If I don't log in quickly enough after boot, or log out and try to log back in again, the login fails with the message: "Oh no! Something has gone wrong." All subsequent attempts to log in fail in the same way. Looking in .xsession-errors, I see: *** glibc detected *** /usr/lib64/gnome-settings-daemon: free(): invalid pointer: 0x00007fcfc4b4f380 *** ======= Backtrace: ========= /lib64/libc.so.6(+0x73476)[0x7fcfde401476] /usr/lib64/libgnomekbd.so.7(gkbd_keyboard_config_equals+0x15f)[0x7fcfc494bb4f] /usr/lib64/gnome-settings-daemon-3.0/libkeyboard.so(+0x4a45)[0x7fcfc4d6ba45] /usr/lib64/gnome-settings-daemon-3.0/libkeyboard.so(+0x4bb4)[0x7fcfc4d6bbb4] /usr/lib64/gnome-settings-daemon-3.0/libkeyboard.so(gsd_keyboard_xkb_init+0x21a)[0x7fcfc4d6c11a] /usr/lib64/gnome-settings-daemon-3.0/libkeyboard.so(+0x41fd)[0x7fcfc4d6b1fd] /lib64/libglib-2.0.so.0(g_main_context_dispatch+0x16a)[0x7fcfde97d9ba] /lib64/libglib-2.0.so.0(+0x47d80)[0x7fcfde97dd80] /lib64/libglib-2.0.so.0(g_main_loop_run+0x6a)[0x7fcfde97e17a] /usr/lib64/libgtk-3.so.0(gtk_main+0x8d)[0x7fcfdf35aa3d] /usr/lib64/gnome-settings-daemon(main+0x1ea)[0x40386a] /lib64/libc.so.6(__libc_start_main+0xed)[0x7fcfde3af32d] /usr/lib64/gnome-settings-daemon[0x4039d9] This bug appeared immediately after doing an update. Previously gnome-desktop3-3.3.91-1.mga2 was installed, which didn't show this problem. I've tested this with a newly created user (i.e. empty home directory), and the same problem occurs.
I can confirm this bug. My test system has both KDE4 and Gnome installed and was defaulted to boot up into KDE4. In KDE4 I logged in, updated, then rebooted then logged into KDE4 again, successfully. I then logged out of KDE4 and attempted to log-in using Gnome and got the same full screen error message reported here. Subsequent testing tells me that so long as Auto Log-in is disabled I can successfully log in to KDE4. For now I am going to leave Auto log-in turned off.
CC: (none) => wilcal.int
should be fixed in libgnomekbd-3.4.0-2.mga2
Status: NEW => RESOLVEDCC: (none) => olavSee Also: (none) => https://bugzilla.gnome.org/show_bug.cgi?id=672729Resolution: (none) => FIXEDAssignee: bugsquad => olavSource RPM: gnome-desktop3-3.3.92-1.mga2 => libgnomekbd
Seems not fixed (after update) or I am missing something. [bogdan@localhost ~]$ rpm -qa | grep libgnomekbd libgnomekbd-common-3.4.0-2.mga2 Looking in the upstream bug discussion, first they picked a wrong commit ... wonder in these circumstances whether your fix was correct. Thanks Olav for your commitment. Bug 5097 and Bug 5099 seems to be (at least partly) related.
Status: RESOLVED => REOPENEDCC: (none) => gruescubogdanResolution: FIXED => (none)
I agree; this is not fixed for me either.
Version 3.4.0.1 you made now available, but still no joy (well, Alt+F4 and GNOME will appear ... sort of). [bogdan@localhost ~]$ rpm -qa | grep gnomekbd libgnomekbd-common-3.4.0.1-1.mga2 lib64gnomekbd-gir3.0-3.4.0.1-1.mga2 lib64gnomekbd7-3.4.0.1-1.mga2 [bogdan@localhost ~]
Could you attach ~/.xsession-errors again with the latest libgnomekbd updates (3.4.0.1)? In case something crashed, I'd like a stack trace. Meaning: from a terminal do something like: gdb /usr/lib64/gnome-settings-daemon then gdb starts; it might suggest some debug packages; install them within gdb do: r, <enter> then when something crashes: bt, <enter> attach what bt gives please.
Created attachment 1850 [details] .xsessions-errors after failed login with 3.4.0.1
Created attachment 1851 [details] gdb output as requested The crash occurs immediately after typing r, <enter>.
Attachment 1851 mime type: application/octet-stream => text/plain
Cool, quick response! Could you install the following and provide another stack trace? - libgnomekbd-debug - gnome-settings-daemon-debug - glib2.0-debug Thanks
Sorry Olav your second request was not answered promptly, probably the stack trace is no longer necessary. Olivier Blin has added a patch (thanks) and GNOME is operational again (and hope it will stay so with the forthcoming version 3.4.0.2). I think we can assume this bug and the duplicated Bug 5097 as fixed (didn't dare to close them :-)).
Sorry Olav, your second request arrived after I'd left for work. I can confirm that the latest updates have fixed this bug. Thanks for your efforts.
Status: REOPENED => RESOLVEDResolution: (none) => FIXED
Latest updates for me have fixed this bug
*** Bug 5097 has been marked as a duplicate of this bug. ***
CC: (none) => nextwavecs