Throw a 'restart' signal; the main entrypoint loop will catch this and actually perform the restart.

Note that we *could* do the VM reset (via restartGame()) here, but there's an advantage to doing it in the main loop: we won't be in the stack context of whatever command we're performing. If we did the restart here, it's possible that some useless objects would survive the VM reset just because they're referenced from within a caller's stack frame. Those objects would immediately go out of scope when we get back to the main loop, but they might survive long enough to create apparent inconsistencies. In particular, if we did a firstObj/nextObj loop, we could discover those objects and re-establish more lasting references to them, which we certainly don't want to do. By deferring the VM reset until we get back to the main loop, we'll ensure that objects won't survive the reset just because they're on the stack momentarily here.

Credits :   SystemAction