ACL and RSYNC and -a

[Marc Perkel]

Paul Slootman wrote:

>On Thu 11 Aug 2005, Marc Perkel wrote:
>
>  
>
>>Forgive me if this has already been asked but - shouldn't the '-a' 
>>switch include -X and -A for ACLs and extended attributes?
>>
>>I thought that -a means everything - so when everything grows -a should 
>>grow with it.
>>    
>>
>
>-a does not mean everything... e.g. it does not include -H as that
>incurs a lot more processing. -X and -A aren't (yet?) understood by
>standard rsync versions, so I don't think it's useful to include those
>in -a yet.
>
>
>Paul Slootman
>  
>

Paul,

I can appreciate the difficulty of figuring out what to include in -a and what not to include. Perhaps if ACLs were more widely used and common it would be easier. In my case just starting to use them I have to go back and add -A -X to all my scripts now. In contrast cp -a just works without me having to change anything.

I also have a netware background and I'm used to having fine grained permissions. In fact when I started in Linux it was shocking that Linux permission were so crude. In fact - even with ACLs it still doesn't match where Netware was 15 years ago. But - it's a giant step forward.

Having said that - maybe there could be a way with environment variables to set what -a means. But it seems to me that if -a includes some permissions and not others that is logically inconsistent. In the future I'm hoping that people realize the power of ACLs and it becomes standard.


-------------- next part --------------
HTML attachment scrubbed and removed