Closed Bug 1080685 Opened 9 years ago Closed 8 years ago

Intermittent test_playback.html | vp9cake.webm timed out, should get 'ended'

Categories

(Core :: Audio/Video, defect)

Product:
Core
Component:
Audio/Video
Platform:
ARM
Gonk (Firefox OS)
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla40
Tracking Flags:
Tracking Status
firefox38 --- fixed
firefox39 --- fixed
firefox40 --- fixed
firefox-esr31 --- unaffected

People

(Reporter: RyanVM, Assigned: jwwang)

Details

(Keywords: intermittent-failure)

Attachments

(1 file, 2 obsolete files)

1080685_longer_timeout-v2.patch
4.32 KB, patch
jwwang
: review+
Details | Diff | Splinter Review 
Probably fallout from one of the test fixes landed today.

https://treeherder.mozilla.org/ui/logviewer.html#?job_id=2888170&repo=mozilla-inbound

builder 	b2g_emulator_vm mozilla-inbound debug test mochitest-debug-5
buildid 	20141009042406
builduid 	a0a4ea85d0804e9d9065e9e541a58aaa
results 	warnings (1)
revision 	6bf40bd453b08ea76224527e4340d0e590b43560
slave 		tst-linux64-spot-200
starttime 	Thu Oct 09 2014 08:48:52 GMT-0400 (Eastern Standard Time)

07:28:50 INFO - 418 INFO [finished test-4-quad.opus-44] remaining= vp9cake.webm-35
07:28:50 INFO - 419 INFO TEST-PASS | /tests/content/media/test/test_playback.html | [finished test-4-quad.opus-44] Length of array should match number of running tests
07:28:50 INFO - 420 INFO TEST-PASS | /tests/content/media/test/test_playback.html | [started test-5-5.0.opus-45] Length of array should match number of running tests
07:28:50 INFO - 421 INFO TEST-PASS | /tests/content/media/test/test_playback.html | test-5-5.0.opus: Name should match #3
07:28:50 INFO - 422 INFO TEST-PASS | /tests/content/media/test/test_playback.html | test-5-5.0.opus: Name should match #1
07:28:50 INFO - 423 INFO TEST-PASS | /tests/content/media/test/test_playback.html | test-5-5.0.opus duration (7.558104) should be around 7.558
07:28:50 INFO - 424 INFO TEST-UNEXPECTED-FAIL | /tests/content/media/test/test_playback.html | vp9cake.webm timed out, should get 'ended' - expected PASS
07:29:38 INFO - [Child 787] WARNING: stride not available, assuming width: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 408
07:29:38 INFO - [Child 787] WARNING: slice height not available, assuming height: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 413
07:29:38 INFO - [Child 787] WARNING: rotation not available, assuming 0: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 432
07:29:40 INFO - [Child 787] WARNING: stride not available, assuming width: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 408
07:29:40 INFO - [Child 787] WARNING: slice height not available, assuming height: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 413
07:29:40 INFO - [Child 787] WARNING: rotation not available, assuming 0: file ../../../../gecko/content/media/omx/OmxDecoder.cpp, line 432
07:31:05 INFO - 425 INFO TEST-OK | /tests/content/media/test/test_playback.html | took 315149ms
add logs to debug vp9cake.webm timeout in test_playback.html.
Assignee: nobody → jwwang
Status: NEW → ASSIGNED
Attachment #8540510 - Flags: review?(cpearce)
Comment on attachment 8540510 [details] [diff] [review]
1080685_add_logs_vp9cake.webm.patch

Review of attachment 8540510 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/media/test/manifest.js
@@ +221,5 @@
> +  { name:"bogus.duh", type:"bogus/duh", duration:Number.NaN },
> +
> +  // Test playback of a WebM file with vp9 video
> +  //{ name:"vp9.webm", type:"video/webm", duration:4 },
> +  { name:"vp9cake.webm", type:"video/webm", duration:7.966 }

Move this file to the bottom of the list to ensure logs for vp9cake.webm are not overwritten by other logs.
Attachment #8540510 - Flags: review?(cpearce) → review+
Let's see if these logs help us catch the problem.
Keywords: checkin-needed, leave-open
Attached patch 1080685_longer_timeout.patch (obsolete) — — Splinter Review 
Add more logs for debugging.
Attachment #8540510 - Attachment is obsolete: true
Comment on attachment 8591526 [details] [diff] [review]
1080685_longer_timeout.patch

Review of attachment 8591526 [details] [diff] [review]:
-----------------------------------------------------------------

We have peaks on test_playback.html recently (See bug 832768 comment 447 ~ bug 832768 comment 474). Most of them are on Android API 9 which are slow machines. I wonder if it will be green when given more time to run the tests.

::: dom/media/test/manifest.js
@@ +113,5 @@
>  // Used by test_playback, which expects no error event and one ended event.
>  var gPlayTests = [
> +  // Test playback of a WebM file with vp9 video
> +  //{ name:"vp9.webm", type:"video/webm", duration:4 },
> +  { name:"vp9cake.webm", type:"video/webm", duration:7.966 },

Since most timeouts fall in this file, I will move this file to the top to know if it timed out because we have some problems in vp9cake.webm or it is just because there isn't enough time to finish the tests.

::: dom/media/test/test_playback.html
@@ +11,5 @@
>  <script class="testbody" type="text/javascript">
>  
> +function isSlowPlatform() {
> +  return SpecialPowers.Services.appinfo.name == "B2G" ||
> +         navigator.userAgent.indexOf("Mobile") != -1 && androidVersion == 9;

Give more time for slow machines to finish the tests.
Attachment #8591526 - Flags: review?(edwin)
Comment on attachment 8591526 [details] [diff] [review]
1080685_longer_timeout.patch

Need to add more logs.
Attachment #8591526 - Flags: review?(edwin)
Comment on attachment 8591526 [details] [diff] [review]
1080685_longer_timeout.patch

Review of attachment 8591526 [details] [diff] [review]:
-----------------------------------------------------------------

https://treeherder.mozilla.org/logviewer.html#?job_id=6476718&repo=try
The logs show that the test is still running after timeout. I think longer timeout will definitely help.

::: dom/media/test/test_playback.html
@@ +11,5 @@
>  <script class="testbody" type="text/javascript">
>  
> +function isSlowPlatform() {
> +  return SpecialPowers.Services.appinfo.name == "B2G" ||
> +         navigator.userAgent.indexOf("Mobile") != -1 && androidVersion == 9;

It should be |androidVersion == 10| instead of |androidVersion == 9|. I will fix this nit in next patch.
Attachment #8591526 - Flags: review?(edwin)
Fix nits.
Attachment #8591526 - Attachment is obsolete: true
Attachment #8592060 - Flags: review+
https://hg.mozilla.org/mozilla-central/rev/7cb645a5d38e
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
status-firefox40: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla40
You need to log in before you can comment on or make changes to this bug.