[660] in athena10
Re: attach -e
daemon@ATHENA.MIT.EDU (Greg Price)
Tue Dec 2 16:34:05 2008
Date: Tue, 2 Dec 2008 16:33:20 -0500
From: Greg Price <price@MIT.EDU>
To: Evan Broder <broder@mit.edu>
Cc: Anders Kaseorg <andersk@mit.edu>, Bill Cattey <wdc@mit.edu>,
Jonathan Reed <jdreed@mit.edu>, athena10@mit.edu
Message-ID: <20081202213317.GI10836@vinegar-pot.mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4935A807.3000002@mit.edu>
On Tue, Dec 02, 2008 at 04:26:31PM -0500, Evan Broder wrote:
> This should all work about the way it did before if I implement my pony
> - our hacked attach and detach, even though they don't directly
> manipulate /mit, still change the attachtab, so if pyHesiodFS uses the
> attachtab as its cache, then attach and detach start working the same
> way they did again.
Yuck. So now `detach gnu` would actually remove /mit/gnu from my
namespace? Then we'd be back to the bad old days of having to attach
everything before relying on it being around, rather than using /mit/
as a namespace that always works. That's more complicated to explain
to users and a source of bugs in scripts.
I like what Anders said better -- if you want to take something out of
your PATH, then take it out of your PATH. We need ln -s or attach -e
so power users can maintain replicated lockers, but it should be
something that a power user has to choose to do.
Greg
