#29 was done without unit tests, need to fill those in

1. Manually create two projects in env0 called "TestProject"
2. Archive one of those projects
3. Create env0_project data in a TF file with name = "TestProject":

data "env0_project" "test_project" {
  name = "TestProject"
}

4. terraform plan fails with "Found multiple Projects for name"

This data should completely ignore archived projects, as they are technically "destroyed"

Getting an error

Provider terraform-registry.env0.com/env0/env0 v6.6.6 does not have a package
available for your current platform, darwin_amd64.

this can be fixed by changing tests/harness.go in buildFakeTerraformRegistry making the registry_dir end with darwin_amd64 instead of linux_amd64, but we should have this resolved automatically.

Create/Update configuration variables fails with body is not an array, as we aren't really sending the request as an array and API schema has changed recently.

We need to add gitignore to that folder.

resource_project
data_sshkey
resource_sshkey

Are in the files but there is no documentation about them. Probably better to do after #11

On Mac :

ok  	github.com/env0/terraform-provider-env0/client/http	(cached)
=== RUN   TestUnitOrganizationData
--- PASS: TestUnitOrganizationData (16.59s)
=== RUN   TestUnitProjectData
--- PASS: TestUnitProjectData (16.04s)
=== RUN   TestProvider
--- PASS: TestProvider (0.00s)
=== RUN   TestMissingConfigurations
--- PASS: TestMissingConfigurations (0.00s)
=== RUN   TestUnitProjectResource
--- PASS: TestUnitProjectResource (27.87s)
PASS
ok  	github.com/env0/terraform-provider-env0/env0	61.365s

via GIPHY

Looks like we are missing documentation in the READ.me file for SSH keys, and for Project resource (There is a link but it leads to no-where)

When we release there is a github action that builds all the binaries.
But the TF Registry will pickup on the release too soon and we get

Could not find all platform assets for this release yet (0.0.5) — 2 minutes ago

Then we need to manually re-sync it (from the registry UI)

We should add an API call in the release flow to do that when the binaries are ready.

1. Create a Cloud Credential
2. Create the TF config for it
3. Do a terraform import on it
4. Apply- you will be prompted that the credential will be replaced

This is probably because we don't get the full state of the credential on the import.

Create Pull Request Description template

There is a script to generate the docs (./generate-docs.sh) which does everything
But that needs to be run manually and the output committed and pushed.
We should have some kind of pipeline to do that automatically, probably as part of the release (in GitHub actions)

https://github.com/env0/terraform-provider-env0/blob/main/client/model.go is going to get bigger and bigger, should be split up per class, preferably to go along the API methods themselves.

Currently, description is an optional field of data "env0_project". If you create a data using this description, it is completely ignored.

We should either:

1. Remove it
2. Make it computed (and compute it on read)

something like terraform-provider-env0
if possible should include the provider version, tf version, maybe the architecture or something like that

1. Apply this tf

resource "env0_project" "my_project" {
  name        = "Provider Project"
  description = "Project Created By Providerz"
}

resource "env0_template" "test_template" {
  name        = "Created by provider"
  description = "This is the best template."
  repository  = "https://github.com/env0/templates"
  path        = "aws/hello-world"
  revision = "master"
  project_ids = [ env0_project.my_project.id ]
}

2. The template+project are created and assigned
3. Change to project_ids = []
4. Apply. The apply makes it look like the project_ids are really cleared.
5. Go into the env0 UI - the template is still assigned to the project

The retry regexes just do nothing

This will tie into the current bug #13 with creating projects, but we would like the capability to tie org templates to newly created projects.

I would propose some new type of resource called something like "env0_project_template_attachment" that leverages the https://api.env0.com/blueprints/id/projects API endpoint to tie templates to projects

The TF code could look like:

data "env0_template" "aws-s3" {
  name = "aws-s3"
}

resource "env0_project" "project" {
  name = "My Project"
}

resource "env0_project_template_attachment" "attach" {
  template_id = data.env0_template.s3.id
  project_id = env0_project.project.id
}

Only if there are no environments in the project it should be archived.

If there are - you should get an error.
There should be a flag for "force_destroy" (like in aws_s3_bucket) , when it's true then we should archive the project even if it has environments in it.

