[Haskell-cafe] haddock-2.3.0 literate comments discarded from .lhs input

[Alistair Bayley]

I have this test case for Haddock (2.3.0):

--------------------------------------------------

|
Module      :  Test.Haddock
Copyright   :  (c) 2009 Alistair Bayley
License     :  BSD-style
Maintainer  :  alistair at abayley.org
Stability   :  stable
Portability :  portable

Test case for Haddock.


> module Test.Haddock
>   (
>   -- $named_block
>   Fail(..)
>   )
> where
> data Fail = Fail | Succeed

$named_block

This is some hadock documentation.

--------------------------------------------------

This fails with:

[1 of 1] Compiling Test.Fail        ( Test\Fail.hs, Test\Fail.o )

Test\Fail.hs:11:26:
    Can't make a derived instance of `Typeable Fail'
      (You need -XDeriveDataTypeable to derive an instance for this class)
    In the data type declaration for `Fail'

If I manually unlit, then Haddock is happy:

--------------------------------------------------

-- |
-- Module      :  Test.Haddock2
-- Copyright   :  (c) 2009 Alistair Bayley
-- License     :  BSD-style
-- Maintainer  :  alistair at abayley.org
-- Stability   :  stable
-- Portability :  portable
-- 
-- Test case for Haddock.

module Test.Haddock2
  (
  -- $named_block
  Fail(..)
  )
where
data Fail = Fail | Succeed

-- $named_block
-- 
-- This is some hadock documentation.

--------------------------------------------------


so it looks as though it's discarding the literate comments. Is this
intended? I was under the impression that because it used the ghc
parser, it could now properly handle .lhs input.

Ona related note, we have a nice unlitter in cabal that would preserve
these comments before invoking haddock. Is this still used with
haddock2, or does it now assume haddock2 will do the right thing?

Alistair