[Dovecot] Extraneous information
Not technically a bug, but a small waste of bandwidth: it seems (at
least as of 1.2.10) that if neither CONDSTORE/QRESYNC are enabled
dovecot still sends HIGHESTMODSEQ info:
a examine Testing.AAA
- FLAGS (\Answered \Flagged \Deleted \Seen \Draft $Forwarded impflag0
impflag1 impflag2) - OK [PERMANENTFLAGS ()] Read-only mailbox.
- 309 EXISTS
- 0 RECENT
- OK [UNSEEN 5] First unseen.
- OK [UIDVALIDITY 1225053009] UIDs valid
- OK [UIDNEXT 387] Predicted next UID
- OK [HIGHESTMODSEQ 136] Highest a OK [READ-ONLY] Select completed.
I'm guessing this would only happen if using a client that supports
QRESYNC (like IMP) and then switching later to a client that doesn't
support.
On 1.2.2010, at 9.52, Michael M. Slusarz wrote:
Not technically a bug, but a small waste of bandwidth: it seems (at least as of 1.2.10) that if neither CONDSTORE/QRESYNC are enabled dovecot still sends HIGHESTMODSEQ info:
It's intentional. From rfc 4551:
This document adds two new response codes, HIGHESTMODSEQ and NOMODSEQ. One of those response codes MUST be returned in the OK untagged response for a successful SELECT/EXAMINE command.
Quoting Timo Sirainen tss@iki.fi:
On 1.2.2010, at 9.52, Michael M. Slusarz wrote:
Not technically a bug, but a small waste of bandwidth: it seems (at
least as of 1.2.10) that if neither CONDSTORE/QRESYNC are enabled
dovecot still sends HIGHESTMODSEQ info:It's intentional. From rfc 4551:
This document adds two new response codes, HIGHESTMODSEQ and NOMODSEQ. One of those response codes MUST be returned in the OK untagged response for a successful SELECT/EXAMINE command.
Yeah, you're right. I wasn't thinking about that part properly - I
was thinking about it in terms of a user *using* CONDSTORE vs. a
server *supporting* CONDSTORE. I guess that info does always need to
be sent because it is possible to be using CONDSTORE info without
sending an enabling command.
Sorry for the noise.
michael
participants (2)
-
Michael M. Slusarz
-
Timo Sirainen