Re: Patch: don't generate dead bytecode

[Per Bothner <per at bothner dot com>]

Tom Tromey wrote:

>Today I ran `make class-check' on libgcj.  One of the problems is a
>bug in the front end (previously reported).  All the other problems
>turned out to be dead byte code which was generated.  It turns out we
>generate code for a statement like this:
>
>    if (false) { do something }
>
>This patch removes all that dead bytecode.  It does so by simply not
>generating code along a dead branch.
>
The attached patch seems to work.  Could you see if it works as well as
your patch?
Index: parse.y
===================================================================
RCS file: /cvs/gcc/gcc/gcc/java/parse.y,v
retrieving revision 1.318
diff -u -p -r1.318 parse.y
--- parse.y	2001/10/09 05:40:35	1.318
+++ parse.y	2001/10/10 00:02:47
@@ -15023,11 +15011,22 @@ patch_if_else_statement (node)
       return error_mark_node;
     }
   
+  if (TREE_CODE (expression) == INTEGER_CST)
+    {
+      if (integer_zerop (expression))
+	node = TREE_OPERAND (node, 2);
+      else
+	node = TREE_OPERAND (node, 1);
+      if (CAN_COMPLETE_NORMALLY (node) != can_complete_normally)
+	{
+	  node = build (COMPOUND_EXPR, void_type_node, node, empty_stmt_node);
+	  CAN_COMPLETE_NORMALLY (node) = can_complete_normally;
+	}
+      return node;
+    }
   TREE_TYPE (node) = void_type_node;
   TREE_SIDE_EFFECTS (node) = 1;
-  CAN_COMPLETE_NORMALLY (node)
-    = CAN_COMPLETE_NORMALLY (TREE_OPERAND (node, 1))
-    | CAN_COMPLETE_NORMALLY (TREE_OPERAND (node, 2));
+  CAN_COMPLETE_NORMALLY (node) = can_complete_normally;
   return node;
 }