Resurrect battery_status_dispatcher_unittest.

Resurrect battery_status_dispatcher_unittest.

While still in the code base the battery_status_dispatcher_unittest.cc is
not included in content_unittests. This patch rewrites this test to match
the new mojo-implementation of the battery_status_dispatcher and includes
it in content_unittests.

BUG=

Committed: https://crrev.com/88d1c9a4473a40c07751a839ff22425eaa97a999
Cr-Commit-Position: refs/heads/master@{#310103}

[timvolodine, 2014-12-08 19:30:06]

timvolodine@chromium.org changed reviewers:
+ mvanouwerkerk@chromium.org

[Michael van Ouwerkerk, 2014-12-08 20:26:17]

lgtm but some bots are failing to build

[timvolodine, 2015-01-06 16:42:46]

On 2014/12/08 20:26:17, Michael van Ouwerkerk wrote:
> lgtm but some bots are failing to build

thanks Michael, the bots seem to be fine now (except one which seems unrelated).

[timvolodine, 2015-01-06 16:42:50]

The CQ bit was checked by timvolodine@chromium.org

[commit-bot: I haz the power, 2015-01-06 16:43:16]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/768113006/20001

[commit-bot: I haz the power, 2015-01-06 17:31:54]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-01-06 17:32:38]

Patchset 2 (id:??) landed as
https://crrev.com/88d1c9a4473a40c07751a839ff22425eaa97a999
Cr-Commit-Position: refs/heads/master@{#310103}

[darin (slow to review), 2015-01-07 06:39:28]

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

[darin (slow to review), 2015-01-07 06:39:29]

It occurs to me that this test is not really all that interesting. It seems to
just test conversion between device::BatteryStatus and blink::WebBatteryStatus.

The test has the side-effect of forcing BatteryStatusDispatcher to have to deal
with the ServiceRegistry not being present, which is not something that ever
happens when running inside Chrome.

Another approach might be to dependency inject the ServiceRegistry and then
provide one in the unit test along with a fake BatteryMonitor implementation.

[timvolodine, 2015-01-07 14:53:50]

On 2015/01/07 06:39:29, darin wrote:
> It occurs to me that this test is not really all that interesting. It seems to
> just test conversion between device::BatteryStatus and
blink::WebBatteryStatus.
> 
> The test has the side-effect of forcing BatteryStatusDispatcher to have to
deal
> with the ServiceRegistry not being present, which is not something that ever
> happens when running inside Chrome.
> 
> Another approach might be to dependency inject the ServiceRegistry and then
> provide one in the unit test along with a fake BatteryMonitor implementation.

Yes on second thought this is even somewhat covered by the browser-test we have
for battery, so technically we could probably remove this unittest altogether.

However it may still be interesting to have the unittest by means of an example.
I'll try to investigate the dependency injection suggestion as it would simplify
the BatteryStatusDispatcher code (especially wrt your new patch). I am assuming
by dependency injection you meant passing the ServiceRegistry in the constructor
of BatteryStatusDispatcher?