Minor bug in EllipticCurve's global_minimal_model()

[Benjamin Matschke]

Computing the global minimal model of an elliptic curve over a number field isomorphic to QQ...

sage: K.<t> = NumberField(x-1)
sage: E = EllipticCurve(K,[0,2])
sage: E.global_minimal_model()

... raises an unexpected TypeError:

TypeError: round() takes no arguments (1 given).

(Run on Sage 9.0, Linux Mint 19.2, 64 bit)

It's not dramatic as one can simply work over QQ and coerce, but I think the above code should also work and yield the same result, as sometimes it is convenient to treat QQ as an instance of NumberField.

Thanks,
Benjamin

[John Cremona]

Yes it's a bug.  The rounding is only used when the base field has degree 1, but the test should be that it is actually QQ.  A one-line fix in line 895 of  sage/schemes/elliptic_curves/ell_number_field.py would sort this.