Passing BackgroundColorSpan and UnderlineSpan from Clank to Blink.

Passing BackgroundColorSpan and UnderlineSpan from Clank to Blink.

This CL enables custom color to be specified for IME composition.
Its Blink counterpart is https://codereview.chromium.org/313233002/ . Details:

- Prerequesite: the Blink change, since we assume WebCore::CompositionUnderline and
blink::WebCompositionUnderline have been updated.
- Prerequesite CL: https://codereview.chromium.org/319553002/ to allow
structures to temporarily mismatch, without triggering compiler assert.
- Adding background_color to ui::CompositionUnderline.
- To pass data from Android to Native C++, using JNI generator with callback: In C++, ImeAdapterAndroid::SetComposingText() calls Java to iterate over spans, then dispatch BackgroundColorSpan and UnderlineSpan data to C++ code and populate list of blink::WebCompositionUnderline.

We'll need to split this CL when we commit.

BUG=135900
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=278794

[huangs, 2014-06-04 22:30:54]

PTAL, noting the Blink counterpart https://codereview.chromium.org/313233002/ .

[huangs, 2014-06-06 15:49:49]

Ping.

[yosin_UTC9, 2014-06-09 01:34:37]

Rubber stump ACK

I'm not familiar with files in here. But, it seems OK. Please wait aurimas@ for
reviewing Android Java.

[aurimas (slooooooooow), 2014-06-10 00:30:49]

https://chromiumcodereview.appspot.com/313053007/diff/20001/content/browser/r...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://chromiumcodereview.appspot.com/313053007/diff/20001/content/browser/r...
content/browser/renderer_host/ime_adapter_android.cc:72: jclass
clsSpannableString = env->FindClass("android/text/SpannableString");
In Chrome for Android we avoid using JNI FindClass and GetMethodID calls
directly with hardcoded strings as it often leads to bugs. Instead we using JNI
generator to create safe accessors. There is some info on this here:
http://www.chromium.org/developers/design-documents/android-jni

See comments below how to fix it.

https://chromiumcodereview.appspot.com/313053007/diff/20001/content/browser/r...
content/browser/renderer_host/ime_adapter_android.cc:201:
GetUnderlinesFromSpanArray(env, spannableString, spanObjArray, &underlines);
Here you should call Java_ImeAdapter_getUnderlinesFromSpan(spannableString,
&underlines).

In java you would write a function something like this:
@CalledByNative
void static getUnderlinesFromSpan(SpannableString string, long underlines) {
  for (Span s: ...) {
    nativeAppendUnderline(underlines, s.getStart(), s.getEnd(), s.getColor());
  }
}

And finally create a static function in ime_adapter_android.cc AppendUnderline
that would do push_back on that array with given values.

[huangs, 2014-06-10 04:39:13]

Updated, PTAL.

https://codereview.chromium.org/313053007/diff/20001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/20001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:72: jclass
clsSpannableString = env->FindClass("android/text/SpannableString");
On 2014/06/10 00:30:48, aurimas wrote:
> In Chrome for Android we avoid using JNI FindClass and GetMethodID calls
> directly with hardcoded strings as it often leads to bugs. Instead we using
JNI
> generator to create safe accessors. There is some info on this here:
> http://www.chromium.org/developers/design-documents/android-jni
> 
> See comments below how to fix it.

Done.

https://codereview.chromium.org/313053007/diff/20001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:201:
GetUnderlinesFromSpanArray(env, spannableString, spanObjArray, &underlines);
Done.  I'm looping over Object[] because in the near future I want to generalize
this code to handle UnderlineSpan as well.

[huangs, 2014-06-10 18:06:35]

Updated CL to make AppendBackgroundColorSpan static.

Added support of UnderlineSpan, which is passed by Google Japan Input.
- made underline color of BackgroundColorSpan transparent to promote separation
of concern.
- inclusion of UnderlineSpan fixes the "2 pixel gap" problem between successive
instances of BackgroundColorSpan.

PTAL.

[huangs, 2014-06-10 18:09:47]

https://codereview.chromium.org/313053007/diff/40001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.h (right):

https://codereview.chromium.org/313053007/diff/40001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.h:45: void
AppendBackgroundColorSpan(JNIEnv*,
Cannot forward declare "public static" methods, since this is "extern", which
conflicts with "static" declarations produced by JNI Generator.  Filed this
issue as  http://crbug.com/382985 .

[aurimas (slooooooooow), 2014-06-10 18:36:19]

https://codereview.chromium.org/313053007/diff/40001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.h (right):

https://codereview.chromium.org/313053007/diff/40001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.h:45: void
AppendBackgroundColorSpan(JNIEnv*,
On 2014/06/10 18:09:47, huangs1 wrote:
> Cannot forward declare "public static" methods, since this is "extern", which
> conflicts with "static" declarations produced by JNI Generator.  Filed this
> issue as  http://crbug.com/382985 .

