pan-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Pan-users] Article Retrieval Issue - Pan will not download headers


From: Duncan
Subject: Re: [Pan-users] Article Retrieval Issue - Pan will not download headers from one group
Date: Wed, 31 Aug 2016 03:30:09 +0000 (UTC)
User-agent: Pan/0.141 (Tarzan's Death; GIT 04f6c8932)

Ted Sone posted on Mon, 29 Aug 2016 11:00:23 -0400 as excerpted:

> Long time Usenet User, (circa 1990), short term Pan User, (circa last
> Saturday).
> 
> I am utilizing the WIndows version and it seems to run fine.  My primary
> reason for using Pan is to poll newsgroups as I already have Newsbin to
> manage the actual downloads.  Newsbin is great for downloading but it is
> not great for seeing what is in the group whereas Pan is similar to the
> Forte Agent style which provides a great interface for seeing what is in
> the Newsgroup.
> 
> I subscribed to a couple of newsgroups and one was quite small and it
> works fine.  The second one, (alt.binaries.tv), is quite large and Pan
> crashed during the original header download. Since that time Pan no
> longer downloads anything from that group.  I have uninstalled Pan,
> deleted the associated directories etc. and then upon re-installation
> Pan continues to ignore the group.  Clearly Pan is obtaining historical
> information from a file that is not within the User profile.  Anyone
> have any thoughts as to where that file might be?

[As anyone on the various lists and newsgroups I participate in can tell 
you, I tend to be quite helpful, in part because I often take the time to 
explain stuff others just give a rote answer to, but that means I often 
take the long way around.  So I have here, there's an answer below, but 
some explanation to get to it first...  Patience...]

FWIW I personally don't... and can't... do proprietaryware, including 
both MS and Apple proprietary platforms, because I do not and cannot 
agree to the EULA; among other things I cannot agree to waive my rights 
to damages should code I don't have the right to examine do damage to 
something, and of course that's exactly what most common licenses, 
including many free/libre software licenses (minus the can't examine 
part), require -- the difference being that with free software, you can 
actually examine what the code does before agreeing to the waiver, it's 
not a black-box they want to require that I agree to the responsibility 
for damages for, without letting me see what's in the box so I can 
actually know what the code does that I'm agreeing to waive my damages 
rights on.

Not that I'm a big coder or anything, but I can let those I trust with 
more expertise in the area examine it for me, too, which is where Linux 
and free software distros come in.

Tho it wasn't always that way.  I ran on MS (and was even quite active in 
the IE/OE betas and newsgroups) until they introduced the remote-lock 
malware in eXPrivacy, as I could predict where it was headed (where 10 
actually ended up!), and paying for malware that had no purpose but to 
thwart the user was a line I simply wasn't going to cross.  Which of 
course left me with two choices, staying with MS and pirating it, or 
staying legal with Linux and free software, which actually respects user 
rights.  So you see, it was actually MS that pushed me to Linux. =:^)

Tho I recognize it as a personal decision, and wouldn't dream of 
demanding anyone else conform to what I've found works best for me.  
Indeed, it took me two years to switch over after I had pretty much 
decided I probably would, eventually, and were someone to have tried to 
push me too hard during that time, all it would have accomplished is 
making it take longer, and perhaps changing the outcome to me staying on 
MS, even if I had to pirate it to avoid the eXPrivacy malware.  So one 
thing I do NOT do is try to push someone who's not ready to take that 
step on their own.

Which being the helpful sort of guy I try to be, means I'll even help you 
try to work thru problems on the platform, however personally distasteful 
for me it might be.  Tho there's a pretty strict limit -- I don't and 
won't spend time learning how newer MS works, because it's dead history 
to me, and I don't actually work on other people's MS based systems any 
more either, but if I can help within those limits without spending time 
on or researching the platform, I often will.

Anyway... all that to say that I won't, indeed, can't, touch pan for MS, 
since I legally can't do MS platform anything on my systems, as long as 
it remains proprietary, even were I to want to for some reason, and thus 
don't have a clue what pan's layout is there.

What I /can/ tell you, however, is what the file you're probably looking 
for is, and where it'd be located by default on freedomware Linux.  Which 
should help you find it on MS as well.

The file you're looking for should be a newsrc style file.  There should 
be one for each server you have configured, and you can find the specific 
names by looking at the server entries in the servers.xml file in pan's 
data dir ($HOME/.pan2/servers.xml on Linux/Unix style filesystem layouts).

These newsrc files track the messages pan has already seen on each group 
on each server, using the per-server, per-group message numbers (as 
opposed to the globally universal message-ids that nzb files and pan's 
cache use).  Because these numbers are unique to each server and group, 
the newsrc files are only valid for a single server, tracking all groups 
you've visited that said server carries, whether you're subscribed to the 
group, or not.

The way it normally works is this.  When your client (pan, for us) 
fetches new messages, it queries the server asking for the messages 
available in each group.  The server returns three numbers, the lowest 
available numbered message (the low water mark, normally the oldest since 
the numbers are more or less sequential), the highest numbered message 
(the high water mark, normally the newest), and an estimate of the number 
of messages available between these numbers.  (Due to message cancels and 
various other reasons, some messages within the range between low-water 
and high-water may not be available, so a client can't simply do the 
math, subtracting low from high to get the number available, tho that's 
one way the server can arrive at its estimate if it wants.)

The idea is that a client can compare the given numbers to those it knows 
it has already seen, and by doing so, know if there's any new messages -- 
if the highest message number reported by the server is higher than the 
highest message number the client already knew about.

It's these numbers that are tracked in the newsrc file, and these numbers 
that can get out of whack if the server "resets" and starts over with its 
numbering, thus causing the client to think that it has seen all the new 
messages as the number the client has is far larger than what the server 
has right after such a reset.

So it's not unusual to have to reset a group, or even all groups for a 
particular server, if the server resets, or if the newsrc file otherwise 
(as in a crash, or if you switch servers and just change the existing 
server info so it uses the same newsrc, instead of creating a new server 
and thus create a new newsrc for it) somehow gets out of sync with the 
server.  You can reset all groups by simply deleting the newsrc file for 
that server, or just one group by simply deleting just the line for that 
group, or better yet, only the numbers recorded after the group name and 
separator, leaving that (which some clients use to track subscribed 
groups, but pan doesn't as it tracks that at a level above the individual 
server level) there.

Of course you can edit the number ranges yourself, too, if desired, to be 
a bit more efficient, perhaps, if there's a bunch of skipped numbers and 
therefore a choppy bunch of ranges instead of a nice single range, lowest 
to highest.

A google (on Google itself, or other googling engines) has some excellent 
resources with more info on the somewhat standardized newsrc file format, 
should you wish more detail.  =:^)

Now you just have to either find that file yourself, or wait for someone 
with a bit more knowledge of the pan on MS layout to provide the layout-
specific path, and you should be able to edit it as necessary using a 
standard text editor. =:^)

-- 
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




reply via email to

[Prev in Thread] Current Thread [Next in Thread]