[MPlayer-dev-eng] Mplayer & directfb

[Curtis Magyar]

On Tue, 2004-05-18 at 16:34 +0200, Jiri Svoboda wrote:
> Simple test to determinne what is broken is to run exaple programs coming
> with directfb (try them run also on overlay). If they fail it's propably
> directfb fault otherwise the problem is on mplayer side.

I have done this already, I can't find any program that fails like this
other than MPlayer.  They all appear fine.  Also, I just realized that
the problem also occurs with the fbdev vo, but strangely (to me anyway)
not the fbdev2 vo.  I always thought this was a problem only with the
directfb vo.  Perhaps I never tried fbdev under 2.6 before.


> Also test if -(no)dr switch has any effect. If Yes there is a problem in
> vo directfb. In this case send me a full log please and I'll try to figure
> out whats wrong...
> From the first look at images it looks like pixelformat/stride
> problem.

I did not see any difference from -dr or -nodr.  I'll attach a logfile
for directfb under 2.6.  I can't find any difference between the logs
from fbdev under 2.4 and 2.6 (even though one works, and one doesn't),
but when I diff the directfb logs though, I see one interesting
difference:

In 2.4, I see:
DirectFB: Format query: BGR 24-bit
DirectFB: Test format - layer 0 scale/pos 0
DirectFB: Test format - added layer 0 scale/pos 0
VO: [directfb] 720x480 => 720x540 BGR 24-bit  [fs] [vm]

And in 2.6, I see:
DirectFB: Format query: BGR 32-bit
DirectFB: Test format - layer 0 scale/pos 0
DirectFB: Test format - added layer 0 scale/pos 0
VO: [directfb] 720x480 => 720x540 BGR 32-bit  [fs] [vm]

Maybe this is a difference in the capability of the fb drivers?

Thanks very much for the reply.  

--
Curtis