broken doctests in magma.py for integers in QQ and conversion of number field elements

[mstreng]

A later patch at #12006 does everything this patch does, and more. Now that ticket is closed, so this one can be closed too.

In 4.7.2, I get

        sage -t -only-optional=magma devel/sage/sage/schemes/elliptic_curves/ell_generic.py # 1 doctests failed
        sage -t -only-optional=magma devel/sage/sage/schemes/plane_conics/con_field.py # 1 doctests failed
        sage -t -only-optional=magma devel/sage/sage/interfaces/magma.py # 11 doctests failed
        sage -t -only-optional=magma devel/sage/sage/rings/polynomial/multi_polynomial.pyx # 1 doctests failed

Most of these are because 1 in QQ is now converted to Magma as "1/1" instead of "1". The other one is a simple Sage error introduced in #11580 (x is redefined as a variable in a multivariate polynomial ring and then used again as if this never happened).

Component: interfaces

Keywords: magma interface, rational, relative number field

Reviewer: Marco Streng

Issue created by migration from https://trac.sagemath.org/ticket/12154

[mstreng]

comment:1

See also http://groups.google.com/group/sage-release/browse_thread/thread/dea19a0a4a154740

[mstreng]

comment:2

I'm writing a patch.

Apparently some breaking of magma doctests went unnoticed in both #6329 and #11580.

[mstreng]

comment:3

Another cause of failures is conversion via _repr_, which is the default, but is quite unstable, as variable names are sometimes already initialized in Magma, so I will remove that from relative number field elements.

[mstreng]

Attachment: 12154-magma-doctest.patch.gz

[mstreng]

Changed work issues from change doctests only to none

[mstreng]

Changed keywords from none to magma interface, rational, relative number field

[mstreng]

Author: Marco Streng

[mstreng]

comment:6

The new _magma_init_ in number_field_element.pyx gives me "1 out of 1 hunks FAILED"

[mstreng]

Reviewer: Marco Streng

[mstreng]

Changed author from Marco Streng to none