1124924 - If CONDSTORE is usable, use "uid fetch 1:* flags (CHANGDEDSINCE modseq)" for new mail check by Biff instead of "uid fetch NextUID:* flags", in order to resolve problem like Bug 693204, Bug 517461

Status: NEW

(MailNews Core :: Networking: IMAP, enhancement)

Description

[WADA:World Anti-bad-Duping Agency]

+++ This bug was initially created as a clone of Bug #1124569 +++

If CONDSTORE is usable, use "uid fetch 1:* flags (CHANGEDSiNCE modseq)" for new mail check by Biff instead of "uid fetch NextUIS:* flags", in order to resolve problem like Bug 693204, Bug 517461.

Note:
Bug 693204 is for any flag state change, and CONDSTORE support is essentilally irrelevant, so this bug is fully applicable to Bug 693204.
Bug 517461,is for Unread mail count only, and for issue when CONDSTORE is used. So following problem is also involved in Bug 517461 and this bug is partiaal solution of Bug 517461.
   if CONDSTORE is used, msgDBHdr of "deleted/expunged mail by other client" won't be removed from Tb.
   if the non-removed mail was Unread mail, "incorrect Unread mail count" is surely visible for users.

Comment 1

[WADA:World Anti-bad-Duping Agency]

FYI.
If Gmail IMAP, following is needed. (with CONDSTORE only. without QRESYNC)
   After "Select Mbox" for re-sync         :  uid fetch 1:* Flags X-GM-MSGID  X-GM-THRID X-GM-LABELS
   Upon new mail check by Biff             : "uid fetch 1:* Flags X-GM-MSGID  X-GM-THRID X-GM-LABELS (CHANGEDSiNCE modseq)
   Wen new mail is detected via IDLE   : "uid fetch nn  Flags X-GM-MSGID  X-GM-THRID X-GM-LABELS

Comment 2

[Wayne Mery (:wsmwk)]

Gene, do you think this is feasible? It blocks quite a few other issues.

Comment 3

[gene smith]

(In reply to Wayne Mery (:wsmwk) from comment #2)

Gene, do you think this is feasible? It blocks quite a few other issues.

Not sure if this change is necessary, at least for Bug 693204 which the summary claims this will fix. That bug was because gmail didn't report flag changes (e.g., message has been read) via imap IDLE. Gmail does this now since 2014 and I just tested it to be sure.
Also, this claims to help when CONDSTORE is enable which, by default, it currently isn't. I have a patch to "improve" condstore and add qresync which is somewhat fermenting (but not quite bit rotted, bug 1747311) and, if I ever get back to it, this change might be useful, but I would have to research it some.
Also, my condstore/qresync patch may help with the unread count issue, not sure.