You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@spamassassin.apache.org by jh...@apache.org on 2011/02/21 18:19:54 UTC

svn commit: r1073076 - /spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Author: jhardin
Date: Mon Feb 21 17:19:54 2011
New Revision: 1073076

URL: http://svn.apache.org/viewvc?rev=1073076&view=rev
Log:
Allow obfu rule to be published

Modified:
    spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Modified: spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
URL: http://svn.apache.org/viewvc/spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf?rev=1073076&r1=1073075&r2=1073076&view=diff
==============================================================================
--- spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf (original)
+++ spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf Mon Feb 21 17:19:54 2011
@@ -25,5 +25,5 @@ describe      URI_OBFU_TLD      URI top-
 
 body          URI_DEOBFU_INSTR  /(?:delete|remove|take\sout)(?:\sthe)?\sspaces/i
 describe      URI_DEOBFU_INSTR  How to deobfuscate this URI
-tflags        URI_DEOBFU_INSTR  nopublish
+#tflags        URI_DEOBFU_INSTR  nopublish
 



Re: svn commit: r1073076 - /spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Posted by John Hardin <jh...@impsec.org>.
On Mon, 21 Feb 2011, Warren Togami Jr. wrote:

> On 02/21/2011 10:06 PM, Warren Togami Jr. wrote:
>>  I agree it would be wise to allow this to auto-promote, but given the
>>  small size of our ham corpus and the fact that this pattern could rarely
>>  but legitimately appear in non-spam, perhaps we should manually cap its
>>  score to be on the safe side.
>>
>>  To throw out an arbitrary number I'd suggest 0.9 points?
>>
>>  Warren
>
> http: //ruleqa.spamassassin.org/20110221-r1072884-n/T_URI_DEOBFU_INSTR/detail
> http: //ruleqa.spamassassin.org/20110221-r1072884-n/URI_OBFU_TLD/detail
>
> On second thought, examine the overlap of these rules.  Nearly all such cases 
> are already caught by high scoring rules like PYZOR_CHECK or SOUGHT.  Given 
> that both rules cannot be 100% correct and the fact that they are numerically 
> redundant, I advise caution in allowing these to be auto-promoted and 
> especially auto-scored.
>
> Perhaps we are better off leaving easy to catch temporary campaign patterns 
> like these to tools better equipped to handle them like SOUGHT or PYZOR.

This I disagree with. How do you know the obfuscations these rules test 
for are only temporary? Obfuscating URLs with spaces has a _long_ history.

-- 
  John Hardin KA7OHZ                    http://www.impsec.org/~jhardin/
  jhardin@impsec.org    FALaholic #11174     pgpk -a jhardin@impsec.org
  key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C  AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
   Health Care _is_ a right - the government has no business keeping
   you from getting it. But forcing somebody else to pay for your
   health care at gunpoint (i.e. through taxation) is _not_ a right.
-----------------------------------------------------------------------
  Today: George Washington's 279th Birthday

Re: svn commit: r1073076 - /spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Posted by "Warren Togami Jr." <wt...@gmail.com>.
On 02/21/2011 10:06 PM, Warren Togami Jr. wrote:
> I agree it would be wise to allow this to auto-promote, but given the
> small size of our ham corpus and the fact that this pattern could rarely
> but legitimately appear in non-spam, perhaps we should manually cap its
> score to be on the safe side.
>
> To throw out an arbitrary number I'd suggest 0.9 points?
>
> Warren

http://ruleqa.spamassassin.org/20110221-r1072884-n/T_URI_DEOBFU_INSTR/detail
http://ruleqa.spamassassin.org/20110221-r1072884-n/URI_OBFU_TLD/detail

On second thought, examine the overlap of these rules.  Nearly all such 
cases are already caught by high scoring rules like PYZOR_CHECK or 
SOUGHT.  Given that both rules cannot be 100% correct and the fact that 
they are numerically redundant, I advise caution in allowing these to be 
auto-promoted and especially auto-scored.

