jacobi.pa@XEROX.COM (08/12/89)
Is this a known bug? I believe this bug is in the X server, but how do I prove it... I'm using a PutImage protocol request with format=bitmap aimed at a window on the color screen (8 bit per pixel) of a Sun Sparcstation. (X11R3 (MIT); SunOs 4.0.3) Executing the request partly works: Bits are painted first, but then the rightmost part (looks like left_pad pixels) of the image is restored again to something which looks like the previous contents but blurred a little. It also looks a little funny that the window is painted in two steps. I know I have not given a reasonable test case. Excuse me, I'm not a C programmer; my Cedar test case might not help you anyway. Christian Jacobi Xerox PARC
mvh@cfa250.harvard.edu (Mike Van Hilst) (08/16/89)
Since no one else responded to jacobi.pa@XEROX.COM, here are my two bits: > I'm using a PutImage protocol request with format=bitmap ... on > a color screen (8 bit per pixel) of a Sun Sparcstation. ... > ... Bits are painted first, but then the rightmost part (looks > like left_pad pixels) of the image is restored again to something > which looks like the previous contents but blurred a little. I can only guess at this problem: On Sparcstations (and MIPS machines) there are no real char operations. If you are using chars (or shorts) to construct your bitmaps, be aware of the other bytes, especially for shift and complement (<<, >>, ~) operations. Use "|= 0xFF" in place of "~". > It also looks a little funny that the window is painted in two steps. The foreground and background are both colors. It is not as simple as setting the foreground bits and not the background ones. The foreground must be painted one color, and the background another (as per the GC). Mike VanHilst Smithsonian Astorphysical Observatory