Error with multiplication of points on elliptic curves over Integers(n)

[tornaria]

sage: E = EllipticCurve(Integers(11*19), [1,0])
sage: P = E(93,3)
sage: P + P
(5 : 118 : 1)
sage: 2 * P
...
AttributeError: 'EllipticCurve_generic_with_category' object has no attribute 'base_field'

CC: @yyyyx4

Component: number theory

Author: Gonzalo Tornaría

Branch/Commit: 9bbacfe

Reviewer: Lorenz Panny

Issue created by migration from https://trac.sagemath.org/ticket/34681

[tornaria]

Commit: 6d31aad

[tornaria]

Author: Gonzalo Tornaría

[tornaria]

Branch: u/tornaria/34681

[sagetrac-git]

Changed commit from 6d31aad to 9bbacfe

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

eb5833f	#34681: fix multiplication of points over Integers(n)
9bbacfe	#34681: do not redo multiplication when pari fails

[yyyyx4]

comment:3

Related: #33228

[tornaria]

comment:4

Replying to Lorenz Panny:

Related: #33228

Thanks for the info. Looks quite interesting. However, this fix here seems trivial -- can we get this in? At the very least the first commit.

Also, given that pari ellmul works with integer mod composite (in the sense that it either gives the right answer or gives a Fp_inv error) it seems a pity to fall back to a generic implementation as in #33228.

Same as mentioned in #33228, this was spotted by a student trying to implement ECM. It seems an important application... If point multiplication is implemented via ellmul this means the students might even have a shot at getting good speed if they can figure out how to choose the parameters.

[yyyyx4]

comment:5

Replying to Gonzalo Tornaría:

However, this fix here seems trivial -- can we get this in?

Indeed, why not.

[yyyyx4]

Reviewer: Lorenz Panny

[vbraun]

Changed branch from u/tornaria/34681 to 9bbacfe