I don't think this is just flake. The test has failed twice in the last three runs after not failing for 200 runs.

These two net related changes landed on the first failed build:

"[Android] Delete org.chromium.net.test.BaseTestServer and friends." https://chromium.googlesource.com/chromium/src/+/9d7d9653e3f6c4ecbc970a9c1b558120cb9cc6da

net: remove ExplicitUserRequest() method from URLRequestThrottlerEntry https://chromium.googlesource.com/chromium/src/+/97161cd9d32b7b1e7707fa6a0442ab0961d5c0bd


+tfarina@ & +jbudorick@ - could either of those have made this test flaky?

I don't think so -- nothing was using the code I deleted.

Agree with jbudorick@, this failure is definitely a flake related to a change xunjieli@ landed earlier today. I expect it to stop happening now that https://chromium.googlesource.com/chromium/src/+/a4461011bd1331b61f27ad8fc2e65e53cad92436 landed.

These two CLs also landed in the first failed run of Android Cronet Marshmallow 64 bit:
https://build.chromium.org/p/chromium.android/builders/Android%20Cronet%20Marshmallow%2064bit%20Perf/builds/1992

Along with These two CLs also landed in the first failed run of Android Cronet Marshmallow 64 bit:
"[Cronet] Pass metrics information from C++ BidirectionalStream to Java"
https://build.chromium.org/p/chromium.android/builders/Android%20Cronet%20Marshmallow%2064bit%20Perf/builds/1992

which seems most likely.

xunjieli@ - will you have a fix soon or should we revert that CL?

Thanks mgersh@ - I didn't realize a fix was already in flight.

-tfarina@ and jbudorick@

Thanks for looking into this. Cl is landed (
https://chromium.googlesource.com/chromium/src/+/a4461011bd1331b61f27ad8fc2e65e53cad92436)
The bots should cycle green soon.

Actually...this looks like a different failure from the one that that change fixed. It's this line that's failing: https://chromium.googlesource.com/chromium/src/+/master/components/cronet/android/test/javatests/src/org/chromium/net/MetricsTestUtil.java#147

I'm not sure why it just started showing up now. Helen, any ideas?

CL which will actually make it stop failing is in the CQ now.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/015d9b586eab4ba642c3e882d345312765a1dbf2

commit 015d9b586eab4ba642c3e882d345312765a1dbf2
Author: mgersh <mgersh@chromium.org>
Date: Thu Oct 06 21:39:06 2016

Disable BidirectionalStreamTest#testFailures

It's flaky, with more failures than not, because of new checks added in
a CL from earlier today. Disabling for now, to be fixed for real
tomorrow.

BUG= 653654 

Review-Url: https://codereview.chromium.org/2398273002
Cr-Commit-Position: refs/heads/master@{#423694}

[modify] https://crrev.com/015d9b586eab4ba642c3e882d345312765a1dbf2/components/cronet/android/test/javatests/src/org/chromium/net/BidirectionalStreamTest.java

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9be380db89ba5afe42e956b6cf240bd6036e51e6

commit 9be380db89ba5afe42e956b6cf240bd6036e51e6
Author: xunjieli <xunjieli@chromium.org>
Date: Mon Oct 10 18:25:18 2016

Fix BidirectionalStreamTest#testFailures flake

When we cancel the request in BidirectionalStream#onStreamReady(), it's not
guaranteed that the request is stopped right away, so we might have non-null
connect timing.
This CL makes the test assertion looser.

BUG= 653654 

Review-Url: https://codereview.chromium.org/2404923002
Cr-Commit-Position: refs/heads/master@{#424188}

[modify] https://crrev.com/9be380db89ba5afe42e956b6cf240bd6036e51e6/components/cronet/android/test/javatests/src/org/chromium/net/BidirectionalStreamTest.java

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a4461011bd1331b61f27ad8fc2e65e53cad92436

commit a4461011bd1331b61f27ad8fc2e65e53cad92436
Author: xunjieli <xunjieli@chromium.org>
Date: Thu Oct 06 20:20:43 2016

Fix BidirectionalStreamTest#testFailures

Response start can be null or non-null in
BidirectionalStream.Callback#onStreamReady() step.

TBR=mgersh@chromium.org

BUG= 653654 

Review-Url: https://codereview.chromium.org/2400803002
Cr-Commit-Position: refs/heads/master@{#423649}

[modify] https://crrev.com/a4461011bd1331b61f27ad8fc2e65e53cad92436/components/cronet/android/test/javatests/src/org/chromium/net/BidirectionalStreamTest.java

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/015d9b586eab4ba642c3e882d345312765a1dbf2

commit 015d9b586eab4ba642c3e882d345312765a1dbf2
Author: mgersh <mgersh@chromium.org>
Date: Thu Oct 06 21:39:06 2016

Disable BidirectionalStreamTest#testFailures

It's flaky, with more failures than not, because of new checks added in
a CL from earlier today. Disabling for now, to be fixed for real
tomorrow.

BUG= 653654 

Review-Url: https://codereview.chromium.org/2398273002
Cr-Commit-Position: refs/heads/master@{#423694}

[modify] https://crrev.com/015d9b586eab4ba642c3e882d345312765a1dbf2/components/cronet/android/test/javatests/src/org/chromium/net/BidirectionalStreamTest.java

[Automated comment] removing mislabelled merge-merged-2840