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

Issue 597843 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner:
Last visit > 30 days ago
Closed: Sep 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 3
Type: Bug



Sign in to add a comment

YT MSE tests need more than current default timeout on Android K OS

Project Member Reported by yini...@chromium.org, Mar 24 2016

Issue description

Version: 51.0.2687.0
OS: Android M

These tests fail with spitzer off. So it has lower priority. Half of these tests fail on Android L as well. (see 568686 that track Android L)
since our goal is Android has better user experience with spitzer, we may want to fix these API, but lower priority.

failed tests on Webm:
14 ~ 20, 23, 24, 29, 30, 32 ~ 54

failed tests on mp4:
14 ~ 20, 23, 24, 29, 30, 32 ~ 54

What steps will reproduce the problem?
(1) navigate to yt-dash-mse-test.commondatastorage.googleapis.com/unit-tests/tip.html
(2) click Run All

What is the expected output?
all tests pass

What do you see instead?
half tests failed

 
Labels: MSEscrubbed
Cc: yini...@chromium.org
Labels: Needs-Feedback
yiningc@, thanks for filing this. When you repro, are you keeping to the default test timeout, or are you increasing it? If using the default, can you try with a much larger value and report back?
matt, that's good information. After set timeout to 60000, all tests pass (except #51 which is tracked in cr/574514).
can we resolve this bug as is OR should we ask thom to update default timeout to 60000?
Cc: wolenetz@chromium.org
Labels: -Needs-Feedback
Owner: tdedecko@chromium.org
Summary: YT MSE tests need more than current default timeout on Android K OS (was: YT MSE tests fail on Android K OS)
Thanks for checking again with longer timeout.
Over to tdedecko@, since I think increasing the default timeout would help suppress such failures.
Please see: http://yt-dash-mse-test.commondatastorage.googleapis.com/unit-tests/instructions.html

You can increase the timeout by appending the param ?timeout=60000 


oh sorry, I just read the rest of the bug. I could increase the default but I think giving the option to specify the timeout helps resolve some of these issues. If I were to change the default would 40000 or 50000 be acceptable defaults? Or is 60000 the magic number in most cases?
yiningc@: Can you see if nothing regresses at a shorter timeout on the slowest device we test with? Say, 40000 or 50000? If not, I'd say just go with 60000 to prevent spurious flakes on these slower devices.
Matt, I need make a correction. It seems the failed tests on Android K now pass is NOT because of the increased timeout. All these tests pass on chrome 51.0.2693.2 magically even with the default 30000 timeout time. 
I also tried Android L with the same chrome build, unfortunately all the failed tests (tracked in 568686) still fail, increase timeout time to 60000 doesn't help.
So there is no need to change default timeout value now. I will resolve this bug but still need investigate cr/568686 when you have time. 
Can you confirm on Android K run (that's now passing with default timeout) that you're using Spitzer? IIRC, Dale switched some of the logic around for this last week.
Matt, I confirmed on Android K run, the spitzer is enabled. Actually, no matter spitzer is enabled or disabled (through chrome://flags, I also verified its enable/disable state on chrome://version), all MP4 and Webm tests pass in default timeout.
I tried M50 on the same device, there are still some tests fail. So it does have some magic changes happen on the new dev build 51.0.2693.2.
Cc: -yini...@chromium.org tdedecko@chromium.org
Owner: yini...@chromium.org
yiningc@ is this still an issue?
Status: WontFix (was: Assigned)
not see this issue anymore

Sign in to add a comment