[Pan-users] Re: upgrading from 0.11.4 to 0.14.0

[Duncan]

Anthony posted <address@hidden>, excerpted
below,  on Fri, 31 Oct 2003 01:12:29 -0800:

> I've just upgraded to SuSE 9.0, which comes with pan 0.14.0. I still have
> all my old configuration, and haven't yet run pan 0.14.0. What should I
> keep backups of?

Consider this a practical, but non-authoritative, answer.

If you have room, probably best would be to simply copy the entire ~/.pan/
dir to something like ~/.pan.bak/  If you don't have room and your cache
is large (the largest the GUI would take back then was 1 gig, but I ran
larger by directly entering 4 gig in the config file -- the max size from
the GUI now is 20 gig, per my request for a larger cache option, but I
still run 4 gig), process most of the current messages if necessary so a
smaller cache won't leave you frustrated, deleting most or all of the
messages, then shrink the cache accordingly, and THEN copy it.

A **LOT** has happened to PAN between the Gnome one 0.11.x version, and
the GTK+2 0.14.x.  I believe some of the config files changed format in
there somewhere, possibly twice.  Of course, there's the change in
Gnome/GTK+ libraries, also.  Thus, you may wish to write down your server,
group, and sig settings, entirely wipe the ~/.pan/ dir, and start over,
since a lot of it will be different anyway.  That would keep you from
possibly having an inconsistency in file format or something cause a
problem, and since so much of it will probably need reconfigured anyway,
you might as well.

0.14.0 was probably the most stable version since the switch to GTK2, so
that IS a good version to upgrade to if you don't upgrade often, as it
seems you don't.  Among the new features you may find useful..  

1.  Spell checking, which may or may not be enabled in SuSE's version by
default. 

2.  Article scoring, new in 0.14.0.  (One of the reasons it was so
stable was because there were more betas than normal in ordered to work
out the scoring kinks, which gave them time to work out a larger share of
the others as well.)  

3.  As mentioned, an option to run a larger cache than b4.

4.  New network handling, using gnet.  This initially fools some folks
into thinking PAN is only running one d/l connection at a time.  Rather,
the way it works now is that it puts all four (or as many as you have
configured for that server) connections to work on the SAME task, if there
are enough parts to the task to do so, of course.  When that task is
finished, it goes onto the next, and so on.  The biggest benefit here from
my perspective is that you don't have to queue up four separate tasks to
get PAN to use all four connections as you did b4.  Now, you just queue up
a single task, and PAN will put all connections to work on it.  The only
time you need to queue more tasks to get PAN to use more connections is if
you have two different servers.  You still have to queue the tasks for
each server separately in ordered to get PAN working on both servers at
once.  Still, having to manage two separate queues is far better than
having to manage EIGHT, so it's getting there.

5.  PAN is now available for MSWormOS (using the GTK+2 for MSWormOS
libraries).

Still on the roadmap..

1.  Binary attachment posting.

2.  Reworking the database backend, to use SQLite libraries, similar to
what they did for GNet with the networking.

3.  Prerequisite on the db rewrite, automatic server merges, so once
subscribed, PAN will fetch articles from groups on multiple servers
automatically, without the need for separate queues as now.

4.  A command line mode suitable for scripting PAN, so it can be run from
a cron job to d/l new messages and have them already pre-cached when you
fire PAN up to view them.

5.  Of course, complete documentation/help, and a bit of polish on the
interface, suitable to bring it to a 1.0 version.

-- 
Duncan - List replies preferred.   No HTML msgs.
"They that can give up essential liberty to obtain a little
temporary safety, deserve neither liberty nor safety." --
Benjamin Franklin