Standard conformance question

[Paul THOMAS]

Andrew,

> This code is invalid and has been already filed as PR 22571.
> http://gcc.gnu.org/PR22571

I have realised that this and PR22571 are both fixable without adding formal arguments to gsymbols:

Both the dummy and the actual argument are resolvable as needing to be sequence types; the first because it finds itself in a non-contained procedure and the latter because the referenced procedure has no formal argument list, having no explicit interface. 
> 
> Oh and this has nothing to do with allocatable components TR.

I think that Salvatore knew that. ;)

Thanks

Paul