http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48820
--- Comment #8 from Tobias Burnus <burnus at gcc dot gnu.org> 2012-03-01
19:07:28 UTC ---
(In reply to comment #7)
> Created attachment 26799 [details]
> Patch for TYPE(*)
Just a warning to testers: I have the impression that for scalar allocatables
and pointers, the wrong pointer address is passed (the one of the pointer
instead of the pointer target.) Also, shape/size but not lbound/ubound seem to
be mishandled and I had some issue with optional and null pointers ("null()",
c_null_ptr, nonallocated/nonassociated [(non)polymorphic {scalar/array}]
allocatable/pointer.)
[And to be really useful, one needs to implement the relaxed constraints of
F2008 and TS29113 regarding BIND(C) and C_LOC. gfortran has currently a rather
strict - some times too strict (even for F2003) checking.]
