[Interp] Produce non-empty waveforms for combinational DUTs

[ngernest]

Note: merge #195 before reviewing this PR (I merged #195 into this PR since I found the Cargo aliases defined there helpful for running the interpreter executable, so he changed .toml files come from #195). The only changed file that is relevant to this PR is interpreter.rs.

When debugging why the monitor was failing the adder_d0 (combinational adder) round-trip test (#187), I noticed that the waveform produced by the interpreter was empty (Surfer had trouble opening the .fst file that was produced by the interpreter). I think the issue was that previously, the simulator's sim_step() method was only called when the next_threads queue is non-empty:

// scheduler.rs           
           ...
           // setup the threads for the next cycle
            if !self.next_threads.is_empty() {
                // advance simulation for next step
                info!("Stepping...");
                self.evaluator.sim_step();
                 ...
             }

However, for combinational DUTs like adder_d0, the next_threads queue is empty since all protocols defined with the same .prot file finish within the same cycle. As a result, the simulator never gets stepped, resulting in an empty waveform.

The fix is to add an extra flag to the Scheduler struct that tracks whether a step() statement was encountered, and to call sim_step() on the simulator whenever a step() is encountered (regardless of whether the next_threads queue is empty or not).

This allows the interpreter to produce a non-empty waveform for adder_d0!

All other interpreter tests still pass after this change.