sync: Expose DirectoryDebugInfoEmitters in engine

sync: Expose DirectoryDebugInfoEmitters in engine

Changes the lifetime of DirectoryTypeDebugInfoEmitters,
DirectoryCommitContributors, and DirectoryUpdateHandlers.  Under the old
lifetime management design, we had no state that lived beyond
a configure cycle.  This would be bad, since we want the debug info
counter values to remain after the initial configure is done.  The
refactoring ensures that the DirectoryTypeDebugInfoEmitters will live as
long as their data type is enabled.  As a side effect, the commit
contributors and update handlers get to live that long, too.

Plumbs the DirectoryTypeDebugInfoEmitters through the commit contributor
and update handler classes.  This prepares them for updating counters
and emitting events from sync/engine/ code, although this functionality
won't be implemented until the next patch in the series.

BUG=328606
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=267740

[rlarocque, 2014-04-30 00:24:22]

Please review.

I'm not too happy about how the reconfiguration logic turned out, but I don't
see any other choice.  The old code was simpler, but it meant that we would wipe
out the counters with every configure cycle.  That was a problem, since it's
useful to not have the counters return to zero as the initial configuration ends
and we enter normal mode.

In theory, I could have refactored the lifetimes of the DebugInfoEmitters only,
but then we would have had two sets of reconfiguration logic living side by
side.  That didn't seem much better.

[rlarocque, 2014-04-30 23:16:03]

Updated reconfiguration logic as discussed.  PTAL.

This version should handle some edge cases better than the first attempt, and
it's a smaller diff, too.

[tim (not reviewing), 2014-04-30 23:24:49]

https://codereview.chromium.org/260613002/diff/40001/sync/sessions/model_type...
File sync/sessions/model_type_registry.cc (right):

https://codereview.chromium.org/260613002/diff/40001/sync/sessions/model_type...
sync/sessions/model_type_registry.cc:34: // The DebugInfoEmitters are not
deleted, since we want to keep their counts
I wouldn't use the term "not deleted" or "never deleted" (below); just note that
their lifetime is tied to the ModelTypeRegistry.

That said, I was imaginging we'd remove the emitter for disabled types.  I don't
think I'd expect the numbers on about:sync to include types that were disabled?

[rlarocque, 2014-04-30 23:44:13]

Patch updated.

https://codereview.chromium.org/260613002/diff/40001/sync/sessions/model_type...
File sync/sessions/model_type_registry.cc (right):

https://codereview.chromium.org/260613002/diff/40001/sync/sessions/model_type...
sync/sessions/model_type_registry.cc:34: // The DebugInfoEmitters are not
deleted, since we want to keep their counts
On 2014/04/30 23:24:50, timsteele wrote:
> I wouldn't use the term "not deleted" or "never deleted" (below); just note
that
> their lifetime is tied to the ModelTypeRegistry.
> 
> That said, I was imaginging we'd remove the emitter for disabled types.  I
don't
> think I'd expect the numbers on about:sync to include types that were
disabled? 

I figured that including disabled types would be a good thing.  Otherwise, some
"strictly incrementing" counters might be decremented when we disable a type. 
For example, just because bookmarks sync is disabled now doesn't mean that we
didn't receive X updates for that type back when it was enabled.

Regardless, we can make the decision elsewhere.  The data reported back to the
UI thread and the WebUI will include type information and enabled state.  We can
filter out disabled types there if we want to.  Assuming we split this into
aggregated and per-type stats, we might make different decisions about whether
or not to include disabled types in each of those counters.

[tim (not reviewing), 2014-05-01 17:42:16]

lgtm

[rlarocque, 2014-05-01 17:47:43]

The CQ bit was checked by rlarocque@chromium.org

[commit-bot: I haz the power, 2014-05-01 17:48:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rlarocque@chromium.org/260613002/60001

[commit-bot: I haz the power, 2014-05-01 18:48:09]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-05-01 18:48:09]

Try jobs failed on following builders:
  win_chromium_rel on tryserver.chromium

[rlarocque, 2014-05-01 23:55:01]

The CQ bit was checked by rlarocque@chromium.org

[commit-bot: I haz the power, 2014-05-01 23:56:19]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/rlarocque@chromium.org/260613002/60001

[commit-bot: I haz the power, 2014-05-02 04:36:13]

Change committed as 267740