Been digging through the rfc. Thought it would be good to clarify some things to get on the same page and offer some dev ideas....
RFC Notes: - Messages can be identified by either message-id or unique-id. - Message-id: Is a sequential number unique to a mailbox(folder). It is not system-wide by the rfc. It can change per session.. meaning it could be calculated on each request. 10 messages in a mailbox sequence them 1-10. - Unique-id: A 32-bit value assigned to each message. that MUST NOT refer to any other message in the mailbox. Unique identifiers MUST be strictly ascending in the mailbox at all times. This does not indicate sequential just ascending. It notes that the ids could reference date/time base for uniqueness. This should work across multiple servers. "Note: Ideally, unique identifiers SHOULD persist at all times. Although this specification recognizes that failure to persist can be unavoidable in certain server environments...." and that it should be reordered if necessary. Anyway... based on this it looks like changes could be made to allow for a multimaster environment with syncing and no collisions while maintaining RFC. Development Ideas: 1. The message-id could be removed from the db and just calculated on a mailbox list request. Its just their sequence. 2. The unique-id could be calculated with a date/time/sec/millsec base combined with a machine id added to the dbmail configuration to assure ascending and unique. - Kevin Kevin
