LGTM so far, with comments.
In json-parser.h we integrated string hashing into the loop that does the
parsing to speed up json parsing. We still use uc32 characters there. I
think
that part needs to be changed as well.
https://codereview.chromium.org/11593007/diff/1/src/heap-inl.h
File src/heap-inl.h (right):
https://codereview.chromium.org/11593007/diff/1/src/heap-inl.h#newcode105
src/heap-inl.h:105: if (chars == str.length()) return
AllocateAsciiSymbol(str, hash_field);
Maybe we could encapsulate this logic by replacing the if-condition with
something like AllocateInternalSymbolHelper::IsOneByte(str, chars) and
make sure this gets inlined.
https://codereview.chromium.org/11593007/diff/1/src/heap.h
File src/heap.h (right):
https://codereview.chromium.org/11593007/diff/1/src/heap.h#newcode771
src/heap.h:771: MUST_USE_RESULT MaybeObject* AllocateExternalSymbol(
Let's remove this function header, while we are at it. Seems to have no
implementation or any use.
https://codereview.chromium.org/11593007/
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
