Dave's Brain

Browse - Programming Tips - What's the format of a tar file?

Date: 2006sep30
OS: *nix

Q. What's the format of a tar file?

A. There is no header for the entire file it.
The format is:

[Header for file 1] [Contents of file 1] [Header for file 2] [Contents of file 2] ... [EOF marker]
This means the only way to see how many files are in a tar file is to read the entire file. The header is always 512 bytes. Unused area at the end (of the header) is padded with zeros. The EOF marker is two 512 blocks of zero (ie 1024 zeros). The file header is this:
typedef struct { char name[100]; char mode[8]; char owner_id[8]; char group_id[8]; char size[12]; char modtime[12]; char checksum[8]; char link; // '5' means directory char linked_file[100]; } CLASSIC_TAR_HEADER; typedef struct { char ustar[6]; char version[2]; char owner[32]; char group[32]; char major[8]; char minor[8]; char prefix[155]; } USTAR_HEADER; typedef struct { CLASSIC_TAR_HEADER; USTAR_HEADER; } TAR_HEADER;
Notice there is only one time: modify time. No create or access times. There are pretty tight restrictions on length of names. More info: http://en.wikipedia.org/wiki/Tar_%28file_format%29

Add a comment

Sign in to add a comment
Copyright © 2008-2017, dave - Code samples on Dave's Brain is licensed under the Creative Commons Attribution 2.5 License. However other material, including English text has all rights reserved.