Mail for domain mailboxes should not be treated differently than any other message during the SMTP transaction. In fact it can't - during HELO no RCPT is even known, and a message could have several RCPTs of which not all are destined to a domain mailbox.

It could be that some daemon, for instance GrayWall, intercepts a message before the rule triggers, though.

You can in many cases use the log action (L) to test a transaction rule.

/Rolf 

I have been using Mercury for some years no and it has always worked perfectly. I always worked with local domains, but recently I needed to use a domain mailbox, so now I have a system with 2 local domains and one domain mailbox. I also use SMTP transaction-level filtering extensively which works fine for the local domains. However, the filtering rules seem to be ignored when mail is sent to the domain mailbox.

Are transaction-level filtering rules not applied to domain mailboxes? I cannot find anything wrong with the filtering rules since the same rules work for mail targeted at a local domain (mostly rules based on RCPT of HELO using wildcards) . If they should be applied to domain mailboxes is there a way to determine (or debug) why the filtering rules do not seen to be applied?

Thanks,

Deagol