On Tue, Apr 24, 2001 at 11:32:42AM +0200, Arthur Korn wrote: > Richard Braakman schrieb: > > In that case the right "repository" could be a bugreport to the package > > involved. That way the diff submission is guaranteed. > > I agree with you that _something_ has to be done about > catastrophal NMUs, but just stopping to NMU and only submitting > diffs, even on packages where it is clear that the maintainer > stopped working on them some years ago can't be the solution.
Oh, I didn't mean _only_ sending the diff. It was an addendum to Shaleh's suggestion. The idea is to send the diff, then have someone else look at it, and then do the NMU. I don't think we have to appoint any official diff reviewers, it could be enough to say that the person doing the NMU has to be someone other than the person who prepared the diff. That way you make sure that the package is buildable and the NMU is considered sane by at least two people. Ironically, it won't prevent the problem that sparked this thread, namely a weird build environment on the machine where the NMU is compiled. > Why is "abbreviation" such a long word? So that you can practice on it. Richard Braakman
