>
>
> > BTW=2C I heard from Ian that Only Ben still in fltk2.0's development=2C h=
> ow can I contact with Ben? Thanks.
>
> Ben does still read all the emails on this list=2C but it's somewhat diffic=
> ult to devote hours to investigating things while I'm doing (paid) work -- =
> apparently people complain at me when I do that. =3B-)
>
> I'll take a look at the 4*w*h thing when I get the chance=3B it could be fl=
> tk2 doing something stupid.
> I do agree with Albrecht and Ian and the rest of this mailing list though=
> =3B you'd be far better off porting to 1.3. I've still got some work to do =
> with 3.0 before there's the magical api in place that will save the re-writ=
> e=2C and my dev. on fltk has slowed to a crawl thanks to things getting in =
> the way.
>
> Regards=2C
> Ben
> =
Hi Ben,
I have made one modification on fltk2.0 source code, and now it will not
consume w*h*4 Bytes when new a empty window.
My modification is as follows:
in file src/x11/run.cxx, function Window::flush()
// Copy the backbuffer to the window:
// On Irix, at least, it is much slower unless you cut the rectangle
// down to the clipped area. Seems to be a pretty bad implementation:
Rectangle r(w(),h());
intersect_with_clip(r);
XCopyArea(xdisplay, i->backbuffer, frontbuffer, gc,
r.x(), r.y(), r.w(), r.h(), r.x(), r.y());
if (i->overlay) draw_overlay();
clip_region(0);
+ free_backbuffer();
} else {
// Single buffer drawing
I don't know add this free_backbuffer(); will has problem or not, could you
please help me to check? Thanks. Meanwhile I will also test in my side.
Best regards,
Leo
_______________________________________________
fltk mailing list
fltk@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk
