Re: [PROPOSAL] Extended attributes for Posix security extensions

[Ragnar Kj�rstad]

On Tue, Oct 24, 2000 at 05:48:46PM +0100, Stephen C. Tweedie wrote:
> > I'm talking about inheritence of extended attributes, not ACLs.
> > 
> > I'm not sure, but I think it would be practical for the extended
> > attribute interface to implement some kind of inheritence. If this in
> > insufficent for ACLs; fine, the ACL module can just turn it off and
> > implement it itself.
> 
> The whole point of the EA API was to remove implementation details
> completely from the API!  We already have existing semantics for a lot
> of filesystems where inheritance is forced on us by the existing
> implementation: we *can't* enforce it in the API without breaking
> implementations.  This really ought to be a filesystem-specific thing,
> not an API thing.

It doesn't help the user that he can use the same interface to store
data in multiple filesystems if totally different things happen.

How can an application store data in extended attributes without knowing
if it's inherited or not?

At the very least the user need to know - of course even better if he
can choose.


-- 
Ragnar Kj�rstad
Big Storage
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [EMAIL PROTECTED]