Initialize Cast Media for cast_media_unittests

Initialize Cast Media for cast_media_unittests

This allows removal of the lazy initialisation workaround recently
added to vendor media implementation.

BUG=

Committed: https://crrev.com/82c38ab5d0988f1bf46cecf7ab0a4affa670cd2e
Cr-Commit-Position: refs/heads/master@{#330136}

[halliwell, 2015-05-15 04:41:26]

halliwell@chromium.org changed reviewers:
+ erickung@chromium.org, gunsch@chromium.org

[gunsch-google, 2015-05-15 06:19:39]

On 2015/05/15 04:41:33, halliwell wrote:

Don't mention specific vendors in commit messages.

[halliwell, 2015-05-15 14:04:00]

On 2015/05/15 06:19:39, gunsch-google wrote:
> On 2015/05/15 04:41:33, halliwell wrote:
> 
> Don't mention specific vendors in commit messages.

Commit message updated.

[gunsch, 2015-05-15 16:29:11]

lgtm

[gunsch, 2015-05-15 16:29:51]

Actually, on second thought---why not put this in
chromecast/media/cma/run_all_unittests.cc instead?

[halliwell, 2015-05-15 16:49:58]

On 2015/05/15 16:29:51, gunsch wrote:
> Actually, on second thought---why not put this in
> chromecast/media/cma/run_all_unittests.cc instead?

Are you suggesting doing that in order to have the VideoPlane live for the
entire lifetime of the test process?  I did think about that, my pro/con list
was:
* Benefit is that vendor implementations don't have to support instantiating and
destroying VideoPlane repeatedly.
* Disadvantage is that there is more external state hanging around from previous
tests.  e.g. if one test changes the VideoPlane resolution, and another test
relies on it

Based on those, I thought it better to bring up separately for each individual
test.  Let me know what you think.

[gunsch, 2015-05-15 17:00:11]

discussed offline, but this lgtm % a comment in the public API documenting that
we expect to be able to initialize/finalize multiple times per process

[halliwell, 2015-05-15 17:38:35]

On 2015/05/15 17:00:11, gunsch wrote:
> discussed offline, but this lgtm % a comment in the public API documenting
that
> we expect to be able to initialize/finalize multiple times per process

Commenting updated on CastMediaShlib.  I remembered also that we will want to
initialize/finalize multiple times even in cast_shell, to support the resource
grant/revoke model correctly.

[erickung1, 2015-05-15 17:45:14]

On 2015/05/15 17:38:35, halliwell wrote:
> On 2015/05/15 17:00:11, gunsch wrote:
> > discussed offline, but this lgtm % a comment in the public API documenting
> that
> > we expect to be able to initialize/finalize multiple times per process
> 
> Commenting updated on CastMediaShlib.  I remembered also that we will want to
> initialize/finalize multiple times even in cast_shell, to support the resource
> grant/revoke model correctly.

LGTM

[halliwell, 2015-05-15 18:03:40]

The CQ bit was checked by halliwell@chromium.org

[halliwell, 2015-05-15 18:03:40]

The patchset sent to the CQ was uploaded after l-g-t-m from gunsch@chromium.org
Link to the patchset: https://codereview.chromium.org/1137183003/#ps20001
(title: "Comment update on CastMediaShlib")

[commit-bot: I haz the power, 2015-05-15 18:06:51]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1137183003/20001

[commit-bot: I haz the power, 2015-05-15 18:20:35]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-05-15 18:21:22]

Patchset 2 (id:??) landed as
https://crrev.com/82c38ab5d0988f1bf46cecf7ab0a4affa670cd2e
Cr-Commit-Position: refs/heads/master@{#330136}