Support long option only for certain switches about fluent-command-line-parser

[siywilliams]

+1 from me for optional short or long switches instead of only optional long switches.

I wonder if it may simplify the call if we provided two setup overloads, one taking in a single switch name which the switch type can then be inferred from the length of the value e.g. a single char is defined as a short switch and more than that is treated as a long switch. This would make it less onerous as you don't need to provide empty or null values for things you don't.

The signature could look like this Setup<T>(string longOrShort) which could be called for example with either Setup<bool>("s") which would setup a short switch or Setup<bool>("silent") which would setup a long switch.

An overload for providing both a short and long switch would still remain e.g. Setup<T>(string short, string long)

@jen20 any thoughts?

I appreciate the work on the tests by the way!

[jen20]

Sounds like a good plan re: reducing work. For short names, since we're using only one letter, why not actually take a char? So we'd have:

Setup<T>(char short)
Setup<T>(string long)
Setup<T>(char short, string long)

Appreciate that this is a breaking API change though.

[jen20]

The other thing that would be nice to include here is multiple long options (say, int-ip and internal-ip referring to the same option in different ways). Think this would probably need some changes to the formatter and parser. It could be implemented with two extra setup overloads:

Setup<T>(params string long)
Setup<T>(char short, params string long)

[jen20]

Have split the comments into separate issues, will close this one now.