288534 - Autosaved mails never removed from draft folder

Status: RESOLVED FIXED

(MailNews Core :: Networking: IMAP, enhancement)

Description

[Christian :Biesinger (don't email me, ping me on IRC)]

when I autosave my mail every X minutes, I get one mail per saving in my drafts
folder. I would expect that later autosaves overwrite the previous versions...

It would also be nice if sending the mail would remove the draft(s) it saved.

Comment 1

[Reuben Farrelly]

Seeing this one here also, on trunk from last night (08/05).

I'm using Windows, and signing with a secure certificate.  Otherwise a fairly
standard install.

Suggest this bug gets marked as a 'normal' priority rather than an enhancement,
marked as a cross platform bug, and made a blocker for aviary 1.1 ?

Comment 2

[Reuben Farrelly]

Some observations:

* It makes no difference if I use certificates or not, so this is not a
certificate issue

* I can replicate the behaviour by manually selecting File->Save.  It's not
autosave related, as such, every time a save occurs another copy is created.

* The Drafts folder is, along with the other folders in this account, stored on
an IMAP server.  I think this is the killer here, because if the Drafts folder
is stored locally, this bug does not show up, and only one copy appears in the
local Drafts folder.

So, it seems like the bug is actually a case of bad handling of stuff that is
written to the Drafts folder rather than anything to do with autosave, but only
if that folder is not stored locally.  Can the original bug reporter
([e-mail]) please confirm that this is the case there too?

Comment 3

[Christian :Biesinger (don't email me, ping me on IRC)]

Oh, that's true. I can confirm that: A drafts folder in Local Folders does not
show this bug.

To networking:imap then, I guess.

(fwiw, you should not set blocking+; that's only set by aviary project
management, whoever that is)

Comment 4

[David :Bienvenu]

Does this happen for every message you save as draft? Can you generate an imap
protocol log and attach or send to me?

http://www.mozilla.org/quality/mailnews/mail-troubleshoot.html#imap

Comment 5

[Christian :Biesinger (don't email me, ping me on IRC)]

yes, it happens for every message. I sent the log to your bugzilla mail address.
my IMAP server is dovecot btw (as can be seen in the log too).

Comment 6

[Christian :Biesinger (don't email me, ping me on IRC)]

oh, I should mention...
the actions leading to it were: compose a new mail message, enter some text.
click Save. (enter passwords for S/MIME and the IMAP server). enter some more
text. click save again. enter more text and click save a third time.

I ended up with three mails in my Drafts folder.

Comment 7

[David :Bienvenu]

thx for the log. It looks like dovecot doesn't support uidplus, which is
optional, but it also doesn't seem to support SEARCH, which is our fallback for
servers that don't suport uidplus. Does search in general work against this
server, or does it not support SEARCH at all? The problem is that after
appending to the draft folder, we try to find out what the UID of the message we
appended was by searching for it, using SEARCH, and the server doesn't return
the UID. So it seems like the server is broken.

Comment 8

[Christian :Biesinger (don't email me, ping me on IRC)]

hm... search does seem to be supported ("Search Messages" for subject substrings
does show results)

is the problem maybe the SEEN part of the search command?

Comment 9

[David :Bienvenu]

Oy, good catch. Someone changed it so new drafts are not marked read, so they
wouldn't be seen...

Comment 10

[David :Bienvenu]

Attachment: proposed fix

stop searching for just SEEN messages. Also, issue NOOPS before searching for
appended drafts, or deleting appended drafts. This fixes problems I reproduced
with several courier servers. NOOPS are cheap, so this should be harmless.

Comment 11

[Scott MacGregor]

Comment on attachment 183314 [details] [diff] [review]
proposed fix

gotta love the Noop.

Comment 12

[Christian :Biesinger (don't email me, ping me on IRC)]

the patch is working great. thanks!

Comment 13

[chris hofmann]

Comment on attachment 183314 [details] [diff] [review]
proposed fix

a=chofmann

Comment 14

[Scott MacGregor]

*** Bug 292302 has been marked as a duplicate of this bug. ***

Comment 15

[Adam Guthrie]

*** Bug 307028 has been marked as a duplicate of this bug. ***