Maintainers Guide

Directory structure

• libraries/sdk: the source folder for the main SDK
• libraries/sdk-ring: source folders for the ring adapter specific code
• libraries/sdk-http-kit: source folders for the http-kit adapter specific code
• libraries/sdk-malli-schemas: self explanatory...
• libraries/sdk-brotli: brotli write profiles
• src/bb: tasks used run a repl, tests...
• src/bb-example: bb examples
• src/dev: dev utils, examples
• src/test: centralized tests for all the libraries
• test-resources: self explanatory

bb tasks

Release tasks

• bb bump-version patch/minor/major: to bump a version component across all libs
• bb set-version x.x.x: to set the version component across all libs
• bb jar:all: Build jars artifacts for all of the libs
• bb clean: Clean all build artifacts
• bb install:all: Build jars artifacts for all of the libs and installs them locally
• bb install:clean will delete ~/.m2/repository/dev/data-star/clojure
• bb jar:<lib-name>: Build jars artifacts for one of the libs
• bb publish:all: Publish the artifacts to clojars.org

Development tasks bb run dev

• bb run dev: start a repl with the dev nss, the test nss the malli schemas, ring-jetty and Http-kit on the classpath
• bb run dev:rj9a: same as basic dev task expect for ring-jetty being replaced with rj9a.
• bb run dev:bb: start a bb repl with the core SDK, http-kit adapter and malli-schemas in the classpath.

[!note] You can add additional deps aliases when calling these tasks: bb run dev :debug will add a Flowstorm setup

Test tasks bb run test

• bb run test:all: run all test for the SDK, the Http-kit adapter and the ring adapter using ring-jetty.
• bb run test:rj9a: run all test for the SDK and the ring adapter using rj9a.
• bb run test:bb: run unit tests for the SDK in Babashka.
• bb run test:sdk-common: start the server used to run the SDKs' common tests.
• bb run test:sdk-common: run the common SDK test (go program)

Cljdoc tasks

In order to test how cljdoc will ingest our libraries there are babashka to run a local instance.

To do so you need to have either docker or podman installed, the babashka tasks we provide will work with either one.

Next you need to run:

docker pull cljdoc/cljdoc

Once that is done you can use the following bb tasks:

• bb run cljdoc:server to start a cljdoc local instance on port 8000
• bb run install:all to install locally the jars to ingest
• bb run cljdoc:ingest <SDK-lib-name> to tell cljdoc to ingest a sdk library. The possible values for <SDK-lib-name> are: :sdk :http-kit :ring :malli-schemas :brotli
• bb run cljdoc:clean will delete ./.cljdoc-preview, the directory where the local cljdoc server stores its data.

Release

• The library artifacts are published to Clojars under the dev.data-star.clojure namespace.
• The Clojars account and deploy token are managed by Ben Croker, and added to this repo as GitHub action secrets:
  • Secret name: CLOJARS_USERNAME Value: the clojars account username
  • Secret name: CLOJARS_PASSWORD Value: the clojars deploy token
• The libraries' versions are bumped in lockstep so that there is no confusion over which version of the common lib should be used with an adapter lib.

The Github Actions CI workflow for clojure will always run the tests and produce jar artifacts.

Triggering a deployment to clojars is a manual process. A Datastar core contributor must trigger the Release Clojure SDK workflow with the publish input boolean set to true.

Release process:

1. Use bb set-version or bb bump-version to update the library versions in lockstep
2. Commit those changes and push to GitHub
3. A core contributor must trigger the workflow manually setting publish to true

Test

Running tests

• for the unit and smoke tests see the bb tasks.
• for the generic go SDK tests:
  1. Start the test server with bb run test:sdk-common
  2. Run bb run test:sdk-common-go

webdriver config

You can modify the webdriver config when testing by creating a test-resources/test.config.edn file. It contains a map whose keys are:

• :drivers: etaoin webdriver types to run
• :webdriver-opts: a map of webdriver type to webriver specific options

For instance on my machine:

{:drivers [:firefox :chrome]
 :webdriver-opts {:chrome {:path-driver "/snap/bin/chromium.chromedriver"}}}