Use XInput2 events for keyboard events.

Use XInput2 events for keyboard events.

Patch set 1: Use XInput2 events for keyboard events.

XI2 keyboard events are generated, rewritten to core events, and consumed. EventRewriter tests pass with XI2 key events.

Patch set 2: Move source_device_id_ up to |ui::Event|.

Patch set 3: EventRewriter is no longer a PlatformEventObserver or DeviceHierarchyObserver.

|chromeos::EventRewriter| now uses |ui::KeyEvent::source_device_id()| instead of tracking XI2 key events itself as a chromeos::DeviceHierarchyObserver and |ui::PlatformEventObserver|.

Patch set 4: Convert Alt + Left Button rewriter from XI2 events to|ui::Event|s.

BUG=368750
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=282717

[Daniel Erat, 2014-06-25 17:13:25]

this generally lgtm with some comments

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:253: ui::KeyEvent*
rewritten_key_event = 0;
nit: initialize to NULL instead of 0

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:263: xkeyevent.xkey =
xev->xkey;
nit: mind adding a blank line after this one to set this if/else apart from the
rest of the block?

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:387: int device_id =
key_event.source_device_id();
nit: i don't think gets used later in the method, so it'd probably be cleaner to
just do:

  if (key_event.source_device_id() != ui::ED_UNKNOWN_DEVICE)

on the next line instead

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:877: // Remap modifier masks.
nit: don't need this comment now

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:883: // Then, remap Alt+Button1
to Button3.
nit: remove the "Then, " here; it doesn't make much sense now that it's at the
top of a method

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter.h (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.h:84: static void
RewrittenKeyEvent(const ui::KeyEvent& key_event,
did you mean to name this RewriteKeyEvent()? an earlier comment refers to that,
and it seems like a better name

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.h:171: std::map<int, DeviceType>
device_id_to_type_;
nit: add a blank line below this one

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter_unittest.cc (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter_unittest.cc:301:
CheckKeyTestCase(1000 + i, &rewriter, pc_keyboard_tests[i]);
nit: mind adding a SCOPED_TRACE() call at the beginning of this block to make it
easier to debug test failures? or is the ID already enough for that? (if you add
this, there are a few more loops farther down that should probably get it too)

https://codereview.chromium.org/336403005/diff/80001/ui/events/event.h
File ui/events/event.h (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/event.h#newcod...
ui/events/event.h:242: int source_device_id_;
nit: is the "source" prefix here necessary, or would a shorter "device_id" name
be equally understandable?

(after writing this, i saw that this member already exists and is being moved,
so feel free to ignore that suggestion if it'd require updating a bunch of
existing code)

https://codereview.chromium.org/336403005/diff/80001/ui/events/keycodes/keybo...
File ui/events/keycodes/keyboard_code_conversion_x.cc (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/keycodes/keybo...
ui/events/keycodes/keyboard_code_conversion_x.cc:450: XKeyEvent* xkey;
nit: mind initializing this to NULL here?

https://codereview.chromium.org/336403005/diff/80001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/x/events_x.cc#...
ui/events/x/events_x.cc:59: NOTREACHED();
nit: probably don't need this, right? there's a default case above, and
presumably the compiler would complain if it were possible to reach this line
(since it isn't returning anything)

[kpschoedel, 2014-06-25 19:59:49]

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:253: ui::KeyEvent*
rewritten_key_event = 0;
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: initialize to NULL instead of 0

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:263: xkeyevent.xkey =
xev->xkey;
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: mind adding a blank line after this one to set this if/else apart from
the
> rest of the block?

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:387: int device_id =
key_event.source_device_id();
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: i don't think gets used later in the method, so it'd probably be cleaner
to
> just do:
> 
>   if (key_event.source_device_id() != ui::ED_UNKNOWN_DEVICE)
> 
> on the next line instead

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:877: // Remap modifier masks.
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: don't need this comment now

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.cc:883: // Then, remap Alt+Button1
to Button3.
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: remove the "Then, " here; it doesn't make much sense now that it's at the
> top of a method

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter.h (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.h:84: static void
RewrittenKeyEvent(const ui::KeyEvent& key_event,
On 2014/06/25 17:13:24, Daniel Erat wrote:
> did you mean to name this RewriteKeyEvent()? an earlier comment refers to
that, and it seems like a better name

This is the function that constructs a new key event for a new key code and/or
event flags, and is called at the end of both EventRewriter::RewriteKeyEvent()
and KeyboardDrivenEventRewriter::Rewrite(). I've changed the name and comment to
try to make that clearer, and will happily accept further suggestions to that
end.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.h:171: std::map<int, DeviceType>
device_id_to_type_;
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: add a blank line below this one

Done.

