456601 - Thunderbird loses track of most of the messages in a folder, because all our cached connections are used

Status: RESOLVED FIXED

(MailNews Core :: Backend, defect)

Description

[Dan Mosedale (:dmosedale, :dmose)]

In my case, the INBOX.  davida has seen this as well.  It's likely to freak users out, so this should block beta1.  I've sent bienvenu a protocol log, and he has a patch that he thinks might do the trick.

Comment 1

[David :Bienvenu]

Attachment: possible fix

I think this should prevent this from happening again - I'm still looking into recovering from this situation, but this patch should help. It should prevent those "foundit" assertions you've been seeing as well. The autosync stuff exposed an existing problem in how cached connections and condstore played together, by basically using all our cached connections for its work. This fix makes it so when we re-use a cached connection, in particular, its flag state, that we start off the flag state in "partial" mode.

Comment 2

[Dan Mosedale (:dmosedale, :dmose)]

Comment on attachment 339992 [details] [diff] [review]
possible fix

r+sr=dmose

Comment 3

[David :Bienvenu]

fix checked in, changeset:   415:0873795c1cc4

I'd still like to improve the sanity checking that would have ameliorated this bug, but that requires a bit more thought.

Comment 4

[Dan Mosedale (:dmosedale, :dmose)]

Presumably, the patch that's gone in should be good enough for beta 1.  Moving the target milestone out; feel free to reset if my presumption is wrong.

Comment 5

[David :Bienvenu]

Yes, I believe the patch that went in will fix the underlying issue - if it doesn't completely fix it, then this will have to move back to b1. I will try to have a patch that adds more sanity checking in hand, though. We already do sanity checking if the mailbox changes at all, which helps mostly with the inbox. And we sanity check all the time if you're using the imap delete model.

Comment 6

[David :Bienvenu]

moving to b2 - I haven't heard any more reports of this.

Comment 7

[David :Bienvenu]

moving past b2

Comment 8

[David :Bienvenu]

taking off blocker list.

Comment 9

[Wayne Mery (:wsmwk)]

(In reply to David :Bienvenu from comment #5)
> Yes, I believe the patch that went in will fix the underlying issue - if it
> doesn't completely fix it, then this will have to move back to b1. I will
> try to have a patch that adds more sanity checking in hand, though. We
> already do sanity checking if the mailbox changes at all, which helps mostly
> with the inbox. And we sanity check all the time if you're using the imap
> delete model.

should we close this and file new bug?  Or has the above already been fixed elsewhere?

Comment 10

[Wayne Mery (:wsmwk)]

(In reply to Wayne Mery (:wsmwk) from comment #9)
> (In reply to David :Bienvenu from comment #5)
> > I will
> > try to have a patch that adds more sanity checking in hand, though. We
> > already do sanity checking if the mailbox changes at all, which helps mostly
> > with the inbox. And we sanity check all the time if you're using the imap
> > delete model.

David, do you think we still need a followup bug?

Comment 11

[Wayne Mery (:wsmwk)]

(In reply to Wayne Mery (:wsmwk) from comment #10)
> (In reply to Wayne Mery (:wsmwk) from comment #9)
> > (In reply to David :Bienvenu from comment #5)
> > > I will
> > > try to have a patch that adds more sanity checking in hand, though. We
> > > already do sanity checking if the mailbox changes at all, which helps mostly
> > > with the inbox. And we sanity check all the time if you're using the imap
> > > delete model.
> 
> David, do you think we still need a followup bug?

should we PM david about this one?

Comment 12

[:Irving Reid (No longer working on Firefox)]

If we're not getting bug reports related to a 5 year old change, I think we should just drop it.

Comment 13

[Kent James (:rkent)]

I agree with Irving, if there are no new reports of this then we should move on.