On 09/12/2017 03:36 PM, Brandon Long via mailop wrote:
Thinking about this, it makes sense. It's not a short term temporary condition, it's something that could take hours to resolve, and having your mail sit in your outbox for who knows how long retrying, and maybe the user doesn't notice and shuts down the computer or whatever, it never goes out.The immediate permanent failure lets the user know that the mail isn't going out, and the user can try and send it later.
Temp failures at the end of data are probably even worse, imagine a user sending an 150MB message, no reason to upload the whole thing and fail and retry. Or even smaller messages on mobile devices.
Reading this reminded me of how I used to have a server configured. Namely the MSA (mostly) accepted anything and was configured to smart host through the rest of the email infrastructure (as opposed to sending to the world.)
This allowed me to have better control of what message the clients saw, because I could control what was in the DSN.
This also allowed me to easily scan outbound email from all clients, even with attachments on dial up modems (back in the day) without timeouts on scanners. This worked because clients took as long as they needed to to upload the file to the MSA, which would then burst it over to the MTA and filter it in < 30 seconds.
If the MTA / filters found a problem with the outbound message, the MSA would generate a DSN and send it back to the purported sender. - Which I could trust because the sender authenticated to the MSA.
Sending the bounce ended up providing more flexibility and avoided the UI issue.
-- Grant. . . . unix || die
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ mailop mailing list mailop@mailop.org https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
