internal virtual int GetChars(byte[] bytes, int byteIndex, int byteCount, char[] chars, int charIndex, UTF8Decoder decoder)
{
if (bytes == null || chars == null)
{
throw new ArgumentNullException(bytes == null ? "bytes" : "chars",
"ArgumentNull_Array");
}
if (byteIndex < 0 || byteCount < 0)
{
throw new ArgumentOutOfRangeException((byteIndex < 0 ? "byteIndex" : "byteCount"),
"ArgumentOutOfRange_NeedNonNegNum");
}
if (bytes.Length - byteIndex < byteCount)
{
throw new ArgumentOutOfRangeException("bytes",
"ArgumentOutOfRange_IndexCountBuffer");
}
if (charIndex < 0 || charIndex > chars.Length)
{
throw new ArgumentOutOfRangeException("charIndex",
"ArgumentOutOfRange_Index");
}
int bits = 0;
int trailCount = 0;
bool isSurrogate = false;
// Indicate the current chunk of bytes is a 2-byte, 3-byte or 4-byte UTF8 sequence.
// This is used to detect non-shortest form.
// It will be reset to 0 when the 2nd byte of the UTF8 sequence is read, so that
// we don't check for non-shortest form again.
int byteSequence = 0;
if (decoder != null)
{
bits = decoder.bits;
trailCount = decoder.trailCount;
isSurrogate = decoder.isSurrogate;
byteSequence = decoder.byteSequence;
}
int byteEnd = byteIndex + byteCount;
int charStart = charIndex;
try {
while (byteIndex < byteEnd)
{
byte b = bytes[byteIndex++];
if (trailCount == 0)
{
//
// We are not at a trailing byte.
//
if ((b & 0x80) == 0)
{
// This is the ASCII case.
// 1 7 0vvvvvvv
//
// Found an ASCII character.
//
chars[charIndex++] = (char)b;
}
else
{
// Check if this is a valid starting byte.
byte temp = (byte)b;
while ((temp & 0x80) != 0)
{
temp <<= 1;
trailCount++;
}
switch (trailCount)
{
case 1:
trailCount = 0;
break;
case 2:
// Make sure that bit 8 ~ bit 11 is not all zero.
// 110XXXXx 10xxxxxx
if ((b & 0x1e) == 0)
{
trailCount = 0;
}
break;
case 3:
byteSequence = 3;
break;
case 4:
//
// This is a surrogate unicode pair
//
byteSequence = 4;
break;
default:
trailCount = 0;
break;
}
if (trailCount == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
}
else
{
isSurrogate = (trailCount == 4);
bits = temp >> trailCount;
trailCount--;
}
}
}
else
{
// We are expecting to see bytes like 10vvvvvv
if ((b & 0xC0) != 0x80)
{
// If not, this is NOT a valid sequence.
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
// At this point, we are seeing an invalid trailing byte.
// However, this can be a valid starting byte for another UTF8 byte sequence (e.g.
// this character could be under 0x7f, or a valid leading byte like 110xxxxx).
// So let's put the current byte back, and try to see if this is a valid byte
// for another UTF8 byte sequence.
byteIndex--;
bits = 0;
trailCount = 0;
}
else
{
switch (byteSequence)
{
case 3:
// Check 3-byte sequence for non-shortest form.
// 1110XXXX 10Xxxxxx 10xxxxxx
if (bits == 0 && (b & 0x20) == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
// Rest byteSequence to zero since we are done with non-shortest form check.
byteSequence = 0;
break;
case 4:
// Check 4-byte sequence for non-shortest form.
// 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx
if (bits == 0)
{
if ((b & 0x30) == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
}
else if ((bits & 0x04) != 0)
{
// Make sure that the resulting Unicode is within the valid surrogate range.
// The 4 byte code sequence can hold up to 21 bits, and the maximum valid code point range
// that Unicode (with surrogate) could represent are from U+000000 ~ U+10FFFF.
// Therefore, if the 21 bit (the most significant bit) is 1, we should verify that the 17 ~ 20
// bit are all zero.
// I.e., in 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx,
// XXXXX can only be 10000.
if ((bits & 0x03) != 0 || (b & 0x30) != 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
}
byteSequence = 0;
break;
}
if (--trailCount >= 0)
{
bits = bits << 6 | (b & 0x3F);
if (trailCount == 0)
{
if (!isSurrogate)
{
chars[charIndex++] = (char)bits;
}
else
{
//
// bits >= 0x10000, use surrogate.
