Re: [gentoo-science] [sage-on-gentoo] numpy-1.4.1/scipy-0.8.0

[François Bissey]

> On 09/09/2010 03:47:27 AM, François Bissey wrote:
> > Hi Steve,
> >
> > just to be clear, when you enter the data of the test manually in sage
> > there is no problems?
> > We already have one test like that, and there are a few reports of
> > phantom failing tests on sage-devel so it is possible.
> >
> > Francois
>
> François ,
>
> When the following
>
>              sage: x = polygen(GF(121, 'a'))
>              sage: C = HyperellipticCurve(x^5 + x - 1, x^2 + 2)
>              sage: len(C.points())
>              122
>
> is entered manually I get the expected result. The last line of the sage
> input is where
>
>   sage -t -force_lib
> "devel/sage/sage/schemes/hyperelliptic_curves/hyperelliptic_finite_field.py
> "
>
> fails. I downgraded to
>
> sage-doc-4.5.2-r1
> numpy-1.3.0-r2
> scipy-0.7.2-r1
> sage-4.5.2-r1
>
> and the test passed. I then upgraded to the latest sage-doc, numpy, scipy
> and sage and the test passed, but only the first time. On subsequent tries
> the test has failed.
>
The important bit, I think, is that when you enter the data manually you get 
the expected results. There is quite possibly something wrong with the 
doctests framework. Some of it may come from the port to Gentoo, but
as I said there are other people in the sage community that wonder about 
the same thing.

We probably should enter that as a recorded issue on github, other people
may get it from time to time.

Francois