Petter Reinholdtsen wrote: > But I believe it is the maintainers responsibility to make sure his > packages do enter testing, and if he fail to do so for _300_ days, the > package is not well maintained and can probably be removed. > > If there is a problem with one of the depends, he should assist the > maintainer of the problematic package, or drop the depend on the > package. There are several options available for a maintainer with a > package stuck in unstable, and waiting until the problem fixes itself > is not the best of them. Well let's see. I have some packages depending on the new perl, which has been blocked from testing for the past half a year or so. I have done a lot to try to get the new perl in -- exhaustive researching of its dependencies, mails to debian-devel, debian-perl, its maintainer, the release manager. Dozens of NMUs of packages which were blocking it from testing. At one point or another, it has been almost ready to go in, except for things blocking *it*, like glibc. I think I've done some fairly helpful work on a package that I don't even maintain after all, and it's really annoying and rather unmotivational for someone to call me a bad maintainer because perl is still not in testing. I hadn't considered rewriting my perl modules to not depend on perl. That is an interesting idea. :-P > And there is not very many packages older then 300 days listed on > <URL:http://developer.skolelinux.no/info/cdbygging/distdiff-all.html.gz>. > Most of us are doing a good job. It is just a few problematic > packages. Your list is incomplete, since it seems to count any new release of the package to unstable as something that can reset the counter. If we followed your plan, we would just get a lot of archive churn with changelog entries like this: libterm-slang-perl (0.9.10.1) unstable; urgency=low * Oh er, I haven't uploaded a new release in 299 days, and I see it is about to be removed from unstable because perl is *still* keeping it out of testing. No changes, let's reset that counter! -- Joey Hess <joeyh@debian.org> Fri, 11 Apr 2003 13:11:46 -0400 > > Anyway, common sense should be (and is, by the RM,) applied. > > Of course. Thank goodness common sense will never let your idea happen, indeed. If you're really interested in getting more packages into testing, you could do worse than try to fix the three RC bugs in perl, or apply some positive motivation to get someone else to fix them. -- see shy jo
Attachment:
pgp8Hyypx8541.pgp
Description: PGP signature