splitOn for Data.ByteString/Data.ByteString.Lazy

[ezyang]

As per http://stackoverflow.com/questions/1398322/split-bytestring-on-a-bytestring-instead-of-a-word8-or-char this is a commonly requested piece of functionality. Here is one possible implementation for Data.ByteString:

splitOn :: Str -> Str -> [Str]
splitOn needle haystack0 = go haystack0
  where
    l = length needle
    breaker = breakSubstring needle -- save preprocessing
    go s = let (pre, post) = breaker s
           in pre : if null post
                       then []
                       else go (drop l post)

However, for lazy bytestrings, we will need an implementation of breakSubstring for them.

[sjakobi]

Related to #30.

[Bodigrim]

The common objection is that different people have different expectations for splitOn with regards to keeping delimiters and skipping empty chunks. But given that text already sports splitOn, it seems entirely reasonable to introduce splitOn to bytestring with the same semantics.