New experimental GP parser release 2.4.1.bill-3

[Bill Allombert <allomber@math.u-bordeaux.fr>]

Hello PARI-dev,

I have made available a new version of my experimental GP parser:
<http://pari.math.u-bordeaux.fr/~bill/pari-2.4.1.bill-3.tar.gz>

This version fix a bug with the handling of real constants.

My current plan is to rewrite the interpretor part as a bytecode
generator and a bytecode interpretor. 

The plan calls for a new PARI object the closure:

A closure C are t_VEC with two components: 
gel(C, 1) is a t_VECSMALL ("bytecode")
gel(C, 2) is a t_VEC      ("data")".

The quirk is that C[2] can contains closures as components, so
this is recursive.

A closure is a perfectly standard GEN object and can be handled
in the same way.

Cheers,
Bill.