Change the ServiceWorkerCacheStorage's loader from a ref ptr to scoped.

Change the ServiceWorkerCacheStorage's loader from a ref ptr to scoped and guarantee that callbacks passed to the loader are called by making functions static and passing WeakPtrs.

Part of the following CL chain to guarantee callbacks are called throughout ServiceWorkerCacheStorage:

* 1. Guarantee CacheLoader callbacks: https://codereview.chromium.org/504233003/
2. Guarantee LazyInit callbacks: https://codereview.chromium.org/512743002/
3. Guarantee Create/Delete callbacks: https://codereview.chromium.org/511893002/
4. Reorganize the methods: https://codereview.chromium.org/510813002/

BUG=392621
Committed: https://crrev.com/54d803feb271128003f73c3d5d30e4a2cc93471f
Cr-Commit-Position: refs/heads/master@{#292180}

[jkarlin, 2014-08-26 18:42:08]

jkarlin@chromium.org changed reviewers:
+ falken@chromium.org

[jkarlin, 2014-08-26 18:42:08]

falken: Please review all, thanks!

[jkarlin, 2014-08-27 11:08:17]

jkarlin@chromium.org changed reviewers:
+ horo@chromium.org, michaeln@chromium.org
- falken@chromium.org

[jkarlin, 2014-08-27 11:08:37]

horo is OOO

michaeln and horo: Please review everything.

[horo, 2014-08-27 16:01:45]

falken@ is OOO but I'm not OOO :)

lgtm

[jkarlin, 2014-08-27 16:10:52]

On 2014/08/27 16:01:45, horo wrote:
> falken@ is OOO but I'm not OOO :)
> 
> lgtm

er, yes, thanks :)

[jkarlin, 2014-08-27 16:10:58]

The CQ bit was checked by jkarlin@chromium.org

[commit-bot: I haz the power, 2014-08-27 16:11:46]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/jkarlin@chromium.org/504233003/60001

[commit-bot: I haz the power, 2014-08-27 17:16:08]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  mac_chromium_rel_swarming on tryserver.chromium.mac
(http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)
  win_chromium_rel_swarming on tryserver.chromium.win
(http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[commit-bot: I haz the power, 2014-08-27 18:03:46]

Committed patchset #4 (id:60001) as 6b3fb16ce1106c885062b3c67fac8628c43c0733

[michaeln, 2014-08-27 22:09:43]

> Part of the following CL chain to guarantee callbacks are called throughout
ServiceWorkerCacheStorage:

I'm not sure this is a good goal? The caller has deleted the object, why is it
good to invoke the callers pending completion callback. It opens the door to
very unusual code paths when destructors call out. Additional code has to be
written to handle being invoked at these odd times (else bugs).

[michaeln, 2014-08-27 22:19:37]

What is the motivation for this set of changes? Is there a particular problem
they're trying to solve?

The current composition of these classes seems reasoanble and i'm not so sure
these changes are an improvement. Maybe there's another way to solve whatever
problem these changes are meant to solve.

[jkarlin, 2014-08-28 11:02:31]

On 2014/08/27 22:19:37, michaeln wrote:
> What is the motivation for this set of changes? Is there a particular problem
> they're trying to solve?
> 
> The current composition of these classes seems reasoanble and i'm not so sure
> these changes are an improvement. Maybe there's another way to solve whatever
> problem these changes are meant to solve.

It seemed cleaner and easier to deal with from a callers perspective to know
that the callbacks would be called.  However, it's not required and I can see
your concern about added complexity (in destructors especially).  I'll set these
aside for now.  Removing reviewers.

[jkarlin, 2014-08-28 11:02:39]

jkarlin@chromium.org changed reviewers:
- horo@chromium.org, michaeln@chromium.org

[jkarlin, 2014-08-28 11:05:47]

jkarlin@chromium.org changed reviewers:
+ horo@chromium.org, michaeln@chromium.org

[jkarlin, 2014-08-28 11:07:09]

On 2014/08/28 11:05:47, jkarlin wrote:
> mailto:jkarlin@chromium.org changed reviewers:
> + mailto:horo@chromium.org, mailto:michaeln@chromium.org

Kept the reviewers on this CL but removed them from the followups. I think that
this particular CL should not be reverted as I prefer the scoped_ptr CacheLoader
to reference counting.

[commit-bot: I haz the power, 2014-09-10 02:52:38]

Patchset 4 (id:??) landed as
https://crrev.com/54d803feb271128003f73c3d5d30e4a2cc93471f
Cr-Commit-Position: refs/heads/master@{#292180}