Bug 222422 - REGRESSION (r273461): [ macOS ] imported/w3c/web-platform-tests/media-source/mediasource-duration.html is timing out
Summary: REGRESSION (r273461): [ macOS ] imported/w3c/web-platform-tests/media-source/...
Status: NEW
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2021-02-25 10:01 PST by Robert Jenner
Modified: 2021-02-26 14:14 PST (History)
3 users (show)

See Also:


Attachments
Patch (2.17 KB, patch)
2021-02-25 14:45 PST, Robert Jenner
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Robert Jenner 2021-02-25 10:01:53 PST
imported/w3c/web-platform-tests/media-source/mediasource-duration.html

is timing out in macOS Catalina and Big Sur, in release and Debug, for Wk1 and Wk2. 

HISTORY URL:
https://results.webkit.org/?suite=layout-tests&test=imported%2Fw3c%2Fweb-platform-tests%2Fmedia-source%2Fmediasource-duration.html

DIFF URL:
https://build.webkit.org/results/Apple-BigSur-Release-WK2-Tests/r273484%20(510)/imported/w3c/web-platform-tests/media-source/mediasource-duration-pretty-diff.html

DIFF TEXT:
--- /Volumes/Data/slave/bigsur-release-tests-wk2/build/layout-test-results/imported/w3c/web-platform-tests/media-source/mediasource-duration-expected.txt
+++ /Volumes/Data/slave/bigsur-release-tests-wk2/build/layout-test-results/imported/w3c/web-platform-tests/media-source/mediasource-duration-actual.txt
@@ -1,11 +1,5 @@
+#PID UNRESPONSIVE - WebKitTestRunner (pid 27787)
+FAIL: Timed out waiting for notifyDone to be called
 
-FAIL Test seek starts on duration truncation below currentTime The object is in an invalid state.
-FAIL Test appendBuffer completes previous seek to truncated duration The object is in an invalid state.
-FAIL Test endOfStream completes previous seek to truncated duration The object is in an invalid state.
-PASS Test setting same duration multiple times does not fire duplicate durationchange
-PASS Test setting the duration to less than the highest starting presentation timestamp will throw
-PASS Truncating the duration throws an InvalidStateError exception when new duration is less than the highest buffered range start time of one of the track buffers
-PASS Truncating the duration throws an InvalidStateError exception when new duration is less than a buffered coded frame presentation time
-PASS Increasing the duration does not trigger any SourceBuffer update
-PASS Increasing the duration during media playback does not trigger any SourceBuffer update
-
+#EOF
+#EOF
Comment 1 Radar WebKit Bug Importer 2021-02-25 10:02:19 PST
<rdar://problem/74748843>
Comment 2 Robert Jenner 2021-02-25 10:06:20 PST
Timeouts started at changeset r273461:

CHANGESET URL:
https://trac.webkit.org/changeset/273461/webkit

Changes appear related to the test that is timing out. At this time we cannot reproduce due to issues with the builder bots. The builder bots went down yesterday, and there hasn't been new builds available since yesterday at 1PM. Will test when bots come back up.
Comment 3 Robert Jenner 2021-02-25 12:39:46 PST
This test has been constantly failing for a very long time. The test expectations for it, were set to expect a failure since r206264
CHANGESET URL:
https://trac.webkit.org/changeset/206264/webkit

Starting at r273461, the test started to timeout. At this point, I'm not certain if we just need to update test expectations again, revert changes made at r273461, or modify the test.
Comment 4 Robert Jenner 2021-02-25 13:41:22 PST
Moving forward we are going to treat this as a regression. I was able to reproduce the timeout at tip of tree using the following test:

run-webkit-test imported/w3c/web-platform-tests/media-source/mediasource-duration.html  --iterations 100 -f 

Now, the tricky thing about this one, is that it has been failing for a while, but the test expectations were marked so it passed on the dashboard. So we need to decide if we revert the changes at r273461, or try and fix this test.
Comment 5 Robert Jenner 2021-02-25 14:45:08 PST
Created attachment 421568 [details]
Patch
Comment 6 Truitt Savell 2021-02-25 14:47:11 PST
Comment on attachment 421568 [details]
Patch

Clearing flags on attachment: 421568

Committed r273514 (234589@main): <https://commits.webkit.org/234589@main>
Comment 7 Robert Jenner 2021-02-25 14:47:41 PST
Updated test expectations to include Slow to determine if changes made in r273461 are a regression, or if it made the test run better. 

Before the change the test would fail in 21 seconds. Now, it times out at 60 seconds. Changes made may have fixed part of this test, and that is what we are looking to determine.
Comment 8 Truitt Savell 2021-02-26 14:14:04 PST
It looks like the work to fix this test is happening in https://bugs.webkit.org/show_bug.cgi?id=222462