Allow IO for the computation of PrefHashCalculator::legacy_device_id_instance_.

Allow IO for the computation of PrefHashCalculator::legacy_device_id_instance_.

Explicitly AllowIO in the scope of this computation. This isn't ideal, but there is no other good way until we can get rid of this legacy artifact in a few milestones.

Note that I have to skip a presubmit error for (2): "Banned functions were used. New code should not use ScopedAllowIO." :(.

BUG=361687
R=erikwright@chromium.org

Originally committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=262995
Reverted: https://src.chromium.org/viewvc/chrome?view=rev&revision=262995

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=263831

[gab, 2014-04-09 20:21:50]

Erik, PTAL.

We may want to merge this to M35 for the performance benefits of calculating the
legacy device ID once per process rather than once per PrefHashCalculator.

Thanks!
Gab

[erikwright (departed), 2014-04-10 00:26:10]

LGTM.

[erikwright (departed), 2014-04-10 00:28:43]

I don't think this merits a merge request.

[gab, 2014-04-10 13:14:03]

On 2014/04/10 00:28:43, erikwright wrote:
> I don't think this merits a merge request.

I think a merge is interesting from a performance point of view; especially with
SegregatedPrefStore where there will be 2 PrefHashCalculators per profile.

It doesn't hurt to ask IMO, I wouldn't push it though, WDYT?

Cheers,
Gab

[gab, 2014-04-10 13:14:08]

The CQ bit was checked by gab@chromium.org

[commit-bot: I haz the power, 2014-04-10 13:14:15]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/gab@chromium.org/231673002/40001

[commit-bot: I haz the power, 2014-04-10 14:04:49]

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

[commit-bot: I haz the power, 2014-04-10 14:04:50]

Retried try job too often on chromium_presubmit for step(s) presubmit
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=chromium_p...

[gab, 2014-04-10 15:37:20]

Committed patchset #3 manually as r262995 (tree was closed).

[gab, 2014-04-10 15:41:44]

On 2014/04/10 15:37:20, gab wrote:
> Committed patchset #3 manually as r262995 (tree was closed).

(sorry about the commit while tree was closed; tree was green when I began the
dcommit (but it took several minutes because of weird "-empty_dir: trunk/..."
processings) and it succeeded at the end despite the tree being closed again
because I had to dcommit with --bypass-hooks for reasons explained in the CL
description).

[gab, 2014-04-10 18:02:09]

A revert of this CL has been created in
https://codereview.chromium.org/228763007/ by gab@chromium.org.

The reason for reverting is: PrefHashCalculatorTest.TestLegacyDeviceIdAlgorithm
is failing in some configurations..

[gab, 2014-04-10 18:58:06]

Erik, PTAL. I remember now why I hadn't taken the singleton LazyInstance
approach in the first place: it doesn't work with unit tests which run multiple
tests in the same process.

I don't think it's a big deal to leave it as-is, especially if we're not going
to merge it to M35 as this code will run even less on M36 and probably be
removed in M37.

Cheers!
Gab

[gab, 2014-04-14 12:15:18]

On 2014/04/10 18:58:06, gab wrote:
> Erik, PTAL. I remember now why I hadn't taken the singleton LazyInstance
> approach in the first place: it doesn't work with unit tests which run
multiple
> tests in the same process.
> 
> I don't think it's a big deal to leave it as-is, especially if we're not going
> to merge it to M35 as this code will run even less on M36 and probably be
> removed in M37.
> 
> Cheers!
> Gab

ping erikwright ^^

[erikwright (departed), 2014-04-15 13:13:08]

LGTM.

[gab, 2014-04-15 13:21:12]

Committed patchset #5 manually as r263831 (tree was closed).