Spam

[Gert van den Berg]

How does the same spammer get to post more messages after its first
messages have been deleted? Any reason why spam forum accounts are not
simply banned, IPs blocked on IPTABLES, etc after a single spam post?

(The relevant user havew 2 non-deleted messages on the forum and 6 in
total (which are all somewhere wasting space in my mailbox...))

[dimesio]

How does the same spammer get to post more messages after its first
messages have been deleted? Any reason why spam forum accounts are not
simply banned, IPs blocked on IPTABLES, etc after a single spam post?

(The relevant user havew 2 non-deleted messages on the forum and 6 in
total (which are all somewhere wasting space in my mailbox...))

I've deleted the messages, but I don't have the power to ban accounts.

[Martin Gregorie]

On Tue, 2010-03-16 at 05:43 -0500, dimesio wrote:

Gert van den Berg wrote:

How does the same spammer get to post more messages after its first
messages have been deleted? Any reason why spam forum accounts are not
simply banned, IPs blocked on IPTABLES, etc after a single spam post?

(The relevant user havew 2 non-deleted messages on the forum and 6 in
total (which are all somewhere wasting space in my mailbox...))

I've deleted the messages, but I don't have the power to ban accounts.

Is it possible to put a spam filter in the loop? Something like:

Forum postings ---------+
mail from mail list-----+
|
v
spamfilter
| |
ham spam
v |
forum web server <------+ v
mail list <-------------+ silently deleted

There has to be a central switch which sees all incoming messages or
forum posts couldn't appear on the mail list and vice versa. It should
be possible to add in a spam filter at this point, but it would be nice
for somebody who understands the system to confirm that.

Martin

[Gert van den Berg]

On Tue, Mar 16, 2010 at 12:43, dimesio <[email protected]> wrote:

I've deleted the messages, but I don't have the power to ban accounts.

Ah, ok. moderator access vs. admin access?

(I read via mailing list, where the deletes don't show up ...) (I
have checked it on the forum though...) The issue is if UI enable
gmail's spam filter, the list unsubscribes me due to bounces...

For the forum, it might also be useful to use those bad IP blacklists
and ask for more captchas for users from those IPs. (banning those IPs
are annoying, I'm behind a "transparent" ISP proxy which I cam't turn
it off and the proxy's IPs are often banned from some sites...)
(Something like the SBL, DBL and XBL from Spamhaus might be valid
sources...)

Martin Gregorie wrote:
Is it possible to put a spam filter in the loop?

I'm not sure about the mailing list / integration, but for a standard
phpBB 2, I'm not aware of such a point (which does not involve editing
the code...) It does have a profanity filter, which might be useful
for destroying links regularly spammed and to URL shortening services.
(Which hides URLs and is unnecessary outside Twitter)

It should be possible for the mailing list part though, which would
reduce the impact... (Keep mailing list spam of the forum (and mailing
list) and forum spam of the mailing list...)

My _*guess*_ of the of the architecture for the integration would be:
Form posts gets submitted to forum database and mailing list (from the
special wineforum-user account).
Mailing list posts gets sent out to subscribers. It is received
(probably by a special subscriber) by the integration, that adds it to
the forum's database. Mailing list software archives it.
If this is the case, they run mostly independant, and the intergation
just ensures that everything appears everywhere (using different
mechanisms in each direction... This would also require seperate
filtering for the mailing list and forum )

Gert

[Martin Gregorie]

On Tue, 2010-03-16 at 17:36 +0200, Gert van den Berg wrote:

I'm not sure about the mailing list / integration, but for a standard
phpBB 2, I'm not aware of such a point (which does not involve editing
the code...) It does have a profanity filter, which might be useful
for destroying links regularly spammed and to URL shortening services.
(Which hides URLs and is unnecessary outside Twitter)

How is the profanity filter implemented (built into the code or an
external lookup server? If the latter, substituting Spamassassin should
be fairly simple. Php would pass the message to spamc the spamd client,
spamd scans it, adds headers containing the spam score and returns it to
spamc. Filtering is done downstream of spamc by using the spam score. It
would be trivial to make Spamassassin do double duty and serve as the
profanity filter too.

As all spam scoring is done by spamd, the same server instance can be
simultaneously used by the mailing list. Most MTAs can call it already.

Martin

[Gert van den Berg]

On Tue, Mar 16, 2010 at 23:45, Martin Gregorie wrote:

How is the profanity filter implemented (built into the code or an
external lookup server? If the latter, substituting Spamassassin should
be fairly simple. Php would pass the message to spamc the spamd client,
spamd scans it, adds headers containing the spam score and returns it to
spamc. Filtering is done downstream of spamc by using the spam score. It
would be trivial to make Spamassassin do double duty and serve as the
profanity filter too.

probably internally... There is a quite widely used censor_text
function.... (I can't find its implementation on code.phpbb.com
though... Probably my Googling that is problematic..)

Gert