WIP: IsFFEMatrix, IsFFEMatrixGroup etc. for MatrixObj

[fingolfin]

This is work in progress towards #3568. It is not yet ready in any form but I don't want it to get lost or forgotten.

[fingolfin]

OK, I guess the IsOrdinaryMatrix was here to avoid confusion about the meaning when applied to an IsLieMatrix (I think for MatrixObj we agreed that we only want "ordinary" MatrixObj; implementing Lie algebra elements, if desired, should be wrappers around ordinary MatrixObj. But I digrees)

So what do we do:

1. ignore this issue and just switch to IsMatrixOrMatrixObj (i.e., what this patch does right now)
2. Introduce a new filter IsOrdinaryMatrixOrMatrixObj and wire that in suitably, and use it here.
3. leave this DeclareOperation at IsOrdinaryMatrix and add a second DeclareOperation invocation, i.e.:

Suggested change

	[IsRing,IsMatrixOrMatrixObj,IsPosInt]);
	[IsRing,IsOrdinaryMatrix,IsPosInt]);
	DeclareOperation("CharacteristicPolynomialMatrixNC",
	[IsRing,IsMatrixOrMatrixObj,IsPosInt]);

[ThomasBreuer]

A test case for matrix groups over finite fields whose elements are matrix objects will be groups of IsGenericMatrixRep or IsPlistMatrixRep objects whose entries are from StandardFF fields.
Note that these field elements are not FFEs, hence the matrices are not in danger to be FFECollColls, but groups of these matrices should behave as FFEMatrixGroups.