[Bug 4544] New: SUBJECT_DRUG_GAP_* rules - false positives

[bu...@bugzilla.spamassassin.org]

http://bugzilla.spamassassin.org/show_bug.cgi?id=4544

           Summary: SUBJECT_DRUG_GAP_* rules - false positives
           Product: Spamassassin
           Version: 3.0.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P5
         Component: Rules
        AssignedTo: dev@spamassassin.apache.org
        ReportedBy: birder@ozemail.com.au


The "gappy version of <drug>" rules are probably too over-zealous.

I receive regular emails with the subject "Birdline Victoria", which triggers
the SUBJECT_DRUG_GAP_L rule:  "birdLinE VIcToRiA" (matches in capitals.) 
Despite training these messages as ham, they continue to be marked as spam.

The "gappy" rules probably should be changed to only match if non-alphanumeric
characters are used in the gaps.  (Any spammer attempting to hide drug names
with alphanumeric characters will most likely render them unrecognisable to the
recipient.)



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4544] SUBJECT_DRUG_GAP_* rules - false positives

[bu...@bugzilla.spamassassin.org]

http://bugzilla.spamassassin.org/show_bug.cgi?id=4544





------- Additional Comments From birder@ozemail.com.au  2005-08-18 05:05 -------
Addendum:  An alternate heuristic would be to convert accented characters to
unaccented characters, reduce repeated characters to a single instance, and
ignore case.

For example "VIAGRR�" (actual spam example, which didn't trigger a DRUG rule)
would be converted to "viagra" - easily matched by the existing drug rules.
Heuristics which selected only upper or lower case could also be used.

At some point the obfuscation techniques of the spammer will cancel out the
readability to the end-user i.e. we win.  :)



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.