New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 608002 link

Starred by 3 users
Status: WontFix
Owner:
█ lpique@chromium.org
Last visit > 30 days ago
Closed: May 2016
Cc:
jen...@chromium.org
█ h...@chromium.org
afakhry@chromium.org
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: Bug



Sign in to add a comment

Linux ChromeOS Buildspec Tests (build 639) failure with release-R51-8172.B

Project Member Reported by lpique@chromium.org, Apr 29 2016 
From https://uberchromegw.corp.google.com/i/chromeos.chrome/builders/Linux%20ChromeOS%20Buildspec%20Tests/builds/639

When running browser_tests there is an error in ExistingUserControllerPublicSessionTest.AutoLoginNoDelay

[23369:23369:0428/154449:FATAL:user_session_manager.cc(1208)] Check failed: host.

Looking at the source code for the user_session_manager.cc file, I see:

1206: void UserSessionManager::ActivateWizard(const std::string& screen_name) {
1207:   LoginDisplayHost* host = LoginDisplayHost::default_host();
1208:   CHECK(host);

I'll sync down that version of Chrome and try to repro the failure on my Chromebook.

Comment 1 by lpique@chromium.org, Apr 30 2016

Labels: OS-Chrome
I initially did a "git checkout 51.0.2704.30", but was unable to reproduce the problem. I tried running the individual test, as well as the whole suite.

I went further and figured out how to sync usingthe DEPS manifest at https://chrome-internal.googlesource.com/chrome/tools/buildspec/+/master/releases/51.0.2704.30/DEPS

The first time I built and ran the individual test after doing that, I reproduced the same failure.

I then started a bisect with 51.0.2704.29 ... and it ended up pointing at the "Incrementing VERSION to 51.0.2704.30" commit, which didn't make sense.

I retried the test after syncing back to 51.0.2704.30 (including trying a force sync with the DEPS manifest, and rm-ing out/), and did not get the test to again fail.

I also tried running the single test twenty times, but was still unable to reproduce the failure.

It is either a flake, or something was cached/updated on my system that makes it no longer fail.

Since it is late on a Friday, I'll keep it open and give it one more try on Monday, syncing back to Chrome 50 running it, then trying 51 again.

Comment 2 by lpique@chromium.org, May 2 2016

Cc: afakhry@chromium.org jen...@chromium.org
Status: WontFix (was: Started)
I tried syncing back, and also syncing forwards (Fridays HEAD), but I was unable to reproduce it again. Closing.
►

Sign in to add a comment