Fix crash in SiteEngagementServiceHelper when switching render view hosts.

Fix crash in SiteEngagementServiceHelper when switching render view hosts.

This CL fixes a crash where the RVH would switch on a WebContents that
had input tracking paused. The host was not being switched out for a
paused InputTracker, causing the tracking to be added to the stale RVH.

BUG=544211
Committed: https://crrev.com/b1893018fff3234b87c000c16f2b76926b8cbb6a
Cr-Commit-Position: refs/heads/master@{#354930}

[calamity, 2015-10-19 03:14:48]

Patchset #1 (id:1) has been deleted

[calamity, 2015-10-19 03:41:17]

calamity@chromium.org changed reviewers:
+ benwells@chromium.org

[benwells, 2015-10-19 03:56:39]

Can we remove is_active_ and have IsActive() { return host_; } instead?

That might let us remove some assumptions that are currently enforced via
DCHECK.

[calamity, 2015-10-19 06:11:00]

On 2015/10/19 03:56:39, benwells wrote:
> Can we remove is_active_ and have IsActive() { return host_; } instead?
> 
> That might let us remove some assumptions that are currently enforced via
> DCHECK.

Done.

[benwells, 2015-10-19 06:14:06]

lgtm

[calamity, 2015-10-20 00:13:06]

The CQ bit was checked by calamity@chromium.org

[calamity, 2015-10-20 00:13:06]

The patchset sent to the CQ was uploaded after l-g-t-m from
benwells@chromium.org
Link to the patchset: https://codereview.chromium.org/1411123002/#ps100001
(title: "fix compile")

[commit-bot: I haz the power, 2015-10-20 00:13:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1411123002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1411123002/100001

[commit-bot: I haz the power, 2015-10-20 00:19:38]

Committed patchset #5 (id:100001)

[commit-bot: I haz the power, 2015-10-20 00:20:36]

Patchset 5 (id:??) landed as
https://crrev.com/b1893018fff3234b87c000c16f2b76926b8cbb6a
Cr-Commit-Position: refs/heads/master@{#354930}