> From: "rODbegbie" <[EMAIL PROTECTED]> > > 3) Analyse the HTML source. eg. If the graphic is hosted on > bigandbouncey.com, there's a high chance of spam. >
Could also come up with new blacklist catagories: blacklist_url and blacklist_site where if there is a URL that starts out the same as a blacklisted URL, it will get a point boost, and if there's URL that refers to a site that matches a blacklisted one (by IP or name), it'll get tagged for extra points. Though, I'd also like to have a handy-dandy filter that would take HTML as input and convert it to plain text. Doesn't bother me that fancy formatting would be lost. For links it would put the URL in parenthesis after the link text (most mail readers that are html aware will also create links out of url's in plain text messages). For images, it would probably put "[Image http://url.to.image]";, things like that (or the image's ALT text in place of "Image"). For the types of messages were're talking about, that'd probably leave just a series of "[Image URL](URL) [Image URL](URL) ..." left over. That'd be a pretty easy to match pattern, I'd think. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
