WebDriverTestharnessExecutor should own bidi v. classic decision

[gsnedders]

Currently this is in the ChromeDriverTestharnessExecutor:

wpt/tools/wptrunner/wptrunner/executors/executorchrome.py

Lines 279 to 284 in 8fde6da

	require_webdriver_bidi = kwargs.get("browser_settings", {}).get(
	"require_webdriver_bidi", None)
	if require_webdriver_bidi:
	self.protocol_cls = ChromeDriverBidiProtocol
	else:
	self.protocol_cls = ChromeDriverProtocol

Coming from:

wpt/tools/wptrunner/wptrunner/browsers/chrome.py

Line 284 in 8fde6da

self._require_webdriver_bidi = test.testdriver_features is not None and 'bidi' in test.testdriver_features

I am doubting whether this is the right place for this logic to live; the RFC wasn't in any way Chrome specific, but we do somehow need to know if the browser supports bidi and whether or not we want to enable it.

cc @sadym-chromium

[sadym-chromium]

Good point. The main motivation to put the specific protocol selector logic in ChromeDriverTestharnessExecutor was to minimize the potential impact on other implementations.

I started drafting solution for the concern in #50626.