On Wed, 13 Feb 2013 05:02:21 -0500 (EST)
Kamil Paral <kparal(a)redhat.com> wrote:
> As I'm sure you may have figured out with the email from
> reviewboard, I
> was finally able to get a code review system working with both email
> and a post-commit workflow.
> -
http://reviewboard-tflink.rhcloud.com/
Great work, Tim.
On a side note, I've been having success with the rbtools for
submitting code reviews as well. I'm planning to submit a config file
and add to the blocker tracking app's docs soon.
I just wonder, maybe fedora admins would be willing to give you
sufficient rights to manage the fedorahosted reviewboard instance
[1]? More teams would benefit from a properly working RB, not just
us. We would definitely need to do something about its speed, but I
assume that's something our infra team can do easily (move it to a
more powerful machine).
[1]
https://fedorahosted.org/reviewboard/
The last I talked to them, the plan was to get rid of the fedorahosted
instance and move more towards per-project or per-group instances. That
might be in the form of openshift or cloud instances.
Either way, until FAS integration works better I don't think that would
happen. Rumor has it that there will be an intern working on OpenID
functionality for reviewboard this summer, so hopefully it won't be
long.
I'm not against the idea of helping to make code reviews in fedorahosted
better but it sounds like those plans are pretty much on hold until
proper OpenID support exists upstream.
> As far as the namespace/DNS entries go, I figured that we could
> leave it as is for now. If we want to create an qa-devel openshift
> account, I'm happy to get that set up - now that I've figured out
> how to get the
> thing set up I can make a new one without much effort.
Assuming we go ahead with out private RB instance, is it possible to
set up a redirect from
reviewboard.qa.fedoraproject.org while keeping
the URL in the browser bar (i.e. not just a simple http redirect)?
Excuse my technical ignorance :-)
Yes, AFAIK, that is possible once we figure out where it'll live for
the long term. Whether that method scales well for infra is another
question, though. If we go that route, I'd rather not have the instance
in my account just for SPOF/busrule reasons.
My thought right now is that we're best off with a wait-and-see
approach. While not perfect, I think the current setup will work for
now and we can figure out a more permanent home once FAS auth is
possible.
Tim