UTF-8 encoder/decoder in Lua

utf-8.lua

local
function utf8_to_codepoint (stream, i)
    -- returns the codepoint and the index of the next character in the stream.
    local msB = string.byte(stream, i)
    local b2, b1, b0 

    if msB < 128 then return msB, i + 1

    elseif msB < 192 then
        error("Byte values between 0x80 to 0xBF cannot start a multibyte sequence")

    elseif msB < 224 then
        b0 = string.byte(stream, i + 1)
        assert(b0 > 191)
        return (msB - 192)*64 + b0, i + 2

    elseif msB < 240 then
        b1, b0 = string.byte(stream, i + 1, i + 2)
        assert(b0 > 191 and b1 > 191)
        local res = (msB - 224)*4096 + b1%64*64 + b0%64
        if 55296 <= res and res <= 57343 then 
            error("UTF-16 surrogate lead are not valid codepoints")
        end
        return res, i + 3

    elseif msB < 248 then
        b2, b1, b0 = string.byte(stream, i + 1, i + 2, 1 + 3)
        assert(b0 > 191 and b1 > 191 and b2 > 191)
        local res = (msB - 240)*262144 + b3%64*4096 + b2%64*64 + b1%64
        if res < 1114112 then
            return res, i + 4
        end
    end
    error("Invalid UTF-8 character")
end

local
function codepoint_to_utf8(c)
    assert((55296 > c or c > 57343) and c < 1114112, "Bad Unicode code point: "..c..".")
    if     c < 128 then
        return                                                          string.char(c)
    elseif c < 2048 then
        return                                     string.char(192 + c/64, 128 + c%64)
    elseif c < 55296 or 57343 < c and c < 65536 then
        return                    string.char(224 + c/4096, 128 + c/64%64, 128 + c%64)
    elseif c < 1114112 then
        return string.char(240 + c/262144, 128 + c/4096%64, 128 + c/64%64, 128 + c%64)
    end
end

1. the assertions are wrong (continuation bytes are 10xxxxxx, so 0x80-0xbf are possible)
2. use hexadecimal literals (they make more sense, for example 0xd000 and 0xdfff instead of 55296 and 57343)
3. line 27: "1 + 3", that should be "i + 3" - get a better font ;)
4. line 29: "b3" to "b1" are used, which is an off-by-one error. should be "b2" to "b0"
5. line 27: string.byte usage is wrong, it's actually string,byte(s, i, [j]) and it returns from i to j. it should be string.byte(stream, i + 1, i + 3)

anyway, after making these changes it works for me. nicer than my old way of doing it :)

There still some bug, i think https://github.com/Stepets/utf8.lua is better choice !

for encoding, decoding: https://github.com/luapower/utf8