[quote user="Supur"]You can now create an empty file called MSGIDS.MER in any mailbox
directory (i.e, a directory where a .CNM file gets created), and this
signals to Mercury that it should suppress duplicate messages in that
mailbox. Duplicate detection is based on a combination of sender and
message-ID, and only the last 200 messages delivered to the mailbox are
actually remembered.
It is a bit unclear what does it mean "only the last 200 messages delivered to the mailbox are
actually remembered"? I have thousands of messages in these folders - for which I would like to stay there (and not to be.. deleted). Probably it means that a new incomming message will be compared for duplicates only with last 200 messages in folder. Which would be good.[/quote]
We have msgids.mer in place for each of our local user mailbox accounts. Works fine as long as you don't use additional "public mailboxes". In that case Mercury has some problems in identifying of duplicates.
As soon as you put an empty msgids.mer into the user's mailbox directory, Mercury starts holding the incoming mails (for that user) additionally in a kind of cache. But this "cache" is limited to 200 mails to prevent an extensive memory using and/or processing duration. This "cache", after built up, will permanently renewed where new mails will be added and old mails will be removed. Whenever a new mail will be retrieved by MercuryD for that user, Mercury is checking its "last-200-mail-cache" whether this new mail was already delivered in past. And in case it's a duplicate, Mercury will drop this new mail.