[issue9025] Non-uniformity in randrange for large arguments.

[Antoine Pitrou]

Antoine Pitrou <pitrou at free.fr> added the comment:

> Some considerations for the PRNG are:
> * equidistribution (for quality)
> * repeatability from the same seed (even in multithreaded environments)

I believe a reasonable (com)promise would be to guarantee repeatability
accross a given set of bugfix releases (for example, accross all 2.6.x
releases). We shouldn't necessarily commit to repeatability accross
feature releases, especially if it conflicts with desireable
opportunities for improvement.

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue9025>
_______________________________________