Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!usc!sdd.hp.com!hp-pcd!hpcvlx!harry From: harry@hpcvlx.cv.hp.com (Harry Phinney) Newsgroups: comp.sys.hp Subject: Re: Xserver causing problems on hp9000s300 Message-ID: <101950177@hpcvlx.cv.hp.com> Date: 23 Jan 91 19:15:59 GMT References: Organization: Hewlett-Packard Co., Corvallis, OR, USA Lines: 49 Nick Williams writes: > * Windows will not be refreshed after they have been > overlapped by another window and then exposed. > * When a window is exposed, the rectangle that should be redrawn is > filled with a random colored garbage. This is most noticeable with > Motif text widgets. This is purely a guess. Try turning off TWM's use of backingStore and SaveUnders and see if that improves or changes the behavior. > * TWM title bars occasionally "lose" their decoration (resize button, > etc). Most often happens with applications that change their title > during run. Don't have any good idea of what's happening here. The one guy here who uses twm hasn't seen this problem. > * TWM cursors get left behind in motif windows --- when you opaque > move a window (a TWM function which moves the window itself as opposed > to an outline of the window), several images of the current cursor are > left on the window, and never cleaned up. This happens mostly on motif > windows. Has been exhibited once or twice on other, random windows. This certainly sounds like a reasonably straightforward server bug. What display system is in your 350? The most helpful would be the number from the graphics card plugged into the SPU box (e.g. 98550A). > * The keys' auto-repeat is turned off on startup every now and again. The only time I've seen this is when the HIL loop has been disturbed while the server is running. Plugging and/or unplugging the keyboard, mouse, or other input devices while the server is running will definitely turn off auto-repeat. It can be turned back on with xset. > * The "middle" mouse button (on 2 button mouse) has its threshold time > (the difference between the times each button goes down to be > considered the middle button) decreased to some time which is > impossible to achieve. I've never seen this happen. If you could somehow better pin-down the conditions under which it happens we'd be glad to chase this. > I can also reproducibly get the server to core dump. Could you tell us how to reproduce it? Harry Phinney harry@hp-pcd.cv.hp.com