jockey-kde crashed with AttributeError in ui_main_loop()

[Stefan Lorenz]

I am using Ubuntu 8.04 Hardy .
starting jockey-kde 0.2-0ubuntu6 nothing happened. crash report was generated by appport.

ProblemType: Crash
Architecture: i386
Date: Tue Feb 19 13:21:07 2008
DistroRelease: Ubuntu 8.04
ExecutablePath: /usr/bin/jockey-kde
InterpreterPath: /usr/bin/python2.5
NonfreeKernelModules: nvidia
Package: jockey-kde 0.2-0ubuntu6
PackageArchitecture: all
ProcCmdline: /usr/bin/python /usr/bin/jockey-kde --check
ProcCwd: /home/stefan
ProcEnviron:
 LANGUAGE=de_DE:de:en_GB:en
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=de_DE.UTF-8
 SHELL=/bin/bash
PythonArgs: ['/usr/bin/jockey-kde', '--check']
SourcePackage: jockey
Title: jockey-kde crashed with AttributeError in ui_main_loop()
Uname: Linux stefan-laptop 2.6.24-8-generic #1 SMP Thu Feb 14 20:40:45 UTC 2008 i686 GNU/Linux
UserGroups: adm admin audio cdrom dialout dip floppy lpadmin plugdev video

[Starbetrayer (starbetrayer)]

i confirm this one

[James Westby (james-w)]

Hi,

The attached patch should fix this issue.

Thanks,

James

[Martin Pitt (pitti)]

Thanks, James, but calling ui_init() in check is the wrong way to fix it IMHO with the current code structure, since ui_init() will bring up the main window. But --check should just keep the tray icon alive and process events there. In gtk this can be achieved with gtk.main(). Does Qt have a similar 'static' function?

If not, can we change ui_main_loop() to check for an initialized mw first, and if that does not exist, run some exec() function on the tray icon instead?

If that is not suitable for the Qt implementation, I could also split ui_init() into ui_init() and ui_show_main() upstream, if that's better?

[James Westby (james-w)]

On Mon, 2008-03-03 at 09:19 +0000, Martin Pitt wrote:
> Thanks, James, but calling ui_init() in check is the wrong way to fix it
> IMHO with the current code structure, since ui_init() will bring up the
> main window. But --check should just keep the tray icon alive and
> process events there. In gtk this can be achieved with gtk.main(). Does
> Qt have a similar 'static' function?
>
> If not, can we change ui_main_loop() to check for an initialized mw
> first, and if that does not exist, run some exec() function on the tray
> icon instead?
>
> If that is not suitable for the Qt implementation, I could also split
> ui_init() into ui_init() and ui_show_main() upstream, if that's better?
>

Hi Martin,

I have a confession to make. I know nothing about qt, I just looked at
the code and tried to see what would make the attribute be set, and
went for the simple fix.

I guess you'll need to contact someone with qt experience to get the
right solutions.

Sorry to waste your time.

Thanks,

James

[Starbetrayer (starbetrayer)]

it still crash after the update of today

[James Westby (james-w)]

Hi,

Sorry, I didn't realise this was still assigned to
me.

I don't have the necessary skills to fix it, so I am
unassigning it.

Thanks,

James

[Martin Pitt (pitti)]

This broke the entire --check functionality.

Fixed in trunk.

[Launchpad Janitor (janitor)]

This bug was fixed in the package jockey - 0.3.2-0ubuntu2

---------------
jockey (0.3.2-0ubuntu2) hardy; urgency=low

  * Final 0.3.2 release (previous Ubuntu upload had a botched version number)
    with bug fixes only (no new features). This time also with orig.tar.gz
    again.
    - fglrx, nvidia handlers: Create screen section if it does not exist.
      (LP: #200832)
    - Abort gracefully with a proper error message if cache directory does not
      exist. (LP: #198433)
    - ui.py, check(): Intercept ValueError from package query. This can happen
      if the daily apt cron job runs while jockey is running as well and
      temporarily causes inconsistent package indexes. (LP: #200089)
    - kde/jockey-kde, comfirm_action(): Fix string formatting. (LP: #197777)
    - jockey/oslib.py: Fix calling of kdesu.
    - kde/jockey-kde: Make --check notifications actually work (LP: #193985)
    - Fix the KDE interface test suite.
  * debian/rules: Fix apport hook file name.
  * debian/apport_hook.py: Add missing import.
  * data/handlers/b43.py: Consider driver enabled if firmware files are
    installed; testing whether b43-fwcutter is installed (the default
    behaviour of ModulePackageHandler) is incorrect here. (LP: #198341)

 -- Martin Pitt <email address hidden> Tue, 18 Mar 2008 17:28:32 +0100

[jiri (jirikraus)]

jockey-kde Version: 0.3.3-0ubuntu7 crashes for me:

Traceback (most recent call last):
  File "/usr/bin/jockey-kde", line 367, in <module>
    sys.exit(u.run())
  File "/usr/lib/python2.5/site-packages/jockey/ui.py", line 266, in run
    self.ui_init()
  File "/usr/bin/jockey-kde", line 126, in ui_init
    self.update_tree_model()
  File "/usr/bin/jockey-kde", line 317, in update_tree_model
    is_enabled = handler.enabled()
  File "/usr/share/jockey/handlers/nvidia.py", line 50, in enabled
    devices = self.xorg_conf.getSections('device')
AttributeError: 'tuple' object has no attribute 'getSections'

[James Westby (james-w)]

Hi jiri,

That is a different problem to the one that this bug report
is about, could you please open another bug report
about it?

  https://bugs.launchpad.net/ubuntu/+source/jockey/+filebug

Thanks,

James

[Martin Pitt (pitti)]

jiri, don't bother reporting this. This is bug 214914, and it has more than enough duplicates already.