On http://kripken.github.io/mloc_emscripten_talk/#/2 if you press right-arrow or space, you'll see some links, to BananaBread, Qt, and Ruby. In Chrome, those animate when hovered over, and when clicked, they open the link. In Firefox, they look like plain text, and do not function as links, nor does the animation show.

The basic issue is that that location on screen doesn't seem to hit-test to that element. See what happens if you try to select it with the element picker in the devtools inspector. Now _why_ that is, I'm not sure yet. A smaller testcase would probably be pretty useful.

I investigated if this was anything specific to my slides, but looks like no, the problem happens on the official reveal.js demo as well, http://lab.hakim.se/reveal-js/#/3 So it likely affects all reveal.js-using projects. I also checked a few versions of Firefox. This runs ok on 44/aurora, so it looks like it is only broken on nightly currently - a recent regression?

[Tracking Requested - why for this release]: OK, this regressed (at least the "not animating" part; the clicking part broke sometime later) between the 2015-11-16 nightly and the 2014-11-17 nightly. That's the range in which bug 1208673 landed, which would be consistent with the transforms involved in this testcase and the other regressions from that bug...

I will look into it.

Still happens on m-c tip, so not a dupe of bug 1226904.

Tracking for 45+ since this is a recent regression.

The change that caused this regression was backed out from Firefox 45. A current Developer Edition build should now work correctly, or you can wait until 45 goes to Beta some time next week. https://hg.mozilla.org/releases/mozilla-aurora/rev/64ec448f156d

Are there plans to back this out of 46 and 47? The regression is still present on at least nightly/47. Meanwhile I found another testcase that is hit by this (test_html5_mouse in the emscripten test suite), let me know if it would be useful to post it.

We're hoping to fix the regressions. Not only is backing out risky (because it landed many moons ago, and there are probably many implicit dependencies by now) but APZ/e10s depends on some of the behaviour from those patches, and backing out could delay APZ/e10s.

Comment on attachment 8712467 [details] [diff] [review] Use correct Z order to sort perspective items Review of attachment 8712467 [details] [diff] [review]: ----------------------------------------------------------------- Thanks Matt!

Yes, thanks Matt! (I should just cc you on all the things, let you swoop in and fix them :p)

Matt, want to request uplift to aurora? Thanks!

Comment on attachment 8712467 [details] [diff] [review] Use correct Z order to sort perspective items Approval Request Comment [Feature/regressing bug #]: Bug 1168263 [User impact if declined]: Incorrect hit testing on some websites. [Describe test coverage new/current, TreeHerder]: Manually tested, added test [Risks and why]: Very low risk. [String/UUID change made/needed]: None

Comment on attachment 8712467 [details] [diff] [review] Use correct Z order to sort perspective items Fix for recent regression. We want links to work! Please uplift to aurora.

Seems this isn't entirely fixed. STR: 1) load http://yonet.github.io/d3-codeclass/#/9 in Dev Edition or Nightly 2) Click "JSFiddle" in the slide EXPECTED: page navigates to http://jsfiddle.net/yonet3d/cw6s96my/ ACTUAL: nothing. The same problem reproduces on http://lab.hakim.se/reveal-js/#/3, when trying to click on http://slides.com

It's probably better to file a new bug for comment 20, to avoid this bug from getting too confusing, since a patch already landed and was uplifted.

OK, happy to do that.