You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by cmvhk <Vi...@bristol.ac.uk> on 2010/01/05 16:19:15 UTC

False positive for LOCAL_BODY_CIALIS

An email sent to me containing a book review in French was recently falsely
classified as spam, largely because it failed the LOCAL_BODY_CIALIS rule:

2.0 LOCAL_BODY_CIALIS      BODY: Mentions viagra clone 'cialis'

I quote offending part of the message:

... de\s sa sortie en 1978,
comme un outil de travail de premier plan pour les spe/cialistes de
langue et d'e/pigraphie e/trusques, mais e/tait devenue avec le temps ....

e/ is a standard way of transliterating e-acute. Could the rule be rewritten
so as not to catch instances such as this? (I recall a rule which used to
object to 'Best wishes, Virginia' because of the proximity of 'best' and
'virgin', which was rewritten so as not to match if the string 'virgin' was
part of 'Virginia'.

Virginia Knight (Dr)
ILRT, University of Bristol

-- 
View this message in context: http://old.nabble.com/False-positive-for-LOCAL_BODY_CIALIS-tp27026636p27026636.html
Sent from the SpamAssassin - Users mailing list archive at Nabble.com.


Re: False positive for LOCAL_BODY_CIALIS

Posted by Ned Slider <ne...@unixmail.co.uk>.
On 01/05/2010 06:39 PM, Joseph Brennan wrote:
>
> Ned Slider <ne...@unixmail.co.uk> wrote:
>
>> body LOCAL_BODY_CIALIS /\bcialis/i
>
>
> That's probably what the rule is, and it will match 'spe/cialistes'.
>
> Joseph Brennan
> Columbia University Information Technology
>
>

Yep, my apologies, I missed the broken spe/cial... in the original post, 
and indeed it also hits my own local common drugs rule based on the same.


Re: False positive for LOCAL_BODY_CIALIS

Posted by Joseph Brennan <br...@columbia.edu>.
Ned Slider <ne...@unixmail.co.uk> wrote:

> body	LOCAL_BODY_CIALIS	/\bcialis/i


That's probably what the rule is, and it will match 'spe/cialistes'.

Joseph Brennan
Columbia University Information Technology


Re: False positive for LOCAL_BODY_CIALIS

Posted by Ned Slider <ne...@unixmail.co.uk>.
On 01/05/2010 03:19 PM, cmvhk wrote:
>
> An email sent to me containing a book review in French was recently falsely
> classified as spam, largely because it failed the LOCAL_BODY_CIALIS rule:
>
> 2.0 LOCAL_BODY_CIALIS      BODY: Mentions viagra clone 'cialis'
>
> I quote offending part of the message:
>
> ... de\s sa sortie en 1978,
> comme un outil de travail de premier plan pour les spe/cialistes de
> langue et d'e/pigraphie e/trusques, mais e/tait devenue avec le temps ....
>
> e/ is a standard way of transliterating e-acute. Could the rule be rewritten
> so as not to catch instances such as this? (I recall a rule which used to
> object to 'Best wishes, Virginia' because of the proximity of 'best' and
> 'virgin', which was rewritten so as not to match if the string 'virgin' was
> part of 'Virginia'.
>
> Virginia Knight (Dr)
> ILRT, University of Bristol
>

Any rule named LOCAL_* is usually a good sign that it's a local rule 
added by the local mail administrator. Check with whomever maintains 
your SpamAssassin installation.

Without seeing your rule to know what it's checking for, I would suggest 
adding a check for a word break '\b' at the start of the word so as to 
avoid false positive hits on specialist. For example,

body	LOCAL_BODY_CIALIS	/\bcialis/i


Re: False positive for LOCAL_BODY_CIALIS

Posted by RW <rw...@googlemail.com>.
On Tue, 5 Jan 2010 07:19:15 -0800 (PST)
cmvhk <Vi...@bristol.ac.uk> wrote:


> 2.0 LOCAL_BODY_CIALIS      BODY: Mentions viagra clone 'cialis'
> 
> ...
>  Could the rule be rewritten so as not to catch instances such as
>  this? 


It's not a default rule and "LOCAL_"  looks like a prefix used by
your admin for local rules, so I'd suggest you take it up with your IT
department.

Re: False positive for LOCAL_BODY_CIALIS

Posted by Matus UHLAR - fantomas <uh...@fantomas.sk>.
On 05.01.10 07:19, cmvhk wrote:
> An email sent to me containing a book review in French was recently falsely
> classified as spam, largely because it failed the LOCAL_BODY_CIALIS rule:
> 
> 2.0 LOCAL_BODY_CIALIS      BODY: Mentions viagra clone 'cialis'
> 
> I quote offending part of the message:
> 
> ... de\s sa sortie en 1978,
> comme un outil de travail de premier plan pour les spe/cialistes de
> langue et d'e/pigraphie e/trusques, mais e/tait devenue avec le temps ....
> 
> e/ is a standard way of transliterating e-acute. Could the rule be rewritten
> so as not to catch instances such as this? (I recall a rule which used to
> object to 'Best wishes, Virginia' because of the proximity of 'best' and
> 'virgin', which was rewritten so as not to match if the string 'virgin' was
> part of 'Virginia'.

another rule depending on languages used.

the fastest workaround should be rule that matches "specialistes" in any form
and meta-rule that gives -2 when LOCAL_BODY_CIALIS and the rule above are
hit.

If french language is detected, this could be also a part of the meta-rule.

-- 
Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
Micro$oft random number generator: 0, 0, 0, 4.33e+67, 0, 0, 0...

Re: False positive for LOCAL_BODY_CIALIS

Posted by Kai Schaetzl <ma...@conactive.com>.
Cmvhk wrote on Tue, 5 Jan 2010 07:19:15 -0800 (PST):

> 2.0 LOCAL_BODY_CIALIS      BODY: Mentions viagra clone 'cialis'

Sure, that this rule is part of standard SA? I can't find it in 3.2.5 or 
3.3.0.
Apart from this, if that message came out as spam with these additional 2 
points it must have already scored unusually high for ham.


Kai

-- 
Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com