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter_unittest.cc (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter_unittest.cc:301:
CheckKeyTestCase(1000 + i, &rewriter, pc_keyboard_tests[i]);
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: mind adding a SCOPED_TRACE() call at the beginning of this block to make
it
> easier to debug test failures? or is the ID already enough for that?

Thanks; I only added the ID because I wasn't aware of SCOPED_TRACE().

https://codereview.chromium.org/336403005/diff/80001/ui/events/event.h
File ui/events/event.h (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/event.h#newcod...
ui/events/event.h:242: int source_device_id_;
On 2014/06/25 17:13:24, Daniel Erat wrote:
> nit: is the "source" prefix here necessary, or would a shorter "device_id"
name
> be equally understandable?
> 
> (after writing this, i saw that this member already exists and is being moved,
> so feel free to ignore that suggestion if it'd require updating a bunch of
> existing code)

Yes, touch events already had this. I'll defer to Sadrul on the name.

https://codereview.chromium.org/336403005/diff/80001/ui/events/keycodes/keybo...
File ui/events/keycodes/keyboard_code_conversion_x.cc (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/keycodes/keybo...
ui/events/keycodes/keyboard_code_conversion_x.cc:450: XKeyEvent* xkey;
On 2014/06/25 17:13:25, Daniel Erat wrote:
> nit: mind initializing this to NULL here?

Done.

https://codereview.chromium.org/336403005/diff/80001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/80001/ui/events/x/events_x.cc#...
ui/events/x/events_x.cc:59: NOTREACHED();
On 2014/06/25 17:13:25, Daniel Erat wrote:
> nit: probably don't need this, right? there's a default case above, and
> presumably the compiler would complain if it were possible to reach this line
> (since it isn't returning anything)

Done.

[Daniel Erat, 2014-06-25 20:06:13]

lgtm

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
File chrome/browser/chromeos/events/event_rewriter.h (right):

https://codereview.chromium.org/336403005/diff/80001/chrome/browser/chromeos/...
chrome/browser/chromeos/events/event_rewriter.h:84: static void
RewrittenKeyEvent(const ui::KeyEvent& key_event,
On 2014/06/25 19:59:48, kpschoedel wrote:
> On 2014/06/25 17:13:24, Daniel Erat wrote:
> > did you mean to name this RewriteKeyEvent()? an earlier comment refers to
> that, and it seems like a better name
> 
> This is the function that constructs a new key event for a new key code and/or
> event flags, and is called at the end of both EventRewriter::RewriteKeyEvent()
> and KeyboardDrivenEventRewriter::Rewrite(). I've changed the name and comment
to
> try to make that clearer, and will happily accept further suggestions to that
> end.

great, thanks! (i mostly just wanted there to be a verb in the name, and "Build"
fits the bill)

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter.cc:254: ui::KeyEvent*
rewritten_key_event = 0;
nit: this should still be NULL

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter_unittest.cc (right):

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter_unittest.cc:296: SCOPED_TRACE(1000
+ i);
nit: probably don't need the "1000" now, right? i think that the actual call to
CheckKeyTestCase will be included in the failure now, so just calling
SCOPED_TRACE(i) should suffice

[kpschoedel, 2014-06-25 21:10:40]

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter.cc:254: ui::KeyEvent*
rewritten_key_event = 0;
On 2014/06/25 20:06:13, Daniel Erat wrote:
> nit: this should still be NULL

Done.

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter_unittest.cc (right):

https://codereview.chromium.org/336403005/diff/100001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter_unittest.cc:296: SCOPED_TRACE(1000
+ i);
On 2014/06/25 20:06:13, Daniel Erat wrote:
> nit: probably don't need the "1000" now, right? i think that the actual call
to
> CheckKeyTestCase will be included in the failure now, so just calling
> SCOPED_TRACE(i) should suffice

Done.

