Martin Atkins

I've added the "blocked" label to this just to represent that (as discussed above) we're intending to write an RFC for this behavior before continuing with it. Hopefully this label...

I've just left a long comment at https://github.com/opentofu/opentofu/issues/300#issuecomment-2387458333 about a design hazard with that feature, and I think a variation of that hazard applies to this feature request too. The...

Hi again! As I alluded to before, this has some overlap with the problem of dynamic provider expansion: it's largely the same thing just with the expansion happening at the...

The main thing that constrains what destroy-time provisioners can refer to is that during destroy work Tofu essentially "works backwards", destroying objects in the reverse order to how they would've...

In https://github.com/opentofu/opentofu/issues/3414 we are intending to implement a new approach to evaluation and execution, which gives us some more options for what we can do about this issue. The details...

Thinking specifically about the problems that are currently being worked around by trying to force the GC to collect leaked file handles, I notice that there was some chatter about...

I'm closing this per our discussion at https://github.com/opentofu/opentofu/issues/3054#issuecomment-3106331106. Thanks!

It's certainly gonna have some stale parts by now, and is unlikely to be written in the same style as the other new content you are referring to, but in...

For normal (non-mocked) resource configurations OpenTofu uses [`convert.Convert`](https://pkg.go.dev/github.com/zclconf/[email protected]/cty/convert#Convert) from upstream `go-cty` to convert the author's given value to the type constraint returned in the provider schema. For the normal evaluation...

For what it's worth, the current requirement that variable default values be static is not so much a hard technical constraint as an intentional design tradeoff: it makes it possible...