Re: [libvirt test] 33157: regressions - trouble: blocked/fail/pass/preparing/queued

[Jim Fehlig <jfehlig@suse.com>]

Ian Campbell wrote:
> On Thu, 2015-01-08 at 09:37 -0700, Jim Fehlig wrote:
>   
>> xen.org wrote:
>>     
>>> flight 33157 libvirt running [real]
>>> http://www.chiark.greenend.org.uk/~xensrcts/logs/33157/
>>>
>>> Regressions :-(
>>>
>>> Tests which did not succeed and are blocking,
>>> including tests which could not be run:
>>>  build-amd64-libvirt             <none executed>              queued
>>>  test-amd64-i386-libvirt         <none executed>              queued
>>>  test-amd64-amd64-libvirt        <none executed>              queued
>>>  build-armhf-libvirt           5 libvirt-build             fail REGR. vs. 32648
>>>  build-i386-libvirt            5 libvirt-build             fail REGR. vs. 32648
>>>   
>>>       
>> The build failure here was
>>
>> make[2]: *** No rule to make target `xenconfig/xen_xl_disk.h', needed by
>> `all'. Stop.
>>
>> Now fixed with libvirt commits edacdb3d and 533349ff.
>>     
>
> Great, thanks for the heads up.
>
> It looks like libvirt has imported the disk syntax parser stuff from
> xen.git/tools/libxl/libxlu_disk_l.l?

Yes.

> Why not just use libxlu.so which
> Xen ships containing that code? We do so pretty much exactly because we
> thought it would be useful to libvirt...
>   

Because I didn't know :-(.  And yes, it would be useful to libvirt, and
would have saved myself and others considerable time.  In hindsight, I
should have cc'd xen-devel on the original patch series.  I think Wei
was the only Xen tools dev to review the series, and he too missed your
suggestion.

> I must confess that I thought you already did use libxlu...
>   

No, not directly.  But I will be doing so now.  I should try to revert
all this nonsense and use libxlu before it ends up in the next libvirt
release.

Thanks a lot Ian!

Regards,
Jim