Consistent handling of Specification.unrestricted() in Specification.not(..)

[peteraisher]

Summary

This PR updates the handling of Specification.unrestricted() (and other specifications whose toPredicate(..) method returns null) in Specification.not(..) to ensure consistent handling across all logical operations:

unrestricted().or(other) // equivalent to other not(unrestricted()).or(other) // equivalent to other *changed* unrestricted().and(other) // equivalent to other not(unrestricted()).and(other) // equivalent to other *changed* not(unrestricted()) // equivalent to unrestricted() *changed* 

Changes

• Updated tests to verify that not(unrestricted()).toPredicate(..) returns null and that no CriteriaBuilder methods are called.
• Updated implementation accordingly.
• Updated documentation to better explain the semantics of a specification whose toPredicate(..) method returns null, in particular the static unrestricted() factory method.
• Mirrored changes across Specification, DeleteSpecification, PredicateSpecification, and UpdateSpecification classes.

PR Checklist

• I have read the Spring Data contribution guidelines.
• I used the code formatters provided here and have them applied to my changes.
• I am submitting updated test cases that back my changes.
• I added myself as author in the headers of the classes I touched, and amended the date range in the Apache license headers where needed.

Closes #4023

[mp911de]

Thank you for your contribution. That's merged, polished, and backported now.