Description of problem: If you install F19 and boot, you'll see a 'sad face' error. I've found out it comes from gnome-initial-setup. I have tried it just in VMs, but for many times and it seems to happen always. I don't know how to debug this and provide more info, instructions welcome. Version-Release number of selected component (if applicable): gnome-initial-setup-0.7-1.fc19.x86_64 gdm-3.7.91-1.fc19.x86_64 gnome-shell-3.7.91-1.fc19.x86_64 How reproducible: always Steps to Reproduce: 1. boot F19 in a VM
Created attachment 712242 [details] failure screenshot
Created attachment 712243 [details] journalctl output
Hmm. This bit looks wrong: if [ $1 -eq 1 ] ; then # Initial installation /usr/bin/systemctl enable gnome-initial-setup.service &>/dev/null || : fi Do we really want g-i-s enabled by default when installed in Fedora at this time? I don't know what the intersection plan is between anaconda/firstboot/g-i-s or if there is even a plan, but regardless I suspect installing the package shouldn't enable it, it should be done explicitly by anaconda.
I think we should keep this bug related to the crash and discuss the startup conditions in bug 922885.
colin: comps has gnome-initial-setup as part of the GNOME group and firstboot in all the other desktop groups. christ alone knows what'll happen if you install two desktops, but it should work correctly in single-desktop cases (or 'any combination not including GNOME' cases). kamil: 0.8 fixes the sadface but gives you https://bugzilla.redhat.com/show_bug.cgi?id=923501 in exchange.
Discussed at 2013-03-20 blocker review meeting: http://meetbot.fedoraproject.org/fedora-blocker-review/2013-03-20/f19alpha-blocker-review-2.2013-03-20-16.00.log.txt . This is obviously a blocker, but we believe it's probably 'fixed' in 0.8 (0.8 breaks differently! see https://bugzilla.redhat.com/show_bug.cgi?id=923501 ), so Kamil is going to verify that 0.8 behaves the same for him as for me, then we will close this bug.
Still just waiting for Kamil's confirmation, but I've tested the current 0.8-3 build and it works fine for me.
With 0.8-3 this crash is gone, but I see bug 923501 instead.