Build with the break: http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Linux%2032/builds/11207.
I think this is my fault, I forgot to update this test as part of my tile culling patch - sorry. The test expectations for this test need to be modified (layer1 should now *not* draw once the opaque layer2 is added, since it is completely occluded by layer2 and thus gets culled). I'll submit a patch this morning, plus add another test case just to confirm that layer1 does draw if layer2 becomes non-opaque.
Created attachment 121697 [details] Patch
Created attachment 121702 [details] Patch
Removed "FAILS_".
Comment on attachment 121702 [details] Patch R=me. I'm surprised EWS bots don't run webkit_unit_tests, will look into that.
Comment on attachment 121702 [details] Patch Clearing flags on attachment: 121702 Committed r104505: <http://trac.webkit.org/changeset/104505>
All reviewed patches have been landed. Closing bug.