Why does it have to be public? It will only be used inside of this class to call
ime_adapter_android.cc so you can make it private.

See example here of nativeIsEditBookmarksEnabled()
https://code.google.com/p/chromium/codesearch#chromium/src/chrome/android/jav...

https://codereview.chromium.org/313053007/diff/60001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/60001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:92: void
AppendBackgroundColorSpan(JNIEnv*,
Please add comments explaining what these functions do.

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/AdapterInputConnection.java
(right):

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/AdapterInputConnection.java:22:
import java.lang.CharSequence;
Is this import needed?

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:534:
SpannableString ss = ((SpannableString) text);
According to the style guide [1] you should not be using cryptic names such as
"ss" or "sObj", "s", etc. Please rename them to something clearer.

Also keep in mind that Java allows 100 character lines.

http://google-styleguide.googlecode.com/svn/trunk/javaguide.html#s5.2.7-local...

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:535:
for (Object sObj : ss.getSpans(0, textStr.length(), Object.class)) {
Both BackgroundColorSpan and UnderlineSpan extend CharacterStyle. Can you used
that instead of plain Object type?

[huangs, 2014-06-10 19:54:53]

Updated, PTAL.

https://codereview.chromium.org/313053007/diff/60001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/60001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:92: void
AppendBackgroundColorSpan(JNIEnv*,
On 2014/06/10 18:36:18, aurimas wrote:
> Please add comments explaining what these functions do.

Done.

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/AdapterInputConnection.java
(right):

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/AdapterInputConnection.java:22:
import java.lang.CharSequence;
Not needed to compile, but I lapsed into "include what you use" mode in C++.  :)
 Removed.

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:534:
SpannableString ss = ((SpannableString) text);
On 2014/06/10 18:36:18, aurimas wrote:
> According to the style guide [1] you should not be using cryptic names such as
> "ss" or "sObj", "s", etc. Please rename them to something clearer.
> 
> Also keep in mind that Java allows 100 character lines.
> 
>
http://google-styleguide.googlecode.com/svn/trunk/javaguide.html#s5.2.7-local...

Done.

https://codereview.chromium.org/313053007/diff/60001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:535:
for (Object sObj : ss.getSpans(0, textStr.length(), Object.class)) {
On 2014/06/10 18:36:18, aurimas wrote:
> Both BackgroundColorSpan and UnderlineSpan extend CharacterStyle. Can you used
> that instead of plain Object type?

Done.

[aurimas (slooooooooow), 2014-06-10 20:06:36]

lgtm for the following files:
ImeAdapter.java
AdapterInputConnection.java
ime_adapter_android.*

+tedchoc for another quick look over Androidy bits
+nona for review in ui/base/ime/*
+palmer for review of content/common/content_param_traits_macros.h
+piman for content/*

https://codereview.chromium.org/313053007/diff/80001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/80001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:532:
if (!(text instanceof SpannableString))
Can we add assert : text instanceof SpannableString if you are not sure that the
correct type will be passed in?

Also in java we either put if statement in one line or use curly braces. In this
case you can move "return;" to the same line as if.

[huangs, 2014-06-10 20:23:56]

Updated.

https://codereview.chromium.org/313053007/diff/80001/content/public/android/j...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/80001/content/public/android/j...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:532:
if (!(text instanceof SpannableString))
I think here we should be robust rather than fragile, since |text| comes form
arbitrary Android IME stuff.

Replaced early "out-of-the-way check" return with surrounding if {}.

[palmer, 2014-06-10 20:32:23]

IPC LGTM. Some comments about correct integer type.

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:97: jint start,
I'm not sure, but should |start| and |end| be jlongs (for when we have 64-bit
Clank)?

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:112: jint start,
Same thing here. (I could be wrong.)

https://codereview.chromium.org/313053007/diff/80001/ui/base/ime/composition_...
File ui/base/ime/composition_underline.h (right):

https://codereview.chromium.org/313053007/diff/80001/ui/base/ime/composition_...
ui/base/ime/composition_underline.h:32: CompositionUnderline(unsigned s,
unsigned e, SkColor c, bool t, SkColor bc)
Ahh, I see. Since these values cross the native/Java barrier, it might be better
to be explicit about their widths. uint32_t and juint, are 32-bit unsigned, for
example.

[Ted C, 2014-06-10 20:47:33]

lgtm w/ some style nits

https://codereview.chromium.org/313053007/diff/100001/content/browser/rendere...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/100001/content/browser/rendere...
content/browser/renderer_host/ime_adapter_android.cc:104: start, end,
SK_ColorTRANSPARENT, false, background_color));
indented 4 too much  same for the underlines push_back below

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:531:
String textStr = text.toString();
You don't need textStr do you?  CharSequence has .length(), and that is the only
thing you use this for below.

