Maintainer Responsibilities

Adam Williamson awilliam at redhat.com
Wed Jun 3 21:44:12 UTC 2009


On Wed, 2009-06-03 at 11:25 +0100, David Woodhouse wrote:
> On Tue, 2009-06-02 at 16:17 -0400, Steve Grubb wrote:
> > 
> > I don't want to start a long thread, but just to ask a couple questions for my 
> > own clarification. Does a maintainer's responsibilities end with packaging 
> > bugs? IOW, if there is a problem in the package that is _broken code_ do they 
> > need to do something about it or is it acceptable for them to close the bug 
> > and say talk to upstream? 
> 
> There are _some_ kinds of bug (feature requests, etc.) which it's
> reasonable for any decent maintainer to punt upstream.
> 
> There are other kinds of bugs (crashes, security issues -- perhaps even
> _anything_ that's a real bug rather than an RFE) which the maintainer
> really _ought_ to deal with directly.
> 
> Opinions vary on precisely where the boundary between those classes
> should be, but I'm fairly adamant it should be 'RFE vs. bug'.

I'm with David on this one. To answer the initial question, this is not
explicitly codified anywhere AFAIK, and that's probably because it
_can't_ be. It's just not possible to reasonably say that all issues
that aren't introduced by actual Fedora packaging should or shouldn't be
reported upstream.

David's right in identifying the types of bug which it generally does
and doesn't make sense to move upstream, and the rest of the thread
obviously illustrates that we have different attitudes on the part of
different maintainers about it as well. Personally with my maintainer
hat on I generally like to keep reports open downstream as it helps me
remember stuff, but for a guy with as much work to do as Kevin on a
project as big as KDE, I can see how that would not be the case.

There's a third element, which is how active upstream is. If you know
damn well that upstream's been dead for six years or doesn't care about
the branch you're packaging any more, you pretty much have to keep the
bugs downstream (I know we try to avoid this situation in Fedora, but it
does occur sometimes).

I think we just have to be explicit about not being explicit here: it's
a judgment call on the part of the maintainer (and the Bugzapper for
that component, if there is one, working together with the maintainer).
As long as everyone bears in mind that the objective is to get the
problem fixed, whichever method is most likely to result in that
happening is fine as long as it's consistently applied.

As far as being nice to reporters goes, attitude is far more important
than process. 

"report this upstream" may not work great. 

"Thanks for reporting this issue: as it's a bug in the latest version of
the code, it will be fixed most quickly if you report it to the original
developers. Their Bugzilla is at http://bugzilla.foobar.com, and you
should report the bug against version X of component foobar." will work
a lot better. You can also add "Once the issue has been resolved
upstream, please re-open this bug to request we include the updated code
in Fedora" if appropriate. Yes, it's a lot of text, but you only have to
write it once, then stick it in a text file or the Stock Responses page
on the Bugzappers wiki section or wherever, and you're good to go for
all future upstreaming requests. Basically, I think that if you're
worried about trying to keep reporters happy, the way you interact with
them matters a lot more than exactly what it is you ask them to do.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net




More information about the devel mailing list