Work around buggy break in List.sort and optimize constant folding of negative values.

dart/frog/leg/ssa/builder.dart

Index: dart/frog/leg/ssa/builder.dart
diff --git a/dart/frog/leg/ssa/builder.dart b/dart/frog/leg/ssa/builder.dart
index 04deac976cc90c3fc947801210c55d635b4e4caa..cffadd05bb0aadf001ff35a9b91f42ff966648c9 100644
--- a/dart/frog/leg/ssa/builder.dart
+++ b/dart/frog/leg/ssa/builder.dart
@@ -1302,7 +1302,15 @@ class SsaBuilder implements Visitor {
         new HStatic(interceptors.getPrefixOperatorInterceptor(op));
     add(target);
     switch (op.source.stringValue) {
-      case "-": push(new HNegate(target, operand)); break;
+      case "-":

[kasperl, 2012/03/01 06:13:00]

This could possible be dealt with in a faster way by not visiting the
node.receiver when it is a literal (thus not creating the initial HLiteral node
at all). That way we also don't have to use 'is double' and 'is int' to
distinguish. Feel free to throw in a TODO(kasperl) so I don't forget about this.

[ahe, 2012/03/01 06:54:15]

Done.

+        if ((operand is HLiteral) && (operand.value is double)) {
+          stack.add(graph.addNewLiteralDouble(-operand.value));
+        } else if ((operand is HLiteral) && (operand.value is int)) {
+          stack.add(graph.addNewLiteralInt(-operand.value));

[ngeoffray, 2012/02/29 20:36:26]

Why is that not done in the constant folder instead?

[ahe, 2012/02/29 20:41:24]

Because it is really slow at doing it.

[ngeoffray, 2012/03/01 08:15:51]

Why is that?

[floitsch, 2012/03/01 10:32:18]

The problem is not really the constant folding, but then replacing the
instruction with the folded one.
Currently we know in which instruction we are used, but not exactly at which
position. For arrays, which have hundreds of inputs, finding the old instruction
to replace it with the new folded one, boils down to doing a linear search.
The worst case is basically an array like: [-1, -2, -3....].
Then each input is replaced and searched for in a linear way yielding n^2
behavior.
This CL is a short-term fix, but we (kasper and I) have discussed better ways to
replace it.

+        } else {
+          push(new HNegate(target, operand));
+        }
+        break;
       case "~": push(new HBitNot(target, operand)); break;
       default: unreachable();
     }