[Haskell-cafe] Smallcheck Depth

[Roman Cheplyaka]

* Omari Norman <omari at smileystation.com> [2014-06-29 18:44:35-0400]
> The docs for smallcheck, and the original paper, say that the depth is the
> depth of nested constructor applications.
> 
> For some reason this behavior has changed in newer versions of Smallcheck
> and maybe someone can shed some light on why.  For instance, I would think
> the depth of True and False is 0, as there are no nested constructors.
>  Sure enough, in Smallcheck 0.2.1,
> 
> (series 0) :: [Bool]
> 
> gives [True, False].  Yet in smallcheck 1.1.1,
> 
> (list 0 series) :: [Bool]
> 
> gives [].

No good reason (for either behavior).

A general principle on how series begin would be good to have. series 0 = []
could be one such principle (but I'm not sure all the instances follow it at the
moment).

Roman
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://www.haskell.org/pipermail/haskell-cafe/attachments/20140629/d1fc4e21/attachment.sig>