Let's make sure we are all talking about the same thing. What is an
"unread article"? One answer - it is an article whose 'identifier'
(usually a sequential number [by newsgroup] kept by the news server)
does not appear in a control list of articles that have been "read".
My point is that souper is run on your own machine. The control list
that it uses is called 'newsrc'. If the HOME environmental variable
is set, souper's 'newsrc' is in _that_ directory. If HOME is not set,
souper's 'newsrc' is in the current dir (when souper is started).
I myself run trn in a shell account on my ISP's machine. The control
list that the ISP's trn uses is called '.newsrc'. It is normally in
the home directory of the (unix) user's account.
Note: souper (on your machine) and trn (on the ISP's machine) use
DIFFERENT control lists (the lists are on two different machines!)
[And different control lists means one list _could_ show an article
as "read", while the other list _could_ not_show the article (i.e.,
on the second list that article would have the status: "unread".)]
The only time the two lists should be equal for any newsgroup is if:
(a) you fetch [or view] _all_ articles in the newsgroup from _both_
souper and trn, and (b) no new articles for that newsgroup arrived
at the server in the interval between when you ran the one program
(souper or trn) and when you ran the other program (trn or souper).
mikus