ACL and RSYNC and -a

[Marc Perkel]

Wayne Davison wrote:

>
>Given a better implementation of the --acls option (i.e. one that
>doesn't complain if a source disk doesn't support ACLs), that might
>make sense.  Another thing to keep in mind, though, is that rsync
>always passes the individual options to the remote rsync, so adding
>new options to -a makes it inconvenient to communicate with a variety
>of rsync versions.  We'll need to balance these issues whenever ACL
>support gets officially included in rsync.
>
>  
>
Yes - ok - so ACLs aren't "officially" there yet. That's actually good 
in a way. Means there's room to do it right.

Ultimately this is an option that should "just work" and be transparent 
as part of permissions. As to communicationg with a variety of rsync 
version perhaps there should be some initial communication about 
capabilities at the initial connection and adjust acordingly. Older 
versions would be assumed to not have ACL capability and that data 
wouldn't be sent. Or if the receiving file system doesn't support ACLs 
then that data is silently removed.
-------------- next part --------------
HTML attachment scrubbed and removed