I was just thinking about it and I bet that, based on our current pace of progress on the new compiler, we could get it to a stage where people can use it on basic-cli for Advent of Code 2025
that wouldn't be 0.1.0 because not everything might be all the way there yet - e.g. release builds, testing, repl, etc. - but I bet static dispatch could be there, so people could try that out for AoC
it's almost exactly 6 months away as of today, and it feels achievable to me if we set our sights on it
I presume getting something basic canonicalized, type checked, and interpreted is the first step on the way there
Seems like we have people working on all of those already
Which is awesome!
What work needs to be figured out that hasn't been architected? Feels like the interaction between the interpreter and basic-cli is the biggest gap here, AKA the shim
Is there a design for the shim? If not, what features do we need to constrain the design with?
that's a good question! I suspect most of the design work there exists in @Brendan Hansknecht's head :smile:
I think the shim will be pretty easy overall, it is really just a single function per entry point that adds type information and calls the interpreter. Maybe some minor weird linking (like loading the compiler itself as a shared library such that we can call into the interpreter), but nothing too special
Okay, then we will just have to ask you to write everything down when the time comes
I'm keen to help work on that soon. I think once we get some basic interpreter infrastructure and typechecking I could have a crack at it under Brendan's guidance.
Last updated: Jul 05 2025 at 12:14 UTC