I know it's complete heresy to say so, but I use laziness very
little in Haskell, while I probably pay quite a lot for it
in CPU time and memory, because of all those thunks which have to be
stored. However I prefer Haskell's type classes, syntax and
purity to, say, Standard ML. So I wonder whet
"C.Reinke" <[EMAIL PROTECTED]> writes:
> So foldl is indeed tail recursive, but this doesn't help if its
> operator isn't strict because the tail recursion only builds up the
> expression to be evaluated. Making strictness explicit by defining a
> variant of foldl that evaluates its accumulator a
"Julian Assange" <[EMAIL PROTECTED]>:
> ..
> | When used with a 170k input file, makemap suffers from a stack
> | overflow. foldl should be tail recursive. What's the score?
"Simon Peyton-Jones" <[EMAIL PROTECTED]>:
> Consider
> foldl (+) 0 [x1,x2,x3,x4,...
>
> This rewrites to
>
>