Record constructor with single write-only parameter should result in properties-based creator, to fix #3897

[yihtserns]

Currently, this is how a 1-arg constructor parameter is evaluated (priority according to the listed order):

	What	Properties-based	Delegating
1	Constructor parameter has @JsonProperty("<some name>")	✅
2	Any field has @JsonValue		✅
3	Any getter method has @JsonValue		✅
4	Constructor parameter has @JacksonInject	✅
5	Constructor parameter name matched a getter method name	✅
6	Constructor parameter name matched a field name	✅
7	Default		✅

This PR changes that to:

	What	Properties-based	Delegating
1	Constructor parameter has @JsonProperty("<some name>")	✅
2	Any field has @JsonValue		✅
3	Any getter method has @JsonValue		✅
4	Constructor parameter has @JacksonInject	✅
5	Constructor parameter name matched a getter method name	✅
6	Constructor parameter name matched a field name	✅
6	Is Record canonical constructor parameter ⭐ NEW	✅
8	Default		✅

[cowtowncoder]

@yihtserns Do you think this would be safe enough to go in 2.15(.1)? I think I'd merge it if you are comfortable.

If you are ok with that, would it be possible to re-based/re-create against 2.15 branch?

[cowtowncoder]

Minor thing but can remove first check and move inside previous if-block (which checks for propDef != null).

[yihtserns]

Actually I agonized over these 2 options:

Option 1	Option 2
private boolean _checkIfCreatorPropertyBased(...) { ... if (propDef != null) { ... if (implName != null && !implName.isEmpty()) { if (propDef.couldSerialize()) { return true; } else if (beanDesc.isRecordType()) { // [databind#3897]: Record canonical constructor will have implicitly named propDef return true; } } } // in absence of everything else, default to delegating return false; }	private boolean _checkIfCreatorPropertyBased(...) { ... if (propDef != null) { ... } // [databind#3897]: Record canonical constructor will have implicitly named propDef if (propDef != null && !propDef.isExplicitlyNamed() && beanDesc.isRecordType()) { return true; } // in absence of everything else, default to delegating return false; }

I chose Option 2 because it looks easier to understand with just a quick glance. But I'll change it if you feel strongly about Option 1.

[yihtserns]

If you are ok with that, would it be possible to re-based/re-create against 2.15 branch?

Actually the branch was created from 2.15, I didn't notice the PR is against 2.16. Let me change that.