Articles spuriously marked as read?

Articles spuriously marked as read?

[Lars Magne Ingebrigtsen]

I've lately had a nagging feeling of having articles disappear on me.
I'll be reading a followup to some article, and find I can't really
remember reading the original article.

So I've started to suspect that Gnus is marking unread articles as
read. 

This is only a vague suspicion, because I have been totally unable to
provoke Gnus into doing anything bad while I'm watching.  It seems to
be waiting for me to look the other way, and then it starts doing odd
things...  Spooky!

Has anybody else noticed anything in this area?  Or is it just me? 

(And is "anybody" singular or plural?)

-- 
  "Yes.  The journey through the human heart 
     would have to wait until some other time."

Re: Articles spuriously marked as read?

[David C Worenklein]

> From: larsi@ifi.uio.no (Lars Magne Ingebrigtsen)
> Date: 19 Mar 1996 02:24:52 +0100
> Organization: Dept. of Informatics, University of Oslo, Norway
> X-From-Line: ding-request@ifi.uio.no  Mon Mar 18 21:10:38 1996
> 
> I've lately had a nagging feeling of having articles disappear on me.
> I'll be reading a followup to some article, and find I can't really
> remember reading the original article.

This may be an nntp problem. Sometimes articles aren't properly
propogated -- site x will get an article that you don't. Then user@x
follows-up to the article, and the follow-up hits your site.

Re: Articles spuriously marked as read?

[Jack Vinson]

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

LMI> So I've started to suspect that Gnus is marking unread articles as
LMI> read. 

I usually credit this to the nature of the Usenet.  Sometimes followups to
articles appear before the original, and sometimes articles expire locally
before I ever see the original.

I've actually noticed the reverse situation, though.  There are articles
(Usenet, never mail) that I am s\b_u\b_r\b_e\b_ I have read once that I see
again.  This usually happens in the larger groups that I read, so it is
entirely possible that people are reposting the same articles because they
didn't see their words immediately.

-- 
Jack Vinson                   jvinson@cheux.ecs.umass.edu
"Narf!" - Pinky               <http://www.cis.upenn.edu/~vinson/home.html>

Re: Articles spuriously marked as read?

[Per Persson]

jvinson@cheux.ecs.umass.edu (Jack Vinson) wrote:

   I've actually noticed the reverse situation, though.  There are
   articles (Usenet, never mail) that I am s\b_u\b_r\b_e\b_ I have
   read once that I see again.  This usually happens in the larger
   groups that I read, so it is entirely possible that people are
   reposting the same articles because they didn't see their words
   immediately.

That could be thanks to a low history-expire date with INN. If your
server has forgotten that it recieved an article the article will get
through a second (and third (and fourth (I've seen it happen three
times once))) time and you'll have to read it once again.

I guess people with a _huge_ disk would like Gnus to keep a record of
all Messages-ID's it has seen (I have a vague feeling that you can do
this with september Gnus already).

-- 
Bahnhof; http://www.bahnhof.se/      <pp@bahnhof.se> | a   difference   of
  (work) phone(voice/fax):      +46 18 100899/103737 | opinion is possible
 PFAWWW; http://pfawww.pp.se/~pp/  <pp@pfawww.pp.se> | ^. .^     anum meum  
 (!work) phone(voice)    :             +46 18 247473 | ( @ )       aperies

Re: Articles spuriously marked as read?

[Lars Magne Ingebrigtsen]

David C Worenklein <dcw@gcm.com> writes:

> This may be an nntp problem. Sometimes articles aren't properly
> propogated -- site x will get an article that you don't. Then user@x
> follows-up to the article, and the follow-up hits your site.

(I forgot to mention that the article(s) that I was *sure* I hadn't
seen really were in the group.)

It is definitely a Gnus bug.  I did the following yesterday:  1)
Process-marked two groups.  2)  `G s b' to brew SOUP packets.  

There were 5 unread articles in the second group, but only 2 of them
ended up being in the SOUP packet.  None of the other three articles
were cross-posted to the first group.

This might be a gnus-soup bug only.  (I certainly hope so.)  However,
when I try to repeat the bug, it just doesn't happen.  I'm starting to
leaning towards problems with nntp server communication when
requesting several articles rapidly, but that's highly speculative. 

-- 
  "Yes.  The journey through the human heart 
     would have to wait until some other time."