Perhaps we are better off leaving easy to catch temporary campaign 
patterns like these to tools better equipped to handle them like SOUGHT 
or PYZOR.

Warren

Re: svn commit: r1073076 - /spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Posted by John Hardin <jh...@impsec.org>.
On Mon, 21 Feb 2011, Warren Togami Jr. wrote:

> I agree it would be wise to allow this to auto-promote, but given the small 
> size of our ham corpus and the fact that this pattern could rarely but 
> legitimately appear in non-spam, perhaps we should manually cap its score to 
> be on the safe side.

How do you manually cap a score vs. manually setting a score to a specific 
value? I'm not aware of such a capability, and would welcome it.

> To throw out an arbitrary number I'd suggest 0.9 points?

Sounds reasonable to me.

> Warren
>
> On 02/21/2011 07:19 AM, jhardin@apache.org wrote:
>>  Author: jhardin
>>  Date: Mon Feb 21 17:19:54 2011
>>  New Revision: 1073076
>>
>>  URL: http://svn.apache.org/viewvc?rev=1073076&view=rev
>>  Log:
>>  Allow obfu rule to be published
>>
>>  Modified:
>>       spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
>>
>>  Modified: spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
>>  URL:
>>  http://svn.apache.org/viewvc/spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf?rev=1073076&r1=1073075&r2=1073076&view=diff
>>  ==============================================================================
>>  --- spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
>>  (original)
>>  +++ spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf Mon Feb
>>  21 17:19:54 2011
>> @@ -25,5 +25,5 @@ describe      URI_OBFU_TLD      URI top-
>>
>>    body          URI_DEOBFU_INSTR
>>    /(?:delete|remove|take\sout)(?:\sthe)?\sspaces/i
>>    describe      URI_DEOBFU_INSTR  How to deobfuscate this URI
>>  -tflags        URI_DEOBFU_INSTR  nopublish
>>  +#tflags        URI_DEOBFU_INSTR  nopublish

-- 
  John Hardin KA7OHZ                    http://www.impsec.org/~jhardin/
  jhardin@impsec.org    FALaholic #11174     pgpk -a jhardin@impsec.org
  key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C  AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
   Health Care _is_ a right - the government has no business keeping
   you from getting it. But forcing somebody else to pay for your
   health care at gunpoint (i.e. through taxation) is _not_ a right.
-----------------------------------------------------------------------
  Today: George Washington's 279th Birthday

Re: svn commit: r1073076 - /spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf

Posted by "Warren Togami Jr." <wt...@gmail.com>.
I agree it would be wise to allow this to auto-promote, but given the 
small size of our ham corpus and the fact that this pattern could rarely 
but legitimately appear in non-spam, perhaps we should manually cap its 
score to be on the safe side.

To throw out an arbitrary number I'd suggest 0.9 points?

Warren

On 02/21/2011 07:19 AM, jhardin@apache.org wrote:
> Author: jhardin
> Date: Mon Feb 21 17:19:54 2011
> New Revision: 1073076
>
> URL: http://svn.apache.org/viewvc?rev=1073076&view=rev
> Log:
> Allow obfu rule to be published
>
> Modified:
>      spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
>
> Modified: spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf
> URL: http://svn.apache.org/viewvc/spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf?rev=1073076&r1=1073075&r2=1073076&view=diff
> ==============================================================================
> --- spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf (original)
> +++ spamassassin/trunk/rulesrc/sandbox/jhardin/20_uri_obfu_ws.cf Mon Feb 21 17:19:54 2011
> @@ -25,5 +25,5 @@ describe      URI_OBFU_TLD      URI top-
>
>   body          URI_DEOBFU_INSTR  /(?:delete|remove|take\sout)(?:\sthe)?\sspaces/i
>   describe      URI_DEOBFU_INSTR  How to deobfuscate this URI
> -tflags        URI_DEOBFU_INSTR  nopublish
> +#tflags        URI_DEOBFU_INSTR  nopublish
>
>
>