//
chars[charIndex++] = (char)(0xD7C0 + (bits >> 10));
chars[charIndex++] = (char)(CharacterInfo.LOW_SURROGATE_START + (bits & 0x3FF));
}
}
}
}
}
}
}
catch (IndexOutOfRangeException) {
throw new ArgumentException("Argument_ConversionOverflow");
}
if (decoder != null)
{
decoder.bits = bits;
decoder.trailCount = trailCount;
decoder.isSurrogate = isSurrogate;
decoder.byteSequence = byteSequence;
}
return(charIndex - charStart);
}
internal virtual int GetCharCount(byte[] bytes, int index, int count, UTF8Decoder decoder)
{
if (bytes == null)
{
throw new ArgumentNullException("bytes",
"ArgumentNull_Array");
}
if (index < 0 || count < 0)
{
throw new ArgumentOutOfRangeException((index < 0 ? "index" : "count"),
"ArgumentOutOfRange_NeedNonNegNum");
}
if (bytes.Length - index < count)
{
throw new ArgumentOutOfRangeException("bytes",
"ArgumentOutOfRange_IndexCountBuffer");
}
int charCount = 0;
int trailCount = 0;
// Indicate the current chunk of bytes is a 2-byte, 3-byte or 4-byte UTF8 sequence.
// This is used to detect non-shortest form.
// It will be reset to 0 when the 2nd byte of the UTF8 sequence is read, so that
// we don't check for non-shortest form again.
int byteSequence = 0;
bool isSurrogate = false;
int bits = 0;
if (decoder != null)
{
trailCount = decoder.trailCount;
isSurrogate = decoder.isSurrogate;
byteSequence = decoder.byteSequence;
bits = decoder.bits;
}
int end = index + count;
while (index < end)
{
byte b = bytes[index++];
if (trailCount == 0)
{
if ((b & 0x80) == 0)
{
// This is an ASCII.
charCount++;
}
else
{
byte temp = b;
while ((temp & 0x80) != 0)
{
temp <<= 1;
trailCount++;
}
switch (trailCount)
{
case 1:
trailCount = 0;
break;
case 2:
// Make sure that bit 8 ~ bit 11 is not all zero.
// 110XXXXx 10xxxxxx
if ((b & 0x1e) == 0)
{
trailCount = 0;
}
break;
case 3:
byteSequence = 3;
break;
case 4:
isSurrogate = true;
byteSequence = 4;
break;
default:
trailCount = 0;
break;
}
if (trailCount == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
}
else
{
bits = temp >> trailCount;
trailCount--;
}
}
}
else
{
// We are expecting to see trailing bytes like 10vvvvvv
if ((b & 0xC0) != 0x80)
{
// If not, this is NOT a valid sequence.
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
index--;
trailCount = 0;
isSurrogate = false;
}
else
{
switch (byteSequence)
{
case 3:
// Check 3-byte sequence for non-shortest form.
// 1110XXXX 10Xxxxxx 10xxxxxx
if (bits == 0 && (b & 0x20) == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
// We are done checking the non-shortest form, reset byteSequence to 0, so that we don't
// do the extra check for the remaining byte of the 3-byte chunk.
byteSequence = 0;
break;
case 4:
// Check 4-byte sequence for non-shortest form.
// 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx
if (bits == 0)
{
if ((b & 0x30) == 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
}
else if ((bits & 0x04) != 0)
{
// Make sure that the resulting Unicode is within the valid surrogate range.
// The 4 byte code sequence can hold up to 21 bits, and the maximum valid code point range
// that Unicode (with surrogate) could represent are from U+000000 ~ U+10FFFF.
// Therefore, if the 21 bit (the most significant bit) is 1, we should verify that the 17 ~ 20
// bit are all zero.
// I.e., in 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx,
// XXXXX can only be 10000.
if ((bits & 0x03) != 0 || (b & 0x30) != 0)
{
if (isThrowException)
{
throw new ArgumentException("Argument_InvalidByteSequence");
}
trailCount = -1;
}
}
byteSequence = 0;
break;
}
if (--trailCount == 0)
{
charCount++;
if (isSurrogate)
{
charCount++;
isSurrogate = false;
}
}
}
}
}
return(charCount);
}
internal virtual int GetCharCount(byte[] bytes, int index, int count, UTF8Decoder decoder) {
if (bytes == null) {
throw new ArgumentNullException("bytes",
Environment.GetResourceString("ArgumentNull_Array"));
}
if (index < 0 || count < 0) {
throw new ArgumentOutOfRangeException((index<0 ? "index" : "count"),
Environment.GetResourceString("ArgumentOutOfRange_NeedNonNegNum"));
}
if (bytes.Length - index < count) {
throw new ArgumentOutOfRangeException("bytes",
Environment.GetResourceString("ArgumentOutOfRange_IndexCountBuffer"));
}
int charCount = 0;
int trailCount = 0;
// Indicate the current chunk of bytes is a 2-byte, 3-byte or 4-byte UTF8 sequence.
