Hello, When I fetch a single non existent message sequence id which is larger then the number of existent message in a mailbox (pretty obvious if it is non existent) then RFC3501 tells us the imap server SHOULD return a tagged BAD response. A tagged NO respons should good as well when we look at the description of the FETCH responses (NO - fetch error: can't fetch that data).
The comparation between UW, Cyrus and Courier shows us some interesting results: On UW: A01 FETCH 121212 BODY[1] A01 BAD Bogus sequence in FETCH On Courier: A01 FETCH 121212 BODY[1] * NO Invalid message sequence number 121212 A01 OK FETCH Completed. On Cyrus 2.2.0: A01 FETCH 121212 BODY[1] A01 NO No matching messages According me the courier response is wrong, uw and cyrus responses are right. Because it does not concerns a UID FETCH I think the server never should return a tagged OK but somehow the courier people have a different opinion and tell me rfc3501 is not clear about that and that a client should not request non existent message sequence id's. The last argument is of course a valid one. Finally, what's the use of untagged NO responses (warnings) when each server has it own specific warnings. When I take the example again and all were untagged NO responses then we have: Bogus sequence in FETCH Invalid message sequence number 121212 No matching messages For an imap client they are of no use because every server has different warning messages for the same issue. Regards, Marc Groot Koerkamp. -- ----------------------------------------------------------------- For information about this mailing list, and its archives, see: http://www.washington.edu/imap/imap-list.html -----------------------------------------------------------------
