Power series: composition and power projections

[vneiger]

Recently a quasi-linear algorithm for composing power series, and also the related power projections, was announced: https://noshi91.hatenablog.com/entry/2024/03/16/224034

And there is now a more complete description on arXiv: https://arxiv.org/pdf/2404.05177.pdf

The algorithm seems rather easy to implement. At first reading, I do not see reasons to think that it will not be fast in practice. Well, at least in some contexts, since this may well depend on where the coefficients lie.

It would be nice to try this in Flint and see how it compares to the current gr_poly_compose_series.