Filed by: wkocher [at] mozilla.com https://treeherder.mozilla.org/logviewer.html#?job_id=74102775&repo=mozilla-inbound https://queue.taskcluster.net/v1/task/N8T74jmXTga8Yyit56DrJA/runs/0/artifacts/public/logs/live_backing.log

I guess this crash is bug 1334583.

I thought initially fix for bug 1333539 raised the probability of this crash, but the fix had not yet merged into inbound at the revision in comment 0. https://hg.mozilla.org/integration/mozilla-inbound/file/86de6a5ecbe550ffe5cbd49c7cc4f40c2576e848/dom/animation/Animation.h#l279

(In reply to Hiroyuki Ikezoe (:hiro) from comment #1) > I guess this crash is bug 1334583. That seems right. If you look at the logcats for these failures, you can see the assertion: https://treeherder.mozilla.org/logviewer.html#?repo=try&job_id=86430564&lineNumber=2074 https://public-artifacts.taskcluster.net/KGxqeiKsQS689iNrhB6gfw/0/public/test_info//logcat-emulator-5554.log 03-25 10:40:10.932 792 807 I GeckoDump: ⰲ겿{"action":"log","time":1490463610934,"thread":null,"pid":null,"source":"mochitest","level":"INFO","message":"SpawnTask.js | Entering test change_duration_and_currenttime","js_source":"TestRunner.js"}ⰲ겿 03-25 10:40:11.321 792 807 I GeckoDump: ⰲ겿{"action":"test_status","time":1490463611321,"thread":null,"pid":null,"source":"mochitest","test":"chrome://mochitests/content/chrome/dom/animation/test/chrome/test_restyles.html","subtest":"undefined assertion name","status":"PASS","js_source":"TestRunner.js"}ⰲ겿 03-25 10:40:11.341 792 807 I GeckoDump: ⰲ겿{"action":"test_status","time":1490463611333,"thread":null,"pid":null,"source":"mochitest","test":"chrome://mochitests/content/chrome/dom/animation/test/chrome/test_restyles.html","subtest":"undefined assertion name","status":"PASS","js_source":"TestRunner.js"}ⰲ겿 03-25 10:40:11.672 792 807 F MOZ_Assert: Assertion failure: !IsNull() (Cannot compute with a null value), at /home/worker/workspace/build/src/obj-firefox/dist/include/mozilla/TimeStamp.h:527 Just a reminder, if we are not going to make progress on bug 1334583, we can effectively avoid these failures by skipping on Android Debug.

I am suspecting the reason why failure frequency of this raised in these days is the same as bug 1351042.

The frequency of this failure increased today or yesterday (now about a half of all runs on autoland in the last 12 hours). Please take a look.

Excitingly, I now have retriggered jobs back to April 24 that show this failure happening really frequently... https://treeherder.allizom.org/#/jobs?repo=autoland&filter-searchStr=15%2B%20debug%20m(c1&fromchange=934221c412fca006f410a77cb64981bcd0b5025f&group_state=expanded&tochange=2e0e49cb827f65d788d8b86a7239a0ef9814bff8&selectedJob=89353153 Some build/releng change, maybe?

possibly :gbrown would know if there was a tooling change. also I see an assert in the failure messages: 04-30 00:38:53.124 F/MOZ_Assert( 805): Assertion failure: !IsNull() (Cannot compute with a null value), at /home/worker/workspace/build/src/obj-firefox/dist/include/mozilla/TimeStamp.h:527 this is right in the mix of when we are failing the test case- possibly related or just a side effect after the crash?

I don't know of any recent Android tooling change. Based on :hiro's earlier comments, I'd say we should push for a fix in bug 1334583 or skip this test on Android Debug.

I'm going to work on a fix for bug 1334583 today. If I don't get something ready by the end of the day let's just skip this test on Android debug.

I didn't quite finish fixing bug 1334583 today so I'll finish it off tomorrow. Feel free to disable to test if it gets too annoying in the meantime.

Just to update here. I have patches up for review for bug 1334583 and a couple of dependent bugs. An initial try run for those looks promising.

overall this seems fixed