> while i believe a free afs client could be built from spec, you need
> a whole lot more: bos, vos, pts, fs, kas, klog, etc. i think it's too
> big a job.
Given a freely-available Rx, it would be a lot easier to build a free
client from spec than it would have been in the past.
As for other tools...
bos, vos, kas, and perhaps others are really only of use to cell
administrators, who presumably have Transarc's server software. Even
so, they might be of some use to users, and equivalent tools could be
eventually built from spec without too much trouble. The same goes for
pts - most of these are really just wrappers around RPC's, which are
documented somewhere.
fs is nothing more than a user-interface to various routines that are
actually implemented by the cache manager. Implementing some of this
functionality would be part of a basic client, and again, not too hard.
klog is unnecessary - use kinit (which gets Kerberos tickets), and aklog
(which obtains an AFS ticket and generates a token for the cache manager).
We already do this, even though we use stock AFS for everything except
authentication.
The truly hard part is not really a technical issue at all. Those of
us who know enough to do a free client wouldn't be able to, because most
of us have AFS source access, which would make it hard to claim that
the new client had been written without reading and/or using Transarc's
code...
-- Jeffrey T. Hutzelman (N3NHS) <[EMAIL PROTECTED]>
Systems Programmer, CMU SCS Research Facility
Please send requests and problem reports to [EMAIL PROTECTED]
