[mikewarot]

It says in the article that everything is immutable, which I found intriguing, but then I see this in the example at mint-lang

fun increment { next { counter = counter + 1 } }

Something doesn't add up.

[kureikain]

I don't know Mint but want to give a data point about that.

The `=` is rebinding, what that mean is the original value doesn't change. If you pass that counter variable somewhere, it has its original value. the `count = counter + 1` here mean that a new memory block is allocated, get a value, and the `counter` in this scope is point to it. The old value in the original memory block won't change.

At least, that's the case of Erlang/Elixir.

[mikewarot]

That's horrible... and yet I see exactly why it is needed for multi-threaded, multi-core systems.

Question: When is it safe to throw the old value away?

[dragonwriter]

Knowing zero about Mint, from the claim and snippet, I assume that “assignments” in a next-block set up bindings for a future state in terms of bindings for the current state. If so, it's not mutation where something could have a reference to the the old counter and have it swapped to the new one.

[fggg444]

ha, that made me chuckle