Function oneUTF16toUTF32

engine/source/string/unicode.cc:428–478

-----------------------------------------------------------------------------

//-----------------------------------------------------------------------------
const UTF32  oneUTF16toUTF32(const UTF16* codepoint, U32 *unitsWalked)
{
   PROFILE_START(oneUTF16toUTF32);
   U8    expectedType;
   U32   unitCount;
   UTF32 ret = 0;
   UTF16 codeunit1,codeunit2;
   
   codeunit1 = codepoint[0];
   expectedType = surrogateLUT[codeunit1 >> 10];
   switch(expectedType)
   {
      case 0: // simple
         ret = codeunit1;
         unitCount = 1;
         break;
      case 1: // 2 surrogates
         codeunit2 = codepoint[1];
         if( surrogateLUT[codeunit2 >> 10] == 2)
         {
            ret = ((codeunit1 & byteMaskLow10 ) << 10) | (codeunit2 & byteMaskLow10);
            unitCount = 2;
            break;
         }
         // else, did not find a trailing surrogate where we expected one,
         // so fall through to the error
      case 2: // error
         // found a trailing surrogate where we expected a codepoint or leading surrogate.
         // Dump the replacement.
         ret = kReplacementChar;
         unitCount = 1;
         break;
   }

   if(unitsWalked != NULL)
      *unitsWalked = unitCount;

   // codepoints in the surrogate range are illegal, and should be replaced.
   if(isSurrogateRange(ret))
      ret = kReplacementChar;

   // codepoints outside the Basic Multilingual Plane add complexity to our UTF16 string classes,
   // we've read them correctly so they wont foul the byte stream,
   // but we kill them here to make sure they wont foul anything else
   // NOTE: these are perfectly legal codepoints, we just dont want to deal with them.
   if(isAboveBMP(ret))
      ret = kReplacementChar;

   PROFILE_END();
   return ret;
}

//-----------------------------------------------------------------------------
const UTF16 oneUTF32toUTF16(const UTF32 codepoint)