On Thu, Mar 27, 2003 at 09:37:46AM +0100, Thomas Viehmann wrote: [snip] > IMHO Debian should develop a way of > a) systematically pushing bug fixes even for non-rc bugs i.e. maybe have > a list of bugs tagged "patch" and a policy that NMU's are admissible > if a maintainer has gone X week without any comment on a patch, even > when it's not rc Um, NMU's already *are* permitted for non-RC bugs. I've NMU'd fixes for wishlist bugs before. It's just not encouraged, so that people's precious time is spent more on the more important RC bugs. Any NMU's are OK as long as proper etiquette is followed (as described in the developers' reference). Of course, it's another story if the maintainer rejects an NMU... but in that case, it's presumably because there's a good reason to do so, and the fact that the maintainer responded means he's not MIA and probably is taking proper care of the package. > b) collecting the work applicants do on bugs or accept that fixing bugs > is not something non-maintainers will be helping out with. While it'd be nice to have the BTS keep a per-person list of who submitted what patch, the applicant *could* just keep a list of bugs he helped with, y'know. The BTS does keep a log of these things; so it's just a matter of keeping track of bug numbers. > For manpages point b) certainly applies. Also, it suffers the same > problem as orphaned packages: It's thoroughly unattractive, which is why > people claiming to maintain packages don't care about them. [snip] It may be unattractive to many, but some people (like yours truly) just like writing manpages. I've submitted several manpages to the BTS, some have been gladly accepted, others have been ignored (mostly due to MIA maintainers, so the package subsequently gets NMU'd). Now having said all that, I do agree that current NMU procedures could do with some massaging. It seems a bit pushing it to require at least 6 weeks to NMU a wishlist fix (2-4 weeks after submitting patch, 2 weeks to confirm maintainer is MIA, 2 weeks in delayed upload queue). Each step in itself is reasonable; you don't want to encourage hasty or careless NMU's, but added together, it seems a bit too tedious. T -- My program has no bugs! Only undocumented features...
Description: PGP signature