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

Issue 809123 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 807370
Owner:
Closed: Feb 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 2
Type: Bug-Regression



Sign in to add a comment

maps_pixel_test on Android failing on chromium.gpu/Android Release (Nexus 5X)

Project Member Reported by sheriff-...@appspot.gserviceaccount.com, Feb 5 2018

Issue description

Filed by sheriff-o-matic@appspot.gserviceaccount.com on behalf of tmartino@chromium.org

maps_pixel_test on Android failing on chromium.gpu/Android Release (Nexus 5X)

Builders failed on: 
- Android Release (Nexus 5X): 
  https://build.chromium.org/p/chromium.gpu/builders/Android%20Release%20%28Nexus%205X%29

We've got two assertion failures so far today on this test. They're few and far between so it's not having a big impact on the build, but probably still worth investigating.
 
Labels: OS-Android
Owner: kbr@chromium.org
Status: Assigned (was: Available)
Assigning to kbr, who has made recent changes to this test.

Comment 2 by kbr@chromium.org, Feb 5 2018

Cc: kbr@chromium.org
Components: Internals>GPU>Testing Blink>WebGL
Owner: ynovikov@chromium.org
From this failing build:
https://ci.chromium.org/buildbot/chromium.gpu/Android%20Release%20%28Nexus%205X%29/1994
https://chromium-swarm.appspot.com/task?id=3b7fc4b746091e10&refresh=10&show_raw=1

We can see the generated image:

http://chromium-browser-gpu-tests.commondatastorage.googleapis.com/view_test_results.html?e082a634cab32b5dbbb20a44c28c4007d4bc805d_Android_Release_Nexus_5X__telemetry

I'm not sure what's gone wrong because the successful runs don't upload their images.

Yuly, any chance you could do a spot check and see what's wrong with these runs? Do we need to mark the test flaky? Because our harness interacts with the Maps team's test, we should be able to understand exactly what's on screen at the end of the test.

Thanks.

Cc: ynovikov@chromium.org
Labels: Type-Bug-Regression
Owner: bpastene@chromium.org
The image in #2 comes from a device in landscape orientation and not in portrait.
Portrait image contains slightly bigger portion of the map.

Catapult should be configuring the device to be in portrait, not sure why it would fail.
We do: https://chrome-internal.googlesource.com/infradata/config/+/master/configs/chromium-swarm/scripts/android.py#245

I don't see any errors in the bot's logs when that gets set. Could just be weird device flake.
FWIW, same device (build384-m1--device7) passed maps_pixel_test here: https://chromium-swarm.appspot.com/task?id=3b39cdca950ea110&refresh=10&show_raw=1

Maybe a race between changing orientation and running test?
I've observed it can take a while for Android to apply orientation change.
Mergedinto: 807370
Status: Duplicate (was: Assigned)
That device's user_rotation had a value=1. I'm fixing that in https://chrome-internal-review.googlesource.com/c/infradata/config/+/570958.

Seems to be a dupe of  bug 807370 ?

Sign in to add a comment