Re: More IMAP testing, please

[Dan Christensen <jdc@uwo.ca>]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Dan Christensen <jdc@uwo.ca> writes:
>
>> My largest group has about 6500 articles, but after a few minutes of
>> using gnus, gnus would only show me the newest 71 articles (which seem
>> to be the ones whose article numbers are within 100 of the latest).
>> `M-g' on the group fixed this, but then I found I had lost *all*
>> marks on the older articles (tick, expired, answered definitely
>> lost, read not lost).
>
> That's not good.
>
> To take the last problem first, Gnus will only send over updates of
> marks when it thinks a mark has changed.  So if Gnus doesn't think that
> article 423 has a tick, it won't de-flag the article on the server,
> either -- it should only send over updates of articles that has changed
> status while in the summary buffer.
>
> I think.
>
> So I don't understand that at all.

I've verified that these flags were removed from the articles on the
server (a total of more than 1000 flags).  And this problem happened
with two different imap groups, at the same time.

All flags except read ("Seen", I think) were removed: Answered, Flagged,
gnus-expire, gnus-save and gnus-forward were the ones that were there.

> The former problem sounds like a parsing problem of some kind, but since
> `M-g' is able to parse the data, I'm not sure how to interpret that...

To me it sounds like a problem involving the strategy of sometimes only
getting mark information for the last 100 articles, and sometimes
getting it for all articles.

So far it hasn't happened again, but I'm still in the same emacs session
so I can provide whatever information you need.  I've included an except
from the *imap log* buffer below, and can send more if that would help.
My INBOX has about 80 articles, but many of them are quite old so their
article numbers aren't in the most recent 100 numbers.

>> A less important problem: for some reason, two of the groups on this
>> server show up with "*" instead of the unread article count.  I think
>> these are on the only two groups that are empty.  If I hit `M g', my
>> summary buffer shows 1 unseen and 1 unread.  But when I try to enter the
>> groups, gnus notices that there aren't any articles.  After that, `M g'
>> shows 0 unread and unseen.
>
> Sounds like an off-by-one error on empty groups.

This is repeatable, so let me know if there's anything I can do to help
debug it.

Dan

16:33:52 9 EXAMINE "INBOX"
16:33:52 10 UID FETCH 5459:* FLAGS
...
16:34:07 35 EXAMINE "INBOX"
16:34:07 36 UID FETCH 5459:* FLAGS
...
16:34:25 59 EXAMINE "INBOX"
16:34:25 60 UID FETCH 5459:* FLAGS
...
16:35:24 84 EXAMINE "INBOX"
16:35:24 85 UID FETCH 5459:* FLAGS
16:35:24 86 SELECT "INBOX"
16:35:24 87 SELECT "INBOX"
16:35:24 88 UID FETCH 5466:5559 (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref X-Spam-Status To Cc Keywords Newsgroups Subject)])
16:35:34 89 UID STORE 4808,5398 -FLAGS.SILENT (gnus-forward)
16:35:34 90 UID STORE 5355,5358,5361:5362,5365,5372,5377:5378,5380:5381,5398,5400,5404,5409,5411,5416:5419,5421,5423,5425,5434:5437,5443 -FLAGS.SILENT (gnus-expire)
16:35:34 91 UID STORE 4357,4808,5358,5418:5419,5425,5436:5437 -FLAGS.SILENT (\Answered)
16:35:34 92 UID STORE 4781,4808,4813,5420 -FLAGS.SILENT (\Flagged)
...
16:38:41 113 EXAMINE "INBOX"
16:38:41 114 UID FETCH 1:* FLAGS
16:38:41 115 SELECT "INBOX"
16:38:41 116 SELECT "INBOX"
16:38:41 117 UID FETCH 123:5559 (UID RFC822.SIZE BODYSTRUCTURE BODY.PEEK[HEADER.FIELDS (Subject From Date Message-Id References In-Reply-To Xref X-Spam-Status To Cc Keywords Newsgroups Subject)])