utf.c

struct DecodeResult {
    uint32_t codepoint;
    uint32_t decodedBytes;
};

static inline struct DecodeResult DecodeOneUTF8(const uint8_t *utf8) {
    static uint8_t byteCountTable[16] = {
        // 0xxx
        [0x0] = 1,
        [0x1] = 1,
        [0x2] = 1,
        [0x3] = 1,
        [0x4] = 1,
        [0x5] = 1,
        [0x6] = 1,
        [0x7] = 1,
        
        // 110x
        [0xc] = 2,
        [0xd] = 2,
        
        // 1110
        [0xe] = 3,
        
        // 1111
        [0xf] = 4
    };
    
    uint8_t byteOne = utf8[0];
    uint8_t byteTwo = utf8[1];
    uint8_t byteThree = utf8[2];
    uint8_t byteFour = utf8[3];
    
    uint32_t codepoints[5] = {
        -1, // dummy
        byteOne,
        ((byteOne & 0x1f) << 6) | (byteTwo & 0x3f),
        ((byteOne & 0x0f) << 12) | ((byteTwo & 0x3f) << 6) | (byteThree & 0x3f),
        ((byteOne & 0x03) << 18) | ((byteTwo & 0x3f) << 12) | ((byteThree & 0x3f) << 6) | (byteFour & 0x3f)
    };
    
    uint8_t byteCount = byteCountTable[byteOne >> 4];
    
    return (struct DecodeResult){ codepoints[byteCount], byteCount };
}

size_t DecodeUTF8(const uint8_t *utf8, uint32_t *destination) {
    size_t length = 0;
    while(*utf8 != '\0') {
        struct DecodeResult result = DecodeOneUTF8(utf8);
        *destination++ = result.codepoint;
        utf8 += result.decodedBytes;
        length++;
    }
    return length;
}