BIN file format

BINs are files used by the game client. They contain typed structured data.

BIN header

The BIN file starts with the following header:

Pos Size Format Description
0 4 PROP magic code
4 4 u32 version

Parsing of following data depends on the version.

Starting from version 2, the header is followed by links to other BIN files (a list of their path):

Pos Size Format Description
0 4 u32 linked files count
4 ?×n string linked files paths

After linked files (or version for older BIN files) are the entry types:

Pos Size Format Description
0 4 u32 entry count
4 4×n u32 entry types

BIN entries

Entry header:

Pos Size Format Description
0 4 u32 entry length
4 4 u32 entry hash
8 4 u16 field count

The header is followed by field data as defined here.

Data types

BIN data is structured and each value is typed. Some types are nested and can contain other values.

The following types are defined:

Value Format Size Nested Description
0 empty 6
1 bool 1
2 s8 1
3 u8 1
4 s16 2
5 u16 2
6 s32 4
7 u32 4
8 s64 8
9 u64 8
10 float 4
11 vec2(float) 8
12 vec3(float) 12
13 vec4(float) 16
14 matrix4x4 64
15 rgba 4 RGBA byte values
16 string 2+n 2-byte size followed by content
17 hash 4 BIN hash value
18 container > 9 ✓ array of values of the same type
19 struct > 10 ✓ structure with named fields
20 embedded > 10 ✓ structure with named fields
21 link 4
22 option 2+n ✓ an optional value
23 map > 10 ✓ key/value pairs
24 flag 1 boolean used as a flag

Note: Value size is either fixed (e.g. numeric types) or provided in the header (e.g. structs). The only exceptions are nested strings whose size cannot be inferred from the header.

Structs and embeddeds

Structs and embeddeds store their values in fields. Each field of is associated to a type and a hash (based on its human-readable name). Fields are optional.

Note: Structs and embeddeds work the same. The difference is internal: structs are stored as fields and embeddeds are pointers to static data.

Data starts with the following header:

Pos Size Format Description
0 4 u32 struct/embedded name hash
4 4 u32 data size (remaining bytes)
8 2 u16 field count

Note: if the name hash is 0, struct is void and ends right after the name hash.

Header is followed by field data. Each field is introduced by a small header followed by field value data. Field header format is describe below:

Pos Size Format Description
0 4 u32 field name hash
4 1 u8 field type

Other nested types

Containers, options and maps are nested types for which all elements are sequenced and have the same type (provided in the header).

Container

Pos Size Format Description
0 1 u8 value type
1 4 u32 data size (remaining bytes)
5 4 u32 value count

Option

Pos Size Format Description
0 1 u8 value type
1 1 u8 1 if option has a value, 0 otherwise

Map

Pos Size Format Description
0 1 u8 key type
1 1 u8 value type
2 4 u32 data size (remaining bytes)
6 4 u32 key/value count

BIN hashes

BIN files use 32-bit hashes to identify fields and reference some data instances.

An FNV-1a hash is applied to the lowercased string value.

Patch files

Patch files are used to override specific BIN entries. Their format is identical to BIN files, except for an additional header, before the usual PROP magic code.

Pos Size Format Description
0 4 PTCH magic code
4 4 u32 unknown (always 1)
8 4 u32 unknown (always 0)
12 4 PROP BIN file magic code followed by usual data
This page is editable on GitHub. Found any mistakes or want to improve this page? Click here to edit this page.