GRAF:Act Of War DAT

DAT

Format type: Archive
Endianness: Little-endian

Format Specifications

char {4} - Header (edat)

uint32 {4} - Length of header after the 10 nulls (18)
uint32 {4} - Unknown
byte {10}    - null
uint16 {2} - Unknown (1)
byte {1}     - null
uint32 {4} - Directory Offset
uint32 {4} - Directory Length
byte {7}     - null

// For each file

byte {X} - File Data

// Directory

// for each group

uint32 {4} - Group/File Indicator (0=file, #=length of this group entry, -#=offset to a previous file entry)

if (group){

uint32 {4} - Length of this group, including the file entries (ie relative offset to next group or null)

char {X} - Group Name

byte {X} - null Group Name Terminator

byte {0-1} - null Padding to a multiple of 2 bytes (ie only exists if groupnameLength+1 is odd)

}

else if (file){

uint32 {4} - Last File Indicator (0=last file in this group, #=length of this file entry)

uint32 {4} - File Offset

uint32 {4} - File Length

byte {1} - Unknown (0) (something to do with the Filename position?)

char {X} - Filename

byte {X} - null Filename Terminator

byte {0-1} - null Padding to a multiple of 2 bytes (ie only exists if filenameLength+2 is odd)

}

else if (previous offset (negative number)){

uint32 {4} - Last File Indicator (0=last file in this group, #=length of this file entry)

uint32 {4} - File Offset

uint32 {4} - File Length

uint16 {2} - null

}

Notes and Comments

The files are stored in a special way so as to reduce the amount of space needed to store the filenames. It does this by grouping the files together that share a common first-part of their name, and then building up the actual filename of each file by joining a whole lot of group names together.

For example, if we have 2 files called "sound11.wav" and "sound22.wav", you can see that they both share 5 common letters at the beginning of the filename (ie "sound"). This archive would have a group with the name "sound" and each of the 2 files would be saved with names "11.wav" and "22.wav" within that group, thus saving space. This space saving is only really benificial for storing directorie names though, as there is a lot of overhead.

Groups can be nested inside each other, so there is no boundary as to the number of groups that can exist in a nested way.

Also, filenames can be reused by using a negative number for the Group/File ID. In this case, take the offset just before the negative Group/File ID, subtract the negative number, and it will point to the filename that will be reused.