4486841: UTF-8 decoder should adhere to corrigendum to Unicode 3.0.1

6636317: Optimize UTF-8 coder for ASCII input

Re-write the UTF-8 charset to obey the standard and improve the performance

Reviewed-by: alanb
This commit is contained in:
Xueming Shen 2008-08-22 14:37:46 -07:00
parent cac629a070
commit 630d73eb0a
2 changed files with 720 additions and 395 deletions

View File

@ -25,34 +25,39 @@
package sun.nio.cs; package sun.nio.cs;
import java.nio.Buffer;
import java.nio.ByteBuffer; import java.nio.ByteBuffer;
import java.nio.CharBuffer; import java.nio.CharBuffer;
import java.nio.BufferOverflowException;
import java.nio.BufferUnderflowException;
import java.nio.charset.Charset; import java.nio.charset.Charset;
import java.nio.charset.CharsetDecoder; import java.nio.charset.CharsetDecoder;
import java.nio.charset.CharsetEncoder; import java.nio.charset.CharsetEncoder;
import java.nio.charset.CoderResult; import java.nio.charset.CoderResult;
import java.nio.charset.CharacterCodingException;
import java.nio.charset.MalformedInputException;
import java.nio.charset.UnmappableCharacterException;
/* Legal UTF-8 Byte Sequences
/* *
* # Bits Bit pattern * # Code Points Bits Bit/Byte pattern
* 1 7 0xxxxxxx * 1 7 0xxxxxxx
* 2 11 110xxxxx 10xxxxxx * U+0000..U+007F 00..7F
* 3 16 1110xxxx 10xxxxxx 10xxxxxx *
* 4 21 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx * 2 11 110xxxxx 10xxxxxx
* 5 26 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx * U+0080..U+07FF C2..DF 80..BF
* 6 31 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx *
* 3 16 1110xxxx 10xxxxxx 10xxxxxx
* U+0800..U+0FFF E0 A0..BF 80..BF
* U+1000..U+FFFF E1..EF 80..BF 80..BF
*
* 4 21 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
* U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
* U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
* U+100000..U10FFFF F4 80..8F 80..BF 80..BF
*
* @author Xueming Shen
* @author Martin Buchholz
* *
* UCS-2 uses 1-3, UTF-16 uses 1-4, UCS-4 uses 1-6
*/ */
class UTF_8 extends Unicode class UTF_8 extends Unicode
{ {
public UTF_8() { public UTF_8() {
super("UTF-8", StandardCharsets.aliases_UTF_8); super("UTF-8", StandardCharsets.aliases_UTF_8);
} }
@ -69,304 +74,250 @@ class UTF_8 extends Unicode
return new Encoder(this); return new Encoder(this);
} }
static final void updatePositions(Buffer src, int sp,
Buffer dst, int dp) {
src.position(sp - src.arrayOffset());
dst.position(dp - dst.arrayOffset());
}
private static class Decoder extends CharsetDecoder { private static class Decoder extends CharsetDecoder {
private Decoder(Charset cs) { private Decoder(Charset cs) {
super(cs, 1.0f, 1.0f); super(cs, 1.0f, 1.0f);
} }
private boolean isContinuation(int b) { private static boolean isNotContinuation(int b) {
return ((b & 0xc0) == 0x80); return (b & 0xc0) != 0x80;
} }
private final Surrogate.Generator sgg = new Surrogate.Generator(); // [C2..DF] [80..BF]
private static boolean isMalformed2(int b1, int b2) {
return (b1 & 0x1e) == 0x0 || (b2 & 0xc0) != 0x80;
}
// [E0] [A0..BF] [80..BF]
// [E1..EF] [80..BF] [80..BF]
private static boolean isMalformed3(int b1, int b2, int b3) {
return (b1 == (byte)0xe0 && (b2 & 0xe0) == 0x80) ||
(b2 & 0xc0) != 0x80 || (b3 & 0xc0) != 0x80;
}
// [F0] [90..BF] [80..BF] [80..BF]
// [F1..F3] [80..BF] [80..BF] [80..BF]
// [F4] [80..8F] [80..BF] [80..BF]
// only check 80-be range here, the [0xf0,0x80...] and [0xf4,0x90-...]
// will be checked by Surrogate.neededFor(uc)
private static boolean isMalformed4(int b2, int b3, int b4) {
return (b2 & 0xc0) != 0x80 || (b3 & 0xc0) != 0x80 ||
(b4 & 0xc0) != 0x80;
}
private static CoderResult lookupN(ByteBuffer src, int n)
{
for (int i = 1; i < n; i++) {
if (isNotContinuation(src.get()))
return CoderResult.malformedForLength(i);
}
return CoderResult.malformedForLength(n);
}
private static CoderResult malformedN(ByteBuffer src, int nb) {
switch (nb) {
case 1:
int b1 = src.get();
if ((b1 >> 2) == -2) {
// 5 bytes 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
if (src.remaining() < 4)
return CoderResult.UNDERFLOW;
return lookupN(src, 5);
}
if ((b1 >> 1) == -2) {
// 6 bytes 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
if (src.remaining() < 5)
return CoderResult.UNDERFLOW;
return lookupN(src, 6);
}
return CoderResult.malformedForLength(1);
case 2: // always 1
return CoderResult.malformedForLength(1);
case 3:
b1 = src.get();
int b2 = src.get(); // no need to lookup b3
return CoderResult.malformedForLength(
((b1 == (byte)0xe0 && (b2 & 0xe0) == 0x80) ||
isNotContinuation(b2))?1:2);
case 4: // we don't care the speed here
b1 = src.get() & 0xff;
b2 = src.get() & 0xff;
if (b1 > 0xf4 ||
(b1 == 0xf0 && (b2 < 0x90 || b2 > 0xbf)) ||
(b1 == 0xf4 && (b2 & 0xf0) != 0x80) ||
isNotContinuation(b2))
return CoderResult.malformedForLength(1);
if (isNotContinuation(src.get()))
return CoderResult.malformedForLength(2);
return CoderResult.malformedForLength(3);
default:
assert false;
return null;
}
}
private static CoderResult malformed(ByteBuffer src, int sp,
CharBuffer dst, int dp,
int nb)
{
src.position(sp - src.arrayOffset());
CoderResult cr = malformedN(src, nb);
updatePositions(src, sp, dst, dp);
return cr;
}
private static CoderResult malformed(ByteBuffer src,
int mark, int nb)
{
src.position(mark);
CoderResult cr = malformedN(src, nb);
src.position(mark);
return cr;
}
private static CoderResult xflow(Buffer src, int sp, int sl,
Buffer dst, int dp, int nb) {
updatePositions(src, sp, dst, dp);
return (nb == 0 || sl - sp < nb)
?CoderResult.UNDERFLOW:CoderResult.OVERFLOW;
}
private static CoderResult xflow(Buffer src, int mark, int nb) {
CoderResult cr = (nb == 0 || src.remaining() < (nb - 1))
?CoderResult.UNDERFLOW:CoderResult.OVERFLOW;
src.position(mark);
return cr;
}
private CoderResult decodeArrayLoop(ByteBuffer src, private CoderResult decodeArrayLoop(ByteBuffer src,
CharBuffer dst) CharBuffer dst)
{ {
// This method is optimized for ASCII input.
