On Tue, 24 Sep 2002, Larry Osterman wrote:
> But what about this scenario:
> C: 1 NOOP
> S: * 5 EXISTS
> S: 1 OK NOOP Completed
> Client 2 now deletes all the messages in the mailbox.  At this point,
> the client thinks that there are 5 items in the mailbox while the server
> knows that there are none.
>
> C: 2 FETCH *
> S: * 1 EXPUNGE
> S: * 1 EXPUNGE
> S: * 1 EXPUNGE
> S: * 1 EXPUNGE
> S: * 1 EXPUNGE
> S: 2 BAD Sequence Number * is illegal on an empty folder

This scenario is specifically forbidden in RFC 2060 section 7.4.1.

RFC 2180 section 4.1 discusses what should be done.

In believe that the best strategies are 4.1.1 and 4.1.4.  Users show a
definite preference for 4.1.1.

Failing that, 4.1.3 is far preferable to 4.1.2.  4.1.3 causes some
temporary user interface confusion but it usually resolves itself quickly
in a compliant client.  4.1.2, on the other hand, definitely confuses
clients.

In UW imapd, the behavior depends upon the underlying mail store and its
mailbox format.  The preferred mbx format has 4.1.1 behavior.  Most other
formats have 4.1.4 behavior, but a few legacy formats have 4.1.3 behavior.

I have gotten *very* accustomed to the 4.1.1 behavior, and would never
again willingly use a server which any other behavior.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.

Reply via email to