[sadrul, 2014-06-26 12:34:20]

https://codereview.chromium.org/336403005/diff/120001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/120001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter.cc:166: #endif
change this line to:
#endif  // defined(USE_X11)

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
File ui/aura/window_tree_host_x11.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
ui/aura/window_tree_host_x11.cc:379: if (base::SysInfo::IsRunningOnChromeOS())
We should avoid using this, because it makes it very difficult to test this code
locally during development.

I believe the XI2 key-events we currently receive come to the X11 root window
(because of SelectXInput2EventsForRootWindow()). If we listen to XI2 key-events
on the WindowTreeHost's |x_window_| instead, then this wouldn't be necessary?

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.cc
File ui/events/event.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.cc#newc...
ui/events/event.cc:571: native_event()->type == GenericEvent);
Also add a check that native_event()->xgeneric.evtype == XI_KeyPress or
XI_KeyRelease

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.h
File ui/events/event.h (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.h#newco...
ui/events/event.h:73: void set_source_device_id(int source_device_id) {
Move this to EventTestApi

https://codereview.chromium.org/336403005/diff/120001/ui/events/keycodes/keyb...
File ui/events/keycodes/keyboard_code_conversion_x.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/keycodes/keyb...
ui/events/keycodes/keyboard_code_conversion_x.cc:906: case XI_KeyPress:
should be 2 more indent

https://codereview.chromium.org/336403005/diff/120001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:63: (native_event->type == KeyRelease));
I suspect this will need to change once you make the changes in
window-tree-host-x11.

[kpschoedel, 2014-06-26 17:24:10]

https://codereview.chromium.org/336403005/diff/120001/chrome/browser/chromeos...
File chrome/browser/chromeos/events/event_rewriter.cc (right):

https://codereview.chromium.org/336403005/diff/120001/chrome/browser/chromeos...
chrome/browser/chromeos/events/event_rewriter.cc:166: #endif
On 2014/06/26 12:34:19, sadrul wrote:
> change this line to:
> #endif  // defined(USE_X11)

Done.

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
File ui/aura/window_tree_host_x11.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
ui/aura/window_tree_host_x11.cc:379: if (base::SysInfo::IsRunningOnChromeOS())
On 2014/06/26 12:34:20, sadrul wrote:
> We should avoid using this, because it makes it very difficult to test this
code
> locally during development.
> 
> I believe the XI2 key-events we currently receive come to the X11 root window
> (because of SelectXInput2EventsForRootWindow()). If we listen to XI2
key-events
> on the WindowTreeHost's |x_window_| instead, then this wouldn't be necessary?

I don't follow how that helps. We still have to choose, somewhere, whether to
process core key events or XI2 key events.

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.cc
File ui/events/event.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.cc#newc...
ui/events/event.cc:571: native_event()->type == GenericEvent);
On 2014/06/26 12:34:20, sadrul wrote:
> Also add a check that native_event()->xgeneric.evtype == XI_KeyPress or
> XI_KeyRelease

Done.

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.h
File ui/events/event.h (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/event.h#newco...
ui/events/event.h:73: void set_source_device_id(int source_device_id) {
On 2014/06/26 12:34:20, sadrul wrote:
> Move this to EventTestApi

Done.

https://codereview.chromium.org/336403005/diff/120001/ui/events/keycodes/keyb...
File ui/events/keycodes/keyboard_code_conversion_x.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/keycodes/keyb...
ui/events/keycodes/keyboard_code_conversion_x.cc:906: case XI_KeyPress:
On 2014/06/26 12:34:20, sadrul wrote:
> should be 2 more indent

Done.

https://codereview.chromium.org/336403005/diff/120001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:63: (native_event->type == KeyRelease));
On 2014/06/26 12:34:20, sadrul wrote:
> I suspect this will need to change once you make the changes in
> window-tree-host-x11.

XI2 handling uses UpdateStateFromXGenericEvent; this one is specifically for
|XKeyEvent|.

[sadrul, 2014-06-26 17:28:07]

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
File ui/aura/window_tree_host_x11.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
ui/aura/window_tree_host_x11.cc:379: if (base::SysInfo::IsRunningOnChromeOS())
On 2014/06/26 17:24:10, kpschoedel wrote:
> On 2014/06/26 12:34:20, sadrul wrote:
> > We should avoid using this, because it makes it very difficult to test this
> code
> > locally during development.
> > 
> > I believe the XI2 key-events we currently receive come to the X11 root
window
> > (because of SelectXInput2EventsForRootWindow()). If we listen to XI2
> key-events
> > on the WindowTreeHost's |x_window_| instead, then this wouldn't be
necessary?
> 
> I don't follow how that helps. We still have to choose, somewhere, whether to
> process core key events or XI2 key events.

