On 04/24/2013 08:03 PM, J. Liles wrote:
On Wed, Apr 24, 2013 at 10:54 AM, ed44 <[email protected]
<mailto:[email protected]>> wrote:
But the same as in the sequencer.
If it was possible to draw the rects in the seq.with one fill an
stroke
in stead of hundreds that would speed up the drawing. But that
can't be done.
Using an image surface can be done, and is faster, only i don't
know if that's possible in NTK.
Yeah. The image surface optimization is going to have to wait a while
due to font rendering. Good to know there's a way to squeeze out a bit
more performance though.
As to stroking a bunch of lines at once... I wonder why this is faster
and if it's also faster when the lines are pixel aligned (no AA). But
anyway, it's rare that one can do that in practice, as one rarely has
the need to draw 100s of lines in the exact same color.
Cairo used to get the destiny pixels from the pixmap or window (xgetimage )
and use them to do compositing/antilias for every stroke or fill, then
write them back
That makes it slow.
Using shm-images is much faster and there is some shm code in the
sourcetree but i don't know
if it is used for rendering yet.
Try resizing the window (preferably bigger) before you click
anything. I haven't tested it in a non-tiiling WM yet.
That did it.
Looks good, cannot switch to the next pattern but scrolling is ok,
zoom
in one direction is smooth so it's drawing fast enough now.
Only the playbar breaks up while playing...
But the speed is ok with this method and that was the goal.
Excellent. Yes, there are still plenty of problems with that branch,
but for me it's at least as fast as the pre-Cairo non-sequencer. Good
to know it's on the right track.
I have some questions regarding the seq. code, i'll ask on freenode...
