Re: [Pan-users] Pan 0.133 refuses to work any longer

[Mike Brown]

On Sat, Aug 25, 2012 at 08:50:37AM +0000, Duncan wrote:
> 1) It's really cool that, of all the apps you could choose to use for 
> that sort of news download volume, you choose to use pan.  There was a 
> time when it would have had a very hard time scaling, and it's great that 
> it not only can do so now, but that you find pan's UI usable and 
> efficient enough to chose it for the task.

I knew nothing about pan before I got it.  A user in a mail list that I am
on mentioned it for Linux users.  So I installed it.  I really do like it.
I don't really download the headers.  Was during the trouble-shooting.

> 2) I think 0.133's too old to have it, but I was /going/ to ask how you 
> like the new auto-"actions" that newer pan has now, that make it possible 
> to set pan up to auto-download new posts (not just headers) based on 
> score, before you've ever seen them.

I'd probably never use it.  The last thing I need is for it to start
downloading something I don't want, wasting precious time for stuff I do
want.

I use binsearch.info to find what I want and then create the nzb files.

That said, I did run across some stuff as a result of loading the headers
of a newsgroup.

> It seems to me, especially since you're moving so much data over a 
> /relatively/ narrowband (4 Mbit, but that's only ~1/2 a MByte/sec and you 
> just called over 11 gigs a day likely not enough, if my math's not wrong 
> that's ~6 hours daily of downloading at full tilt!) connection, that 
> especially if the downloads are predictable enough to score on, it would 
> be REALLY worth your while to get a newer pan with that feature, even if 
> you have to build it from sources yourself.  Then just leave the computer 
> and pan running, and let it download as the posts show up (0.133 can 
> already be set to grab headers as they show up, but not to grab full 
> posts), so they're already waiting for you when you get ready to do 
> whatever you'd do with them.

I get about 400-450 MBytes/sec download using 4 connections.

I've queued stuff such that the amount of time remaining was 4 days, i.e.,
non-stop.  Unless the computer shuts down.  New system and there is a
hardware issue I haven't determined yet.

I have to upgrade from Fedora 14 in order to get a newer pan.  Can't do source
as the developers have decided that a program should be used to determine if
what is needed is installed (pkg-config).  But, because not ALL of the needed
stuff has those entries, the configure script fails, even though what it wants
is there.  Pan really needs to drop the use of that program and use the old,
and proven, method of determining what is on the system, or not.

Maybe it'll compile under Fedora 16, but it certainly doesn't under Federa 14.
I have no clue how anyone has compiled this under Fedora, considering that
not all of the required entries exist for pkg-config.

As an example, the first thing configure dies on is GMIME.  I say first, but
because I can't get past that one, I do not know how many others there are.
It only takes one to make to compile not work.  And yes, the needed version
of GMIME is on my system.

> Of course if the downloaded posts are varied enough that they're not 
> easily pre-scored, that hurts, but I'd guess at least /some/ of it can be 
> automated.

There are some potential possibilities.  But, if pan is busy keeping an eye
on headers, it is eating up bandwidth in which I could be downloading. :-)

MB
-- 
e-mail: address@hidden | address@hidden            /~\ The ASCII
        address@hidden (140 char limit)       \ / Ribbon Campaign
Visit - URL: http://vidiot.com/                           X  Against
             http://vidiot.net/                          / \ HTML Email