// This is used to detect non-shortest form.
// It will be reset to 0 when the 2nd byte of the UTF8 sequence is read, so that
// we don't check for non-shortest form again.
int byteSequence = 0;
bool isSurrogate = false;
int bits = 0;
if (decoder != null) {
trailCount = decoder.trailCount;
isSurrogate = decoder.isSurrogate;
byteSequence = decoder.byteSequence;
bits = decoder.bits;
}
int end = index + count;
while (index < end) {
byte b = bytes[index++];
if (trailCount == 0) {
if ((b & 0x80) == 0) {
// This is an ASCII.
charCount++;
} else {
byte temp = b;
while ((temp & 0x80) != 0) {
temp <<= 1;
trailCount++;
}
switch (trailCount) {
case 1:
trailCount = 0;
break;
case 2:
// Make sure that bit 8 ~ bit 11 is not all zero.
// 110XXXXx 10xxxxxx
if ((b & 0x1e) == 0) {
trailCount = 0;
}
break;
case 3:
byteSequence = 3;
break;
case 4:
isSurrogate = true;
byteSequence = 4;
break;
default:
trailCount = 0;
break;
}
if (trailCount == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), index-1));
}
} else {
bits = temp >> trailCount;
trailCount--;
}
}
} else {
// We are expecting to see trailing bytes like 10vvvvvv
if ((b & 0xC0) != 0x80) {
// If not, this is NOT a valid sequence.
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), index-1));
}
index--;
trailCount = 0;
isSurrogate = false;
} else {
switch (byteSequence) {
case 3:
// Check 3-byte sequence for non-shortest form.
// 1110XXXX 10Xxxxxx 10xxxxxx
if (bits == 0 && (b & 0x20) == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), index-1));
}
trailCount = -1;
}
// We are done checking the non-shortest form, reset byteSequence to 0, so that we don't
// do the extra check for the remaining byte of the 3-byte chunk.
byteSequence = 0;
break;
case 4:
// Check 4-byte sequence for non-shortest form.
// 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx
if (bits == 0) {
if ((b & 0x30) == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), index-1));
}
trailCount = -1;
}
} else if ((bits & 0x04) != 0) {
// Make sure that the resulting Unicode is within the valid surrogate range.
// The 4 byte code sequence can hold up to 21 bits, and the maximum valid code point ragne
// that Unicode (with surrogate) could represent are from U+000000 ~ U+10FFFF.
// Therefore, if the 21 bit (the most significant bit) is 1, we should verify that the 17 ~ 20
// bit are all zero.
// I.e., in 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx,
// XXXXX can only be 10000.
if ((bits & 0x03) != 0 || (b & 0x30) != 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), index-1));
}
trailCount = -1;
}
}
byteSequence = 0;
break;
}
if (--trailCount == 0) {
charCount++;
if (isSurrogate) {
charCount++;
isSurrogate = false;
}
}
}
}
}
return charCount;
}
internal virtual int GetChars(byte[] bytes, int byteIndex, int byteCount, char[] chars, int charIndex, UTF8Decoder decoder) {
if (bytes == null || chars == null) {
throw new ArgumentNullException(bytes == null ? "bytes" : "chars",
Environment.GetResourceString("ArgumentNull_Array"));
}
if (byteIndex < 0 || byteCount < 0) {
throw new ArgumentOutOfRangeException((byteIndex<0 ? "byteIndex" : "byteCount"),
Environment.GetResourceString("ArgumentOutOfRange_NeedNonNegNum"));
}
if ( bytes.Length - byteIndex < byteCount)
{
throw new ArgumentOutOfRangeException("bytes",
Environment.GetResourceString("ArgumentOutOfRange_IndexCountBuffer"));
}
if (charIndex < 0 || charIndex > chars.Length) {
throw new ArgumentOutOfRangeException("charIndex",
Environment.GetResourceString("ArgumentOutOfRange_Index"));
}
int bits = 0;
int trailCount = 0;
bool isSurrogate = false;
// Indicate the current chunk of bytes is a 2-byte, 3-byte or 4-byte UTF8 sequence.
