Added smi check to the stub when called using %_StringCompare.
http://codereview.chromium.org/525115/diff/1/5 File src/ia32/assembler-ia32.h (right): http://codereview.chromium.org/525115/diff/1/5#newcode679 src/ia32/assembler-ia32.h:679: // loop On 2010/01/08 11:33:05, Mads Ager wrote:
Expand comment?
Done. http://codereview.chromium.org/525115/diff/1/2 File src/ia32/codegen-ia32.cc (right): http://codereview.chromium.org/525115/diff/1/2#newcode8450 src/ia32/codegen-ia32.cc:8450: __ and_(ecx, kFlatAsciiString); On 2010/01/08 11:33:05, Mads Ager wrote:
Can't you combine the checks. And the two types, and with the mask,
cmp and
branch?
Not really. As both kStringTag and KSeqStringTag are all 0 bits and kAsciiStringTag is 1 bit neither and nor or can be used. Added kStringTag to the compare for completeness. http://codereview.chromium.org/525115/diff/1/2#newcode9541 src/ia32/codegen-ia32.cc:9541: __ and_(ecx, kStringRepresentationMask | kStringEncodingMask); On 2010/01/08 11:33:05, Mads Ager wrote:
Ditto.
Added missing kIsNotStringMask to mask, otherwise see above. http://codereview.chromium.org/525115/diff/1/3 File src/ia32/codegen-ia32.h (right): http://codereview.chromium.org/525115/diff/1/3#newcode817 src/ia32/codegen-ia32.h:817: // constraints on the registers that cna be passed. On 2010/01/08 11:33:05, Mads Ager wrote:
cna -> can
Done. http://codereview.chromium.org/525115/diff/1/11 File src/v8-counters.h (right): http://codereview.chromium.org/525115/diff/1/11#newcode157 src/v8-counters.h:157: SC(string_yyy, V8.StringYYY) \ On 2010/01/08 11:33:05, Mads Ager wrote:
Remove these?
Done. http://codereview.chromium.org/525115/diff/1/11#newcode161 src/v8-counters.h:161: SC(string_compare_native, V8.StringCompareNative) \ On 2010/01/08 11:33:05, Mads Ager wrote:
Align the '\' on these lines.
Done. http://codereview.chromium.org/525115
-- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
