Add compatibilityChangeExcludes configuration

[simonbasle]

This commit adds the compatibilityChangeExcludes configuration option
which allows to simply turn well-known JApiCompatibilityChange enum
values into compatible changes, effectively excluding all instances
of such violations from the API comparison.

Fixes #43.

[melix]

Thanks for the PR! I think I would prefer, rather than excluding the problems, to actually turn them into accepted changes, so that they show up in reports. Basically, that would mean using the ViolationTransformer rule. WDYT?

[simonbasle]

I have tried doing that, and it doesn't work well for the METHOD_NEW_DEFAULT case at least. I was running into blocker after blocker and found myself having to write a ton of logic, eg. to detect that a class-level violation ("class has incompatible changes" IIRC) with only accepted changes below (the METHOD_NEW_DEFAULT ones) should also be accepted / ignored, etc...

The library has a japicmp.cmp.JarArchiveComparatorOptions.OverrideCompatibilityChange and supports this use case (including in the Maven plugin), and I think that it is a valuable feature and a valuable convenience compared to ViolationTransformer (essentially supporting the METHOD_NEW_DEFAULT case with a configuration one-liner).

Sure this PR is opinionated compared to the Maven plugin, which allows to individually change the isBinaryCompatible() and isSourceCompatible() for each enum value. I refrained from doing that to avoid having to expose an equivalent of OverrideCompatibilityChange to plugin users for now.

Sure, with some elbow grease you could do the same using ViolationTransformers, but I found that's a LOT of elbow grease. And in my case, I gave up and painstakingly listed all the new default methods as methodExcludes instead.

[melix]

Right, indeed it's painful to figure out you need to silence the top level error when all children have been resolved. This is something I'd like to implement directly in the plugin.

[melix]

Given that I didn't have the opportunity to work on a "better" solution, I'll get this into the next release, thank you!