At 2:57 PM -0500 1/6/03, Don Doumakes is rumored to have typed:
That makes sense. Clearly, though, A is somehow returning an error return to the MTA. Mr. A is an AOL subscriber.
(*sigh*) Yippee. That explains a whole lot.
The bounce message from the AOL mailer daemon to the list-request address (addresses munged) says: <snip> Your mail to the following recipients could not be delivered because they are not accepting mail from A@aol.com: B
Whoa...other than, I think, swapping A with B (A was the bouncer, B was the bouncee, yes?), this again appears to be a really malformed mess. It's obvious AOL's filters are trusing the From: header field, but this example you've given, if complete and correct, isn't a valid bounce message. If A and B were really listeed that way, it's no wonder SmartList got confused.
Still, I'm guessing that as an AOL subscriber, A lacks the kind of tools that would make it possible to bounce more politely.
More correctly, at least. I've seen some _really_ lousy bounces from AOL. (Although lately, MSN is taking the cake for this nonsense.) I refuse to use AOL, so I don't know how their new "anti-spam" filters work, but I've seen:
----- The following addresses had permanent fatal errors ----- <xxxxx@aol.com>
----- Transcript of session follows ----- ... while talking to air-xf04.mail.aol.com.:
RCPT To:<xxxxx@aol.com> <<< 550 xxxxx IS NOT ACCEPTING MAIL FROM THIS SENDER 550 <xxxxx@aol.com>... User unknown
...used as a lazy man's unsubscribe method more than once, so I expect to see interesting things coming out of AOL, MSN, and even Norton's Anti-Virus in the future (although I think Norton's just deletes mail instead of bouncing it). I should probably note that in the cases I've received the error message I list above from AOL, SmartList has found and unsubscribed the correct address. I've also not seen anything from AOL that looks exactly like the error message you've included. I seem to be one of the few people who has SmartList send bounce copies to me, and I review them daily for the lists I maintain. I _have_ over the years seen a small number of instances where SmartList was wrong in its guess, but darned few over a whole lot of bounces.
Sooner or later this problem will be back, with a different cast of characters, and it would be nice to solve it technically.
You are likely correct, but I'd argue this is not an obvious flaw in SmartList, but rather with AOL (gee, add another one to the list). If AOL isn't giving proper session transcripts with valid 500-level errors, SmartList is kinda stuck trying to guess what address is really bouncing. I am constantly amazed at how good a job it does at that, even when the mailer sends cutsy garbage like, "Sorry it didn't work out." The only thing I can think to do is to force the issue; I'll see if, over the next week or so, I can get a couple AOL users to help me run some tests to see if the bounce message you've seen, at least, can be duplicated. Charlie