IDP quits without explanation

[Ingmar Dasseville]

IDP should explain why he suddenly stops

#!

>>> Search start
> restarts              : 1
> decisions             : 1  (0 % random)
> conflicts             : 0
> propagations          : 3
> conflict literals     : 0  (-nan % deleted)
> 1 model found
>>> Search done
Starting approximation at 40ms
Applying propagation to structure
Killed

[Broes De Cat]

This is not a bug? IDP went in memory overflow and was externally killed by the OS.

[Bart Bogaerts]

The reason for the memory overflow is that you are using infinite domains (int)

[Ingmar Dasseville]

Joachim said IDP should always give a reason, and this was a bug.

[Bart Bogaerts]

As the program was killed externally, this is not a bug

(no program can give a reason for stopping if you (or the os) kill(s) it brutally (instead of gently asking the program to stop) )

[Ingmar Dasseville]

One could use a wrapper process for the real process, which detects the kill by the OS and displays an appropriate error message.

On a related note: why does the program still say "Killed" instead of just being shut down without any message?

[Bart Bogaerts]

I think it is not IDP printing this message, but rather the terminal who prints this ("killed" gets printed, even if you redirect all IDP's output streams somewhere else

[Broes De Cat]

Killed is printed by the operation system. You could also not wrap IDP, as this results in various other issues related to signal passing to IDP. I do not immediately see a way to allow this, but if anyone knows one, feel free ;)