feat(lib): Expand Output value types

[jsteinich]

Fixes #1088

Terraform output values can be assigned any expression according to https://www.terraform.io/docs/language/values/outputs.html#declaring-an-output-value.

Since we have started having a type to represent expressions, I thought it would make sense to start using it. I added a few things in so that we wouldn't reduce what could be passed to an output. We don't yet have a complete list; however, we do have any which can help cover missing cases.

I also added in ITerraformDependable to just output values to cover outputting an entire resource or module. I assume data sources can also be used that way, but if not, we can easily narrow it. I'm not sure if these can be used in general expressions, but if so, can move to there.

Draft because no new tests yet.

[ansgarm]

Hi @jsteinich! Should I help write some tests so we can merge this? Or do you already have some work in progress tests lying around?

[jsteinich]

Hi @jsteinich! Should I help write some tests so we can merge this? Or do you already have some work in progress tests lying around?

You can if you want; otherwise, I'll get to them at some point. I've just been prioritizing some other things since I created this.

[ansgarm]

@jsteinich fyi: I rebased your branch onto main and force pushed, hope that is okay with you 😇

[ansgarm]

@jsteinich I added a testcase for passing a complete resource to an output (which will follow the isITerraformDependable path in the code).
Is there another test case you'd like to see or had in mind?

[jsteinich]

@jsteinich I added a testcase for passing a complete resource to an output (which will follow the isITerraformDependable path in the code). Is there another test case you'd like to see or had in mind?

Thanks. I added a couple more tests to pick up some of the expression expansion as well.

[skorfmann]

Is this related to to #1279 ? If not, can we piggyback on this or should we create a new PR?

[jsteinich]

Is this related to to #1279 ? If not, can we piggyback on this or should we create a new PR?

Yeah, we should fix that here (it isn't currently fixed).

[jsteinich]

I think we'll need to do something similar to TfExpression.markAsInner to correctly synthesis the value for an output for #1279. I can take care of that in the next couple of days.

[jsteinich]

Alright I updated the synthesis in a way that should take care of #1279 (working in a very similar test case) as well as a couple other cases that users could run into.

[skorfmann]

Cool, thank you!

[github-actions]

I'm going to lock this pull request because it has been closed for 30 days. This helps our maintainers find and focus on the active issues. If you've found a problem that seems related to this change, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.