On Sat, Aug 20, 2011 at 5:36 AM, Rob Beezer <goo...@beezer.cotse.net> wrote: > I'm trying to compute some eigenspaces for matrices over number > fields, mostly because I can, in order to build some doctests. > Eigenspace routine uses the root_field() method on an irreducible > factor of a characteristic polynomial to create a field containing > eigenvalues. Then the routine tries to grab the generator of that > field, which seems to be non-existent. > > Here's as far as I've gone to isolate the problem. This looks to me > like the sequence of commands and constructions incurred by the > eigenspace routine and the error message is similar, if not > identical. The polynomial is just a random one that I've captured for > reproducible input. > > sage: F.<y> = NumberField(x^3+x+1) > sage: z = polygen(F, 'z') > sage: p = (16/3*y + 3/31)*z^2 + (-y^2 + 31)*z - 1/2*y^2 - 1 > sage: p.is_irreducible() > True > sage: M = NumberField(p, 'w') > sage: M > Number Field in w with defining polynomial > (16/3*y + 3/31)*z^2 + (-y^2 + 31)*z - 1/2*y^2 - 1 > over its base field > sage: M.gen(0) > > <BOOM, lots of PARI error messages> > > A bug? Am I doing something wrong? If not, is there a workaround?
Certainly there is a bug. In fact M does have gens: you can assign g=M.gens() and find that g is a tuple of length 2 whose elements are of the correct type. But somehow *displaying* then leads to the error. This is very odd since there are lots of examples of elements of relative number fields in the doctests.l I was pretty sure that this is being caused by the fact that you define M using a non-monic polynomial. But replacing p by p.monic() still raises an error. Even M.absolute_field() crashes. All this probably works as expected when the coefficients of the polynomials are integral. But you clearly do not want that in your case -- you want to be able to find the eigenvalues of a non-integral matrix! So I cannot see a workaround, sorry. John > > Thanks, > Rob > > -- > To post to this group, send an email to sage-devel@googlegroups.com > To unsubscribe from this group, send an email to > sage-devel+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org > -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
