← Back to team overview

openstack-qa-team team mailing list archive

XML Testing in Tempest

 

Hi all,

I've been working on XML API tests for tempest all week, after the
decision in the Nova meeting last week that XML needs to be supported
and tested. When I happened through #openstack-meeting today, I realized
I had missed a QA meeting where it was discussed that Daryl Walleck had
a bunch of yet-to-be-public XML tests, and that "some guy" was
submitting some separately. I'd the "guy" in this case :)

I didn't mean to cause any trouble by doing this -- it seemed to me that
nobody else was doing it (after some discussion with a few folks) and
that there wasn't likely a conflict to be had. I did see the mention of
Darryl's 9-month-old Zodiac tree, but it didn't seem like there was
movement intended or implied there. However, reading back through the
archives, it looks like I may have misinterpreted what Jay said about
Rackspace's intent to move their Zodiac stuff to tempest and submit.

On Monday, I talked to Jay a bit about the test strategy, and settled on
his suggestion of the two dialect tests inheriting from a common test
logic base class, per some prior art in glance testing. The patches I
have thus far are based on that assumption:

https://review.openstack.org/#/q/status:open+project:openstack/tempest+branch:master+topic:xml,n,z

After talking with David Kranz today, I filed a bug (1037693) at his
request with pointers to the patches I have up for review. I'm not sure
if discussion should continue here or there (I'd prefer here). However,
we have several other developers at IBM working on XML tests on top of
this infrastructure that I've started, so I'd surely like to figure out
what the path forward is before we get too much farther down this
path...  :)

Thanks!

-- 
Dan Smith
IBM Linux Technology Center


Follow ups