pidgin hangs on startup

[Steve Magoun]

After upgrading to a recent set of bits in raring, pidgin reliably hangs on startup. This is a regression from 12.10.

To reproduce:
1) Launch pidgin

Expected results:
The Buddy List window loads a list of buddies (IRC channels in my case), and pidgin opens a new chat window to auto-join a set of IRC channels

Actual results:
The Buddy List window is created but it's completely blank - there is no content at all, not even the account status popup menu. The OS greys out the Buddy List window. Pidgin is completely unresponsive.

GDB shows pidgin is hung in purple_gnome_proxy_get_parameter:

Thread 1 (Thread 0x7f3629f809c0 (LWP 2713)):
#0 0x00007f362681339d in poll () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007f362738b87c in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00007f362738bd52 in g_main_loop_run () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007f36273cd3e1 in g_spawn_sync () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007f36273cd858 in g_spawn_command_line_sync () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007f36270942ad in purple_gnome_proxy_get_parameter (gnome_version=<optimized out>, parameter=<optimized out>)
    at /build/buildd/pidgin-2.10.6/./libpurple/proxy.c:270
#6 purple_gnome_proxy_get_parameter (parameter=<optimized out>, gnome_version=<optimized out>)
    at /build/buildd/pidgin-2.10.6/./libpurple/proxy.c:260
#7 0x00007f3627094920 in purple_gnome_proxy_get_info () at /build/buildd/pidgin-2.10.6/./libpurple/proxy.c:307
#8 purple_proxy_get_setup (account=<optimized out>) at /build/buildd/pidgin-2.10.6/./libpurple/proxy.c:2257
#9 0x00007f36270a5102 in initiate_resolving (data=0x7f362bdf6040) at /build/buildd/pidgin-2.10.6/./libpurple/dnsquery.c:896
#10 0x00007f362738c1ab in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#11 0x00007f362738b5a5 in g_main_context_dispatch () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#12 0x00007f362738b8e8 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#13 0x00007f362738bd52 in g_main_loop_run () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#14 0x00007f3628601f17 in gtk_main () from /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0
#15 0x00007f3629ffd119 in main (argc=1, argv=0x7fffaf9259d8) at /build/buildd/pidgin-2.10.6/./pidgin/gtkmain.c:933

ProblemType: Bug
DistroRelease: Ubuntu 13.04
Package: pidgin 1:2.10.6-0ubuntu4
ProcVersionSignature: Ubuntu 3.8.0-2.6-generic 3.8.0-rc4
Uname: Linux 3.8.0-2-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.8-0ubuntu2
Architecture: amd64
Date: Tue Jan 29 15:13:17 2013
InstallationDate: Installed on 2010-09-17 (865 days ago)
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Beta amd64 (20100901.1)
MarkForUpload: True
ProcEnviron:
 TERM=xterm
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: pidgin
UpgradeStatus: Upgraded to raring on 2013-01-25 (4 days ago)

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[zarquon42 (zarquon42)]

Same here, I purged, reinstalled and deleted my pidgin user profile - but in 9 out of 10 times pidgin remains unresponsive (I only have one active jabber account).

[Merelin (merelin)]

The same here. But I seem to have found a workaround. Open a terminal window and run "nohup pidgin &". Pidgin does not hang on startup and remains working after the terminal window is closed.

[zarquon42 (zarquon42)]

It does not work here: still unresponsive

[bwat47 (bwat47)]

I can reproduce this happening this way:

When I first boot ubuntu 13.04 and open pidgin its fine. If I log out, log back in, and open pidgin it always hangs immediately. Also tried clearing my pidgin config and such, no help.

[zarquon42 (zarquon42)]

do you have pidgin-openpgp installed? Removing this package seemed to have solved this problem for me (at least it is working for the last 15 minutes, including two restarts)

[zarquon42 (zarquon42)]

Nope, it hangs again - after 2hours it is back to hang at startup

[zarquon42 (zarquon42)]

Fiddling about brought some insight. When pidgin hangs, kill it and restart indicator-messages-service with the command 'killall indicator-messages-service'. Then, eventually, pidgin works ... for some time.

[zarquon42 (zarquon42)]

After two days of testing, the following workaround seems to work in most cases:
1. When starting pidgin, it always hangs, then kill it,
2. issue the command 'killall indicator-messages-service'
3. wait 2-3 minutes
4. start pidgin

[Arthur H (arthurh)]

A bit of info from my testing:

Updating to 2.10.7 does not resolve the issue (current version in raring repo, as of today's date).

Killing indicator-messages-service as suggested above also does not resolve the issue for me.

'nohup'ping and tossing in the background as suggested above does not help.

