On Thu, Mar 10, 2005, Paul J Stevens <[EMAIL PROTECTED]> said: >> Ok, thanks for reminding me. Let's kick #161 because if we haven't started >> looking into what crashes the mime parser, now ain't the time. > > Well, the crash was triggered by malformed messages with crafted headers > (read: embedded CR within a header, without a LF). Since those will be > pretty rare, and this bug does not represent any regression since it's > been around forever, and since it's not relevant for 2.1 since gmime > handles it ok, we'll simply defer that bug.
Defer is fine, but we really need to get on it soon because the crash prevents the message from being accepted or rejected; it just stays in the queue and helps enforce that MAXCONNECTS option in a bad way... Not that I have time to fix it right now, just to remember that it's bad. Aaron
