Add testdriver support for generate_test_report WebDriver command.

[paulmeyer90]

Also adds Reporting API tests which utilize this command.

[jgraham]

TestDriver additions now require an RFC in https://github.com/web-platform-tests/rfcs/ to allow fedback from across the project.

In this case it's unclear what specifies the underlying WebDriver API since it isn't part of the WebDriver spec. So your RFC should link to the relevant spec text.

[paulmeyer90]

I have opened an RFC issue here: web-platform-tests/rfcs#7

[paulmeyer90]

TestDriver additions now require an RFC in https://github.com/web-platform-tests/rfcs/ to allow fedback from across the project.

In this case it's unclear what specifies the underlying WebDriver API since it isn't part of the WebDriver spec. So your RFC should link to the relevant spec text.

RFC has passed

[paulmeyer90]

@jgraham are you satisfied with this change now?

[jgraham]

I think the testdriver changes are OK, but I shouldn't review the actual tests :)

RFC has passed.

[LukeZielinski]

Lgtm, thanks Paul.

[foolip]

The failing Taskcluster run on this PR is because testdriver.js was modified and tests including it were run. If they weren't stable, well then the job fails. Here are the Chrome failures:

Test	Subtest	Results	Messages
/event-timing/event-timing-timingconditions.html	Event Timing only times certain types of trusted event.	FAIL: 2/10, PASS: 8/10	assert_equals: Should only observe one entry: []. expected 1 but got 0
/pointerevents/pointerevent_pointerout_pen.html		ERROR: 1/10, OK: 9/10
/pointerevents/pointerevent_pointerout_pen.html	pointerout event received	PASS: 9/10, MISSING: 1/10
/speech-api/SpeechSynthesis-pause-resume.tentative.html		OK: 1/10, TIMEOUT: 9/10
/speech-api/SpeechSynthesis-pause-resume.tentative.html	speechSynthesis.pause() and resume() state and events	FAIL: 1/10, TIMEOUT: 9/10	assert_unreached: error event Reached unreachable code;Test timed out

And Firefox:

Test	Subtest	Results	Messages
/pointerevents/pointerevent_setpointercapture_inactive_button_mouse.html		OK: 8/10, TIMEOUT: 2/10
/pointerevents/pointerevent_setpointercapture_inactive_button_mouse.html	setPointerCapture + inactive button state	FAIL: 8/10, MISSING: 2/10	InvalidPointerId: Invalid pointer id.

@paulmeyer90 if none of these are due to changes that you made, which it doesn't seem like, then just let me know when this is ready to merge. I'll file bugs about the remaining failures and admin merge this. Sorry about the inconveniece, this is tracked by #14564 but I don't have a good fix for it.

[paulmeyer90]

@foolip None of those appear to be related to my change, since all the tests added by my change pass, and none of those failing tests use the new API that I added (and those are the only changes in this PR).

[Hexcles]

Force-merged as per the comment above.

[foolip]

@Hexcles can you file the bugs as well?

[Hexcles]

Filed #15486 for pointer events, as well as #15485 for the more general question of skipping testdriver.js altogether (same as testharness.js).

I'm looking into the event-timing and speech-api flakes. Those don't seem to be flaky on master.