http://codereview.chromium.org/13247/diff/1/4 File src/jsregexp.cc (right):
http://codereview.chromium.org/13247/diff/1/4#newcode966 Line 966: FlattenString(subject); The string is also flattened in the IA32 branch (line 889). You could move this outside the switch. http://codereview.chromium.org/13247/diff/1/4#newcode1721 Line 1721: const int max_char = ascii ? 0x7f : 0xffff; Use String::kMaxAsciiCharCode instead of 0x7f? http://codereview.chromium.org/13247/diff/1/4#newcode1781 Line 1781: macro_assembler->CheckCharacterLT(from, &next_range); How about a CheckCharacterRange(from, to, char_is_in_class)? That would allow the assembler to do the unsigned-range-check-hack? "if (x - from < to - from) ..." http://codereview.chromium.org/13247/diff/1/4#newcode1910 Line 1910: if (quarks[j] > 0x7f) { Use String::kMaxAsciiCharCode http://codereview.chromium.org/13247 --~--~---------~--~----~------------~-------~--~----~ v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev -~----------~----~----~----~------~----~------~--~---
