994351 - media-element-source-seek-1 fails randomly on B2G emulator opt

Status: RESOLVED FIXED

(Core :: Audio/Video, defect)

Description

[Randell Jesup [:jesup] (needinfo me)]

Attachment: debugs from the failed runs (locally)

This fails with timeouts with the patches for bug 818822 applied; we're going to disable this test for now.  We should fix this test or find the underlying bug.

We made the test more stable by adding preload="auto" and loop = true; that still left it failing 5-10% of the time (including on local runs of crashtest-remote content/media/test/crashtests)

I'll attach a patch used to debug the problem.

I'll note that the emulator stops working randomly after some number of runs (I got a max of 70 runs), after which you may need to logout, reboot or even pull and build a new repo (I haven't debugged all the issues here).

Comment 1

[Karl Tomlinson (:karlt)]

Attachment: break playing/seeked cycle in test

Perhaps the cycle was growing the number of events exponentially somehow.

Comment 2

[Randell Jesup [:jesup] (needinfo me)]

(In reply to Karl Tomlinson (:karlt) from comment #1)
> Created attachment 8404332 [details] [diff] [review]
> break playing/seeked cycle in test
> 
> Perhaps the cycle was growing the number of events exponentially somehow.

Wouldn't the debugs have shown that then?

Comment 3

[Karl Tomlinson (:karlt)]

Comment on attachment 8404285 [details]
debugs from the failed runs (locally)

(In reply to Randell Jesup [:jesup] from comment #2)
> > Perhaps the cycle was growing the number of events exponentially somehow.
> 
> Wouldn't the debugs have shown that then?

I guess so.

>REFTEST TEST-START | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html
>REFTEST TEST-LOAD | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html | 68 / 72 (94%)
>starting test
>connected source to audioElement
>onloadedmetadata fired
>onloadeddata fired
>onplay fired
>onplaying fired
>onsuspend fired
>onwaiting fired
>onseeked fired
>onplaying fired
>onwaiting fired
>onplaying fired
>onseeked fired
>######################## BrowserElementChildPreload.js loaded
>TEST-UNEXPECTED-FAIL | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html | application timed out after 330 seconds with no output

This log shows a small number of onplaying/onseeked repeats, but the test would have removed reftest-wait in onseeked, so I guess something is busy, blocking the test harness from continuing.

>REFTEST TEST-START | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html
>REFTEST TEST-LOAD | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html | 68 / 72 (94%)
>starting test
>connected source to audioElement
>onloadedmetadata fired
>onloadeddata fired
>onplay fired
>onplaying fired
>onsuspend fired
>TEST-UNEXPECTED-FAIL | http://192.168.1.17:8888/tests/content/media/test/crashtests/media-element-source-seek-1.html | application timed out after 330 seconds with no output

This is quite different.  I tested that a seek on an element that is not playing still generated an onseeked event, so I don't know why the test stopped here.

I'll apply attachment 8404332 [details] [diff] [review] to the test, but it may not be the cause of the timeout.

Comment 4

[Karl Tomlinson (:karlt)]

FWIW I verified that attachment 8404332 [details] [diff] [review] does *not* resolve the timeout.

https://tbpl.mozilla.org/?tree=Try&rev=d3a2ee6aa63f

Comment 5

[Karl Tomlinson (:karlt)]

Landed attachment 8404332 [details] [diff] [review], now that the crash bugs have been fixed.
https://hg.mozilla.org/integration/mozilla-inbound/rev/dc0f0c7c308f

Also re-enabled the test:
https://hg.mozilla.org/integration/mozilla-inbound/rev/be5e437595e3

because try runs say it is now fixed:
https://tbpl.mozilla.org/?tree=Try&rev=f6b00487c3a3
https://tbpl.mozilla.org/?tree=Try&rev=e9a1cdd15719

Probably fixed by fixing the memory corruption in bug 998711 and/or bug 997152.

Comment 6

[Karl Tomlinson (:karlt)]

https://tbpl.mozilla.org/?tree=Try&rev=5d967ba4400a

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/dc0f0c7c308f
https://hg.mozilla.org/mozilla-central/rev/be5e437595e3