Re: tar problem (duplicate files)

[Stepan Kasal]

Hello,

On Mon, Jun 23, 2003 at 10:57:21AM +0200, Dario Spagnolo wrote:
> this morning I ran into a really strange problem. I'm not sure it's
> a bug, but this list seemed the most appropriate.

well, if it were bug, this list would be the right one.
As it is not, we might say that comp.unix.shell is a better group
for this.  But you didn't know.

> $ find /home/mm -size -100k > mm-files
> $ tar -cvf mm.tar -T mm-files
> $ tar -tvf mm.tar | grep "blog/s4.zip"
> -rwxrwxrwx mm/ftp_only    23233 2003-04-05 14:04:10 home/mm/blog/s4.zip
> -rwxrwxrwx mm/ftp_only    23233 2003-04-05 14:04:10 home/mm/blog/s4.zip
> -rwxrwxrwx mm/ftp_only    23233 2003-04-05 14:04:10 home/mm/blog/s4.zip

The explanation is surprisingly simple: the list contains names of
directories, which tell tar to pack the whole tree.

In fact, you did something like this:
        tar -cvf mm.tar home/mm home/mm/blog home/mm/blog/s4.zip

One of possible ways to fix this is to ignore dirnames from backup.
For example this command:
        find /home/mm -type f -size -100k > mm-files

ignores everything except regular files (it ignores also symlinks).

Hope this explains it,
        Stepan Kasal