Filed by: archaeopteryx [at] coole-files.de https://treeherder.mozilla.org/logviewer.html#?job_id=98954674&repo=autoland https://queue.taskcluster.net/v1/task/IsNqlt03Qt6LekiH-bjH6w/runs/0/artifacts/public/test_info//reftest-stylo_errorsummary.log https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://queue.taskcluster.net/v1/task/IsNqlt03Qt6LekiH-bjH6w/runs/0/artifacts/public/test_info//reftest-stylo_errorsummary.log&only_show_unexpected=1
This test is already fuzzy-if'd at https://dxr.mozilla.org/mozilla-central/rev/cad53f061da634a16ea75887558301b77f65745d/layout/reftests/svg/smil/reftest.list#241 - including stylo might be an easy win here.
Whiteboard: [stockwell needswork]
(In reply to Geoff Brown [:gbrown] from comment #5) > This test is already fuzzy-if'd at > https://dxr.mozilla.org/mozilla-central/rev/ > cad53f061da634a16ea75887558301b77f65745d/layout/reftests/svg/smil/reftest. > list#241 - including stylo might be an easy win here. Actually, this failure has a much greater # of different pixels -- all the edges.
:hiro - This started failing shortly after bug 1357295 -- related?
Yes, of course, this is related to bug 1367327, but the failure frequency raised after bug 1367327? It's really odd, since bug 1367327 changed stroke-width parser to match to gecko's behavior.
This started failing on May 14, shortly after bug 1357295 landed, and increased in frequency May 28, a couple of days after bug 1367327 landed (but that was a weekend, and the increase was "only" about 2x, so, its hard to say if the increase was related to 1367327).
this week we are on track for the same failure rate as last week. :hiro, can you pick up fixing this bug? either random-if or skip-if is ok, possibly there is a real fix we can do.
Comment on attachment 8876530 [details] Bug 1364731 - Specify px unit explicitly. https://reviewboard.mozilla.org/r/147856/#review152198 I'm fine with doing this in the interests of enabling this test, but I also feel like we may as well just fix the underlying issue (precision of lengths in transform lists) and not touch this test.
Attachment #8876530 - Flags: review?(bbirtles) → review+
(In reply to Brian Birtles (:birtles) from comment #13) > Comment on attachment 8876530 [details] > Bug 1364731 - Specify px unit explicitly. > > https://reviewboard.mozilla.org/r/147856/#review152198 > > I'm fine with doing this in the interests of enabling this test, Actually, now I see this is an intermittent failure. In that case this is totally fine!
Pushed by email@example.com: https://hg.mozilla.org/integration/autoland/rev/3118b9ee9db8 Specify px unit explicitly. r=birtles
Leaving ni to me here since I should provide a test case which reliably fails due to the precision mismatch.
Whiteboard: [stockwell needswork] → [stockwell fixed:other]
You need to log in before you can comment on or make changes to this bug.