LizeriumDataToolkit

XUTF

All values are stored in little endian format

---

HEADER (17 bytes)

fourCC signature = "XUTF"
byte version = 1
uint32 string block length
uint32 node block length
uint32 data block length

---

VARINT ENCODING

To ensure a unique bit pattern for each integer, each new byte in the varint encoding represents:

the previous minimum value + the new bit encoding

---

Encoding example:

if (u <= 127)
{
    PutByte((byte) u);
}
else if (u <= 16511)
{
    u -= 128;
    PutByte((byte) ((u & 0x7f) | 0x80));
    PutByte((byte) ((u >> 7) & 0x7f));
}
else if (u <= 2113662)
{
    u -= 16512;
    PutByte((byte) ((u & 0x7f) | 0x80));
    PutByte((byte) (((u >> 7) & 0x7f) | 0x80));
    PutByte((byte) ((u >> 14) & 0x7f));
}
else if (u <= 270549118)
{
    u -= 2113663;
    PutByte((byte) ((u & 0x7f) | 0x80));
    PutByte((byte) (((u >> 7) & 0x7f) | 0x80));
    PutByte((byte) (((u >> 14) & 0x7f) | 0x80));
    PutByte((byte) ((u >> 21) & 0x7f));
}
else
{
    u -= 270549119;
    PutByte((byte) ((u & 0x7f) | 0x80));
    PutByte((byte) (((u >> 7) & 0x7f) | 0x80));
    PutByte((byte) (((u >> 14) & 0x7f) | 0x80));
    PutByte((byte) (((u >> 21) & 0x7f) | 0x80));
    PutByte((byte) ((u >> 28) & 0x7f));
}

---

Decoding example:

int b = GetByte();
uint a = (uint) (b & 0x7f);
int extraCount = 0;

// first extra byte
if ((b & 0x80) == 0x80)
{
    b = GetByte();
    a |= (uint) ((b & 0x7f) << 7);
    extraCount++;
}

// second
if ((b & 0x80) == 0x80)
{
    b = GetByte();
    a |= (uint) ((b & 0x7f) << 14);
    extraCount++;
}

// third
if ((b & 0x80) == 0x80)
{
    b = GetByte();
    a |= (uint) ((b & 0x7f) << 21);
    extraCount++;
}

// fourth
if ((b & 0x80) == 0x80)
{
    b = GetByte();
    a |= (uint) ((b & 0x7f) << 28);
    extraCount++;
}

switch (extraCount)
{
    case 1:
        a += 128;
        break;
    case 2:
        a += 16512;
        break;
    case 3:
        a += 2113663;
        break;
    case 4:
        a += 270549119;
        break;
}
return a;

---

STRING BLOCK

Located immediately after the header.

A block of UTF-8 strings with length prefixes (in bytes). The entire block is compressed.

ushort length
byte[length] utf8data

---

NODE BLOCK

varint nameOffset — points to the start of a string in the string block
byte type — 0 for file, 1 for directory

type 0 (file):

varint data offset — from the start of the data block
varint data length

type 1 (directory):

varint node count
node[node count] children

---

DATA BLOCK

Located immediately after the node block.

Represents a binary blob.

This site is open source. Improve this page.