[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Whose bug(s)? tetex-base, tetex-extra, lilypond

On Tue, Dec 24, 2002 at 11:31:26AM +0100, J?rg Johannes wrote:
> Colin Watson schrieb:
> >This kind of thing is always a bug. If a file moves from one package to
> >another, the destination package needs a versioned Replaces: against the
> >original package. If the file is simply shared, then either it should be
> >removed from one of the two packages or the packages must conflict.
> >
> >You should always make sure that a bug is filed when you encounter this
> >kind of thing. If I were you I wouldn't accept excuses from maintainers
> >about not supporting that particular type of upgrade, either; that tends
> >to be plain laziness. :)
> OK, you're right, but it is not very easy to reproduce such a bug when 
> it appears on a major dist-upgrade changing ~200 packages of you sid 
> system (yes, I do know this kind of thing is considered DANGEROUS ;) ).

Yes, but you can always keep scrollback. :) Also you can often reproduce
it on subsequent dpkg runs too.

> But I would not talk about "plain laziness" for the tetex-maintainer (or 
> any package maintainer). Managing this package must be quite a lot of 
> work, and who am I to say "This guy is lazy" while I just lay myself in 
> the pre-warmed bed of a ready-packaged TeX-installation?

True enough; I tend to get annoyed by it because, for the most part,
moving a file from one package to another is always a manual action, and
it shouldn't be *that* hard to adjust one's habits to include adding
Replaces: lines. I got used to doing that back in the days when I was
moving files around among the groff packages regularly.

Forgetting is one thing, but if you report this kind of bug out of habit
you'll run across maintainers who say "we don't support that kind of
upgrade" (discounting cases that Debian as a whole doesn't really
support, like upgrading directly from hamm to woody). And *that*
infuriates me, because the work for the maintainer once informed of the
problem is generally negligible, while the confusion caused to hundreds
of users is much greater.


Colin Watson                                  [cjwatson@flatline.org.uk]

Reply to: