Bug 1001508

Created attachment 790822 [details]
File: backtrace

Created attachment 790823 [details]
File: cgroup

Created attachment 790824 [details]
File: core_backtrace

Created attachment 790825 [details]
File: dso_list

Created attachment 790826 [details]
File: environ

Created attachment 790827 [details]
File: exploitable

Created attachment 790828 [details]
File: limits

Created attachment 790829 [details]
File: maps

Created attachment 790830 [details]
File: open_fds

Created attachment 790831 [details]
File: proc_pid_status

Created attachment 790832 [details]
File: var_log_messages

*** Bug 1032257 has been marked as a duplicate of this bug. ***

Another user experienced a similar problem:

claws-mail has been minimized to the system tray and closed itself.
set 2 on Imap mailboxes and 1 on pop3.

reporter:       libreport-2.1.10
backtrace_rating: 4
cmdline:        claws-mail
crash_function: imap_threaded_noop
executable:     /usr/bin/claws-mail
kernel:         3.12.5-302.fc20.x86_64
package:        claws-mail-3.9.2-7.fc20
reason:         claws-mail killed by SIGSEGV
runlevel:       N 5
type:           CCpp
uid:            1000

Another user experienced a similar problem:

One IMAP server timed out, Claws-Mail displays a pop-up and hangs in such situation. So after some time I came home and closed that window, then tried to open new message in another mailbox (on another server), which made whole client crash.

reporter:       libreport-2.1.12
backtrace_rating: 4
cmdline:        claws-mail
crash_function: imap_threaded_noop
executable:     /usr/bin/claws-mail
kernel:         3.12.9-301.fc20.x86_64
package:        claws-mail-3.9.3-1.fc20
reason:         claws-mail killed by SIGSEGV
runlevel:       N 5
type:           CCpp
uid:            1000

Just a note the bug reporters here, I've given up trying to handle bug reports about Claws Mail at Fedora. Upstream only wants to learn about reproducible bugs. They are aware of old bugs causing crashes as a side-effect, and in my opinion IMAP access instabilities cause side-effects in Claws Mail.

http://www.thewildbeast.co.uk/claws-mail/bugzilla/index.cgi
http://www.claws-mail.org/

Another user experienced a similar problem:

Claws crash in background without any user assistance.

reporter:       libreport-2.1.12
backtrace_rating: 4
cmdline:        claws-mail
crash_function: imap_threaded_noop
executable:     /usr/bin/claws-mail
kernel:         3.12.9-301.fc20.x86_64
package:        claws-mail-3.9.3-1.fc20
reason:         claws-mail killed by SIGSEGV
runlevel:       N 5
type:           CCpp
uid:            1000

> Claws crash in background without any user assistance.

If that's really the only details about this, then it likely fixed upstream by the recent imap_destroy_all related commits.

This will need retesting once a new version of Claws Mail (or a Test Update) will be available.

A new release of Claws Mail (and its dependency libetpan) is available for testing with Fedora 20 and Fedora 19:

  https://admin.fedoraproject.org/updates/search/claws

Please allow for the time it takes for the packages to show up in the "updates-testing" repository and then to be picked up by the world-wide download mirror servers.

This bug happened again, as Abrt pointed me to this report.
Possible duplication: bug #1032257

claws-mail-3.10.1-1.fc20.x86_64

> Abrt pointed me to this report.
> Possible duplication: bug #1032257

That's not helpful, especially as it's not clear how good the dupe detection is.

A full backtrace will be needed before anything could be guessed. There is nothing like "this bug", because it has not been a single specific bug. As the ticket title suggests there have been side-effects (= crashes in various parts of the code) as a result of Claws Mail working on already destroyed/freed memory structures. There have been dozens of different tickets in bugzilla about it. It has taken a lot of time to track down a possible cause and get upstream to look into it, too. As of Claws Mail 3.10.1 several such issues have been fixed, but some [possibly different ones] remain, which might take down Claws Mail in similar ways. There is ongoing work on tracking down non-thread-safe parts, for instance.

With the changes in Claws Mail 3.10.1 we need to restart evaluating it, since we cannot keep all old tickets open which may only be the result of side-effects.

If you could generate a full backtrace in ABRT and attach it in a new ticket, that could lead to something.