Programming Workflow

The following is a suggested workflow for new Pico developers. Related information is available in the following resources:

Programming involves a collection of tools, often referred to as the "tool chain" since they are used more or less in sequence. Here's a graphical representation of the tool chain for programming picos, noting each stage. We'll use this to organize the remainder of this document. 

Develop

To develop KRL for picos, you can use Git (or another versioning tool) and your favorite editor. There are old plug ins for several popular editors.

  1. Create a directory for your project 

    1 2 $ mkdir myproject $ cd myproject/



  2. Initialize the project directory as a Git repository

    1 $ git init



  3. Create a KRL ruleset as a file in the repository using your favorite editor. There may be a KRL plugin for your editor

Build

The build process for rulesets primarily involves editing them and then parsing them to ensure they are correctly formatted. The preferred way to ensure they parse is using a Git pre-commit hook. 

  1. Follow the directions in Setting Up a Git Pre-Commit Hook to Parse KRL to set up the KRL parser and ensure that KRL files are parsed before they are committed. 

  2. Add the file to the repository and commit it. 

    1 2 3 4 5 6 $ git add hello_world.krl $ git commit -a -m "initial commit" All rulessets parse [master (root-commit) 4a2f95c] initial commit 1 file changed, 34 insertions(+) create mode 100644 hello_world.krl

    With the pre-commit hook in place, you're assured that there are no parse errors in your ruleset if it commits.

Deploy

Deploying consists of installing the ruleset in any picos it needs to run in. If the ruleset has already been installed, then it will just need to be flushed. 

Installing

Install the ruleset in your pico by navigating to the Rulesets tab in the developer interface. 

To install the ruleset, you need its URL. If you are committing to Github, then you can use the raw URL from Github.

If you are running locally for testing, then you can install from the file using file:// followed by the full path to the KRL file. Note that you'll end up with something like file:///Users/pjw/... with three slashes after file:

Once the ruleset is installed, it will show up in the Installed Rulesets section of this page. 

Flushing

Because the engine compiles and caches the ruleset, whenever you make changes to the source, you need to flush the ruleset. For development and testing, you can do that manually from the Installed Rulesets section of the develop interface. 

This will flush the ruleset for all picos on this instance of the pico-engine. The hash will change with each new version of the file and is unique for every version of the file. 

If you have not parsed your rulesets when you commit them (as described above) then the developer interface will show you any errors when you install or flush the ruleset. 

Test 

The developer interface includes support for testing your rulesets. See the detailed instructions for more information on how to configure and use that feature. 

The Logging tab in the developer interface is also helpful in debugging and testing rules. 

Release

For local development, there's no need to do anything more than what's specified above for releasing a ruleset. 

If you're releasing a ruleset for broader use beyond your own needs, you'll want to take steps to ensure others see a stable ruleset even though you may continue to work on it. Putting your ruleset on an HTTP server allows it to be registered with multiple engines. 

The use of versions in your ruleset can ensure that picos see the version they need to operate correctly. 

The first step is to decide where you want to host your rulesets. Any HTTP server will do. Here are some tips for hosting code on GitHub and AWS

You can use directories on AWS to create a stable set of code for a given release. GitHub has a feature called "releases" that create a tag for a given version. Both of these work well.