ServiceWorker: Fix registration refcount handling in the renderer

ServiceWorker: Fix registration refcount handling in the renderer

- ServiceWorkerRegistrationHandleReference::Create() should increment registration refcount
- ServiceWorkerMessageFilter::OnStaleSetVersionAttributes shouldn’t decrement registration refcount

TEST=n/a
BUG=452403
Committed: https://crrev.com/e849bd834ff29b8e64380cbbbb8e12dc19ec5f36
Cr-Commit-Position: refs/heads/master@{#313497}

[nhiroki, 2015-01-27 12:16:59]

nhiroki@chromium.org changed reviewers:
+ kinuko@chromium.org

[nhiroki, 2015-01-27 12:16:59]

PTAL, thanks!

[kinuko, 2015-01-28 06:01:24]

Meh, was this an existing bug?

lgtm, but would have been nice if we have a test..

[nhiroki, 2015-01-28 07:06:34]

Thank you for reviewing.

On 2015/01/28 06:01:24, kinuko wrote:
> Meh, was this an existing bug?

Regarding |increment_ref_in_ctor|, that was introduced by the initial CL[1]...

[1]
https://chromium.googlesource.com/chromium/src/+/61752d1c1703c30a55c14675e6e2...

> lgtm, but would have been nice if we have a test..

Strongly agree, but for now I'm not really sure how to write tests between
renderer and browser via ServiceWorkerDispatcherHost. This is blocking other
high-priority CLs, so those tests could be in a separate CLs... anyway, I'll try
it.

[nhiroki, 2015-01-28 13:24:33]

On 2015/01/28 07:06:34, nhiroki (slow) wrote:
> Thank you for reviewing.
> 
> On 2015/01/28 06:01:24, kinuko wrote:
> > Meh, was this an existing bug?
> 
> Regarding |increment_ref_in_ctor|, that was introduced by the initial CL[1]...
> 
> [1]
>
https://chromium.googlesource.com/chromium/src/+/61752d1c1703c30a55c14675e6e2...
> 
> > lgtm, but would have been nice if we have a test..
> 
> Strongly agree, but for now I'm not really sure how to write tests between
> renderer and browser via ServiceWorkerDispatcherHost. This is blocking other
> high-priority CLs, so those tests could be in a separate CLs... anyway, I'll
try
> it.

Subsequent CLs get ready to land, so let me make tests in a separate CL...

[nhiroki, 2015-01-28 13:24:44]

The CQ bit was checked by nhiroki@chromium.org

[commit-bot: I haz the power, 2015-01-28 13:26:02]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/868483008/1

[commit-bot: I haz the power, 2015-01-28 13:29:07]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-01-28 13:30:02]

Patchset 1 (id:??) landed as
https://crrev.com/e849bd834ff29b8e64380cbbbb8e12dc19ec5f36
Cr-Commit-Position: refs/heads/master@{#313497}