[Python-3000] Extension: mpf for GNU MP floating point

[Rob Crowther]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've uploaded the latest code to http://umass.glexia.net/mpf.tar.bz2

Here's a quick rundown of supported functions and operations.

The MPF() constructor accepts a string and an optional keyword
argument, prec, specifying precision (as a Long).

Supported module functions:

mpf_add 	Add two MPF objects
mpf_sub 	Subtract two MPF objects
mpf_div 	Divide two MPF objects
mpf_mul 	Multiply two MPF objects
mpf_sqrt 	Take the square root of an MPF object
mpf_neg 	Get the negative of an MPF object
mpf_abs 	Get the absolute value of an MPF object
mpf_pow 	Raise an MPF object to a power
mpf_ceil 	Round an MPF object to the next highest integer
mpf_floor 	Round an MPF object to the next lower integer
mpf_trunc 	Truncate the decimal portion of an MPF object

Operations supported:
(note that only MPF objects are supported atm)

+ - * / ** abs() and - (negative)

Attributes:

value		A tuple of the form (base, sign, whole, decimal)

Also, it supports a print() representation. No more finagling with
value if you don't want to.

Things to come:

floor divide, support for other python numbers in the number interface

Comments:

This wasn't a case of NIH syndrome. I wrote this extension because
Decimal simply was not fast enough and the builtin floats didn't
provide enough precision for a project. The pre-existing modules were
terrible, didn't compile, etc. Necessity, not NIH syndrome.

Questions:

What features would you like to see?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFG++HqqR5p8HaX4oURAsROAKCfNMxxoa+i0lFWJZPDWH8/lguT5ACfSl7d
eYrrkokoCIjuFmnxTW6f4y4=
=cZ8M
-----END PGP SIGNATURE-----