charlie@wam.umd.edu (Charles William Fletcher) (02/16/91)
While running three or four other applications (most of them hidden) I started up the demo FractalTree. Suddenly (boy that 040 is fast isn't it) as the third window for FT was to come up everything disappeared and I was logged out. I logged in again and found every thing was fine. I have not been able to repeat the feat. Has anyone else seen this. Any guesses what happened? Is this a bug or a "feature" like the 1.0 \\\<Return> logout? -Charlie
melling@cs.psu.edu (Michael D Mellinger) (02/17/91)
I have seen one other message in this group describing this phenomenon, and it has also happen to me on several occasions. It must be an OS bug(perhaps a setup problem), and a nasty one at that. Has anyone else been automagically logged out? -Mike
uskmg@mathcs.emory.edu (Ken Guyton {EUCC}) (02/17/91)
In article <432G31*o@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >I have seen one other message in this group describing this >phenomenon, and it has also happen to me on several occasions. It >must be an OS bug(perhaps a setup problem), and a nasty one at that. >Has anyone else been automagically logged out? > >-Mike There was one occasion when, everytime I typed a key while in the browser (``File Viewer'' now?), my session went *instantly* to the logon screen. (Normally the filename-matching panel would pop up). This behavior was repeatable everytime I logged back on. This was either on a NeXTcube 030 running 2.0 or on a NeXTstation with 2.0. (I wasn't keeping careful notes at the time). The problem went away (I think after a reboot) and has never occurred again. -- Ken Guyton | uskmg@unix.cc.emory.edu PREFERRED Emory University | ...gatech!emoryu1!uskmg UUCP Academic Computing | uskmg@emoryu1 OR uskmg@emuvm1 NON-DOMAIN BITNET Atlanta, GA 30322 | Phone: (404) 727-7685
gessel@ilium.cs.swarthmore.edu (Daniel Mark Gessel) (02/20/91)
This is the window server dying and restarting, I'm pretty sure. I've had similar problems when sending rude postscript through YAP. I'm trying to remember exactly what I was doing. I think it was in a for loop, and I ended up scaling by 1.0/loopvariable, and the loopvariable went through 0. I think that killed the window server (and I can see why, it'd take a while to draw anything with that scaling). Dan -- Daniel Mark Gessel Internet: gessel@cs.swarthmore.edu I do not speak (nor type) representing Swarthmore College.
datran2 (02/20/91)
In article <GESSEL.91Feb19102343@ilium.cs.swarthmore.edu> gessel@ilium.cs.swarthmore.edu (Daniel Mark Gessel) writes: >This is the window server dying and restarting, I'm pretty sure. I've >had similar problems when sending rude postscript through YAP. I'm >trying to remember exactly what I was doing. I think it was in a for >loop, and I ended up scaling by 1.0/loopvariable, and the loopvariable >went through 0. I think that killed the window server (and I can see >why, it'd take a while to draw anything with that scaling). I had the same reaction when I tried to display postscript that had an object scaled to 0. This took a while to track down, because the window server didn't die until I quit Preview. Also this is old data from 1.0. I don't know if the same is true of 2.0. It seems to me that an application such as Preview or Yap should not be able to kill the window server. Perhaps this is a practical impossibility, but is an ideal to strive for. Steve. -- #====#====#====#====#====#====#====#====#====#====#====#====#====#====#====# # Steve Boker # "Badgers, we don't have no stinking badgers" # # smb@data.com # -from Treasure of the Sierra Madre Zoo # #====#====#====#====#====#====#====#====#====#====#====#====#====#====#====#
glenn@heaven.woodside.ca.us (Glenn Reid) (02/20/91)
In article <432G31*o@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >I have seen one other message in this group describing this >phenomenon, and it has also happen to me on several occasions. It >must be an OS bug(perhaps a setup problem), and a nasty one at that. >Has anyone else been automagically logged out? The phenomenon of getting "automagically logged out" usually means that the window server died. The window server is basically the PostScript interpreter. Evidently there are still some conditions (PostScript errors) that can cause the window server to crash. It's not really an OS problem (the OS doesn't crash), and the only reason you're "logged out" is that all of your processes quit when the window server quits. You may notice messages in the log file that say something like "Exiting due to window server death". Anyway, I've had the window server crash during software development work when I'm pushing things, and sometimes when I've got hundreds of fonts loaded, but I can't seem to find a repeatable sequence that will crash it. Sigh. -- Glenn Reid RightBrain Software glenn@heaven.woodside.ca.us NeXT/PostScript developers ..{adobe,next}!heaven!glenn 415-851-1785 (fax 851-1470)
mike@phoenix.Princeton.EDU (Michael S. Mahoney) (02/23/91)
In article <432G31*o@cs.psu.edu> melling@cs.psu.edu (Michael D Mellinger) writes: > >I have seen one other message in this group describing this >phenomenon, and it has also happen to me on several occasions. It >must be an OS bug(perhaps a setup problem), and a nasty one at that. >Has anyone else been automagically logged out? > >-Mike It has happened to me a couple of times. /usr/adm/messages reported a failure in the DPS server, which then provoked a sequence of windowserver failures and shutdowns including logoff. It seemed to occur when switching applications very quickly. Mike