[stdlib] Speculative workaround for ASAN failure

I suspect the fancy _unsafeUpcast implementation, that was supposed to
be more optimizable, was confusing the optimizer, and that is the cause
of ASAN failures we're seeing.  Let's see if this makes ASAN happy
again.

Author: Dave Abrahams

stdlib/public/core/HashedCollections.swift.gyb

@@ -1265,8 +1265,14 @@ func _impossible<T>(_:T.Type) -> T {
 }
 
 func _unsafeUpcast<T, U>(_ x: T, to: U.Type) -> U {
+#if False
+  // disabled temporarily to speculatively work around a supposed optimizer bug
+  // that is triggering ASAN failures.
   _sanityCheck(x is U)
   return x as? U ?? _impossible(U.self)
+#else
+  return x as! U
+#endif
 }
 
 /// Perform a non-bridged upcast that always succeeds.