AssistantClient initialization failing on Mojo ChromiumOS FYI bot |
||||
Issue descriptionExample build: https://ci.chromium.org/buildbot/chromium.fyi/Mojo%20ChromiumOS/30532 Initialization of AssistantClient assumes that media::AudioManager::Get() will return non-null, which does not appear to be the case on the Mojo ChromiumOS FYI bot. It is unclear if this is because the AudioServiceOutOfProcess feature is enabled for the telemetry_perf_unittests task or there is some other cause.
,
Jun 12 2018
,
Jun 12 2018
Hmmm AudioServiceOutOfProcess should be enabled for Linux, but not for ChromeOS. Could you suggest what is this bot and what is it meant for? Also, is it really a P3 bug?
,
Jun 12 2018
AudioServiceOutOfProcess is rolling out asn an experiment on Linux (but not on ChromeOS). If this bot is a test-only config which builds/tests ChromeOS functionality on Linux, then the test can just disable AudioServiceOutOfProcess feature. If it's some configuration which goes into production, then we need to figure out what we should do about it.
,
Jun 12 2018
Adding jam@ who is the owner of this bot. Testing Chrome OS functionality on Linux is pretty common but it doesn't look like telemetry_perf_unittests is included in chromium.chromiumos.json which makes me think that we know something is a little broken about these tests when running them with a Chrome built with the target_os="chromeos" GN flag. That said, the fact that this is failing means we should take a close look at whether this experiment is going to be applied to Chrome OS by accident. Bumping to P2 because this is only an FYI bot. |
||||
►
Sign in to add a comment |
||||
Comment 1 by xiaoh...@chromium.org
, Jun 11 2018