byte[] sa = src.array(); byte[] sa = src.array();
int sp = src.arrayOffset() + src.position(); int sp = src.arrayOffset() + src.position();
int sl = src.arrayOffset() + src.limit(); int sl = src.arrayOffset() + src.limit();
assert (sp <= sl);
sp = (sp <= sl ? sp : sl);
char[] da = dst.array(); char[] da = dst.array();
int dp = dst.arrayOffset() + dst.position(); int dp = dst.arrayOffset() + dst.position();
int dl = dst.arrayOffset() + dst.limit(); int dl = dst.arrayOffset() + dst.limit();
assert (dp <= dl); int dlASCII = dp + Math.min(sl - sp, dl - dp);
dp = (dp <= dl ? dp : dl);
try { // ASCII only loop
while (sp < sl) { while (dp < dlASCII && sa[sp] >= 0)
int b1 = sa[sp]; da[dp++] = (char)sa[sp++];
int b2, b3;
switch ((b1 >> 4) & 0x0f) {
case 0: case 1: case 2: case 3:
case 4: case 5: case 6: case 7:
// 1 byte, 7 bits: 0xxxxxxx
if (dl - dp < 1)
return CoderResult.OVERFLOW;
da[dp++] = (char)(b1 & 0x7f);
sp++;
continue;
case 12: case 13:
// 2 bytes, 11 bits: 110xxxxx 10xxxxxx
if (sl - sp < 2)
return CoderResult.UNDERFLOW;
if (dl - dp < 1)
return CoderResult.OVERFLOW;
if (!isContinuation(b2 = sa[sp + 1]))
return CoderResult.malformedForLength(1);
da[dp++] = ((char)(((b1 & 0x1f) << 6) |
((b2 & 0x3f) << 0)));
sp += 2;
continue;
case 14:
// 3 bytes, 16 bits: 1110xxxx 10xxxxxx 10xxxxxx
if (sl - sp < 3)
return CoderResult.UNDERFLOW;
if (dl - dp < 1)
return CoderResult.OVERFLOW;
if (!isContinuation(b2 = sa[sp + 1]))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = sa[sp + 2]))
return CoderResult.malformedForLength(2);
da[dp++] = ((char)(((b1 & 0x0f) << 12) |
((b2 & 0x3f) << 06) |
((b3 & 0x3f) << 0)));
sp += 3;
continue;
case 15:
// 4, 5, or 6 bytes
int b4, b5, b6, uc, n;
switch (b1 & 0x0f) {
case 0: case 1: case 2: case 3:
case 4: case 5: case 6: case 7:
// 4 bytes, 21 bits
if (sl - sp < 4)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = sa[sp + 1]))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = sa[sp + 2]))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = sa[sp + 3]))
return CoderResult.malformedForLength(3);
uc = (((b1 & 0x07) << 18) |
((b2 & 0x3f) << 12) |
((b3 & 0x3f) << 06) |
((b4 & 0x3f) << 00));
n = 4;
break;
case 8: case 9: case 10: case 11:
// 5 bytes, 26 bits
if (sl - sp < 5)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = sa[sp + 1]))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = sa[sp + 2]))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = sa[sp + 3]))
return CoderResult.malformedForLength(3);
if (!isContinuation(b5 = sa[sp + 4]))
return CoderResult.malformedForLength(4);
uc = (((b1 & 0x03) << 24) |
((b2 & 0x3f) << 18) |
((b3 & 0x3f) << 12) |
((b4 & 0x3f) << 06) |
((b5 & 0x3f) << 00));
n = 5;
break;
case 12: case 13:
// 6 bytes, 31 bits
if (sl - sp < 6)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = sa[sp + 1]))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = sa[sp + 2]))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = sa[sp + 3]))
return CoderResult.malformedForLength(3);
if (!isContinuation(b5 = sa[sp + 4]))
return CoderResult.malformedForLength(4);
if (!isContinuation(b6 = sa[sp + 5]))
return CoderResult.malformedForLength(5);
uc = (((b1 & 0x01) << 30) |
((b2 & 0x3f) << 24) |
((b3 & 0x3f) << 18) |
((b4 & 0x3f) << 12) |
((b5 & 0x3f) << 06) |
((b6 & 0x3f)));
n = 6;
break;
default:
return CoderResult.malformedForLength(1);
}
int gn = sgg.generate(uc, n, da, dp, dl);
if (gn < 0)
return sgg.error();
dp += gn;
sp += n;
continue;
default:
return CoderResult.malformedForLength(1);
while (sp < sl) {
int b1 = sa[sp];
if (b1 >= 0) {
// 1 byte, 7 bits: 0xxxxxxx
if (dp >= dl)
return xflow(src, sp, sl, dst, dp, 1);
da[dp++] = (char)b1;
sp++;
} else if ((b1 >> 5) == -2) {
// 2 bytes, 11 bits: 110xxxxx 10xxxxxx
if (sl - sp < 2 || dp >= dl)
return xflow(src, sp, sl, dst, dp, 2);
int b2 = sa[sp + 1];
if (isMalformed2(b1, b2))
return malformed(src, sp, dst, dp, 2);
da[dp++] = (char) (((b1 << 6) ^ b2) ^ 0x0f80);
sp += 2;
} else if ((b1 >> 4) == -2) {
// 3 bytes, 16 bits: 1110xxxx 10xxxxxx 10xxxxxx
if (sl - sp < 3 || dp >= dl)
return xflow(src, sp, sl, dst, dp, 3);
