Sieve of Eratosthenes in Haskell

Question

I've implemented the Sieve of Eratosthenes algorithm in Haskell. Please take a look. Any suggestions are greatly appreciated :)

import Data.List

-- a list identical with {seq} but for multiples of {p} which are replaced by {-1}
sieve_once :: [Integer] -> Integer -> [Integer]
sieve_once seq p = map (\x -> if x `mod` p == 0 then -1 else x) seq

-- a list of all prime numbers
sieve :: [Integer]
sieve = unfoldr (\acc -> let p:rest = dropWhile (\x -> x == -1) acc in Just(p, sieve_once rest p)) [2..]

-- a list of all prime numbers not greater than {n}
sieve_until :: Integer -> [Integer]
sieve_until n = takeWhile (\x -> x <= n) sieve

Answer 1

I'd suggest to filter out unneeded elements, instead of replacing them by -1. Keeping and processing all the -1 is unnecessary:

sieve_once :: [Integer] -> Integer -> [Integer]
sieve_once seq p = filter (\x -> x `mod` p /= 0) seq

-- a list of all prime numbers
sieve :: [Integer]
sieve = unfoldr (\(p:rest) -> Just (p, sieve_once rest p)) [2..]