Switch CertStore to RenderProcessHostObserver

Switch CertStore to RenderProcessHostObserver.

BUG=170921
TEST=no change

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=236861

[Avi (use Gerrit), 2013-11-18 23:01:02]

This is the vision for services within content: they register to watch creation
of RenderProcessHosts, and when the hosts are created, they start observation of
individual ones.

For services outside of content, they get to watch creation of
RenderProcessHosts via the ContentBrowserClient. I haven't worked out the exact
details but I will.

[Avi (use Gerrit), 2013-11-20 19:53:10]

John: take a look; better.

[jam, 2013-11-21 19:06:48]

https://codereview.chromium.org/68543003/diff/180001/content/browser/renderer...
File content/browser/renderer_data_memoizing_store.h (right):

https://codereview.chromium.org/68543003/diff/180001/content/browser/renderer...
content/browser/renderer_data_memoizing_store.h:81: // NB: This is super-sketchy
and racy; can we start observing the
why not just handle this now in this cl, it seems like that needs to be done
anyway to avoid crashes. i.e. in StartObservingProcess
RenderProcessHost::FromID(process_id) needs to be null checked to avoid races,
and if it's null you can call RenderProcessHostDestroyed manually?

[Avi (use Gerrit), 2013-11-21 19:25:54]

https://codereview.chromium.org/68543003/diff/180001/content/browser/renderer...
File content/browser/renderer_data_memoizing_store.h (right):

https://codereview.chromium.org/68543003/diff/180001/content/browser/renderer...
content/browser/renderer_data_memoizing_store.h:81: // NB: This is super-sketchy
and racy; can we start observing the
On 2013/11/21 19:06:48, jam wrote:
> why not just handle this now in this cl, it seems like that needs to be done
> anyway to avoid crashes. i.e. in StartObservingProcess
> RenderProcessHost::FromID(process_id) needs to be null checked to avoid races,
> and if it's null you can call RenderProcessHostDestroyed manually?

I hadn't thought about doing that. This is a horrible design anyway (touching
both the UI and IO threads) and I was too creeped out by that to think of that
idea.

[jam, 2013-11-22 17:02:02]

lgtm

[commit-bot: I haz the power, 2013-11-22 17:03:31]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/avi@chromium.org/68543003/360001

[commit-bot: I haz the power, 2013-11-22 23:14:31]

Change committed as 236861