On Apr 26, 2012, at 9:34 PM, Tillmann Rendel wrote:
>> 
>>   class Unfoldable t where
>>     unfold :: Unfolder f =>  f a ->  f (t a)
> 
> Just to be sure: That's not a generalization of Data.List.unfoldr, or is it 
> somehow?

Yes, it is. unfoldr is quite specifically tailored to lists, so it doesn't work 
well generically. I did include it in the package, but it does a breadth-first 
search for the first value that has exactly enough positions to store the 
elements ('a's), and there might not be one.

> 
>> Different unfolders provide different ways of generating values, for example:
>>  - Random values
>>  - Enumeration of all values (depth-first or breadth-first)
>>  - Convert from a list
>>  - An implementation of QuickCheck's arbitrary should also be possible 
>> (still working on that)
> 
> Can this be extended to provide a single API that allows testing à la 
> SmallCheck, LazySmallCheck and/or QuickCheck without duplicating properties 
> or instances?


Well, the idea is to unify all ways of unfolding (i.e. all ways of generating 
values). So those parts of the checkers could use the same API, but there's a 
lot more to checking than that.

By the way, I uploaded 0.5.0 a few hours ago, which contains a generic 
arbitrary implementation.

greetings,
--
Sjoerd Visscher
https://github.com/sjoerdvisscher/blog






_______________________________________________
Haskell-Cafe mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply via email to