int b2 = sa[sp + 1];
int b3 = sa[sp + 2];
if (isMalformed3(b1, b2, b3))
return malformed(src, sp, dst, dp, 3);
da[dp++] = (char) (((b1 << 12) ^ (b2 << 6) ^ b3) ^ 0x1f80);
sp += 3;
} else if ((b1 >> 3) == -2) {
// 4 bytes, 21 bits: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
if (sl - sp < 4 || dl - dp < 2)
return xflow(src, sp, sl, dst, dp, 4);
int b2 = sa[sp + 1];
int b3 = sa[sp + 2];
int b4 = sa[sp + 3];
int uc = ((b1 & 0x07) << 18) |
((b2 & 0x3f) << 12) |
((b3 & 0x3f) << 06) |
(b4 & 0x3f);
if (isMalformed4(b2, b3, b4) ||
!Surrogate.neededFor(uc)) {
return malformed(src, sp, dst, dp, 4);
} }
da[dp++] = Surrogate.high(uc);
} da[dp++] = Surrogate.low(uc);
sp += 4;
return CoderResult.UNDERFLOW; } else
} finally { return malformed(src, sp, dst, dp, 1);
src.position(sp - src.arrayOffset());
dst.position(dp - dst.arrayOffset());
} }
return xflow(src, sp, sl, dst, dp, 0);
} }
private CoderResult decodeBufferLoop(ByteBuffer src, private CoderResult decodeBufferLoop(ByteBuffer src,
CharBuffer dst) CharBuffer dst)
{ {
int mark = src.position(); int mark = src.position();
try { int limit = src.limit();
while (src.hasRemaining()) { while (mark < limit) {
int b1 = src.get(); int b1 = src.get();
int b2, b3; if (b1 >= 0) {
switch ((b1 >> 4) & 0x0f) { // 1 byte, 7 bits: 0xxxxxxx
if (dst.remaining() < 1)
case 0: case 1: case 2: case 3: return xflow(src, mark, 1); //overflow
case 4: case 5: case 6: case 7: dst.put((char)b1);
// 1 byte, 7 bits: 0xxxxxxx mark++;
if (dst.remaining() < 1) } else if ((b1 >> 5) == -2) {
return CoderResult.OVERFLOW; // 2 bytes, 11 bits: 110xxxxx 10xxxxxx
dst.put((char)b1); if (limit - mark < 2|| dst.remaining() < 1)
mark++; return xflow(src, mark, 2);
continue; int b2 = src.get();
if (isMalformed2(b1, b2))
case 12: case 13: return malformed(src, mark, 2);
// 2 bytes, 11 bits: 110xxxxx 10xxxxxx dst.put((char) (((b1 << 6) ^ b2) ^ 0x0f80));
if (src.remaining() < 1) mark += 2;
return CoderResult.UNDERFLOW; } else if ((b1 >> 4) == -2) {
if (dst.remaining() < 1) // 3 bytes, 16 bits: 1110xxxx 10xxxxxx 10xxxxxx
return CoderResult.OVERFLOW; if (limit - mark < 3 || dst.remaining() < 1)
if (!isContinuation(b2 = src.get())) return xflow(src, mark, 3);
return CoderResult.malformedForLength(1); int b2 = src.get();
dst.put((char)(((b1 & 0x1f) << 6) | int b3 = src.get();
((b2 & 0x3f) << 0))); if (isMalformed3(b1, b2, b3))
mark += 2; return malformed(src, mark, 3);
continue; dst.put((char) (((b1 << 12) ^ (b2 << 6) ^ b3) ^ 0x1f80));
mark += 3;
case 14: } else if ((b1 >> 3) == -2) {
// 3 bytes, 16 bits: 1110xxxx 10xxxxxx 10xxxxxx // 4 bytes, 21 bits: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
if (src.remaining() < 2) if (limit - mark < 4 || dst.remaining() < 2)
return CoderResult.UNDERFLOW; return xflow(src, mark, 4);
if (dst.remaining() < 1) int b2 = src.get();
return CoderResult.OVERFLOW; int b3 = src.get();
if (!isContinuation(b2 = src.get())) int b4 = src.get();
return CoderResult.malformedForLength(1); int uc = ((b1 & 0x07) << 18) |
if (!isContinuation(b3 = src.get())) ((b2 & 0x3f) << 12) |
return CoderResult.malformedForLength(2); ((b3 & 0x3f) << 06) |
dst.put((char)(((b1 & 0x0f) << 12) | (b4 & 0x3f);
((b2 & 0x3f) << 06) | if (isMalformed4(b2, b3, b4) ||
((b3 & 0x3f) << 0))); !Surrogate.neededFor(uc)) { // shortest form check
mark += 3; return malformed(src, mark, 4);
continue;
case 15:
// 4, 5, or 6 bytes
int b4, b5, b6, uc, n;
switch (b1 & 0x0f) {
case 0: case 1: case 2: case 3:
case 4: case 5: case 6: case 7:
// 4 bytes, 21 bits
if (src.remaining() < 3)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = src.get()))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = src.get()))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = src.get()))
return CoderResult.malformedForLength(3);
uc = (((b1 & 0x07) << 18) |
((b2 & 0x3f) << 12) |
((b3 & 0x3f) << 06) |
((b4 & 0x3f) << 00));
n = 4;
break;
case 8: case 9: case 10: case 11:
// 5 bytes, 26 bits
if (src.remaining() < 4)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = src.get()))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = src.get()))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = src.get()))
