Matt Sergeant wrote: MS> I think the reason my stuff would work quite well is it's specifically MS> designed to make it easy to get the known body text from the document, MS> in the same way that an email client would. The MIME::* stuff seems more MS> generic to me, so you'd have to code that stuff in as extra.
Well, might an analysis of the number and type of MIME parts be useful for identifying spam? MS> It was written for Windows (it's actually part of O'Reilly's WebBoard MS> (now Activa or some other wierd latin name), and I got permission to MS> take that bit of code as my own, because I knew it would be useful in MS> the future). Excellent then; sounds great. MS> Honestly, it would have to be your call - I'll send you the code so you MS> can play with the API and see if it fits your mind-space. I'm not too MS> likely to integrate it because we don't use this stuff in house to parse MS> our emails - we have C code doing it which is beyond my control (we MS> split all the emails up into separate files before passing them to our MS> anti-virus, anti-spam, and anti-porn products). Ok, happy to take a look at it. Certainly sounds like it has a lot of promise. C _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
