Re: Re: Dar and OS X xattrs: Willing to help

On Sep 25, 2005, at 9:55 AM, Denis Corbin wrote:
> I agree with this way of doing, but to avoid loop problem or  
> limitations
> (like a read-only filesystem), the convertion from ACL to text must be
> done in memory, no temporary file must be used.

Yes, that's perfectly possible, none of the acl_*() functions require  
a temp file. Probably I would just modify the OSX_EA_SUPPORT  
implementations of my_l*xattr() so that they pretend as if ACLs were  
a real xattr.

However, some people on #opendarwin are now telling me that the ACL  
support on Mac OS X is a bit flaky. I've already experienced a  
strange bug where enabling ACLs on my filesystem causes xattrs to not  
work properly. So, given that I can't really use ACLs myself, and  
that very few other Mac users actually use them either, I don't have  
motivation to work on supporting them at this time. Apple is expected  
to come out with a fairly large bugfix release (10.4.3) reasonably  
soon; if it fixes the ACL issues I'll revisit this.

Thanks again for all your work on dar, now I have to go create my new  
backup strategy :-)

Dave