Add blueprint parameters as env variables in before/after scripts

[steverobbins]

This is exposes parameters to the before/after scripts.

Previously, if you wanted to pass envs, vars, params to a before/after script you need to do something like

blueprints:
  - stackname: 'something-lambda'
    template: 'lambda.template'
    Capabilities: 'CAPABILITY_IAM'
    parameters:
      S3Bucket: 'mybucket'
      S3KeyPrefix: 'uploads/functions_{tstamp}/'
    before:
      - 'cd function && ../scripts/zip-and-upload-to-s3.sh s3://{var:S3Bucket}/{var:S3KeyPrefix} {var:SomeOtherVar-{env:Environment}}'

Then in your zip-and-upload-to-s3.sh script you need to use $1, $2, etc

perl -pi -e "s/{{SomeOtherVar}}/$2/g" "/path/to/function.js"

aws s3 cp --recursive "/path/to/function.js" "$1"

This is getting messy when my lambda functions start requiring a lot of blueprint vars, resources, etc.

---

With this PR all blueprint params are added to the environment

blueprints:
  - stackname: 'something-lambda'
    template: 'lambda.template'
    Capabilities: 'CAPABILITY_IAM'
    parameters:
      S3Bucket: 'mybucket'
      S3KeyPrefix: 'uploads/functions_{tstamp}/'
      SomeOtherVar: '{var:SomeOtherVar-{env:Environment}}'
    before:
      - 'cd function && ../scripts/zip-and-upload-to-s3.sh'

And I can access them in the script like

perl -pi -e "s/{{SomeOtherVar}}/$SomeOtherVar/g" "/path/to/function.js"

aws s3 cp --recursive "/path/to/function.js" "s3://$S3Bucket/$S3KeyPrefix"

---

This is really handy, for example, when you have a lot of DynamoDB table names that you need to pass to your script. You can use a function like

replaceParametersInFile() {
    while IFS='=' read -r NAME VALUE ; do
        sed -i -- "s~{{$NAME}}~$VALUE~g" "$1"
    done < <(env)
}

to replace all {{PARAM}} style tags with their values as they're added.