[Bug]: Issue with codefresh_pipeline resource where original_yaml_string uses YAML anchors & aliases

[nathanzanon]

What happened?

In our pipeline definitions we're using YAML anchors (and aliases) to avoid repetition.

When using a codefresh_pipeline resource where the file original_yaml_string contains YAML anchors and aliases the apply fails with:

╷
│ Error: error while converting 'steps' YAML to JSON: yaml: unknown anchor 'my_common_envs' referenced
│ 
│   with codefresh_pipeline.indicators,
│   on main.tf line 49, in resource "codefresh_pipeline" "indicators":
│   49: resource "codefresh_pipeline" "indicators" {
│ 
╵

The YAML pipeline definitions I'm using to validate this are the examples in the documentation. So these work fine when configuring through the Codefresh console/GUI, and they also pass validation

❯ codefresh validate plain.yaml
Yaml is valid!

❯ codefresh validate indicators.yaml
Yaml is valid!

Version

Seems to be broken in >= 0.4.0, including current latest 0.10.1
Works in <= 0.3.1

Relevant Terraform Configuration

main.tf

terraform {
  required_version = "1.9.5"

  backend "local" {}

  required_providers {
    codefresh = {
      source  = "codefresh-io/codefresh"
      version = "0.3.1"   # works
      #version = "0.4.0"  # broken
      #version = "0.10.1" # broken
    }
  }
}

provider "codefresh" {
  api_url = "${var.codefresh_api_url}/api"
  token   = var.codefresh_token
}

variable "codefresh_api_url" {
  description = "Codefresh authentication token"
  type        = string
}

variable "codefresh_token" {
  description = "Codefresh authentication token"
  type        = string
}

locals {
  codefresh_project_name = "testing"
}

resource "codefresh_pipeline" "plain" {

  name = "${local.codefresh_project_name}/example-plain"

  original_yaml_string = file("${path.module}/plain.yaml")

  spec {
    variables = {
      TIME = timestamp() # force update of resource on every plan/apply
    }
  }

}

resource "codefresh_pipeline" "indicators" {

  name = "${local.codefresh_project_name}/example-indicators"

  original_yaml_string = file("${path.module}/indicators.yaml")

  spec {
    variables = {
      TIME = timestamp() # force update of resource on every plan/apply
    }
  }

}

---

plain.yaml

version: "1.0"
steps:
  preLoadDatabase:
    title: Loading Data
    image: alpine
    commands:
      - printenv
      - echo "Loading DB"
    environment: &my_common_envs
      - MYSQL_HOST=mysql
      - MYSQL_USER=user
      - MYSQL_PASS=password
      - MYSQL_PORT=3351
  runTests:
    title: Integration tests
    image: alpine
    commands:
      - printenv
      - echo "Running tests"
    environment: *my_common_envs # Same MYSQL_HOST, MYSQL_USER etc.

---

indicators.yaml

version: "1.0"

indicators:
  - environment: &my_common_envs
      - MYSQL_HOST=mysql
      - MYSQL_USER=user
      - MYSQL_PASS=password
      - MYSQL_PORT=3351

steps:
  preLoadDatabase:
    title: Loading Data
    image: alpine
    commands:
      - printenv
      - echo "Loading DB"
    environment: *my_common_envs # Same MYSQL_HOST, MYSQL_USER etc.
  runTests:
    title: Integration tests
    image: alpine
    commands:
      - printenv
      - echo "Running tests"
    environment: *my_common_envs # Same MYSQL_HOST, MYSQL_USER etc.

[ilia-medvedev-codefresh]

@nathanzanon thank you very much for reporting the issue!
The fix has been included in version 0.11.0 of the provider.