δΊ 2013-4-9 23:45, John Ferlan ει:
On 04/09/2013 10:40 AM, John Ferlan wrote:
> On 04/08/2013 06:16 AM, Wenchao Xia wrote:
>> This is a patch follows John's 9 patches for cimtest, after it
>> cimtest should only fail 3 case on RH6.4:
>> HostSystem - 01_enum.py: FAIL
>> HostSystem - 03_hs_to_settdefcap.py: FAIL
>> VirtualSystemManagementService - 19_definenetwork_ers.py: FAIL
>>
>> This patch is only for review and test, it may need adjust and merge
>> with John's patch, and change to author name(not root :|), please
>> do not push directly.
>>
>> root (3):
>> test: common_util, use number to check version
>> test: rasd use int as comparation condtion for libvirt version
>> test: RPCS fix nfs issue
>>
>> .../12_create_netfs_storagevolume_errs.py | 2 +-
>> suites/libvirt-cim/lib/XenKvmLib/common_util.py | 32 ++++++++++++++++---
>> suites/libvirt-cim/lib/XenKvmLib/pool.py | 8 ++--
>> suites/libvirt-cim/lib/XenKvmLib/rasd.py | 7 ++--
>> 4 files changed, 36 insertions(+), 13 deletions(-)
>>
>>
>
> While it seems the change resolves some issues I saw in my initial run,
> I think the official patch needs to describe the problem/symptom and
> resolution more clearly. In particular, is the change because cimtest
> was improperly handling the result of the "virsh -v"? Was this only a
> rhel64 issue?
>
Uh, duh. Should have held off hitting send for just a few minutes.
My version on rhel64 is "0.10.2" while on my f18 system it was
"1.0.3",
so naturally when comparing against "0.4.1" I can "see" why the
change
was necessary. I can also understand why this is a "new" regression
since probably the last time tests were run the virsh version was
"0.9.*" or "0.8.*"...
While I agree what you did resolves some issues - I think the change is
incomplete. You've only changed a few places and my cscope tells me
there are 15 callers to virsh_version().
Let's see what I can come up with...
John
Yes, this fix all bugs show on my machine, maybe there are other
caller, but want to sent this first to make it work. I like
your patches which compare two version string from libvirt. :>
I got only 3 case fail now, may be you can paste some detail
of your fail here.
Something I know to make it work:
1 host must be able to resolve itself's name, that is ping [MACINENAME]
must succeed.
2 two fake image must be created in /var/lib/libvirt/images
3 default disk-pool default must be "virsh pool-undefine".
_______________________________________________
Libvirt-cim mailing list
Libvirt-cim(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-cim
--
Best Regards
Wenchao Xia