kFreeBSD’s future [was: systemd .service file conversion]

[Josselin Mouette <joss@debian.org>]

Le samedi 01 juin 2013 à 11:56 +0200, Marc Haber a écrit : 
> On Fri, 31 May 2013 14:08:01 +0200, Josselin Mouette <joss@debian.org>
> >I disagree with this claim. The wheezy release for kfreebsd is a joke,
> >and we should end it with jessie unless there are real users.
> 
> I find the way you're dismissing other people's work utterly
> offensive. Please think before you write thing like that.

How is that “other people’s” work?
I was among those who spent quite some time to fix packages so that they
build on kfreebsd, yet I can’t say I’m pleased with the result,
especially because nobody ever bothered to test anything.

Just look at popcon: 
* kfreebsd-* reports 183 users, 142 of which use stable.
* sparc reports 183 users, 42 of which use stable.

Graphs are very interesting too: compared to other architectures, there
is a peak of kfreebsd installations after the squeeze and wheezy
releases, something that other architectures don’t notice. So not only
kFreeBSD has just as many users as a dying port (that’s supposing the
same share use popcon), but figures prove these users don’t care about
testing and unstable.

Apart from packages which interact closely with the CPU features, such
as interpreters, there isn’t much doubt that a package that works on
amd64 will work on sparc or ia64. This doesn’t hold for kFreeBSD, for
which every package can fail (and often does fail). Unless we have MANY
more unstable users to ensure our packages work, including users for all
subsystems (not just pf), I don’t think we should release such ports.

Cheers,
-- 
 .''`.      Josselin Mouette
: :' :
`. `'
  `-