Re: Proposal: have NM and DM sign a package maintainer pledge
Hi,
On Sun, 20 Sep 2009, Stefano Zacchiroli wrote:
> > As a package maintainer, I will do my best to help the Debian project
> > release a stable version of our operating system. In particular, I will
> > work together with the release team and I will keep all packages
> > associated to my name free of release critical bugs.
>
> I'd rephrase this last sentence as "... and I will do my best to keep
> all packages ...". IOW you should ask applicants to state their
> "guarantee of invested efforts" rather than "guarantee of result"
> (nobody can give the latter, everybody should give the former).
Ok.
> Beside the above wording remark, I've a problem with the first part of
> the pledge: it is too limited in scope.
It's intended. As I said it must apply to all package maintainers (DM as
well, not only future DD).
> The Release Team, routinely before releases, complaints (rightful!) that
> too few people work on RC bug which are not "theirs". The current NM
> process, IMHO, does not stress enough the fact that being a DD is also
> about something more than your own packages.
If all DD/DM applied what this pledge says, we would not have to rely so
much on having DD fixing someone else's package.
> I would add to the pledge a statement that, once your packages are in
> good condition, you should look forward to fix RC bugs in other
> packages. If we agree that this should be part of the pledge text, I can
> draft a corresponding sentence.
I don't think it's a good idea. And it would only be acceptable for DD,
and not all DD agree with this.
I want this pledge to be uncontroversial so that its message remains clear
and effective.
On Sun, 20 Sep 2009, Michael Banck wrote:
> To be honest, I think it goes into way too much detail about how RC bugs
> should get handled. Of course, RC bugs are important, but if we
> introduce something like this, it seems a bit blown out of proportion.
Well, I drafted it precisely because I'm annoyed of having so many
maintainers ignore their RC bugs.
> What about trying to work gracefully with other members of the Debian
> community, as well as upstream and downstream developers? What about
> treating our users respectfully? What about striving to be excellent in
> your packaging?
If you find a wording for those points that is acceptable to everybody
then I'm fine to incorporate those but I fear it will be difficult.
As I stated, I want this text to be the minimal set of expectations
that we all agree on.
> If we are just worried about RC bugs (or RC bugs as result of overworked
> developers), I think we can handle this just fine right now (NMUs
> followed by forced orphaning from QA in coordination with the DAM).
I think the graphs show clearly that we don't. Having so many NMU and
force orphan by the QA team is a sign of failure of the current
maintainers. I want to avoid that as much as possible in the future.
(The key distinction for me is pro-active vs reactive)
Cheers,
--
Raphaël Hertzog
Reply to: