Re: math/pari: remove unless updated

On 2018/11/24 07:20, Sebastien Marie wrote:
> On Fri, Nov 23, 2018 at 05:27:25PM -0500, Daniel Dickman wrote:
> >
> >
> > > On Nov 23, 2018, at 2:39 PM, Christian Weisgerber <naddy@mips.inka.de> wrote:
> > >
> > > The math/pari port remains at version 2.1.7 from 2005 (!). In the
> > > meantime, upstream has released a stream of new versions up to
> > > 2.11.0 earlier this year. If there is so little interest in
> > > maintaining the port, we might as well delete it.
> >
> > there were a few attempts to update it but their were some perl dependencies that have never been updated if i remember and so pari could not be updated.
> >
> > i'd rather delete those perl dependencies and update pari than delete pari.
> >
>
> There is a somehow "long" chain of reverse dependencies.
>
> - math/pari
> - math/p5-Math-Pari
> - math/p5-Math-BigInt-Pari
> - security/p5-Crypt-Random
> - security/p5-Crypt-DH
> - security/p5-Crypt-Primes
> - security/p5-Crypt-RSA
> - security/p5-Crypt-OpenPGP
>
>
> I see that math/p5-Math-Pari recently got upstream update (15 Sep 2018: 2.030507)
>
> According to the README of Math-Pari, it means the perl module should fully
> support at least pari-2.3.4:
>
> Integer part and the first 4 digits after the point in the version
> of Math::Pari reflect the latest version of GP/PARI it was tested with.
> (2.01080* still fully supports only 2.1.7, but mostly works with 2.3.*
> too.)
>
> It is a good thing, as latest pari update was stopped at the latest 2.1.*
> version because of incomptabilities with 2.3 branch.

All seem pretty dead to me. I'm not a big fan of bundling outdated pari
into p5-Math-Pari unless really necessary, I'd favour removing these p5-*
things (could re-add them later if this proves to be a problem but
it seems unlikely).