Description of problem:
I booted into multi-user.target and run 'systemctl isolate graphical.target'
Version-Release number of selected component:
cmdline: gnome-shell --mode=gdm
runlevel: 3 5
var_log_messages: [System Logs]:\n-- Logs begin at Tue 2014-07-08 10:59:30 CEST, end at Mon 2015-01-26 10:25:36 CET. --
Thread no. 1 (9 frames)
#0 g_str_hash at /lib64/libglib-2.0.so.0
#1 config_hash at /lib64/libmutter.so.0
#2 g_hash_table_lookup at /lib64/libglib-2.0.so.0
#3 meta_monitor_config_apply_stored at /lib64/libmutter.so.0
#4 meta_monitor_manager_constructed at /lib64/libmutter.so.0
#5 g_object_new_internal at /lib64/libgobject-2.0.so.0
#8 meta_backend_real_post_init at /lib64/libmutter.so.0
#9 meta_backend_x11_post_init at /lib64/libmutter.so.0
#10 meta_init at /lib64/libmutter.so.0
Created attachment 984156 [details]
Created attachment 984157 [details]
Created attachment 984158 [details]
Created attachment 984159 [details]
Created attachment 984160 [details]
Created attachment 984161 [details]
Created attachment 984162 [details]
Created attachment 984163 [details]
Created attachment 984164 [details]
Created attachment 984165 [details]
The crash was introduced in mutter-3.15.4-1.fc22 and gnome-shell-3.15.4-1.fc22. The previous versions work fine.
GNOME session aborts with above error upon login from LightDM on a current rawhide system.
The following upstream commit probably fixes this bug:
The 2015-02-02 Workstation live nightly hits the 'Oh no!' screen on boot, with this crash captured by abrt-cli - so this is a clear Alpha blocker: "Release-blocking live images must boot to the expected boot menu, and then to a desktop or to a login prompt where it is clear how to log in to a desktop." , https://fedoraproject.org/wiki/Fedora_22_Alpha_Release_Criteria#Expected_image_boot_behavior .
Hum, well, it does so in a VM, looking at the upstream commit, seems like it'll happen when no EDID is available, which seems most likely to happen in VMs and then maybe on broken monitors and/or TVs and projectors and things. So, maybe just a Beta blocker, but we can discuss it if it's not fixed by the next review meeting.
I also see this in my VM (qxl+spice).
So Adam backported the patch in comment #13 and I'm pretty sure that fixed this. At least, with 2015-02-02 I can't get past the crash, but 2015-02-04 works fine. Please reopen if you still hit this with mutter-3.15.4-2.fc22 or higher.
yeah, I also tested and confirmed the fix, just forgot to update the bug.