<Regression> Mozilla-Inbound-Non-PGO - Dromaeo (CSS) - Ubuntu HW 12.04 - 3.38%

[nob...@cruncher.build.mozilla.org]

Regression: Mozilla-Inbound-Non-PGO - Dromaeo (CSS) - Ubuntu HW 12.04 - 3.38% decrease
--------------------------------------------------------------------------------------
Previous: avg 4081.704 stddev 35.716 of 12 runs up to revision 627c3d68be29
New : avg 3943.854 stddev 51.473 of 12 runs since revision 57cda2d1472f
Change : -137.850 (3.38% / z=3.860)
Graph : http://mzl.la/16iIwCw

Changeset range: http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=627c3d68be29&tochange=57cda2d1472f

Changesets:
* http://hg.mozilla.org/integration/mozilla-inbound/rev/3d58bd987452
: Michael Comella <michael....@gmail.com> - Bug 893910 - Reset FHR submission failure count after timeout. r=rnewman
: http://bugzilla.mozilla.org/show_bug.cgi?id=893910

* http://hg.mozilla.org/integration/mozilla-inbound/rev/57cda2d1472f
: Bobby Holley <bobby...@gmail.com> - Bug 898811 - Use the system error reporter for XBL compilation. r=mrbkap

The current XBL error reporter logs things to the console service, but not to
stderr. And in certain situations (especially if there's a parse error in
frontend XBL that leaves the browser in a half-baked state), it can be very
difficult to get to the error console.

So we should log to stderr too, which is exactly what the system error reporter
does.
: http://bugzilla.mozilla.org/show_bug.cgi?id=898811

Bugs:
* http://bugzilla.mozilla.org/show_bug.cgi?id=898811 - JS method parsing failure is swallowed
* http://bugzilla.mozilla.org/show_bug.cgi?id=893910 - Current day failure count in data submission does not reset after 24-hour period

[Matt Brubeck]

Inspecting the data, it looks more likely that this regression was
caused by:

Bug 899834 - Use a better hash function for the atoms table
https://bugzilla.mozilla.org/show_bug.cgi?id=899834
https://hg.mozilla.org/mozilla-central/rev/bad4a1691369