Henrik Skupin [:whimboo][⌚️UTC+1] Reporter
	Description • 7 years ago

There are a couple of issues with parsing the proxy capabilities. Mostly because some parts of the webdriver spec are still not that clear about, and we don't know how people are actually using the proxy settings.

So what we want to do first is to get an impression how current tests as written for Selenium are using the proxy settings. If it turns out that the spec is offering to many different ways to specify the settings, and some are not used at all, we would like to strip them off.

Here a raw query for proxy tests for different Selenium bindings:

https://github.com/SeleniumHQ/selenium/search?utf8=%E2%9C%93&q=proxy%2Btest&type=

Based on the outcome it would be good to get started in creating wdspec tests first. Once a good set of tests have been put together we will be able to see which changes for geckodriver and Marionette are necessary.

	Henrik Skupin [:whimboo][⌚️UTC+1] Reporter
	Comment 1 • 7 years ago

We are actually pretty good in parsing the proxy capabilities in Marionette now. But we need more tests for webdriver-rust, and especially for the wdspec tests in web-platform-tests. I will start working on that bug once geckodriver 0.19 is out and webdriver-rust is part of the tree.

	Andreas Tolfsen ❲:ato❳
	Comment 2 • 6 years ago

More tests would be nice to have, but as far as I can tell this does
not align with conformance work unless we believe we have bugs in that
area?

	Andreas Tolfsen ❲:ato❳
	Comment 3 • 6 years ago

Also the webdriver crate is now in central, so that unblocks this.