Re: [Xen-devel] [OSSTEST PATCH 3/3] smoke tests: Consider osstest revision when reusing builds

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Ian Campbell writes ("Re: [OSSTEST PATCH 3/3] smoke tests: Consider osstest 
revision when reusing builds"):
> Probably a bad idea, but I wonder: would comparing the hostflags required
> of the two build hosts take care of some of this?
> 
> e.g. some random job I just pulled up:
> 
>     share-build-wheezy-amd64,arch-amd64,suite-wheezy,purpose-build
> 
> This is a bit tenuous though, since really it is $r{$ident_suite} //
> $c{DebianSuite} which matters.

Also I don't think grobbling around in the runvars looking at
all_hostflags etc. is right.

> [...]
> > 3. After a manual force push of an untested osstest, there no suitable

Fixed (and the other typo).

> > builds on either xen-unstable or osstest.  The first
> > xen-unstable-smoke run will have to do all the builds.  However,
> > subsequent xen-unstable-smoke runs can just pick up those builds.
> > These same builds will be reused until a xen-unstable flight using the
> > new osstest produces a passing build.
> 
> 4. After a push from another tree whose gated output is used by xen
> -unstable-smoke (e.g. the linux-X.Y for the default kernel revision) then
> there will be no suitable builds in either the latest xen-unstable or
> osstest (neither of which are likely to have seen the linux push and built
> it before a smoke run occurs) in which case xen-unstable-smoke will do that
> build and then subsequently reuse it until a xen-unstable or osstest flight
> occurs which uses that Linux tree.
> 
> (is that worth mentioning? is it correct?)

No.  This reuse machinery does not consider the versions of anything -
except, now, osstest itself.  This is because the other trees'
versions are supposed to be intercompatible.

> Acked-by: Ian Campbell <ian.campbell@xxxxxxxxxx>

Thanks,
Ian.

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