[Pan-users] Re: 0.129 - a few minor issues

[Duncan]

Jim Henderson <address@hidden> posted
address@hidden, excerpted below, on  Thu, 10 May 2007 19:05:38
+0000:

> I also am trying to figure out how to get it to pre-cache the articles
> it downloads - I use that a lot in 0.14.x, but don't want to set up
> leafnode to do this.  The article cache works fine for my purposes, but
> how do I tell it to prefetch the bodies?

pan doesn't currently have fully automated body downloads, as was 
possible using rules in old-pan.  However, there are two possible current 
solutions, depending on your needs.  You mentioned one, leafnode, or to 
make it generic, running a local news server.  That's the fully automated 
solution, but may not be what you want.

My solution isn't as automated, but works better for me.  I use a larger 
local pan cache, 12 gigs on a separate partition (symlinked), for my 
binary pan instance (see the multiple pan instance discussion on IIRC my 
second post back).  I then do my file saves in two steps, downloading 
headers and setting up the download to cache right before I go to work or 
to bed or whatever, then later, sorting thru what has been downloaded, 
and deleting or saving it as appropriate, then.  This way, I get to save 
the files to a permanent or semi-permanent location while I still have 
all the info in the post headers (subject, poster, date posted, etc.) 
available to me, and can use this to sort the files I save into a more 
descriptive location by subdir.  Folks that just download and decode/save 
everything directly to a scratch dir, without saving to cache and sorting 
thru it later, generally don't get the benefit of all that extra metadata 
that came in the post.

To do this, you'll need to increase your cache size beyond pan's default 
10 MB.  Charles obviously considers this setting, along with a few 
others, advanced enough that he doesn't want to bother confusing newbie 
users with it in the GUI config dialogs.  Thus, while pan honors the 
setting as exposed in its config files, it's not exposed in the GUI 
config, and to set it you must edit the appropriate config file 
directly.  In this case, the setting is in preferences.xml, IIRC as cache-
size-megs, but something similar anyway.  I always just search on cache, 
and hit next until I see the appropriate description.

Some other "advanced" settings include the PAN_HOME environmental 
variable, which can be used to point pan at a location other than the 
default ~/.pan2 (and thus to run multiple pan instances, separated by 
purpose, as I do and explain elsewhere), and the connections per server 
and per server expiration settings in servers.xml.  As it happens, 
connections per server in particular is a GNKSA qualifier, where the 
maximum is four.  To maintain GNKSA compliance, the GUI setting maxes at 
four, but certain paid servers for instance allow up to 8 connections, 
and as quite a number of folks have requested being able to set more 
connections over the years, with the rewrite, Charles exposed the setting 
in the config file itself, for those advanced users that want to use it.  
No longer is it necessary to directly patch the sources as desired and 
compile your own copy. =8^)  On the expire settings, the servers.xml 
value is in days.  I have no idea why Charles didn't make the GUI config 
widget for it a spinbox, but he obviously thinks the handful of specific 
options in a dropdown is simpler, so configuring a setting outside that 
limited handful is an "advanced" option, exposed only to those willing to 
edit the config files directly.  Still, as long as those options remain, 
if only by editing the config files directly, fine by me.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman