PermGroupElem and AbstractAlgebra.Perm clash in documentation/functionality

[thofma]

Some of our matrix documentation references AbstractAlgebra.Perm: https://docs.oscar-system.org/stable/AbstractAlgebra/matrix/#Permutation

While the signature mismatch Perm vs PermGroupElem might be forgivable, with Oscar permutations the function does not work:

julia> perm([1,2])*zero_matrix(QQ, 2, 2);
ERROR: MethodError: no method matching *(::PermGroupElem, ::QQMatrix)
[...]

julia> AbstractAlgebra.perm([1, 2]) * zero_matrix(QQ, 2, 2)
[0   0]
[0   0]

Edit, also the docstring contains G = SymmetricGroup(3) and the documentation page many other references to SymmetricGroup.

Reported on slack.

[fingolfin]

I am not sure in how far this is a "bug", there never was a design goal to let Oscar permutations match behavior of AA permutations in any way.

I find it really odd to allow multiplication of permutations by matrices -- and even odder that it then apparently is only supported from one side, not the other. However, if people find this feature useful, I am not at all opposed to implementing it also for PermGroupElem, it should only require a few lines.

However, that won't help with functions like lu and fflu being documented as taking and/or producing AA permutations or symmetric groups. To me that is the real issue here, and I am not quite sure how to deal with that

[thofma]

I think the bug is that a docstring in the Oscar documentation does not work.