Playwright Technology Preview

Since its inception, CumulusCI has relied on Selenium to provide the foundation of our browser automation keywords.

In 2020, Microsoft introduced a new browser automation tool named Playwright. Playwright is a ground-up reinvention of a browser automation library that aims to address several shortcomings of Selenium. For example, Playwright has built-in support for waiting for elements to appear, for working with the shadow DOM, video capture of a testing session, and so on.

In 2021 the Robot Framework project introduced the Browser library which adds keywords that use the Playwright API.

Starting with CumulusCI version 3.59.0, we are providing experimental support for Playwright and the Browser library in CumulusCI.

In CumulusCI 3.60, we’ve reorganized our keywords so that a test can import the API and performance keywords without importing Selenium keywords. To use Playwright-based keywords, import the resource file SalesforcePlaywright.robot, which imports the non-Selenium keywords along with the keywords in the SalesforcePlaywright library.

Installation

We have not yet bundled Playwright and the Browser library with CumulusCI. However, we have provided a script to make it easy to install or uninstall Playwright and the Browser library while we continue to work on fully supporting it.

Step 1: Install Node.js

Playwright is based on Node.js. If you don’t already have Node.js installed, you can find a Node.js installer for your platform on the Node.js downloads page.

Important

You must have Node.js installed before continuing with these instructions.

Step 2: Run the Playwright installation command

Installing the browser library requires a couple of manual steps, which we’ve automated in a single script. This script does three things:

  • it verifies that Node.js has been installed

  • it downloads and installs the Browser keyword library

  • it downloads and installs the Node.js modules and drivers for playwright.

Note

The installation of Playwright contains drivers for all supported browsers, so there’s no need to manually install drivers such as ChromeDriver. It works right out of the box!

Before you run the script, make sure your working directory is at the root of your repository. You can then run the script with the following command:

$ cci robot install_playwright

Tip

You can use the --dry_run (or -n) option to see what the command will do without actually installing.

Running an example test

As mentioned earlier, this is an experimental release of Playwright integration, so any CumulusCI keywords that rely on Selenium won’t work. However, the following example shows how easy it can be to write Playwright-based tests with off-the-shelf keywords provided by the Browser library

To initialize Playwright support in a test suite, import the SalesforcePlaywright.robot resource file as shown in the following example. It imports the Browser library and defines the keywords Open Test Browser and Delete records and close browser.

*** Settings ***
Resource     cumulusci/robotframework/SalesforcePlaywright.robot

Suite Setup      Open test browser
Suite Teardown   Delete records and close browser

*** Test Cases ***

Go to user profile
    Click    button:has-text("View profile")
    Click    .profile-card-name .profile-link-label

    Wait until network is idle
    Take screenshot

Go to contacts home
    Click            button:has-text("App Launcher")
    Fill text        input[placeholder='Search apps and items...']  Contacts
    Click            one-app-launcher-menu-item:has-text("Contacts")

    Wait until network is idle
    Take screenshot

To run the test, save the above code in a .robot file (e.g. example.robot) and then run it with the standard robot task:

$ cci task run robot --suites example.robot

Things to Notice

This example test is unable to use any of the existing Selenium-based keywords, except for two. We’ve created a new library based on Playwright and the Browser library with two keywords that are similar to existing keywords: Open Test Browser and Delete Records and Close Browser

This test also uses the Browser keyword Wait until network is idle before taking a screenshot. This is a keyword that waits for there to be at least one instance of 500ms of no network traffic on the page after it starts to load. This seems to be more reliable and easier to use method than waiting for a page-specific element to appear.

This test has no explicit waits for the buttons and links that it clicks on. The underlying Playwright engine automatically waits for elements, so there should almost never be a need for keywords such as Wait until page contains element or Wait until element is enabled.

Finally, notice how easy it is to interact with both the app menu and the user profile. Playwright locators are often much easier to write than Selenium locators, which translates to tests and keywords that don’t have to be tweaked when the page markup changes.

Writing keywords in JavaScript

At its core, Playwright is built on Node.js, which makes it possible to write keywords in JavaScript. This is enabled by providing the path to a JavaScript library when the Browser library is imported. CumulusCI comes with a small bootstrap file which will import a Node module from your project, which itself can import any number of other modules.

Tip

For more information about writing keywords in Javascript see Extending Browser library with a JavaScript module in the Browser library documentation.

Creating a keyword file in JavaScript

When you import SalesforcePlaywright.robot into a test, it will import the Browser library and pass it the name of a small bootstrap JavaScript file. This bootstrap file will import the file robot/<ProjectName>/javascript/index.js, exposing the exported functions as keywords.

Example

Who doesn’t love a good “Hello, world” example? Start by creating the javascript folder in your robot folder. For example, if your project is named Food-Bank, you should create a folder named robot/Food-Bank/javascript.

In that folder, create a file named keywords.js with the following content:

// keywords.js
function hello_javascript() {
    return "Hello, JavaScript!";
}

exports.__esModule = true;
module.exports = { hello_javascript };

Note

The Browser documentation for writing keywords in javascript uses async functions in all of the examples, but async functions are not strictly required.

Next, create a file named index.js in the same folder. This file is where you can import the keywords.js file and any other files you need to import. For our simple example index.js should look like the following example.

// index.js
exports.__esModule = true;
module.exports = {
    ...require("keywords.js"),
};

Getting CumulusCI test context

CumumlusCI comes with a node module named cumulusci which provides access to information about the repository and org.

Let’s extend our earlier example to include a keyword for getting the instance URL of the org under test.

The first step is to import the cumulusci module. This module returns data that looks like the following:

{
    "project_config": {
        "repo_name": "Food-Bank",
        "repo_root": "/projects/Food-Bank"
    },
    "org": {
        "name": "dev",
        "instance_url": "https://something-clever.my.salesforce.com",
        "org_id": "00D5C..."
    }
}

We can directly access these member attributes with a keyword. Edit the keywords.js file to look like the following example.

var cci = require("cumulusci");

function hello_javascript() {
    return "Hello, Javascript!";
}

function get_instance_url() {
    return cci.org.instance_url;
}

exports.__esModule = true;
module.exports = {
    hello_javascript,
    get_instance_url,
};

Summary

This is just a preview of things to come. The CumulusCI team will be spending more time evaluating Playwright, with an eye toward making it a viable and more robust replacement for Selenium.

Resources