On Sat, Oct 13, 2012 at 03:10:52PM +0000, Thorsten Glaser wrote: > >includes switching from the well-understood set of calls to debhelper > >and dpkg-dev programs to something hand-rolled. Please *don't* do > >this; it's not explicitly required by policy that packagers use these > >programs, but please consider when other developers might need to work > >on your package (NMUs for important fixes, security updates > >etc.).
I'd like to AOL the above. Using "non-standard" build systems might be
very cool for those working with them today, or might address some very
specific need (such the one you mentioned: building on very slow build
systems), but in the long run they induce a potentially severe community
cost. If you one day stop caring about Debian, for whatever reason,
those who come after you will have to spend more effort understanding
what's going on than "needed". The case of NMUs is a degenerate case of
the very same problem: you might be temporarily unavailable and people
might have to fix a serious bug in your package to kick out a release
(which is in fact, exactly what happens during BSPs). Using something
more "standard" in your package helps tremendously those people who,
ultimately, just want to help you out.
A principle I recommend following in Debian is: make yourself as
replaceable as possible, because sooner or later someone will have to
step in your shoes. We might not see it today, but it will have to
happen, eventually.
> Right. On the other hand, all those things you mentioned will not touch
> the areas where debhelper is not used, and I’ve seen worse, such as dh7
> ('%:\n\tdh $@'), cdbs and Manoj’s buildsystem (no offence).
Sure, but two wrongs don't make one right, ... right? :) Manoj's
buildsystem hasn't seem a lot of love among other people who had to
touch his packages, precisely for that reason. All in all, that might
have also reduced the potential contributions by others that those
packages could have received if something more standard were used.
Please, think of the kittens :).
--
Stefano Zacchiroli . . . . . . . z...@upsilon.cc . . . . o . . . o . o
Maître de conférences . . . . . http://upsilon.cc/zack . . . o . . . o o
Debian Project Leader . . . . . . @zack on identi.ca . . o o o . . . o .
« the first rule of tautology club is the first rule of tautology club »
signature.asc
Description: Digital signature
