Option to truncate integers to 64 bits, part 1 (core VM changes)

runtime/vm/intermediate_language.cc

Index: runtime/vm/intermediate_language.cc
diff --git a/runtime/vm/intermediate_language.cc b/runtime/vm/intermediate_language.cc
index 68cf40125c2256810f757cb0aa8dc01ef8ffe21c..43ce48618b512d2ff0461a004c9f56fde026b8f7 100644
--- a/runtime/vm/intermediate_language.cc
+++ b/runtime/vm/intermediate_language.cc
@@ -1818,7 +1818,7 @@ RawInteger* BinaryIntegerOpInstr::Evaluate(const Integer& left,
     if (is_truncating()) {
       int64_t truncated = result.AsTruncatedInt64Value();
       truncated &= RepresentationMask(representation());
-      result = Integer::New(truncated);
+      result = Integer::New(truncated, Heap::kOld);

[siva, 2017/07/07 23:11:20]

Why is it necessary to allocate in old space? the value seems to be
canonicalized below which will ensure it gets to the old space.

[alexmarkov, 2017/07/10 16:58:12]

It is a bug fix. This code is used to do constant folding in the background
compilation thread, which is unable to allocate objects in the young generation
(causes assertion failure).

       ASSERT(IsRepresentable(result, representation()));
     } else if (!IsRepresentable(result, representation())) {
       // If this operation is not truncating it would deoptimize on overflow.