Revert of Attempt #3 at registry outage: ShellExecute reg delete

Revert of Attempt #3 at registry outage: ShellExecute reg delete (patchset #1 id:1 of https://codereview.chromium.org/2897483003/ )

Reason for revert:
Nuclear cleanup no longer needed (and having reg.exe run in parallel with the browser_tests that are starting can cause flakiness in itself).

Original issue's description:
> Attempt #3 at registry outage: ShellExecute reg delete
> 
> Nuclear option to be reverted after the bots heal up.
> 
> BUG=721245
> R=thakis@chromium.org
> 
> Review-Url: https://codereview.chromium.org/2897483003 .
> Cr-Commit-Position: refs/heads/master@{#472945}
> Committed: https://chromium.googlesource.com/chromium/src/+/f26f6233d2d8be69c638ddac047db13fd6f14062

TBR=thakis@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=721245
Review-Url: https://codereview.chromium.org/2902683003
Cr-Commit-Position: refs/heads/master@{#473898}
Committed: https://chromium.googlesource.com/chromium/src/+/b6d2543cee9b9dff8030ff64e7104fa486cc590d

[gab, 2017-05-23 14:20:54]

The CQ bit was checked by gab@chromium.org

[gab, 2017-05-23 14:20:54]

Created Revert of Attempt #3 at registry outage: ShellExecute reg delete

[commit-bot: I haz the power, 2017-05-23 14:21:09]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-23 15:08:43]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1495549254154960, "parent_rev":
"d86012b80944319678d07adcc26b07b873deb5fa", "commit_rev":
"b6d2543cee9b9dff8030ff64e7104fa486cc590d"}

[commit-bot: I haz the power, 2017-05-23 15:08:53]

Description was changed from

==========
Revert of Attempt #3 at registry outage: ShellExecute reg delete (patchset #1
id:1 of https://codereview.chromium.org/2897483003/ )

Reason for revert:
Nuclear cleanup no longer needed (and having reg.exe run in parallel with the
browser_tests that are starting can cause flakiness in itself).

Original issue's description:
> Attempt #3 at registry outage: ShellExecute reg delete
> 
> Nuclear option to be reverted after the bots heal up.
> 
> BUG=721245
> R=thakis@chromium.org
> 
> Review-Url: https://codereview.chromium.org/2897483003 .
> Cr-Commit-Position: refs/heads/master@{#472945}
> Committed:
https://chromium.googlesource.com/chromium/src/+/f26f6233d2d8be69c638ddac047d...

TBR=thakis@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=721245
==========

to

==========
Revert of Attempt #3 at registry outage: ShellExecute reg delete (patchset #1
id:1 of https://codereview.chromium.org/2897483003/ )

Reason for revert:
Nuclear cleanup no longer needed (and having reg.exe run in parallel with the
browser_tests that are starting can cause flakiness in itself).

Original issue's description:
> Attempt #3 at registry outage: ShellExecute reg delete
> 
> Nuclear option to be reverted after the bots heal up.
> 
> BUG=721245
> R=thakis@chromium.org
> 
> Review-Url: https://codereview.chromium.org/2897483003 .
> Cr-Commit-Position: refs/heads/master@{#472945}
> Committed:
https://chromium.googlesource.com/chromium/src/+/f26f6233d2d8be69c638ddac047d...

TBR=thakis@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=721245

Review-Url: https://codereview.chromium.org/2902683003
Cr-Commit-Position: refs/heads/master@{#473898}
Committed:
https://chromium.googlesource.com/chromium/src/+/b6d2543cee9b9dff8030ff64e710...
==========

[commit-bot: I haz the power, 2017-05-23 15:08:54]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/chromium/src/+/b6d2543cee9b9dff8030ff64e710...

[jam, 2017-05-23 15:15:34]

can you remove all this code? I don't think we need it anymore.

[gab, 2017-05-23 16:15:23]

On 2017/05/23 15:15:34, jam wrote:
> can you remove all this code? I don't think we need it anymore.

I think it's still required as although the code self cleans in clean exits, it
can still leak in test crashes. Or are you not worried that crashes alone will
make the registry grow to 2GB (it took months with an explicit leak in every
single run).

So long as we build in a registry size watcher on the bots or something to
prevent a similar leakage in the future maybe we're fine?

[jam, 2017-05-23 23:55:38]

On 2017/05/23 16:15:23, gab wrote:
> On 2017/05/23 15:15:34, jam wrote:
> > can you remove all this code? I don't think we need it anymore.
> 
> I think it's still required as although the code self cleans in clean exits,
it
> can still leak in test crashes. Or are you not worried that crashes alone will
> make the registry grow to 2GB (it took months with an explicit leak in every
> single run).
> 
> So long as we build in a registry size watcher on the bots or something to
> prevent a similar leakage in the future maybe we're fine?

good point, I guess we can keep that code.