timeout added to start test in qunit goes towards the runtime

Question

I am adding some qunit test cases for a module. Few of the test cases have sync processes which I am using the standard stop() and start() as per docs.

My questions is, isn't the fact that the extra 1 second from setTimeout(function () { start();}, 1000); is added to the runtime of the test run, making the results in accurate?

I am a little not satisfied that +1000ms is added to the runtime as outside of the testsuite, inside the real app that uses that module that process completes without the 1000ms added here to carry out the test.

So when I pass this interface to less technical tester I have to explain in the title of the test to subtract that 1000 from that test before adding them up or whatever to calculate overall speed etc. [I basically want a way to have that extra timeout removed from the results automatically]

Module code below:

    define("tests/admin.connections.tests", ["mods/admin.connections", "datacontext"], function (connections, datacontext) {

    module("ADMIN PAGE CONNECTION LIST MODULE", {
        setup: function () {
            //ok(true, "once extra assert per test for Search Modules");
        }
    });

    test('Module is available?', function () {
        equal(_.isUndefined(connections), false, "connections js module exists");
        equal(_.isObject(connections), true, "connections js module is valid object");
    });

    test('HTML and CSS loading correctly? [Subtract 1 second from time to get the real time lapsed]', function () {
        function testHtml(html) {
            var d = document.createElement('htmlTestDiv');
            d.innerHTML = html;
            return d.innerHTML.replace(/\s+/g, ' ');;
        }
        stop();
        $.get('http://media.concepglobal.com/cbaddons/templates/connections.html', function (data) {
            equal(testHtml(connections.html), data.replace(/\s+/g, ' '), 'Html of the module was correctly loaded');
        });
        $.get('http://media.concepglobal.com/cbaddons/styles/connections.css', function (data) {
            equal(testHtml(connections.css), data.replace(/\s+/g, ' '), 'CSS of the module was correctly loaded');
        });
        setTimeout(function () { start();}, 1000);        
    });

    test('getConnectionsByUserId Async Method [Subtract 1 second from time to get the real time lapsed]', function () {
        function getConnectionsByUserId(successCallback) {
            amplify.request("getConnectionsByUserId", { uid: '0' }, function (data) {
                connections.userConnectionsCallback(data);
                successCallback();
            });
        }
        stop();
        getConnectionsByUserId(function () {
            var connectionsReturnedData = connections.connectionListViewModel.connections();
            expect(2);
            ok(_.isArray(connectionsReturnedData), 'Valid array has been returned for connections: ' + connectionsReturnedData);
            equal(connectionsReturnedData[0].type(), "sitecore", 'First returned object has a type property of "' + connectionsReturnedData[0].type() + '" and we expected it to be "sitecore"');
        });
        setTimeout(function () { start(); }, 1000);
    });

});

Answer 1

QUnit saves the currently running test in QUnit.config.current, this allows you to change the test during it's execution.

What you probably want is to reset the timer of the test after the timeout.

I created a little example to show what I mean (see on jsFiddle):

asyncTest("Long running test with 2s timeout", function () {
    expect(1);
    ok(true);
    setTimeout(function () { 
        QUnit.config.current.started = +new Date();
        start();
    }, 2000);
});

Like that the timer is reset once the timeout is over. This results in more accurate runtime in terms of what is executed. Now only the total time shows how much time was actually used to run all tests.