[GiNaC-devel] more patch

Hi! To keep up my sending of, as Christian put it, "new sets of 'final' patches [...] at a rate of two per day or so", here is a new one. For my previous patch I had forgotten that while a user-defined non-commutative object carrying indices should be non-commutative, this is not the case with a matrix that is given indices. Best, Chris