Terraform Core Configuration Schema

This repository contains CUE schemata representing the structure of the core Terraform configuration language, manually curated from the Terraform language documentation.

Currently, only the fields permitted by the latest language version (v1.4) are present. PRs that improve or expand the schemata selection are extremely welcome!

Using this repo

These schemata are primarily designed for use inside the Cueniform ecosystem, and will change and adapt along with that project's aim of advocating the maintenance of Terraform configurations in CUE (not HCL or JSON). However, nothing stops them being used separately, in isolation, and this README briefly describes how to do that.

These schemata can be used to validate the structure of Terraform configurations expressed in Terraform's JSON syntax. Direct validation of Terraform's HCL variant is not possible. Converting from HCL to the JSON format is therefore required, but is outside the scope of this repo.

Requirements

1. CUE version 0.6.0-alpha.1, or later [Releases]
   • Required Fields are used, which is a v0.6 feature

Usage

1. Initialise a CUE module directory with cue mod init

2. Clone this repo into cue.mod/pkg/cueniform.com/x/terraform-core:

   git clone \
       https://github.com/cueniform/terraform-core-schema \
       cue.mod/pkg/cueniform.com/x/terraform-core
   

3. Choose a Terraform version

   • 1.4.x: for Terraform versions 1.4.0 and above (1.5.0 at your own risk!)

4. From a working directory inside your CUE module directory, run cue:

   cue vet -c -d '#Configuration' \
       cueniform.com/x/terraform-core/<VERSION>/core \
       /path/to/*.tf.json
   

Missing fields and malformed structs will be pointed out by CUE. If you notice things that it misses or gets wrong please do tell us by opening an Issue!

Do be aware that this repository only contains schemata that validate core Terraform configuration structures, and not the individual requirements of specific providers, resources, data-sources, etc.

Integrating the schemata for those components, on top of these core language schemata, is what the Cueniform project was created to do ... and that's still a work in progress!