Add bear settings to green mode project

The [quickstart green project](https://github.com/coala/projects/blob/master/_projects/quickstart_green_mode.md) should also include adding/exposing/standarding bears defaults.

The green mode will go around in circles when it finds different bears have undeclared defaults. We've done a lot of work tackling this problem with the tabs-vs-spaces and max-line-length for each bear, creating explicit settings with a default which shows how the linter works by default.

When these settings are missing, a green mode cant easily see why there are so many failures, and thus cant change a setting to allow the linter to accept the code.

Another type of useful change to bear settings is adding sensible defaults, like https://github.com/coala/coala-bears/pull/2227 . That allows the bear to be enabled more easily, because green mode doesnt need to guess an initial value without any context to know whether it should be a low number or a high number.

Even better is adding setting datatype constraints, where the data type of the setting specifies a valid range. e.g. `int` has a lot of possible values which green mode would need to try. However the range of suitable values might actually be logarithmic. e.g. a max_line_per_file of 1200003 isnt particularly meaningful. The datatype could be 1..1000, 2000, 3000 etc to 10,000 , then incrementing in 10,000s, and then incrementing in 100,000s etc until it reaches the maximum size of `int`.
Another problem is we are using `0` for unlimited chars per line. That is non-nonsensical for ML algorithm. It should be the highest possible `int` instead of `0`.