Bug 1102793

Created attachment 900381 [details]
File: anaconda-tb

Created attachment 900382 [details]
File: anaconda.log

Created attachment 900383 [details]
File: environ

Created attachment 900384 [details]
File: journalctl

Created attachment 900385 [details]
File: lsblk_output

Created attachment 900386 [details]
File: nmcli_dev_list

Created attachment 900387 [details]
File: os_info

Created attachment 900388 [details]
File: program.log

Created attachment 900389 [details]
File: storage.log

Created attachment 900390 [details]
File: ifcfg.log

Another user experienced a similar problem:

boot.iso 20140539 21.38-1 in Virtualbox 

cmdline:        /usr/bin/python  /sbin/anaconda
cmdline_file:   BOOT_IMAGE=vmlinuz initrd=initrd.img inst.stage2=hd:LABEL=Fedora\x20rawhide\x20x86_64 quiet
hashmarkername: anaconda
kernel:         3.15.0-0.rc7.git2.1.fc21.x86_64
package:        anaconda-21.38-1
product:        Fedora
reason:         Error: gtk-builder-error-quark: Invalid signal `toggled' for type `GtkCellRendererToggle' on line 207 (12)
release:        Cannot get release name.
version:        rawhide

Created attachment 900432 [details]
simple reproducer

Adding a small python file that reproduces the problem.

gtk3-3.13.2-1.fc21.x86_64
pygobject3-3.13.2-1.fc21.x86_64

This does not occur with gtk3-3.13.1-2

Also seen on ARM images when launching initial-setup-graphical.

systemctl status initial-setup-graphical -l
��● initial-setup-graphical.service - Initial Setup configuration program
   Loaded: loaded (/usr/lib/systemd/system/initial-setup-graphical.service; enabled)
   Active: inactive (dead) since Thu 2014-05-29 13:48:11 EDT; 5min ago
  Process: 626 ExecStart=/bin/xinit /bin/firstboot-windowmanager /bin/initial-setup -- /bin/Xorg :9 -ac -nolisten tcp (code=exited, status=0/SUCCESS)
  Process: 573 ExecStartPre=/bin/plymouth quit (code=exited, status=0/SUCCESS)
 Main PID: 626 (code=exited, status=0/SUCCESS)

May 29 13:48:09 localhost xinit[626]: File "/usr/lib/python2.7/site-packages/pyanaconda/ui/gui/spokes/datetime_spoke.py", line 122, in __init__
May 29 13:48:09 localhost xinit[626]: GUIObject.__init__(self, *args)
May 29 13:48:09 localhost xinit[626]: File "/usr/lib/python2.7/site-packages/pyanaconda/ui/gui/__init__.py", line 122, in __init__
May 29 13:48:09 localhost xinit[626]: self.builder.add_objects_from_file(self._findUIFile(), self.builderObjects)
May 29 13:48:09 localhost xinit[626]: GLib.Error: gtk-builder-error-quark: Invalid signal `toggled' for type `GtkCellRendererToggle' on line 207 (12)

*** Bug 1103093 has been marked as a duplicate of this bug. ***

Another user experienced a similar problem:

Selected English(UK) and hit continue in the installer

cmdline:        /usr/bin/python  /sbin/anaconda
cmdline_file:   inst.repo=http://www.mirrorservice.org/sites/dl.fedoraproject.org/pub/fedora/linux/development/rawhide/x86_64/os
hashmarkername: anaconda
kernel:         3.15.0-0.rc7.git2.1.fc21.x86_64
package:        anaconda-21.38-1
product:        Fedora
reason:         Error: gtk-builder-error-quark: Invalid signal `toggled' for type `GtkCellRendererToggle' on line 207 (12)
release:        Cannot get release name.
version:        rawhide

*** Bug 1103025 has been marked as a duplicate of this bug. ***

*** Bug 1103267 has been marked as a duplicate of this bug. ***

I've noticed this bug in the gtk builder parser recently - it doesn't parse <signal> properly if the signal is the only thing inside <object>

As a workaround, just add a

<property name="visible">True</property>

before the <signal> element.

In the intrest of moving forward, while gtk3 may not parse <signal> properly, does it "hurt" to just permanently patch anaconda by adding:

<property name="visible">True</property>

before the <signal> element?

Yeah, it would hurt, because the types that contain only a signal child in anaconda's glade files (GtkCellRendererToggle, GtkTreeSelection) don't have visible properties because the types don't inherit from GtkWidget.

Understood.  In that case, is there a fix for gtk3 coming "real soon now" or should this update be backed off?  With this on it makes it a little difficult to do an install other than TUI and, except when doing a kickstart install, I think I would find the TUI install to be a bit difficult.

Cell renderers do have a visible property (I tested the workaround before mentioning it here...). 

The underlying problem has been fixed upstream, I expect a gtk release in 2-3 weeks with that fix.

Two to three weeks with no GUI installs?  What is the plan to work around this because I am sure that two to three weeks of no gui installs willl go over like a lead balloon.

I've backported the fix to rawhide's gtk3-3.13.2-3.fc21 build.

I just built an (Xfce) live image with the updated gtk3, live install completed successfully. Fix looks good.