[Debian-science-sagemath] Status of sagemath?

[E. Madison Bray]

On Tue, Dec 18, 2018 at 1:12 PM Tobias Hansen <thansen at debian.org> wrote:
>
> On 12/18/18 12:34 PM, E. Madison Bray wrote:
> >
> > Regarding the status of porting Sage to use GAP 4.10, I have been
> > working on that, so here is a brief status update:  The port is almost
> > ready from the Sage end of things, but it still required a number of
> > patches to GAP.  I am working with the GAP developers to hopefully
> > release a GAP 4.10.1 containing those patches, hopefully as soon as
> > this week.  If/when that gets done, GAP 4.10.1 should be considered
> > the first release with a truly "usable" libgap API, as what was
> > available in 4.10.0 was not really ready for non-trivial use.
> >
> > If for some reason that release can't be made in time for Debian, the
> > necessary patches can still applied to the Debian package for GAP if
> > necessary.  I'm not sure exactly what the policy is on that, but these
> > are not Sage-specific patches or anything like that, but rather they
> > fix real bugs and shortcomings that made the libgap API not very
> > usable.
> >
> > On the Sage end of things, Sage 8.5 will be released very soon, and
> > will *not* contain support for GAP 4.10.  I will push Volker to try to
> > release a Sage 8.6 soon focused just on the GAP stuff; otherwise it
> > will have to be applied to Sage 8.5 as a patch which I find
> > unfortunate.  I'm going to see if I can push the Sage 8.5 release back
> > but knowing Volker that might be impossible.... :(
>
> Hi,
>
> I followed the ticket. Thanks a lot for making this a priority for us! That was certainly a bunch of work.
>
> I already applied your patches to gap 4.10 and sage 8.4 to do test builds and it works fine. At the moment we are discussing with the gap package maintainer Bill Allombert how we can get the patches to gap available in Debian in time for the freeze. See the bug report [1] which is also in CC. If gap 4.10.1 would have the patches and would come in time that would be great!

Thanks, that's great that you're on top of it.

> Can you comment on whether the writeandcheck.patch is strictly required? We didn't have it before and without it I don't get any additional test failures in sage.

I believe it is necessary but I will double-check.  It's not that
there is specific functionality in Sage that depends on it, but it is
a pretty bad bug that I know can impact Sage.  However, there are some
problems with the current form of the patch as discussed here:
https://github.com/gap-system/gap/pull/3102

I have an idea for a much simpler version of the patch which has fewer
problems that I will try to post shortly.