UTF8Encoding

// System.Text.UTF8Encoding — System.Core (decompiled from Larva HIR)

import System;

into namespace System.Text
{
	public final class UTF8Encoding
	{
		/// Whether `count` bytes from `start` are well-formed UTF-8.
		static IsValid(u8[] bytes, i32 start, i32 count) : bool
		{
			i32 i = start;
			i32 end = start + count;
			loop (i < end)
			{
				u32 b = bytes[i];
				if (b < 128)
				{
					i++;
					continue;
				}
				i32 extra = b >= 245 ? 0 : b >= 240 ? 3 : b >= 224 ? 2 : b >= 194 ? 1 : 0;
				if (extra == 0 || i + extra >= end)
					return false;
				u32 point = b & (extra == 1 ? 31 : extra == 2 ? 15 : 7);
				loop (i32 k = 1..=extra)
				{
					u32 next = bytes[i + k];
					if ((next & 192) != 128)
						return false;
					point = point << 6 | next & 63;
				}
				if (extra == 2 && (point < 2048 || point >= 55296 && point <= 57343) || extra == 3 && (point < 65536 || point > 1114111))
					return false;
				i += extra + 1;
			}
			return true;
		}

		public:
			/// How many bytes `text` takes in UTF-8 (a str is UTF-8 already: its length).
			GetByteCount(str text) : i32 => text.Length;

			/// The bytes of `text` (a copy: strings do not change).
			GetBytes(str text) : u8[]
			{
				u8[] units = text.Units();
				u8[] bytes = new u8[units.Length];
				loop (i32 i = 0..units.Length)
					bytes[i] = units[i];
				return bytes;
			}

			GetString(u8[] bytes) : str => GetString(bytes, 0, bytes.Length);

			/// The text of `count` bytes from `start` (sequences that are not UTF-8 become U+FFFD).
			GetString(u8[] bytes, i32 start, i32 count) : str
			{
				u8[] copy = new u8[count];
				loop (i32 i = 0..count)
					copy[i] = bytes[start + i];
				if (IsValid(copy, 0, count))
					return new str(copy);
				return new str(Unicode.Encode<u8>(Unicode.Decode(copy)));
			}
	}
}