[prev in list] [next in list] [prev in thread] [next in thread] 

List:       haskell-cafe
Subject:    Re: [Haskell-cafe] foldlWhile
From:       dpt () bostoncoop ! net (Dylan Thurston)
Date:       2004-11-20 14:57:03
Message-ID: 20041120145703.GB26099 () bostoncoop ! net
[Download RAW message or body]


On Sat, Nov 20, 2004 at 03:48:23PM +0000, Jorge Adriano Aires wrote:
> > On Sat, Nov 20, 2004 at 12:47:58PM +0300, Serge D. Mechveliani wrote:
> > >   foldlWhile :: (a -> b -> a) -> (a -> Bool) -> a -> [b] -> a
> > >   foldlWhile    f                p              a    bs  =
> > >     case
> > >         (bs, p a)
> > >     of
> > >     ([],    _    ) -> a
> > >     (_,     False) -> a
> > >     (b:bs', _    ) -> foldlWhile f p (f a b) bs'
> >
> > Why not just
> >   foldlWhile f p a bs = takeWhile p $ foldl f a bs
> 
> Quite different. The former stops a foldl when the accumulating parameter no 
> longer satisfies p, the later assumes the accumulating parameter of the foldl 
> is a list, and takes the portion of the list that does satisfy p.

Yes, this was a mistake.

> The following is closer to the original, but doesn't work when the whole list 
> is folded (i.e., p always satisfied):  
> foldlWhile f p a = head . dropWhile p . scanl f a

Serge's version returns the last 'a' that satisfies 'p', while yours
returns the first 'a' that does not satisfy 'p'.  This should be an
equivalent version:

  foldlWhile f p a = tail . takeWhile p . scanl f a

But what about the version with Maybe?  There ought to be a concise way
to write that too.

Peace,
	Dylan

["signature.asc" (application/pgp-signature)]

[prev in list] [next in list] [prev in thread] [next in thread]