Re: [Xen-devel] [xen-unstable test] 15401: regressions - FAIL

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Ian Campbell writes ("Re: [Xen-devel] [xen-unstable test] 15401: regressions - 
FAIL"):
> On Mon, 2013-02-04 at 14:22 +0000, Ian Jackson wrote:
> > No.  Because the problem is essentially a fluke pass, not a fluke
> > fail.
> 
> So you are also proposing to flip something in osstest to erase the
> fluke pass from its memory? Otherwise won't it see all future fails as
> regressions, rather than never pass, due to the fluke pass?

No, right now I'm proposing to do a manual push of (as I proposed)
d1bf3b21f783.  The effect of that is that future pushes will be
regarded as regressions iff their test results are worse than
d1bf3b21f783's.

That is, the test system uses whatever revision non-staging has as the
baseline revision, not whatever it most recently tested.  (And if
non-staging hasn't had a test at all, it will test it.)

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel