What is the issue with the HTML Standard?

This test tests that if a shared-worker takes a lock, a connected document cannot enter BFCache. I wonder if this is really the behaviour we want.

A document in BFCache should not prevent another document from taking a lock (the tests above actually test that such a dodument doesn't even enter BFCache which might be too strict).

In the ShW case, it's unclear to me why the document cannot enter BFCache. The ShW is capable of releasing the lock at any point, e.g. due to a message from another document. The only exception to this would be if the ShW somehow required a message from a specific document (e.g. the one in BFCache) to release the lock however such a SharedWorker is already inherently buggy and will end up holding that lock forever if the document in question is destroyed without signalling to release the lock (either through a bug or a crash).

Am I missing something?

I'd like to propose that we not block entry to BFCache for documents connected to a SharedWorker holding a WebLock. I think the same argument should apply to any othe potentially BFCache-blocking feature that a ShW uses.

@saschanaz @inexorabletash @yoshisatoyanagisawa @llannasatoll