factor is broken for polynomials over relative number fields

[lftabera]

I suspect this is a regression, but I do not have and old sage copy right now

sage: K=QQ[I]['x,t']
sage: K.inject_variables()
Defining x, t
sage: f=t^2+1
sage: f.factor()
(t + (-I)) * (t + (I))
sage: K=QQ[sqrt(5),I]['t']  
sage: K.inject_variables()
Defining t
sage: f=t^2+1   
sage: f.factor()
(x + I) * (x - I)
sage: prod(f.factor())==f
False
sage: f.parent()
Univariate Polynomial Ring in t over Number Field in sqrt5 with defining polynomial x^2 - 5 over its base field
sage: prod(f.factor()).parent()
Univariate Polynomial Ring in x over Number Field in sqrt5 with defining polynomial x^2 - 5 over its base field

Factor of uni-variate polynomials in polynomial rings (uni and multivariate) over towers of number fields are broken, since the output variable is always 'x', so it the output is in the wrong ring.

Component: factorization

Keywords: factorization, tower of number fields

Author: Francis Clarke

Reviewer: Luis Felipe Tabera Alonso

Merged: sage-4.7.1.alpha3

Issue created by migration from https://trac.sagemath.org/ticket/11218

[sagetrac-fwclarke]

Attachment: trac_11218_factorization_fix.patch.gz

[sagetrac-fwclarke]

comment:1

In the attached patch this defect is corrected very easily and a doctest is altered to demonstrate that the problem is solved.

[sagetrac-fwclarke]

Author: Francis Clarke

[lftabera]

comment:2

The problem has been solved. It was an easy mistake in the code. The doctest shows that the problem has disapeared and the documentation looks god.

Buildbot says that doctest pass.

Positive review.

[jdemeyer]

Reviewer: Luis Felipe Tabera Alonso

[jdemeyer]

Merged: sage-4.7.1.alpha3