Provide dummy Android UI thread with looper for C++ unit tests

Provide dummy Android UI thread with looper for C++ unit tests

This is is intended for C++ unit tests (e.g. the net unit tests) that don't run
with the UI thread as their main looper, but test code that, on Android, uses Ui
thread events, so need a running Ui thread.

Committed: https://crrev.com/91ea3f95a19219abd7b1f309db084fddcc5a4ce2
Cr-Commit-Position: refs/heads/master@{#333734}

[aberent, 2015-06-08 13:56:17]

aberent@chromium.org changed reviewers:
+ nyquist@chromium.org, phajdan.jr@chromium.org

[aberent, 2015-06-08 13:56:18]

nyquist@chromium.org: Please review all changes.

phajdan.jr@chromium.org: OWNER review please for base/test

This is motived the need to create tests for Kerberos for Android
(https://codereview.chromium.org/1128043007/).

See also
http://www.chromium.org/developers/testing/android-tests/testing-android-code...;
in particular the TBA at the bottom.

[aberent, 2015-06-08 14:28:09]

aberent@chromium.org changed reviewers:
+ thestig@chromium.org

[aberent, 2015-06-08 14:28:09]

thestig@chromium.org: Please review changes, as OWNER, in base/base.gyp

[aberent, 2015-06-08 14:28:39]

The CQ bit was checked by aberent@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-06-08 14:28:56]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1162763004/20001

[commit-bot: I haz the power, 2015-06-08 16:42:35]

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

[commit-bot: I haz the power, 2015-06-08 16:42:36]

Dry run: This issue passed the CQ dry run.

[Lei Zhang, 2015-06-08 21:15:56]

gyp / gn changes lgtm

[nyquist, 2015-06-08 23:38:19]

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
File base/test/android/java/src/org/chromium/base/TestUiThread.java (right):

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:14: * code that,
on Android, uses Ui thread events, so need a running Ui thread.
Nit: UI*

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:17: static
boolean sStarted = false;
Nit: private

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:21: //
setUiThread can only be called once in a test run, so do this once, and leave it
running.
Nit: @{link ThreadUtils#setUiThread(Looper)}

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:22: if (sStarted)
return;
It's not clear to me that we are guaranteed to get the call to loop() from the
same thread every time. There seems to be a race condition here where multiple
threads could be started and the last call to setUiThread(...) would win.

Could we do something simple like using an AtomicBoolean?

if (sStarted.getAndSet(false)) return;

Or of course some other way you'd prefer to make calls to this method thread
safe.

I'd mildly argue for also adding @ThreadSafe annotation to the class at that
point to make it clear that you've thought about that case.

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:42: //
TODO(aberent): Auto-generated catch block
Nit: Could we log an error message here instead of printStackTrace?

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
File base/test/test_ui_thread_android.h (right):

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
base/test/test_ui_thread_android.h:5: #ifndef BASE_TEST_ANDROID_TEST_UI_THREAD_
Nit: BASE_TEST_TEST_UI_THREAD_ANDROID_H_

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
base/test/test_ui_thread_android.h:14: // UI thread as their main looper, but
test code that, on Android, uses Ui
Nit: UI* here and below

[aberent, 2015-06-09 10:33:13]

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
File base/test/android/java/src/org/chromium/base/TestUiThread.java (right):

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:14: * code that,
on Android, uses Ui thread events, so need a running Ui thread.
On 2015/06/08 23:38:19, nyquist wrote:
> Nit: UI*

Done.

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:17: static
boolean sStarted = false;
On 2015/06/08 23:38:19, nyquist wrote:
> Nit: private

Done.

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:22: if (sStarted)
return;
On 2015/06/08 23:38:19, nyquist wrote:
> It's not clear to me that we are guaranteed to get the call to loop() from the
> same thread every time. There seems to be a race condition here where multiple
> threads could be started and the last call to setUiThread(...) would win.
> 
> Could we do something simple like using an AtomicBoolean?
> 
> if (sStarted.getAndSet(false)) return;
> 
> Or of course some other way you'd prefer to make calls to this method thread
> safe.
> 
> I'd mildly argue for also adding @ThreadSafe annotation to the class at that
> point to make it clear that you've thought about that case.

Done.

https://codereview.chromium.org/1162763004/diff/20001/base/test/android/java/...
base/test/android/java/src/org/chromium/base/TestUiThread.java:42: //
TODO(aberent): Auto-generated catch block
On 2015/06/08 23:38:19, nyquist wrote:
> Nit: Could we log an error message here instead of printStackTrace?

Done.

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
File base/test/test_ui_thread_android.h (right):

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
base/test/test_ui_thread_android.h:5: #ifndef BASE_TEST_ANDROID_TEST_UI_THREAD_
On 2015/06/08 23:38:19, nyquist wrote:
> Nit: BASE_TEST_TEST_UI_THREAD_ANDROID_H_

Done.

https://codereview.chromium.org/1162763004/diff/20001/base/test/test_ui_threa...
base/test/test_ui_thread_android.h:14: // UI thread as their main looper, but
test code that, on Android, uses Ui
On 2015/06/08 23:38:19, nyquist wrote:
> Nit: UI* here and below

Done.

[aberent, 2015-06-09 10:33:24]

The CQ bit was checked by aberent@chromium.org to run a CQ dry run

[aberent, 2015-06-09 10:33:25]

The patchset sent to the CQ was uploaded after l-g-t-m from thestig@chromium.org
Link to the patchset: https://codereview.chromium.org/1162763004/#ps40001
(title: "Respond to nyquist@'s comments.")

[commit-bot: I haz the power, 2015-06-09 10:33:49]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1162763004/40001

[commit-bot: I haz the power, 2015-06-09 11:36:32]

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

[commit-bot: I haz the power, 2015-06-09 11:36:34]

Dry run: This issue passed the CQ dry run.

[nyquist, 2015-06-09 14:08:46]

lgtm

[Paweł Hajdan Jr., 2015-06-10 10:29:53]

LGTM

[aberent, 2015-06-10 13:12:45]

The CQ bit was checked by aberent@chromium.org

[commit-bot: I haz the power, 2015-06-10 13:12:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1162763004/40001

[commit-bot: I haz the power, 2015-06-10 15:03:42]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-06-10 15:04:43]

Patchset 3 (id:??) landed as
https://crrev.com/91ea3f95a19219abd7b1f309db084fddcc5a4ce2
Cr-Commit-Position: refs/heads/master@{#333734}