Introduce a placeholder for min/max parameters

Building an array of length 3 or more is currently not ideal, solutions are the following:

• fc.array(fc.nat()).filter(data => data.length >= 3) - not optimized, will reject too many arrays
• fc.array(fc.nat(), 3, 100) - but why 100? Why do the user has to specify this value?
• fc.tuple(fc.nat(), fc.nat(), fc.nat(), fc.array(fc.nat())).map(([a,b,c,data]) => [a,b,c,...data]) - really to complicated

The idea is to introduced a placeholder whenever possible to be able to do:

• fc.array(fc.nat(), 3, fc._) - min = 3, max = auto
• fc.array(fc.nat(), fc._, 3) - min = auto, max = 3
• fc.array(fc.nat(), fc._, fc._) - min = auto, max = auto

[dubzzz]

@ethanresnick For the moment the approach I wanted to experiment is the one suggested by @SupernaviX. I did try the other one a while ago but it didn’t fit with the framework. The object based syntax has been spread and used in all the recent arbitraries and works perfectly: typings are easier, no need for ‘overloaded’ functions…

[dubzzz]

@SupernaviX Indeed I think I will move to this kind of signature - with min and max as attributes of an object. The placeholder way - fc.array(fc.nat(), 3, fc._) - seems less idiomatic and does not adapt to all situations. The only benefit I see is that it seems lighter for cases where you need to specify both min and max.

I think I will update the branch I started to implement this feature to the {min,max} way.

In any cases, I need some kind of: I want a min not a max, or the contrary. I prefer not forcing the user into specifying a value for min or max - the framework should do the magical part itself 😉