Re: [Pan-users] Trying to use 'new' Pan

[Duncan]

Maurice Batey posted on Tue, 16 Apr 2013 17:18:09 +0100 as excerpted:

> On Tue, 16 Apr 2013 15:48:06 +0000, Duncan wrote:
> 
>> it's possible to set a multi-gig cache and no expiry if you like,
> 
>   How does one do that?  (Can't see option anywhere.)

The expiry is as I said set per server, so it's in server prefs (the 
expire old articles dropdown).

The expiry setting, BTW, is one of several settings that can be set more 
precisely by directly editing the config file (servers.xml in this case), 
as opposed to setting it in the GUI.  In the config file, the expiry is a 
simple number of days and can be set to any reasonable (positive) 
integer.  But in the GUI, the dropdown contains only a relatively limited 
subset of the available choices, two weeks, 1/2/3/6 months, never.

Originally, Charles considered the cache size option too complex to 
bother ordinary users with, so it wasn't available to set in the new-pan 
GUI.  Advanced users who thought they needed to could still set it in the 
config file, however (preferences.xml, search for size).

But Charles stepped down as pan's lead dev some years ago, and after a 
couple years of basically being orphaned, there's now a new set of devs, 
Heinrich being the most active in the last couple years, adding a bunch 
of long missing features, etc.  And Heinrich has if anything the opposite 
approach, exposing all sorts of options that Charles considered too 
complex for ordinary users to deal with.  One of the first ones was the 
cache size.

Take a look in pan prefs, behavior tab, article cache section.  That's 
where the cache size is set, here, tho it can be noted that I'm running 
the git version, but 0.139 is pretty recent so I think it should have 
that setting.

As I said, I've been using a setting of several gigs for years.  For my 
text pan instance, I have a setting of 5 gigs (5120 MB), tho after 
several years actual usage is still under a gig (just, 900-something 
megs).

For my binary pan instance, I run a dedicated partition as cache, 12 gigs 
in size, with pan set to something slightly larger than that so it'll use 
the whole partition.

For binaries, unlike many I do NOT download straight to file.  I download 
to cache and go do something else (sleep, work...) for awhile, then come 
back when everything's available locally, and sort and save off what I 
want at that point, deleting messages as I either save permanently or 
decide I don't want their content after all.  I'll grab a few gigs worth 
and process it, ending up with empty groups as I've deleted all the 
headers I downloaded, then I'll go and either delete any stragglers still 
found in the dedicated partition, or simply mkfs it and start over again, 
ready to fill it up again.

Tho obviously at that size I'm not downloading /too/ many full size 
movies or TV programs... 12 gigs won't even fit a single full blu-ray, 
after all.  But people doing that tend to work with far fewer but far 
larger files, and downloading to saved file directly instead of to cache, 
is thus likely more efficient for them.  If they do use my cache-first 
method, they'd probably need a 100-gig cache at least, if not a dedicated 
TB+ cache drive...

-- 
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