http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4691
------- Additional Comments From [EMAIL PROTECTED] 2005-11-21 07:13 ------- > I thought the reason that body and rawbody were split into arrays was to protect > against worst-case behavior of test regexes. By pasting those lines back > together, this patch appears to forgo that protection. That is the theory. The practice is that spammers have several very trivial patterns that are guaranteed to be impossible to detect in SA, short of writing a separate plugin for each thing you want to test. Essentially, any HTML-level pattern containing one or more line breaks is undetectable. The only current solution is to run 'full' rules over the entire email to test for those patterns. This of course will only work with non-encoded mail bodies. It is perhaps also worth noting that by refusing to allow multiline rawbody, and arbitrarily and unpredictably breaking 'body', you are forcing the use of 'full' rules for things that could be trivially parsed in connected body or rawbody rules. The only conclusion can be that the dev's believe that is is safer and faster to parse the entire email with a regex than it would be to parse a small but useful part of the email with the same regex. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
