Reset some UMA stability stats on a version change.

Reset some UMA stability stats on a version change.

This is a partial revert of some of the logic that was
removed by http://crrev.com/262526, but updated with
new logic to only do it when there is a version change
and we didn't generate an initial stability log. Plus,
the list of prefs that are reset is trimmed to only
the ones that need to be (per my auditing of the code).

BUG=409117
Committed: https://crrev.com/ff3e2a6552f825070f5fa830ba14faae00558556
Cr-Commit-Position: refs/heads/master@{#295574}

[Alexei Svitkine (slow), 2014-09-18 16:58:48]

Patchset #1 (id:1) has been deleted

[Alexei Svitkine (slow), 2014-09-18 16:59:15]

asvitkine@chromium.org changed reviewers:
+ isherman@chromium.org

[Ilya Sherman, 2014-09-18 18:30:20]

LGTM % nits.  Thanks, Alexei!

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
File components/metrics/metrics_provider.h (right):

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
components/metrics/metrics_provider.h:46: // Called to indicate that saved
stability prefs should be cleared, for e.g.
nit: "for e.g. if" -> "e.g. because" -- at least, that reads more naturally to
me

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
File components/metrics/metrics_service.cc (right):

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
components/metrics/metrics_service.cc:598: // reset stability stats from the
previous version.
nit: I'd go into a bit more detail here on why this might happen -- something
like "There is an edge case where the initial stability log can fail to be
generated even if the previous version crashed: It's possible that the previous
version crashed before it was able to save off a system profile. In this case,
we can't correctly identify what version the saved stability metrics correspond
to; but since we detected that the version changed, we at least know that it's
not the current version."

... but, you know, shorter would be better.  Please improve that further; I just
think the general idea is important to get across.

[Alexei Svitkine (slow), 2014-09-18 20:20:58]

Patchset #2 (id:40001) has been deleted

[Alexei Svitkine (slow), 2014-09-18 20:21:14]

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
File components/metrics/metrics_provider.h (right):

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
components/metrics/metrics_provider.h:46: // Called to indicate that saved
stability prefs should be cleared, for e.g.
On 2014/09/18 18:30:19, Ilya Sherman wrote:
> nit: "for e.g. if" -> "e.g. because" -- at least, that reads more naturally to
> me

Done.

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
File components/metrics/metrics_service.cc (right):

https://codereview.chromium.org/580093004/diff/20001/components/metrics/metri...
components/metrics/metrics_service.cc:598: // reset stability stats from the
previous version.
On 2014/09/18 18:30:20, Ilya Sherman wrote:
> nit: I'd go into a bit more detail here on why this might happen -- something
> like "There is an edge case where the initial stability log can fail to be
> generated even if the previous version crashed: It's possible that the
previous
> version crashed before it was able to save off a system profile. In this case,
> we can't correctly identify what version the saved stability metrics
correspond
> to; but since we detected that the version changed, we at least know that it's
> not the current version."
> 
> ... but, you know, shorter would be better.  Please improve that further; I
just
> think the general idea is important to get across.

Done.

[Alexei Svitkine (slow), 2014-09-18 20:21:18]

The CQ bit was checked by asvitkine@chromium.org

[commit-bot: I haz the power, 2014-09-18 20:22:40]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/580093004/60001

[commit-bot: I haz the power, 2014-09-18 22:01:57]

Committed patchset #2 (id:60001) as e6ce3740ce7eadf8bd1e5e6cf7cb7a56fdce0faf

[commit-bot: I haz the power, 2014-09-18 22:02:26]

Patchset 2 (id:??) landed as
https://crrev.com/ff3e2a6552f825070f5fa830ba14faae00558556
Cr-Commit-Position: refs/heads/master@{#295574}