Richard Bang zei: > Hi, > > I've been following this thread with interest and now I cant decide if > my implementation is compliant or not. > > Mark Crispin wrote: > --------------------------------------------- > You could keep track of STATUS vs. > SELECT, and only increase the UIDVALIDITY on SELECT. Or (and this is > probably better) increase the UIDVALIDITY subsequent to the SELECT so > that > STATUS will show the UIDVALIDITY of the next SELECT. > --------------------------------------------- > > My implementation will return the same UIDVALIDITY for each SELECT > unless its internal UIDNEXT becomes invalid due to some other external > influence. > > Mark statement implies that UIDVALIDITY should increase UIDVALIDITY on > select rather than when the UIDNEXT goes out of scope. > > Which is correct? >
Your implementation. If you increase UIDVALIDITY on each select an imap-client will never be able to cache uid's/messages. If the UIDVALIDITY stays the same, the client knows that the cached messsages/headers together with the UID's are still valid and do not need to be refreshed. Regards, Marc Groot Koerkamp.
