Some MTAs have taken to adding a Delivered-To header to mail as it is delivered, and using that header to perform primitive loop detection. This means that when you receive a mail once, then redirect it to a list to which you are subscribed, some MTA configurations will then object to the fact that you're receiving a message which appears to have already been delivered to you once, and will bounce the message. It has been asserted that an MUA should be removing Delivered-To: headers when redirecting a mail message, and that pine is at fault, not the MTA which bounced the message. I disagree with this assertion, which was not backed up by any RFC reference. But as we can easily work around this particular brokenness in pine, perhaps we should?
Delivered-To: is a djb-ism. Relying on it for mail system behaviour will all end in tears. There may be a case for using it in mailing list loop detection and similar cases, however the MTA enforcing it on intermediate or final delivery is clearly wrong and not supportable by RFCs (other than you being able to make a fool of yourself on your own system). MUAs should not be spending their time second guessing wierd MTA behaviour. If this was an MS MTA that decided to bounce messages like this would we still be seeking to fix the MUA?
This is a bug in the MTA. PINE is flawless. No bugs. No bugs I tell you. Please use memtest86 to be sure you do not have bad RAM. If that checks out ok, try going to http://www.zombo.com