Description
What is the issue with the HTML Standard?
After #10818 merges, we will have an implementation-defined step which establishes subtreesToSearch in "find a navigable by target name".
@kjmcnee states:
WebKit and Chromium search the requesting window's subtree then search from the top. Firefox iterates and searches from each ancestor. If there's a way to express in the spec that the implementation-defined behavior is fixed for the instance of the user agent, then that'd be appropriate here.
We should probably standardize on the 2/3 behavior here, unless Gecko has a strong argument for why theirs is better. This issue tracks doing so.
The only hard part will be writing appropriate web platform tests. For inspiration on how to write such tests, you can see Kevin's web-platform-tests/wpt#49631 . We just need to establish a setup where the two strategies give different results and then assert the WebKit/Chromium behavior.