Re: [Xen-devel] [xen-unstable test] 12353: tolerable FAIL

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Jan Beulich writes ("Re: [Xen-devel] [xen-unstable test] 12353: tolerable 
FAIL"):
> Would it really make sense to block hypervisor/tools pushes when
> kernel builds fail?

In principle not but ...

> There are no shared sources, so the builds should
> be completely independent. Of course, it tests can't be run because
> of the build failure, that is a blocking reason,

... in general the system doesn't regard blocked tests as blockers.
It expects the depended-upon action to be a blocker in itself.

>  but then the question
> is whether everything should be built from scratch in the first place
> (rather than using pre-existing binaries either from an earlier
> successful run of the same tests or from the most recent successful
> run against the corresponding tree).

That would be possible and I have some code to do this for bisections
but it's more complicated here.

In practice these builds fail rarely enough that it's not a problem.
If they do fail because the upstream sites are down they are failing
for other people too and I would think that the best way to deal with
that is to reduce the number of sites we depend on.

Ian.

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