Thanks for the review! 🙂 and yeah, I’ve seen others compare these concepts to debugging too. It’s a tradeoff though, debuggers are amazing and the whole stop and step over / inspect is great. But they work more like coroutines I guess? Whereas the REPL way is always run in real time but inspect the live thing, don’t pause the world and inspect the memory.

The REPL approach involves one concept: Executing code. There’s not much else to learn and you can extend that protocol infinitely (I think?). So stepping debuggers are sort of a slightly locked down specialisation of the concept + the ability to stop the world.

if I could investigate react state from it, would be nice

This is one reason to override eval btw, starting a sub-repl at a point in order to debug.