Initialize metrics service after cast service is initialized.

Initialize metrics service after cast service is initialized.

Metrics service uses sys info to determine if user opted out metrics
reporting.

BUG=internal b/19432329

Committed: https://crrev.com/93077dd7090480873991c75dbf51e6e5790a57fd
Cr-Commit-Position: refs/heads/master@{#317191}

[byungchul, 2015-02-19 23:12:38]

byungchul@chromium.org changed reviewers:
+ gfhuang@chromium.org, gunsch@chromium.org, lcwu@chromium.org

[lcwu1, 2015-02-19 23:20:29]

lgtm

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
File chromecast/browser/cast_browser_main_parts.cc (right):

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
chromecast/browser/cast_browser_main_parts.cc:236: // CastNetworkDelegate may
use components initialized by cast service.
nit: remove 'may' because we know for a fact that they do use components
initialized by the cast service. :-)

[gunsch, 2015-02-19 23:21:46]

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
File chromecast/browser/cast_browser_main_parts.cc (right):

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
chromecast/browser/cast_browser_main_parts.cc:236: // CastNetworkDelegate may
use components initialized by cast service.
On 2015/02/19 23:20:29, lcwu1 wrote:
> nit: remove 'may' because we know for a fact that they do use components
> initialized by the cast service. :-)

Doesn't that depend on the particular CastService implementation?

I don't suppose there's another way we can enforce that in
CastMetricsServiceClient...

[byungchul, 2015-02-19 23:29:01]

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
File chromecast/browser/cast_browser_main_parts.cc (right):

https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
chromecast/browser/cast_browser_main_parts.cc:236: // CastNetworkDelegate may
use components initialized by cast service.
On 2015/02/19 23:21:46, gunsch wrote:
> On 2015/02/19 23:20:29, lcwu1 wrote:
> > nit: remove 'may' because we know for a fact that they do use components
> > initialized by the cast service. :-)
> 
> Doesn't that depend on the particular CastService implementation?
> 
> I don't suppose there's another way we can enforce that in
> CastMetricsServiceClient...

May be possible if sys info supports Initialized() and
PlatformIsReportingEnabled does DCHECK(). Anyway, that's out of scope of this
CL.

Le-chun, do you agree not to remove "may" because it depends?

[lcwu1, 2015-02-19 23:44:48]

On 2015/02/19 23:29:01, byungchul wrote:
>
https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
> File chromecast/browser/cast_browser_main_parts.cc (right):
> 
>
https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
> chromecast/browser/cast_browser_main_parts.cc:236: // CastNetworkDelegate may
> use components initialized by cast service.
> On 2015/02/19 23:21:46, gunsch wrote:
> > On 2015/02/19 23:20:29, lcwu1 wrote:
> > > nit: remove 'may' because we know for a fact that they do use components
> > > initialized by the cast service. :-)
> > 
> > Doesn't that depend on the particular CastService implementation?
> > 
> > I don't suppose there's another way we can enforce that in
> > CastMetricsServiceClient...
> 
> May be possible if sys info supports Initialized() and
> PlatformIsReportingEnabled does DCHECK(). Anyway, that's out of scope of this
> CL.
> 
> Le-chun, do you agree not to remove "may" because it depends?

Sure.

[gunsch, 2015-02-19 23:45:55]

On 2015/02/19 23:44:48, lcwu1 wrote:
> On 2015/02/19 23:29:01, byungchul wrote:
> >
>
https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
> > File chromecast/browser/cast_browser_main_parts.cc (right):
> > 
> >
>
https://codereview.chromium.org/940163005/diff/1/chromecast/browser/cast_brow...
> > chromecast/browser/cast_browser_main_parts.cc:236: // CastNetworkDelegate
may
> > use components initialized by cast service.
> > On 2015/02/19 23:21:46, gunsch wrote:
> > > On 2015/02/19 23:20:29, lcwu1 wrote:
> > > > nit: remove 'may' because we know for a fact that they do use components
> > > > initialized by the cast service. :-)
> > > 
> > > Doesn't that depend on the particular CastService implementation?
> > > 
> > > I don't suppose there's another way we can enforce that in
> > > CastMetricsServiceClient...
> > 
> > May be possible if sys info supports Initialized() and
> > PlatformIsReportingEnabled does DCHECK(). Anyway, that's out of scope of
this
> > CL.
> > 
> > Le-chun, do you agree not to remove "may" because it depends?
> 
> Sure.

lgtm, and looking forward to an internal follow-up CL that can, at the very
least, DCHECK to keep this ordering correct :)

[byungchul, 2015-02-20 00:09:46]

The CQ bit was checked by byungchul@chromium.org

[commit-bot: I haz the power, 2015-02-20 00:10:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/940163005/1

[commit-bot: I haz the power, 2015-02-20 00:33:42]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-02-20 00:34:50]

Patchset 1 (id:??) landed as
https://crrev.com/93077dd7090480873991c75dbf51e6e5790a57fd
Cr-Commit-Position: refs/heads/master@{#317191}