1014808 - 2.18% CART regression on OSX 10.6 from may 21st on fx-team (fx32)

Status: RESOLVED FIXED

(Testing :: Talos, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

I found a slight cart regression on fx-team:
http://graphs.mozilla.org/graph.html#tests=[[309,64,21]]&sel=1400626859000,1400799659000&displayrange=7&datatype=running

I did some retriggers:
https://tbpl.mozilla.org/?tree=Fx-Team&fromchange=9f7d9d5c9261&tochange=164f8a2bcfba&jobname=Rev4%20MacOSX%20Snow%20Leopard%2010.6%20fx-team%20talos%20svgr

and it came down to this push:
https://hg.mozilla.org/integration/fx-team/pushloghtml?changeset=84c572a9c878

I am more concerned about:
https://hg.mozilla.org/integration/fx-team/rev/a9329165ace3 (bug 759252)
https://hg.mozilla.org/integration/fx-team/rev/84c572a9c878 (bug 1008355)

Of course there are some developer tools and android stuff mixed in there as well.

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

osx 10.8 shows a regression as well:
* tart 2.92% regression: http://mzl.la/1lX7gWx

:jaws, can you take a look at this?

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

it appears that we also have a 2.55% regression on linux32 around this time, looking as though this set of changes is the root cause:
http://mzl.la/1sZlWad

Comment 4

[Jared Wein [:jaws] (please needinfo? me)]

This is probably fixed by the patch in bug 1016942.

Comment 5

[Joel Maher ( :jmaher ) (UTC -8)]

cart has improved as of yesterday:
http://graphs.mozilla.org/graph.html#tests=[[309,64,21]]&sel=none&displayrange=7&datatype=running

thanks for linking this together