There are a number of blocking details to sort out here, but I figured it was worth posting this to get things moving and get some initial feedback.

1. There is ongoing, unresolved discussion in Tentative testdriver methods rfcs#226 as to whether these TestDriver methods need to be prefixed with tentative_. In short, RFC 127 just says "an existing WebDriver endpoint", which isn't explicit about a formal spec or where it needs to be specified. If an explainer is sufficient, we have Explainer: WebDriver Extension for Accessible Nodes, etc. (potential solution for #197) WICG/aom#203.
2. In https://github.com/web-platform-tests/rfcs/pull/226/files#r2411669032, @cookiecrook questioned whether we should be removing accessible children/parent ids from this first pass in order to unblock it sooner. I feel this is a fundamental detail we need to sort out because it is a primary goal of the work here. However, pragmatism is important, so I'm happy to go with the consensus here. I guess that would also involve removing verifyAccessibilityTree from this first pass.
3. executormarionette being the only implementation in this PR is a bit odd. On the other hand, it's difficult to implement (and verify the implementation of) this in executorwebdriver because we don't have a browser implementation yet to use.

From our last meeting I think we can omit "tentative" prefixes in the API names and instead assert that the methods are only used in tentative tests. I've implemented that in this PR.

(All my comments should be considered optional suggestions! I think this is fine already.)

@cookiecrook Re-requesting review because your review is still showing as changes requested. let me know if there's anything else you would like adjusted here. Thanks.

@jcsteh is this ready to merge?

@jcsteh is this ready to merge?

Currently, these tests aren't passing on wpt.fyi even in Firefox because I need to get the updates landed in Firefox and then relevant updates need to reach the WPT infra. I was holding off on that until we stabilised the method names, etc., but I'll land those patches now. I think it'd be good to see this passing in Firefox at least before we land the PR in order to show that it actually works, but I'm also happy to land it sooner and file follow-ups if you think that's more effective.

The tests are now passing on wpt.fyi for Firefox! I don't understand these Azure Pipelines CI failures though. It says that test_list_tests_missing_manifest is failing, but I've just spent quite some time digging through dashboards and logs and I cannot find any useful information about what the actual problem is.

I asked @jgraham about this and apparently this Azure Pipelines error can be disregarded.

🎉