GNOME Bugzilla – Bug 707884
Runtime error: cannot register existing type EABContactDisplay
Last modified: 2013-09-12 07:00:03 UTC
I have setup a SOGo server and I'm using Evolution under Ubuntu Raring. Basic usage seems to work reasonably well, but every time I try to update (or add, or delete) any e-mail addresses in any CardDAV contact, Evolution crashes. The information I get in the console after starting with WEBDAV_DEBUG=all is: ---------------------------- 8< ---------------------------- (evolution:13349): GLib-GObject-WARNING **: cannot register existing type `EABContactDisplay' (evolution:13349): GLib-CRITICAL **: g_once_init_leave: assertion `result != 0' failed ** (evolution:13349): ERROR **: eab-contact-commit-duplicate-detected.ui: Invalid type function on line 61: 'eab_contact_display_get_type' Trace/breakpoint trap (core dumped) ---------------------------- >8 ---------------------------- I'm not sure if it's easy to upgrade Evolution to the latest version to give it a try, or if there's anything else you would like me to do to investigate the problem...
If you could try to test 3.8 on a VM with Fedora 19 and see if the problem still exists there that would be awesome.
(In reply to comment #0) > (evolution:13349): GLib-GObject-WARNING **: cannot register existing type > `EABContactDisplay' Thanks for a bug report. The above is the main issue, once you fix it, you get things working properly. I do not think it has anything to do with the WebDAV backend, because it's evolution crashing for you, not the e-addressbook-factory (where all the WebDAV communication happens. The thing is that this can be an issue on a build side, and/or also on a runtime side. Please see bug #659890 for more information, though there did not land any complete fix for it yet. *** This bug has been marked as a duplicate of bug 659890 ***
In case it helps, it seems to work perfectly fine in Evolution 3.8.3 under Fedora 19's live CD... Should I add this information to bug 659890?
Yeah, that's the thing why I suspect a build system being involved here. Fedora's build environment doesn't suffer of the issue, while some other (like on Ubundu) do suffer of this. It sometimes happens that what Fedora survives without issues, other build environments claim as errors (like undefined symbols - I noticed claims from Debian and Ubuntu). If I don't forget, I will merge the two source parts into one for 3.12, which, I hope, will (partially) fix the issue (as mentioned in the bug #659890; I would do it earlier, but I always forget.)