Removing ~/.purple and re-configuring the account does seem to allow pidgin to work (tried twice) for that launch. Exiting and relaunching then exhibits the hang.

[zarquon42 (zarquon42)]

Sometimes I have to try the 'indicator-killing' several times until pidgin is working.
The strange thing, however, is the erratic behaviour.
I consider this bug relatively critical, as there is no true alternative to pidgin - empathy is just a pile of garbage: huge buddy icons make it unpossible to use it when more than 5 buddies are online, no encryption, no customisability at all (the new gnome usability axiom)

[bwat47 (bwat47)]

@zarquon, agreed, and even worse empathy is constantly unreliable when it comes to connecting to accounts, it loves to randomly disconnect or refuse to connect at all, and its gotten even more unreliably since the introduction of UOA's buggy single-sign-on. And in raring empathy/UOA is also broken as all hell right now leaving no good IM solution for raring: https://bugs.launchpad.net/ubuntu/+source/empathy/+bug/1116449

I might have to end up skipping 13.04 entirely if it gets released with these bugs unfixed.

[Arthur H (arthurh)]

I seem to have no hang issues if I don't add an XMPP account.

After adding only AIM, Yahoo and MSN accounts, I can use pidgin, close pidgin and relaunch pidgin without issue.

After adding an XMPP account, I can use the account until I close and relaunch pidgin, at which point, I get the crash. Even tried to use finch and the behaviour seems the same. Looks like it's with the XMPP connection routines / libs, at least for me. I'll try and get stack traces and whatnot when I have a bit more time.

[zarquon42 (zarquon42)]

Hi, i disabled all xmpp accounts and only enabled irc - it hangs less often, but still hangs. Always at:
...
(20:35:17) Session Management: Using pidgin as command
(20:35:17) pidgin-libnotify: Updating status set from pidgin to 'offline'
(20:35:17) util: requesting to fetch a URL

[zarquon42 (zarquon42)]

OK - when I disable network connections, pidgin starts without hanging, but hangs, as soon as I connect to the internet and set pidgin to online too (sometimes, DE becomes irresponsible and only ctrl+alt+f1 and killall -9 pidgin makes can solve it)

[Arthur H (arthurh)]

One more tidbit --

Pidgin hangs with the 3 above accounts configured after sleep and resume. I have to kill pidgin and restart for it to re-connect.

[zarquon42 (zarquon42)]

This seems to be a duplicate of https://bugs.launchpad.net/pidgin/+bug/1108056, and pidgin is waiting for gsettings, which is non-responsive (also see https://developer.pidgin.im/ticket/15493)

[Simon K (octav14n)]

I can confirm, that in !my! case https://bugs.launchpad.net/pidgin/+bug/1108056 would have been the right bug.

[Arthur H (arthurh)]

Confirmed. This is also the root cause of my issue.

[bwat47 (bwat47)]

I recently installed xubuntu raring and pidgin works totally fine there, so definitely something to do with gnome/unity.

[Arthur H (arthurh)]

@bwat47, looks like it is the issue linked by zarquon42, although things are now working after latest round of apt-get updates so they might have fixed the gnome based gsettings bug that was causing the issue with pidgin. All my accounts seem to be good and well at the moment, with the exception of after the sleep/resume cycle, at which point pidgin hangs and the screen gets grayed out.

[Arthur H (arthurh)]

Scratch that, it's broken again. I've been working all day without issue, multiple reboots, closes and relaunches, no hiccups, and as I swtiched back my pidgin window from writing my last post, it was gray.

[bwat47 (bwat47)]

Do you think the empathy issue I linked earlier could be related? The symtoms sound similar, at first the empathy accounts work fine, and then after a while they totally refuse to work, and trying to set one up in UOA its just totally blank under "enter account details". Maybe they could be hitting the same gsettings bug?

[Arthur H (arthurh)]

I'm not sure. I'm running into a few problems. I JUST ran an update (doing so every few hours) and got a new pidgin and supporting libraries. Empathy seemed to let me view the AIM account credentials in the account manager now. Pidgin still hangs randomly.

Pidgin seems to be freezing at a few different places (by way of the -d argument to pidgin). Either at the URL fetch, or doing SRV lookups, or sometimes when trying to resolve my gateway's hostname.

When it hangs at the fetch, debug output gives me:

util: requesting to fetch a URL

When it hangs at the SRV lookup (for xmpp), I hang at:

dnssrv: found 1 SRV entries

On my last launch -- at the hang point -- I got the gdb full thread stack trace (thread apply all bt):

Thread 5 (Thread 0x7fffd6009700 (LWP 4823)):
#0 0x00007ffff484b3cd in poll () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007ffff53c415c in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00007ffff53c4284 in g_main_context_iteration ()
   from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007ffff53c42e1 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007ffff53e8075 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007ffff4b2df8e in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#6 0x00007ffff4857e1d in clone () from /lib/x86_64-linux-gnu/libc.so.6

Thread 4 (Thread 0x7fffd680a700 (LWP 4821)):
#0 0x00007ffff4b3205e in pthread_cond_timedwait@@GLIBC_2.3.2 ()
   from /lib/x86_64-linux-gnu/libpthread.so.0
#1 0x00007ffff5403bf5 in g_cond_wait_until () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00007ffff5399bb1 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007ffff539a1fa in g_async_queue_timeout_pop ()
   from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#4 0x00007ffff53e8872 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007ffff53e8075 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#6 0x00007ffff4b2df8e in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#7 0x00007ffff4857e1d in clone () from /lib/x86_64-linux-gnu/libc.so.6

Thread 3 (Thread 0x7fffeaa55700 (LWP 4820)):
#0 0x00007ffff484b3cd in poll () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007ffff53c415c in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00007ffff53c463a in g_main_loop_run () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007ffff2f0b726 in ?? () from /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0
#4 0x00007ffff53e8075 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007ffff4b2df8e in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#6 0x00007ffff4857e1d in clone () from /lib/x86_64-linux-gnu/libc.so.6

Thread 2 (Thread 0x7fffeb256700 (LWP 4819)):
#0 0x00007ffff484b3cd in poll () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007ffff53c415c in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#2 0x00007ffff53c4284 in g_main_context_iteration ()
   from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#3 0x00007fffeb25d9cd in ?? ()
   from /usr/lib/x86_64-linux-gnu/gio/modules/libdconfsettings.so
#4 0x00007ffff53e8075 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5 0x00007ffff4b2df8e in start_thread () from /lib/x86_64-linux...

[zarquon42 (zarquon42)]

@pablo,
this has no effect in my case in 3 out of 4 cases pidgin still hangs either at
dnssrv: found 2 SRV entries or

util: requesting to fetch a URL

[Pablo Castellazzi (pcastellazzi)]

@zarquon42
you are right, at first it seemed to work, but i think the problem is much more complicated than we originaly though. In my case the hang up is not related to a network issue. If i run pidgin -n it starts ok, but when i press Ctrl-C on the terminal sometimes it hang up. strace output is the same in both cases (with or without -n).

[pid 3085] restart_syscall(<... resuming interrupted call ...> <unfinished ...>
[pid 3082] restart_syscall(<... resuming interrupted call ...> <unfinished ...>
[pid 3081] restart_syscall(<... resuming interrupted call ...> <unfinished ...>
[pid 3078] restart_syscall(<... resuming interrupted call ...> <unfinished ...>

[Roy Kishun (roy-kishun)]

I have the same symptom where pidgin launches and the buddy list just hangs.... The workaround for me is to run it in debug mode "pidgin -d" and it works fine.

[Arthur H (arthurh)]

@Roy I think the debug flag working around the deadlock is coincidental, I had several instances where everything worked and then stopped (multiple times, and worked for several hours on each). You may want to be aware of this. In any event, I can't provide any more info on this bug as I've switched distributions and will be unsubscribing from this bug. Thanks for the help, guys!

[garnus (garnus)]

Hi,
im on 12.04 and i have the same problem.

[Fedik (fedikw)]

same here ...

BUT

it alway (here) works if run from terminal:
~$ pidgin -d

but stuck on startup if:
~$ pidgin

[aso (plasiek)]

Same in me, sometimes Pidgin start, sometimes not.

(15:36:51) account: Connecting to account (my XMPP)/.
(15:36:51) connection: Connecting. gc = 0x7f66324d6300
(15:36:51) gg: // gg_resolver_run() count = 4
(15:36:51) dnssrv: querying SRV record for aqq.eu: _xmpp-client._tcp.aqq.eu
(15:36:51) util: requesting to fetch a URL

And freeze.
This in Debian 7.0 and Pidgin 2.10.7 (libpurple 2.10.7), I post it here because Ubuntu is similar to Debian.

[JaSK (jask)]

I have the same issue (crash on startup, sometimes it works but then it will stop after a few hours (maybe when it reconnects an account?)

deleted ~/.purple/blist.xml and now it works

no idea if coincidence or not, but some people wrote somewhere that deleting .purple helps so maybe it's one of the files there getting corrupted.
I don't see how this can be related to unity indicator stuff though...
Will post again whether that helps next time it crashes too.

[JaSK (jask)]

god damn, sorry, wrong thread. didn't see the "duplicate" flag
sorry, never reported a bug before :p