I was notified (by an early elaborator of that assignment) that RISCAL often crashes when running the game model to produce the traces to winning games.
Debugging this issue, it seems to me that the problem is not in RISCAL itself but in the underlying graphical toolkit (the SWT library) which (in its current version) for some reason crashes, if too much (or too quickly?) output is produced. While I cannot really do anything about this now, some mitigations:
- Setting in the model "val all = false;" (only a single winning game is printed, not all of them) lets the crash occur much more rarely. This might be sufficient for the purpose of this assignment.
- Additionally, choosing in the RISCAL "Analysis" panel the red button "Start Logging" writes the screen output also into a file; here much more output is logged than shown in the screen until the crash.
With (1) or (1+2) it should be possible to produce the expected results. If not, just report the error and what game moves you could deduce from the partial output.