LGTM
http://codereview.chromium.org/503079/diff/1/4 File bleeding_edge/src/codegen.h (right): http://codereview.chromium.org/503079/diff/1/4#newcode311 bleeding_edge/src/codegen.h:311: return OpField::encode(op_) | OverwriteField::encode(overwrite_); Bingo! http://codereview.chromium.org/503079/diff/1/4#newcode319 bleeding_edge/src/codegen.h:319: : "GenericUnaryOpStub_Alloc"; This is annoying in profiles. Please make it produce a more informative name. See the GenericBinaryOpstub for one way to do it. http://codereview.chromium.org/503079/diff/1/3 File bleeding_edge/src/ia32/codegen-ia32.cc (right): http://codereview.chromium.org/503079/diff/1/3#newcode7702 bleeding_edge/src/ia32/codegen-ia32.cc:7702: if (op_ == Token::SUB) { If I were doing this I would divide this function into two. It's too big. http://codereview.chromium.org/503079/diff/1/3#newcode7781 bleeding_edge/src/ia32/codegen-ia32.cc:7781: __ cvtsi2sd(xmm0, Operand(ecx)); This is good, so we should do it for SHR too (search for fstp_d). http://codereview.chromium.org/503079/diff/1/2 File bleeding_edge/test/mjsunit/bit-not.js (right): http://codereview.chromium.org/503079/diff/1/2#newcode45 bleeding_edge/test/mjsunit/bit-not.js:45: testBitNot(100); We should have a loop so we can test the case where the allocation fails. http://codereview.chromium.org/503079 -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
