On Thu, 30 Jan 2014 13:41:16 +0000 Wookey <wookey@wookware.org> wrote: > +++ Neil Williams [2014-01-30 13:25 +0000]: > > On Thu, 30 Jan 2014 13:51:40 +0100 > > Olivier Berger <olivier.berger@telecom-sudparis.eu> wrote: > > > > > I myself isn't very motivated to sponsor packages which I don't > > > use... but maybe I'm not noticing at all (or didn't read > > > how-can-i-help messages ;-). > > > > That's the crux of it - a sponsor needs some level of interest in > > the package, beyond whether it is scheduled for removal. Fly-by > > sponsoring is not helpful, the sponsor needs to keep an eye on the > > package - and work with the maintainer - for the long term. > > In general I'd agree with that, but there are exceptions. Defining those exceptions isn't easy. 0: It would seem to be good to allow quick sponsoring for packages with an active maintainer but equally it is better for the maintainer to have a regular sponsor, especially if DM is being considered. 1: stale packages could go to Orphaned - at which point it becomes simple to do quick sponsoring as a QA upload. 2: sponsoring a new upstream is enough work that a sponsor needs some kind of familiarity and interest in the package. 3: minimal changes may not attract interest by being too "trivial". 4: During a BSP, there should be sponsors available for any RC bug fix but not for more intrusive changes. 5: the workload of sponsoring an RC bug fix is roughly equivalent to verifying a patch in the BTS and doing the upload yourself. > Yes, they should get DM status, but there will always be some > sponsored uploads first. We probably lose (good) people when they do > this sort of work and then just get ignored. Quite likely, but DM status is less likely if the each sponsored upload was done by a different sponsor. > The problem for a maintainer with no particular interest in a package > is determining whether the above is the case for a particular > package/request or not. > > I've sponsored a couple of things where clearly someone had done a > reasonable job of bug-fixing/updating/backporting a poorly-maintained > package and despite me having never heard of it or them before, doing > an upload for them was helpful. > > Clearly long-term relationships and maintainership is better, and > uploaders remain responsible for what they upload, but getting more > (competent) people involved and making it easy for them to get > started, and just simply uploading stuff that needs uploading when > bugs have been fixed, is good too. > > So I disagree that 'fly-by sponsoring' is always 'not helpful'. > Sometimes it _is_ helpful. Any tools that made it easier to judge when > would be good. Just being on the mentors list gives a poor overview of > how much is slipping through the net. Personally, I'm wondering how much of what is "slipping through the net" is actually worth worrying about. There are still quite a lot of packages in Debian which are worth removing. e.g. I'm more tempted to think about sponsoring fixes for packages already in Debian than any ITA's. -- Neil Williams ============= http://www.linux.codehelp.co.uk/
Attachment:
signature.asc
Description: PGP signature