layout/reftests/transform-3d/matrix3d-1a.html and layout/reftests/transform-3d/matrix3d-2a.html both test that a div with a particular matrix3d() transform is equivalent to another div. But the div just contains a line of text, and the transform sends the text off the screen to the top in both cases, so it's not actually testing that they're the same. In particular, matrix3d-2a.html is testing that "matrix3d(1,0,0,0, 0,2,0,0, 0,0,3,0, 4,5,6,1)" is equivalent to "scale3d(1,2,3) translate3d(4px,5px,6px)", but it's not -- it's equivalent to "translate3d(4px,5px,6px) scale3d(1,2,3)". These were added in part 13 of the patchset for bug 505115. Patch to be attached in a moment.
Created attachment 593916 [details] [diff] [review] Patch Reftests in this directory pass both before and after this patch. If I undo the reversal of translate3d() and scale3d() in matrix3d-2a.html, but still add background: gold, I get a TEST-UNEXPECTED-FAIL.
Autoland Patchset: Patches: 593916 Branch: mozilla-central => try Destination: http://hg.mozilla.org/try/rev/cff0c83c75b0 Try run started, revision cff0c83c75b0. To cancel or monitor the job, see: https://tbpl.mozilla.org/?tree=Try&rev=cff0c83c75b0
Try run for cff0c83c75b0 is complete. Detailed breakdown of the results available here: https://tbpl.mozilla.org/?tree=Try&rev=cff0c83c75b0 Results (out of 207 total builds): exception: 2 success: 177 warnings: 27 failure: 1 Builds (or logs if builds failed) available at: http://firstname.lastname@example.org