Carl Mäsak

Here is some prior art drop from the Babel toolchain: [regenerator](https://github.com/facebook/regenerator). I don't know why, but I find this project fascinating. Particularly reading [the emit.js file](https://github.com/facebook/regenerator/blob/e4d19a8126c5ee48211c296154e3120045ee341f/packages/regenerator-transform/src/emit.js) is inspirational. I really...

I was thinking of what would be the absolute ideal in terms of code re-use, and realized that we should be able to just plug _code generation_ (#145) into the...

Oh! Remember the "clones get different states" semantics from #207. It applies here also. (Edit: Re-reading this in 2021, it seems that OP fully realized this. G'ah, I feel like...

The [`once` type from the package `sync`](https://pkg.go.dev/sync#Once) is an interesting take on this, essentially reducing the challenge to an object that does its own synchronization. As long as it has...

> Speaking of encapsulation, note that `state` with an assignment has `once` semantics. (Think of how often the assignment to 0 in `state s = 0;` above ought to run.)...

For that matter, why are we constructing staged code in this example? We're in two nested `fun`s and have access to the two `Int`s `z1` and `z2`. Wouldn't it be...

> > you're never really dealing with ASTs; only with code quotes and regular values. Refreshing! > > Isn't that pretty similar to rakudo's current macros impl :) I'm out...

Ouch. 😉 But even that jest is too charitable, I think. A non-introspectable AST is effective neither as an AST nor as a runtime value.

I struggle to put into words what makes staging different from compilation. (And therefore also, I can't fully map out the shape of my enthusiasm for it, in this issue...

Yes, I think we're making, or reaching for, the same point. > It's not like `eval`, where we juggle strings, drop some of them on the floor, lose one under...