[nem-en] [proposal] foldr1/foldl1

[Sandro Magi]

I have implementations of map, filter and foldr for IEnumerable in my
FP# library:

http://fpsharp.svn.sourceforge.net/viewvc/fpsharp/trunk/FP.Fn/Fn.cs?view=markup

Perhaps they might be of use.

Sandro

Kamil Skalski wrote:
> Very nice functions indeed. It would be also nice to have a general
> extension methods operating on IEnumerable [T] doing the same things,
> so the functionality won't be limited to lists (which should have
> their own specialized implementation as you provided).
> 
> Will you commit the changes or prepare a patch?
> 
> 2007/7/1, Elifant <elifantu at mail.ru>:
>> Hello all.
>>
>> I propose to add functions like Haskell's foldr1 and foldl1 to standard
>> library. They are very convenient
>> when result has the same type and element type has no "zero" value with
>> respect to the applied operation.
>>
>> Example:
>> Within macro call I have a list of expressions (with type PExpr) "e1,
>> e2, ..., eN" and I need to get expression
>> "((e1 + e2) + ...) + eN". FoldLeft can't be used directly, since <[]>
>> can't be used as a starting value.
>> With FoldLeft: def sum = FoldLeft(exprs.Tail, exprs.Head, (el, er) => <[
>> $el + $er ]>)
>> With FoldLeft1: def sum = FoldLeft1(exprs, (el, er) => <[ $el + $er ]>)
>>
>> Don't know whether these functions must have a different name or just be
>> an overload of FoldLeft and FoldRight.
>> Reference implementation:
>>
>> def FoldLeft1 ['a'] (l: list['a], f: 'a * 'a -> 'a): 'a
>>         FoldLeft(l.Tail, l.Head, f)
>>
>> // This one may be implemented more efficiently, I think...
>> def FoldRight1 ['a'] (l: list['a], f: 'a * 'a -> 'a): 'a
>>         def (rest, last) = DivideLast(l)
>>         FoldRight(rest, last, f)
>>
>>
>> _______________________________________________
>> https://nemerle.org/mailman/listinfo/devel-en
>>
> 
>