/// <summary>
/// Calculates the number of UTF-16 code units (<see cref="char"/>s) that would result from converting
/// the provided UTF-8 string to UTF-16 representation.
/// </summary>
/// <param name="inputBuffer">The buffer containing UTF-8 text.</param>
/// <param name="charCount">
/// If this method returns <see langword="true"/>, contains the equivalent <see cref="char"/> count of the input string.
/// If this method returns <see langword="false"/>, the value is undefined.
/// </param>
/// <returns>
/// <see langword="true"/> on success, <see langword="false"/> if the input is not a well-formed UTF-8 string.
/// </returns>
public static bool TryGetUtf16CharCount(ReadOnlySpan <byte> inputBuffer, out int charCount)
{
if (Utf8Util.GetIndexOfFirstInvalidUtf8Sequence(inputBuffer, out int runeCount, out int surrogatePairCount) < 0)
{
// can't overflow because UTF-16 code unit count is always <= UTF-8 code unit count for well-formed strings
charCount = runeCount + surrogatePairCount;
return(true);
}
private static void GetIndexOfFirstInvalidUtf8Sequence_Test_Core(byte[] input, int expectedRetVal, int expectedRuneCount, int expectedSurrogatePairCount)
{
// Arrange
var inputBytes = NativeMemory.GetProtectedReadonlyBuffer(input);
// Act
var indexOfFirstInvalidChar = Utf8Util.GetIndexOfFirstInvalidUtf8Sequence(inputBytes, out int actualRuneCount, out int actualSurrogatePairCount);
// Assert
Assert.Equal(expectedRetVal, indexOfFirstInvalidChar);
Assert.Equal(expectedRuneCount, actualRuneCount);
Assert.Equal(expectedSurrogatePairCount, actualSurrogatePairCount);
}
/// <summary>
/// Peeks at the first UTF-8 sequence in the input buffer and returns information about that
/// sequence. If the sequence is well-formed, returns <see cref="SequenceValidity.WellFormed"/>
/// and sets the <paramref name="scalarValue"/> output parameter to the scalar value encoded by
/// the sequence. If the return value is anything other than <see cref="SequenceValidity.WellFormed"/>,
/// sets the <paramref name="scalarValue"/> output parameter to <see cref="UnicodeScalar.ReplacementChar"/>.
/// In all cases, the <paramref name="numBytesConsumed"/> output parameter will contain the
/// number of UTF-8 code units read from the input buffer in order to make the determination.
/// </summary>
public static SequenceValidity PeekFirstSequence(ReadOnlySpan <byte> data, out int numBytesConsumed, out UnicodeScalar scalarValue)
{
// This method is implemented to match the behavior of System.Text.Encoding.UTF8 in terms of
// how many bytes it consumes when reporting invalid sequences. The behavior is as follows:
//
// - Some bytes are *always* invalid (ranges [ C0..C1 ] and [ F5..FF ]), and when these
// are encountered it's an invalid sequence of length 1.
//
// - Multi-byte sequences which are overlong are reported as an invalid sequence of length 2,
// since per the Unicode Standard Table 3-7 it's always possible to tell these by the second byte.
// Exception: Sequences which begin with [ C0..C1 ] are covered by the above case, thus length 1.
//
// - Multi-byte sequences which are improperly terminated (no continuation byte when one is
// expected) are reported as invalid sequences up to and including the last seen continuation byte.
scalarValue = UnicodeScalar.ReplacementChar;
if (data.IsEmpty)
{
// No data to peek at
numBytesConsumed = 0;
return(SequenceValidity.Empty);
}
byte firstByte = data[0];
if (IsAsciiValue(firstByte))
{
// ASCII byte = well-formed one-byte sequence.
scalarValue = UnicodeScalar.CreateWithoutValidation(firstByte);
numBytesConsumed = 1;
return(SequenceValidity.WellFormed);
}
if (!Utf8Util.IsInRangeInclusive(firstByte, (byte)0xC2U, (byte)0xF4U))
{
// Standalone continuation byte or "always invalid" byte = ill-formed one-byte sequence.
goto InvalidOneByteSequence;
}
// At this point, we know we're working with a multi-byte sequence,
// and we know that at least the first byte is potentially valid.
if (data.Length < 2)
{
// One byte of an incomplete multi-byte sequence.
goto OneByteOfIncompleteMultiByteSequence;
}
byte secondByte = data[1];
if (!IsUtf8ContinuationByte(secondByte))
{
// One byte of an improperly terminated multi-byte sequence.
goto InvalidOneByteSequence;
}
if (firstByte < (byte)0xE0U)
{
// Well-formed two-byte sequence.
scalarValue = UnicodeScalar.CreateWithoutValidation((((uint)firstByte & 0x1FU) << 6) | ((uint)secondByte & 0x3FU));
numBytesConsumed = 2;
return(SequenceValidity.WellFormed);
}
if (firstByte < (byte)0xF0U)
{
// Start of a three-byte sequence.