return CoderResult.malformedForLength(3);
if (!isContinuation(b5 = src.get()))
return CoderResult.malformedForLength(4);
uc = (((b1 & 0x03) << 24) |
((b2 & 0x3f) << 18) |
((b3 & 0x3f) << 12) |
((b4 & 0x3f) << 06) |
((b5 & 0x3f) << 00));
n = 5;
break;
case 12: case 13:
// 6 bytes, 31 bits
if (src.remaining() < 5)
return CoderResult.UNDERFLOW;
if (!isContinuation(b2 = src.get()))
return CoderResult.malformedForLength(1);
if (!isContinuation(b3 = src.get()))
return CoderResult.malformedForLength(2);
if (!isContinuation(b4 = src.get()))
return CoderResult.malformedForLength(3);
if (!isContinuation(b5 = src.get()))
return CoderResult.malformedForLength(4);
if (!isContinuation(b6 = src.get()))
return CoderResult.malformedForLength(5);
uc = (((b1 & 0x01) << 30) |
((b2 & 0x3f) << 24) |
((b3 & 0x3f) << 18) |
((b4 & 0x3f) << 12) |
((b5 & 0x3f) << 06) |
((b6 & 0x3f)));
n = 6;
break;
default:
return CoderResult.malformedForLength(1);
}
if (sgg.generate(uc, n, dst) < 0)
return sgg.error();
mark += n;
continue;
default:
return CoderResult.malformedForLength(1);
} }
dst.put(Surrogate.high(uc));
dst.put(Surrogate.low(uc));
mark += 4;
} else {
return malformed(src, mark, 1);
} }
return CoderResult.UNDERFLOW;
} finally {
src.position(mark);
} }
return xflow(src, mark, 0);
} }
protected CoderResult decodeLoop(ByteBuffer src, protected CoderResult decodeLoop(ByteBuffer src,
@ -377,10 +328,8 @@ class UTF_8 extends Unicode
else else
return decodeBufferLoop(src, dst); return decodeBufferLoop(src, dst);
} }
} }
private static class Encoder extends CharsetEncoder { private static class Encoder extends CharsetEncoder {
private Encoder(Charset cs) { private Encoder(Charset cs) {
@ -391,141 +340,126 @@ class UTF_8 extends Unicode
return !Surrogate.is(c); return !Surrogate.is(c);
} }
private final Surrogate.Parser sgp = new Surrogate.Parser(); public boolean isLegalReplacement(byte[] repl) {
return ((repl.length == 1 && repl[0] >= 0) ||
super.isLegalReplacement(repl));
}
private static CoderResult overflow(CharBuffer src, int sp,
ByteBuffer dst, int dp) {
updatePositions(src, sp, dst, dp);
return CoderResult.OVERFLOW;
}
private static CoderResult overflow(CharBuffer src, int mark) {
src.position(mark);
return CoderResult.OVERFLOW;
}
private Surrogate.Parser sgp;
private CoderResult encodeArrayLoop(CharBuffer src, private CoderResult encodeArrayLoop(CharBuffer src,
ByteBuffer dst) ByteBuffer dst)
{ {
char[] sa = src.array(); char[] sa = src.array();
int sp = src.arrayOffset() + src.position(); int sp = src.arrayOffset() + src.position();
int sl = src.arrayOffset() + src.limit(); int sl = src.arrayOffset() + src.limit();
assert (sp <= sl);
sp = (sp <= sl ? sp : sl);
byte[] da = dst.array(); byte[] da = dst.array();
int dp = dst.arrayOffset() + dst.position(); int dp = dst.arrayOffset() + dst.position();
int dl = dst.arrayOffset() + dst.limit(); int dl = dst.arrayOffset() + dst.limit();
assert (dp <= dl); int dlASCII = dp + Math.min(sl - sp, dl - dp);
dp = (dp <= dl ? dp : dl);
try {
while (sp < sl) {
char c = sa[sp];
if (c < 0x80) {
// Have at most seven bits
if (dp >= dl)
return CoderResult.OVERFLOW;
da[dp++] = (byte)c;
sp++;
continue;
}
if (!Surrogate.is(c)) {
// 2 bytes, 11 bits
if (c < 0x800) {
if (dl - dp < 2)
return CoderResult.OVERFLOW;
da[dp++] = (byte)(0xc0 | ((c >> 06)));
da[dp++] = (byte)(0x80 | ((c >> 00) & 0x3f));
sp++;
continue;
}
if (c <= '\uFFFF') {
// 3 bytes, 16 bits
if (dl - dp < 3)
return CoderResult.OVERFLOW;
da[dp++] = (byte)(0xe0 | ((c >> 12)));
da[dp++] = (byte)(0x80 | ((c >> 06) & 0x3f));
da[dp++] = (byte)(0x80 | ((c >> 00) & 0x3f));
sp++;
continue;
}
}
//ASCII only loop
while (dp < dlASCII && sa[sp] < '\u0080')
da[dp++] = (byte) sa[sp++];
while (sp < sl) {
int c = sa[sp];
if (c < 0x80) {
// Have at most seven bits
if (dp >= dl)
return overflow(src, sp, dst, dp);
da[dp++] = (byte)c;
} else if (c < 0x800) {
// 2 bytes, 11 bits
if (dl - dp < 2)
return overflow(src, sp, dst, dp);
da[dp++] = (byte)(0xc0 | ((c >> 06)));
