More fine-grained exclusions (static vs non-static method problems)

[jrudolph]

Changes to static forwarder generation in scalac may require exclusion for DirectMissingMethodProblem but those are quite heavy-handed and may mask actual problems later on (e.g. see https://github.com/akka/akka-http/pull/2656/files#diff-0744dca53576695333760d7659be094fR48). Instead it would be good if you could distinguish static vs. normal methods.

If it would only be for DirectMissingMethodProblem I guess it could be solved by a subtype DirectMissingStaticMethodProblem that could be excluded specifically. Another possibility could be to allow adding some modifiers to the exclusion string (that would help when you would like to filter other problems in a similar way).

[dwijnand]

I'm happy to add a subtype thing, for this case. Could you craft a minimiser, please?

[dwijnand]

While MiMa could most definitely use "more fine-grained exclusions", I'm going to close this in favour of the more specific issue on handling the change in static forward generation #388, which doesn't have any implementation decisions (or compromises) like we were discussing here.

[dwijnand]

While MiMa could most definitely use "more fine-grained exclusions", I'm going to close this in favour of [...]

Actually, I've changed my mind. Though let's keep the scope limited to the static/non-static use-case presented.