https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78152
--- Comment #5 from Steve Kargl <sgk at troutmask dot apl.washington.edu> ---
On Tue, Nov 01, 2016 at 08:32:46PM +0000, pault at gcc dot gnu.org wrote:
> >
> > Someone needs to submit an interpretation request to J3.
>
>
> If you bypass the error for associate names, does such code run correctly?
The code compiles with the rather trivial patch (watch for
cut-n-paste whitespace corruption).
Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c (revision 241667)
+++ gcc/fortran/resolve.c (working copy)
@@ -14630,7 +14630,9 @@ resolve_symbol (gfc_symbol *sym)
|| (sym->ns->save_all && !sym->attr.automatic)
|| sym->ns->proc_name->attr.flavor == FL_MODULE
|| sym->ns->proc_name->attr.is_main_program
- || sym->attr.function || sym->attr.result || sym->attr.use_assoc))
+ || sym->attr.function || sym->attr.result
+ || sym->attr.use_assoc
+ || sym->attr.associate_var))
{
gfc_error ("Variable %qs at %L is a coarray and is not ALLOCATABLE, SAVE
"
"nor a dummy argument", sym->name, &sym->declared_at);
The original testcase and my altered code are rather trivial
and don't do anything within the ASSOCIATE constructed. I
suspect that it will be get optimized out.
> Also, what do the other brands do?
Fortunately, I use FreeBSD as my operating system, which
unfortunately limits me to gfortran. I posted to c.l.f,
but haven't got much feedback.
