On Fri, Sep 30, 2016 at 02:02:37PM +0200, Ludovic Court??s wrote:
> John Darrington <j...@gnu.org> skribis:
>
> > Another draft for review ...
>
> Could you please include an iteration number in the subject line, and a
> terse summary of the changes compared to the previous iteration?
>
> That would be greatly helpful???I???m getting lost in a maze of unrelated
> patch series and sometimes have a hard time remembering where we are and
> what it is that I???m doing here. ;-)
>
OK
> Please also mention the idmap things, the doc/guix.texi changes, etc.
OK
> ???The @dfn{global security system} (GSS) daemon provides ??????
OK
>
> > +(define-record-type* <pipefs-configuration>
> > + pipefs-configuration make-pipefs-configuration
> > + pipefs-configuration?
> > + (mount-point pipefs-configuration-mount-point
> > + (default default-pipefs-dir)))
>
> Seems to me we don???t even need <pipefs-configuration>; a string would be
> enough, no?
We could. But then if somebody wanted to mount it in a non-standard
place they wouldn't be able to override that from /etc/config.scm
> Overall this seems to be almost ready, no? Since this is a pretty
> involved service composition, I think it would be fruitful in the future
> to add a full test case in (gnu tests nfs) where we would export an NFS
> tree and mount it.
I agree. However that will mean having to first implement the server side
NFS stuff which is a whole other can of worms....
J'
