Fix composition key-code after switching to a new input field.

Fix composition key-code after switching to a new input field.

A line of code that resets the composition state upon
switching to a new field got removed by mistake during the
clean-up of dead code in the last CL.  The adds it back and
creates a test to make sure the switching works.

BUG=118639
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=286510

[jdduke (slow), 2014-07-28 14:41:42]

https://codereview.chromium.org/411383002/diff/1/content/public/android/java/...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/411383002/diff/1/content/public/android/java/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:439:
mLastComposeText = null;
Should this go before the native handle check?

[bcwhite, 2014-07-28 14:45:40]

https://codereview.chromium.org/411383002/diff/1/content/public/android/java/...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/411383002/diff/1/content/public/android/java/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:439:
mLastComposeText = null;
On 2014/07/28 14:41:41, jdduke wrote:
> Should this go before the native handle check?

Everything else is always done after...  but there's no reason not to always do
it; makes it fail-safe.  Done.

[bcwhite, 2014-07-29 14:50:27]

Ping?  All good now?

[bcwhite, 2014-07-29 14:57:47]

+aurimas (guru for IME)

[jdduke (slow), 2014-07-29 14:58:05]

On 2014/07/29 14:50:27, bcwhite wrote:
> Ping?  All good now?

Hmm, looks like your patch description is off? You'll want the title on the
first line, then the actual "description" below that (preferably limited to at
most 80 char column width).

[bcwhite, 2014-07-29 16:23:14]

> Hmm, looks like your patch description is off? You'll want the title on the
> first line, then the actual "description" below that (preferably limited to at
> most 80 char column width).

It seems that codereview.chromium.org has changed.  It used to auto-wrap the
description.  Inserting line-breaks made it look worse.

[aurimas (slooooooooow), 2014-07-29 16:26:39]

https://codereview.chromium.org/411383002/diff/20001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/411383002/diff/20001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:438:
mLastComposeText = null;
I'm not 100% convinced that every keyboard will call finishComposingText(). Are
the any other calls that might need to reset mLastComposeText? Maybe when we
construct a new AdapterInputConnection?

[bcwhite, 2014-07-29 19:55:19]

https://codereview.chromium.org/411383002/diff/20001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/411383002/diff/20001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:438:
mLastComposeText = null;
On 2014/07/29 16:26:39, aurimas wrote:
> I'm not 100% convinced that every keyboard will call finishComposingText().
Are
> the any other calls that might need to reset mLastComposeText? Maybe when we
> construct a new AdapterInputConnection?

Done.  Two additional resets added during initialization.

[aurimas (slooooooooow), 2014-07-29 20:07:45]

lgtm

[jdduke (slow), 2014-07-29 20:54:48]

On 2014/07/29 20:07:45, aurimas wrote:
> lgtm

rubberstamp lgtm

[bcwhite, 2014-07-30 12:40:34]

The CQ bit was checked by bcwhite@chromium.org

[commit-bot: I haz the power, 2014-07-30 12:41:39]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/bcwhite@chromium.org/411383002/40001

[commit-bot: I haz the power, 2014-07-30 14:16:52]

Change committed as 286510