private static bool ValidateUtf8(ref Utf8ValidatorState state, ReadableBuffer payload, bool fin)
{
// Walk through the payload verifying it
var offset = 0;
foreach (var mem in payload)
{
var span = mem.Span;
for (int i = 0; i < span.Length; i++)
{
var b = span[i];
if (!state._withinSequence)
{
// This is the first byte of a char, so set things up
var width = _utf8Width[b];
state._remainingBytesInChar = width - 1;
if (state._remainingBytesInChar < 0)
{
// Invalid first byte
return(false);
}
// Use the width (-1) to index into the mask and min tables.
state._currentDecodedValue = b & _utf8Mask[width - 1];
state._minCodePoint = _utf8Min[width - 1];
state._withinSequence = true;
}
else
{
// Add this byte to the value
state._currentDecodedValue = (state._currentDecodedValue << 6) | (b & 0x3F);
state._remainingBytesInChar--;
}
// Fast invalid exits
if (state._remainingBytesInChar == 1 && state._currentDecodedValue >= 0x360 && state._currentDecodedValue <= 0x37F)
{
// This will be a UTF-16 surrogate: 0xD800-0xDFFF
return(false);
}
if (state._remainingBytesInChar == 2 && state._currentDecodedValue >= 0x110)
{
// This will be above the maximum Unicode character (0x10FFFF).
return(false);
}
if (state._remainingBytesInChar == 0)
{
// Check the range of the final decoded value
if (state._currentDecodedValue < state._minCodePoint)
{
// This encoding is longer than it should be, which is not allowed.
return(false);
}
// Reset state
state._withinSequence = false;
}
offset++;
}
}
// We're done.
// The value is valid if:
// 1. We haven't reached the end of the whole message yet (we'll be caching this state for the next message)
// 2. We aren't inside a character sequence (i.e. the last character isn't unterminated)
return(!fin || !state._withinSequence);
}
public static bool ValidateUtf8(ReadableBuffer payload)
{
var state = new Utf8ValidatorState();
return(ValidateUtf8(ref state, payload, fin: true));
}
