New issue
Advanced search Search tips

Issue 635508 link

Starred by 1 user

Issue metadata

Status: Untriaged
Owner: ----
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 3
Type: Bug



Sign in to add a comment

ZipFileInstallerTest.GoodZip (and some other unit tests) fails

Project Member Reported by ckulakow...@opera.com, Aug 8 2016

Issue description

Version: 53
OS: Linux

What steps will reproduce the problem?
(1) Build target unit_tests
(2) Start unit_tests --gtest_filter=ZipFileInstallerTest.GoodZip
(3)

What is the expected output?
Test passes

What do you see instead?
Test fails (freezes)

It is caused by https://bugs.chromium.org/p/chromium/issues/detail?id=628492.

 
Project Member

Comment 1 by bugdroid1@chromium.org, Aug 9 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/36f85398960e7836d5325e1a58298209ced34f33

commit 36f85398960e7836d5325e1a58298209ced34f33
Author: rockot <rockot@chromium.org>
Date: Tue Aug 09 02:26:52 2016

[mojo-edk] Revert ObserveProxy retransmission behavior

Gets rid of ObserveProxy retransmission which was only required
to facilitate excessive randomization in tests. In practice this
behavior can race with normal port closure and lead to busy
loops.

Gets rid of excessive randomization in tests. The new PortsTest
environment more closely emulates reality: Nodes operate
independently but there are some basic and trivially satisfiable
ordering guarantees.

Gets rid of deferred AcceptIncomingMessages pumping in
NodeController, which was only added to facilitate
ObserveProxy retransmission.

BUG= 628492 ,635508

Review-Url: https://codereview.chromium.org/2219733005
Cr-Commit-Position: refs/heads/master@{#410535}

[modify] https://crrev.com/36f85398960e7836d5325e1a58298209ced34f33/mojo/edk/system/node_controller.cc
[modify] https://crrev.com/36f85398960e7836d5325e1a58298209ced34f33/mojo/edk/system/node_controller.h
[modify] https://crrev.com/36f85398960e7836d5325e1a58298209ced34f33/mojo/edk/system/ports/node.cc
[modify] https://crrev.com/36f85398960e7836d5325e1a58298209ced34f33/mojo/edk/system/ports/node.h
[modify] https://crrev.com/36f85398960e7836d5325e1a58298209ced34f33/mojo/edk/system/ports/ports_unittest.cc

Yes, that solved the problem. Thanks :).
Project Member

Comment 3 by bugdroid1@chromium.org, Aug 10 2016

Labels: merge-merged-2785
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/573af98298255787657627c25a78a5de1676da73

commit 573af98298255787657627c25a78a5de1676da73
Author: Ken Rockot <rockot@chromium.org>
Date: Wed Aug 10 23:16:16 2016

[mojo-edk] Revert ObserveProxy retransmission behavior

Gets rid of ObserveProxy retransmission which was only required
to facilitate excessive randomization in tests. In practice this
behavior can race with normal port closure and lead to busy
loops.

Gets rid of excessive randomization in tests. The new PortsTest
environment more closely emulates reality: Nodes operate
independently but there are some basic and trivially satisfiable
ordering guarantees.

Gets rid of deferred AcceptIncomingMessages pumping in
NodeController, which was only added to facilitate
ObserveProxy retransmission.

BUG= 628492 ,635508

Review-Url: https://codereview.chromium.org/2219733005
Cr-Commit-Position: refs/heads/master@{#410535}
(cherry picked from commit 36f85398960e7836d5325e1a58298209ced34f33)

Review URL: https://codereview.chromium.org/2236473003 .

Cr-Commit-Position: refs/branch-heads/2785@{#558}
Cr-Branched-From: 68623971be0cfc492a2cb0427d7f478e7b214c24-refs/heads/master@{#403382}

[modify] https://crrev.com/573af98298255787657627c25a78a5de1676da73/mojo/edk/system/node_controller.cc
[modify] https://crrev.com/573af98298255787657627c25a78a5de1676da73/mojo/edk/system/ports/node.cc
[modify] https://crrev.com/573af98298255787657627c25a78a5de1676da73/mojo/edk/system/ports/node.h
[modify] https://crrev.com/573af98298255787657627c25a78a5de1676da73/mojo/edk/system/ports/ports_unittest.cc

Components: Tests

Sign in to add a comment