Fix unicode conversion on Android / Emscripten

This also moves functions in utfsjis.cpp to encoding.cpp.

src/encoding.cpp

@@ -1,5 +1,13 @@
+#include <stdlib.h>
+#include <string.h>
+#include "common.h"
 #include "encoding.h"
-#include "utfsjis.h"
+
+namespace {
+
+#include "s2utbl.h"
+
+} // namespace
 
 class SjisEncoding : public Encoding {
 public:
@@ -16,10 +24,116 @@ public:
 		return sjis_to_unicode(code);
 	}
 
+	char* fromUtf8(const char* str)
+	{
+		unsigned char* src = (unsigned char*)str;
+		unsigned char* dst = (unsigned char*)malloc(strlen(str) + 1);
+		unsigned char* dstp = dst;
+
+		while (*src) {
+			if (*src <= 0x7f) {
+				*dstp++ = *src++;
+				continue;
+			}
+
+			int u;
+			if (*src <= 0xdf) {
+				u = (src[0] & 0x1f) << 6 | (src[1] & 0x3f);
+				src += 2;
+			} else if (*src <= 0xef) {
+				u = (src[0] & 0xf) << 12 | (src[1] & 0x3f) << 6 | (src[2] & 0x3f);
+				src += 3;
+			} else {
+				*dstp++ = '?';
+				do src++; while ((*src & 0xc0) == 0x80);
+				continue;
+			}
+
+			if (u > 0xff60 && u <= 0xff9f) {
+				*dstp++ = u - 0xff60 + 0xa0;
+			} else {
+				int c = unicode_to_sjis(u);
+				if (c) {
+					*dstp++ = c >> 8;
+					*dstp++ = c & 0xff;
+				} else {
+					*dstp++ = '?';
+				}
+			}
+		}
+		*dstp = '\0';
+		return (char*)dst;
+	}
+
+	char* toUtf8(const char* str)
+	{
+		unsigned char* src = (unsigned char*)str;
+		unsigned char* dst = (unsigned char*)malloc(strlen(str) * 3 + 1);
+		unsigned char* dstp = dst;
+
+		while (*src) {
+			if (*src <= 0x7f) {
+				*dstp++ = *src++;
+				continue;
+			}
+
+			int c;
+			if (*src >= 0xa0 && *src <= 0xdf) {
+				c = 0xff60 + *src - 0xa0;
+				src++;
+			} else {
+				c = s2u[*src - 0x80][*(src+1) - 0x40];
+				src += 2;
+			}
+
+			if (c <= 0x7f) {
+				*dstp++ = c;
+			} else if (c <= 0x7ff) {
+				*dstp++ = 0xc0 | c >> 6;
+				*dstp++ = 0x80 | (c & 0x3f);
+			} else {
+				*dstp++ = 0xe0 | c >> 12;
+				*dstp++ = 0x80 | (c >> 6 & 0x3f);
+				*dstp++ = 0x80 | (c & 0x3f);
+			}
+		}
+		*dstp = '\0';
+		return (char*)dst;
+	}
+
 private:
 	bool is_2byte(unsigned char c) {
 		return (0x81 <= c && c <= 0x9f) || 0xe0 <= c;
 	}
+
+	static int unicode_to_sjis(int u) {
+		for (int b1 = 0x80; b1 <= 0xff; b1++) {
+			if (b1 >= 0xa0 && b1 <= 0xdf)
+				continue;
+			for (int b2 = 0x40; b2 <= 0xff; b2++) {
+				if (u == s2u[b1 - 0x80][b2 - 0x40])
+					return b1 << 8 | b2;
+			}
+		}
+		return 0;
+	}
+
+	static uint16 sjis_to_unicode(uint16 code)
+	{
+		// ASCII characters.
+		if (code < 0x80)
+			return code;
+		// 1-byte kana characters.
+		if (code >= 0xa0 && code <= 0xdf)
+			return 0xff60 + code - 0xa0;
+		// Gaiji characters.
+		if (0xeb9f <= code && code <= 0xebfc)
+			return code - 0xeb9f + GAIJI_FIRST;
+		if (0xec40 <= code && code <= 0xec9e)
+			return code - 0xec40 + 94 + GAIJI_FIRST;
+
+		return s2u[(code >> 8) - 0x80][(code & 0xff) - 0x40];
+	}
 };
 
 class Utf8Encoding : public Encoding {
@@ -64,6 +178,16 @@ public:
 		*str = s;
 		return code;
 	}
+
+	char* fromUtf8(const char* s)
+	{
+		return strdup(s);
+	}
+
+	char* toUtf8(const char* s)
+	{
+		return strdup(s);
+	}
 };
 
 int Encoding::mbslen(const unsigned char* s)