Optimize special cases in utf8 and utf16 where input can be copied to output.

utils/string_encoding/unicode_core.dart

 // Copyright (c) 2012, the Dart project authors.  Please see the AUTHORS file
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 
 #library("unicode_core");
 
 /*
  * Test for presence of bug related to the use of UTF-16 code units for
  * Dart compiled to JS.
  */
@@ skipping 43 matching lines @@
     throw new IllegalArgumentException("length");
   }
 
   int end = length != null ?
       Math.min(codepoints.length, offset + length) :
       codepoints.length;
 
   int encodedLength = 0;
   for (int i = offset; i < end; i++) {
     int value = codepoints[i];
-    if ((value >= 0 && value < UNICODE_UTF16_RESERVED_LO) || 
+    if ((value >= 0 && value < UNICODE_UTF16_RESERVED_LO) ||
         (value > UNICODE_UTF16_RESERVED_HI && value <= UNICODE_PLANE_ONE_MAX)) {
       encodedLength++;
-    } else if (value > UNICODE_PLANE_ONE_MAX && 
+    } else if (value > UNICODE_PLANE_ONE_MAX &&
         value <= UNICODE_VALID_RANGE_MAX) {
       encodedLength += 2;
     } else {
       encodedLength++;
     }
   }
-  
+
   void addReplacementCodepoint(List<int> codepointBuffer, int offset,
       int replacementCodepoint) {
-    if(replacementCodepoint != null) {
+    if (replacementCodepoint != null) {
       codepointBuffer[offset] = replacementCodepoint;
     } else {
       throw new IllegalArgumentException("Invalid encoding");
     }
   }
+
   List<int> codeUnitsBuffer = new List<int>(encodedLength);
   int j = 0;
   for (int i = offset; i < end; i++) {
     int value = codepoints[i];
-    if ((value >= 0 && value < UNICODE_UTF16_RESERVED_LO) || 
+    if ((value >= 0 && value < UNICODE_UTF16_RESERVED_LO) ||
         (value > UNICODE_UTF16_RESERVED_HI && value <= UNICODE_PLANE_ONE_MAX)) {
       codeUnitsBuffer[j++] = value;
-    } else if (value > UNICODE_PLANE_ONE_MAX && 
+    } else if (value > UNICODE_PLANE_ONE_MAX &&
         value <= UNICODE_VALID_RANGE_MAX) {
       int base = value - UNICODE_UTF16_OFFSET;
       codeUnitsBuffer[j++] = UNICODE_UTF16_SURROGATE_UNIT_0_BASE +
           ((base & UNICODE_UTF16_HI_MASK) >> 10);
       codeUnitsBuffer[j++] = UNICODE_UTF16_SURROGATE_UNIT_1_BASE +
           (base & UNICODE_UTF16_LO_MASK);
     } else {
       addReplacementCodepoint(codeUnitsBuffer, j++, replacementCodepoint);
     }
   }
@@ skipping 11 matching lines @@
   }
 
   if (!(length == null || length >= 0)) {
     throw new IllegalArgumentException("length");
   }
 
   int end = length != null ?
       Math.min(utf16CodeUnits.length, offset + length) :
       utf16CodeUnits.length;
 
-  void addReplacementCodepoint(void f(int v), int replacementCodepoint) {
-    if(replacementCodepoint != null) {
-      f(replacementCodepoint);
-    } else {
-      throw new IllegalArgumentException("Invalid encoding");
-    }
-  }
-
-  void apply(void f(int v)) {
+  void decode(void f(int v)) {
     int i = offset;
     // skip the first entry if it is a BOM.
     if (end > 0 && utf16CodeUnits[0] == UNICODE_BOM) {
       i++;
     }
     while (i < end) {
       int value = utf16CodeUnits[i++];
       if (value < 0) {
-        addReplacementCodepoint(f, replacementCodepoint);
+        f(null);
         continue;
       }
-      if (value < UNICODE_UTF16_RESERVED_LO || 
+      if (value < UNICODE_UTF16_RESERVED_LO ||
           (value > UNICODE_UTF16_RESERVED_HI &&
           value <= UNICODE_PLANE_ONE_MAX)) {
         // transfer directly
         f(value);
       } else if (value < UNICODE_UTF16_SURROGATE_UNIT_1_BASE && i < end) {
         // merge surrogate pair
         int nextValue = utf16CodeUnits[i++];
-        if (nextValue >= UNICODE_UTF16_SURROGATE_UNIT_1_BASE && 
+        if (nextValue >= UNICODE_UTF16_SURROGATE_UNIT_1_BASE &&
             nextValue <= UNICODE_UTF16_RESERVED_HI) {
           value = (value - UNICODE_UTF16_SURROGATE_UNIT_0_BASE) << 10;
           value += UNICODE_UTF16_OFFSET +
               (nextValue - UNICODE_UTF16_SURROGATE_UNIT_1_BASE);
           f(value);
         } else {
           if (nextValue >= UNICODE_UTF16_SURROGATE_UNIT_0_BASE &&
              nextValue < UNICODE_UTF16_SURROGATE_UNIT_1_BASE) {
             i--;
           }
-          addReplacementCodepoint(f, replacementCodepoint);
+          f(null);
           continue;
         }
       } else {
-        addReplacementCodepoint(f, replacementCodepoint);
+        f(null);
         continue;
       }
     }
   }
+
+  // First pass through data to 1) size the output buffer and 2) check for 
+  // special case optimization where A) the length stays the same and B)
+  // no special replacement characters are used. If these criteria are met
+  // we can just copy input to the output.
   int codepointBufferLength = 0;
-  apply(void _(int value) {
+  bool hasReplacements = false;
+  decode(void _(int value) {
       codepointBufferLength++;
+      if (value == null) {
+        hasReplacements = true;
+      }
   });
 
+  // If the string calls for replacements, but we the method is called

[Søren Gjesse, 2012/02/06 22:13:29]

Remove 'we' in 'but we the'

[dcarlson, 2012/02/13 16:37:51]

weird typo -- corrected to "when"

+  // with replacementCodepoint explicitly set to null, then throw an exception.
+  if (hasReplacements && replacementCodepoint == null) {
+    throw new IllegalArgumentException("Invalid encoding");
+  }
+
+  int _length = end - offset;
   List<int> codepointBuffer = new List<int>(codepointBufferLength);
-  int i = 0;
-  apply(void _(int value) {
-      codepointBuffer[i++] = value;
-  });
+  if (_length == codepointBufferLength && !hasReplacements) {
+    codepointBuffer.setRange(0, _length, utf16CodeUnits, offset);
+  } else {
+    int i = 0;
+    decode(
+      void _(int value) {
+        if (value != null) {
+          codepointBuffer[i++] = value;
+        } else {
+          codepointBuffer[i++] = replacementCodepoint;
+        }
+      }
+    );
+  }
   return codepointBuffer;
 }