Add optimised SmallestMovedPoint implementation

[ChrisJefferson]

This just adds a SmallestMovedPoint C implementation, as I had a case where the time taken mattered.

We leave a GAP-level implementation for permutations which aren't PERM2s or PERM4s, as SmallestMovedPoint is used on permutations which aren't internal rep.

Note (if interested) we don't have a GAP-level implementation for LargestMovedPoint, as there isn't a sensible algorithm for arbitrary permutation implementations (as we don't have a "largest candidate" to start from, as we do with internal permutations).

[coveralls]

Coverage increased (+0.0004%) to 84.507% when pulling ed2cbe1 on ChrisJefferson:smallest-moved-point into 71194d3 on gap-system:master.

[fingolfin]

This is of course overall quite sensible, but I do have some nitpicks ;-)

[fingolfin]

I don't think this adds any useful information, as the 2 line implementation makes this info clear anyway. So I'd just remove it.

Suggested change

** GAP-level wrapper for 'SmallestMovedPointPerm'.

[fingolfin]

The reason for these tests was to make 100% sure we test those kernel function, and not some other installed methods.

I guess these days, I'd move those tests to tst/testinstall/kernel/permutat.tst

[ChrisJefferson]

Thank you both for your comments.they mostly look accurate but I may not be able to address them until Sunday.

[fingolfin]

"Request changes" to mark this issue as not approved, to make sure it doesn't get merged by accident while still waiting for changes by @ChrisJefferson (Chris, this is not meant to bother you into acting, I simply don't want to keep looking at this PR every couple of days, wondering "gee, why is this approved PR not yet merge?", due to my bad memory ;-)

[ChrisJefferson]

Yes, sorry, I'm trying to do no non-required programming until a paper gets finished :)

[wilfwilson]

Got a minor question, I'll be happy to approve once you answer it one way or another 🙂

[ChrisJefferson]

I'll add a test like #3738 which tests SMALLEST_MOVED_POINT_PERM once that PR is merged (I want that one to PR to be merged in first, so it cleanly merges back to 4.11).