Cached permutation groups

[Rob Beezer]

Consider

sage: G = SymmetricGroup(2)
sage: elements = G.list()
sage: elements.remove(G("()"))
sage: K = SymmetricGroup(2)
sage: K.list()
[(1,2)]

Notice K lacks the identity permutation. I'd at least expect K to
be a "new" undisturbed version of S_2.

Is there some overly-aggressive caching here? Should there be some
protection against the above? Or am I missing something?

Using elements = copy(G.list()) causes K to have all of its
elements, but the docstring for list() doesn't say anything about
making a copy, or being an "in-place" version, etc.

Rob

[Robert Bradshaw]

Yes, it's a cached_method returning a mutable object, which is bad.
This is a case where we would be better of return a tuple, and letting
the user create a list out of it if they wanted.

http://trac.sagemath.org/sage_trac/ticket/9155

- Robert

[Rob Beezer]

Robert,

Thanks for the confirmation - I figured that was one of the
possibilities.

Rob

On Jun 5, 9:35 pm, Robert Bradshaw <rober...@math.washington.edu>
wrote: