Revision history [back]

Its basically this bug: http://trac.sagemath.org/ticket/10046

We don't have an arbitrary precision polytopes, so sooner or later numerical errors in floating point numbers will give you the wrong answer.