[libvirt] [PATCH v2 00/10] Make loading domains with invalid XML possible
John Ferlan
jferlan at redhat.com
Fri Dec 11 12:15:22 UTC 2015
[...]
>>
>> Beyond the few noted spots changes look good to me. Implicit ACK for
>> those not specifically noted.
>>
>
> Thanks a lot, but there is still the security issue and a crash
> mentioned by Luyao. I know how to deal with only a part of it. Anyway,
> this will need another version, so I'll include all the nits pointed out
> in there, but it'll take some time again because from my POV this is
> just a nice-to-have feature and nobody is asking for this, so there are
> other, more pressing, things in the priority queue and hence this one
> will have to wait again. Anyway, thanks again for checking this out.
right understood - although if you wanted to make at least some progress
- it seems patches 2-5 are separable. Patch 2&3 adds a "nice to have"
--reason output for the virsh table which is useful without the invalid
XML output...
Although, as I thought about 2-3 a bit more, I have to think there are
other tests out there (like virttest/avacado) which compare output of
commands to stock/previous output and complain about failures when
something is slightly different. I recall having to fix up a bunch when
the output changed to add a single space before the header and data
rows... E.G from:
Id ...
--- ...
1 ...
to
Id ...
--- ...
1 ...
So those 3 missing '-' could cause agita for someone.
John
More information about the libvir-list
mailing list