Thanks Gyula,
it's very interesting.
I see the problem in my case, though, because I don't know in advance what
will be the exact names of these X-headers. All I know is that they are
constructed as a combination of two randomly chosen words. These headers hit
Exchange information store, overflow the headers storage, and create DoS
conditions. Here are some examples (very weird):
abdominothoracic-birthplace
absolute-anteconsonantal
abundancy-canorously
accelerando-armilla
accretal-bedark
acidosteophyte-adipopexia
acronically-bronchotome
adjunction-Argean
adjunctively-bullfighting
adularescence-aval
adverseness-akepiro
aeronef-betoil
aeruginous-alalus
afghanistan-bacterial
afterwards-bicornuate
Agathis-albinuria
agentship-allegorister
aigialosaur-cartmaking
aire-bocher
airward-ceraunoscope
aliphatic-Albireo
allalinite-calibered
Amaryllidaceae-Bahaullah
Amblyopsidae-autoscopic
aminoacetophenone-cannonry
Amitabha-bardlike
amminolytic-abox
amulla-cadaverously
anally-Caesarism
Anybody ever encountered that?
> Boris,
>
[quoted text clipped - 16 lines]
> >
> > Any advise or link will be highly appreciated.
Karakas, Gyula [vamsoft] - 17 Nov 2004 10:36 GMT
This rather seems to be malformed spam emails, these random words are
used to confuse Bayesian methods. Your problem might be that the actual
message is bad, i.e. the message does not contain an empty line after
the headers so these words are treated as header information.
To confirm this you should examine the full MIME message. This cannot
be done with outlook as it only shows partial header information, but
using another MUA - like outlook express - you can get the email in
question via POP3 or IMAP.
If this is the problem I do not think too much can be done - use some
kind of spam protection, if you have not done this already.
Gyula Karakas
orf support
www.vamsoft.com/orf
> I see the problem in my case, though, because I don't know in advance
> what will be the exact names of these X-headers. All I know is that
[quoted text clipped - 10 lines]
> acidosteophyte-adipopexia
> acronically-bronchotome