I built a Fedora 19 live image from current repos, adding gnome-initial-setup-0.8-1.fc19.x86_64.rpm . GDM is gdm-3.7.91-1.fc19.x86_64 .
On first boot after install (with SELinux disabled - some SELinux borkage nerfs everything if you leave it enabled), I just see what looks like GDM with no users: a big dark grey rectangle with a panel at the top showing date, accessibility, and sound.
However, logs indicate that gnome-initial-setup is running; there are various messages from it in /var/log/messages and stuff. If I hit 'alt-tab', at least the first couple of times, I see a 'Initial Setup' icon in the alt-tab switcher.
So it looks like g-i-s is actually running, but it's somehow behind GDM? Just a guess.
This is in a virt-manager VM, with qxl/SPICE graphics. If I create a user at console and reboot, I still see the same blank grey screen - so what I'm seeing isn't exactly the GDM login screen, I guess.
AlphaBlocker: "A system installed with a graphical package set must boot to the 'firstboot' utility on the first boot after installation. The firstboot utility must be able to create a working user account."
OK, we need to re-word that for initial-setup, but it's clearly breaching the spirit.
I'm uploading my latest live image to assist in debugging this. It will be complete at:
in approx. 50 minutes (it's currently there, but still uploading, so incomplete). sha256sum is:
With this kind of stuff, the most useful is the journalctl output.
oh, sorry, thought I'd attached them. Will do shortly.
Created attachment 713305 [details]
journalctl from a system displaying the bug
GDM starts right around 12:00:07. I am in enforcing=0 , so the SELinux issues shouldn't be causing the problem, though obviously we'd want to fix them too.
The crash at 12:00:47 is me trying to fpaste this log and it being too big for the server's liking.
I'll file the eventual g-i-s crash (12:01:14) if I can get abrt to play ball.
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 . Accepted as a blocker per criterion cited in comment #1: it's possible that this might not happen on bare metal, if that's the case, we'll re-visit it. But for now we're assuming it's a showstopper.
https://bugzilla.redhat.com/show_bug.cgi?id=923908 is my best effort at filing the crash.
This patch from Jasper:
fixes this bug for me. However, the crash - https://bugzilla.redhat.com/show_bug.cgi?id=923908 - still happens: g-i-s runs for a while, then crashes. Also, the fourth panel of g-i-s - where you create a user - renders properly very very briefly (a split second), then the whole window just goes grey. I'll see if I can find any output specifically related to that.
The latest build fixes this, I've tested it, just waiting a bit until it hits the public repos before closing.
What's "the latest build"? I still see this problem with 0.8-3.
This is verified with g-s-i 0.8-3 AND gnome-shell 3.97.2-2.
Those are both in the repos now.