What is the issue with the HTML Standard?

The iframe removing steps destroy the child navigable, and I am wondering how one should deal with the case where the window is still accessible from script, and then script attempts to create an auxiliary using the window open steps of that window.

Minimal test case here

In Firefox, the test seems to hit a non-standard early return.

The spec seems silent on that issue.

Perhaps the window open steps should check whether sourceDocument still has a browsing context associated with it(destroying a document sets the bs to null see step 8 of destroy a document), and if not, do an early return?