[volunteers] (forw) [Mailman-owner] Auto-discard notification

[Rick Moen]

Quoting Don Marti (dmarti at zgp.org):

> Subscribing to a Mailman list is much less work than
> your typical comment spammer will go through to get
> a link from a high-PageRank site.  I have a feeling
> we're going to see a lot more of this stuff.

One reason I _hope_ not to see that is BOfHish listadmins' efforts to
keep the price high.  For example, I have a case open with The Planet / 
EV1servers.net's Abuse Department, where I've amply documented this
person's ToS violations, and have requested termination of his commercial
hosting, there.

Additionally, I have just now excised all of the spam from the SVLUG
list's cumulative, and am rebuilding the HTML archive from the mbox.
Sometimes, manual cleanup of the mbox and a second rebuild is required,
in these cases, so don't be alarmed if for a while today people see a
couple of partial, garbled messages at the end of the HTML archive:
These result from text lines within messages in the mbox that have the
capitalised word "From" flush-left. 

Hmm, the archive-builder is harfing on some character enconding in the
mbox -- October 2000, right around posting #3728.  Drat.  I'm going to
have to track that down, and edit _that_ out manually, too.  Grrr.

Marc Merlin has some more-robust mbox-parsing software for these
situations, and I may have to resort to asking him for that.  Meanwhile,
I've just managed to take the back-postings Web-archive for
svlug at lists.svlug.org offline for the time-being.  Sorry about that.
(If I've completely screwed up, we can get the archive back from
backups.)

Anyhow, what I was _trying_ to, of course, is remove any significant
advantage the spammer might have gained from hitting our main mailing
list.  Unfortunately, I forgot that rebuilding from a dodgy,
non-cleaned-up mbox file a mailing list as ancient as SVLUG's can be
fraught with parsing problems. 


Getting back to the matter of cheeky spammers barging into mailing
lists, if that starts to happen, then our fallback measure would
logically be setting the flag to make new subscribers be initially
moderated, then remove that flag manually after the poster has sent one
or two real, provably-human postings.