reassign 572848 phonon-backend-gstreamer tags 572848 + upstream thanks Hello,
2010/4/15 Sebastian Dröge <sl...@circular-chaos.org>: > On Thu, 2010-04-15 at 11:57 +0300, George Kiagiadakis wrote: >> This has also been reported to nokia as a bug in the phonon backend, with a >> trivial fix, but I'm still not sure if the bug is in phonon or gstreamer. The >> fix looks like a hack. >> >> http://bugreports.qt.nokia.com/browse/QTBUG-8737 > > The Nokia bug is a problem in phonon's gstreamer backend... > > It claims to support all kinds of RGB formats but in fact only supports > xRGB in native endianness. Hmm, thanks for the explanation, I think I understand the issue now and what a proper fix would be. > Could you add that to the Nokia bugreport? The code in question is in > qwidgetvideosink.cpp: > >> >> static GstStaticPadTemplate template_factory_yuv = >> GST_STATIC_PAD_TEMPLATE("sink", >> GST_PAD_SINK, >> GST_PAD_ALWAYS, >> GST_STATIC_CAPS("video/x-raw-yuv, " >> "framerate = (fraction) [ 0, MAX >> ], " >> "width = (int) [ 1, MAX ], " >> "height = (int) [ 1, MAX ]," >> "bpp = (int) 32")); >> >> static GstStaticPadTemplate template_factory_rgb = >> GST_STATIC_PAD_TEMPLATE("sink", >> GST_PAD_SINK, >> GST_PAD_ALWAYS, >> GST_STATIC_CAPS("video/x-raw-rgb, " >> "framerate = (fraction) [ 0, MAX >> ], " >> "width = (int) [ 1, MAX ], " >> "height = (int) [ 1, MAX ]," >> "bpp = (int) 32")); > > Not only does it claim to support all 32bpp RGB formats but also *all* YUV > formats (and a "bpp" field is not used for YUV at all). That's not true, it claims either RGB or YUV depending on the template parameter. Note that QWidgetVideoSink is a template ;) The "bpp" for YUV may not be needed but it doesn't hurt either, plus this is completely unrelated to this bug since the code in question uses the RGB version of the template. > If they need some help with fixing this they can contact me via mail but > I don't want to subscribe to yet another bugtracker... So, from what I understand, the RGB caps need to specify the endianess so that it matches the endianess used by QImage. I'll try to fix it and inform them. If I need help I'll contact you. Thanks :) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
