Issue Summary

Currently, if you use the Pipes script (atlassian/trigger-pipeline) to trigger a Pipelines build to another repository while passing a JSON file as variables, Pipelines can only parse variables with the correct casing. It will fail if the incorrect casing has been used in the JSON format.

For example:

  {
    "Key": "SOME_KEY",
    "Value": "SOME_VARIABLE"
  }

Steps to Reproduce

1. Use the Pipes script to trigger a build - atlassian/trigger-pipeline
2. Pass JSON variables (with the example format above) to the PIPELINE_VARIABLES variable (or using the below step configuration)
3. Run a build

         - step:
             name: Some Step
             script:
               - export VARIABLES=$(envsubst < ./some_data.json)
               - pipe: atlassian/trigger-pipeline:5.1.0
                 variables:
                   BITBUCKET_USERNAME: $BITBUCKET_USERNAME
                   BITBUCKET_APP_PASSWORD: $BITBUCKET_APP_PASSWORD
                   REPOSITORY: 'some_repo'
                   REF_TYPE: 'branch'
                   REF_NAME: 'some_branch'                
                   CUSTOM_PIPELINE_NAME: 'some_custom_pipelines_name'
                   PIPELINE_VARIABLES: $VARIABLES
                   WAIT: 'true'
                   DEBUG: 'true' 
   

Expected Results

The JSON file will be parsed even if it's case insensitive

Actual Results

The JSON file is not being parsed by Pipelines due to case insensitivity

Workaround

Update the JSON format by using lower casing.