Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!rutgers!netnews.upenn.edu!grip.cis.upenn.edu!bradley From: bradley@grip.cis.upenn.edu (John Bradley) Newsgroups: comp.windows.x Subject: X11R4 on an IBM RT (megapel) Message-ID: <20053@netnews.upenn.edu> Date: 7 Feb 90 01:08:49 GMT Sender: news@netnews.upenn.edu Reply-To: bradley@grip.cis.upenn.edu (John Bradley) Organization: Grasp Lab Lines: 49 Configuration: IBM RT 125, with megapel display, running AOS 4.3. X11R4 code with fix-1 applied. TWM. Problems: The IBM megapel server code seems to have a major bug in it, as shipped. The code seems unable to correctly draw a stippled rectangle (such as a scrollbar from an R3 client), a bitmap (icons), or a pixmap (xgif). All three things exhibit the same behavior, which is why I suspect it's only one bug. bitmaps are come out drawn as a square split along the top-left to bottom-right line. (ie, a pair of right triangles) The lower-left triangle is drawn in the background color, and the upper-right triangle is drawn in black. This happens on all the icons that have bitmaps (xload, xclock, xrn), but not the ones that just have text. (xcalendar, xterm, xmh). Scrollbars on R3 clients are drawn as a series of small pairs-of-triangles (as described above). If I do an 'xsetroot', the root is tiled with a pattern of black-and-white triangles, same as you'd expect... xsetroot (default root window pattern) appears solid black. xgif didn't display ANYTHING in it's window. Interesting side note: R4 scrollbars (xterm, at least) DO draw correctly. A change in the way they're drawn, perhaps? Anyhow, I compiled the server with hc2.1d, and it compiled fine except for a compiler error in 'miarc.c', which I hand-compiled with pcc. I've compiled with and without optimization, and the server behaves the same. I attempted to compile the whole server with pcc, but got hung up on some 'out of tree space; simplify expression' errors, so I decided pcc wasn't the way to do it. Other problems: I see no (text) cursor in my xrn or xmh edit windows. These are R3 clients running on another machine. Gray buttons (R3 xmh) display wrong. (more of this diagonal banding problem) Plenty of other stuff also displays this problem. HELP!!! John Bradley - University o' Pennsylvania - bradley@cis.upenn.edu