[virt-tools-list] [Libvirt-ci] Build failed in Jenkins: virt-manager-test » libvirt-centos-7 #1474

[Daniel P. Berrange]

This latest virt-manager failure looks genuine:

On Thu, Jul 28, 2016 at 09:12:18AM +0000, ci at centos.org wrote:
> --- tests/cli-test-xml/compare/virt-install-kvm-f14-url.xml
> +++ Generated Output
> @@ -50,7 +50,7 @@
>        <mac address="00:11:22:33:44:55"/>
>        <model type="virtio"/>
>      </interface>
> -    <input type="tablet" bus="usb"/>
> +    <input type="virtio1.0-input"/>
>      <graphics type="spice" port="-1" tlsPort="-1" autoport="yes">
>        <image compression="off"/>
>      </graphics>
> @@ -109,7 +109,7 @@
>        <mac address="00:11:22:33:44:55"/>
>        <model type="virtio"/>
>      </interface>
> -    <input type="tablet" bus="usb"/>
> +    <input type="virtio1.0-input"/>
>      <graphics type="spice" port="-1" tlsPort="-1" autoport="yes">
>        <image compression="off"/>
>      </graphics>

Looks like there's two separate issues here

 - The test case is assuming usb tablet, but in fact pulling data
   from libosinfo. Fedora 23 records support for virtio1.0-input
   devices, so this seems to have tripped up the test assumptions

 - It looks like the code is just blindly copying the device
   name from libosinfo, into the guest XML, but this is not valid
   since there's not a 1-1 match between dev names in libosinfo
   and device model names in libvirt.

eg, a name virtio1.0-input in libosinfo maps to <input type="virtio">
in libvirt.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|