See bug 157681 comment 71 for steps to reproduce!

Note that the work-around in attachment 612095 [details] [diff] [review] should be removed once this bug is fixed.

Is this still a problem, or can we resolve it as WFM? It looks to me like "the work-around in attachment 612095 [details] [diff] [review]" never actually landed in the tree....

The workarounds landed as fuzzy-if annotations: https://dxr.mozilla.org/mozilla-central/search?q=742176 I guess you can remove them and see if the tests fail to see if this is still an issue...

(In reply to :Ehsan Akhgari (not reading bugmail, needinfo? me!) from comment #3) > The workarounds landed as fuzzy-if annotations: > https://dxr.mozilla.org/mozilla-central/search?q=742176 > > I guess you can remove them and see if the tests fail to see if this is > still an issue... See try run in comment 4. Note that only one of the tests failed, and it's apparently less "fuzzy" than it used to be, judging by the numbers in the annotation. Also, checking the reftest analyzer images, the discrepancy is NOT at the boundaries of the area being drawn, as we'd expect if it's a problem with visual overflow areas. Rather, it's one of the mid-word glyphs that is rendered with slightly different subpixel antialiasing between the testcase and reference. That's not an indication of overflow-area problems.

It's very well possible that the original issue has been fixed, or it still exists but this test case doesn't hit it any more. I don't think this bug is actionable any more.

OK, I propose we tighten up the annotations on these tests, and then we'll close this as WorksForMe given that we're not seeing overflow-area issues here any more.