[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Reporting bugs upstream

Dave Jones wrote:
On Wed, Jan 18, 2006 at 03:41:16PM -0500, John Ellson wrote:

> I more interested in the general user, or a user unfamiliar with a > particular package, having a single-point-of-contact.
 > I'm interested in not having reports discarded and lost just
 > because they not immediately fixable by the Fedora maintainer.

For the kernel package, I did a bit of both. Sometimes I play
internet traffic cop, routing the reporter to the right person/mailing list,
and other times (especially when theres >1 report) I'll take it
upstream myself. Neither of these are CLOSED->UPSTREAM though.

The only time I CLOSE->UPSTREAM reports is when the bug is against
something we don't actively support (like XFS).  Having the bug left
open in the Fedora bugzilla isn't going to solve anything at all.
It doesn't let the Fedora maintainer "solve" anything, no, but I think it is still valuable.
It is the single-point-of-contact record of the user's bug report.
To the user the problem it is still OPEN, even if it can't be fixed by the Fedora maintainer
(which is what OPEN->UPSTREAM would indicate).

(BTW. Dave, since we're having a one-on-one discussion here, I would like to say that I've never had a problem with your treatment of any of my kernel bug reports. I think thats because, effectively, you are a part of the upstream team and so the issue
doesn't arise.)

The only chance that bug is going to get fixed is inheriting it on an
upstream rebase if the upstream authors have been made aware of the problem.
Right. I just want this to be the maintainer's responsibility. I don't think that Fedora maintainers should simply close a bug expecting the user to deal with some completely different organization and interface. Its not unreasonable to assume the maintainer is already
familiar with the interface to upstream.

 > >What would be *really* awesome would be the possibility of having
 > >a facility in bugzilla to escalate a bug to upstream
> That would be great, but I wonder if its overly complicated and perhaps > thats why its not happening?

I think the biggest hurdle is probably that different bugzillas
have different input forms, so data needs to be marshalled into a form
the other end can understand.

I'm not surprised. Thats why I propose just a URL reference instead of an active coupling.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]