A correction to what I said earlier: It looks like WPT uses the same technique as Selenium, to randomise the port, but only if the `port` kwarg is `None`: https://searchfox.org/mozilla-central/source/testing/web-platform/tests/tools/wptrunner/wptrunner/webdriver_server.py#109 So the easiest fix for bug 1571426 is to ensure `defaults.DRIVER_PORT` is set to `None`, though this comes at the slight risk of a race condition.
Bug 1574097 Comment 3 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
A correction to what I said earlier: It looks like WPT uses the same port randomisation technique as Selenium for other test types, by setting leaving the `port` kwarg as `None`, which will call `get_free_port()`: https://searchfox.org/mozilla-central/source/testing/web-platform/tests/tools/wptrunner/wptrunner/webdriver_server.py#109 So the easiest fix for bug 1571426 is to ensure `defaults.DRIVER_PORT` is set to `None`, though this comes at the slight risk of a race condition.
It looks like WPT uses the same port randomisation technique as Selenium for other test types by setting leaving the `port` kwarg as `None`, which will call `get_free_port()`: https://searchfox.org/mozilla-central/source/testing/web-platform/tests/tools/wptrunner/wptrunner/webdriver_server.py#109 With wdspec this gets set to `defaults.DRIVER_PORT`, so likely the easiest fix for bug 1571426 is to ensure that is set to `None`. This comes at the slight risk of a race condition, in that the port is bound by `get_free_port()` and subsequently released. This is not atomic, and another process could potentially bind to the port before geckodriver has time to.