da[dp++] = (byte)(0x80 | (c & 0x3f));
} else if (Surrogate.is(c)) {
// Have a surrogate pair // Have a surrogate pair
int uc = sgp.parse(c, sa, sp, sl); if (sgp == null)
if (uc < 0) sgp = new Surrogate.Parser();
int uc = sgp.parse((char)c, sa, sp, sl);
if (uc < 0) {
updatePositions(src, sp, dst, dp);
return sgp.error(); return sgp.error();
if (uc < 0x200000) {
if (dl - dp < 4)
return CoderResult.OVERFLOW;
da[dp++] = (byte)(0xf0 | ((uc >> 18)));
da[dp++] = (byte)(0x80 | ((uc >> 12) & 0x3f));
da[dp++] = (byte)(0x80 | ((uc >> 06) & 0x3f));
da[dp++] = (byte)(0x80 | ((uc >> 00) & 0x3f));
sp += sgp.increment();
continue;
} }
assert false; if (dl - dp < 4)
return overflow(src, sp, dst, dp);
da[dp++] = (byte)(0xf0 | ((uc >> 18)));
da[dp++] = (byte)(0x80 | ((uc >> 12) & 0x3f));
da[dp++] = (byte)(0x80 | ((uc >> 06) & 0x3f));
da[dp++] = (byte)(0x80 | (uc & 0x3f));
sp++; // 2 chars
} else {
// 3 bytes, 16 bits
if (dl - dp < 3)
return overflow(src, sp, dst, dp);
da[dp++] = (byte)(0xe0 | ((c >> 12)));
da[dp++] = (byte)(0x80 | ((c >> 06) & 0x3f));
da[dp++] = (byte)(0x80 | (c & 0x3f));
} }
return CoderResult.UNDERFLOW; sp++;
} finally {
src.position(sp - src.arrayOffset());
dst.position(dp - dst.arrayOffset());
} }
updatePositions(src, sp, dst, dp);
return CoderResult.UNDERFLOW;
} }
private CoderResult encodeBufferLoop(CharBuffer src, private CoderResult encodeBufferLoop(CharBuffer src,
ByteBuffer dst) ByteBuffer dst)
{ {
int mark = src.position(); int mark = src.position();
try { while (src.hasRemaining()) {
while (src.hasRemaining()) { int c = src.get();
char c = src.get(); if (c < 0x80) {
// Have at most seven bits
if (c < 0x80) { if (!dst.hasRemaining())
// Have at most seven bits return overflow(src, mark);
if (!dst.hasRemaining()) dst.put((byte)c);
return CoderResult.OVERFLOW; } else if (c < 0x800) {
dst.put((byte)c); // 2 bytes, 11 bits
mark++; if (dst.remaining() < 2)
continue; return overflow(src, mark);
} dst.put((byte)(0xc0 | ((c >> 06))));
dst.put((byte)(0x80 | (c & 0x3f)));
if (!Surrogate.is(c)) { } else if (Surrogate.is(c)) {
if (c < 0x800) {
// 2 bytes, 11 bits
if (dst.remaining() < 2)
return CoderResult.OVERFLOW;
dst.put((byte)(0xc0 | ((c >> 06))));
dst.put((byte)(0x80 | ((c >> 00) & 0x3f)));
mark++;
continue;
}
if (c <= '\uFFFF') {
// 3 bytes, 16 bits
if (dst.remaining() < 3)
return CoderResult.OVERFLOW;
dst.put((byte)(0xe0 | ((c >> 12))));
dst.put((byte)(0x80 | ((c >> 06) & 0x3f)));
dst.put((byte)(0x80 | ((c >> 00) & 0x3f)));
mark++;
continue;
}
}
// Have a surrogate pair // Have a surrogate pair
int uc = sgp.parse(c, src); if (sgp == null)
if (uc < 0) sgp = new Surrogate.Parser();
int uc = sgp.parse((char)c, src);
if (uc < 0) {
src.position(mark);
return sgp.error(); return sgp.error();
if (uc < 0x200000) {
if (dst.remaining() < 4)
return CoderResult.OVERFLOW;
dst.put((byte)(0xf0 | ((uc >> 18))));
dst.put((byte)(0x80 | ((uc >> 12) & 0x3f)));
dst.put((byte)(0x80 | ((uc >> 06) & 0x3f)));
dst.put((byte)(0x80 | ((uc >> 00) & 0x3f)));
mark += sgp.increment();
continue;
} }
assert false; if (dst.remaining() < 4)
return overflow(src, mark);
dst.put((byte)(0xf0 | ((uc >> 18))));
dst.put((byte)(0x80 | ((uc >> 12) & 0x3f)));
dst.put((byte)(0x80 | ((uc >> 06) & 0x3f)));
dst.put((byte)(0x80 | (uc & 0x3f)));
mark++; //2 chars
} else {
// 3 bytes, 16 bits
if (dst.remaining() < 3)
return overflow(src, mark);
dst.put((byte)(0xe0 | ((c >> 12))));
dst.put((byte)(0x80 | ((c >> 06) & 0x3f)));
dst.put((byte)(0x80 | (c & 0x3f)));
} }
return CoderResult.UNDERFLOW; mark++;
} finally {
src.position(mark);
} }
src.position(mark);
return CoderResult.UNDERFLOW;
} }
protected final CoderResult encodeLoop(CharBuffer src, protected final CoderResult encodeLoop(CharBuffer src,
@ -536,7 +470,5 @@ class UTF_8 extends Unicode
else else
return encodeBufferLoop(src, dst); return encodeBufferLoop(src, dst);
} }
} }
} }

View File

@ -0,0 +1,393 @@
/*
* Copyright 2008 Sun Microsystems, Inc. All Rights Reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
* CA 95054 USA or visit www.sun.com if you need additional information or
* have any questions.
