Revert of Revert of Revert of Disable a test that fails under ASan and MSan

Revert of Revert of Revert of Disable a test that fails under ASan and MSan (patchset #1 id:1 of https://codereview.chromium.org/1011083002/)

Reason for revert:
Seems like the underlying problem should now be fixed. Reverting the reverted revert, so the end state should be that the test is no longer disabled. Will look out for problems on the bots.

Original issue's description:
> Revert of Revert of Disable a test that fails under ASan and MSan (patchset #1 id:1 of https://codereview.chromium.org/1018443002/)
>
> Reason for revert:
> The underlying issue is still not fixed, re-enabling the test causes bots to turn red.
>
> Please wait for the underlying issue to be resolved before re-enabling.
>
> Original issue's description:
> > Revert of Disable a test that fails under ASan and MSan (patchset #1 id:1 of https://codereview.chromium.org/1007293002/)
> >
> > Reason for revert:
> > This test is a victim of a bigger issue of main renderer message loop being used after it is deleted. Details are in crbug.com/467369.
> >
> > I've also run the test with ASAN locally and didn't reproduce after a few runs.
> >
> > Original issue's description:
> > > Disable a test that fails under ASan and MSan
> > >
> > > It can be flaky in other configurations as well
> > >
> > > BUG=467483
> > > TBR=nasko
> > >
> > > Committed: https://crrev.com/992ff715f61e90562983139ac31392b72e716822
> > > Cr-Commit-Position: refs/heads/master@{#320713}
> >
> > TBR=timurrrr@chromium.org
> > NOPRESUBMIT=true
> > NOTREECHECKS=true
> > NOTRY=true
> > BUG=467483
> >
> > Committed: https://crrev.com/216981c97018a8b8dea097cbde58889bf4e68c1a
> > Cr-Commit-Position: refs/heads/master@{#320755}
>
> TBR=nasko@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=467483
>
> Committed: https://crrev.com/186c9ac9e0bc352e0971923584ca3ae26491ece0
> Cr-Commit-Position: refs/heads/master@{#320890}

TBR=nasko@chromium.org,timurrrr@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=467483
Committed: https://crrev.com/58167ebbb84499812f67ca93ad587048d3004214
Cr-Commit-Position: refs/heads/master@{#321053}

[benwells, 2015-03-18 02:06:48]

Created Revert of Revert of Revert of Disable a test that fails under ASan and
MSan

[commit-bot: I haz the power, 2015-03-18 02:07:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1014183002/1

[commit-bot: I haz the power, 2015-03-18 02:07:57]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-03-18 02:08:50]

Patchset 1 (id:??) landed as
https://crrev.com/58167ebbb84499812f67ca93ad587048d3004214
Cr-Commit-Position: refs/heads/master@{#321053}

[benwells, 2015-03-18 06:23:33]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/1016783004/ by benwells@chromium.org.

The reason for reverting is: The problem still exists. Reverting this revert^3.
.

[nasko, 2015-03-18 16:05:03]

On 2015/03/18 06:23:33, benwells wrote:
> A revert of this CL (patchset #1 id:1) has been created in
> https://codereview.chromium.org/1016783004/ by mailto:benwells@chromium.org.
> 
> The reason for reverting is: The problem still exists. Reverting this
revert^3.
> .

Is it possible to disable the test only for the MSan bot? I'm not sure if this
is flaky test in general or flaky on MSan, so I'd like to enable it on the other
platforms and see if I get any flakiness.

[benwells, 2015-03-19 07:07:54]

On 2015/03/18 16:05:03, nasko wrote:
> On 2015/03/18 06:23:33, benwells wrote:
> > A revert of this CL (patchset #1 id:1) has been created in
> > https://codereview.chromium.org/1016783004/ by mailto:benwells@chromium.org.
> > 
> > The reason for reverting is: The problem still exists. Reverting this
> revert^3.
> > .
> 
> Is it possible to disable the test only for the MSan bot? I'm not sure if this
> is flaky test in general or flaky on MSan, so I'd like to enable it on the
other
> platforms and see if I get any flakiness.

Yes, it is indeed. To do this use #if defined(MEMORY_SANITIZER).

It will mean we can't keep this amazing revert chain going, but maybe that's a
good thing.

[Timur Iskhodzhanov, 2015-03-19 07:37:50]

This test also fails under ASan, so ADDRESS_SANITIZER should also be used.

