GNOME Bugzilla – Bug 513389
Evolution Mail and Calendar: Evolution crash sending ...
Last modified: 2013-09-13 00:59:15 UTC
Version: 2.22.x Evolution crash sending e-mails from an IMAP account. Starting evolution from a console, this is what I get: daniele@besa:~$ evolution (evolution:19204): GLib-GObject-WARNING **: IA__g_object_weak_unref: couldn't find weak ref 0xb64c9bd0(0x8b83168) (evolution:19204): GLib-GObject-WARNING **: IA__g_object_weak_unref: couldn't find weak ref 0xb64c9bd0(0x8c66478) (evolution:19204): GLib-GObject-WARNING **: IA__g_object_weak_unref: couldn't find weak ref 0xb64c9bd0(0x8c35cc8) ** Message: Response 1 (evolution:19204): GLib-GObject-WARNING **: IA__g_object_weak_unref: couldn't find weak ref 0xb64c9bd0(0x8c78b18) ** Message: Response 1 camel-CRITICAL **: camel_url_new: assertion `*url_string != 0' failed aborting... Trace/breakpoint trap (core dumped) Version is Evolution 2.21.90 Distribution: Ubuntu 8.04 (hardy) Gnome Release: 2.21.90 2008-01-29 (Ubuntu) BugBuddy Version: 2.20.1
Same problem here. Confirming. Really nasty bug :-/
gjc or daniele, can you provide a useful trace, please? (running evolution in gdb with "thread apply all bt")
Hi Andre, I find out that the same problem, with the same error: camel-CRITICAL **: camel_url_new: assertion `*url_string != 0' failed happens trying to add a new account. This should be easy to reproduce, I had it on 2 different boxes running Ubuntu Hardy. I found that starting evolution with: evolution --disable-eplugin solves the problem. I'll try to run the gdb. Thanks! -- Daniele
Created attachment 104213 [details] [review] proposed eds patch for evolution-data-server; This has been cause by fix for bug #500233. Unfortunately on some distros the G_DEBUG can be set too "sensitive", so it crashes. Anyway, it is possible to create only-to-send account in evo, in which case it has url as empty string. This patch only checks for this case and returns safely.
Daniele, can you try to get stack trace for the weak_unref message, please? Maybe with 'export G_DEBUG=fatal_warnings' it will stop immediately when that happened. Can you provide also steps for it? I guess those messages come earlier then in send process. Thanks in advance.
G_DEBUG is not "too sensitive". it's up to every developer to carefully review his patches.
Milan, I think in this case, you can go ahead with if-null-return for both the ones. I don't see any issue. So change the other one also and commit.
Thanks to everyone, latest Ubuntu Hardy packages solved the problem. So the problem is now gone. Thanks -- Daniele
Committed to trunk (slightly modified). Committed revision 8453.
*** Bug 517131 has been marked as a duplicate of this bug. ***
*** Bug 516610 has been marked as a duplicate of this bug. ***
*** Bug 516598 has been marked as a duplicate of this bug. ***
*** Bug 513493 has been marked as a duplicate of this bug. ***