I’m trying to test some jQuery ajax code using QUnit and Mockjax and have it return different JSON for different tests, like this:
$(document).ready(function() {
function functionToTest() {
return $.getJSON('/echo/json/', {
json: JSON.stringify({
"won't": "run"
})
});
}
module("first");
test("first test", function() {
stop();
$.mockjax({
url: '/echo/json/',
responseText: JSON.stringify({
hello: 'HEYO!'
})
});
functionToTest().done(function(json) {
ok(true, json.hello);
start();
});
});
test("second test", function() {
stop();
$.mockjax({
url: '/echo/json/',
responseText: JSON.stringify({
hello: 'HELL NO!'
})
});
functionToTest().done(function(json) {
ok(true, json.hello);
start();
});
});
});
Unfortunately it returns the same response for each call, and order can’t be guaranteed, so was wondering how I could set it up so that it was coupled to the actual request and came up with this:
$.mockjax({
url: '/echo/json/',
response: function(settings) {
if (JSON.parse(settings.data.json).order === 1) {
this.responseText = JSON.stringify({
hello: 'HEYO!'
});
} else {
this.responseText = JSON.stringify({
hello: 'HELL NO!'
});
}
}
});
This relies on parameters being sent to the server, but what about requests without parameters, where I still need to test different responses? Is there a way to use QUnit’s setup/teardown to do this?
It looks like you need to call
$.mockjaxClear();before you create another mock handler.Mockjax works by changing the $.ajax method.
At the bottom of its source code we can see that the $.mockjax method you are using, which is one of its exposed methods, just appends more handlers to the mockHandlers array.
In the source code for the $.ajax replacement, we can see:
Your problem is due to the $.ajax method being satisfied with the first handler (therefore not the latest) mock handler in the array for the
/echo/json/url.Here is my fork of your fiddle, just adding a
$.mockjaxClear()line.Edit:
A more flexible solution:
<your handlerID variable>) to this backup. In the module’s teardown, use $.mockjaxClear(<your handlerID variable>) to delete the module, and then set<your handlerID variable>to something else.setupand in the individual test functions.But don’t take my word for it. Check out the fiddle.
It’s a lot more flexible like this. You can change that one handler and leave all the other handlers you might have intact.
It does seem to be potentially error-prone, so I would advise to be careful.