Bug#336090: xserver-xorg: [via] xvideo with high pre-scaling width has corruption on edges of video playback
On Tue, Jan 24, 2006 at 11:51:05PM -0500, Bryan Donlan wrote:
> On 12/17/05, Luc Verhaegen <email@example.com> wrote:
> > On Sat, Dec 17, 2005 at 03:53:29PM -0500, Bryan Donlan wrote:
> > >
> > > No bug there. That said, I've recently cleaned out a lot of dust off
> > > my heatsinks and vents - overheating could've caused the problem. I'll
> > > try to reproduce again with the driver version that was causing me
> > > problems before.
> > >
> > My KM400A 2d engine locked today, after many days of steady use. My
> > guess is still that this is temperature related, as it can get pretty
> > cold up in this room in winter.
> > About your issue, if it does still exist with the new version, after
> > the cleanout, then you might want to try halving the search space each
> > time. Take a CVS tag (look at the cvs log of any file; cvs -z9 log
> > via_driver.h > /tmp/bleh for instance) somewhere halfway between working
> > and not working. Then, depending of the result of this one, go with a
> > tag halfway before and after, until you have the two tags (or 1 commit)
> > between which the problem started appearing.
> > Thanks for putting in so much effort btw, most users would've given up
> > ages ago.
> > Luc Verhaegen.
> Hm, an update with the original problem of corrupted edges. I just
> tried playing video of such a resolution again today, and I'm seeing
> the problem with totem-gstreamer, mplayer, and xine.
> ii mplayer-k6 1.0-pre7cvs200 The Ultimate Movie Player For Linux
> ii totem-gstreame 1.2.1-3 A simple media player for the Gnome desktop
> ii xine-ui 0.99.3-1.3 the xine video player, user interface
> ii xserver-xorg 6.9.0.dfsg.1-4 the X.Org X server
> Also, if the output size is below about 800 pixels in width, or the
> original size is, there is no corruption.
> Since all three players show this bug I again suspect it's a bug in
> the driver. :|
Please send me a gzipped log, so that i can tell wether this is this a
bug with the stock driver, or with my code.
Stock: old Xv code which nobody dared touch.
My code: clean Xv implementation (only took ~3 months).