// This is used to detect non-shortest form.
// It will be reset to 0 when the 2nd byte of the UTF8 sequence is read, so that
// we don't check for non-shortest form again.
int byteSequence = 0;
if (decoder != null) {
bits = decoder.bits;
trailCount = decoder.trailCount;
isSurrogate = decoder.isSurrogate;
byteSequence = decoder.byteSequence;
}
int byteEnd = byteIndex + byteCount;
int charStart = charIndex;
try {
while (byteIndex < byteEnd) {
byte b = bytes[byteIndex++];
if (trailCount == 0) {
//
// We are not at a trailing byte.
//
if ((b & 0x80) == 0) {
// This is the ASCII case.
// 1 7 0vvvvvvv
//
// Found an ASCII character.
//
chars[charIndex++] = (char)b;
} else {
// Check if this is a valid starting byte.
byte temp = (byte)b;
while ((temp & 0x80) != 0) {
temp <<= 1;
trailCount++;
}
switch (trailCount) {
case 1:
trailCount = 0;
break;
case 2:
// Make sure that bit 8 ~ bit 11 is not all zero.
// 110XXXXx 10xxxxxx
if ((b & 0x1e) == 0) {
trailCount = 0;
}
break;
case 3:
byteSequence = 3;
break;
case 4:
//
// This is a surrogate unicode pair
//
byteSequence = 4;
break;
default:
trailCount = 0;
break;
}
if (trailCount == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), byteIndex-1));
}
} else {
isSurrogate = (trailCount == 4);
bits = temp >> trailCount;
trailCount--;
}
}
} else {
// We are expecting to see bytes like 10vvvvvv
if ((b & 0xC0) != 0x80) {
// If not, this is NOT a valid sequence.
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), byteIndex-1));
}
// At this point, we are seeing an invalid trailing byte.
// However, this can be a valid starting byte for another UTF8 byte sequence (e.g.
// this character could be under 0x7f, or a valid leading byte like 110xxxxx).
// So let's put the current byte back, and try to see if this is a valid byte
// for another UTF8 byte sequence.
byteIndex--;
bits = 0;
trailCount = 0;
} else {
switch (byteSequence) {
case 3:
// Check 3-byte sequence for non-shortest form.
// 1110XXXX 10Xxxxxx 10xxxxxx
if (bits == 0 && (b & 0x20) == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), byteIndex-1));
}
trailCount = -1;
}
// Rest byteSequence to zero since we are done with non-shortest form check.
byteSequence = 0;
break;
case 4:
// Check 4-byte sequence for non-shortest form.
// 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx
if (bits == 0) {
if ((b & 0x30) == 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), byteIndex-1));
}
trailCount = -1;
}
} else if ((bits & 0x04) != 0) {
// Make sure that the resulting Unicode is within the valid surrogate range.
// The 4 byte code sequence can hold up to 21 bits, and the maximum valid code point ragne
// that Unicode (with surrogate) could represent are from U+000000 ~ U+10FFFF.
// Therefore, if the 21 bit (the most significant bit) is 1, we should verify that the 17 ~ 20
// bit are all zero.
// I.e., in 11110XXX 10XXxxxx 10xxxxxx 10xxxxxx,
// XXXXX can only be 10000.
if ((bits & 0x03) != 0 || (b & 0x30) != 0) {
if (isThrowException) {
throw new ArgumentException(String.Format(Environment.GetResourceString("Argument_InvalidByteSequence"), byteIndex-1));
}
trailCount = -1;
}
}
byteSequence = 0;
break;
}
if (--trailCount >= 0) {
bits = bits << 6 | (b & 0x3F);
if (trailCount == 0) {
if (!isSurrogate) {
chars[charIndex++] = (char)bits;
}
else {
//
// bits >= 0x10000, use surrogate.
//
chars[charIndex++] = (char)(0xD7C0 + (bits >> 10));
chars[charIndex++] = (char)(CharacterInfo.LOW_SURROGATE_START + (bits & 0x3FF));
}
}
}
}
}
}
} catch (IndexOutOfRangeException) {
throw new ArgumentException(Environment.GetResourceString("Argument_ConversionOverflow"));
}
if (decoder != null) {
decoder.bits = bits;
decoder.trailCount = trailCount;
decoder.isSurrogate = isSurrogate;
decoder.byteSequence = byteSequence;
}
return charIndex - charStart;
}
public UTF8Encoding()
{
encoder = new UTF8Encoder();
decoder = new UTF8Decoder();
}
