[PATCH][WIP] Make vfs_acl_xattr use hash of the posix ACL

[Andrew Bartlett]

On Tue, 2012-10-16 at 09:47 +0200, Christian Ambach wrote:
> Hi Andrew,
> 
> On 10/12/2012 10:59 PM, Andrew Bartlett wrote:
> 
> >
> > I think you did.  I hope what I've said above helps explain it
> > better, but the actual implementation module will be on the list
> > again for all to review once I sort out the issues with the previous
> > patch and those that cause this work in progress to fail.
> 
> If this is still work in progress (and it is because there has not been
> an agreement yet if this approach is right or not), I wonder why parts
> of it already hit master? Such work should live in a private branch
> until it has been accepted and then be pushed to master. IMHO, master is
> not a place for half-baked ideas.

At the time, I wasn't aware of any of the concerns Simo has raised.  So
far, I've only implemented the parts that were discussed on the mailing
list over the past months, particularly when we discussed the issue in
May.  

The approach was accepted then (the ACL-blob design was proposed), and
again just before RC1 when I was able to get the VFS layer changes
merged. 

I'm quite happy with that state of the proposal, and absent Simo's
concerns (which I've been unable to pin down into changes I can make), I
would be proceeding to finish the details of the vfs_acl_common changes
and build the testsuite that it requires. 

> I know that you have asked for feedback on your private branch before
> but did not get much of attention. However, I do not see this as a
> permission to push it.

In case it was missed, I got an ACK from the maintainer of the smbd file
server ACLs.  I missed that Simo had concerns beyond formatting, but we
have already covered that ground, and I'm not sure what more is to be
said here. 

Despite this, and probably due to the issues on the gpfs side, I seem to
have annoyed you significantly with this patch series, and I'm sorry
about that.  

In all the work I do on Samba, like others I have to make small and
large changes in preparation for the actual changes I desire to
implement, and generally implement these (with review where appropriate)
in advance of the actual change I want.  Typically (but not in this
particular case) these changes provide some incremental benefit on their
own.  

Occasionally these ideas don't work out, and I've been involved in
removing some failed experiments that didn't develop into useful
features.

That said, I'm really not sure what more you want me to do from here.  

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org