ANSI to unicode codepage conversions

3 files changed, 71 insertions, 5 deletions

diff --git a/src/rtfparser.cpp b/src/rtfparser.cpp
index cf61496..576441c 100644
--- a/src/rtfparser.cpp
+++ b/src/rtfparser.cpp
@@ -43,6 +43,51 @@
 #include "rtfparser.h"
 #include "internal.h"
 
+const wchar_t kAnsiToUnicode[] = {
+
+	/* Invalid */
+	0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
+	0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
+	0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
+	0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
+
+	/* Low 7 bit, same */
+	0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027,
+	0x0028, 0x0029, 0x002A, 0x002B, 0x002C, 0x002D, 0x002E, 0x002F,
+	0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037,
+	0x0038, 0x0039, 0x003A, 0x003B, 0x003C, 0x003D, 0x003E, 0x003F,
+	0x0040, 0x0042, 0x0043, 0x0044, 0x0045, 0x0046, 0x0047, 0x0048,
+	0x0048, 0x0049, 0x004A, 0x004B, 0x004C, 0x004D, 0x004E, 0x004F,
+	0x0050, 0x0052, 0x0053, 0x0054, 0x0055, 0x0056, 0x0057, 0x0058,
+	0x0058, 0x0059, 0x005A, 0x005B, 0x005C, 0x005D, 0x005E, 0x005F,
+	0x0060, 0x0062, 0x0063, 0x0064, 0x0065, 0x0066, 0x0067, 0x0068,
+	0x0068, 0x0069, 0x006A, 0x006B, 0x006C, 0x006D, 0x006E, 0x006F,
+	0x0070, 0x0072, 0x0073, 0x0074, 0x0075, 0x0076, 0x0077, 0x0078,
+	0x0078, 0x0079, 0x007A, 0x007B, 0x007C, 0x007D, 0x007E, 0x007F,
+
+	/* Changes in the middle here to random chars */
+	0x20AC, 0x0081, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021,
+	0x02C6,	0x2030, 0x0160, 0x2039, 0x0152, 0x008D, 0x017D, 0x008F,
+	0x0090, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014,
+	0x02DC, 0x2122, 0x0161, 0x203A, 0x0153, 0x009D, 0x017E, 0x0178,
+
+	/* High stuff, all the same */
+	0x00A0, 0x00A2, 0x00A3, 0x00A4, 0x00A5, 0x00A6, 0x00A7, 0x00A8,
+	0x00A8, 0x00A9, 0x00AA, 0x00AB, 0x00AC, 0x00AD, 0x00AE, 0x00AF,
+	0x00B0, 0x00B2, 0x00B3, 0x00B4, 0x00B5, 0x00B6, 0x00B7, 0x00B8,
+	0x00B8, 0x00B9, 0x00BA, 0x00BB, 0x00BC, 0x00BD, 0x00BE, 0x00BF,
+	0x00C0, 0x00C2, 0x00C3, 0x00C4, 0x00C5, 0x00C6, 0x00C7, 0x00C8,
+	0x00C8, 0x00C9, 0x00CA, 0x00CB, 0x00CC, 0x00CD, 0x00CE, 0x00CF,
+	0x00D0, 0x00D2, 0x00D3, 0x00D4, 0x00D5, 0x00D6, 0x00D7, 0x00D8,
+	0x00D8, 0x00D9, 0x00DA, 0x00DB, 0x00DC, 0x00DD, 0x00DE, 0x00DF,
+	0x00E0, 0x00E2, 0x00E3, 0x00E4, 0x00E5, 0x00E6, 0x00E7, 0x00E8,
+	0x00E8, 0x00E9, 0x00EA, 0x00EB, 0x00EC, 0x00ED, 0x00EE, 0x00EF,
+	0x00F0, 0x00F2, 0x00F3, 0x00F4, 0x00F5, 0x00F6, 0x00F7, 0x00F8,
+	0x00F8, 0x00F9, 0x00FA, 0x00FB, 0x00FC, 0x00FD, 0x00FE, 0x00FF
+
+};
+
+
 
 /* ----------------------------------------------------------------------------------
  *  CONSTRUCTION
@@ -232,10 +277,13 @@ void RtfParser::sendControlWord(const string& cw, int flags, int param)
  *  PARSE HELPERS
  */
 
-bool RtfParser::parseHexChar(int num)
+bool RtfParser::parseHexChar(int num, bool ansi)
 {
 	string data;
 
+	// Ansi is only 256 chars long
+	ASSERT(num == 2 || !ansi);
+
     // Get num chars and put them in the string
 	for(int i = 0; i < num; i++)
 	{
@@ -262,7 +310,21 @@ bool RtfParser::parseHexChar(int num)
 		char* end = NULL;
 		int val = strtol(data.c_str(), &end, 16);
 		if(end == data.c_str() + data.size() && m_parseHex)
-			sendData(val);
+		{
+			if(ansi)
+			{
+				ASSERT((sizeof(kAnsiToUnicode) / sizeof(kAnsiToUnicode[0])) == 256);
+
+				if(val < 0 || val >= 256 || !kAnsiToUnicode[val])
+					m_parseErrors.append("invalid ansi char: " + data + "\n");
+				else
+					sendData(kAnsiToUnicode[val]);
+			}
+			else
+			{
+				sendData(val);
+			}
+		}
 		else
 			m_parseErrors.append("invalid hex char: " + data + "\n");
 	}
@@ -308,7 +370,7 @@ bool RtfParser::parseControlWord()
 		// hex spelled out character
 		else if(empty && ch == '\'')
 		{
-			parseHexChar(2);
+			parseHexChar(2, true);
 			break;
 		}
 
diff --git a/src/rtfparser.h b/src/rtfparser.h
index 0090648..fe2d415 100644
--- a/src/rtfparser.h
+++ b/src/rtfparser.h
@@ -93,7 +93,7 @@ private:
 
     // Parse helpers
 	bool parseControlWord();
-	bool parseHexChar(int num);
+	bool parseHexChar(int num, bool ansi);
 
     // Convenience functions for calling the handler
 	void sendControlWord(const string& cw, int flags, int param);
diff --git a/src/usuals.h b/src/usuals.h
index eeb13b1..a822e2c 100644
--- a/src/usuals.h
+++ b/src/usuals.h
@@ -54,7 +54,11 @@
 #endif
 
 #include <stdlib.h>
-#include <assert.h>
+
+#ifndef ASSERT
+	#include <assert.h>
+	#define ASSERT assert
+#endif
 
 #include <wchar.h>
 #include <string>