On Sun, Oct 30, 2011 at 06:17:51PM +0100, Thomas Spura wrote:
On Sun, 30 Oct 2011 10:07:02 -0700
Toshio Kuratomi wrote:
> On Sun, Oct 30, 2011 at 12:58:02AM +0200, Thomas Spura wrote:
> > On Thu, 27 Oct 2011 15:20:43 -0700
> > Toshio Kuratomi wrote:
> > > This release is intended mostly for getting the majority of
> > > hotfixes that are in infrastructure's puppet repository into an
> > > rpm package as well as pulling some other features (like a more
> > > readable captcha) into production.
> > Also the bugzilla plugin for changing the mail address there?
> > It would be really nice to have that feature.
> It would be but as far as I know, no one has done any work to create
> that. Any interested takers? It's not an easy fix but it would be a
> very, very helpful feature to have.
And what's this?
When I was looking into it, I wasn't able to get a local fas instance
running. (But was some really long time ago)
I hadn't seen that before... Mike, is that code finished and ready, just
needs to be packaged?
We'd also need to figure out how to get that information out to the bugzilla
syncing scripts/other users who need to access this.... We could special
case it on the servers (kinda defeats the purpose of having it in a plugin),
add it to the client code in python-fedora (but, at least with the methods
I see exposed right now, that means an extra query per user returned which
will be unacceptably slow -- probably needs a method to return all the
bugzilla emails which would mean client-side we'd have to wait for one extra
query per request -- that could be cached on the client-side which would
make it faster for scripts that make many calls to one fas objects but still
slower for scripts that either run once and then exit or have to use
separate fas instances (for instance, for threadsafety)).
(Now that I'm going over the problems... I have a vague recollection that
Mike might have queried me about these same topics long ago and pinged me
about it... if that's correct, maybe we never came up with good answers to
this and that's why we've never deployed this.)
Another option would be to break backwards compatibility and tell people
that bugzilla_email addresses have to be queried separately. It's been
quite a while since we had an API break so it's not out of the question to
do that. We might want to look at whether there's some other low hanging
fruit that we've put off due to API breakage at the same time, though.