Draft: Use text interpolation in Ci::Yaml::Loader

What does this MR do and why?

Use text interpolation in Ci::Yaml::Loader

This commit introduces text interpolation to CI YAML loading under a ci_text_interpolation feature flag. This form of interpolation should enable array and hash type inputs, fix existing bugs with input types, and be faster.

Relates to: #433002 (closed)

How to set up and validate locally

1. In any project, create a file included.yml in the root directory. Give it this content:

spec:
  inputs:
    after_script:
    allow_failure:
      type: boolean
    cleanup:
    deploy_job_name:
    parallel:
      type: number
    test_script_name:
---
.scripts:
  rspec:
    - echo "Runs rspec"
  cleanup:
    - echo "$[[ inputs.cleanup ]]"

test_job:
  after_script: $[[ inputs.after_script ]]
  parallel: $[[ inputs.parallel ]]
  script: !reference [.scripts, "$[[ inputs.test_script_name ]]"]

$[[ inputs.deploy_job_name ]]:
  allow_failure: $[[ inputs.allow_failure ]]
  script: echo "deploy"

2. In the project's .gitlab-ci.yml, include included.yml like this:

   include:
     - local: included.yml
       inputs:
         after_script: "!reference [.scripts, cleanup]"
         allow_failure: true
         cleanup: Cleanup things
         deploy_job_name: inputs_deploy
         parallel: 8
         test_script_name: rspec

3. See that the configuration is valid. Save it and see that a pipeline is created as expected

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.

Related to #433002 (closed)