Text 25856, 71 rader
Skriven 2012-02-26 14:53:03 av Richard Webb (1:116/901.0)
  Kommentar till text 25778 av Dale Shipp (1:261/1466.0)
Ärende: Fidonews v29, n8 mes
============================
HEllo DAle,

On Sat 2012-Feb-25 22:57, Dale Shipp (1:261/1466) wrote to Richard Webb:

 RW> Might be.  HOwever, only the first one appeared here when I
 RW> read the message base after tossing them.  I don't know if
 RW> they arrived and were killed as dupes here by squish or

DS> Squish certainly has the ability to kill them as dupes, even though
DS> they were not duplicate messages.  IF you have the DupeCheck key
DS> word set to check Header, then it will check the message header to
DS> decide if it is a duplicate.  If the header is a duplicate, then
DS> Squish will call the message a duplicate -- even if the MSGID is
DS> different.

Usually doesn't have, there are more than one ways to check
for dupes in Squish.
Mine is set up to use both header and msgid which usually
prevents that.

 RW> what. I do use the dupelongheader keyword, however that's
 RW> actually written in squish which should prevent that 
 RW> though, and it usually doesn't have a problem, but fyi only 
 RW> the
 RW> first one of that issue appeared here as well.
DS>  
DS> DupeLongHeader would not have prevented this error by SQUISH since
DS> the entire header was a duplicate for all of the messages.

Right, the msgids must have been ... ah I don't know,
waiting for coffee.

DS> If you have KillDupes set, then the dupes are gone.  If you have
DS> that keyword commented out, then the dupes would have been placed
DS> into your DUPE area as defined by your setup.   You could look for
DS> them there. 

YEs, I'm doing the killdupes thing.  I just read it when I
received it next day as a file.  AT least it doesn't kill
the other so I get any articles that are in it too.  I've
always known the dupe checking routines in squish are rather strict.    They
usually don't get me, and I've read quite a
few echoes with multi part posts that you'd think squish
would nail.  I note Janis says she's tweaked the posting
routines in her sw that she's designed for that to tweak it
a bit.  Mark LEwis's postit seems to not get caught by it,
and I do post multi part things from here on occasion using
it.

 RW>  <hmmm>  Begs the question.  WAs it only squish users that
 RW> didn't see the remainder of the messages?

DS> Squish users could certainly have had the problem, depending on how
DS> they are set up.  I have no knowledge about what other types of
DS> system might be subject to the same flaw in their dupe checking
DS> logic.

Yep, and this gets into how msgids are defined and created
as well.  Can't speak for Mark, but i think that's why he
was so careful with his postit utility on msgid creation.
I've seen a dissertation by him, iirc in ftsc_public re
creating msgids a couple years ago, and saw reference to it
by him this thread iirc.
He's doing his weekend work marathon thing though so you
won't see him comment until tomorrow or Tuesday when he
catches up <g>.

Regards,
           Richard
---
 * Origin:  (1:116/901)