Re: [Xen-devel] [xen-unstable test] 22184: regressions - trouble: broken/fail/pass

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Ian Campbell writes ("Re: [Xen-devel] [xen-unstable test] 22184: regressions - 
trouble: broken/fail/pass"):
> On Mon, 2013-12-02 at 09:36 +0000, Jan Beulich wrote:
...
> > But then again - are the tests here using the newer qemu already?
> > QEMU_TAG in ./Config.mk seems to suggest they wouldn't (yet I
> > thought we got a push on the qemu-upstream tree)...
> 
> http://www.chiark.greenend.org.uk/~xensrcts/logs/22184/build-amd64/info.html 
> shows built_revision_qemuu b97307ecaad98360f41ea36cd9674ef810c4f8cf, which 
> appears to follow the qemu-1.6 merge changeset, so I think we are testing the 
> new version despite what Config.mk says. Which is good argument for updating 
> Config.mk such that users actually get what we have tested!

Indeed.

osstest has its own push gate for the "qemu upstream" branch.

And, and in general, osstest overrides Config.mk's selections of which
components to use.  This is so that it controls and knows exactly what
is being tested.

For the xen-unstable tests, it uses the tested qemu upstream (that is,
the push gate output).

I think Config.mk's upstream qemu should be updated to refer to
master.

Ian.

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