Case independent back references.

src/interpreter-irregexp.cc

 // Copyright 2008 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 11 matching lines @@
 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 // A simple interpreter for the Irregexp byte code.
 
 
 #include "v8.h"
+#include "unicode.h"
 #include "utils.h"
 #include "ast.h"
 #include "bytecodes-irregexp.h"
 #include "interpreter-irregexp.h"
 
 
 namespace v8 { namespace internal {
 
 
+static unibrow::Mapping<unibrow::Ecma262Canonicalize> canonicalize;
+
+
+static bool BackRefMatchesNoCase(int from,

[Christian Plesner Hansen, 2008/11/25 08:18:45]

If we were concerned about performance we could consider comparing compares
old_char and new_char before canonicalization, as an expected fast case.

+                                 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++];
+    canonicalize.get(old_char, '\0', &old_char);
+    canonicalize.get(new_char, '\0', &new_char);
+    if (old_char != new_char) {
+      return false;
+    }
+  }
+  return true;
+}
+
+
 #ifdef DEBUG
 static void TraceInterpreter(const byte* code_base,
                              const byte* pc,
                              int stack_depth,
                              int current_position,
                              int bytecode_length,
                              const char* bytecode_name) {
   if (FLAG_trace_regexp_bytecodes) {
     PrintF("pc = %02x, sp = %d, current = %d, bc = %s",
             pc - code_base,
@@ skipping 261 matching lines @@
               pc = code_base + Load32(pc + 2);
               break;
             }
           }
           if (i < len) break;
           current += len;
         }
         pc += BC_CHECK_NOT_BACK_REF_LENGTH;
         break;
       }
+      BYTECODE(CHECK_NOT_BACK_REF_NO_CASE) {
+        int from = registers[pc[1]];
+        int len = registers[pc[1] + 1] - from;
+        if (current + len > subject.length()) {
+          pc = code_base + Load32(pc + 2);
+          break;
+        } else {
+          if (BackRefMatchesNoCase(from, current, len, subject)) {
+            pc += BC_CHECK_NOT_BACK_REF_NO_CASE_LENGTH;
+          } else {
+            pc = code_base + Load32(pc + 2);
+          }
+        }
+        break;
+      }
       default:
         UNREACHABLE();
         break;
     }
   }
 }
 
 
 bool IrregexpInterpreter::Match(Handle<ByteArray> code_array,
                                 Handle<String> subject16,
                                 int* registers,
                                 int start_position) {
   ASSERT(StringShape(*subject16).IsTwoByteRepresentation());
   ASSERT(subject16->IsFlat(StringShape(*subject16)));
 
 
   AssertNoAllocation a;
   const byte* code_base = code_array->GetDataStartAddress();
   return RawMatch(code_base,
                   Vector<const uc16>(subject16->GetTwoByteData(),
                                      subject16->length()),
                   registers,
                   start_position);
 }
 
 } }  // namespace v8::internal