Generate timing diagram

How it looks like? (maybe)

entry |> State1 |> exit

state State2 {
  State21 >< activity

  initial -> State21
  State21 -> State22 @ Event
}

initial -> State1
State1 -> State2 @ Event[guard] |> action
State2 -> State1 @ Reset |> action

From 👆 to 👇 Screenshot_20190706_065814

{signal: [
  {name: 'execution', wave: '03524.02350', data: ['entry', 'exit', 'action', 'activity', 'action', 'entry', 'exit']},
  {name: 'trigger',   wave: '0..=0.==0..', data: ['Event [<i>guard</i>]', 'Event', 'Reset']},

  {name: 'State1',    wave: '01.x...1..0'},
  ['State2',
   {name: 'State21',  wave: '0x.1..x...0'},
   {name: 'State22',  wave: '0x....1x..0'},
  ],
],
  config: { hscale: 3 }
}

Paste 👆 to https://wavedrom.com/editor.html

I think I should move action, entry, and exit inside state timeline 🤔. Also separate activity to different lifeline, especially when in the compound state.

However, for easier implementation, it should lay flat. Imagine having an action that executed many times, how long the data will be 😅

Something needs to watch out!

waveform doesn't automatically fit the diagram with the texts
probably there will be a need for sorting (I hope not because I use wave: 0.x..0) if I want to support Arrows Splines
activity inside Compound State could be tricky to visualize. Maybe activity should be separated as a different lifeline

Other approachs

transpile into LLHD then visualize using gtkwave or alike
generate VCD file

My background is electrical engineering so it might get influence from that. There are several way to analyze FSM in ASIC design (see [here](FSM-viz Verilog)). Let's move the discussion in #27.

Yeah I had a look at the link. It is true that chips are on a clock, so it makes sense to visualize it that way. Did you ever heard about Lucid Synchrone or Esterel? Those are synchronous languages compiling to fsm with model checking and all the jazz. They embed a notion of clock and use clocked streams for everything.

For the FSM I deal with, I think the interesting information for tracing is the sequence itself , i.e. instead of visualizing [(x_i, t_i] i.e. a sequence of x occurring at time t it is more interesting to visualize just the [x_i] directly on the graph visualization. But I haven't given too much thoughts to that.

DrSensor / scdlang