linux mandoline: Fix double characters in omnibox.

linux mandoline: Fix double characters in omnibox.

In https://codereview.chromium.org/476283006, I added a quick hack that
fixed double characters when an IME was not present on Linux. This
removes that hack now that we initialize our IME subsystem.

BUG=none

Committed: https://crrev.com/6bf197c4f3436f32e7ebd8dc14cd85d42f018163
Cr-Commit-Position: refs/heads/master@{#338744}

[Elliot Glaysher, 2015-07-13 23:04:14]

erg@chromium.org changed reviewers:
+ ben@chromium.org

[Ben Goodger (Google), 2015-07-13 23:08:27]

ben@chromium.org changed reviewers:
+ sadrul@chromium.org

[Ben Goodger (Google), 2015-07-13 23:08:27]

replacing myself with sadrul since he recently added this

[sadrul, 2015-07-13 23:31:38]

I actually thought we would need something like InputMethodEventHandler (see
https://codereview.chromium.org/1209663002) for WindowTreeHostMojo.

[Elliot Glaysher, 2015-07-14 21:48:33]

On 2015/07/13 23:31:38, sadrul wrote:
> I actually thought we would need something like InputMethodEventHandler (see
> https://codereview.chromium.org/1209663002) for WindowTreeHostMojo.

So I'm a little confused here after reading that patch.

After building synthetic events, it appears on ash that we have an EventHandler
which just filters out synthetic events from ever reaching the IME. But it
doesn't stop propagation so I assume these synthetic events then get handled
somewhere else for some other reasons. Do those reasons hold outside of
something chromeos like?

The double character in omnibox bug that I'm seeing only exists on Linux;
whatever Windows is doing, it doesn't repro there.

[sadrul, 2015-07-14 21:49:48]

lgtm

[sadrul, 2015-07-14 21:56:20]

On 2015/07/14 21:48:33, Elliot Glaysher wrote:
> On 2015/07/13 23:31:38, sadrul wrote:
> > I actually thought we would need something like InputMethodEventHandler (see
> > https://codereview.chromium.org/1209663002) for WindowTreeHostMojo.
> 
> So I'm a little confused here after reading that patch.
> 
> After building synthetic events, it appears on ash that we have an
EventHandler
> which just filters out synthetic events from ever reaching the IME. But it
> doesn't stop propagation so I assume these synthetic events then get handled
> somewhere else for some other reasons. Do those reasons hold outside of
> something chromeos like?
> 
> The double character in omnibox bug that I'm seeing only exists on Linux;
> whatever Windows is doing, it doesn't repro there.

/yeah

[sadrul, 2015-07-14 21:57:53]

On 2015/07/14 21:48:33, Elliot Glaysher wrote:
> On 2015/07/13 23:31:38, sadrul wrote:
> > I actually thought we would need something like InputMethodEventHandler (see
> > https://codereview.chromium.org/1209663002) for WindowTreeHostMojo.
> 
> So I'm a little confused here after reading that patch.
> 
> After building synthetic events, it appears on ash that we have an
EventHandler
> which just filters out synthetic events from ever reaching the IME. But it
> doesn't stop propagation so I assume these synthetic events then get handled
> somewhere else for some other reasons. Do those reasons hold outside of
> something chromeos like?
> 
> The double character in omnibox bug that I'm seeing only exists on Linux;
> whatever Windows is doing, it doesn't repro there.

The fix you have look correct. The issue I was thinking of regarding
IME/InputMethodEventHandler should affect all platforms; but since we don't have
double-echo on Windows (or android with Peng's patch), that's probably not the
issue here.

[Elliot Glaysher, 2015-07-14 22:00:52]

The CQ bit was checked by erg@chromium.org

[commit-bot: I haz the power, 2015-07-14 22:01:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1232703006/1

[commit-bot: I haz the power, 2015-07-14 22:14:35]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-07-14 22:16:14]

Patchset 1 (id:??) landed as
https://crrev.com/6bf197c4f3436f32e7ebd8dc14cd85d42f018163
Cr-Commit-Position: refs/heads/master@{#338744}