Test contexts

Version
See buster-test

A test context is Buster's internal test case/specification data format. Specifically, a test context is the kind of object that the test runner knows how to run. This document describes the data format in detail - what features it and the runner supports as well as how to create test contexts from external front-ends.

The idea behind test contexts is to separate the syntax you use to write a test case/specification (the front-end) from the tools that run and visualize results of these tests. Test contexts have enabled Buster to ship with two rather different ways of writing tests for the same runner; buster.testCase and buster.spec. They also enable us to provide adapters for other test frameworks, such as in buster-jstestdriver.

If you are looking to run tests written for another testing library with Buster, you have come to the right place. Supporting other test frameworks front-ends is a simple matter of converting the objects/functions created by the library to a test context object. For inspiration, see the implementation of the aforementioned buster.testCase, buster.spec and buster-jstestdriver.

NB! Test contexts are not intended to be hand-written, rather they represent the data test specification DSLs should produce in order to use the Buster test runner.

testContext

Represents a test case (xUnit terminology) or a specification (BDD terminology). Contains a set of tests to run, optional setup and teardown methods as well as optional nested contexts. Refer to the example for more information on how this can be used in practice.

Setup and teardown methods can be asynchronous, see the section on asynchronous tests.

name
The test context name/description as a string
setUp
Optional. A function that will run before each test. If using nested contexts, the setup function will also be executed before each test (and their local setup, if any) in nested contexts.
tearDown
Optional. A function that will run after each test. If using nested contexts, the teardown function will also be executed after each test (and their local teardown, if any) in nested contexts.
tests
An array of test objects.
contexts
An array of testContext objects. In other words, the test context data format (and thus, the test runner) supports arbitrarily nested contexts.
parent
The parent testContext, if any.
testCase
Optional. Prototype object used for this when running tests. This object may define helper methods and properties to use in tests. The test runner creates a new instance from this object with buster.create for each test. The created object is shared as this in all setup and teardown methods as well as in the test. If this object is not provided, an "empty" object is created and used as this when running tests.

test

Represents a test function.

name
The test function name, as a string.
func
The test function. See the section on asynchronous tests for how to mark it as - well, asynchronous.
context
Optional. The context to which the test belongs.
deferred
Optional. If this property is set to true, the test will not be run, but the test runner will emit an event for it, allowing reporters to communicate tests that should eventually run (and hopefully pass).

Asynchronous tests

There is no flag to mark tests as asynchronous even though the test runner supports both synchronous and asynchronous tests. The reason is that it cannot be determined up-front if a test is asynchronous or not.

To create asynchronous tests (i.e. ones that the runner will wait for), the test function can either explicitly accept a single argument, which is a function, or return a thenable promise.

The argument passed to the test is a function. When the function is called, the asynchronous test is deemed done. The idiomatic way of creating asynchronous tests using this arguments looks like the following:

function someAsyncTestFunction(done) {
    setTimeout(function () {
        buster.assert(true);
        done();
    }, 100);
}

This assumes that the assertion framework can fail without throwing an error (as an error would be intercepted as uncaught in the above example, if intercepted at all). If this is not the case, you can make your assertions in a callback to the done function:

function someAsyncTestFunction(done) {
    setTimeout(function () {
        done(function () {
            buster.assert(true);
        });
    }, 100);
}

Tests can also be made asynchronous by way of returning a promise. The test runner considers any object with a then method a promise:

function someAsyncTestFunction() {
    var promise = {
        then: function (callback) {
            this.callbacks = this.callbacks || [];
            this.callbacks.push(callback);
        }
    };
    setTimeout(function () {
        buster.assert(true);
        var callbacks = promise.callbacks || [];
        for (var i = 0, l = callbacks.length; i < l; ++i) {
            callbacks[i]();
        }
    }, 100);
    return promise;
}

Note that this does not work entirely as expected unless your assertion framework of choice is able to notify the runner of failure without throwing an exception. If the assertion fails (and throws an exception), the promise will never be resolved, thus the runner will fail the test with a timeout, not an assertion error.

The above example is very verbose, simply to illustrate the duck-typed nature of promises. You can do better by using e.g. when.js:

function someAsyncTestFunction() {
    var deferred = when.defer();
    setTimeout(function () {
        buster.assert(true);
        deferred.resolver.resolve();
    }, 100);
    return deferred.promise;
}

Setup and teardown functions can use the same mechanism to be asynchronous.

Example

Say you have a test case like the following (warning: fictional front-end, this is just to explain what goes where in the generated context):

testCase("Circle tests", {
    createCircle: function (radius) {
        return {
            diameter: function () {
                return radius * 2;
            }
        };
    },
    "diameter should equal twice the radius": function () {
        var circle = this.createCircle(6);
        buster.assert.equals(circle.diameter(), 12);
    }
});

The test case has a single test and a helper method defined on the same object (thus accessed through this in the test). This simple test case can be represented as a Buster runnable context the following way:

var simpleContext = {
    name: "Circle tests",
    testCase: {
        createCircle: function (radius) {
            return {
                diameter: function () {
                    return radius * 2;
                }
            };
        }
    },
    tests: [{
        name: "diameter should equal twice the radius",
        func: function () {
            var circle = this.createCircle(6);
            buster.assert.equals(circle.diameter(), 12);
        }
    }]
};

The following example is reproduced from the official QUnit docs and shows a fairly typical QUnit test.

test("a basic test example", function () {
    ok(true, "this test is fine");
    var value = "hello";
    equals("hello", value, "We expect value to be hello");
});
module("Module A");
test("first test within module", function () {
    ok(true, "all pass");
});
test("second test within module", function () {
    ok(true, "all pass");
});
module("Module B");
test("some other test", function() {
    expect(2);
    equals(true, false, "failing test");
    equals(true, true, "passing test");
});

The corresponding test context would look like:

var qunitContext = {
    name: "" // Top level module was nameless
    tests: [{
        name: "a basic test example",
        func: function () {
            ok(true, "this test is fine");
            var value = "hello";
            equals("hello", value, "We expect value to be hello");
        }
    }],
    contexts: [{
        name: "Module A",
        tests: [{
            name: "first test within module",
            func: function () {
                ok(true, "all pass");
            }
        }, {
            name: "second test within module",
            func: function () {
                ok(true, "all pass");
            }
        }]
    }, {
        name: "Module B",
        tests: [{
            name: "some other test",
            func: function() {
                expect(2);
                equals(true, false, "failing test");
                equals(true, true, "passing test");
            }
        }]
    }]
};

NB! Test contexts are not intended to be hand-written, rather they represent the data test specification DSLs should produce in order to use the Buster test runner.