In bug 1353461 we are in the process of switching reftest to "run-by-manifest" mode. This means that between every manifest of tests we will restart Firefox. In general this leads to improved stability, though in a few cases existing tests that happened to depend on side-effects of previous tests can start failing. This is one of those cases. In this instance, there is a perma memory leak after running the tests in layout/reftests/webcomponents/reftest.list. Here is a push showing this: https://treeherder.mozilla.org/#/jobs?repo=try&revision=e53ddd2a3870117db53f466219fcfed9ffade9bd Example log: https://treeherder.mozilla.org/logviewer.html#?job_id=163442959&repo=try&lineNumber=38234 We'll need to fix this before landing run-by-manifest, or else we'll need to skip the entire manifest on OSX debug.

Emilio, is this something you might be able to help with?

That doesn't look like a memory leak at all, it's an assertion in nspr: Assertion failure: 0 == rv, at /builds/worker/workspace/build/src/nsprpub/pr/src/pthreads/ptthread.c:277 Looks somewhat related to bug 1399746. Glandium, any idea?

Thanks, my mistake. Feel free to dupe if it's the same.

This is somewhat good news, because it means we can actually finally reproduce this more reliably, with a Firefox build.

Actually, it *is* a leak. The nspr assert comes *afterwards*, during shutdown. It does help with reproducing that nspr assert, though, since it's so reliably happening.

I'm stunned, the leak message actually goes away when bug 1399746 is fixed.