Hi,
I'm using Mutt 1.5.21+46 (b01d63af6fea) (2011-07-01) on a Debian
experimental, amd64 system. Built from source, using openssl instead
of gnutls to make debugging a little easier. muttrc contains
set folder=imaps://[email protected]@imap.gmail.com/
set spoolfile==INBOX
set header_cache=~/.mutt/cache/headers
set message_cachedir=~/.mutt/cache/bodies
set certificate_file=~/.mutt/certificates
Generally that works quite well. This message is about a small
but annoying quirk: often (after a timeout?) if I try to switch
folders, I will get the message "Error opening mailbox" and the list
of messages will go blank. When I reopen the folder I was in, all the
messages I had marked deleted are no longer marked deleted.
Trying to debug, I used "mutt -d 5" and added some extra debugging
messages:
[2011-07-17 16:07:33] Handling FETCH
[2011-07-17 16:07:33] FETCH response ignored for this message
[2011-07-17 16:07:33] imap_cmd_step: normal return, rc=1
[2011-07-17 16:07:33] 4< * 23 FETCH (UID 27078 FLAGS (\Seen))
[2011-07-17 16:07:33] Handling FETCH
[2011-07-17 16:07:33] FETCH response ignored for this message
[2011-07-17 16:07:33] imap_cmd_step: normal return, rc=1
[2011-07-17 16:07:33] SSL error: Connection reset by peer
[2011-07-17 16:07:33] imap_cmd_step: Error reading server response.
[2011-07-17 16:07:33] imap_cmd_step: rc=-1, msgno=23
[2011-07-17 16:07:33] imap_cmd_step error: msgno=23 rc=-1 mfhrc=0
[2011-07-17 16:07:38] Error opening mailbox
[2011-07-17 16:07:39] mutt_index_menu[601]: Got op 146
That is, mutt_socket_readln() fails when called from the
imap_cmd_step() loop, because the server decided to close the
connection.
The question: is there any way to recover from this? How are IMAPS
clients supposed to react to a reset connection? Is this a bug,
and if so, is it known?
Using Debian OpenSSL 1.0.0d-3; similar things happen with the stock
Debian mutt package (which uses gnutls), too.
Thanks for mutt, by the way. I love it.
Regards,
Jonathan
