Combinatorics

[Alex Martelli]

Peter Dobcsanyi wrote:

> Alex Martelli <aleax at aleax.it> wrote:
>> Incidentally, if you do a lot of combinatorics and need reasonable
>> performance, you may want to look at gmpy.sf.net:
> 
> Actually, I do. Thank you for the pointer.

Peter was also kind enough in followup mail to point out that
the gmpy he downloaded doesn't handle GMP 4.* (which changed
a field name in a struct) -- so it seems the code I have in the
sourceforge CVS gmpy.c, i.e., only relevant snippet:

            if(resob) {
#if __GNU_MP__ >= 4
                mpz_set(resob->z, randstate->_mp_seed);
#else
                mpz_set(resob->z, randstate->seed);
#endif
                result = (PyObject*)resob;

didn't make it to the last release -- I apologize and I'll try to make
another release package next week -- meanwhile I suggest that
people who want to try GMPY and have problems with GMP 4 could
try getting the CVS gmpy (it's stable, solid, tested) and build that.
Sorry for the inconvenience, everybody!-(


Alex