Fixes #16

• We maintain backwards compatibility by continuing to allow the camelCase names in addition to the kebab-case names during argument parsing.

• When a an explicit name = ??? is given to the @main or @arg annotation, that takes precedence over everything, and is not affected by the name mapping,

• Name mapping is configurable by passing in nameMapper = mainargs.Util.snakeCaseNameMapper or nameMapper = mainargs.Util.nullNameMapper when you call ParserForClass or ParserForMethods

• I had to add a whole bunch of annoying shims to maintain binary compatibility when threading the new nameMapper through all our method signatures. That would be resolved by a proposal like https://contributors.scala-lang.org/t/can-we-make-adding-a-parameter-with-a-default-value-binary-compatible/6132/3, which alas does not exist yet in the Scala implementation

• The duplication in method argument lists is getting very annoying. Again, this would be solved by a proposal like https://contributors.scala-lang.org/t/unpacking-classes-into-method-argument-lists/6329, which still doesn't exist in the language

• Bumping to 0.6.0 since we cannot maintain bincompat for Scala 3 and Scala 2 simultaneously

  • There is no way to continue to evolve the case classes that is compatible with both Scala 2 and Scala 3, due to differing method signature requirements. e.g. def unapply(x: MyCaseClass): Option[Tuple] vs def unapply(x: MyCaseClass): MyCaseClass.
  • The choice is either to break bincompat in Scala 2 or break bincompat in Scala 2, and I ended up choosing to do so in Scala 2 since those would have the larger slower-moving codebases with more of a concern for binary compatibility

• Updated the docs and added coverage in the unit tests

• I intend to release this as 0.5.5 once it lands