LGTM.
I don't like how CopySmiToDoubleElements(), CopyPackedSmiToDoubleElements()
and
CopyObjectToDoubleElements() share what feels like 90% of their code, but
their
differences are such that I don't see an element way to unify them more :-/
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/elements.cc
File src/elements.cc (right):
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/elements.cc#newcode379
src/elements.cc:379: to_end = to_start + copy_size;
I have a suspicion that the Windows compiler might wish for a
static_cast<uint32_t>(copy_size) here (and again below).
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/elements.cc#newcode425
src/elements.cc:425: if (hole_or_object == the_hole) {
Why this change? Is it faster?
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/elements.cc#newcode625
src/elements.cc:625: bool is_packed = FLAG_packed_arrays &&
Another unnecessary FLAG check.
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/objects.cc
File src/objects.cc (right):
https://chromiumcodereview.appspot.com/10538104/diff/5001/src/objects.cc#newcode10026
src/objects.cc:10026: if (FLAG_packed_arrays && IsJSArray()) {
This flag check isn't necessary, as we can't have PACKED elements kinds
without the flag anyway. Same below.
https://chromiumcodereview.appspot.com/10538104/
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
