this is probably the same bug as bug 1017256, filing as separate as the platform is different and the test is as well. here is a graph of the failure: http://graphs.mozilla.org/graph.html#tests=[[281,131,25]]&sel=1401190795000,1401363595000&displayrange=7&datatype=running I did some retriggers to help reduce and mistakes due to noise: https://tbpl.mozilla.org/?tree=Mozilla-Inbound&fromchange=d94346125453&tochange=e368bba6dbbc&jobname=Windows%207%2032-bit%20mozilla-inbound%20talos%20svgr we go from 246-248 -> 255-260 when the 3 patches landed.
Unfortunately this is a different regression (with a different cause) than bug 1017256. And bug 1016437 does not affect Windows, so this must have been caused by bug 1009679.
And the regression is that layerizing the scroll indicator slowed down pageload + firstpaint of these four pages: https://hg.mozilla.org/build/talos/raw-file/tip/talos/page_load_test/svgx/composite-scale.svg https://hg.mozilla.org/build/talos/raw-file/tip/talos/page_load_test/svgx/composite-scale-opacity.svg https://hg.mozilla.org/build/talos/raw-file/tip/talos/page_load_test/svgx/composite-scale-rotate.svg https://hg.mozilla.org/build/talos/raw-file/tip/talos/page_load_test/svgx/composite-scale-rotate-opacity.svg Avi, do we have on-Talos profiling support for these tests? How can I use it?
when we uplifted to aurora, we seem to have a net win: http://graphs.mozilla.org/graph.html#tests=[[281,131,25],[281,52,25]]&sel=1397237120718,1405013120718&displayrange=90&datatype=running
We are approaching 3 months on this with no action, fx32 has shipped, Any concerns with closing this bug as wontfix?