Bug 1451966 ("Intermittent [u'svg/foreignObject-img-form-theme.html', u'==', u'svg/foreignObject-img-form-theme-ref.html'] | assertion count 1 is more than expected 0 assertions") is bad enough, but even more horrifying is the duplicate bug 1452044 ("Intermittent (u'svg/foreignObject-img-form-theme.html', u'==', u'svg/foreignObject-img-form-theme-ref.html') | assertion count 1 is more than expected 0 assertions") (where, if like me you're not seeing it, the difference is () versus []). https://hg.mozilla.org/mozilla-central/rev/3585c6359fa2 seems to have touched output handling a bit, so I'll start my suspicions with bug 1353461

Made the mistake of reading this on the weekend. Better needinfo myself so I don't forget to take a look on Monday.

I'm fairly sure this wasn't caused by bug 1353461, but I found the fix, so not going to bother trying to find the actual culprit (assuming this is even a regression).

Comment on attachment 8966262 [details] Bug 1452432 - Convert 'test' field to a string in assertionCount StructuredLog.jsm messages, https://reviewboard.mozilla.org/r/235012/#review240660

Pushed by ahalberstadt@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/1e4c7821c157 Convert 'test' field to a string in assertionCount StructuredLog.jsm messages, r=gbrown

Ah, my mistake: looks like the regression, more like 6 months old, came from fixing the way we used to report an assertion count failure in, e.g., svg/foreignObject-img-form-theme.html == svg/foreignObject-img-form-theme-ref.html, as just "svg/foreignObject-img-form-theme-ref.html | assertion count..." without mentioning the test filename, only the reference filename.