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:532:
if (text instanceof SpannableString) {
Even if we don't add the assert, I think we should leave it as an early out
(with the return on the same line as Aurimas suggested).

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:539:
spannableString.getSpanEnd(span),
+4 indent on these two lines (same for the line below for
nativeAppendUnderlineSpan)

[piman, 2014-06-10 23:00:07]

LGTM modulo existing style comments (run through git cl format?)

[huangs, 2014-06-11 02:41:02]

Updated.  Please note that this CL won't even compile without the prerequesite
WebKit CL

https://codereview.chromium.org/313233002/

I'd appreciate help review that as well.  :)

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:97: jint start,
SpannableString.getSpanStart(), SpannableString.getSpanEnd() return int, and
blink::WebCompositionUnderline stores |startOffset| and |endOffset| as unsigned.

So I think the interface should be jint (instead of juint), and we shouldn't
make the data 64-bit and then 32-bit again?

However, I'll add explicit static_cast<unsigned>.

https://codereview.chromium.org/313053007/diff/80001/content/browser/renderer...
content/browser/renderer_host/ime_adapter_android.cc:112: jint start,
Same thing.

https://codereview.chromium.org/313053007/diff/80001/ui/base/ime/composition_...
File ui/base/ime/composition_underline.h (right):

https://codereview.chromium.org/313053007/diff/80001/ui/base/ime/composition_...
ui/base/ime/composition_underline.h:32: CompositionUnderline(unsigned s,
unsigned e, SkColor c, bool t, SkColor bc)
Please note the comment below in line 51 stating why "unsigned" was used.

Previously std::vector<ui::CompositionUnderlines> was reinterpret_casted to
std::vector<blink::WebCompositionUnderline>.  *However*, in

https://codereview.chromium.org/319553002

I got rid of the code.  So I'll go with using uint32 (since this is used often
in the module), and do explicit conversion in render_widget_host_view_aura.cc.

Also updated imm32_manager.cc to reflect this.

https://codereview.chromium.org/313053007/diff/100001/content/browser/rendere...
File content/browser/renderer_host/ime_adapter_android.cc (right):

https://codereview.chromium.org/313053007/diff/100001/content/browser/rendere...
content/browser/renderer_host/ime_adapter_android.cc:104: start, end,
SK_ColorTRANSPARENT, false, background_color));
On 2014/06/10 20:47:33, Ted C wrote:
> indented 4 too much  same for the underlines push_back below

Done.

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:531:
String textStr = text.toString();
On 2014/06/10 20:47:33, Ted C wrote:
> You don't need textStr do you?  CharSequence has .length(), and that is the
only
> thing you use this for below.

Done.

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:532:
if (text instanceof SpannableString) {
Done.

https://codereview.chromium.org/313053007/diff/100001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:539:
spannableString.getSpanEnd(span),
On 2014/06/10 20:47:33, Ted C wrote:
> +4 indent on these two lines (same for the line below for
> nativeAppendUnderlineSpan)

Done.

[huangs, 2014-06-11 02:42:00]

Adding sadrul@chromium.org for OWNER review for

ui\aura\remote_window_tree_host_win.cc

[sadrul, 2014-06-11 21:49:39]

rslgtm

https://codereview.chromium.org/313053007/diff/160001/ui/aura/remote_window_t...
File ui/aura/remote_window_tree_host_win.cc (right):

https://codereview.chromium.org/313053007/diff/160001/ui/aura/remote_window_t...
ui/aura/remote_window_tree_host_win.cc:80:
composition_text->underlines[i].background_color = SK_ColorTRANSPARENT;
I assume you are not adding a background-color to metro_viewer::UnderlineInfo
because it also lacks a foreground-color?

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
File ui/base/ime/composition_underline.h (right):

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
ui/base/ime/composition_underline.h:23: background_color(0) {}
SK_ColorTRANSPARENT?

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
ui/base/ime/composition_underline.h:31: background_color(0) {}
ditto

[palmer, 2014-06-11 22:31:46]

When you do your explicit conversion between signed and unsigned, it might make
sense to do a little sanity check at that time (e.g. >= 0 and <
SOME_REASONABLE_MAX).

