Add native X11 testing to chromeos::EventRewriter unit tests

ui/events/test/events_test_utils_x11.cc

Index: ui/events/test/events_test_utils_x11.cc
diff --git a/ui/events/test/events_test_utils_x11.cc b/ui/events/test/events_test_utils_x11.cc
index 76b4bce7511b33b20008e38c95d041a64c1d28f6..560689558dc94eacad0018b725b9bc883e18acf3 100644
--- a/ui/events/test/events_test_utils_x11.cc
+++ b/ui/events/test/events_test_utils_x11.cc
@@ -23,6 +23,10 @@ unsigned int XEventState(int flags) {
       ((flags & ui::EF_CONTROL_DOWN) ? ControlMask : 0) |
       ((flags & ui::EF_ALT_DOWN) ? Mod1Mask : 0) |
       ((flags & ui::EF_CAPS_LOCK_DOWN) ? LockMask : 0) |
+      ((flags & ui::EF_ALTGR_DOWN) ? Mod5Mask : 0) |
+      ((flags & ui::EF_COMMAND_DOWN) ? Mod4Mask : 0) |
+      ((flags & ui::EF_MOD3_DOWN) ? Mod3Mask : 0) |
+      ((flags & ui::EF_NUMPAD_KEY) ? Mod2Mask : 0) |

[sadrul, 2014/05/14 01:23:55]

We don't map Mod3 to MOD3 and Mod2 to NUMPAD in GetEventFlagsFromXState() (in
ui/events/x/events_x.cc). Should we?

[kpschoedel, 2014/05/14 02:23:57]

Mod3Mask, yes; it's set when the caps lock key is down, as distinct from
CapsLock, which is set when caps locking is active, which I didn't realize until
it caused a regression, so that is already in another CL:
https://codereview.chromium.org/277443004/

Mod2Mask is not present on input because Chrome doesn't manage NumLock state. It
does get set in the outgoing X11 events.

       ((flags & ui::EF_LEFT_MOUSE_BUTTON) ? Button1Mask: 0) |
       ((flags & ui::EF_MIDDLE_MOUSE_BUTTON) ? Button2Mask: 0) |
       ((flags & ui::EF_RIGHT_MOUSE_BUTTON) ? Button3Mask: 0);