Testing

Only available in figwheel.main >= 0.1.9-SNAPSHOT

Also keep in mind that the testing process helpers below are only intended to work with [:optimizations config option] set to its’ default :none

Testing ClojureScript

If you haven’t written tests in ClojureScript yet, see the testing guide over at clojurescript.org.

Auto testing

The easiest way to start working with tests in Figwheel is to write some tests in a namespace that is on your classpath, and set the :auto-testing config option to true. Figwheel will automatically discover your tests and display them with cljs-test-display at the HTTP endpoint /figwheel-extra-main/auto-testing. No test runner is required and the tests will be re-run every time you save a watched source code file.

The :auto-testing config option can also take a map with the following keys:

• :cljs-test-display can have a boolean value indicating if you want to display the tests
• :namespaces takes a vector of namespaces that you want to test

Custom testing with :extra-main-files

If the testing setup provided by :auto-testing doesn’t work for you, the extra mains feature has you covered. We’ll quickly walk through an example of using :extra-main-files to provide a custom testing setup with relatively little effort.

First you would add a test runner namespace of some sort. I’m going to assume we are using cljs.test and cljs-test-display.

(ns example.test-runner
  (:require
    [cljs-test-display.core]
    [cljs.test :refer-macros [run-tests]]
    ;; require all the namespaces that have tests in them
    [example.core-test]
    [example.other-test]))

(run-tests (cljs-test-display.core/init! "app-testing")
           'example.core-test
           'example.other-test)

Now we’ll configure this as an extra main entry point in our dev.cljs.edn file:

^{:extra-main-files {:testing {:main example.test-runner}}}
{:main example.core}

Now when you start your build with the standard clojure -m figwheel.main -b dev -r command you will now be able to see your custom testing at the /figwheel-extra-main/testing endpoint.

Please note that the "app-testing" above refers to the DOM id of the DIV that is available by default on the page served by the extra main endpoint. This id needs to correspond to the name of your extra main. In this case our extra main is named :testing. If your extra main was named :tests you would need to use "app-tests" in the cljs-test-display/init! call.

You can also utilize the figwheel.main.testing/run-tests macro that will automatically find all the testing namespaces in your source files.

(ns example.test-runner
  (:require
    [cljs-test-display.core]
    [figwheel.main.testing :refer-macros [run-tests]]
    ;; require all the namespaces that have tests in them
    [example.core-test]
    [example.other-test]))

(run-tests (cljs-test-display.core/init! "app-testing"))

And as usual you can always create your own HTML host page for your extra main. In this case you would have to require the target/public/cljs-out/dev-main-testing.js bootstrap script.

Running tests from the command line

Running tests from the command line is an important feature that allows you to confirm that your tests are passing without having to boot up and kill a REPL. Being able to run tests from the command line in a process that will non-zero exit on test failure is an essential feature that allows us to integrate testing into modern CI and cloud testing services.

Up until now most CLJS developers have had to rely on using a Node process to drive a headless JavaScript environment to run their ClojureScript tests. This is a shame considering that we are already using a mediating process when we run Clojure to start our ClojureScript REPLs and run our scripts.

Figwheel can now run an asynchronous main script and report failures with a non-zero exit status.

Here is an example test runner that uses the new asynchronous main script functionality.

(ns example.test-runner
  (:require
    [cljs.test :refer-macros [run-tests] :refer [report]]
    [figwheel.main.async-result :as async-result]
    ;; require all the namespaces that have tests in them
    [example.core-test]
    [example.other-test]))

;; tests can be asynchronous, we must hook test end
(defmethod report [:cljs.test/default :end-run-tests] [test-data]
  (if (cljs.test/successful? test-data)
    (async-result/send "Tests passed!!")
    (async-result/throw-ex (ex-info "Tests Failed" test-data))))

(defn -main [& args]
  (run-tests 'example.core-test 'example.other-test)
  ;; return a message to the figwheel process that tells it to wait
  [:figwheel.main.async-result/wait 5000])

This is just an example but given that the test namespaces are on the classpath, you can execute a test runner main script like the one above on the command line as follows:

clj -m figwheel.main -m example.test-runner

The above example is a bit verbose so Figwheel has provided a helper that will find all your testing namespaces and execute them taking advantage of the new asynchronous -main support.

Let’s repeat the above example using the figwheel.main.testing/run-tests-async helper.

(ns example.test-runner
  (:require
    [figwheel.main.testing :refer-macros [run-tests-async]]
    ;; require all the namespaces that have tests in them
    [example.core-test]
    [example.other-test]))

(defn -main [& args]
  ;; this needs to be the last statement in the main function so that it can
  ;; return the value `[:figwheel.main.async-result/wait 10000]`
  (run-tests-async 10000))

If by any chance your tests are synchronous please feel free to use the figwheel.main.testing/run-tests helper which will throw an exception if tests fail.

Running tests in a headless environment

CI and testing environments almost always require being able to run tests in a headless JavaScript environment.

Figwheel now has the :launch-js configuration option which should allow you to launch an arbitrary JavaScript environment.

:launch-js allows three different ways to launch a JavaScript environment.

• a string representing a shell script that will be supplied a URL or a path to your compiled ClojureScript. Example: "headless-chrome"
• a vector representing a shell command. Example: ["headless-chrome" :open-url]
• a namespaced symbol that represents a Clojure function that will be supplied a map. Example: user/run-test-env

I’m going to show how you can use the :launch-js config option to launch a headless Chrome environment to run your tests using all 3 different methods.

Shell Script method

On my Mac in my ~/bin directory I have an executable shell script named headless-chrome that has the following content:

#!/bin/sh
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --headless --disable-gpu --repl "$@"

The above script takes a single URL argument.

Now we just have to add :launch-js "headless-chrome" to our Figwheel options. We can do this on the command line while we run our tests like so:

clojure -m figwheel.main -fwo '{:launch-js "headless-chrome"}'  -m example.test-runner

We can also add the option in a tests.cljs.edn build file if we are using it:

^{:launch-js "headless-chrome"}
{:main example.normal-test-runner}

And then use it like this:

clojure -m figwheel.main -co tests.cljs.edn  -m example.test-runner

The command vector method

The command vector method is pretty much the same as the shell script method but it allows you to execute arbitrary shell commands without needing to write a script.

The following will launch headless Chrome on my system:

:launch-js ["/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
            "--headless" "--disable-gpu" "--repl" :open-url]

Using a Clojure function

This time let’s use a Clojure function to launch headless Chrome. First we’ll place a headless-js-env function in our user.clj file.

(ns user
   (:require [clojure.java.shell :as shell]))

(defn headless-js-env [{:keys [open-url]}]
  (shell/sh "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
            "--headless" "--disable-gpu" "--repl" open-url))

Now we will configure :launch-js to take a user/headless-js-env symbol:

clojure -m figwheel.main -fwo '{:launch-js user/headless-js-env}'  -m example.test-runner