*/
/*
* @test
* @bug 4486841
* @summary Test UTF-8 charset
*/
import java.nio.charset.*;
import java.nio.*;
import java.util.*;
public class TestUTF8 {
static char[] decode(byte[] bb, String csn, boolean testDirect)
throws Exception {
CharsetDecoder dec = Charset.forName(csn).newDecoder();
ByteBuffer bbf;
CharBuffer cbf;
if (testDirect) {
bbf = ByteBuffer.allocateDirect(bb.length);
cbf = ByteBuffer.allocateDirect(bb.length*2).asCharBuffer();
bbf.put(bb).flip();
} else {
bbf = ByteBuffer.wrap(bb);
cbf = CharBuffer.allocate(bb.length);
}
CoderResult cr = dec.decode(bbf, cbf, true);
if (cr != CoderResult.UNDERFLOW)
throw new RuntimeException("Decoding err: " + csn);
char[] cc = new char[cbf.position()];
cbf.flip(); cbf.get(cc);
return cc;
}
static CoderResult decodeCR(byte[] bb, String csn, boolean testDirect)
throws Exception {
CharsetDecoder dec = Charset.forName(csn).newDecoder();
ByteBuffer bbf;
CharBuffer cbf;
if (testDirect) {
bbf = ByteBuffer.allocateDirect(bb.length);
cbf = ByteBuffer.allocateDirect(bb.length*2).asCharBuffer();
bbf.put(bb).flip();
} else {
bbf = ByteBuffer.wrap(bb);
cbf = CharBuffer.allocate(bb.length);
}
return dec.decode(bbf, cbf, true);
}
static byte[] encode(char[] cc, String csn, boolean testDirect)
throws Exception {
ByteBuffer bbf;
CharBuffer cbf;
CharsetEncoder enc = Charset.forName(csn).newEncoder();
if (testDirect) {
bbf = ByteBuffer.allocateDirect(cc.length * 4);
cbf = ByteBuffer.allocateDirect(cc.length * 2).asCharBuffer();
cbf.put(cc).flip();
} else {
bbf = ByteBuffer.allocate(cc.length * 4);
cbf = CharBuffer.wrap(cc);
}
CoderResult cr = enc.encode(cbf, bbf, true);
if (cr != CoderResult.UNDERFLOW)
throw new RuntimeException("Encoding err: " + csn);
byte[] bb = new byte[bbf.position()];
bbf.flip(); bbf.get(bb);
return bb;
}
static CoderResult encodeCR(char[] cc, String csn, boolean testDirect)
throws Exception {
ByteBuffer bbf;
CharBuffer cbf;
CharsetEncoder enc = Charset.forName(csn).newEncoder();
if (testDirect) {
bbf = ByteBuffer.allocateDirect(cc.length * 4);
cbf = ByteBuffer.allocateDirect(cc.length * 2).asCharBuffer();
cbf.put(cc).flip();
} else {
bbf = ByteBuffer.allocate(cc.length * 4);
cbf = CharBuffer.wrap(cc);
}
return enc.encode(cbf, bbf, true);
}
static char[] getUTFChars() {
char[] cc = new char[0x10000 - 0xe000 + 0xd800 + //bmp
(0x110000 - 0x10000) * 2]; //supp
int pos = 0;
int i = 0;
for (i = 0; i < 0xd800; i++)
cc[pos++] = (char)i;
for (i = 0xe000; i < 0x10000; i++)
cc[pos++] = (char)i;
for (i = 0x10000; i < 0x110000; i++) {
pos += Character.toChars(i, cc, pos);
}
return cc;
}
static int to3ByteUTF8(char c, byte[] bb, int pos) {
bb[pos++] = (byte)(0xe0 | ((c >> 12)));
bb[pos++] = (byte)(0x80 | ((c >> 06) & 0x3f));
bb[pos++] = (byte)(0x80 | ((c >> 00) & 0x3f));
return 3;
}
static void checkRoundtrip(String csn) throws Exception {
System.out.printf(" Check roundtrip <%s>...", csn);
char[] cc = getUTFChars();
byte[] bb = encode(cc, csn, false);
char[] ccO = decode(bb, csn, false);
if (!Arrays.equals(cc, ccO)) {
System.out.printf(" non-direct failed");
}
bb = encode(cc, csn, true);
ccO = decode(bb, csn, true);
if (!Arrays.equals(cc, ccO)) {
System.out.printf(" (direct) failed");
}
System.out.println();
}
static void check6ByteSurrs(String csn) throws Exception {
System.out.printf(" Check 6-byte Surrogates <%s>...%n", csn);
byte[] bb = new byte[(0x110000 - 0x10000) * 6];
char[] cc = new char[(0x110000 - 0x10000) * 2];
int bpos = 0;
int cpos = 0;
for (int i = 0x10000; i < 0x110000; i++) {
Character.toChars(i, cc, cpos);
bpos += to3ByteUTF8(cc[cpos], bb, bpos);
bpos += to3ByteUTF8(cc[cpos + 1], bb, bpos);
cpos += 2;
}
char[] ccO = decode(bb, csn, false);
if (!Arrays.equals(cc, ccO)) {
System.out.printf(" decoding failed%n");
}
ccO = decode(bb, csn, true);
if (!Arrays.equals(cc, ccO)) {
System.out.printf(" decoding(direct) failed%n");
}
}
static void compare(String csn1, String csn2) throws Exception {
System.out.printf(" Diff <%s> <%s>...%n", csn1, csn2);
char[] cc = getUTFChars();
byte[] bb1 = encode(cc, csn1, false);
byte[] bb2 = encode(cc, csn2, false);
if (!Arrays.equals(bb1, bb2))
System.out.printf(" encoding failed%n");
char[] cc1 = decode(bb1, csn1, false);
char[] cc2 = decode(bb1, csn2, false);
if (!Arrays.equals(cc1, cc2)) {
