Need regex code for counting newsgroups

Aug 17, 2013 10 Replies

There was a thread some time ago on ways to kill postings with too many cross-posts. Some of these methods involved regular expressions, most likely counting commas in the newsgroups header. Can anyone provide a pointer to the thread, or related threads?



Thanks,



Joe Gwinn


You can set that in Agent 6.0 in folder properties. Today, I've been using the Message-ID field for the new flock of bozos. When I see a message id containing "theremailer" or "dont-email.me", it's a pretty good bet for the bin.

Pete Keillor

Except that I cannot use Agent, because there is no Mac version.

I am using Thoth under MacOS, and Thoth follows Perl's version of regex.

Joe Gwinn

Runs great under Crossover Office on Linux, others have reported the MAC version is just as good. Turning on a signature I normally use only in the Forte Agent beta group...

How are you doing that, Pete? I only see a Message-ID field in filtering emails, not Usenet posts.

The way I do it is based on the commas in the "Newgsgroups: " header.

Two newgroups .*,.* Three newsgroups .*,.*,.* Four newsgroups .*,.*,.*,.*

Where: ".*" means any number of any characters. ',' means itself -- a plain old comma.

Exactly how you tell your newsreader to use those varies. I would do something like "-10" for each one of those, so two newsgroups would be -10, three would be -20, four (or more) would be -30, and set the auto-kill threshold to -15. (Three is too many in cross-posting.) I also use + scores on the "Subject: " header for the very few things which I want to see which are cross-posted, such as the "What Is It" weekly puzzle posting thread.

Enjoy, DoN.

The third link in google for terms newsgroup filter crosspost count looks like an r.c.m thread from 31 Aug 2011, subject = "Agent kill filter help please". See eg

You could also add a term for your specific newsreader.

H key to show the headers, then make a new filter with the message id in the filter expression like below.

Message-ID: {googlegroups\.com}

The "\" is needed in front of the "." in the regular expression language so that it isn't interpreted as a wildcard.

Pete Keillor

Aha! Peachy. I'll give that a try. Thanks. Maybe this will work for reference headers, too. If I can catch the orig. message (he's filtered in my reader) when he's being referred to, I can end the spam created when good folks (guilttripping Gunner, et al) reply to said spammers.

I've been asking Agent techs (Beck, Gold, Prince) for text (or more header) filtering for a decade now, to no avail.

That's what I was trying to remember. I knew there was a simple, battle-tested solution. Thanks

I use Thoth, which uses the Perl regex engine PCRE .

Wonder if there is a direct way to count commas and kill if count exceeds some threshold.

I can also put the subject header test earlier in the processing, and terminate filter processing for the correct subject.

Joe Gwinn

I did do that and got flooded with suggestions that I use some other newsreader, but I was looking for actual regex code, because the reader I currently use (Thoth) does accept regex, so step one is to try that.

Hmm. Thoth is uncommon, so I didn't think to try that.

But I just tried it, using "thoth regex cross posting" (without quotes), and it yielded another regex expression "x-ref matches Regular Expression" from

and

Joe Gwinn

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required