HHH-19115 - implement ordered multiloading with natural ids

[jrenaat]

---

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license
and can be relicensed under the terms of the LGPL v2.1 license in the future at the maintainers' discretion.
For more information on licensing, please check here.

---

https://hibernate.atlassian.net/browse/HHH-19115

[beikov]

This skip blocks here and below are a bit confusing IMO as there is no javadoc comment on why the skips are necessary. Also, I'm a bit surprised that a multi-load of 3 composite natural ids only requires 3 parameters. What is rendered SQL for this? I would expect this always requires 6 parameters i.e. where (ssn,ssn2) in ( (?,?), (?,?), (?,?) ) because tuples can't be passed as a single parameter AFAIK.

[jrenaat]

The regex that i'm using for the tests that exclude Sybase, SQLServer, DB2 and hsqldb are looking at the (?,?) pattern, ergo, it checks whether there are 3
Sybase, SQLServer, DB2 and hsqldb produce different sql queries that effectively contain 6 '?', but using MultiKeyLoadHelper.supportsSqlArrayType for this didn't do the trick
I'll add a comment to the tests to clarify this.

[jrenaat]

@beikov Wrt to our conversation in zulip, i checked whether using a @DialectFeatureCheck as you suggested would be viable, but IMHO it isn't because I don't have acccess for now to the element that differentiates the dialects; it would be much uglier to hack something than to just leave the test as it was conceived.
Again, I want to have a clear assert for those tests that support row value syntax in IN lists that the generated sql effectively contains '(?,?)' for each compound natural id, I think it is important.