We don't have to make such decisions for mouse events. Why would we have to make
this decision for key events?

[kpschoedel, 2014-06-26 19:10:59]

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
File ui/aura/window_tree_host_x11.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
ui/aura/window_tree_host_x11.cc:379: if (base::SysInfo::IsRunningOnChromeOS())
On 2014/06/26 17:28:07, sadrul wrote:
> We don't have to make such decisions for mouse events. Why would we have to
make
> this decision for key events?

OK... I see that no core ButtonPress events arrive for mice, but searching the
code I don't see why not. Do you happen to know?

[kpschoedel, 2014-06-26 21:50:40]

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
File ui/aura/window_tree_host_x11.cc (right):

https://codereview.chromium.org/336403005/diff/120001/ui/aura/window_tree_hos...
ui/aura/window_tree_host_x11.cc:379: if (base::SysInfo::IsRunningOnChromeOS())
Never mind, got it now.

[sadrul, 2014-07-10 16:03:06]

https://codereview.chromium.org/336403005/diff/320001/ui/events/keycodes/keyb...
File ui/events/keycodes/keyboard_code_conversion_x.h (right):

https://codereview.chromium.org/336403005/diff/320001/ui/events/keycodes/keyb...
ui/events/keycodes/keyboard_code_conversion_x.h:34: void
InitXKeyEventFromXIDeviceEvent(const XEvent& src, XEvent* dst);
You likely need to EXPORT this?

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:80: void UpdateStateFromXGenericEvent(const
base::NativeEvent& native_event) {
We are duping some amount of code and comment here. Can we keep them in the same
function?

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:96: if (native_event->type == XI_KeyPress)
native_event->type is going to be GenericEvent here, right?

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:387: } else {
Remove the else here, and just do a 'return ET_MOUSE_MOVED;', because otherwise
it looks like there's fall-through code in here.

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:652: return copy;
We should not make a copy of a XI2 event here. What code needs this? (e.g.
simply copying XIDeviceEvent may not be sufficient for copying
XIButtonState/XIValuatorState correctly)

[kpschoedel, 2014-07-10 17:22:40]

https://codereview.chromium.org/336403005/diff/320001/ui/events/keycodes/keyb...
File ui/events/keycodes/keyboard_code_conversion_x.h (right):

https://codereview.chromium.org/336403005/diff/320001/ui/events/keycodes/keyb...
ui/events/keycodes/keyboard_code_conversion_x.h:34: void
InitXKeyEventFromXIDeviceEvent(const XEvent& src, XEvent* dst);
On 2014/07/10 16:03:06, sadrul wrote:
> You likely need to EXPORT this?

Done.

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc
File ui/events/x/events_x.cc (right):

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:80: void UpdateStateFromXGenericEvent(const
base::NativeEvent& native_event) {
On 2014/07/10 16:03:06, sadrul wrote:
> We are duping some amount of code and comment here. Can we keep them in the
same
> function?

Done.

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:96: if (native_event->type == XI_KeyPress)
On 2014/07/10 16:03:06, sadrul wrote:
> native_event->type is going to be GenericEvent here, right?

Done.

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:387: } else {
On 2014/07/10 16:03:06, sadrul wrote:
> Remove the else here, and just do a 'return ET_MOUSE_MOVED;', because
otherwise
> it looks like there's fall-through code in here.

Done.

https://codereview.chromium.org/336403005/diff/320001/ui/events/x/events_x.cc...
ui/events/x/events_x.cc:652: return copy;
On 2014/07/10 16:03:06, sadrul wrote:
> We should not make a copy of a XI2 event here. What code needs this?

Evidently none that's covered by unit_tests ash_unittests ui_unittests or
events_unittests. Removed.

[sadrul, 2014-07-11 18:26:01]

LGTM

[kpschoedel, 2014-07-11 18:57:06]

The CQ bit was checked by kpschoedel@chromium.org

[commit-bot: I haz the power, 2014-07-11 18:58:21]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/kpschoedel@chromium.org/336403005/340001

[commit-bot: I haz the power, 2014-07-11 22:25:16]

Change committed as 282717