WinDll.Lib.NativeMapping_Debug

Contains the list of native types and their mapping to their equivalent FFI types

Documentation

class LLD m a b c | b -> c where

Typeclase to allow Left LoaD transform. It is basically to allow a transformation to take place at the last argument/return type of the function. This is because most of the functions are in IO.

Methods

lld :: m a b -> m a c

Instances

LLD (->) a b (IO b)

class FFIType phi ix where

A class that manages the conversion between the normal and type supported by ffi. Minimal implementation requires atleast one of the pair toNativetoFFI and fromNativefromFFI. The implementation will almost always call fromNative and toNative because all exported functions are in IO since they all might have side-effects. The only exception to this is for the defaults provided in this module.

Methods

toFFI :: Stackable st => st -> phi -> ix

fromFFI :: Stackable st => st -> ix -> phi

fromList :: Stackable st => st -> CInt -> ix -> IO phi

fromNative :: Stackable st => st -> ix -> IO phi

toNative :: Stackable st => st -> phi -> IO ix

freeFFI :: Stackable st => st -> phi -> ix -> IO ()

Instances

FFIType Bool Bool

Numeral values are all also already FFI values, If I've read the documentation correctly Due to GHC matching only the instance heads this instance can't unfortunately be used. (Booo bad GHC) instance Num a => FFIType a a where toFFI = const id fromFFI = const id

Booleans are by default already an FFI value

FFIType Bool Int8

Convert booleans to Int8 to save space for use when using the ccall or stdcall conventions

FFIType Bool Word8

Convert booleans to Word8 to save space for use when using the ccall or stdcall conventions

FFIType Bool CInt

Convert booleans to Cints for use when using the ccall or stdcall conventions

FFIType Char CChar

FFIType Char CWchar

FFIType Float CFloat

Instance for unboxed integers, which are first boxed then returned instance FFIType FastInt CInt where toFFI = toFFI . iBox fromFFI = iUnbox . fromFFI

Fix float instances

FFIType Int CInt

Fix integers from the machine dependend values to fixed 32bit values

(Num a, Integral a) => FFIType Integer a

Convertion instance for Integer types to CLLongs (long long)

FFIType Rational CDouble

FFIType String CWString

I decided to use a CAString because on windows this gives me a constant 16 value

FFIType () ()

Default values needed to satisfy .NET marshaller when having unused structures. class FFIType phi ix => FFIDefaults phi ix where class Default phi where nDefault :: phi

Wrapper functions for dealing with FunPtrs wrapFn :: (FFIType (IO a) ca, FFIType b (IO cb)) => (a -> IO b) -> (ca -> IO cb) wrapFn fn = fromFFI >=> fn >=> toFFI

Dedicated instance for ()

FFIType CWString CWString

Another simple identity instance, I really need to get that overlapping instances looked at.

FFIType CInt CInt

Another simple identity instance, I really need to get that overlapping instances looked at.

FFIType CLLong CLLong

Another simple identity instance, I really need to get that overlapping instances looked at.

FFIType CDouble CDouble

Another simple identity instance, I really need to get that overlapping instances looked at.

FFIType CWchar CWchar

Another simple identity instance, I really need to get that overlapping instances looked at.

FFIType FastString CWString

Convert between FastString and CWString

Storable a => FFIType a (Ptr a)

Any class implementing Storable has implemented enough to be considered a FFIType

Storable a => FFIType [a] (Ptr a)

Cover lists to array convertion IF the type is also an FFI type

(Storable a, FFIType b a) => FFIType [b] (Ptr a)

Intermediate conversion instance for storing values of arrays

FFIType (StablePtr a) (StablePtr a)

A StorablePtr instance

FFIType (FunPtr a) (FunPtr a)

A FunPtr instance

(Functor f, FFIType a b) => FFIType (f a) (f b)

Instance for Functor classes

(FFIType a b, Storable b) => FFIType [a] (Ptr CInt -> IO (Ptr b))

One way instance for returning lists as the result of a function call. We assume to have an int* as an argument and then fill that in with the length

(FFIType a b, FFIType c d) => FFIType (a -> c) (b -> d)

Tranform functions to and from the correct types TODO: Update this variant to use the impure variants, This will be an issue..

(FFIType a b, FFIType c d, Storable b, Storable d) => FFIType (a, c) (Tuple2Ptr b d)

(FFIType a b, FFIType c d) => FFIType (a, c) (Tuple2 b d)

(FFIType a b, FFIType c d, FFIType e f, Storable b, Storable d, Storable f) => FFIType (a, c, e) (Tuple3Ptr b d f)

(FFIType a b, FFIType c d, FFIType e f) => FFIType (a, c, e) (Tuple3 b d f)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, Storable b, Storable d, Storable f, Storable h) => FFIType (a, c, e, g) (Tuple4Ptr b d f h)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h) => FFIType (a, c, e, g) (Tuple4 b d f h)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, Storable j, Storable b, Storable d, Storable f, Storable h) => FFIType (a, c, e, g, i) (Tuple5Ptr b d f h j)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j) => FFIType (a, c, e, g, i) (Tuple5 b d f h j)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l, Storable j, Storable l, Storable b, Storable d, Storable f, Storable h) => FFIType (a, c, e, g, i, k) (Tuple6Ptr b d f h j l)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l) => FFIType (a, c, e, g, i, k) (Tuple6 b d f h j l)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l, FFIType m n, Storable n, Storable j, Storable l, Storable b, Storable d, Storable f, Storable h) => FFIType (a, c, e, g, i, k, m) (Tuple7Ptr b d f h j l n)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l, FFIType m n) => FFIType (a, c, e, g, i, k, m) (Tuple7 b d f h j l n)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l, FFIType m n, FFIType o p, Storable n, Storable p, Storable j, Storable l, Storable b, Storable d, Storable f, Storable h) => FFIType (a, c, e, g, i, k, m, o) (Tuple8Ptr b d f h j l n p)

(FFIType a b, FFIType c d, FFIType e f, FFIType g h, FFIType i j, FFIType k l, FFIType m n, FFIType o p) => FFIType (a, c, e, g, i, k, m, o) (Tuple8 b d f h j l n p)

translate :: Defs -> Type -> Type

Tuples are not FFI compatible, As such i'll translate them to a build in tuple datatype . This function translates the embedded types of a Ty to the correct forms using the function translate' (see below)

translatePartial :: Defs -> Type -> Type

Translate everything but applied types. e.g. Foo Token -> FooPtr Token And lists, since lists are implicitly an applied type: e.g [Token] -->> [] Token -->> Ptr Token

isSimpleType :: Type -> Bool

Check to see if the next type is a Simple type. e.g. A TyVar or TyCon

translatePrimitive :: Defs -> Type -> Type

Contrary to translate translatePrimitive will only transform the defined primitive types in the convList below. This is because while a transformed signature should only be partially transformed till the first application (Since that'll be the main pointer) we should pre-transform the primitive types into their well known static forms.

translateP :: Defs -> String -> String

Helper function to define translatePrimitive. It attemps to lookup the type in convList but in the case where it's not found the search query is returned.

translatePForm :: Defs -> Type -> Type

Translate Partial Form, This is basically translatePrimitive . translatePartial

translate' :: Defs -> String -> String

Look up the FFI type representation of the given type. Moreover when the type is not found it is assumed to be a new structure and it is assumed to be a pointer value.

trim :: String -> String

Remove all spaces from a sentence

split :: Eq a => [a] -> a -> [[a]]

A function to split a list of elements by the given seperator