<Regression> Mozilla-Inbound-Non-PGO - a11y Row Major MozAfterPaint - Ubuntu HW 12.04 - 3.47%

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

Regression: Mozilla-Inbound-Non-PGO - a11y Row Major MozAfterPaint - Ubuntu HW 12.04 - 3.47% increase
-----------------------------------------------------------------------------------------------------
Previous: avg 348.542 stddev 2.965 of 12 runs up to revision 8ed0d31309bf
New : avg 360.625 stddev 2.993 of 12 runs since revision 47057b236b1a
Change : +12.083 (3.47% / z=4.075)
Graph : http://mzl.la/18prS4S

Changeset range: http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=8ed0d31309bf&tochange=47057b236b1a

Changesets:
* http://hg.mozilla.org/integration/mozilla-inbound/rev/61a824c4e122
: Geoff Brown <gbr...@mozilla.com> - Bug 868681 - Remove undefined reference to chromeWin in OfflineApps.js; r=mfinkle
: http://bugzilla.mozilla.org/show_bug.cgi?id=868681

* http://hg.mozilla.org/integration/mozilla-inbound/rev/cf94844ab83f
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 1 - Import a vanilla copy of r148833 of Blink core's Decimal class from http://src.chromium.org/viewvc/blink/trunk/Source/core/platform/. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/cc31f6926e56
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 2 - Add update.sh script to mfbt/decimal/. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/4bbed5d03005
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 3 - Fix bugs in Decimal::floor() and Decimal::ceiling(). r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/574bda36a38c
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 4 - Fix bug in Decimal serialization so that it won't display zero as 0.00 etc. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/101889486438
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 5 - Fix the handling of NaN in Decimal::operator== and Decimal::operator!=. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/84eb9e72b338
: Jonathan Watt <jw...@jwatt.org> - Bug 854531, part 6 - Add MFBT_API markers to Decimal methods. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=854531

* http://hg.mozilla.org/integration/mozilla-inbound/rev/1ca314d5783b
: Jonathan Watt <jw...@jwatt.org> - Bug 853525, part 7 - Make mfbt/decimal/Decimal{.h/.cpp} use Mozilla/std dependencies instead of Blink core dependencies. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=853525

* http://hg.mozilla.org/integration/mozilla-inbound/rev/1c2a11eb7db0
: Jonathan Watt <jw...@jwatt.org> - Bug 853525, part 8 - Enable building of mfbt/decimal. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=853525

* http://hg.mozilla.org/integration/mozilla-inbound/rev/f40200a44162
: Jonathan Watt <jw...@jwatt.org> - Bug 853525, part 9 - Changes to Decimal code as a series of patches for update.sh. r=Waldo
: http://bugzilla.mozilla.org/show_bug.cgi?id=853525

* http://hg.mozilla.org/integration/mozilla-inbound/rev/c35ac9704caa
: Jonathan Watt <jw...@jwatt.org> - Bug 853525 - Convert much of HTMLInputElement (step handling, validation, events, etc.) to use mozilla::Decimal (a Binary Coded Decimal type) to avoid many rounding issues when it has fractional step values. r=mounir
: http://bugzilla.mozilla.org/show_bug.cgi?id=853525

* http://hg.mozilla.org/integration/mozilla-inbound/rev/47057b236b1a
: Jonathan Watt <jw...@jwatt.org> - Bug 853525 - Mochitests for <input type=range> rounding issues. r=mounir
: http://bugzilla.mozilla.org/show_bug.cgi?id=853525

Bugs:
* http://bugzilla.mozilla.org/show_bug.cgi?id=868681 - Offline apps doorhanger not shown
* http://bugzilla.mozilla.org/show_bug.cgi?id=854531 - Import WebCore's Decimal class
* http://bugzilla.mozilla.org/show_bug.cgi?id=853525 - Convert much of HTMLInputElement (step handling, validation, some event handling, etc.) to use mozilla::Decimal to avoid many rounding issues with fractional step values

[Matt Brubeck]

It looks like this has already returned to normal. The un-regression
range is:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchanget=4995d566bd6d&tochangeset=8e98863ce994

This might have just been a random fluctuation; the regression was
almost within the noise level of this test. Other platforms did not see
the same regression. (I'm getting a little suspicious of the Ubuntu
machines, which seem to report more mysterious regressions than other
platforms.)