This is only implemented with ssh_key_names right now, but the template also needs the ID of the ssh key

1. Create a template resource with path = "shoes"
2. Apply
3. Remove the path and apply again
4. The plan says - path = "shoes" -> null
5. Nothing really changes

Expected behaviour would be to update path with an empty string when it's not defined

Currently ConfigurationVariableCreate and ProjectCreate receives flat parameters while all other resources get a payload interfaces. We should align them to have such interfaces.

These following files are not covered by UTs:

1. data_aws_credentials
2. resource_aws_credentials
3. resource_template_project_assignment

1. Create new project, with description in TF, like -

resource "env0_project" "my_project" {
  name        = "Provider Project"
  description = "Project Created By Provider"
}

2. apply
3. The description isn't applied.

If you update, the description will be there, but not on create.

Our current "Documentation" page is empty
We need to https://www.terraform.io/docs/registry/providers/docs.html

We need some run this as pre-commit git hook on this repo

1. go fmt (husky & prettier style)

Considering those too (maybe with a different hook? commit can be slow)
2. generate mocks
3. generate docs

#77 is an example but it's only one resource

terraform import "env0_configuration_variable.test1" b94f954c-0b06-4fb4-8cb5-ed5161c99480
env0_configuration_variable.test1: Importing from ID "b94f954c-0b06-4fb4-8cb5-ed5161c99480"...

Error: Not implemented

:(

1. Create a template in TF
2. Apply
3. Apply second time, without any change

Get -

Note: Objects have changed outside of Terraform

Terraform detected the following changes made outside of Terraform since the last "terraform apply":

  # env0_template.test_template has been changed
  ~ resource "env0_template" "test_template" {
      + github_installation_id = 0
        id                     = "0495debd-96de-4c82-a6fd-3d607d373d65"
        name                   = "Created by providerz"
      + project_ids            = []
      + retries_on_deploy      = 0
      + retries_on_destroy     = 0
        # (6 unchanged attributes hidden)
    }

Unless you have made equivalent changes to your configuration, or ignored the relevant attributes using ignore_changes, the following plan may include actions to undo or respond to these changes.

I guess we need to fix the types better so that TF will understand those default values are not changes.

example TF for recreate

resource "env0_template" "test_template" {
  name        = "Created by providerz"
  description = "This is the best template."
  repository  = "https://github.com/env0/templates"
  path        = "aws/hello-world"
  revision = "master"
}

examples/003_configuration_variable/main.tf relies on the organization having

• A configuration variable called AWS_DEFAULT_REGION in the organization scope.
• A configuration variable called AWS_DEFAULT_REGION in the project scope, in the Default Organization Project project.

This should at least be documented but preferably fixed so the integration tests can run on a "clean" organization and not depend on prior manual setup, besides the API Key.

Terraform Version:

Terraform v1.0.0
on linux_amd64
+ provider registry.terraform.io/env0/env0 v0.0.9

Ubuntu version:

20.04.2

Hi, I'm start trying Env0 Terraform Provider and I'm getting an error when I try to run a Terraform Apply.

Here are the steps that I did:

1. Create an Env0 API Key.

2. Successfully tested the Credentials using curl:

curl \
  --silent \
  --header 'Accept: application/json' \
  --user ${ENV0_API_KEY?}:${ENV0_API_SECRET?} \
  "https://api.env0.com/blueprints?organizationId=${ENV0_ORGANIZATION?}" | jq

3. Set Environment Variables for test:

ENV0_ORGANIZATION=de6d1bb0-f746-4974-ab74-1ff73d0e2445
ENV0_API_SECRET=api_secret_value_here
ENV0_API_KEY=api_key_value_here

4. Create a Simple Env0 Project worked:

# main.tf
terraform {
  required_providers {
    env0 = {
      source = "env0/env0"
      version = "0.0.9"
    }
  }
}

provider "env0" {
}

resource "env0_project" "example" {
  name        = "example"
  description = "Example project"
}

5. When I'm trying to create a new env0_template and an env0_template_project_assignment using this:

terraform {
  required_providers {
    env0 = {
      source = "env0/env0"
      version = "0.0.9"
    }
  }
}

provider "env0" {
}

resource "env0_project" "example" {
  name        = "example"
  description = "Example project"
}

resource "env0_template" "example" {
  name        = "mytemplate"
  description = "Example template"
  repository  = "https://github.com/env0/templates"
  path        = "aws/hello-world"
}

resource "env0_template_project_assignment" "assignment" {
  template_id = env0_template.example.id
  project_id  = env0_project.example.id
}

Plan:

Terraform used the selected providers to generate the following execution plan. Resource actions are indicated with the following symbols:
  + create

