Adding Gregor Herrmann to this because he and I were looking to work on #672198 but we both were swamped with other work. On Friday, February 07, 2014 00:02:16 Sebastian Reichel wrote: > On Thu, Feb 06, 2014 at 03:27:51PM -0500, Chris Knadle wrote: > > On Thursday, February 06, 2014 13:59:59 Sebastian Reichel wrote: > > [...] > > > > NMUs don't necessarily need to be minimalistic -- for instance packaging > > new versions is something that can be done with NMUs. This is admittedly > > not terribly clear and I raised this question in #672198 which hasn't had > > any activity for almost two years. > > Yes, but its advised to change as little as possible. I think > changing the packaging style from cdbs to debhelper and similar > changes are not ok (note: this is just an example). I know; I agree with you and I think the text is a bit misleading -- by stating that you shouldn't change the packaging style it seems to indicate that NMUs are supposed to be minimalistic, but a situation in which the maintainer of a package disappears for an extended period is exactly a situation in which a minimalistic change approach won't work. > > For cases where the maintainer is unresponsive for an extended period, I'd > > recommend requesting a new version via a 'wishlist' bug, then releasing a > > new version as a -0.1 NMU. Others (myself included) have done this > > successfully. > > I opened the wishlist bug entry for that (#723781) in September and > agree, that uploading a -0.1 NMU would solve the issue of the new > upstream version. When I last did this in #728545 for mumble, the situation was rather serious because it had been removed from jessie due to package dependency issues and needed to get fixed ASAP. So I opened a wishlist bug, then waited about a week, then uploaded a package for review to mentors.debian.net and started hunting for a DD sponsor. I contacted the prior maintainer, who examined the package and decided it was good enough and uploaded it to the DELAYED/5 queue. Then I wrote to the bug to notify the maintainer in case he needed more time to respond and review the package if needed. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=728545 I didn't change the packaging style in doing this, but just about everything else did. ;-) Obviously I was willing to support the package if there were problems brought by my sponsored upload, and as long as you keep this in mind as well then I think this practice should work. > OTOH having an active maintainer is more helpful than lots of NMUs > IMHO. Thus it makes more sense to take over packages or add at least > add a Co-Maintianer for this. Right, exactly. But to start with you may not want to do that; the maintainer normally gives approval for adding a co-maintainer. After you've done several NMU uploads and tried to contact the maintainer via the MIA team, then after that I think the next logical step I think is to add one's self onto the list of Uploaders... basically only because I know of no better option rather than that being "the right thing to do". Because it's not reasonable to be expected to do minimalistic changes for long periods of time. So NMUs can solve things in the short-term, but between NMUs and "where to go from there" is still a limbo I haven't yet gotten good answers for. There's been a lot of debate on [debian-devel] about this and NMUs are generally one of the answers, but there are situations that don't quite fit any standard situation. Like for instance a maintainer might be MIA but ignoring one particular package for a long period of time, thus the MIA team can't say that the maintainer is really MIA, yet the package isn't getting maintenance, and thus no next logical step to take. That's why I'm suggesting that adding one's self to Uploaders after some number of NMUs seems to make sense. :-/ Again not necessarily right, just "the least worst" next step I can think of. > > As always, thanks for your continued work in Debian. ;-) > > You are welcome. Cheers. If you take the suggestion to do a new version NMU, keep in touch with me and let me know how it works out. [Likewise if you're able figure out "the right path forward", let me know, because I'm likewise in a situation with a package where I need to know what the right solution for this is. ;-)] -- Chris -- Chris Knadle Chris.Knadle@coredump.us
Attachment:
signature.asc
Description: This is a digitally signed message part.