This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

fix fortran debug/9864

The problem to me appears to be that the FUNCTION_DECL for QUUS
winds up in a local binding level for BAZ, rather than at the
global binding level.  The following appears to fix the problem.

I'm somewhat concerned about the assert at the beginning of
start_decl and finish_decl, but I couldn't trigger it.  Toon,
would you mind running this across a larger set of Fortran
input and see if Something Bad Happens?


	* com.c (ffecom_sym_transform_): Install FFEINFO_whereGLOBAL
	symbols in the global binding level.

Index: f/com.c
RCS file: /cvs/gcc/gcc/gcc/f/com.c,v
retrieving revision 1.202
diff -c -p -d -u -r1.202 com.c
--- f/com.c	8 Jun 2003 14:42:05 -0000	1.202
+++ f/com.c	12 Jun 2003 21:09:09 -0000
@@ -8078,8 +8078,8 @@ ffecom_sym_transform_ (ffesymbol s)
 	  DECL_EXTERNAL (t) = 1;
 	  TREE_PUBLIC (t) = 1;
-	  t = start_decl (t, FALSE);
-	  finish_decl (t, NULL_TREE, FALSE);
+	  t = start_decl (t, TRUE);
+	  finish_decl (t, NULL_TREE, TRUE);
 	  if ((g != NULL)
 	      && ((ffeglobal_type (g) == FFEGLOBAL_typeSUBR)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]