Bug: Circular Descent Check in WeylGroups

[nathanwilliams]

The following code breaks:

  WeylGroup(['A',2]).long_element().has_left_descent(1)

[tscrim]

This is because the Weyl group element implements has_descent, which is the standard entry point, and the default implementations do the following call structure:

has_descent -> has_left_descent <-> has_right_descent

since I think most code calls has_descent and gives the shortest path function call path if someone only implements has_left_descent. So the solution would be to explicitly have has_left_descent(i) call has_descent(i, side='left').

[nathanwilliams]

The suggestion is essentially a good one, but fails because CoxeterGroups().example() is in a category where this again will cause an infinite loop. Changing has_right_descent(i) in the way suggested works.

---

New commits:

​5716df5	Fixed according to Travis's suggestions
​4404133	Fixed according to Travis's suggestions, and now actually running I think
​8fe1bb8	Tab fixed

[chapoton]

see #15456

[nathanwilliams]

Oy. Veni, vidi, reliqui.

[chapoton]

I think this really needs to be solved.

Either you go this way, but add a doctest, or you use the other ticket.

[nathanwilliams]

Replying to chapoton:

I think this really needs to be solved.

Either you go this way, but add a doctest, or you use the other ticket.

I had added a test, so could I ask you to be more specific about what you would like?

[git]

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

​ab4999f

more intelligent fix a la stump

[chapoton]

ok, this does not seem to break anything. Let it be