Running Tests from the Local System

The tests are designed to be run from your local computer.

Install WPT

If you haven’t already, clone the web-platform-tests repository:

git clone https://github.com/web-platform-tests/wpt.git
cd wpt

System Setup

Running the tests requires python and pip as well as updating the system hosts file.

WPT requires Python 3.8 or higher.

The required setup is different depending on your operating system.

Linux Setup

If not already present, use the system package manager to install python, and pip.

On Ubuntu:

sudo apt-get install python3 python3-pip python3-venv

It is important to have a package that provides a python binary. On Fedora, for example, that means installing the python-unversioned-command package. On Ubuntu Focal and later, the package is called python-is-python3.

macOS Setup

The system-provided Python can be used, while pip can be installed for the user only:

python -m ensurepip --user
export PATH="$PATH:$( python3 -m site --user-base )/bin"

To make the PATH change persistent, add it to your ~/.bash_profile file or wherever you currently set your PATH.

See also additional setup required to run Safari.

Windows Setup

Download and install Python 3. The installer includes pip by default.

Add C:\Python39 and C:\Python39\Scripts to your %Path% environment variable.

The standard Windows shell requires that all wpt commands are prefixed by the Python binary i.e. assuming python is on your path the server is started using:

python wpt serve

Windows Subsystem for Linux

Optionally on Windows you can use the Windows Subsystem for Linux (WSL). If doing so, installation and usage are similar to the Linux instructions. Be aware that WSL may attempt to override /etc/hosts each time it is launched, which would then require you to re-run hosts File Setup. This behavior can be configured.

hosts File Setup

To get the tests running, you need to set up the test domains in your hosts file.

On Linux, macOS or other UNIX-like system:

./wpt make-hosts-file | sudo tee -a /etc/hosts

And on Windows (this must be run in a PowerShell session with Administrator privileges):

python wpt make-hosts-file | Out-File $env:SystemRoot\System32\drivers\etc\hosts -Encoding ascii -Append

If you are behind a proxy, you also need to make sure the domains above are excluded from your proxy lookups.

Via the browser

The test environment can then be started using

./wpt serve

This will start HTTP servers on two ports and a websockets server on one port. By default the web servers start on ports 8000 and 8443 and the other ports are randomly-chosen free ports. Tests must be loaded from the first HTTP server in the output. To change the ports, create a config.json file in the wpt root directory, and add port definitions of your choice e.g.:

{
  "ports": {
    "http": [1234, "auto"],
    "https":[5678]
  }
}

After your hosts file is configured, the servers will be locally accessible at:

http://web-platform.test:8000/
https://web-platform.test:8443/ *

To use the web-based runner point your browser to:

http://web-platform.test:8000/tools/runner/index.html
https://web-platform.test:8443/tools/runner/index.html *

This server has all the capabilities of the publicly-deployed version–see Running the Tests from the Web.

*See Trusting Root CA

Via the command line

Many tests can be automatically executed in a new browser instance using

./wpt run [browsername] [tests]

This will automatically load the tests in the chosen browser and extract the test results. For example to run the dom/historical.html tests in a local copy of Chrome:

./wpt run chrome dom/historical.html

Or to run in a specified copy of Firefox:

./wpt run --binary ~/local/firefox/firefox firefox dom/historical.html

For details on the supported products and a large number of other options for customising the test run:

./wpt run --help

A complete listing of the command-line arguments is available here.

Browser-specific instructions

Running in parallel

To speed up the testing process, use the --processes option to run multiple browser instances in parallel. For example, to run the tests in dom/ with six Firefox instances in parallel:

./wpt run --processes=6 firefox dom/

But note that behaviour in this mode is necessarily less deterministic than with a single process (the default), so there may be more noise in the test results.

Output formats

By default, ./wpt run outputs test results and a summary in a human readable format. For debugging, --log-mach can give more verbose output. (In particular, it will show the console output from the browser and driver; by default, those are not shown) For example:

./wpt run --log-mach=- --log-mach-level=info firefox dom/

A machine readable JSON report can be produced using --log-wptreport. This together with --log-wptscreenshot is what is used to produce results for wpt.fyi. For example:

./wpt run --log-wptreport=report.json --log-wptscreenshot=screenshots.txt firefox css/css-grid/

(See wpt.fyi documentation for how results are uploaded.)

Expectation data

For use in continuous integration systems, and other scenarios where regression tracking is required, the command-line interface supports storing and loading the expected result of each test in a test run. See Expectations Data for more information on creating and maintaining these files.

Testing polyfills

Polyfill scripts can be tested using the --inject-script argument to either wpt run or wpt serve. See Testing Polyfills for details.