Could you attach a sample of your old scorefile that doesn't work?

%Score created by Pan on Thu Aug 10 09:25:23 2006
[^cox.*]
Score: =100
From: ^sean <sean@no\.spam\.plz>$

That's the one I traced it down with.  As you can see, it's newly created.  It works as it is above.  On the cox line, replace the * after the dot with a \ before it, making the line [^cox\.] (begins with cox., where the dot is a dot, not any char) and it no longer applies.  I wondered if it had to match the entire thing so tried [^cox\..*] as well, no luck.  However, as I said, [^cox.*] works.

You mentioned on another bug you have access to Cox.  This scoring rule is based on a post to the cox.internet.discussion.unix group from back on July 09, the BIND on Cox network thread, you have the author above, message-id: <e41sg.154502$k%3.147287@dukeread12>

...

BTW, it doesn't seem significant above, but I noticed new-pan was creating the score line with a double colon after Score (the example above was originally Score:: =100), while old-pan used a single colon.  Both appear to work and the single colon is just how it happened to be left when I decided it wasn't the colon and moved on to trying something else.  Is there a significance to the difference?  Which one matches the other client score files, or are they ambiguous as well (assuming there's not a semantic meaning I failed to deduce, thus both are legal but specify slightly different things)?  IOW, which one's the bug?  <g>

It's inconvenient, but "old pan" scorefiles using that notation
are wrong.  The problem is that the groups are listed as wildmat
strings, not regular expressions.  slrn does /not/ recognize the
exmaple you attached.

As for the BTW: the one or two colons indicate whether only
one (`::') or all of the lines (`:') inside the group need to
match for the group to pass.  So for single-line groups,
they are equivalent.

=======================================================================

Since you're one of the keepers of Pan wisdom,
here's where to get the canonical information on scoring:
http://www.slrn.org/docs/score.txt is /the/ scorefile bible.
http://xnews.remarqs.net/scoring.txt is helpful, but secondary.

Pan aims for equivalence with the first reference,
with two exceptions:

(1) Use of PCRE regular expresions, rather than S-Lang REs.
This is because of PCRE's overwhelming mindshare.
Also, S-Lang 2.2 is slated to migrate to PCRE.

(2) Pan adds the Xnews tweak of allowing case sensitivity:
"If you want the  expression-to-match to be case sensitive,
then put an equal sign instead of of a : after the keyword, e.g.
  Score: -9999
  ~Subject= .*[a-z]

Thanks for the tips.  I have a couple new bookmarks now, the links of which I have a feeling I'll be posting from time to time. =8^)