Terraform will perform the following actions:

  # env0_project.example will be created
  + resource "env0_project" "example" {
      + description = "Example project"
      + id          = (known after apply)
      + name        = "example"
    }

  # env0_template.example will be created
  + resource "env0_template" "example" {
      + description       = "Example template"
      + id                = (known after apply)
      + name              = "mytemplate"
      + path              = "aws/hello-world"
      + repository        = "https://github.com/env0/templates"
      + terraform_version = "0.15.1"
      + type              = "terraform"
    }

  # env0_template_project_assignment.assignment will be created
  + resource "env0_template_project_assignment" "assignment" {
      + id          = (known after apply)
      + project_id  = (known after apply)
      + template_id = (known after apply)
    }

Plan: 3 to add, 0 to change, 0 to destroy.

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

Saved the plan to: .terraform.plan

To perform exactly these actions, run the following command to apply:
    terraform apply ".terraform.plan"

Error:

env0_project.example: Creating...
env0_template.example: Creating...
env0_project.example: Creation complete after 6s [id=a56c4001-e98d-4c66-9487-8724c4465073]
╷
│ Error: could not create template: 400 Bad Request: ["/body/retry/onDestroy must have required property times","/body/retry/onDestroy must be null","/body/retry/onDestroy must match a schema in \"anyOf\"","/body/retry/onDestroy must be null","/body/retry/onDestroy must match a schema in \"anyOf\"","/body/retry/onDeploy must have required property times","/body/retry/onDeploy must be null","/body/retry/onDeploy must match a schema in \"anyOf\"","/body/retry/onDeploy must be null","/body/retry/onDeploy must match a schema in \"anyOf\"","/body/retry must be null","/body/retry must match a schema in \"anyOf\""]
│ 
│   with env0_template.example,
│   on main.tf line 18, in resource "env0_template" "example":
│   18: resource "env0_template" "example" {
│ 
╵

Am I missing something?

Thanks.

We currently have some integration tests which are great, but we also want unit tests that can be used more frequently during development see #2 for some reference

It seems like we are not sending the Organization id to the API, so we get an UnAuthorized error from the API.
We need to add it to the project resource.

https://github.com/env0/terraform-provider-env0/blob/main/env0tfprovider/resource_project.go#L35

We can throw it in Github actions should be trivial

Under the current instructions I just get
The "source" attribute must be in the format "[hostname/][namespace/]name"

so

terraform {
  required_providers {
    env0 = {
      source = "terraform-registry.env0.com/"
    }
  }
}

doesn't really work.

We need to provide instructions on how to install and use the provider locally.
After that we can think of publishing it to a registry

1. Add optional github_installation_id attribute to template resource + data schemas.
2. Send it on Create/Update operations.
3. Attach to object on Read. (both resource/data)

Merged #28 to resolve #17 - but no ut since we haven't merged in the testbed.
We should add a proper unit test once the testbed's in place

Currently we allow fetching the project data type by name, but projects aren't unique by name on env0.
Solution should include throwing an error if /projects API returns more than a single project with the same name.

We currently use terraform's testing framework for our unit tests, but this framework is genuinely intended for E2E testing (aka Acceptance tests).
This should basically replaces our somewhat problematic harness tests we have today.

1. Create a TF file with an env0_project data with a non-existent ID

data "env0_project" "non_existent" {
  id = "non-existent-id"
}

2. terraform apply
3. The failure is unclear, mentions failure to query template: "Could not query template: 404 Not found: not found"

Merged #25 to resolve #13 - but no ut since we haven't merged in the testbed.
We should add a proper unit test once the testbed's in place

1. Create an "env0_project" resource with a description, and terraform apply it

resource "env0_project" "example" {
	name = "example"
	description = "Remove this later"
}

2. Remove that description line and re-run terraform apply
3. The description is not really removed from the env0 project

1. Create resource "env0_template" without ssh_keys and apply it
2. Go to env0, attach an SSH key to the template, and save
3. run terraform plan, no changes are detected

Merged #37 to resolve #21 - but no ut since we haven't merged in the testbed.
We should add a proper unit test once the testbed's in place

1. Add optional description attribute to project resource + data schemas.
2. Send it on Create/Update operations.
3. Attach to object on Read. (both resource/data)

1. Apply a template with retries_on_deploy
2. Remove retries_on_deploy
3. Apply - the plan says retries_on_deploy = 4 -> null
4. The retries don't actually change