Users of cain please raise your voice (Was: Bug#948473: cain: should this package be removed?)
Hi,
when I did my first attempt to port cain I learned that it is really
complex to do and while I gained some ideas what to try next in another
attempt I need to fully agree with Sandro that the maintenance will be
hard.
I hereby explicitly CC Tony from BioLinux (please also spread the word
in the BioLinux community which unfortunately is not rising their
opinion here very frequently) and also Ivo who is listed as other
Uploader. I think if we do not hear from either of you in the next two
weeks I agree with Sandro and a removal seems to be the best idea.
Kind regards
Andreas.
On Thu, Jan 09, 2020 at 12:40:14AM -0500, Sandro Tosi wrote:
> control: retitle -1 keep cain out of testing
>
> On Thu, Jan 9, 2020 at 12:12 AM Andreas Tille <tille@debian.org> wrote:
> >
> > Hi Sandro,
> >
> > On Wed, Jan 08, 2020 at 09:28:16PM -0500, Sandro Tosi wrote:
> > >
> > > * python2-only
> > > * no upstream release since July 2012 (not even advertized on their HP)
> > > * low popcon (but it's also probably part of a niche package)
> >
> > I once gave it a try with 2to3 but failed. I intend to give
> > it another try later.
> >
> > > * one of the only 2 reverse-dependency for python-scipy
> >
> > For the moment I think removing cain from testing and remove
> > Python2 version of python-scipy is OK. Cain definitely should
> > not block progress in scipy.
>
> i'll ask the RT to remove it from testing from the time being, this RC
> bug will prevent it from re-entering.
>
> > > It is my opinion we should remove this package from Debian; if i dont hear back
> > > within a week with a good reason to keep this package around, i'll file for its
> > > removal.
> >
> > If you do not agree with the method above I'm fine with the
> > removal.
>
> i mean, it's fine, but is it maintainable? this is a big piece of code
> that no upstream developer has supported in 7 and a half years. you're
> gonna end up with a gigantic 2to3 patch and then what? what if 3.8
> breaks it, or 3.9? or the next numpy, etc: whos gonna maintain it?
>
> are there any unittests to guarantee the 2to3 conversion is ok? from a
> quick look at the package it looks like there are exactly 0 tests, not
> comforting :( https://salsa.debian.org/med-team/cain/blob/master/debian/rules#L41
>
> the ultimate question is: is it worth to keep this software in debian
> since there's nobody outside willing to commit time and effort to
> maintain and evolve it? are you, alone, gonna do it?
>
> i dont think it's a sustainable approach to just 2to3 and live to see
> another day: this py2removal is also a good opportunity to get rid of
> old/unused/unmaintained packages; to my eyes src:cain look like one of
> them
>
> cheers,
> --
> Sandro "morph" Tosi
> My website: http://sandrotosi.me/
> Me at Debian: http://wiki.debian.org/SandroTosi
> Twitter: https://twitter.com/sandrotosi
>
--
http://fam-tille.de
Reply to: