RAR version 4.20 - Technical information
               RAR version 4.20 - Technical information


                         RAR archive file format

   Archive file consists of variable length blocks. The order of these
blocks may vary, but the first block must be a marker block followed by
an archive header block.

   Each block begins with the following fields:

HEAD_CRC       2 bytes     CRC of total block or block part
HEAD_TYPE      1 byte      Block type
HEAD_FLAGS     2 bytes     Block flags
HEAD_SIZE      2 bytes     Block size
ADD_SIZE       4 bytes     Optional field - added block size

   Field ADD_SIZE present only if (HEAD_FLAGS & 0x8000) != 0

   Total block size is HEAD_SIZE if (HEAD_FLAGS & 0x8000) == 0
and HEAD_SIZE+ADD_SIZE if the field ADD_SIZE is present - when
(HEAD_FLAGS & 0x8000) != 0.

   In each block the followings bits in HEAD_FLAGS have the same meaning:

  0x4000 - if set, older RAR versions will ignore the block
           and remove it when the archive is updated.
           if clear, the block is copied to the new archive
           file when the archive is updated;

  0x8000 - if set, ADD_SIZE field is present and the full block
           size is HEAD_SIZE+ADD_SIZE.

  Declared block types:

HEAD_TYPE=0x72          marker block
HEAD_TYPE=0x73          archive header
HEAD_TYPE=0x74          file header
HEAD_TYPE=0x75          old style comment header
HEAD_TYPE=0x76          old style authenticity information
HEAD_TYPE=0x77          old style subblock
HEAD_TYPE=0x78          old style recovery record
HEAD_TYPE=0x79          old style authenticity information
HEAD_TYPE=0x7a          subblock

   Comment block is actually used only within other blocks and doesn't
exist separately.

   Archive processing is made in the following manner:

1. Read and check marker block
2. Read archive header
3. Read or skip HEAD_SIZE-sizeof(MAIN_HEAD) bytes
4. If end of archive encountered then terminate archive processing,
   else read 7 bytes into fields HEAD_CRC, HEAD_TYPE, HEAD_FLAGS,
5. Check HEAD_TYPE.
   if HEAD_TYPE==0x74
     read file header ( first 7 bytes already read )
     read or skip HEAD_SIZE-sizeof(FILE_HEAD) bytes
     if (HEAD_FLAGS & 0x100)
       read or skip HIGH_PACK_SIZE*0x100000000+PACK_SIZE bytes
       read or skip PACK_SIZE bytes
     read corresponding HEAD_TYPE block:
       read HEAD_SIZE-7 bytes
       if (HEAD_FLAGS & 0x8000)
         read ADD_SIZE bytes
6. go to 4.

                               Block Formats

   Marker block ( MARK_HEAD )

HEAD_CRC        Always 0x6152
2 bytes

HEAD_TYPE       Header type: 0x72
1 byte

HEAD_FLAGS      Always 0x1a21
2 bytes

HEAD_SIZE       Block size = 0x0007
2 bytes

   The marker block is actually considered as a fixed byte
sequence: 0x52 0x61 0x72 0x21 0x1a 0x07 0x00

   Archive header ( MAIN_HEAD )

HEAD_CRC        CRC of fields HEAD_TYPE to RESERVED2
2 bytes

HEAD_TYPE       Header type: 0x73
1 byte

HEAD_FLAGS      Bit flags:
2 bytes
                0x0001  - Volume attribute (archive volume)
                0x0002  - Archive comment present
                          RAR 3.x uses the separate comment block
                          and does not set this flag.

                0x0004  - Archive lock attribute
                0x0008  - Solid attribute (solid archive)
                0x0010  - New volume naming scheme ('volname.partN.rar')
                0x0020  - Authenticity information present
                          RAR 3.x does not set this flag.

