Static factory methods for Association

[odrotbohm]

Introduced Association.forAggregate(…) and ….forId(…) to easily create Association instances without having to explicitly implement it.

[odrotbohm]

Thinking about this, I wonder how we should treat equality for Associations in general and the ones created by the factory methods in particular. I.e. should this be true or not:

SomeIdentifier identifier = new SomeIdentifier(…);
SomeAggregate aggregate = new SomeAggregate(identifier);

assertThat(Associations.forId(identifier)).isEqualTo(Association.forAggregate(aggregate));

While they're logically the same as they're pointing to the same aggregate for technical reasons, they're currently not as they're implemented using lambdas. that would be resolvable by introducing a dedicated (probably hidden) type. Even with that in place we'd have to answer how that's going to work when comparing those ad hoc Associations with custom Association implementations:

class SomeAssociation implements Association<SomeAggregate, SomeIdentifier> { /* … */ }

assertThat(SomeAssociation.of(identifier)).isEqualTo(Association.ofId(identifier));

We're never gonna get this to work symmetrically. I.e. different Association implementations can never be equal(…) in the first place. I am still leaning towards requiring that two different calls to Association.ofId(…) should produce instances that are equal. Also the assertion at the very top should be (mixing aggregate and identifier instances ultimately pointing to the same identifier).

WDYT?

[odrotbohm]

I've updated the feature branch with the package protected SimpleAssociation to ensure equality as described above and introduced a pointsToSameAggregateAs(…) for sole identifier comparison.

[odrotbohm]

Unless there are objections, I'd merge this later tonight or early morning tomorrow.

[odrotbohm]

That's merged.