Bug 206437 - REGRESSION: ( r254256 ) [ Mojave wk2 ] http/wpt/mediarecorder/MediaRecorder-AV-audio-video-dataavailable-gpuprocess.html is a flaky failure
Summary: REGRESSION: ( r254256 ) [ Mojave wk2 ] http/wpt/mediarecorder/MediaRecorder-A...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebRTC (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: youenn fablet
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2020-01-17 13:15 PST by Truitt Savell
Modified: 2020-01-20 07:07 PST (History)
10 users (show)

See Also:

Attachments
Patch (4.97 KB, patch)
2020-01-20 02:51 PST, youenn fablet 		no flags Details | Formatted Diff | Diff
Patch for landing (4.97 KB, patch)
2020-01-20 06:13 PST, youenn fablet 		no flags Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Truitt Savell 2020-01-17 13:15:04 PST 
http/wpt/mediarecorder/MediaRecorder-AV-audio-video-dataavailable-gpuprocess.html 

Description:
This test has been failing sense its introduction in https://trac.webkit.org/changeset/254256/webkit

History:
https://results.webkit.org/?suite=layout-tests&test=http%2Fwpt%2Fmediarecorder%2FMediaRecorder-AV-audio-video-dataavailable-gpuprocess.html

Diff:

--- /Volumes/Data/slave/mojave-release-tests-wk2/build/layout-test-results/http/wpt/mediarecorder/MediaRecorder-AV-audio-video-dataavailable-gpuprocess-expected.txt
+++ /Volumes/Data/slave/mojave-release-tests-wk2/build/layout-test-results/http/wpt/mediarecorder/MediaRecorder-AV-audio-video-dataavailable-gpuprocess-actual.txt
@@ -1,4 +1,4 @@
 
 
-PASS Verify MediaRecorder is working in GPUProcess 
+FAIL Verify MediaRecorder is working in GPUProcess assert_true: the blob should contain some buffers expected true got false
Comment 1 Radar WebKit Bug Importer 2020-01-17 13:16:00 PST 
<rdar://problem/58692880>
Comment 2 youenn fablet 2020-01-20 02:20:16 PST 
Cannot reproduce currently.
One hypothesis is that we call stopRecording which after sometime and hopping through threads will set m_isStopped and m_hasStartedWriting to false.
Any further call to fetchData will then fail.
This might happen with GPUProcess since the stopRecording and fetchData calls are two IPC messages.
Comment 3 youenn fablet 2020-01-20 02:51:20 PST 
Created attachment 388222 [details]
Patch
Comment 4 Eric Carlson 2020-01-20 05:58:32 PST 
Comment on attachment 388222 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=388222&action=review

> Source/WebCore/ChangeLog:10
> +        If too much time happened betweent he two call,s stopRecording will trigger setting m_isStopped and m_hasStartedWriting to false.

s/call,s/calls/
Comment 5 youenn fablet 2020-01-20 06:13:17 PST 
Created attachment 388234 [details]
Patch for landing
Comment 6 WebKit Commit Bot 2020-01-20 07:07:45 PST 
Comment on attachment 388234 [details]
Patch for landing

Clearing flags on attachment: 388234

Committed r254823: <https://trac.webkit.org/changeset/254823>
Comment 7 WebKit Commit Bot 2020-01-20 07:07:47 PST 
All reviewed patches have been landed.  Closing bug.