[Sync] Fix invalidations on Android.

[Sync] Fix invalidations on Android.

This bug was preventing Chrome from registering for GCM invalidations on Android, meaning that sync was never told to download new data.

The bug was introduced by http://crrev.com/1048423005, where the ensureGcmIsInitialized line was removed from the SyncController constructor. The onFirstStart call was supposed to handle it, but from a clean start sync is disabled, so that case is not hit.

This patch moves the ensureGcmIsInitialized call back to the constructor, and the test ensures that the call is made at some point during browser initialization.

BUG=475299
Committed: https://crrev.com/dc18c01f1c98bc24fd88e19ff331c9dd573fd12c
Cr-Commit-Position: refs/heads/master@{#325359}

[nyquist, 2015-04-15 22:49:45]

nyquist@chromium.org changed reviewers:
+ nyquist@chromium.org

[nyquist, 2015-04-15 22:49:46]

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
File
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java
(right):

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java:85:
mChromeSigninController.ensureGcmIsInitialized();
Could you add a test for this?

Possibly just add an accessor for test in ChromeSigninController to see if GCM
has been (or should have been) initialized.

Or if you want to go further, mock out the GCMRegistrar, and ensure that we do
the right thing.

[nyquist, 2015-04-15 23:06:26]

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
File
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java
(right):

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java:267:
mChromeSigninController.ensureGcmIsInitialized();
Why is this still here?

[maxbogue, 2015-04-15 23:23:25]

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
File
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java
(right):

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java:85:
mChromeSigninController.ensureGcmIsInitialized();
On 2015/04/15 22:49:46, nyquist wrote:
> Could you add a test for this?
> 
> Possibly just add an accessor for test in ChromeSigninController to see if GCM
> has been (or should have been) initialized.
> 
> Or if you want to go further, mock out the GCMRegistrar, and ensure that we do
> the right thing.

Done.

https://codereview.chromium.org/1081283003/diff/1/chrome/android/java/src/org...
chrome/android/java/src/org/chromium/chrome/browser/sync/SyncController.java:267:
mChromeSigninController.ensureGcmIsInitialized();
On 2015/04/15 23:06:26, nyquist wrote:
> Why is this still here?

Wasn't sure whether to leave it in for the fix or not. Removed now.

[nyquist, 2015-04-15 23:56:30]

lgtm, but could you please expand the commit description?

Some things to add to the commit message:
- What did not work before?
- Why was it not working?
- How does this patch address the issue?
- How do we ensure that this never happens again?

[maxbogue, 2015-04-16 00:08:03]

The CQ bit was checked by maxbogue@chromium.org

[commit-bot: I haz the power, 2015-04-16 00:08:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1081283003/40001

[commit-bot: I haz the power, 2015-04-16 00:37:07]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-04-16 00:38:36]

Patchset 3 (id:??) landed as
https://crrev.com/dc18c01f1c98bc24fd88e19ff331c9dd573fd12c
Cr-Commit-Position: refs/heads/master@{#325359}