[patch] Code cleanup: Introduce allocate_optimized_out_value

[patch] Code cleanup: Introduce allocate_optimized_out_value

[Jan Kratochvil]

Hi,

this is a code cleanup.  Additionally it makes more optimized_out values lazy.
IIRC I had some compatibility problems before with it but I no longer see any
regressions, maybe since:
	[patch 1/2] Make values more lazy if possible
	http://sourceware.org/ml/gdb-patches/2011-01/msg00123.html
or thanks to some other patch.

allocate_optimized_out_value gets used more by my later entryval patch.

No regressions on {x86_64,x86_64-m32,i686}-fedora15-linux-gnu.
I will check it in in some time if no comments appear.


Thanks,
Jan


gdb/
2011-07-12  Jan Kratochvil  <jan.kratochvil@redhat.com>

	Code cleanup making also optimized out values lazy.
	* dwarf2loc.c (dwarf2_evaluate_loc_desc_full): Use
	allocate_optimized_out_value.  Twice.
	(loclist_read_variable)  Use allocate_optimized_out_value.  Once.
	* findvar.c (read_var_value): Likewise.
	* value.c (allocate_optimized_out_value): New function.
	* value.h (allocate_optimized_out_value): New declaration.

--- a/gdb/dwarf2loc.c
+++ b/gdb/dwarf2loc.c
@@ -1094,12 +1094,7 @@ dwarf2_evaluate_loc_desc_full (struct type *type, struct frame_info *frame,
     invalid_synthetic_pointer ();
 
   if (size == 0)
-    {
-      retval = allocate_value (type);
-      VALUE_LVAL (retval) = not_lval;
-      set_value_optimized_out (retval, 1);
-      return retval;
-    }
+    return allocate_optimized_out_value (type);
 
   baton.frame = frame;
   baton.per_cu = per_cu;
@@ -1247,9 +1242,7 @@ dwarf2_evaluate_loc_desc_full (struct type *type, struct frame_info *frame,
 
 	case DWARF_VALUE_OPTIMIZED_OUT:
 	  do_cleanups (value_chain);
-	  retval = allocate_value (type);
-	  VALUE_LVAL (retval) = not_lval;
-	  set_value_optimized_out (retval, 1);
+	  retval = allocate_optimized_out_value (type);
 	  break;
 
 	  /* DWARF_VALUE_IMPLICIT_POINTER was converted to a pieced
@@ -2829,11 +2822,7 @@ loclist_read_variable (struct symbol *symbol, struct frame_info *frame)
 
   data = dwarf2_find_location_expression (dlbaton, &size, pc);
   if (data == NULL)
-    {
-      val = allocate_value (SYMBOL_TYPE (symbol));
-      VALUE_LVAL (val) = not_lval;
-      set_value_optimized_out (val, 1);
-    }
+    val = allocate_optimized_out_value (SYMBOL_TYPE (symbol));
   else
     val = dwarf2_evaluate_loc_desc (SYMBOL_TYPE (symbol), frame, data, size,
 				    dlbaton->per_cu);
--- a/gdb/findvar.c
+++ b/gdb/findvar.c
@@ -577,10 +577,7 @@ read_var_value (struct symbol *var, struct frame_info *frame)
       break;
 
     case LOC_OPTIMIZED_OUT:
-      v = allocate_value_lazy (type);
-      VALUE_LVAL (v) = not_lval;
-      set_value_optimized_out (v, 1);
-      return v;
+      return allocate_optimized_out_value (type);
 
     default:
       error (_("Cannot look up value of a botched symbol."));
--- a/gdb/value.c
+++ b/gdb/value.c
@@ -728,6 +728,18 @@ allocate_computed_value (struct type *type,
   return v;
 }
 
+/* Allocate NOT_LVAL value for type TYPE being OPTIMIZED_OUT.  */
+
+struct value *
+allocate_optimized_out_value (struct type *type)
+{
+  struct value *retval = allocate_value_lazy (type);
+
+  set_value_optimized_out (retval, 1);
+
+  return retval;
+}
+
 /* Accessor methods.  */
 
 struct value *
--- a/gdb/value.h
+++ b/gdb/value.h
@@ -207,6 +207,8 @@ extern struct value *allocate_computed_value (struct type *type,
                                               struct lval_funcs *funcs,
                                               void *closure);
 
+extern struct value *allocate_optimized_out_value (struct type *type);
+
 /* If VALUE is lval_computed, return its lval_funcs structure.  */
 
 extern struct lval_funcs *value_computed_funcs (struct value *value);

Re: [patch] Code cleanup: Introduce allocate_optimized_out_value

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

Jan> this is a code cleanup.  Additionally it makes more optimized_out values lazy.
Jan> IIRC I had some compatibility problems before with it but I no longer see any
Jan> regressions, maybe since:
Jan> 	[patch 1/2] Make values more lazy if possible
Jan> 	http://sourceware.org/ml/gdb-patches/2011-01/msg00123.html
Jan> or thanks to some other patch.

It looks ok to me.

Tom

Re: [patch] Code cleanup: Introduce allocate_optimized_out_value

[Jan Kratochvil]

On Tue, 12 Jul 2011 23:10:02 +0200, Tom Tromey wrote:
> >>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:
> 
> Jan> this is a code cleanup.  Additionally it makes more optimized_out values lazy.
> Jan> IIRC I had some compatibility problems before with it but I no longer see any
> Jan> regressions, maybe since:
> Jan> 	[patch 1/2] Make values more lazy if possible
> Jan> 	http://sourceware.org/ml/gdb-patches/2011-01/msg00123.html
> Jan> or thanks to some other patch.
> 
> It looks ok to me.

It is true it would be definitely a bug in a code incompatible with such
change so one may rather fix that possible regression there.

Checked it in:
	http://sourceware.org/ml/gdb-cvs/2011-07/msg00124.html


Thanks for the review,
Jan