Add some support for the code-point code-unit distinction.

runtime/vm/unicode.cc

 // Copyright (c) 2011, 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.
 
 #include "vm/unicode.h"
 
 #include "vm/allocation.h"
 #include "vm/globals.h"
 #include "vm/object.h"
 
@@ skipping 47 matching lines @@
 }
 
 
 static bool IsAsciiSequenceStart(uint8_t code_unit) {
   // Check is codepoint is <= U+007F
   return (code_unit <= Utf8::kMaxOneByteChar);
 }
 
 
 static bool IsSmpSequenceStart(uint8_t code_unit) {
-  // Check is codepoint is >= U+10000.
+  // Check the UTF-8 code unit to determine if it is a sequence start for a
+  // code point >= U+10000.
   return (code_unit >= 0xF0);
 }
 
 
-// Returns true if the code point is a high- or low-surrogate.
-static bool IsSurrogate(uint32_t code_point) {
-  return (code_point & 0xfffff800) == 0xd800;
-}
-
-
 // Returns true if the code point value is above Plane 17.
 static bool IsOutOfRange(uint32_t code_point) {
-  return (code_point > 0x10FFFF);
+  return (code_point > Utf16::kMaxCodePoint);
 }
 
 
 // Returns true if the byte sequence is ill-formed.
 static bool IsNonShortestForm(uint32_t code_point, size_t num_bytes) {
   return code_point < kOverlongMinimum[num_bytes];
 }
 
 
-void Utf8::ConvertUTF32ToUTF16(int32_t codepoint, uint16_t* dst) {
-  ASSERT(codepoint > kMaxBmpCodepoint);
-  ASSERT(dst != NULL);
-  dst[0] = (Utf8::kLeadOffset + (codepoint >> 10));
-  dst[1] = (0xDC00 + (codepoint & 0x3FF));
-}
-
-
-// Returns a count of the number of UTF-8 trail bytes.
-intptr_t Utf8::CodePointCount(const uint8_t* utf8_array,
-                              intptr_t array_len,
-                              Type* type) {
+// Returns a count of the number of UTF-16 code units represented by this UTF-8
+// array.  Type is kASCII for 7-bit-only.  If there are surrogate pairs then
+// the type is kSMP.  Otherwise it is kBMP.
+intptr_t Utf8::CodeUnitCount(const uint8_t* utf8_array,
+                             intptr_t array_len,
+                             Type* type) {
   intptr_t len = 0;
   Type char_type = kAscii;
   for (intptr_t i = 0; i < array_len; i++) {
     uint8_t code_unit = utf8_array[i];
     if (!IsTrailByte(code_unit)) {
       ++len;
     }
     if (!IsAsciiSequenceStart(code_unit)) {  // > U+007F
       if (IsSmpSequenceStart(code_unit)) {  // >= U+10000
         char_type = kSMP;
-        ++len;
+        ++len;  // Surrogate pair in the UTF-16 encoding.
       } else if (char_type == kAscii) {
         char_type = kBMP;
       }
     }
   }
   *type = char_type;
   return len;
 }
 
 
@@ skipping 13 matching lines @@
           ch = (ch << 6) + code_unit;
         } else {
           return false;
         }
       }
       ch -= kMagicBits[num_trail_bytes];
       if (!((is_malformed == false) &&
             (j == num_trail_bytes) &&
             !IsOutOfRange(ch) &&
             !IsNonShortestForm(ch, j) &&
-            !IsSurrogate(ch))) {
+            !Utf16::IsSurrogate(ch))) {
         return false;
       }
     }
     i += j;
   }
   return true;
 }
 
 
 intptr_t Utf8::Length(int32_t ch) {
@@ skipping 55 matching lines @@
     }
     Utf8::Encode(ch, &dst[pos]);
     pos += num_bytes;
   }
   return pos;
 }
 
 
 intptr_t Utf8::Decode(const uint8_t* utf8_array,
                       intptr_t array_len,
-                      int32_t* dst) {
+                      uint32_t* dst) {
   uint32_t ch = utf8_array[0] & 0xFF;
   intptr_t i = 1;
   if (ch >= 0x80) {
     int32_t num_trail_bytes = kTrailBytes[ch];
     bool is_malformed = false;
     for (; i < num_trail_bytes; ++i) {
       if (i < array_len) {
         uint8_t code_unit = utf8_array[i];
         is_malformed |= !IsTrailByte(code_unit);
         ch = (ch << 6) + code_unit;
       } else {
-        *dst = -1;
+        *dst = kInvalidCodePoint;
         return 0;
       }
     }
     ch -= kMagicBits[num_trail_bytes];
     if (!((is_malformed == false) &&
           (i == num_trail_bytes) &&
           !IsOutOfRange(ch) &&
           !IsNonShortestForm(ch, i) &&
-          !IsSurrogate(ch))) {
-      *dst = -1;
+          !Utf16::IsSurrogate(ch))) {
+      *dst = kInvalidCodePoint;
       return 0;
     }
   }
   *dst = ch;
   return i;
 }
 
 
 bool Utf8::DecodeToAscii(const uint8_t* utf8_array,
                          intptr_t array_len,
@@ skipping 13 matching lines @@
 
 
 bool Utf8::DecodeToUTF16(const uint8_t* utf8_array,
                          intptr_t array_len,
                          uint16_t* dst,
                          intptr_t len) {
   intptr_t i = 0;
   intptr_t j = 0;
   intptr_t num_bytes;
   for (; (i < array_len) && (j < len); i += num_bytes, ++j) {
-    int32_t ch;
+    uint32_t ch;
     bool is_smp = IsSmpSequenceStart(utf8_array[i]);
     num_bytes = Utf8::Decode(&utf8_array[i], (array_len - i), &ch);
-    if (ch == -1) {
+    if (ch == kInvalidCodePoint) {
       return false;  // invalid input
     }
     if (is_smp) {
-      ConvertUTF32ToUTF16(ch, &(dst[j]));
-      j = j + 1;
+      dst[j] = Utf16::LeadFromCodePoint(ch);
+      dst[j + 1] = Utf16::TrailFromCodePoint(ch);
+      ++j;
     } else {
       dst[j] = ch;
     }
   }
   if ((i < array_len) && (j == len)) {
     return false;  // output overflow
   }
   return true;  // success
 }
 
 
 bool Utf8::DecodeToUTF32(const uint8_t* utf8_array,
                          intptr_t array_len,
                          uint32_t* dst,
                          intptr_t len) {
   intptr_t i = 0;
   intptr_t j = 0;
   intptr_t num_bytes;
   for (; (i < array_len) && (j < len); i += num_bytes, ++j) {
-    int32_t ch;
+    uint32_t ch;
     num_bytes = Utf8::Decode(&utf8_array[i], (array_len - i), &ch);
-    if (ch == -1) {
+    if (ch == kInvalidCodePoint) {
       return false;  // invalid input
     }
     dst[j] = ch;
   }
   if ((i < array_len) && (j == len)) {
     return false;  // output overflow
   }
   return true;  // success
 }
 
 }  // namespace dart