use libgap for abelian subgroups

[fchapoton]

instead of gap pexpect

CC: @tscrim @kwankyu @slel

Component: group theory

Author: Frédéric Chapoton

Branch/Commit: da3bc66

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/34308

[fchapoton]

Branch: u/chapoton/34308

[fchapoton]

Commit: da3bc66

[fchapoton]

New commits:

da3bc66

refactor abelian subgroup with libgap

[fchapoton]

comment:2

green bot, so please review

[tscrim]

comment:3

I recall that len([1 for g in gen]) is faster than sum(1 for g in gen), but it is not as efficient in memory. Is this why you went this route? Do we expect these lists to all be small? I vaguely recall there is a Python count function/method; could that be faster?

[fchapoton]

comment:4

hello Travis,

I am just trying to get rid of the gap pexpect interface here. I am surprised that I finally managed to do that (it must be my third or fourth tentative).

I am not sure it is worth optimising further. The whole abelian group setup is IMHO a complete mess, that would deserve a complete rewriting.

[tscrim]

comment:5

Of course. I was just wondering if there was something specific behind this choice.

[tscrim]

Reviewer: Travis Scrimshaw

[vbraun]

Changed branch from u/chapoton/34308 to da3bc66