Re: [Xen-devel] [xen-4.2-testing test] 15450: regressions - FAIL - PUSHED

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Jan Beulich writes ("Re: [Xen-devel] [xen-4.2-testing test] 15450: regressions 
- FAIL - PUSHED"):
> >>> On 09.02.13 at 01:08, xen.org <ian.jackson@xxxxxxxxxxxxx> wrote:
> > flight 15450 xen-4.2-testing real [real]
> > http://www.chiark.greenend.org.uk/~xensrcts/logs/15450/ 
> > 
> > Regressions :-(
> 
> So why did this do a push?

I have investigated this and it was due to my recent efforts to try to
make the push gate look at bisection reports too.  I had earlier
introduced a bug in sg-report-flight, which produces the report and
looks for regressions.  When it looks at a flight other than the one
we are mainly looking at, it is still prepared to write "tolerable" to
the output file which controls the push.  This wasn't a problem when
the other flights were other full flight tests of the same version,
but when it looks at a bisection flight it sees only one column of the
test matrix.

Sorry.  I've fixed this bug now.

We need to decide what to do with xen-4.2-testing.hg non-staging.  I
would be tempted to rewind it, if that isn't a horrible thing to do.
It's not as bad with hg as it is with git: aiui if you pull from a
rewound repo, nothing bad happens.

I think this can be done with hg strip.

Ian.

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