                0x0040  - Recovery record present
                0x0080  - Block headers are encrypted
                0x0100  - First volume (set only by RAR 3.0 and later)

                other bits in HEAD_FLAGS are reserved for
                internal use

HEAD_SIZE       Archive header total size including archive comments
2 bytes

RESERVED1       Reserved
2 bytes

RESERVED2       Reserved
4 bytes

   File header (File in archive)

HEAD_CRC        CRC of fields from HEAD_TYPE to FILEATTR
2 bytes         and file name

HEAD_TYPE       Header type: 0x74
1 byte

HEAD_FLAGS      Bit flags:
2 bytes
                0x01 - file continued from previous volume
                0x02 - file continued in next volume
                0x04 - file encrypted with password

                0x08 - file comment present
                       RAR 3.x uses the separate comment block
                       and does not set this flag.

                0x10 - information from previous files is used (solid flag)
                       (for RAR 2.0 and later)

                bits 7 6 5 (for RAR 2.0 and later)

                     0 0 0    - dictionary size   64 KB
                     0 0 1    - dictionary size  128 KB
                     0 1 0    - dictionary size  256 KB
                     0 1 1    - dictionary size  512 KB
                     1 0 0    - dictionary size 1024 KB
                     1 0 1    - dictionary size 2048 KB
                     1 1 0    - dictionary size 4096 KB
                     1 1 1    - file is directory

               0x100 - HIGH_PACK_SIZE and HIGH_UNP_SIZE fields
                       are present. These fields are used to archive
                       only very large files (larger than 2Gb),
                       for smaller files these fields are absent.

               0x200 - FILE_NAME contains both usual and encoded
                       Unicode name separated by zero. In this case
                       NAME_SIZE field is equal to the length
                       of usual name plus encoded Unicode name plus 1.

                       If this flag is present, but FILE_NAME does not
                       contain zero bytes, it means that file name
                       is encoded using UTF-8.

               0x400 - the header contains additional 8 bytes
                       after the file name, which are required to
                       increase encryption security (so called 'salt').

               0x800 - Version flag. It is an old file version,
                       a version number is appended to file name as ';n'.

              0x1000 - Extended time field present.

              0x8000 - this bit always is set, so the complete
                       block size is HEAD_SIZE + PACK_SIZE
                       (and plus HIGH_PACK_SIZE, if bit 0x100 is set)

HEAD_SIZE       File header full size including file name and comments
2 bytes

PACK_SIZE       Compressed file size
4 bytes

UNP_SIZE        Uncompressed file size
4 bytes

HOST_OS         Operating system used for archiving
1 byte                 0 - MS DOS
                       1 - OS/2
                       2 - Windows
                       3 - Unix
                       4 - Mac OS
                       5 - BeOS

FILE_CRC        File CRC
4 bytes

FTIME           Date and time in standard MS DOS format
4 bytes

UNP_VER         RAR version needed to extract file
1 byte
                Version number is encoded as
                10 * Major version + minor version.

METHOD          Packing method
1 byte
                0x30 - storing
                0x31 - fastest compression
                0x32 - fast compression
                0x33 - normal compression
                0x34 - good compression
                0x35 - best compression

NAME_SIZE       File name size
2 bytes

ATTR            File attributes
4 bytes

HIGH_PACK_SIZE  High 4 bytes of 64 bit value of compressed file size.
4 bytes         Optional value, presents only if bit 0x100 in HEAD_FLAGS
                is set.

HIGH_UNP_SIZE   High 4 bytes of 64 bit value of uncompressed file size.
4 bytes         Optional value, presents only if bit 0x100 in HEAD_FLAGS
                is set.

FILE_NAME       File name - string of NAME_SIZE bytes size

SALT            present if (HEAD_FLAGS & 0x400) != 0
8 bytes

EXT_TIME        present if (HEAD_FLAGS & 0x1000) != 0
variable size

other new fields may appear here.

                              Application notes

   1. To process an SFX archive you need to skip the SFX module searching
for the marker block in the archive. There is no marker block sequence (0x52
0x61 0x72 0x21 0x1a 0x07 0x00) in the SFX module itself.

   2. The CRC is calculated using the standard polynomial 0xEDB88320. In
case the size of the CRC is less than 4 bytes, only the low order bytes
are used.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License