System.out.printf(" decoding failed%n");
}
bb1 = encode(cc, csn1, true);
bb2 = encode(cc, csn2, true);
if (!Arrays.equals(bb1, bb2))
System.out.printf(" encoding (direct) failed%n");
cc1 = decode(bb1, csn1, true);
cc2 = decode(bb1, csn2, true);
if (!Arrays.equals(cc1, cc2)) {
System.out.printf(" decoding (direct) failed%n");
}
}
// The first byte is the length of malformed bytes
static byte[][] malformed = {
// One-byte sequences:
{1, (byte)0xFF },
{1, (byte)0xC0 },
{1, (byte)0x80 },
{1, (byte)0xFF, (byte)0xFF}, // all ones
{1, (byte)0xA0, (byte)0x80}, // 101x first byte first nibble
// Two-byte sequences:
{1, (byte)0xC0, (byte)0x80}, // invalid first byte
{1, (byte)0xC1, (byte)0xBF}, // invalid first byte
{1, (byte)0xC2, (byte)0x00}, // invalid second byte
{1, (byte)0xC2, (byte)0xC0}, // invalid second byte
{1, (byte)0xD0, (byte)0x00}, // invalid second byte
{1, (byte)0xD0, (byte)0xC0}, // invalid second byte
{1, (byte)0xDF, (byte)0x00}, // invalid second byte
{1, (byte)0xDF, (byte)0xC0}, // invalid second byte
// Three-byte sequences
{1, (byte)0xE0, (byte)0x80, (byte)0x80}, // 111x first byte first nibble
{1, (byte)0xE0, (byte)0x80, (byte)0x80 }, // U+0000 zero-padded
{1, (byte)0xE0, (byte)0x81, (byte)0xBF }, // U+007F zero-padded
{1, (byte)0xE0, (byte)0x9F, (byte)0xBF }, // U+07FF zero-padded
{1, (byte)0xE0, (byte)0xC0, (byte)0xBF }, // invalid second byte
{2, (byte)0xE0, (byte)0xA0, (byte)0x7F }, // invalid third byte
{2, (byte)0xE0, (byte)0xA0, (byte)0xC0 }, // invalid third byte
{1, (byte)0xFF, (byte)0xFF, (byte)0xFF }, // all ones
{1, (byte)0xE0, (byte)0xC0, (byte)0x80 }, // invalid second byte
{1, (byte)0xE0, (byte)0x80, (byte)0xC0 }, // invalid first byte
// Four-byte sequences
{1, (byte)0xF0, (byte)0x80, (byte)0x80, (byte)0x80 }, // U+0000 zero-padded
{1, (byte)0xF0, (byte)0x80, (byte)0x81, (byte)0xBF }, // U+007F zero-padded
{1, (byte)0xF0, (byte)0x80, (byte)0x9F, (byte)0xBF }, // U+007F zero-padded
{1, (byte)0xF0, (byte)0x8F, (byte)0xBF, (byte)0xBF }, // U+07FF zero-padded
{1, (byte)0xFF, (byte)0xFF, (byte)0xFF, (byte)0xFF }, // all ones
{1, (byte)0xF0, (byte)0x80, (byte)0x80, (byte)0x80}, // invalid second byte
{1, (byte)0xF0, (byte)0xC0, (byte)0x80, (byte)0x80 }, // invalid second byte
{2, (byte)0xF0, (byte)0x90, (byte)0xC0, (byte)0x80 }, // invalid third byte
{3, (byte)0xF0, (byte)0x90, (byte)0x80, (byte)0xC0 }, // invalid third byte
{1, (byte)0xF1, (byte)0xC0, (byte)0x80, (byte)0x80 }, // invalid second byte
{2, (byte)0xF1, (byte)0x80, (byte)0xC0, (byte)0x80 }, // invalid third byte
{3, (byte)0xF1, (byte)0x80, (byte)0x80, (byte)0xC0 }, // invalid forth byte
{1, (byte)0xF4, (byte)0x90, (byte)0x80, (byte)0xC0 }, // out-range 4-byte
{1, (byte)0xF4, (byte)0xC0, (byte)0x80, (byte)0xC0 }, // out-range 4-byte
{1, (byte)0xF5, (byte)0x80, (byte)0x80, (byte)0xC0 }, // out-range 4-byte
// Five-byte sequences
{5, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x80}, // invalid first byte
{5, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x80 }, // U+0000 zero-padded
{5, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x81, (byte)0xBF }, // U+007F zero-padded
{5, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x9F, (byte)0xBF }, // U+07FF zero-padded
{5, (byte)0xF8, (byte)0x80, (byte)0x8F, (byte)0xBF, (byte)0xBF }, // U+FFFF zero-padded
{1, (byte)0xF8, (byte)0xC0, (byte)0x80, (byte)0x80, (byte)0x80},
{2, (byte)0xF8, (byte)0x80, (byte)0xC0, (byte)0x80, (byte)0x80 },
{3, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0xC1, (byte)0xBF },
{4, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x9F, (byte)0xC0 },
// Six-byte sequences
{6, (byte)0xFC, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x80 }, // U+0000 zero-padded
{6, (byte)0xFC, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x81, (byte)0xBF }, // U+007F zero-padded
{6, (byte)0xFC, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x9F, (byte)0xBF }, // U+07FF zero-padded
{6, (byte)0xFC, (byte)0x80, (byte)0x80, (byte)0x8F, (byte)0xBF, (byte)0xBF }, // U+FFFF zero-padded
{1, (byte)0xF8, (byte)0xC0, (byte)0x80, (byte)0x80, (byte)0x80, (byte)0x80 },
{2, (byte)0xF8, (byte)0x80, (byte)0xC0, (byte)0x80, (byte)0x80, (byte)0x80 },
{3, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0xC1, (byte)0xBF, (byte)0x80 },
