[pm-h] Ideas for March Meeting?

[Charles Boyd]

I could give a talk about the Perl interface to Pari/GP I've been
working on, if anybody is interested.

The code lives here: https://github.com/FreeMonad/GPP

The main goals of this library are:

(1) To make it easy to write a shell for the GP interpreter in perl

[ link: https://github.com/FreeMonad/GPP/blob/master/bin/gpp.pl ]

(2) To provide a way for Perl scripts to call native libpari
functions, similar to the Math::Pari module but without handling
tricky conversions between Perl and Pari/GP data structures or
overloading Perl operators.

[ link: https://github.com/FreeMonad/GPP/blob/master/examples/ellinit.pl ]

These have already been implemented (for details, see the links), but
I think there is a lot of room for improvement - so feedback and
suggestions would be useful.

One of the specific technical challenges is to solve the problem of
having a reasonable build environment for a Perl module that derives
functionality from a native C library which is simply a wrapper around
a large and complex external library - which may or may not already be
installed on the system (and also may or may not be handled by the
operating system's package management system).

My current solution is to simply assume the user has already compiled
Pari/GP (>= 2.6.0) from source and that libpari and its header files
were installed to the default locations on the file system. This
works, but it is not a safe assumption.

I gave a talk on this last month for Pari/GP developers at Institut de
Mathématiques de Bordeaux:
http://pari.math.u-bordeaux.fr/Events/PARI2013/

The previous talk was focused on how to use this Perl interface to do
number theory.

This time, I would be more interested to discuss the design and
implementation details of how to glue a Perl module to a huge library
like Pari/GP.

- Charles

On Mon, Feb 25, 2013 at 10:20 PM, G. Wade Johnson <gwadej at anomaly.org> wrote:
> Our next meeting is on March 14 (about 2.5 weeks).
>
> Any suggestions or volunteers to present for that meeting?
>
> G. Wade

-- 
2048D/E67424BF (GnuPG)