// Need to check for overlong or surrogate sequences.
uint scalar = (((uint)firstByte & 0x0FU) << 12) | (((uint)secondByte & 0x3FU) << 6);
if (scalar < 0x800U || Utf8Util.IsLowWordSurrogate(scalar))
{
goto OverlongOutOfRangeOrSurrogateSequence;
}
// At this point, we have a valid two-byte start of a three-byte sequence.
if (data.Length < 3)
{
// Two bytes of an incomplete three-byte sequence.
goto TwoBytesOfIncompleteMultiByteSequence;
}
else
{
byte thirdByte = data[2];
if (IsUtf8ContinuationByte(thirdByte))
{
// Well-formed three-byte sequence.
scalar |= (uint)thirdByte & 0x3FU;
scalarValue = UnicodeScalar.CreateWithoutValidation(scalar);
numBytesConsumed = 3;
return(SequenceValidity.WellFormed);
}
else
{
// Two bytes of improperly terminated multi-byte sequence.
goto InvalidTwoByteSequence;
}
}
}
{
// Start of four-byte sequence.
// Need to check for overlong or out-of-range sequences.
uint scalar = (((uint)firstByte & 0x07U) << 18) | (((uint)secondByte & 0x3FU) << 12);
if (!Utf8Util.IsInRangeInclusive(scalar, 0x10000U, 0x10FFFFU))
{
goto OverlongOutOfRangeOrSurrogateSequence;
}
// At this point, we have a valid two-byte start of a four-byte sequence.
if (data.Length < 3)
{
// Two bytes of an incomplete four-byte sequence.
goto TwoBytesOfIncompleteMultiByteSequence;
}
else
{
byte thirdByte = data[2];
if (IsUtf8ContinuationByte(thirdByte))
{
// Valid three-byte start of a four-byte sequence.
if (data.Length < 4)
{
// Three bytes of an incomplete four-byte sequence.
goto ThreeBytesOfIncompleteMultiByteSequence;
}
else
{
byte fourthByte = data[3];
if (IsUtf8ContinuationByte(fourthByte))
{
// Well-formed four-byte sequence.
scalar |= (((uint)thirdByte & 0x3FU) << 6) | ((uint)fourthByte & 0x3FU);
scalarValue = UnicodeScalar.CreateWithoutValidation(scalar);
numBytesConsumed = 4;
return(SequenceValidity.WellFormed);
}
else
{
// Three bytes of an improperly terminated multi-byte sequence.
goto InvalidThreeByteSequence;
}
}
}
else
{
// Two bytes of improperly terminated multi-byte sequence.
goto InvalidTwoByteSequence;
}
}
}
// Everything below here is error handling.
InvalidOneByteSequence:
numBytesConsumed = 1;
return(SequenceValidity.Invalid);
InvalidTwoByteSequence:
OverlongOutOfRangeOrSurrogateSequence:
numBytesConsumed = 2;
return(SequenceValidity.Invalid);
InvalidThreeByteSequence:
numBytesConsumed = 3;
return(SequenceValidity.Invalid);
OneByteOfIncompleteMultiByteSequence:
numBytesConsumed = 1;
return(SequenceValidity.Incomplete);
TwoBytesOfIncompleteMultiByteSequence:
numBytesConsumed = 2;
return(SequenceValidity.Incomplete);
ThreeBytesOfIncompleteMultiByteSequence:
numBytesConsumed = 3;
return(SequenceValidity.Incomplete);
}
/// <summary> /// Returns the index of the first byte of the first invalid UTF-8 sequence in <paramref name="data"/>, /// or -1 if <paramref name="data"/> is a well-formed UTF-8 string. /// </summary> public static int GetIndexOfFirstInvalidUtf8Sequence(ReadOnlySpan <byte> data) => Utf8Util.GetIndexOfFirstInvalidUtf8Sequence(data, out _, out _);
