The haddock for the function streamDecodeUtf8With says that it takes a byte stream, but in truth it takes a byte array!

[kindaro]

See:

λ import Data.Text.Encoding
λ import Data.Text.Encoding.Error
λ import Data.ByteString.Lazy (pack)
λ :type pack
pack ∷ [GHC.Word.Word8] → Data.ByteString.Lazy.Internal.ByteString
λ :type streamDecodeUtf8With lenientDecode . pack

<interactive>:1:38: error:
    • Couldn't match type ‘Data.ByteString.Lazy.Internal.ByteString’
                     with ‘bytestring-0.11.4.0:Data.ByteString.Internal.Type.ByteString’
      Expected: [GHC.Word.Word8]
                → bytestring-0.11.4.0:Data.ByteString.Internal.Type.ByteString
        Actual: [GHC.Word.Word8]
                → Data.ByteString.Lazy.Internal.ByteString
      NB: ‘bytestring-0.11.4.0:Data.ByteString.Internal.Type.ByteString’
            is defined in ‘Data.ByteString.Internal.Type’
          ‘Data.ByteString.Lazy.Internal.ByteString’
            is defined in ‘Data.ByteString.Lazy.Internal’
    • In the second argument of ‘(.)’, namely ‘pack’
      In the expression: streamDecodeUtf8With lenientDecode . pack

This is what the haddock says:

Decode, in a stream oriented way, a lazy ByteString containing UTF-8 encoded text.

As I understand, my code should work — the pack imported from Data.ByteString.Lazy should compose with streamDecodeUtf8With lenientDecode. It is confusing that we have the same name ByteString for two different things — byte stream and byte array. Maybe I am terribly confused. Please help me understand this situation.

[Bodigrim]

(bytestring offers type synonyms StrictByteString and LazyByteString, let me use them to reduce confusion)

It's just that documentation for streamDecodeUtf8With is wrong since inception. It takes a StrictByteString indeed. "Stream-oriented" means that you are supposed to pass further StrictByteStrings one by one to Decoding.

Could you possibly raise a PR to fix haddocks?