[Pkg-haskell-maintainers] Bug#661192: Haddock is stricter than GHC about newtypes in FFI code

[Joachim Breitner]

Package: ghc-haddock
Version: 7.0.4-8
Severity: important

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

this bug makes haskell-glut FTBFS and is forwarded upstream as
http://trac.haskell.org/haddock/ticket/194, this is just for reference:

Hi,

this is Haddock version 2.10.0 as shipped with GHC 7.4.1. The package
GLUT-2.1.2.1 contains FFI code that GHC warns about, but haddock fails:

	[ 9 of 21] Compiling Graphics.UI.GLUT.Window ( Graphics/UI/GLUT/Window.hs, dist-ghc/build/Graphics/UI/GLUT/Window.o )

	Graphics/UI/GLUT/Window.hs:113:1:
	    Warning: newtype `Window' is used in an FFI declaration,
		     but its constructor is not in scope.
		     This will become an error in GHC 7.6.1.
	    When checking declaration:
	      foreign import ccall unsafe ""static glutCreateWindow"" glutCreateWindow
		:: CString -> IO Window
vs.
	Running Haddock for GLUT-2.1.2.1...
	Running hscolour for GLUT-2.1.2.1...
	Preprocessing library GLUT-2.1.2.1...
	Warning: The documentation for the following packages are not installed. No
	links will be generated to these packages: rts-1.0
	Preprocessing library GLUT-2.1.2.1...
	[..]
	Graphics/UI/GLUT/Window.hs:113:1:
	    Unacceptable result type in foreign declaration: IO Window
	    When checking declaration:
	      foreign import ccall unsafe ""static glutCreateWindow"" glutCreateWindow
		:: CString -> IO Window

But something is fishy, the [..] part contains a few of the Warning: newtype
`CInt' is used in an FFI declaration, warnings, so maybe something else is
amiss here.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEARECAAYFAk9ICkoACgkQ9ijrk0dDIGyfRACdGhvYLrpVJ0HV/qBJCQC6etHz
iegAnRnzowAUJ3516VnDj6vZk0Hxk2W7
=q7t+
-----END PGP SIGNATURE-----