Created attachment 817212 [details] python traceback Description of problem: Starting virt-manager as root silently fails. Using --no-fork shows the attached traceback. Version-Release number of selected component (if applicable): virt-manager-0.10.0-4.git79196cdf.fc21.noarch How reproducible: 100% Steps to Reproduce: 1. become root 2. execute virt-manager Actual results: silent failure (or traceback if --no-fork is specified_ Additional info: See attachment for traceback
Aargh, it seems like this isn't limited to root; hopefully this is PBKAC, but virt-manager is non-functional on my rawhide system. I can provide access to the VM.
The exception is raised by gtk.TreeViewColumn.set_cell_data_func when the callback function is set to NULL, though I am not sure if it is an error there since userdata is also NULL so the error doesn't make sense. On the other hand, if I understand correctly the code, setting the callback to NULL is a NOP so we can just avoid this in virt-manager by skipping the call to set_cell_data_func when the callback function is NULL. Patch proposed upstream: https://www.redhat.com/archives/virt-tools-list/2013-October/msg00166.html
fixed in pygobject upstream: https://bugzilla.gnome.org/show_bug.cgi?id=711173
Here's the commit: commit 79aea2655db11bc9d2c0ad75c87862b2b66da594 Author: Simon Feltman <sfeltman.org> Date: Mon Nov 4 03:29:57 2013 -0800 Remove overzealous argument checking for callback userdata It just missed pygobject 3.11.1
I think this is the same error, but this extract from the log file might be useful: [Sat, 09 Nov 2013 17:01:32 virt-manager 30088] DEBUG (cli:217) Launched with command line: /usr/share/virt-manager/virt-manager [Sat, 09 Nov 2013 17:01:32 virt-manager 30088] DEBUG (virt-manager:180) virt-manager version: 0.10.0 [Sat, 09 Nov 2013 17:01:32 virt-manager 30088] DEBUG (virt-manager:181) virtManager import: <module 'virtManager' from '/usr/share/virt-manager/virtManager/__init__.pyc'> [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (virt-manager:226) GTK version: 3.11.0 [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] ERROR (importer:51) Could not find any typelib for AppIndicator3 [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (engine:439) libguestfs inspection support: True [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (inspection:82) waiting [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (systray:163) Showing systray: False [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (engine:231) About to connect to uris ['qemu:///system'] [Sat, 09 Nov 2013 17:01:32 virt-manager 30089] DEBUG (cli:212) Uncaught exception: Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/engine.py", line 133, in _activate self.show_manager() File "/usr/share/virt-manager/virtManager/engine.py", line 862, in show_manager self._do_show_manager(None) File "/usr/share/virt-manager/virtManager/engine.py", line 806, in _do_show_manager manager = self.get_manager() File "/usr/share/virt-manager/virtManager/engine.py", line 767, in get_manager obj = vmmManager() File "/usr/share/virt-manager/virtManager/manager.py", line 173, in __init__ self.init_stats() File "/usr/share/virt-manager/virtManager/manager.py", line 276, in init_stats self.toggle_host_cpu_usage_visible_widget() File "/usr/share/virt-manager/virtManager/manager.py", line 1026, in toggle_host_cpu_usage_visible_widget self.host_cpu_usage_img, "menu_view_stats_host_cpu") File "/usr/share/virt-manager/virtManager/manager.py", line 1003, in _toggle_graph_helper col.set_cell_data_func(img, datafunc, None) File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1320, in set_cell_data_func super(TreeViewColumn, self).set_cell_data_func(cell_renderer, func, func_data) TypeError: When passing None for a callback userdata must also be None
Fixed in virt-manager-0.10.0-5.git1ffcc0cc.fc21
This does indeed fix that traceback, unfortunately I get another one now: [Sun, 10 Nov 2013 10:37:51 virt-manager 1365] DEBUG (cli:212) Uncaught exception: Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/manager.py", line 710, in add_conn row = self._append_conn(vmlist.get_model(), conn) File "/usr/share/virt-manager/virtManager/manager.py", line 691, in _append_conn _iter = model.append(None, row) File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1194, in append return self._do_insert(parent, -1, row) File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1187, in _do_insert treeiter = self.insert_with_values(parent, position, columns, row) AttributeError: 'TreeStore' object has no attribute 'insert_with_values' (for each connection I have defined) and it starts, but I can't connect to anything, and the window is blank. Shall I open a new bug on that?
