback on an alternative name.
> > 
> > Thanks in advance.
> 
> I think we can sidestep the naming issue by adding a different function.
> 
> // ui/events/ozone/evdev/cursor_delegate_evdev.h
> // Get bounds of the display the cursor is under.
> virtual gfx::RectF GetCursorDisplayBounds() = 0;

Er, bounds of the display the cursor is over.

[robert.bradford, 2014-11-19 19:19:05]

Hi Michael, thanks for the speedy feedback.

Here is an updated version that adds that method and uses it. I needed to
overload MoveCursorTo too so that it has a version that does not need a widget.

My gut feeling is that this is better than providing a getter for the cursor
window as i'm not sure it is useful for anything else outside this use case.

Do let me know if you'd prefer it that way and i'll change it (the method is
already available in DriCursor but not exposed through the cursor delegate.)

Thanks.

[spang, 2014-11-19 19:27:45]

lgtm

[robert.bradford, 2014-11-20 13:31:40]

The CQ bit was checked by robert.bradford@intel.com

[commit-bot: I haz the power, 2014-11-20 13:32:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/721823002/100001

[commit-bot: I haz the power, 2014-11-20 14:47:02]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2014-11-20 14:47:46]

Patchset 6 (id:??) landed as
https://crrev.com/d5e789eaa18748652ffb4211a73485fb59d2d886
Cr-Commit-Position: refs/heads/master@{#305013}