[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Effectively filtering bulk mail from lynx-dev (was Re: LYNX-DEV Your Web
From: |
Jonathan Sergent |
Subject: |
Effectively filtering bulk mail from lynx-dev (was Re: LYNX-DEV Your Web Site's Findability ) |
Date: |
Sat, 26 Jul 1997 21:11:38 -0500 |
I'd be interested to know what portion lynx-devvers want some form of
filtering for junk mail to be set up on the list.
Respond privately to that question *and I will summarize to the list!*
... this discussion is dumping too much mail in our boxes already.
In message <address@hidden>, David Woolley writes:
] >
] > Maybe we should divert messages that don't contain the word "lynx"
] > (case-insensitive) in the message to a moderator, excluding
]
] That would probably mean modifying a shared copy of MajorDomo, so not
] something that could be done on the basis of just Lynx. The alternative
] amounts to duplicating a significant part of MajorDomo.
Not really.
Majordomo aliases look like this (this is a list on this machine):
# SigOS
sigos: "|/home/majordom/wrapper resend -l sigos sigos-list"
sigos-list: :include:/home/majordom/lists/sigos
sigos-owner: sergent,oboyle
sigos-request: "|/home/majordom/wrapper majordomo -l sigos"
owner-sigos: sigos-owner
sigos-approval: sigos-owner
Changing the resend line from "resend" to "resend-lynx-dev", and making
a "resend-lynx-dev" script [modified resend] wouldn't be monumental.
] > Received: headers. That would get rid of most of the commercial
] ^^^^^^^^
]
] When spamming is happening, the Received headers are the most important
] things to PRESERVE! Rather than suppress them to the moderator, they
] should be preserved to the list (I think this is a per list option on
] MajorDomo, so could be done).
I'm not saying to get rid of them, silly, just that the Received header
will most likely have "lynx-dev" in it with current versions of sendmail's
default HReceived: line...
HReceived: $?sfrom $s $.$?_($?s$|from $.$_)
$.by $j ($v/$Z)$?r with $r$. id $i$?u
for $u; $|;
$.$b
(the 'for $u' would say 'for <address@hidden>' as generated by
the sendmail on the system running majordomo in the Received: header
added just before sending the message (through a pipe) to resend;
have also seen 'for delivery to $u')
... sig.net seems to have their headers modified slightly, and resend
does appear to be killing old Received headers. I don't know that this
matters all that much. I really don't think complaining is worthwhile;
most of the outfits are working with people like AGIS these days.
As I said before, IMHO the effective way to do things would be to look
for case-insensitve "lynx" anywhere other than a Received header.
This would work because spam generally does not have the recipient's
address anywhere in a header like To: or Cc:, only in the envelope address.
I say to not look at Received when doing this because Received often
contains the envelope address.
Searching for lynx in To or Cc would probably work as well but searching
the entire body might reduce false positive matches.
We can work on non-bulk junkmail (I call it clueless-mail) later. Checking
to make sure there is more than one non-blank, non-quoted line in the
message would be a good idea.
Most of our clueless-mail is either one line with key banging on it or the
quoted lynx.browser.org page with nothing added).
If some subscribers do not want the filtering, it's possible to set that
up ... alias something like
lynx-dev: "|/path/to/wrapper resend
lynx-dev-people-who-do-not-want-filtering-outgoing", "|/path/to/wrapper
modified_resend_which_does_filtering
lynx-dev-people-who-do-want-filtering-outgoing"
then two :include: files and two majordomo config files...
less verbose names hopefully :-)
Geez, I'm spending time writing this message instead of working on
the stupid tempspace code. Blah. You'd think I'd have sat down and
just done it by now, but I haven't.
--jss.
;
; To UNSUBSCRIBE: Send a mail message to address@hidden
; with "unsubscribe lynx-dev" (without the
; quotation marks) on a line by itself.
;
Re: LYNX-DEV Your Web Site's Findability, Robert Bonomi, 1997/07/26