We have a rather big set of end-to-end tests on Protractor. We are following the Page Object pattern which helps us to keep our tests clean and modular. We also have a set of helper functions which help us to follow the DRY principle.
The Problem:
A single spec may require multiple page objects and helper modules. For instance:
"use strict";
var helpers = require("./../../helpers/helpers.js");
var localStoragePage = require("./../../helpers/localStorage.js");
var sessionStoragePage = require("./../../helpers/sessionStorage.js");
var loginPage = require("./../../po/login.po.js");
var headerPage = require("./../../po/header.po.js");
var queuePage = require("./../../po/queue.po.js");
describe("Login functionality", function () {
beforeEach(function () {
browser.get("/#login");
localStoragePage.clear();
});
// ...
});
You can see that we have that directory traversal in every require statement: ./../..
. This is because we have a specs
directory where we keep the specs and multiple directories inside grouped by application functionality under test.
The Question:
What is the canonical way to approach the relative path problem in Protractor?
In other words, we'd like to avoid traversing the tree, going up to import modules. It would be much cleaner to go down from the base application directory instead.
Attempts and thoughts:
There is a great article about approaching this problem: Better local require() paths for Node.js, but I'm not sure which of the options is a recommended one when developing tests with Protractor.
We've also tried to use require.main
to construct the path, but it points to the node_modules/protractor
directory instead of our application directory.