Dreaming about cooperation with upstream [Was: Re: Upstream error reporting]

[Nils Philippsen]

On Thu, 2008-03-20 at 17:13 +0100, Matej Cepl wrote: 
> On 2008-03-19, 22:51 GMT, Philip Ashmore wrote:
> > I just filed
> >
> > https://bugzilla.redhat.com/show_bug.cgi?id=438269
> >
> > Forgive me if I'm wrong, but isn't there a policy in place for Fedora to 
> > be the "point of contact" for upstream bug reports so that
> > those wishing to improve the Fedora experience don't have to create 
> > multiple accounts with multiple upstream sources in order to report a bug?
> 
> yeah, life sucks. Sometimes.
> 
> and our tools are pitiful. Quite often.
> 
> In my ideal world, the workflow would be like this.
> 
> 1) You as a reporter file a bug in Fedora bugzilla and that would 
>    be mostly end of the story for you (barring some NEEDINFOs) --    
>    then you should just sit, relax and watch programmers to fix 
>    the bug.
> 2) Bug triager and/or developer finds out that this is not a bug 
>    in the distro packaging, configuration, etc., and in the same 
>    time it is not time-sensitive, security, super-important bug, 
>    so that it should be fixed upstream.
>    She clicks the button "UPSTREAM" and that is for some time the 
>    end of the story for her (unless, which is quite often the 
>    case, he is also working at upstream, so he can deal with the 
>    bug there).
> 3) Our bugzilla collects all information from the bug itself, and 
>    from other sources (fixing backtraces with our debuginfos, 
>    providing versions of the packages in question, default 
>    configuration, etc.) and files (via XML-RPC or something like 
>    that) a bug in the upstream database.
> 4) Upstream bug triagers/developers check that the bug is not 
>    duplicate, and do other necessary diaper-changing for the new 
>    bug.
>    Upstream developers when working on the resolving the issue, 
>    could file a NEEDINFO, which would be automagically sent to 
>    our bugzilla, and from that to you as a NEEDINFO from our 
>    bugzilla. Your reply would go other way around, of course. You 
>    are not logged in to the upstream bugzilla, nor you have an 
>    account there.
>    On every step (or every comment?) the upstream bugzilla sends 
>    some information to our bugzilla, so we (you as a reporter and 
>    our developers) can follow the progress of the bug.
> 5) Finally, when the bug is closed upstream, our package 
>    maintainer (or all package maintainers of distributions 
>    collaboratin with that particular upstream?) get a message 
>    (through our bugzilla of course, so that it is logged) that 
>    the issue was resolved and she should make an updated package 
>    for our distro.
> 6) When it is done, our bug closes (with possible QA, Errata 
>    process, etc. as usual).
> 
> Unfortunately, so far this is just a pipedream and from point 2) 
> to 5) it doesn't exist at all.
> 
> Unfortunately, there is a huge value for developer to have actual 
> real reporter (and not a fake one like Fedora bug triager) 
> available. Upstream reporter may need some information, he will 
> certainly need at least one attempt to reproduce the bug with his 
> fix included (which in turn may require cooperation of the 
> downstream package maintainer; what a mess :-(), he may (and he 
> probbaly will) find out that some assumption original reporter, 
> and bug triager had on their minds when filing the upstream bug 
> were wrong, etc.

Well, we already have some information about several upstream bugzillas
(let's keep this simple for the moment and ignore trac, sourceforge e.a.
bug trackers). In theory it should simply be a case of asking "Please
<link>create a bugzilla ticket upstream</link>. If you don't have an
account at $upstream_bugzilla, <link>register here</link>." Clicking on
the first link would drop the user into a pre-filled form that
automatically Cc's the Fedora maintainer etc. We could store the
identities at other Bugzillas with the user's account to make it even
smoother. Or maybe something like OpenID can be used for that.

Nils
-- 
     Nils Philippsen    /    Red Hat    /    nphilipp at redhat.com
"Those who would give up Essential Liberty to purchase a little Temporary
 Safety, deserve neither Liberty nor Safety."  --  B. Franklin, 1759
 PGP fingerprint:  C4A8 9474 5C4C ADE3 2B8F  656D 47D8 9B65 6951 3011