Re: [patch] fix for c++/2416

[Aleksandar Ristovski <aristovski at qnx dot com>]

Michael Snyder wrote:
> On Wed, 2008-02-27 at 13:03 -0500, Aleksandar Ristovski wrote:
> > Hello,
> >
> > As described in the bug report 2416, the problem is with casting to a reference. 
> > The attached patch should fix this.
> >
> > Testing on head, no change in pass rate encountered (probably suggesting there 
> > is no test case for this - sample code in the bug report should be a good 
> > starting point).
>
> I don't understand the change in unpack_long.  
> You want to treat structs and unions as longs?

Hmm... I thought I put it there to handle cases when we end up with these type 
codes; in that case, I thought, value contains the address and we intepret it as 
a long...

However, now I am having trouble reproducing this case, so maybe I had "covered" 
this scenario while still searching for a more general solution... in any case, 
tests give the same results without unpack_long change and the new diff is attached.

Thanks,

Aleksandar


2008-02-27  Aleksandar Ristovski <aristovski@qnx.com>

	* eval.c (evaluate_subexp_standard): UNOP_CAST use
	value_cast_pointers when casting reference to reference. Print
	error when reference/non-reference mix.
	* value.c (value_as_address): Call coerce_array only on arrays.


Index: gdb/eval.c
===================================================================
RCS file: /cvs/src/src/gdb/eval.c,v
retrieving revision 1.80
diff -u -p -r1.80 eval.c
--- gdb/eval.c	4 Feb 2008 00:23:04 -0000	1.80
+++ gdb/eval.c	27 Feb 2008 19:02:41 -0000
@@ -1985,8 +1985,18 @@ evaluate_subexp_standard (struct type *e
       arg1 = evaluate_subexp (type, exp, pos, noside);
       if (noside == EVAL_SKIP)
 	goto nosideret;
-      if (type != value_type (arg1))
-	arg1 = value_cast (type, arg1);
+      if (type != value_type (arg1)) 
+	{
+	  if (TYPE_CODE (value_type (arg1)) == TYPE_CODE_REF
+	      && TYPE_CODE (type) == TYPE_CODE_REF)
+	    arg1 = value_cast_pointers (type, arg1);
+	  else if (TYPE_CODE (value_type (arg1)) != TYPE_CODE_REF
+		   && TYPE_CODE (type) != TYPE_CODE_REF)
+	    arg1 = value_cast (type, arg1);
+	  else /* We can not do much here.  */
+	    error (_("Attempt to cast to reference type from non-reference "\
+		     "type or vice versa."));
+	}
       return arg1;
 
     case UNOP_MEMVAL:
Index: gdb/value.c
===================================================================
RCS file: /cvs/src/src/gdb/value.c,v
retrieving revision 1.57
diff -u -p -r1.57 value.c
--- gdb/value.c	18 Jan 2008 17:07:40 -0000	1.57
+++ gdb/value.c	27 Feb 2008 19:02:44 -0000
@@ -1039,7 +1039,8 @@ value_as_address (struct value *val)
       || TYPE_CODE (value_type (val)) == TYPE_CODE_METHOD)
     return VALUE_ADDRESS (val);
 
-  val = coerce_array (val);
+  if (TYPE_CODE (value_type (val)) == TYPE_CODE_ARRAY)
+    val = coerce_array (val);
 
   /* Some architectures (e.g. Harvard), map instruction and data
      addresses onto a single large unified address space.  For