Re: [Pan-users] Google groups users breaking threads

[Duncan]

Derek Turner posted on Fri, 10 Jun 2011 18:06:24 +0100 as excerpted:

> A recent annoyance is the new interface at google groups always starting
> a new thread.
> 
> Is there any way to add a scoring rule that will ignore anyone using
> that interface, please? (I have googled to no avail)

There should be, but it's likely to apply only after you've downloaded 
the message, because the required header isn't likely to be in the 
overviews/headers as delivered by your server, that pan uses for pre-
download scoring (unless the Message-IDs have a common bit you can filter 
on, as they might).

Normally, the overviews downloaded to populate the article list have only 
a rather limited set of headers -- those necessary to display the article 
in the list (subject/from/newsgroups/size), the one necessary for proper 
threading (references), if cross-posted, one listing the per-server 
article sequence number for each group in ordered to track the message in 
all groups (x-ref), and the one used to ID the message and often to 
request it (message-ID).

If, by looking at the message-id headers (toggle view headers on and 
compare them), you can find a common element in the google-groups posts 
that's unlikely to false-positive on other posts, you can use that for 
scoring.  (You'll need to edit the scorefile manually, but you can make 
it easy by creating a references score, saving that, then editing that 
entry in the scorefile to message-id instead of references.)

If not, KHaley mentioned recently that current pan now has the ability to 
score on ALL headers, NOT just those in the overview.  I've not tried it 
yet, and there's no GUI for it, you'll have to directly edit the scorefile 
as appropriate, but it should work, with the caveat above about having to 
download the message in ordered to get the full headers, of course.

Also, I'm not sure whether that feature made it into pan 0.134 or not.  
It may be that you must compile from git sources in ordered to get it (or 
get it from someone that did).

Here's pan's git URL (lostcoder is KHaley's handle, KHaley being pan's 
primary developer since CKerr lost interest and asked for someone else to 
take over, some years ago).

git://github.com/lostcoder/pan2.git

And here's some documentation on the scorefile format pan uses, which 
happens to be pretty close to that of slrn:

http://www.slrn.org/docs/score.txt

-- 
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