• Prepare latest Bos Wars upstream, ask for a sponsor: done.

    Bos Wars is a futuristic real time strategy game.

    It used to have copyright and license troubles, but upstreams kindly addressed that.


  • Finish libpuzzle packaging: done.

    Libpuzzle is a library to compare images, which also comes with a simple puzzle-diff binary scoring image similarity.

    Next todo: Wait for a review of the PHP bindings by Emmanuel Bouthenot (who already kindly rewrote upstream's manpages in a proper manner), then ask for a sponsor.

    Another todo: Check upstream status of CLSimilarImages, see whether it might be worth packaging it, and point the upstream authors to each others.

  • Check LuxRender's upstream status: done.

    LuxRender is a rendering system for physically correct, unbiased image synthesis.

    The packaging has been ready for weeks, waiting for a released tarball (rather than working on a CVS snapshot). RC5 is on its way.


  • Suggest an option for fork()-based rendering to Blender upstreams: done.

    Blender is a free open source 3D content creation suite, including modelling and animating tools, as well as an internal renderer.

    For some reasons, one might want to use a fork()-based solution rather than one (only) based on threads (limited to 8 by default through a #define), when it comes to using a cluster.

    Next todo: try and hack it myself.

  • Check appletouch's Debian bug #465278 with latest kernel: done.

    appletouch is the mouse driver on iBook G4 machines, and tends to break badly when resuming from suspend2ram: it's unusuable until the next reboot. Not any better with 2.6.25-rc5-powerpc.

    Next todo: Report it upstream, following Maximilian's request.

  • Check the status of a nasty xfwm4 bug: done.

    When the “switch back to previous workspace” option is enabled, and one keeps the key for switching to another workspace pressed too long, xfwm4 starts flickering between both workspaces, endlessly. Restarting X helps. That was previously reported to happen only on “low” machines, but it's easy to trigger this bug by lowering the delay before character repetition in the “keyboard” settings.

Of course, that's just an excerpt of the actual todo-list…