最近发现Java的UTF转换函数有BUG，在某些手机上（如K700）会出现乱码，与是上网查询了一下，在国外论坛上找到了一个UTF8转换函数，在我们公司所有测试机上都能正常转换，觉得效果还可以
private final String readUnicodeFileUTF8(String filename) {
StringBuffer sb = new StringBuffer(256);
try {
int[] surrogatePair = new int[2];
InputStream is = this.getClass().getResourceAsStream(filename);
int val = 0;
int unicharCount = 0;
while ((val = readNextCharFromStreamUTF8(is))!=-1) {
unicharCount++;
if (val <= 0xFFFF) {
// if first value is the Byte Order Mark (BOM), do not add
if (! (unicharCount == 1 && val == 0xFEFF)) {
sb.append((char)val);
}
} else {
supplementCodePointToSurrogatePair(val, surrogatePair);
sb.append((char)surrogatePair[0]);
sb.append((char)surrogatePair[1]);
}
}
is.close();
} catch (Exception e) {};
return new String(sb);
}
private final static int readNextCharFromStreamUTF8(InputStream is) {
int c = -1;
if (is==null) return c;
boolean complete = false;
try {
int byteVal;
int eXPecting=0;
int composedVal=0;
while (!complete && (byteVal = is.read()) != -1) {
if (expecting > 0 && (byteVal & 0xC0) == 0x80) { /* 10xxxxxx */
expecting--;
composedVal = composedVal ((byteVal & 0x3F) << (expecting*6));
if (expecting == 0) {
c = composedVal;
complete = true;
//System.out.println("appending: U+" + Integer.toHexString(composedVal) );
}
} else {
composedVal = 0;
expecting = 0;
if ((byteVal & 0x80) == 0) { /* 0xxxxxxx */
// one byte character, no extending byte expected
c = byteVal;
complete = true;
//System.out.println("appending: U+" + Integer.toHexString(byteVal) );
} else if ((byteVal & 0xE0) == 0xC0) { /* 110xxxxx */
expecting = 1; // expecting 1 extending byte
composedVal = ((byteVal & 0x1F) << 6);
} else if ((byteVal & 0xF0) == 0xE0) { /* 1110xxxx */
expecting = 2; // expecting 2 extending bytes
composedVal = ((byteVal & 0x0F) << 12);
} else if ((byteVal & 0xF8) == 0xF0) { /* 11110xxx */
expecting = 3; // expecting 3 extending bytes
composedVal = ((byteVal & 0x07) << 18);
} else {
// non conformant utf-8, ignore or catch error
}
}
}
} catch (Exception e) {
System.out.println(e.toString());
}
return c;
}
private final static void supplementCodePointToSurrogatePair(int codePoint, int[] surrogatePair) {
int high4 = ((codePoint >> 16) & 0x1F) - 1;
int mid6 = ((codePoint >> 10) & 0x3F);
int low10 = codePoint & 0x3FF;
surrogatePair[0] = (0xD800 (high4 << 6) (mid6));
surrogatePair[1] = (0xDC00 (low10));
}
（出处：http://www.knowsky.com）
