https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&searchStr=linux%2Cx64%2Casan%2Cweb%2Cplatform%2Ctests%2Cwith%2Ce10s%2Ctest-linux64-asan%2Fopt-web-platform-tests-e10s-7%2Cw-e10s%28wpt7%29&selectedJob=223482216&revision=2072956ca0421c3eb1835522546556bb00bf8953

There are several SEGV on Linux asan for encrypt-media, possible that they show up after bug 1519862 got fixed.

https://hg.mozilla.org/mozilla-central/rev/008355f71ca0

https://hg.mozilla.org/mozilla-central/rev/5d64e252185c

I suspect this and bug 1522547 are manifestations of the same issue.

Do you recall why this patch changed testing/web-platform/meta/websockets/dir.ini? It seems unrelated to the encrypted-media directory.

Sorry that this got into that changeset. That was a leak modification to fix: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&resultStatus=success%2Ctestfailed%2Cbusted%2Cexception&tochange=5d64e252185c4521d414bd197e4159a2971c1617&fromchange=f052949e151b1f2cfd9203a507a94ddd42392ec4&searchStr=wpt7%2Casan&selectedJob=223716749

(In reply to Sebastian Hengst [:aryx] (needinfo on intermittent or backout) from comment #9)

Sorry that this got into that changeset. That was a leak modification to fix:

Thanks for the explanation. The leak-threshold property only affects debug builds, not ASan, so it wouldn't actually have an effect here. That directly just has a permaleak in it that is being ignored, and the tree herder highlighter just shows it in the log summary. It doesn't actually make the tree orange. I'm probably going to file an issue about cleaning up this file anyways, and I wanted to understand why this file is the way it is.

I think fixing bug 1522547 should address the issue here: try run. Will put up a patch to re-enable tests.

Bug 1522547 fixed the issue causing these tests to crash on shutdown, which was
also resulting in asan crashes, so these can be re-enabled.

https://hg.mozilla.org/mozilla-central/rev/0d1df88e93f6

Looks like there are no more failures. Can this bug be closed as fixed now?

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+1] from comment #15)

Looks like there are no more failures. Can this bug be closed as fixed now?

Sounds sensible.