This is erratic-window-server-crashes in view mode; [Up]
Date: Sun 14-May-1990 09:23:00 From: joerg@krynos.mcs.anl.gov Subject: erratic window server crashes I am dying!!! I have been working on a large complicated app for quite a while now (MindShock, Inc. will be releasing it soon -- start getting excited) and it is crashing the Window Server very inconsistently. That is, if I just run the app, do stuff, and quit, maybe 10-15 times or so over the course of an hour, I can place 50-50 odds that the window server will crash. If I launch the app 20 times or so all at once, the 17th or 18th invocation will crash the window server. I say it's a window server crash because after the crash the cursor repositions itself to (100,100) from the UL corner (hooray for NextAnswers!!). I have been frustrating myself to frenzy, combing through all my code, about 90K (tightly written, too) of it, looking for DPSAddTimedEntry's without the corresponding *Remove*'s and for lockFocus's without unlockFocus's, but to no avail. It may help to note that the crash doesn't always occur while I'm in my app. I could, for instance, just start scrolling some DL text, after having run my app a few times (and quitting after each) and suddenly blammo! I get the loginwindow. At a more abstract level, I ask: in what way is the window server trusting my program so that a mistake on my part would crash the whole system? Since a window server crash terminates everything else I'm running, it makes this bug VERY hard to track down (a molecule in a landfill would be easier). Perhaps one of the very nice people at NeXT could produce a list of what kinds of things make the window server crash. Also, the errors I get on the console after the crash are not very informative; just messages from other apps (loginwindow, workspace manager, and other apps I have running at the time) that their connection to the window server was suddenly cut off. Really enormous big thanks in advance, David S. Joerg MindShock, Inc. Internet : joerg@alliant.mcs.anl.gov
Date: Sun 18-May-1990 23:34:29 From: glenn@heaven.woodside.ca.us (Glenn Reid) Subject: Re: erratic window server crashes In article <4000014@krynos> joerg@krynos.mcs.anl.gov writes: > > I am dying!!! > I have been working on a large complicated app for quite a while now >(MindShock, Inc. will be releasing it soon -- start getting excited) >and it is crashing the Window Server very inconsistently. That is, >if I just run the app, do stuff, and quit, maybe 10-15 times or so >over the course of an hour, I can place 50-50 odds that the window >server will crash. If I launch the app 20 times or so all at once, >the 17th or 18th invocation will crash the window server. I say it's >a window server crash because after the crash the cursor repositions itself >to (100,100) from the UL corner (hooray for NextAnswers!!). > [some stuff deleted] > At a more abstract level, I ask: in what way is the window server trusting >my program so that a mistake on my part would crash the whole system? Since >a window server crash terminates everything else I'm running, it makes >this bug VERY hard to track down (a molecule in a landfill would be easier). > Perhaps one of the very nice people at NeXT could produce a list of >what kinds of things make the window server crash. > Also, the errors I get on the console after the crash are not very >informative; just messages from other apps (loginwindow, workspace manager, >and other apps I have running at the time) that their connection to the >window server was suddenly cut off. The same thing has been happening to me, and I haven't yet gotten around to really looking hard for it. After about 15-20 runs of my program, suddenly the fonts on the various windows start to get destroyed and look like just random dots or solid blocks. Shortly thereafter, the window server crashes. I agree with you that a random application should not be able to crash the window server. I am going to look hard for a "grestore" without a matching "gsave" and things like that. I also am making a call to "setcacheparams" to adjust the font cache slightly that I may take out to see if that fixes it. It does look a lot like the font cache getting screwed up. If anybody can shed some light on this, I'd appreciate it, too. Glenn Reid glenn@heaven.woodside.ca.us
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.