Testsuite patch for PR fortran/16511

[Billinghurst, David (CALCRTS)]

I now think the current gfortran behaviour is standard conforming.
See the PR for discussion.

OK to commit.

2004-07-14  David Billinghurst (David.Billinghurst@riotinto.com)

	PR fortran/16511
	* gfortran.dg/g77/19990905-0.f:  Add dg-error. Remove xfail
	* gfortran.dg/g77/README: Update

Index: gfortran.dg/g77/README
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/g77/README,v
retrieving revision 1.1
diff -u -r1.1 README
--- gfortran.dg/g77/README      13 Jul 2004 13:27:52 -0000      1.1
+++ gfortran.dg/g77/README      14 Jul 2004 03:38:56 -0000
@@ -74,7 +74,7 @@
 19990525-0.f                   Y
 19990826-1.f                   Y
 19990826-3.f                   Y
-19990905-0.f                   Y  XFAIL PR 16511
+19990905-0.f                   Y  Correctly gives error with gfortran
 19990905-2.f                   Y
 20000412-1.f                   Y
 20000511-1.f                   Y
Index: gfortran.dg/g77/19990905-0.f
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gfortran.dg/g77/19990905-0.f,v
retrieving revision 1.1
diff -u -r1.1 19990905-0.f
--- gfortran.dg/g77/19990905-0.f        13 Jul 2004 07:08:22 -0000      1.1
+++ gfortran.dg/g77/19990905-0.f        14 Jul 2004 03:38:56 -0000
@@ -1,8 +1,11 @@
 c { dg-do compile }
 * =foo0.f in Burley's g77 test suite.
+c
+c Gives error with gfortran.  This is correct.
+c See Fortran 77 5.5.1 and Fortran 95 5.1.2.4.1 & 7.1.6.2
       subroutine sub(a)
       common /info/ iarray(1000)
       equivalence (m,iarray(100)), (n,iarray(200))
-      real a(m,n) ! { dg-bogus "Variable 'm' cannot appear" "Variable 'm' cannot appear" { xfail *-*-* } } PR 16511 
+      real a(m,n) ! { dg-error "Variable 'm' cannot appear" "Variable 'm' cannot appear" }
       a(1,1) = a(2,2)
       end