0.6.0 released

[Ben Escoto]

--==_Exmh_-2120795788P
Content-Type: text/plain; charset=us-ascii

>>>>> "JH" == Jamie Heilman <jamie@audible.transient.net>
>>>>> wrote the following on Fri, 15 Mar 2002 20:20:24 -0800

  JH> Just out of idle curriosity do you have a policy wrt protocol
  JH> changes and what versions will talk to what?  I haven't read the
  JH> page lately I'm just wondering if maybe that would be a good
  JH> thing to document.  I know rsync tends to maintain a certain
  JH> amount of backwards compatibility (where certain == some
  JH> undefined amount evidently - I'm having ugly problems with 2.3.x
  JH> talking to 2.5.x) but I seem to recall a more stringent version
  JH> homogeny in rdiff-backup being mentioned a few messages back.

I suppose this matters because it is a pain for people to upgrade with
every new release; instead they upgrade every so often, after a fixed
amount of time instead of a fixed release.  So what would be an
appropriate length of time, 4-6 months?  If we tried not to break
compatibility inside a stable series (e.g. 0.6.x) and then didn't
start a new stable series more often than every 4-6 months, that might
work out OK, and (in-)compatibility would be easy to figure out.

  JH> And in other random thoughts that would most certainly break
  JH> backwards compatibilty, have you considered TAI format for
  JH> datestamps?  They would solve some of the special character
  JH> issues across platforms (not that I advocate writing code with
  JH> Windows in mind, quite the opposite actually, I'm a total Unix
  JH> snob).  http://cr.yp.to/proto/utctai.html for more information
  JH> and resources; not sure if there's a python library for it.  I
  JH> dunno, maybe make it a switch or something - probalby just
  JH> useless bloat but I thought I'd throw it out there.

TAI vs UTC is interesting, and I had never thought of it before.  If
both the current two time format choices are unusuable on some
platform, then maybe a third option should be added.  But I'm not sure
the time on files like:

increments.867715200.dir

would be very easy to read for the average person.


--
Ben Escoto

--==_Exmh_-2120795788P
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Exmh version 2.5 01/15/2001

iD8DBQE8k6K2+owuOvknOnURAv5KAJ4lwk4S16T3MguPhQIesBQqXBEaIgCdHCBa
3W7c4rzzhILeB5Q/YNA5p+E=
=vsTe
-----END PGP SIGNATURE-----

--==_Exmh_-2120795788P--