uuencoded attachments produce woe

[David Relson]

At 03:43 PM 12/6/02, Greg Louis wrote:

>Gary Robinson had occasion to send me a .pdf file recently; it arrived
>uuencoded, and was deemed spam by bogofilter; the training had only seen
>uuencoded attachments in spam before.  Running bogofilter with the -R
>option showed about 7,000 unhelpful tokens had been generated.
>
>Reading the comments in lexer.l, it seems that BASE64 is being avoided,
>but not uuencoding.  I have no idea whether it would be hard or easy to
>change that; anyone care to comment?

Greg,

bogofilter doesn't currently have special handling for encoded attachments, 
as they deserve.  It's on the TODO list.

If you care to forward the message to me (privately), I'll see what's 
needed to avoid it like bogofilter currently does with BASE64.

David