While this PR is ready for review, I want to wait a few days for importlib_resources 5.3 to be in the wild to ensure there aren't any major emergent concerns.

I've requested a review of this change with Barry and Brett, both of whom have had substantial influence on this interface and both of whom are on the SC. In this discussion, I've made the case that the deprecation is appropriate, but others have made the case that the deprecation happens too quickly, mainly because there won't be one interface that one could depend on across Python 3.13 and 3.8 without adopting a backport.

If either agree the deprecation is too aggressive, I'll back it out to a PendingDeprecationWarning and let it linger indefinitely.

While this PR is ready for review, I want to wait a few days for importlib_resources 5.3 to be in the wild to ensure there aren't any major emergent concerns.

There have been no reported issues after issuing the deprecation warning in importlib_resources 5.3.

I'd like for this change to be present in early releases of Python 3.11 to limit any surprise and get feedback early. I'd prefer to get a review of this change by next week. Please let me know if you would like more time to consider. If I don't hear anything by 2021-11-20 AOE, I'll plan to merge, but without prejudice; the change can be rolled back or dialed back as needed.

I don't have time to really weigh in on this, so consider me neutral.

I've decided to move forward with this change, with my rationale here.