Skip to main content

How to use Trunc and Split functions in Pipelines as Code

Introduction

Using Pipelines as Code (PaC) is an important feature that helps create Spinnaker pipelines quickly and efficiently. By using Dinghyfiles strategically, Spinnaker users can use the same PaC code for different applications, saving a lot of development time and making Continuous Deployment workflows run smoother.

As pipelines get more complex, Dinghyfiles become more advanced, so advanced functions can help create consistency across different pipelines and their stages. This article shows users how to use Trunc and Split functions to set up pipelines in a detailed and effective way.

Prerequisites

  • Pipelines as Code plugin/service available in Spinnaker (Installation and configuration instructions available here)* ARM-CLI used for debugging purposes and validation - ARM-CLI

Instructions

In this example, we want to provision three different wait stages based on some string manipulation using trunc and split commands

{
"application": "splitexample",
"pipelines": [
{
"application": "splitexample",
"name": "my-pipeline-name",
"stages": [
{{ $count := 0 }}
{{ $a := trunc 11 "foo$bar$baz "}}
{{ $b := split "$" $a }}
{{ range $b }}
{{
module "dinghy-modules/wait.stage.module"
"waitname" .
}}
{{ $count = add $count 1 }},
{{ end }}
{{
module "dinghy-modules/wait.stage.module"
"waitname" "Final Wait"
}}
]
}]}

Starting from the "foo$bar$baz  " string, users need to truncate the whitespace.  This can be accomplished by using the trunc function as seen above.  After applying the function, the data will be appended to the "foo$bar$baz" string.  This string can be remapped into a list by using the split function with the "$" operator.

  • Trunc accepts a numeric value as a parameter.  The value denotes the number of characters to be trimmed either from the start of the string (using positive integers) or from the end of the string (using negative integers).* Split splits a string into substrings based on a delimiter defined as a positional parameter. The final pipeline JSON definition would look like this after all the logic has been applied: 
{
"application": "splitexample",
"pipelines": [
{
"application": "splitexample",
"name": "my-pipeline-name",
"stages": [




{
"name": "foo",
"type": "wait",
"waitTime": 100
}
,

{
"name": "bar",
"type": "wait",
"waitTime": 100
}
,

{
"name": "ba",
"type": "wait",
"waitTime": 100
}
,

{
"name": "Final Wait",
"type": "wait",
"waitTime": 100
}
]
}]}

  For more information on Sprig functions, please visit: https://masterminds.github.io/sprig/strings.html