Fix invalidations after ChromeInvalidationClientService introduction.

Fix invalidations after ChromeInvalidationClientService introduction.

This change, http://crrev.com/2297193003, introduced a chrome
specific InvalidationClientService, but it did not properly update
all the appropriate references.

This pulls the registered service from the AndroidManifest.xml if
applicable.  Does it through this indirection as components/ is
unaware of chrome/.

BUG=647774
Committed: https://crrev.com/c622eb2571e5a9c659c2aee70086af636472032f
Cr-Commit-Position: refs/heads/master@{#420718}

[Ted C, 2016-09-20 00:22:48]

tedchoc@chromium.org changed reviewers:
+ dcheng@chromium.org, maxbogue@chromium.org, nyquist@chromium.org

[Ted C, 2016-09-20 00:22:49]

dcheng@ && nyquist@ - OWNERS (and any historical insight of other places I
missed)

maxbogue@ - FYI and other points to investigate.

More context (for the non-nyquist@ folk), my patch
(https://codereview.chromium.org/2297193003) was the first step to remove the
browser specific bits from ChromeApplication.  Applications in Android are
initialized for every process, so we were doing browser specific stuff for every
renderer, gpu, utility processes.

As such, I had to do the early initialization "somewhere".  For sync, I couldn't
have the InvalidationClientService call into Chrome code, so I introduced
ChromeInvalidationClientService to handle this browser specific initialization:
https://cs.chromium.org/chromium/src/chrome/android/java/src/org/chromium/chr....

Unfortunately, I didn't realize that we hardcoded the .class name in the code so
many places and were using that to start the service.  To work around this, I
pull the registered service from the application and use it (which allows
components/ to be unaware of the chrome specific bits).

I didn't update the tests as of yet, but wanted to get your first round of
thoughts if you thought this was reasonable.

[nyquist, 2016-09-20 00:45:02]

this looks sad, but reasonable to me.

[Ted C, 2016-09-20 16:57:06]

Updated w/ the test fixes

[Ted C, 2016-09-20 16:57:11]

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

[commit-bot: I haz the power, 2016-09-20 16:57:46]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-20 17:27:09]

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

[commit-bot: I haz the power, 2016-09-20 17:27:10]

Dry run: This issue passed the CQ dry run.

[maxbogue, 2016-09-20 18:37:16]

lgtm; this is somehow both incredibly hacky and less hacky than before.

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
File
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java
(right):

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java:116:
public static Class<? extends InvalidationClientService>
getInvalidationClientServiceClass() {
nit: Feel free to ignore but personally I'd just call the method getClass() or
getClassFromManifest() since
InvalidationClientService.getInvalidationClientServiceClass() is a bit
redundant.

[Ted C, 2016-09-20 20:03:40]

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
File
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java
(right):

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java:116:
public static Class<? extends InvalidationClientService>
getInvalidationClientServiceClass() {
On 2016/09/20 18:37:15, maxbogue wrote:
> nit: Feel free to ignore but personally I'd just call the method getClass() or
> getClassFromManifest() since
> InvalidationClientService.getInvalidationClientServiceClass() is a bit
> redundant.

How about getRegisteredClass?  That keeps it inline with the comment.

[maxbogue, 2016-09-20 20:41:56]

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
File
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java
(right):

https://codereview.chromium.org/2352943002/diff/20001/components/invalidation...
components/invalidation/impl/android/java/src/org/chromium/components/invalidation/InvalidationClientService.java:116:
public static Class<? extends InvalidationClientService>
getInvalidationClientServiceClass() {
On 2016/09/20 20:03:40, Ted C wrote:
> On 2016/09/20 18:37:15, maxbogue wrote:
> > nit: Feel free to ignore but personally I'd just call the method getClass()
or
> > getClassFromManifest() since
> > InvalidationClientService.getInvalidationClientServiceClass() is a bit
> > redundant.
> 
> How about getRegisteredClass?  That keeps it inline with the comment.

sgtm!

[dcheng, 2016-09-21 01:08:40]

I'm not very familiar with the Java side of things here, so I'll defer to
nyquist@

[maxbogue, 2016-09-23 18:40:41]

Just a heads up that invalidations are completely broken on Canary until this
lands!

[nyquist, 2016-09-23 19:43:23]

The CQ bit was checked by nyquist@chromium.org

[nyquist, 2016-09-23 19:43:24]

lgtm

[nyquist, 2016-09-23 19:43:24]

The patchset sent to the CQ was uploaded after l-g-t-m from
maxbogue@chromium.org
Link to the patchset: https://codereview.chromium.org/2352943002/#ps40001
(title: "Change method name")

[commit-bot: I haz the power, 2016-09-23 19:44:07]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-23 20:34:11]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-09-23 20:35:42]

Description was changed from

==========
Fix invalidations after ChromeInvalidationClientService introduction.

This change, http://crrev.com/2297193003, introduced a chrome
specific InvalidationClientService, but it did not properly update
all the appropriate references.

This pulls the registered service from the AndroidManifest.xml if
applicable.  Does it through this indirection as components/ is
unaware of chrome/.

BUG=647774
==========

to

==========
Fix invalidations after ChromeInvalidationClientService introduction.

This change, http://crrev.com/2297193003, introduced a chrome
specific InvalidationClientService, but it did not properly update
all the appropriate references.

This pulls the registered service from the AndroidManifest.xml if
applicable.  Does it through this indirection as components/ is
unaware of chrome/.

BUG=647774

Committed: https://crrev.com/c622eb2571e5a9c659c2aee70086af636472032f
Cr-Commit-Position: refs/heads/master@{#420718}
==========

[commit-bot: I haz the power, 2016-09-23 20:35:43]

Patchset 3 (id:??) landed as
https://crrev.com/c622eb2571e5a9c659c2aee70086af636472032f
Cr-Commit-Position: refs/heads/master@{#420718}