public unsafe ValueString CreateValueStringFromUtf8(byte[] arr, int start, int length)
{
if (start + length > arr.Length)
{
throw new ArgumentException();
}
EnsureSpace(length * 2);
fixed(char *ptr = str)
{
var ptrx = ptr + used;
var offset = used;
foreach (var codePoint in new Utf8String.CodePointEnumerable(arr, start, length))
{
int charsEncoded;
if (!Utf16LittleEndianEncoder.TryEncodeCodePoint(codePoint, ptrx, out charsEncoded))
{
// TODO: Change Exception type
throw new Exception("invalid character");
}
ptrx += charsEncoded;
}
var strlen = (int)(ptrx - ptr - used);
used += strlen;
return(new ValueString(str, offset, strlen));
}
}
public override string ToString()
{
// get length first
// TODO: Optimize for characters of length 1 or 2 in UTF-8 representation (no need to read anything)
// TODO: is compiler gonna do the right thing here?
// TODO: Should we use Linq's Count()?
int len = 0;
foreach (var codePoint in CodePoints)
{
len++;
if (!UnicodeCodePoint.IsBmp(codePoint))
{
len++;
}
}
unsafe
{
Span <byte> buffer;
char * stackChars = null;
char[] characters = null;
if (len <= 256)
{
char *stackallocedChars = stackalloc char[len];
stackChars = stackallocedChars;
buffer = new Span <byte>(stackChars, len * 2);
}
else
{
// HACK: Can System.Buffers be used here?
characters = new char[len];
buffer = characters.Slice().Cast <char, byte>();
}
foreach (var codePoint in CodePoints)
{
int bytesEncoded;
if (!Utf16LittleEndianEncoder.TryEncodeCodePoint(codePoint, buffer, out bytesEncoded))
{
// TODO: Change Exception type
throw new Exception("invalid character");
}
buffer = buffer.Slice(bytesEncoded);
}
// TODO: We already have a char[] and this will copy, how to avoid that
return(stackChars != null
? new string(stackChars, 0, len)
: new string(characters));
}
}
public override string ToString()
{
// get length first
// TODO: Optimize for characters of length 1 or 2 in UTF-8 representation (no need to read anything)
// TODO: is compiler gonna do the right thing here?
// TODO: Should we use Linq's Count()?
int len = 0;
foreach (var codePoint in CodePoints)
{
len++;
if (UnicodeCodePoint.IsSurrogate(codePoint))
{
len++;
}
}
char[] characters = new char[len];
unsafe
{
fixed(char *pinnedCharacters = characters)
{
Span <byte> buffer = new Span <byte>((byte *)pinnedCharacters, len * 2);
foreach (var codePoint in CodePoints)
{
int bytesEncoded;
if (!Utf16LittleEndianEncoder.TryEncodeCodePoint(codePoint, buffer, out bytesEncoded))
{
// TODO: Change Exception type
throw new Exception("invalid character");
}
buffer = buffer.Slice(bytesEncoded);
}
}
}
// TODO: We already have a char[] and this will copy, how to avoid that
return(new string(characters));
}