A while ago I asked: > Does anyone know what > (!) DirectFB/Unichrome: Timeout waiting for idle command regulator! > means? When I see it, the machine normally locks up requiring a hard reset.
My debugging suggests that it occurs if I blit a significant distance outside of the surface. Blitting a few pixels outside is OK, but beyond a certain point it will start to fail. Question: in general, is it supposed to be OK to do a blit that goes outside the surface, i.e. < 0 or > height or width? i.e. is something in the hardware or software supposed to clip to the surface for me? More specific question: In the Unichrome source code (gfxdrivers/unichrome/Readme.txt) it says: "Blitting outside the screen is buggy (e.g with negative surface coordinates). The result is clipped, but does not look right." This doesn't say "crashes", so is what I'm seeing this bug, or something else? Finally: you may be wondering why I am blitting outside of the surface. Here's the situation. If you can suggest an alternative solution that doesn't require blitting outside the surface, please let me know as that would certainly solve this problem. - I have an image, say 100x100 pixels, which I am displaying zoomed 10 times using a strechblit. I can move the image around on the screen using the mouse. I can move it partially off the screen. - If it is at the left of the screen, I stretchbit image[(0,0) (99,99)] to screen[(0,0) (990,990)]. - If I move the mouse ten pixels to the left, then I can stretchbit image[(1,0) (99,99)] to screen[(0,0) (980,990)]. No problem. - However, if I only moved it 5 pixels off the left of the screen, then I would want to stretchbit image[(0.5,0) (99,99)] to screen[(0,0) (985,990)]. But I can't supply fractional parameters. So instead I have to stretchbit image[(0,0) (99,99)] to screen[(-5,0) (985,990)]. Any ideas? Cheers, Phil. _______________________________________________ directfb-dev mailing list [email protected] http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
