[Debian-med-packaging] mcl 22-282
Joost van Baal-Ilić
joostvb-debian at mdcc.cx
Fri May 5 07:51:32 BST 2023
Hi,
If it's about ocaml-clue within Debian, we could always ask our nice colleagues
@ https://lists.debian.org/debian-ocaml-maint/ if they could do some tests for
us; or help us to get going, couldn't we?
BTW: Andreas: thanks for keeping me in Cc; appreciated!
And FWIW: I myself don't have time to dive deep into this; busy with other work
for Debian.
Bye,
Joost
On Thu, May 04, 2023 at 10:09:18PM +0200, Andreas Tille wrote:
>
> Hi again,
>
> I would really appreciate if you would contact pplacer authors in an issue. If we will be able to build pplacer again we can run the unit test we have developed before.
>
> Kind regards
> Andreas.
> --
> http://fam-tille.de
>
>
>
> 4. Mai 2023, 18:18 von micans at gmail.com:
>
> > hi Andreas,
> >
> > Thanks for adding context. Interesting to learn about pplacer and covid, and I appreciate and value the mission.
> > pplacer originates from the same place as > https://github.com/fhcrc/mcl> (Fred Hutchinson -- > https://matsen.fhcrc.org/pplacer/> ). From there it links to > https://github.com/matsen/pplacer> , which accepts issues. This is a logical place to enquire; I will do that.
> >
> > Then I (or someone from fhcrc) may be able to port the patch to the new mcl source tree, but I do not have the capability to test it. Ideally someone from Fred Hutchinson can do this, or would you/Debian be able to do that when/if it comes to that?
> >
> > Best wishes,
> > Stijn
> >
> >
> >
> > On Thu, May 4, 2023 at 4:36 PM Andreas Tille <> tillea at tutanota.com> > wrote:
> >
> >> Hi Stijn,
> >>
> >>
> >>
> >> 4. Mai 2023, 14:17 von >> micans at gmail.com>> :
> >>
> >> > The original work apparently is in > >> https://github.com/fhcrc/mcl
> >> > This repo seems very inactive (last activity 9 years ago), and has no issues, and not a great following (7 stars). I understand popularity is not everything of course and that losing functionality is an anti-goal. On the other hand, I have occasionally removed helper/sibling programs from mcl if I felt they were too obscure or wayward, more or less reasoning I'd hear from people if that caused problems.
> >> >
> >> I perfectly understand your point and I confirm I do not feel good to depend from some unmaintained code in a language that nobody in our team is speaking. That's pretty much the worst situation and somehow it would be sensible to say, that we can't maintain it. On the other hand several people of our team worked hard to get pplacer[1] into Debian which was on the list of software really needed for fighting COVID19.
> >>
> >> If we as a team of Debian developers have problems to maintain some code (something we are trained to do) we can assume that users also might have trouble. The Debian Med team was started to help our users out of trouble and thus we would give up on our mission if we would easily drop some software that is of some use.
> >>
> >>
> >> > An example is mcxdeblast - I'll reinstate it upstream in the new release.
> >> > If I recall correctly other packages depend on this ocaml patch. I do wonder how much use this ocaml patch and dependent packages have seen; there is an aspect that developer/maintainer time needs to be weighed against actual usage. I'm slightly alert, having inspected the > >> https://github.com/fhcrc/mcl>> > repository, that it seems a complete desert now.
> >> >
> >>
> >> The fact that nothing happened in the Ocaml patch repository might mean that there was no need to change anything - as long as mcl had no new release this was probably true. So measuring by inactivity is not really a valid means a lot. The fact that issues are disable makes a bit hard to communicate to the author of the patch, thought.
> >>
> >> If you see some chance to reduce the diff of mcl version 22 to version 14 this might help getting back the patch with less effort and might motivate someone to look into it again.
> >>
> >> Kind regards
> >> Andreas.
> >>
> >>
> >> [1] >> https://tracker.debian.org/pkg/pplacer
> >> --
> >> >> http://fam-tille.de
> >>
> >>
> >>
> >> > On Thu, May 4, 2023 at 10:01 AM Stijn van Dongen <> >> micans at gmail.com>> > > wrote:
> >> >
> >> >> Hello,
> >> >>
> >> >> >> I assume the original contributor of that ocaml patch is not available to
> >> >> >> update it.
