> > Shouldn't it read 0.4 ? Also, the bufferflipping seems to be totally
> > messed up. Hard to describe but it seems like the changes made
> > doesn't always carry over as they did when using VESA.
>
> it's considered unstable, experimental and unfinished for a good
> reason.  We'd be happy if this could improve but we only have limited
> resources and will need your help.

I guessed as much :)

Anyway, after doing some testing, this is what I found ...

Case 1: Using savage framebuffer (with or without acceleration or DirectFB
gfx driver)
- Draw vertical line
- Flip buffer
= The vertical line is visible
- Draw horizontal line
- Flip buffer
= Horizontal line is visible (no verical)
- Draw a circle
- Flip buffer
= Vertical line and circle is visible but not horizontal

Case 2: Same procedure but with vesa framebuffer:
- Draw vertical line
- Flip buffer
= The vertical line is visible
- Draw horizontal line
- Flip buffer
= Horizontal and vertical line is visible
- Draw a circle
- Flip buffer
= Vertical & horizontal line aswell as the circle is visible

Is case #1 correct or case #2? I'd stipulate case #2 :) But since I'm no
longer sure, I need to verify this with someone who knows the innerworkings
of DirectFB and framebuffer a bit better than me.

No other artifacts (like this) appears when using the console or mplayer
with fbdev as output. I'm running DirectFB 0.9.14 (release) and kernel
2.4.19

/Henric



-- 
Info:  To unsubscribe send a mail to [EMAIL PROTECTED] with 
"unsubscribe directfb-dev" as subject.

Reply via email to