Re: [slrn] scorefile syntax to eliminate crossposting ?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* Peter J Ross <pjr@xxxxxxxxxxxxxxx> [2007-10-16]:
In news.software.readers on Mon, 15 Oct 2007 15:24:06 GMT, Steven
Schneider <steven_schneider@xxxxxxxxx> wrote:

* Peter J Ross <pjr@xxxxxxxxxxxxxxx> [2007-10-15]:
In news.software.readers on Sun, 14 Oct 2007 03:03:28 GMT, Steven
Schneider <steven_schneider@xxxxxxxxx> wrote:

So this would be overkill then? (assuming > 3 x-postings)

Xref: telusplanet.*\C[a-z]:.*\C[a-z]:.*\C[a-z]:.*\C[a-z]:.* *$

It would also not match group names that don't end with a letter, such
as comp.os.plan9 or comp.lang.c++.

I see. That pretty much confirms what I'd been reading in the slrn
manual then; I like doing things the hard way. Well, not really,
but sometimes it seems that way.

I like "doing things the hard way" for the sake of accuracy.

The other problems with your regex are

1. \C is unnecessary because regexes in slrn scorefiles are
case-insensitive by default.

I don't recall, but I think I knew that at the time. I probably
threw it in there for completeness, or was it for anal retentiveness?
:-)

2. Even if \C were useful, it wouldn't need to be repeated unless \c
intervened.

See above.

3. ".* *$" does nothing that ".*" at the end wouldn't do.

Again, above. Though, as you pointed out earlier, for all my
efforts to be "exhaustive," I was still missing cases where groups
didn't end in a letter.

So far, it seems to have worked. I've probably just been lucky,
though.

Counting colons (or commas when filtering on the Newsgroups header)
ought to be quite good enough.

Xref: :.*:.*:.*: is more elegant and easy to understand.

Hey, I'm always open to learning something new. Especially when it
will save me some work, time, or it makes more sense. :-)

Sometimes the complicated method is best in order to avoid false
positives.

Here's an example of how I test complicated scorefile rules in slrn:

#v+

[*]
Score: -8888
Message-Id: \c^<n?n?q?\.?[0-9]\{10\}\.[0-9]\{6\}\.[0-9]\{4,6\}@[a-z]?[0-9]\{1,2\}[a-z]20[01]0[a-z]\{3\}\.googlegroups\.com>$

[*]
Score: +8888
Message-Id: \.googlegroups\.com>

#v-

The two rules cancel each other out if both match.

Anything that matches the first rule also matches the second.

So if the first rule doesn't match, there will be a huge positive
score to highlight the fact that something in the first rule is
broken.

Let me see if I understand this: you test the validity of a
complicated rule against a simpler rule, or vice versa. If they
match then they both functionally do the same thing.

Is this a test you perform to see if you can simplify a complicated
rule then? When would you want to use a more complicated version of
simpler rule? Even if they're both functionally equivalent in the
test, are there cases where the more complicated rule may be more
desirable?

Am I perhaps missing the point? :-/

Btw, what's with all the noise in the group all of a sudden? Is the
group being attacked for some reason, or is it just random drive-by
idiocy?

Somebody didn't like having his spambot software criticised and
decided to throw a tantrum. The small child in question posts as
"Almond". At least he knows he's a nut.


Ah, thanks for the clarification. :-)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (OpenBSD)

iD8DBQFHFr6syJgTRdM3i6gRAumgAJ0R+K6ja6Vqsgydnb/e6TA1OyPMhACcDtXD
efNLKA8YGzZN4QZi+4A9XOs=
=qEDY
-----END PGP SIGNATURE-----

--
W. Steven Schneider <steven_schneider@xxxxxxxxx>
.

Relevant Pages

  • Re: [slrn] scorefile syntax to eliminate crossposting ?
    ... It would also not match group names that don't end with a letter, ... That pretty much confirms what I'd been reading in the slrn ... Here's an example of how I test complicated scorefile rules in slrn: ... Anything that matches the first rule also matches the second. ...
    (news.software.readers)
  • how to catch up to current mail
    ... I am missing a feature of slrn that I often used when I still read most ... mailing lists via gmane.org. ...
    (comp.mail.mutt)
  • Re: reductio ad falsum versus reductio ad absurdum
    ... the second form seems like the first rule with an unnecessary tilde ... am i missing something? ... Prev by Date: ...
    (sci.logic)
  • Re: Slrn pgp signature and more headers
    ... There used to be a script by Emmanuele Bassi that provided PGP support ... but his website went missing a while ago. ... It may or may not work with recent versions of slrn. ...
    (news.software.readers)
Loading