Scrolling in EVAS

Carsten Haitzler (The Rasterman) raster at
Thu Sep 18 04:46:44 CEST 2008

On Wed, 17 Sep 2008 14:54:40 +0200 Thomas Köckerbauer <tkoeck at>

> I'm currently working on an evas/edje based audioplayer, and have a few 
> questions (which go to raster ;) about scrolling in evas. I've been searching 
> around and found this post about the topic: 
> Is this or something similar already in evas? (I guess no, because scrolling 
> objects had the same "speed" on X using framebuffer and on Xglamo when I 
> tested it).
> Is it still planed implement something like this? (It sounds like something 
> that would make much sense with glamo).

nothing has been done - no time, and the complexity means either you lose out
on features and general rendering (eg being now able to layer things over your
scroll region like shadows or have a static bg etc.) for a specific case just
for the glamo which frankly is probably one of the worst gfx chips i have come
across performance-wise. the imageon on my zaurus c860 beats it... and that was
renowned for how bad it was.

also it's really not easy because of dithering. you then either lose dithered
output (and get 16bpp banding) or you get scrolling - as now the dither
patterns won't match. as such on a normal embedded display that has a directly
mapped framebuffer - the difference between blit and redraw is so little that
you may as well redraw. on a fully accelerated display (eg if opengl was used
and sane) a redraw again is saner than a blit.

so it's just not worth the effort for me. the million things i have still to do
are more important than any specific-case scrolling just for glamo due to its
pretty low level of usefulness and slow write speeds. i have much more hardware
to support than glamo and since its more or less a blip in the scheme of
things, i'm focusing on what is going to win the most per effort spent (nb
i'd do a scale cache long before a scroller/blit specific object) :)

------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at

More information about the community mailing list