public static void writeUTF8StringAsJSON(byte[] b, int s, int l, OutputStream os)
throws IOException {
int utfLength = UTF8StringUtil.getUTFLength(b, s);
int position =
s
+ UTF8StringUtil.getNumBytesToStoreLength(
utfLength); // skip 2 bytes containing string size
int maxPosition = position + utfLength;
os.write('"');
while (position < maxPosition) {
char c = UTF8StringUtil.charAt(b, position);
int sz = UTF8StringUtil.charSize(b, position);
switch (c) {
// escape
case '\b':
os.write('\\');
os.write('b');
position += sz;
break;
case '\f':
os.write('\\');
os.write('f');
position += sz;
break;
case '\n':
os.write('\\');
os.write('n');
position += sz;
break;
case '\r':
os.write('\\');
os.write('r');
position += sz;
break;
case '\t':
os.write('\\');
os.write('t');
position += sz;
break;
case '\\':
case '"':
os.write('\\');
default:
switch (sz) {
case 1:
if (c <= (byte) 0x1f || c == (byte) 0x7f) {
// this is to print out "control code points" (single byte UTF-8 representation,
// value up to 0x1f or 0x7f) in the 'uXXXX' format
writeUEscape(os, c);
++position;
sz = 0; // no more processing
}
break;
case 2:
// 2-byte encodings of some code points in modified UTF-8 as described in
// DataInput.html#modified-utf-8
//
// 110xxxxx 10xxxxxx
// U+0000 00000 000000 C0 80
// U+0080 00010 000000 C2 80
// U+009F 00010 011111 C2 9F
switch (b[position]) {
case (byte) 0xc0:
if (b[position + 1] == (byte) 0x80) {
// special treatment for the U+0000 code point as described in
// DataInput.html#modified-utf-8
writeUEscape(os, c);
position += 2;
sz = 0; // no more processing
}
break;
case (byte) 0xc2:
if (b[position + 1] <= (byte) 0x9f) {
// special treatment for the U+0080 to U+009F code points
writeUEscape(os, c);
position += 2;
sz = 0; // no more processing
}
break;
}
break;
}
while (sz > 0) {
os.write(b[position]);
++position;
--sz;
}
break;
}
}
os.write('\"');
}
