Dear Tobias, > > OK for the trunk? >
OK. >> >> PS: There are still issues with polymophic coarrays; in particular, >> argument passing [cf. coarray/poly_run_1.f90] and SELECT TYPE still fail in >> various ways. > It is remarkable just how many ways [OOP] in any shape or form can fail! Adding co-arrays can only make it more complicated! As I said to you in an off-list message, I am making progress with expressions involving typebound operators, with or without class objects. With a seriously horrible kludge, I can make Arjen's example of a general PDE solver work correctly. I am doing a pincer movement between resolve.c and trans-expr.c to find the origin of the problem. Many thanks for the patch. Paul
