dleyna-control very broken on python-gi 3.8.1

[jku]

TreeIter userdata handling has been changed in ways that completely break our TreeModel in python-gi 3.8: when value 0 is inserted into TreeIter.user_data it becomes None. I consider this a bug and maintainer seemed to agree and was willing to back that out.

Depending on distro uptake and speed of updates we may have to add a workaround to GenericModel that avoids value 0.

[markdryan]

Hi jku,

I just tried dleyna-control on Ubuntu 13.04 and it crashes on startup. This is what I get

50 rows fetched (50/54 rows cached)
/usr/lib/python2.7/dist-packages/gi/types.py:113: Warning: g_object_set_property: assertion `G_IS_VALUE (value)' failed
  return info.invoke(*args, **kwargs)
/usr/lib/python2.7/dist-packages/gi/types.py:113: Warning: g_value_unset: assertion `G_IS_VALUE (value)' failed
  return info.invoke(*args, **kwargs)

(dleyna-control.py:14134): Gtk-WARNING **: /build/buildd/gtk+3.0-3.6.4/./gtk/gtktreemodel.c:639: Invalid path None passed to gtk_tree_path_new_from_string
Traceback (most recent call last):
  File "/usr/local/lib/python2.7/dist-packages/msd/msd_generic_model.py", line 286, in do_get_path
    return Gtk.TreePath((None,))
  File "/usr/lib/python2.7/dist-packages/gi/overrides/Gtk.py", line 1121, in __new__
    raise TypeError("could not parse subscript '%s' as a tree path" % path)
TypeError: could not parse subscript 'None' as a tree path
Segmentation fault

Is this the problem you are referring to here?

[jku]

They are on 3.8 already? Damn, I was hoping we could wait for the bug fix releases since not too many distros were using 3.8... But it seems I was wrong about Ubuntu.

In any case: Yes, that is the bug I'm refering to. It has been fixed (https://bugzilla.gnome.org/show_bug.cgi?id=698366) and backported to 3.8. The fix has not been released though and there's no telling when Ubuntu will update anyway. So I guess we need to do the workaround.

[lferrandis]

Ubuntu 13.04 is officially on Gnome 3.6
Gnome provide a ppa to upgrade to 3.8

I think the problem is more when the maintainers of this ppa will
include the patch.

On 03/05/2013 09:49, Jussi Kukkonen wrote:

They are on 3.8 already? Damn, I was hoping we could wait for the bug
fix releases since not too many distros were using 3.8... But it seems I
was wrong about Ubuntu.

In any case: Yes, that is the bug I'm refering to. It has been fixed
(https://bugzilla.gnome.org/show_bug.cgi?id=698366) and backported to
3.8. The fix has not been released though and there's no telling when
Ubuntu will update anyway. So I guess we need to do the workaround.

—
Reply to this email directly or view it on GitHub
#36 (comment).

Ludovic Ferrandis
Open Source Technology Center
Intel Corporation

[jku]

@lferrandis that's what I was counting on as well, but http://packages.ubuntu.com/raring/python-gi tells me python-gi is 3.8 so apparently it does not follow the normal GNOME versioning, at least not in ubuntu.

For additional information: the fix will be in the 3.8.2 release (and is in 3.9.1 already).

[markdryan]

Yep it looks like Ubuntu 13.04 is uses python-gi 3.8

dpkg -l | grep python-gi
ii  python-gi                                 3.8.0-2                                    amd64        Python 2.x bindings for gobject-introspection libraries
ii  python-gi-cairo                           3.8.0-2                                    amd64        Python Cairo bindings for the GObject library

Since dleyna-control is complete broken on Ubuntu 13.04 I'll test and apply this patch today, unless any one has any objections.

[markdryan]

jku, I can confirm that your patch fixes the problem on Ubuntu 13.04. Could you send a PR?

[markdryan]

Fixed by PR #37