On Wed, 2 May 2007, Thorsten Büker wrote:
Thanks for this clearification! After comparing some cyrus.header files from
different folders and "my" seen file, I think I understood ;-)
While copying the mailbox the message UIDs change, at least as far as I can
see. So, one might be able to replace the mailbox IDs in the seen file be the
mailbox IDs of the copied version, but the message UIDs referenced to are
out-dated. Or do I miss something?
You are correct that the message UIDs will change anytime a message is
copied. Also the mailbox ID will be different for each new mailbox. If
you want to maintain the seen state during the copy, you'll need a fancier
script than the one I wrote. :)
Another attempt:
Are there any reasons, why something like
ORIG=`grep -R -e '^Message-ID:' $originalmbox | awk '{print $2}' | sort`
KOPIE=`grep -R -e '^Message-ID:' $copyofmbox | awk '{print $2}' | sort`
some diff and some improved awking (spaces won't work in case of spaces in
folder names) is not a sufficient way to create a list of files, which
might be deleted in the original mailbox?
That sounds like a reasonable approach, but obviously you should be careful
anytime you modify the cyrus files directly. :)
Well, actually only the files/message themselves are touched. Ok, they're not
simply touched but deleted ;-)
Further Cyrus files (cyrus.*, seen state, subscription list)
are not affected. Furthermore one might assume that they correspond to the
mailbox contents after deleting any "wrong" files/messages.
I hope to find a reasonable approach to create a list of unnecessary
Message-IDs and their corresponding files, which might be piped to remove.
Keep you informed.
If you can figure out the message-IDs to delete, then it should work fine.
Andy
----
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html