Description of problem: [tasaka1@localhost ~]$ deluge [INFO ] 14:45:38 main:84 Deluge ui 0.9.03 [DEBUG ] 14:45:38 main:85 options: {'config': None, 'logfile': None, 'ui': None} [DEBUG ] 14:45:38 main:86 args: [] [DEBUG ] 14:45:38 configmanager:44 ConfigManager started.. [INFO ] 14:45:38 main:89 Starting ui.. [DEBUG ] 14:45:38 ui:44 UI init.. [DEBUG ] 14:45:38 configmanager:88 Getting config 'ui.conf' [DEBUG ] 14:45:38 config:47 Config created with filename: ui.conf [DEBUG ] 14:45:38 config:48 Config defaults: {'default_ui': 'gtk'} [INFO ] 14:45:38 ui:60 Starting GtkUI.. [DEBUG ] 14:45:38 client:54 CoreProxy init.. [DEBUG ] 14:45:39 configmanager:69 get_config_dir: /home/tasaka1/.config/deluge Traceback (most recent call last): File "/usr/bin/deluge", line 8, in <module> load_entry_point('deluge==0.9.03', 'console_scripts', 'deluge')() File "/usr/lib/python2.5/site-packages/deluge/main.py", line 90, in start_ui UI(options, args) File "/usr/lib/python2.5/site-packages/deluge/ui/ui.py", line 62, in __init__ ui = GtkUI(args) File "/usr/lib/python2.5/site-packages/deluge/ui/gtkui/gtkui.py", line 106, in __init__ gtk.gdk.threads_init() SystemError: error return without exception set Version-Release number of selected component (if applicable): deluge-0.9.03-2.fc10.i386 How reproducible: 100% Additional info: 0.9.04 won't start either.
Created attachment 313165 [details] Workaround patch for deluge 0.9.04 gtk ui While I am not sure this is a correct method, for me the attached patch can be a workaroud. ! Note: I created this patch against 0.9.04 although I think this patch can be applied for 0.9.03.
Umm... This issue happended also on reviewing libhocr (bug 455555) such as: $ hocr-gtk Traceback (most recent call last): File "/usr/bin/hocr-gtk", line 1310, in <module> gtk.gdk.threads_init() SystemError: error return without exception set Once switching the component to pygtk2.
*** Bug 457657 has been marked as a duplicate of this bug. ***
Affects Miro as well. Is PyGTK2 2.12 incompatible with gtk2-2.13.x? The bug should probably be reported upstream.
It seems Mandriva fixed this issue 3 hours ago: https://qa.mandriva.com/show_bug.cgi?id=42467 http://cia.vc/stats/project/Mandriva/.message/456d2e6 http://svn.mandriva.com/svn/packages/cooker/python-gobject/current/SOURCES/pygobject-fix-api-breakage.patch I tried to apply the patch by Mandriva and for me deluge and hocr-gtk seem to be working. Once changing the component to pygobject2.
Fixed in pygobject2-2.15.2-2.fc10.
*** Bug 458398 has been marked as a duplicate of this bug. ***
*** Bug 457987 has been marked as a duplicate of this bug. ***
As noted in the Mandriva bug report that Mamoru linked to, 2.15.2 is still not a complete fix: Miro silently freezes in Rawhide the same way it does in Mandrake Cooker. Also, conduit does not work against pygobject2 2.15, failing in Python/ceval.c with the following: PyEval_EvalCodeEx: Assertion `tstate != ((void *)0)' failed Downgrading pygobject2 to 2.14 from F9-updates fixes the conduit problem (but not a solution due to API changes: Miro no longer works).
Actually fixed in pygobject2-2.15.2-3.fc10, which due to the security intrusion affecting Fedora has not been pushed to the Rawhide repository yet. The conduit error is a different pygobject2 bug; filing separately: https://bugzilla.redhat.com/show_bug.cgi?id=459861