I've had a few spam messages slip through SA recently because they consist of
little or no text, with the entire content of the message in a graphic using an
<IMG> tag.

I think virtually any message sent as one big graphic would be spam, but I
can't think of a good way to detect it using a regular expression. Perhaps you
could look for <IMG...height= followed by a number greater than x (300?). Or
maybe a better way would be to analyze the content to HTML ratio of the
message - i.e. the message contains 500 total characters, 450 of which
constitute HTML tags and 50 the actual content between the tags, therefore it's
spam?

--
michael moncur   mgm at starlingtech.com   http://www.starlingtech.com/
"I believe in getting into hot water; it keeps you clean."
                -- G. K. Chesterton


_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to