Thanks for checking Kevin, we can just reuse this report. There doesn't seem to be any simple workaround here. I've filed an upstream gnome report: https://bugzilla.gnome.org/show_bug.cgi?id=711797
For me - I'd to downgrade gtk3 As for my today's rawhide virt-manager works with: gtk3-devel-3.10.3-1.fc20.x86_64 gtk3-3.10.3-1.fc20.i686 gtk3-debuginfo-3.11.0-1.fc21.x86_64 gtk3-3.10.3-1.fc20.x86_64 virt-manager-0.10.0-5.git1ffcc0cc.fc21.noarch virt-manager-common-0.10.0-5.git1ffcc0cc.fc21.noarch pygobject3-base-3.11.1-1.fc21.x86_64 pygobject3-3.11.1-1.fc21.x86_64
Any news here? Having virt-manager broken for almost a month is pretty sad. ;(
Kevin, on my rawhide system virt-manager was fixed on 2013-11-10 when Cole originally closed this BZ and has been working fine afaik ever since. I just updated rawhide and confirmed that virt-manager starts ok. Cole, are you able to reproduce what Kevin is seeing?
To be clear, it starts fine, I just can't see/use any of my connections... It gives: Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/manager.py", line 710, in add_conn row = self._append_conn(vmlist.get_model(), conn) File "/usr/share/virt-manager/virtManager/manager.py", line 691, in _append_conn _iter = model.append(None, row) File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1194, in append return self._do_insert(parent, -1, row) File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1187, in _do_insert treeiter = self.insert_with_values(parent, position, columns, row) AttributeError: 'TreeStore' object has no attribute 'insert_with_values' for every connection it should be showing.
(In reply to Kevin Fenzi from comment #12) > To be clear, it starts fine, I just can't see/use any of my connections... > > It gives: > > Traceback (most recent call last): > File "/usr/share/virt-manager/virtManager/manager.py", line 710, in > add_conn > row = self._append_conn(vmlist.get_model(), conn) > File "/usr/share/virt-manager/virtManager/manager.py", line 691, in > _append_conn > _iter = model.append(None, row) > File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1194, > in append > return self._do_insert(parent, -1, row) > File "/usr/lib64/python2.7/site-packages/gi/overrides/Gtk.py", line 1187, > in _do_insert > treeiter = self.insert_with_values(parent, position, columns, row) > AttributeError: 'TreeStore' object has no attribute 'insert_with_values' > > for every connection it should be showing. That's why I'm keeping downgraded my gtk3 libs...
(In reply to Kevin Fenzi from comment #12) > To be clear, it starts fine, I just can't see/use any of my connections... My bad, sorry; I'm seeing that as well. Sorry, I saw that the first bug was resolved and mistakenly assumed everything was working. Giuseppe is going to follow up with the gtk folks.
I've filed a pygobject bug: https://bugzilla.gnome.org/show_bug.cgi?id=719488 The error seems to be in the Python GTK+ bindings. Even the example fails with the same error: https://developer.gnome.org/gnome-devel-demos/3.5/treeview_treestore.py.html.en
As noted in: https://bugzilla.gnome.org/show_bug.cgi?id=719488 This is caused by an out of date GObject-Introspection not picking up the new syntax for "rename-to" used in GTK+. Updating GI should fix the problem. See also: https://bugzilla.gnome.org/show_bug.cgi?id=711341 https://bugzilla.gnome.org/show_bug.cgi?id=711797 https://bugzilla.gnome.org/show_bug.cgi?id=713044
Thanks for clarifying Simon. gobject-introspection 1.39 is in rawhide now, but the latest gtk3 build failed so things are still busted AFAICT: http://koji.fedoraproject.org/koji/taskinfo?taskID=6239961 I'll follow up tomorrow with the desktop guys to try and get things fixed.
I just rebuilt pygobject3, so with that and latest gtk3 things should be working. I'm still hitting some weirdness but it might be ssh -X related so I'm doing some more testing.
I can confirm it works here with the latest pygobject3 and gtk3 from koji. ;)
Thanks Kevin, closing then. I had to use koji latest-build and koji download-build to get latest bits, so for others listening in a simple 'yum update' probably won't fix things yet. Wait a day or so