Martin Atkins

Continuing the trend of "wait, we already did something like this!" :grinning: today I remembered that we also already have [`globalref.Reference`](https://pkg.go.dev/github.com/opentofu/opentofu/internal/lang/globalref#Reference) which is _even more_ like the `addrs.AbsReference` I added...

Hi @sblask, I think the more direct documentation about the situation you're describing is in [Type Constraints](https://www.terraform.io/docs/language/expressions/type-constraints.html), which directly discusses the meaning of Terraform's type constraints and the type conversion...

I think at we've reached the point where we've agreed that Terraform is working as designed but that you disagree with this particular design decision. That suggests that we turn...

When thinking about these things I think it's important to realize that "map" and "object" are not types, but are instead _kinds of types_. `object({ foo = string })` is...

An existing way to be more strict than Terraform would naturally be is to disable Terraform's normal constraint/conversion behavior by setting the constraint to just `any`, and then implementing whatever...

Hi @Kobzol! Thanks for jumping on this. I tried this in my `no_std` crate and it did indeed work as expected.

**EDIT:** I wrote this at the end of my day and it didn't occur to me that I was reacting only to the single sentence "Ideally, this would also let...

Out of curiousity I tried a few different approaches to see if it'd be possible to incorporate Windows results without having to spin up a whole separate Windows runner. My...

I was considering picking this up, but once again I found that my Windows test setup is broken 🙄 so I wouldn't currently be able to test the changes to...

Hi! I'm not an HCL maintainer so this comment is not an answer on behalf of the HCL project. However, I _am_ the maintainer of the `cty` library that HCL...