diff options
Diffstat (limited to 'deps/v8/src/regexp/interpreter-irregexp.cc')
| -rw-r--r-- | deps/v8/src/regexp/interpreter-irregexp.cc | 60 |
1 files changed, 32 insertions, 28 deletions
diff --git a/deps/v8/src/regexp/interpreter-irregexp.cc b/deps/v8/src/regexp/interpreter-irregexp.cc index ea748e4e55..14834d512a 100644 --- a/deps/v8/src/regexp/interpreter-irregexp.cc +++ b/deps/v8/src/regexp/interpreter-irregexp.cc @@ -4,6 +4,8 @@ // A simple interpreter for the Irregexp byte code. +#ifdef V8_INTERPRETED_REGEXP + #include "src/regexp/interpreter-irregexp.h" #include "src/ast/ast.h" @@ -13,38 +15,32 @@ #include "src/unicode.h" #include "src/utils.h" +#ifdef V8_I18N_SUPPORT +#include "unicode/uchar.h" +#endif // V8_I18N_SUPPORT + namespace v8 { namespace internal { - typedef unibrow::Mapping<unibrow::Ecma262Canonicalize> Canonicalize; -static bool BackRefMatchesNoCase(Canonicalize* interp_canonicalize, - int from, - int current, - int len, - Vector<const uc16> subject) { - for (int i = 0; i < len; i++) { - unibrow::uchar old_char = subject[from++]; - unibrow::uchar new_char = subject[current++]; - if (old_char == new_char) continue; - unibrow::uchar old_string[1] = { old_char }; - unibrow::uchar new_string[1] = { new_char }; - interp_canonicalize->get(old_char, '\0', old_string); - interp_canonicalize->get(new_char, '\0', new_string); - if (old_string[0] != new_string[0]) { - return false; - } - } - return true; +static bool BackRefMatchesNoCase(Isolate* isolate, int from, int current, + int len, Vector<const uc16> subject, + bool unicode) { + Address offset_a = + reinterpret_cast<Address>(const_cast<uc16*>(&subject.at(from))); + Address offset_b = + reinterpret_cast<Address>(const_cast<uc16*>(&subject.at(current))); + size_t length = len * kUC16Size; + return RegExpMacroAssembler::CaseInsensitiveCompareUC16( + offset_a, offset_b, length, unicode ? nullptr : isolate) == 1; } -static bool BackRefMatchesNoCase(Canonicalize* interp_canonicalize, - int from, - int current, - int len, - Vector<const uint8_t> subject) { +static bool BackRefMatchesNoCase(Isolate* isolate, int from, int current, + int len, Vector<const uint8_t> subject, + bool unicode) { + // For Latin1 characters the unicode flag makes no difference. for (int i = 0; i < len; i++) { unsigned int old_char = subject[from++]; unsigned int new_char = subject[current++]; @@ -522,13 +518,16 @@ static RegExpImpl::IrregexpResult RawMatch(Isolate* isolate, pc += BC_CHECK_NOT_BACK_REF_BACKWARD_LENGTH; break; } + BYTECODE(CHECK_NOT_BACK_REF_NO_CASE_UNICODE) BYTECODE(CHECK_NOT_BACK_REF_NO_CASE) { + bool unicode = + (insn & BYTECODE_MASK) == BC_CHECK_NOT_BACK_REF_NO_CASE_UNICODE; int from = registers[insn >> BYTECODE_SHIFT]; int len = registers[(insn >> BYTECODE_SHIFT) + 1] - from; if (from >= 0 && len > 0) { if (current + len > subject.length() || - !BackRefMatchesNoCase(isolate->interp_canonicalize_mapping(), - from, current, len, subject)) { + !BackRefMatchesNoCase(isolate, from, current, len, subject, + unicode)) { pc = code_base + Load32Aligned(pc + 4); break; } @@ -537,13 +536,16 @@ static RegExpImpl::IrregexpResult RawMatch(Isolate* isolate, pc += BC_CHECK_NOT_BACK_REF_NO_CASE_LENGTH; break; } + BYTECODE(CHECK_NOT_BACK_REF_NO_CASE_UNICODE_BACKWARD) BYTECODE(CHECK_NOT_BACK_REF_NO_CASE_BACKWARD) { + bool unicode = (insn & BYTECODE_MASK) == + BC_CHECK_NOT_BACK_REF_NO_CASE_UNICODE_BACKWARD; int from = registers[insn >> BYTECODE_SHIFT]; int len = registers[(insn >> BYTECODE_SHIFT) + 1] - from; if (from >= 0 && len > 0) { if (current - len < 0 || - !BackRefMatchesNoCase(isolate->interp_canonicalize_mapping(), - from, current - len, len, subject)) { + !BackRefMatchesNoCase(isolate, from, current - len, len, subject, + unicode)) { pc = code_base + Load32Aligned(pc + 4); break; } @@ -619,3 +621,5 @@ RegExpImpl::IrregexpResult IrregexpInterpreter::Match( } // namespace internal } // namespace v8 + +#endif // V8_INTERPRETED_REGEXP |