HHH-16710, HHH-16742, HHH-16747 implicit instantiation for HQL result types

[gavinking]

This patch incorporates several different improvements to the handling of query "result types".

1. When a result type is passed to createSelectionQuery(), the select new Xxx(...) is now optional, and results will be happily packaged as a Map, List, or record, just like what happens already with Tuple.
2. Similarly, for native queries when a result type is passed to createNativeQuery().
3. Fixed some bugs in Tuple instantiation.
4. Fixed, as well as possible for now, issue https://hibernate.atlassian.net/browse/HHH-16747, so that creating a query with no explicit result type is always that same as either and entity result type, or the result type Object[], and never has a weird third behavior which cannot be recovered for any explicit result type.
5. Deprecated the createXxxxQuery() method with no explicit result type, since, even so, they still have pretty unintuitive behavior when the select clause is missing. I propose to eventually undeprecate them and have them throw an exception when the select clause is missing (except, perhaps, in extremely unambiguous cases).
6. Properly documented all of this stuff, including in the Javadoc.

[hibernate-github-bot]

Thanks for your pull request!

This pull request does not follow the contribution rules. Could you have a look?

❌ All commit messages should start with a JIRA issue key matching pattern HHH-\d+
    ↳ Offending commits: [50f2b73, 2e78325, 2f9071c]

› This message was automatically generated.

[gavinking]

There's now too much stuff rolled into this PR, and I'm starting to suffer from merge conflicts, so I'm going to merge this soon.

We're going to have to make some time to discuss the fallout, however.