> >> >> >>
> >> >> >
> >> >> > That's a real shame since we lost other packages where we have put quite
> >> >> > some effort into.
> >> >>
> >> >> To be clear, I have no information about this and I do not know the provenance of that patch. The description has: "This patch has been generated by taking the diff of an mcl clone that has implemented OCaml support and upstream's version of mcl 12-068, stripping out unnecessary changes and refreshing it over the Debian mcl packed version 14-137."
> >> >>
> >> >> I don't know who made that mcl clone, is there any information about this? Is Shayan Doust still involved, might he have some insight?
> >> >> I can see that it patches some bit of ocaml so that it can call mcl by linking to some of mcl's libraries, but then there is a huge chunk of ocaml code (todo with plugins perhaps) that does not seem very specific to enabling mcl linking.
> >> >>
> >> >> However, if it is possible to "to introduce some mcl-with-ocaml 14-137 in addition to recent mcl" by some person X, where X is not me, in *that case* it suggests the problem is to do with changes that happened in mcl, and I might be able to help, Unless "some mcl-with-ocaml 14-137 depends on both 'old mcl' and 'old ocaml'. If "some mcl-with-ocaml 14-137" could be made with 'old mcl' + 'current ocaml' then it sounds like 'new mcl' + 'current ocaml' could perhaps be attempted by Joost and me.
> >> >>
> >> >> I hope I've expressed myself (or my questions) a bit more clearly; I lack the Debian perspective so perhaps some of the above does not make sense.
> >> >> Best wishes,
> >> >> Stijn
> >> >>
> >> >>
> >> >>
> >> >> On Wed, May 3, 2023 at 9:31 PM Andreas Tille <>> >> tillea at tutanota.com>> >> > wrote:
> >> >>
> >> >>> Hi Stijn,
> >> >>>
> >> >>> >
> >> >>> > Am Wed, May 03, 2023 at 01:00:57PM +0100 schrieb Stijn van Dongen:
> >> >>> >
> >> >>> >> what is the current state of things - will mcl be stuck at version 14-137,
> >> >>> >> or may it be ejected from (the current/edge) release of Debian?
> >> >>> >>
> >> >>> >
> >> >>> > I confirm that Debian 12 will be released with mcl 22-282[1]
> >> >>> >
> >> >>> >> I assume the original contributor of that ocaml patch is not available to
> >> >>> >> update it.
> >> >>> >>
> >> >>> >
> >> >>> > That's a real shame since we lost other packages where we have put quite
> >> >>> > some effort into. For Debian 12 it is too late now. If we could be
> >> >>> > *really* sure that there is no chance at all to upgrade the Ocaml patch
> >> >>> > it might make sense to introduce some mcl-with-ocaml 14-137 in addition
> >> >>> > to recent mcl. I did not made up my mind yet - but I would prefer to
> >> >>> > have the Ocaml support working with recent mcl.
> >> >>> >
> >> >>> > Kind regards
> >> >>> > Andreas.
> >> >>> >
> >> >>> > [1] >>> >> https://tracker.debian.org/pkg/mcl
> >> >>> >
> >> >>> >> On Tue, Jan 24, 2023 at 6:29 PM Andreas Tille <>>> >> andreas at an3as.eu>> >>> > wrote:
> >> >>> >>
> >> >>> >> > Am Mon, Jan 23, 2023 at 09:56:58AM +0100 schrieb Joost van Baal-Ilić:
> >> >>> >> > >
> >> >>> >> > > I tried to find out which are the three mcl reverse depends we now lost,
> >> >>> >> > but
> >> >>> >> > > failed. Andreas: do you happen to know which packages these are?
> >> >>> >> >
> >> >>> >> > pplacer, sepp and tipp
> >> >>> >> >
> >> >>> >> > Some other packages of Debian Med test-depend from these packages.
> >> >>> >> > All are relevant also for COVID-19 so it is relevant.
> >> >>> >> >
> >> >>> >> > Kind regards
> >> >>> >> > Andreas.
> >> >>> >> >
> >> >>> >> > --
> >> >>> >> > >>> >> http://fam-tille.de
> >> >>> >> >
> >> >>> >>
> >> >>> >
> >> >>> > --
> >> >>> > >>> >> http://fam-tille.de
> >> >>> >
> >> >>>
> >>
>
More information about the Debian-med-packaging
mailing list