[gfortran] Fix PR 24643

[Toon Moene]

Tobi wrote:

	PR 24643 is a case where we rejected vallid code because we wouldn't accept a
	substring of a aymbol with FL_VARIABLE but BT_UNKNOWN, in other words
	substrings of implicitly typed character variables wouldn't work if the
	variable had been used before.  This patch makes sure that we give them the
	right type before looking for the substring specifier.

	Bubblestrapped and tested, two new testcases attached.  Ok?

	- Tobi

	2005-11-09  Tobias Schl"uter  <tobias.schlueter@physik.uni-muenchen.de>

		PR fortran/24643
		* primary.c (match_varspec): Check for implicitly typed CHARACTER
		variables before matching substrings.

	Index: fortran/primary.c
	===================================================================
	--- fortran/primary.c   (revision 106651)
	+++ fortran/primary.c   (working copy)
	@@ -1634,6 +1634,15 @@ match_varspec (gfc_expr * primary, int e
	     }

	 check_substring:
	+  if (primary->ts.type == BT_UNKNOWN)
	+    {
	+      if (gfc_get_default_type (sym, sym->ns)->type == BT_CHARACTER)
	+       {
	+         gfc_set_default_type (sym, 0, sym->ns);
	+         primary->ts = sym->ts;
	+       }
	+    }
	+
	   if (primary->ts.type == BT_CHARACTER)
	     {
	       switch (match_substring (primary->ts.cl, equiv_flag, &substring))

Yeah, not great (in the sense that I do not think that this is the
place where this decision should be), but It Just Works.

Now all I'm left with compiling the 1.2 million lines of HIRLAM is
non-standard code, that'll weed out tomorrow.

Thanks for looking into this !

Cheers,

-- 
Toon Moene - e-mail: toon@moene.indiv.nluug.nl - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/