{4, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x9F, (byte)0xC0, (byte)0x80 },
{5, (byte)0xF8, (byte)0x80, (byte)0x80, (byte)0x9F, (byte)0x80, (byte)0xC0 },
};
static void checkMalformed(String csn) throws Exception {
boolean failed = false;
System.out.printf(" Check malformed <%s>...%n", csn);
for (boolean direct: new boolean[] {false, true}) {
for (byte[] bins : malformed) {
int mlen = bins[0];
byte[] bin = Arrays.copyOfRange(bins, 1, bins.length);
CoderResult cr = decodeCR(bin, csn, direct);
String ashex = "";
for (int i = 0; i < bin.length; i++) {
if (i > 0) ashex += " ";
ashex += Integer.toBinaryString((int)bin[i] & 0xff);
}
if (!cr.isMalformed()) {
System.out.printf(" FAIL(direct=%b): [%s] not malformed.\n", direct, ashex);
failed = true;
} else if (cr.length() != mlen) {
System.out.printf(" FAIL(direct=%b): [%s] malformed[len=%d].\n", direct, ashex, cr.length());
failed = true;
}
}
}
if (failed)
throw new RuntimeException("Check malformed failed " + csn);
}
static boolean check(CharsetDecoder dec, byte[] utf8s, boolean direct, int[] flow) {
int inPos = flow[0];
int inLen = flow[1];
int outPos = flow[2];
int outLen = flow[3];
int expedInPos = flow[4];
int expedOutPos = flow[5];
CoderResult expedCR = (flow[6]==0)?CoderResult.UNDERFLOW
:CoderResult.OVERFLOW;
ByteBuffer bbf;
CharBuffer cbf;
if (direct) {
bbf = ByteBuffer.allocateDirect(inPos + utf8s.length);
cbf = ByteBuffer.allocateDirect((outPos + outLen)*2).asCharBuffer();
} else {
bbf = ByteBuffer.allocate(inPos + utf8s.length);
cbf = CharBuffer.allocate(outPos + outLen);
}
bbf.position(inPos);
bbf.put(utf8s).flip().position(inPos).limit(inPos + inLen);
cbf.position(outPos);
dec.reset();
CoderResult cr = dec.decode(bbf, cbf, false);
if (cr != expedCR ||
bbf.position() != expedInPos ||
cbf.position() != expedOutPos) {
System.out.printf("Expected(direct=%5b): [", direct);
for (int i:flow) System.out.print(" " + i);
System.out.println("] CR=" + cr +
", inPos=" + bbf.position() +
", outPos=" + cbf.position());
return false;
}
return true;
}
static void checkUnderOverflow(String csn) throws Exception {
System.out.printf(" Check under/overflow <%s>...%n", csn);
CharsetDecoder dec = Charset.forName(csn).newDecoder();
boolean failed = false;
byte[] utf8s = new String("\u007f\u07ff\ue000\ud800\udc00").getBytes("UTF-8");
int inlen = utf8s.length;
for (int inoff = 0; inoff < 20; inoff++) {
for (int outoff = 0; outoff < 20; outoff++) {
int[][] Flows = {
//inpos, inLen, outPos, outLen, inPosEP, outposEP, under(0)/over(1)
{inoff, inlen, outoff, 1, inoff + 1, outoff + 1, 1},
{inoff, inlen, outoff, 2, inoff + 3, outoff + 2, 1},
{inoff, inlen, outoff, 3, inoff + 6, outoff + 3, 1},
{inoff, inlen, outoff, 4, inoff + 6, outoff + 3, 1},
{inoff, inlen, outoff, 5, inoff + 10,outoff + 5, 0},
// underflow
{inoff, 1, outoff, 5, inoff + 1, outoff + 1, 0},
{inoff, 2, outoff, 5, inoff + 1, outoff + 1, 0},
{inoff, 3, outoff, 5, inoff + 3, outoff + 2, 0},
{inoff, 4, outoff, 5, inoff + 3, outoff + 2, 0},
{inoff, 5, outoff, 5, inoff + 3, outoff + 2, 0},
{inoff, 6, outoff, 5, inoff + 6, outoff + 3, 0},
{inoff, 7, outoff, 5, inoff + 6, outoff + 3, 0},
{inoff, 8, outoff, 5, inoff + 6, outoff + 3, 0},
{inoff, 9, outoff, 5, inoff + 6, outoff + 3, 0},
{inoff, 10, outoff, 5, inoff + 10,outoff + 5, 0},
// 2-byte underflow/overflow
{inoff, 2, outoff, 1, inoff + 1, outoff + 1, 0},
{inoff, 3, outoff, 1, inoff + 1, outoff + 1, 1},
// 3-byte underflow/overflow
{inoff, 4, outoff, 2, inoff + 3, outoff + 2, 0},
{inoff, 5, outoff, 2, inoff + 3, outoff + 2, 0},
{inoff, 6, outoff, 2, inoff + 3, outoff + 2, 1},
// 4-byte underflow/overflow
{inoff, 7, outoff, 4, inoff + 6, outoff + 3, 0},
{inoff, 8, outoff, 4, inoff + 6, outoff + 3, 0},
{inoff, 9, outoff, 4, inoff + 6, outoff + 3, 0},
{inoff, 10, outoff, 4, inoff + 6, outoff + 3, 1},
};
for (boolean direct: new boolean[] {false, true}) {
for (int[] flow: Flows) {
if (!check(dec, utf8s, direct, flow))
failed = true;
}
}}}
if (failed)
throw new RuntimeException("Check under/overflow failed " + csn);
}
public static void main(String[] args) throws Exception {
checkRoundtrip("UTF-8");
check6ByteSurrs("UTF-8");
//compare("UTF-8", "UTF-8-OLD");
checkMalformed("UTF-8");
checkUnderOverflow("UTF-8");
}
}