Is this a bug in Polyhedron class (RDF vs AA)?

[jplab]

Hi everyone,

Is the following behavior normal:

sage: P = polytopes.regular_polygon(5)
sage: a_vertex = P.vertices()[0]
sage: for facet in P.Hrepresentation(): print facet.contains(a_vertex), facet.interior_contains(a_vertex)

True False
True True
True False
True True
True True
sage: P = polytopes.regular_polygon(5, base_ring=RDF)
sage: a_vertex = P.vertices()[0]
sage: for facet in P.Hrepresentation(): print facet.contains(a_vertex), facet.interior_contains(a_vertex)

True True
True True
True True
True True
True True

Doing this containment test in RDF seems to break. I guess that there are many methods that depend on the containment methods that also break then?

[Nathann Cohen]

Hellooooooo,

Is the following behavior normal:

Well... In the first case you work on an exact ring, and in the second case you compare the output of >= and > on an inexact ring.

I do not know if there is something wrong somewhere, but I do not expect float computations to be exact either, so I do not know...

Nathann

[Vincent Delecroix]

Yeap. This is exactly it. You have vertices over an inexact ring and
equations over an inexact ring. You should not expect anything beyond
the precision of the ring. Including subtle difference between < and <=.

sage: 0.9 == 0.3 + 0.3 + 0.3
False
sage: 0.9 > 0.3 + 0.3 + 0.3
True

[jplab]

Hi,

Thank you for your quick answer!

Ok, that makes sense. I have a follow-up question. I will put it on sage-devel as it is related to the implementation of a ticket.

Best,
JP