Currently, it's quite slow to create an EllipticCurveIsogeny object, because it precomputes a huge amount of information about itself. (This blocks the original request at #7262, for instance.) It seems like it would be easy enough (and generally useful!) to speed up creation of isogeny objects, and have them lazily evaluate the appropriate bits as needed. Then one could unify the multiplication_by_m and multiplication_by_m_isogeny methods, for instance (maybe by deprecating the former?).

That said, I haven't spent much time with the isogeny code -- if there's some obvious reason this is a bad idea, please comment on the ticket.

CC: @JohnCremona @categorie @shumow @yyyyx4

Component: elliptic curves

Issue created by migration from https://trac.sagemath.org/ticket/8014