Author: Sune Fischer
Date: 04:13:29 04/14/04
Go up one level in this thread
On April 14, 2004 at 07:04:17, Fabien Letouzey wrote: >On April 14, 2004 at 06:46:56, Sune Fischer wrote: > >>I'm testing it with Arena which crashes on me in multipv mode. > >!!! > >>My guess that there is a problem with printing to stdout after the pipe has been >>closed. > >Yes, that's "broken pipe". >In theory we should be ready for it anytime, but nobody checks the return value >of printf(). > >Note that there should be two pipes for input and output, but that does not >change anything. What do I do when it doesn't return at all? >>I record all communication to a log file and what I see is that the search stops >>between iterations, frenzee hangs at 0 cpu load, Arena pops up an access >>violation box and stops responding so I have to kill it with the task manager. > >That's "OK", but the OS should close Arena's side of the pipes then. >That's how you could notice (in an OS-independent way) that the GUI is gone. > >>Occasionally I get a few hundred runtime error boxes which nearly crashes the >>entire system, luckily these close themselves after a few seconds. > >It might just "outputting" to the broken pipe. >If the output pipe is closed, each printf() generates its own error. >It should be nowhere near crashing the system of course. > >It might not be that because the xboard protocol is not that verbose. > >>I suspect frenzee is trying to print information to Arena which no longer >>recieves input. Who is crashing who here I don't know, but it would be nice if >>frenzee exitted after I kill Arena with the task manager. > >You should be able to notice that by seeing the input pipe beeing closed. >It might not be true on Windows though ... > >Does Frenzee quit peacefully in a console when you press ctrl-Z? No I get ^Z in the console window. It exits fine when I type ctrl-C however, that's the normal kill command so maybe that's what you mean? >>I will try some more experiments to see if I can figure out what is going on. > >I have got problems with PolyGlot and many engines, I think our problems are >related. I've never had this problem before that I can recall, but I guess each crash is unique in its own way. -S. >Fabien.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.