https://bugs.kde.org/show_bug.cgi?id=364506
--- Comment #5 from Jan Kundrát <j...@kde.org> --- The server's idea of UIDs is this: [781, 1163, 1254, 1255, 1256, 1257, 1258, 1259, 1260, 1635, 1639] IOW, the first 9 messages remain unchanged, all others are gone, and two more are added. The server apeears to be Dovecot 2.1.x (it's got SPECIAL-USE but not BINARY). > Well, my reading of > https://tools.ietf.org/html/rfc7162#section-3.2.10.1 would > lead me to conclude that it doesn't have to, as it only needs to look at the > UID set parameter, which Trojita doesn't pass. I think that you're confusing https://tools.ietf.org/html/rfc7162#section-3.2.5.1 which we indeed don't send because we always maintain a full seq-to-UID mapping with https://tools.ietf.org/html/rfc7162#section-3.2.5.2 which we do send out. The former is defined to default to 1:* in this context, so by us not sending it, we're actually asking Dovecot to potentially extend the range of data to be returned. The latter tells Dovecot something about our idea of the mapping, so it has an idea on our state. But anyway, if we didn't send this partial UID-to-seq mapping, Dovecot would have to send more data. -- You are receiving this mail because: You are watching all bug changes.