Move files around to prepare for assimilation into core

[libs/core.git] / source / stringcodec / utf8.cpp

diff --git a/source/stringcodec/utf8.cpp b/source/stringcodec/utf8.cpp
new file mode 100644 (file)
index 0000000..c7e1705
--- /dev/null
+++ b/source/stringcodec/utf8.cpp
@@ -0,0 +1,92 @@
+/* $Id$
+
+This file is part of libmspstrings
+Copyright © 2006-2007 Mikko Rasa
+Distributed under the LGPL
+*/
+
+#include "utf8.h"
+
+using namespace std;
+
+namespace Msp {
+namespace Codecs {
+
+void Utf8::Encoder::encode_char(UnicodeChar ch, string &buf)
+{
+       if(ch<0 || ch>0x10FFFF)
+               return error(ch, buf, "Can't express character in UTF-8");
+
+       unsigned bytes = 1;
+       if(ch>0xFFFF)
+               bytes = 4;
+       else if(ch>0x7FF)
+               bytes = 3;
+       else if(ch>0x7F)
+               bytes = 2;
+
+       if(bytes==1)
+               buf += ch;
+       else
+       {
+               char utf[4];
+
+               utf[0] = 0xFF<<(8-bytes) | ch>>(bytes*6-6);
+               for(unsigned j=bytes-1; j>0; --j)
+               {
+                       utf[j] = 0x80 | (ch&0x3F);
+                       ch >>= 6;
+               }
+
+               buf.append(utf, bytes);
+       }
+}
+
+void Utf8::Encoder::transliterate(UnicodeChar, string &buf)
+{
+       buf.append("\357\277\275", 3);  // � U+FFFE Replacement Character
+}
+
+
+UnicodeChar Utf8::Decoder::decode_char(const string &str, string::const_iterator &i)
+{
+       if(i==str.end())
+               return error("No input");
+
+       if((*i&0xC0)==0x80)
+       {
+               UnicodeChar result = error("UTF-8 tail byte found when expecting head");
+               ++i;
+               return result;
+       }
+       else if(*i&0x80)
+       {
+               unsigned bytes = 2;
+               unsigned mask = 0x20;
+               for(; *i&mask; mask>>=1)
+                       ++bytes;
+
+               string::const_iterator j = i;
+
+               UnicodeChar result = (*j++)&(mask-1);
+
+               unsigned k;
+               for(k=1; (k<bytes && j!=str.end() && (*j&0xC0)==0x80); ++k)
+                       result = (result<<6) | ((*j++)&0x3F);
+
+               if(k<bytes)
+                       result = error("Incomplete UTF-8 character");
+               else if(!(result>>(bytes*5-4)) || !(result>>7))
+                       result = error("Denormalized UTF-8 multibyte sequence");
+               else if(result>0x10FFFF || (result>=0xD800 && result<=0xDFFF))
+                       result = error("Invalid Unicode code point");
+
+               i = j;
+               return result;
+       }
+       else
+               return *i++;
+}
+
+} // namespace Codecs
+} // namespace Msp