https://treeherder.mozilla.org/#/jobs?repo=try&revision=53497702c4a5d1ff17229725d5675dd3935318fe&selectedJob=70411050 is an attempt at enabling leak checking in web-platform-tests. It seems like we leak the world, but there are lots of track element tests in the log , so I guess we are doing something bad there.  https://public-artifacts.taskcluster.net/QJCv9MleRHqal87i3xUpdQ/0/public/logs/live_backing.log
https://treeherder.mozilla.org/#/jobs?repo=try&revision=7c97e3f7f5eef37e1c0d0c875d775d5f0b52d12a Let's see if the redirection fix in Bug 973384 fixed the leak.
The previous try is not applied the patch in Bug 1333114 for enabling wpt leak checking. See the following one: https://treeherder.mozilla.org/#/jobs?repo=try&revision=13808d6fcb777dd06d3baa00739fefbf88c5002c
Henry, did bug 973384 fix this?
I don't know actually. I was told that my fix for Bug 973384 *probably* also fix this bug so I had a try but ended up with not knowing how to interpret the try result. Not sure how "enabling wpt leaking test is going so far". Since Bug 973384 has landed, maybe :jgraham could suggest how we can try to enable wpt leaking test and see if it's been fixed. p.s. I am not certain if the patch I applied to enable wpt leak checking (in comment 2) is still valid.
Flags: needinfo?(hchang) → needinfo?(james)
You enable leak checking with |mach wpt --leak-check|. To get a try run with it on, you will need to add the --leak-check argument in testing/mozharness/configs/web_platform_tests/prod_config[_windows].py
You need to log in before you can comment on or make changes to this bug.