Add Cookie#setAttribute, v2

[BalusC]

As requested by @gregw here #399 (comment)

[gregw]

I like the semantics of this one more.... but slightly concerned with the efficiency of having to always create the map... basically on every request with a cookie.

This would be OK, so long as an implementation could override methods with a more efficient implementation for cookies that it knows have no special attributes. But this version will not let the underlying attributes map not be created.

In reality, it was a mistake that this was not an interface, with readonly and read/write versions... ie there is no need to call a setter on a Cookie obtained from a request.

As this can effect trillions of requests, every nano second we can save here will save many Kgs of carbon. So give me a day or two and I'll see if I can come up with a version that keeps the API, but can be more efficient for the most frequent cases.

[gregw]

I've created #401 that lazily creates the attribute map, so it will not be instantiated on the most frequent usage of an inbound cookie

[gregw]

@markt-asf @stuartwdouglas should we be creating unit tests for any API that we update/change? If so... is there a test framework we should use?

[stuartwdouglas]

I think unit tests are a good idea. My preference would be JUnit 5.

[BalusC]

Closing off. This is superseded by #401