Integrate the Irregexp Regular Expression Engine.

runtime/lib/string_patch.dart

Index: runtime/lib/string_patch.dart
diff --git a/runtime/lib/string_patch.dart b/runtime/lib/string_patch.dart
index 3a1fd711b6cd06e4a46f44ad383d6258a8d47a76..a65a6bc26257ffe2311814041b4893e739f348cc 100644
--- a/runtime/lib/string_patch.dart
+++ b/runtime/lib/string_patch.dart
@@ -739,6 +739,27 @@ class _OneByteString extends _StringBase implements String {
     return super.split(pattern);
   }
 
+  // The codeUnitsAt family of functions is inlined to an unchecked
+  // LoadCodeUnitsInstr, therefore `this` must be a valid string, and `index`
+  // must be in bounds.
+  int _oneCodeUnitAt(int index) => _codeUnitsAt(index, 1);
+  int _twoCodeUnitsAt(int index) => _codeUnitsAt(index, 2);
+  int _fourCodeUnitsAt(int index) => _codeUnitsAt(index, 4);
+
+  // Loads up to 4 code units into a single integer.
+  int _codeUnitsAt(int index, int count) {
+    assert(index >= 0);
+    assert(index + count - 1 < this.length);
+    assert(0 < count && count <= 4);
+
+    int codeUnits = 0;
+    for (int i = 0; i < count; i++) {
+      codeUnits |= this.codeUnitAt(index + i) << (i * 8);

[Ivan Posva, 2014/10/31 07:32:37]

This looks pretty endian specific to me. No?

[zerny-google, 2014/10/31 11:54:26]

Well spotted. The irregexp code works despite this because we have no big-endian
systems that use this code-unit packing (mips does not support unaligned reads).

After some discussion with Florian, we have decided to avoid this dart code and
instead make it native and generate IR instructions instead. The endian issue
remains however between the current C++ code and the assembly code. I'll either
fix that or assert that packing is not used on big-endian systems.

+    }
+
+    return codeUnits;
+  }
+
   // All element of 'strings' must be OneByteStrings.
   static _concatAll(List<String> strings, int totalLength) {
     // TODO(srdjan): Improve code below and raise or eliminate the limit.
@@ -1003,6 +1024,23 @@ class _TwoByteString extends _StringBase implements String {
   bool operator ==(Object other) {
     return super == other;
   }
+
+  int _oneCodeUnitAt(int index) => _codeUnitsAt(index, 1);
+  int _twoCodeUnitsAt(int index) => _codeUnitsAt(index, 2);
+
+  // Loads up to 2 code units into a single integer.
+  int _codeUnitsAt(int index, int count) {
+    assert(index >= 0);
+    assert(index + count - 1 < this.length);
+    assert(0 < count && count <= 2);
+
+    int codeUnits = 0;
+    for (int i = 0; i < count; i++) {
+      codeUnits |= this.codeUnitAt(index + i) << (i * 16);

[Ivan Posva, 2014/10/31 07:32:37]

ditto.

+    }
+
+    return codeUnits;
+  }
 }