add support to detect accesses of fields by specific methods

[leonardhusmann]

This PR adds a method to the public API of FieldsShould to check if a field has been accessed by methods matching a given predicate. This is checked by a new ArchCondition.

I've diverged a little bit from the original suggestion of this issue: I chose to stay consistent with the current API and return an ArchCondition<JavaField> to allow method chaining, e.g.

fields().that().inSomeCondition()
    .should().notBeAccessedByMethodsThat(...)
    .andShould().notBeFinal()
    .andShould()....

What do you think? Any feedback is highly appreciated :)

Resolves: #857

[codecholeric]

Thanks a lot for adding this! 😃 And sorry that it took me so long to look at it! First of all, what do you mean by

| return an ArchCondition<JavaField> to allow method chaining

? I don't see how returning that would allow chaining as in the example? But in any case, what you did looks consistent, so thanks 🙂 Only thing I noticed is that for all the other methods we have a positive and negative form and I realized that there might be use cases where users want to combine this with a noFields start of a rule. So I added the inverted version (and adjusted the test, because I think that was a little too much mingling, i.e. it deserves it's own test). What do you think?

Independently of that, retrospectively I should have added something smarter like a generic not() function to the API 🙈 Instead of "duplicating" every method. But it's already a little inconsistent between classes and fields 😒 So maybe at some point we will release a 2.0.0 and clean up / consolidate the rules API again...

[codecholeric]

Anyway, what do you think about the review changes in my commit? If everything looks good to you I'd squash everything into one commit...

[leonardhusmann]

Hi, thank your for your feedback 😃

First of all, what do you mean by

This is meant with regards to the initial suggestion in the issue from April '22. What I meant by this is that the arguments (DescribedPredicate) are passed as parameters and the function returns the ArchCondition. Just as you did it in your first answer in the thread of the issue 😃.

Apart from that: thank your for the review changes, they make a lot of sense :) So everything looks good to me 👍

[codecholeric]

Cool, then thanks a lot again, squashed, rebased and now merged 😃