Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Talos Regression :( Dromaeo (DOM) decrease 2.45% on XP Mozilla-Inbound-Non-PGO

2 views
Skip to first unread message

nob...@cruncher.build.mozilla.org

unread,
Sep 1, 2012, 2:00:28 PM9/1/12
to dev-tree-...@lists.mozilla.org
Regression :( Dromaeo (DOM) decrease 2.45% on XP Mozilla-Inbound-Non-PGO
------------------------------------------------------------------------
Previous: avg 207.792 stddev 2.129 of 30 runs up to revision 7a86d43be76f
New : avg 202.705 stddev 0.814 of 5 runs since revision 77c30c1388ee
Change : -5.087 (2.45% / z=2.389)
Graph : http://mzl.la/PU324L

Changeset range: http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=7a86d43be76f&tochange=77c30c1388ee

Changesets:
* http://hg.mozilla.org/integration/mozilla-inbound/rev/078f3af11d13
: Boris Zbarsky <bzba...@mit.edu> - Bug 786105. Setting inline style properties to null should remove them, just like setting them to empty string does. r=peterv
: http://bugzilla.mozilla.org/show_bug.cgi?id=786105

* http://hg.mozilla.org/integration/mozilla-inbound/rev/f51b4777dbf8
: Siddharth Agarwal <sid.bu...@gmail.com> - Bug 784910 - Pymake: adjust sys.path so that native commands can load other modules from their own directories. r=ted
: http://bugzilla.mozilla.org/show_bug.cgi?id=784910

* http://hg.mozilla.org/integration/mozilla-inbound/rev/77c30c1388ee
: Siddharth Agarwal <sid.bu...@gmail.com> - Bug 787600 - Pymake: Shell commands in submakes don't honour exported PATH and don't receive exported envvars. r=gps
: http://bugzilla.mozilla.org/show_bug.cgi?id=787600

Bugs:
* http://bugzilla.mozilla.org/show_bug.cgi?id=787600 - Shell commands in submakes don't honour exported PATH and don't receive exported envvars
* http://bugzilla.mozilla.org/show_bug.cgi?id=786105 - Setting .style.foo to null no longer removes the property
* http://bugzilla.mozilla.org/show_bug.cgi?id=784910 - Pymake: adjust sys.path so that native commands can load other modules from their own directories

nob...@cruncher.build.mozilla.org

unread,
Sep 5, 2012, 9:00:39 PM9/5/12
to dev-tree-...@lists.mozilla.org
Regression :( Dromaeo (DOM) decrease 2.49% on XP Mozilla-Inbound-Non-PGO
------------------------------------------------------------------------
Previous: avg 207.792 stddev 2.129 of 30 runs up to revision 7a86d43be76f
New : avg 202.626 stddev 0.822 of 5 runs since revision 078f3af11d13
Change : -5.166 (2.49% / z=2.426)
Graph : http://mzl.la/Slk3rB

Changeset range: http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=7a86d43be76f&tochange=078f3af11d13

Changesets:
* http://hg.mozilla.org/integration/mozilla-inbound/rev/078f3af11d13
: Boris Zbarsky <bzba...@mit.edu> - Bug 786105. Setting inline style properties to null should remove them, just like setting them to empty string does. r=peterv
: http://bugzilla.mozilla.org/show_bug.cgi?id=786105

Bugs:

Philip Ringnalda

unread,
Sep 6, 2012, 12:32:49 AM9/6/12
to
In case anyone's puzzled by that four-days-later message blaming one
push for a regression that was previously blamed on two pushes, if you
just load up self-serve for the push which didn't get a run of that
talos job, find where it's listed in the Builds section despite not
having actually run on that push, and rebuild it, then it will be run on
that push.

The regression itself looks like just noise to me, but we are not stuck
with the effects of coalescing, we can retrigger around them, even
though, https://bugzilla.mozilla.org/show_bug.cgi?id=787447#c10, I think
the regression script should be the one to do the retriggering before it
sends anything, not us after it sends something bogus.
0 new messages