https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109512

--- Comment #7 from Steve Kargl <kargl at gcc dot gnu.org> ---
(In reply to Jerry DeLisle from comment #6)
> (In reply to Steve Kargl from comment #5)
> > Created attachment 63363 [details]
> > Updated patch
> > 
> > The previous diff ran into issue when %C to reference the current locus,
> > which was NULL.  The new diff uses &c->loc and simply points to the 
> > offending line in the program.
> 
> Just completed regression testing here and all tests good. I will prepare a
> commit and post to list. Since we are in stage 4 I will request release
> manager approval. This one is not too onerous I think.

There's a new diff.  The original patch only handled subroutines.
The new also handles functions.  That is,

function foo(bar)
   implicit none (external)
   foo = bar(1)
end function foo

should issue an error.  It took m3 a few hours to figure out how 
to handle this case because gfortran inserts helper functions
into the namespace, which do not have a symtree and do not have
a locus set.

Reply via email to