чт, 19 марта 2015, 10:07, null <benwells@chromium.org>:

> On 2015/03/18 16:05:03, nasko wrote:
> > On 2015/03/18 06:23:33, benwells wrote:
> > > A revert of this CL (patchset #1 id:1) has been created in
> > > https://codereview.chromium.org/1016783004/ by
> > mailto:benwells@chromium.org.
> > >
> > > The reason for reverting is: The problem still exists. Reverting this
> > revert^3.
> > > .
>
> > Is it possible to disable the test only for the MSan bot? I'm not sure if
> > this
> > is flaky test in general or flaky on MSan, so I'd like to enable it on
> the
> other
> > platforms and see if I get any flakiness.
>
> Yes, it is indeed. To do this use #if defined(MEMORY_SANITIZER).
>
> It will mean we can't keep this amazing revert chain going, but maybe
> that's a
> good thing.
>
> https://codereview.chromium.org/1014183002/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[nasko, 2015-03-19 14:07:48]

On 2015/03/19 07:37:50, Timur Iskhodzhanov wrote:
> This test also fails under ASan, so ADDRESS_SANITIZER should also be used.
> 
> чт, 19 марта 2015, 10:07, null <mailto:benwells@chromium.org>:
> 
> > On 2015/03/18 16:05:03, nasko wrote:
> > > On 2015/03/18 06:23:33, benwells wrote:
> > > > A revert of this CL (patchset #1 id:1) has been created in
> > > > https://codereview.chromium.org/1016783004/ by
> > > mailto:benwells@chromium.org.
> > > >
> > > > The reason for reverting is: The problem still exists. Reverting this
> > > revert^3.
> > > > .
> >
> > > Is it possible to disable the test only for the MSan bot? I'm not sure if
> > > this
> > > is flaky test in general or flaky on MSan, so I'd like to enable it on
> > the
> > other
> > > platforms and see if I get any flakiness.
> >
> > Yes, it is indeed. To do this use #if defined(MEMORY_SANITIZER).
> >
> > It will mean we can't keep this amazing revert chain going, but maybe
> > that's a
> > good thing.
> >
> > https://codereview.chromium.org/1014183002/

Thanks for the info! Unfortunately, the problem is a bit bigger and the *San
bots slow down execution enough to expose it. I've diagnosed it and will be
rewriting the test.

[Timur Iskhodzhanov, 2015-03-19 15:26:15]

FYI from my experience if a test misbehaves due to *San slowdown, it's
likely to be flaky on normal VM bots as well.

On Thu, Mar 19, 2015 at 5:07 PM <nasko@chromium.org> wrote:

> On 2015/03/19 07:37:50, Timur Iskhodzhanov wrote:
> > This test also fails under ASan, so ADDRESS_SANITIZER should also be
> used.
>
> > чт, 19 марта 2015, 10:07, null <mailto:benwells@chromium.org>:
>
> > > On 2015/03/18 16:05:03, nasko wrote:
> > > > On 2015/03/18 06:23:33, benwells wrote:
> > > > > A revert of this CL (patchset #1 id:1) has been created in
> > > > > https://codereview.chromium.org/1016783004/ by
> > > > mailto:benwells@chromium.org.
> > > > >
> > > > > The reason for reverting is: The problem still exists. Reverting
> > this
> > > > revert^3.
> > > > > .
> > >
> > > > Is it possible to disable the test only for the MSan bot? I'm not
> > sure if
> > > > this
> > > > is flaky test in general or flaky on MSan, so I'd like to enable it
> on
> > > the
> > > other
> > > > platforms and see if I get any flakiness.
> > >
> > > Yes, it is indeed. To do this use #if defined(MEMORY_SANITIZER).
> > >
> > > It will mean we can't keep this amazing revert chain going, but maybe
> > > that's a
> > > good thing.
> > >
> > > https://codereview.chromium.org/1014183002/
>
> Thanks for the info! Unfortunately, the problem is a bit bigger and the
> *San
> bots slow down execution enough to expose it. I've diagnosed it and will be
> rewriting the test.
>
>
> https://codereview.chromium.org/1014183002/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[nasko, 2015-03-19 16:03:38]

On 2015/03/19 15:26:15, Timur Iskhodzhanov wrote:
> FYI from my experience if a test misbehaves due to *San slowdown, it's
> likely to be flaky on normal VM bots as well.

Yes, it is mostly the case, but for this one there has been only one failure
I've seen on other bots, so the slow down makes a huge difference. It is due to
the nature of the test, but it is getting rewritten anyway :), so thanks for
your help!