Mercurial > projects > dil

diff trunk/src/dil/Unicode.d @ 764:4579e8505d5e

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Fixed unittests and removed dil.File. Fixed Converter.UTF16toUTF8(). Fixed an encode() function in dil.Unicode.
author Aziz K?ksal <aziz.koeksal@gmail.com>
date Sat, 16 Feb 2008 03:28:39 +0100
parents 00f872d949ea
children 5e3ef1b2011c
line wrap: on
line diff
--- a/trunk/src/dil/Unicode.d	Sat Feb 16 01:47:39 2008 +0100
+++ b/trunk/src/dil/Unicode.d	Sat Feb 16 03:28:39 2008 +0100
@@ -157,7 +157,7 @@
   char[6] b = void;
   if (c < 0x80)
     str ~= c;
-  if (c < 0x800)
+  else if (c < 0x800)
   {
     b[0] = 0xC0 | (c >> 6);
     b[1] = 0x80 | (c & 0x3F);
@@ -211,8 +211,7 @@
   if (c < 0x10000)
     str ~= cast(wchar)c;
   else
-  {
-    // Encode with surrogate pair.
+  { // Encode with surrogate pair.
     wchar[2] pair = void;
     c -= 0x10000; // c'
     // higher10bits(c') | 0b1101_10xx_xxxx_xxxx
@@ -225,7 +224,7 @@
 
 /++
   Returns a decoded character from a UTF-16 sequence.
-  In case of an error in the sequence 0xD800 is returned.
+  In case of an error in the sequence ERROR_CHAR is returned.
   Params:
     str = the UTF-16 sequence.
     index = where to start from.
@@ -243,7 +242,7 @@
   {
     wchar c2 = str[index+1];
     if (0xDC00 <= c2 && c2 <= 0xDFFF)
-    {
+    { // Decode surrogate pair.
       // (c - 0xD800) << 10 + 0x10000 ->
       // (c - 0xD800 + 0x40) << 10 ->
       c = (c - 0xD7C0) << 10;
@@ -257,7 +256,7 @@
 
 /++
   Returns a decoded character from a UTF-16 sequence.
-  In case of an error in the sequence 0xD800 is returned.
+  In case of an error in the sequence ERROR_CHAR is returned.
   Params:
     p = start of the UTF-16 sequence.
     end = one past the end of the sequence.