gmpy 1.01 rc near... anybody wanna test>

Alex Martelli aleax at mail.comcast.net
Mon Nov 7 16:54:33 CET 2005


Paul Rubin <http://phr.cx@NOSPAM.invalid> wrote:

> aleax at mail.comcast.net (Alex Martelli) writes:
> > gmpy users able to download and build from sourceforge's cvs are
> > encouraged to test the current CVS version.
> 
> Oh cool, I wondered whether any gmpy maintenance was still going on.

Only when I get guilt-tripped into working a bit on it;-).  gmpy as it
stands mostly meets my modest personal needs, but I'd still like to keep
it useful to others, too, of course.

> I'll see if I can give the new version a try.

Thanks.  The floordiv and truediv still need tests, and there's a
mysterious report of them not working at all (which I suspect is due to
a cvs update somehow going awry -- I've now added a gmpy._cvsid()
accessor to return the CVS $Id$ string to facilitate checking for such
issues); divm shouldn't leak memory any more, but still fails when
arguments aren't mutually prime (haven't looked into that yet).  I will
be working on these issues presently.

I haven't even looked at what's new in GMP 4.1.4 to see if there's
anything there which I should wrap in gmpy (or use to reimplement some
of gmpy's existing functionality in a more optimized way...).  I'm open
to suggestions for any such minor features or optimizations, although my
main goal for 1.01 is to remove bugs and leaks -- any _major_ new
feature will have to wait for a future 1.1 or whatever...

Current gmpy's biggest single issue right now is that I can't build it
on Windows, having no Windows machine, and even less any GMP 4.1.4 for
Windows...


Alex



More information about the Python-list mailing list