[huangs, 2014-06-12 20:25:35]

Updated. With regard to explicit conversion: there are some contexts:
- For color: any --> any: no check needed, since we really care about 32-bit
patterns.
- For positions: jint --> unsigned: should check >= 0, since upper-bound works.
- uint32 --> unsigned: we assume they're the same size, so no bounding?

https://codereview.chromium.org/313053007/diff/160001/ui/aura/remote_window_t...
File ui/aura/remote_window_tree_host_win.cc (right):

https://codereview.chromium.org/313053007/diff/160001/ui/aura/remote_window_t...
ui/aura/remote_window_tree_host_win.cc:80:
composition_text->underlines[i].background_color = SK_ColorTRANSPARENT;
Originally yes, so I dismissed it as irrelevant.  I just saw the comment in

ui/metro_viewer/ime_types.h

states that metro_viewer::UnderlineInfo is a copy of ui::CompositionUnderline. 
But doing so would require dependency on

#include "third_party/skia/include/core/SkColor.h"

I think adding support will open a can or worms, and I prefer to save the
cleanup to a future CL if you're okay with that.

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
File ui/base/ime/composition_underline.h (right):

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
ui/base/ime/composition_underline.h:23: background_color(0) {}
Done, also replaced color(0).

https://codereview.chromium.org/313053007/diff/160001/ui/base/ime/composition...
ui/base/ime/composition_underline.h:31: background_color(0) {}
On 2014/06/11 21:49:39, sadrul wrote:
> ditto

Done.

[huangs, 2014-06-19 18:34:53]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2014-06-19 18:37:21]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/huangs@chromium.org/313053007/200001

[huangs, 2014-06-19 21:53:18]

Ping nona@ for OWNER review of 

ui/base/ime/composition_text_util_pango.cc
ui/base/ime/composition_text_util_pango_unittest.cc
ui/base/ime/composition_underline.h
ui/base/ime/input_method_chromeos.cc
ui/base/ime/input_method_chromeos_unittest.cc
ui/base/ime/win/imm32_manager.cc

PTAL.  Thanks!

[commit-bot: I haz the power, 2014-06-19 23:07:48]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  android_aosp on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/android_aosp/builds/8...)
  chromium_presubmit on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/chromium_presubmit/bu...)
  linux_chromium_chromeos_clang_dbg on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/linux_chromium_chrome...)
  linux_chromium_chromeos_rel on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/linux_chromium_chrome...)

[commit-bot: I haz the power, 2014-06-19 23:12:26]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-06-19 23:12:28]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/chromium_presubmit/bu...)
  linux_chromium_chromeos_clang_dbg on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/linux_chromium_chrome...)

[Seigo Nonaka, 2014-06-20 06:39:13]

lgtm

lgtm with few comments :)

https://chromiumcodereview.appspot.com/313053007/diff/200001/content/public/a...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://chromiumcodereview.appspot.com/313053007/diff/200001/content/public/a...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:22:
import java.lang.CharSequence;
nit: You no need to import java.lang.* packages :)

https://chromiumcodereview.appspot.com/313053007/diff/200001/ui/base/ime/comp...
File ui/base/ime/composition_underline.h (right):

https://chromiumcodereview.appspot.com/313053007/diff/200001/ui/base/ime/comp...
ui/base/ime/composition_underline.h:26: CompositionUnderline(uint32 s, uint32 e,
SkColor c, bool t)
nit: Can we put only background color for composition?
If so, I think this is no longer Underline, probably good to rename to
CompositionDecoration.
No need to do now but could you file an issue and leave issue ID here?

[huangs, 2014-06-20 14:28:25]

Updated. Gonna commit!

https://codereview.chromium.org/313053007/diff/200001/content/public/android/...
File
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java
(right):

https://codereview.chromium.org/313053007/diff/200001/content/public/android/...
content/public/android/java/src/org/chromium/content/browser/input/ImeAdapter.java:22:
import java.lang.CharSequence;
On 2014/06/20 06:39:13, Seigo Nonaka wrote:
> nit: You no need to import java.lang.* packages :)

Done.

https://codereview.chromium.org/313053007/diff/200001/ui/base/ime/composition...
File ui/base/ime/composition_underline.h (right):

https://codereview.chromium.org/313053007/diff/200001/ui/base/ime/composition...
ui/base/ime/composition_underline.h:26: CompositionUnderline(uint32 s, uint32 e,
SkColor c, bool t)
Filed issue:
http://crbug.com/387032

A more ambitious refactoring would be to:
- unify the duplicate data structures to a unique structure that eveyrone uses
- create a container class (instead of just vector of CompositionDecorations) so
we can use polymorphism (or as flyweight) to store different types of
Decorations.

[huangs, 2014-06-20 14:58:00]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2014-06-20 15:00:04]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/huangs@chromium.org/313053007/240001

[huangs, 2014-06-20 15:45:52]

The CQ bit was unchecked by huangs@chromium.org

[huangs, 2014-06-20 15:58:35]

Turns out we need 

import java.lang.CharSequence

in ImeAdapter.java.  Otherwise android_aosp produces JNI error.

[huangs, 2014-06-20 15:58:49]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2014-06-20 16:00:42]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/huangs@chromium.org/313053007/260001

[commit-bot: I haz the power, 2014-06-20 19:23:52]

Change committed as 278794