What does the Storable typeclass do in Haskell

Question

I have been reading some documentation on Haskell's C FFI. And I've just encountered a typeclass called Storable, which I don't understand very well.

Are instances of this typeclass, those types are supposed to have a "pointer" to them while interfacing with the C code?

Also what do the individual functions sizeOf, alignment, peek, poke do? It seems peek and poke are used to read data from or write data to a place in memory pointed to by Ptr a. Is this right?

But I don't know what sizeOf and alignment mean at all. Can someone give examples to clarify their use?

Answer 1

Haskell stores values in memory in a way that is very incompatible to C. As a result, it is not possible to call a C function from haskell and pass to it haskell values directly. Instead, you have to create a copy of the value, but not an exact copy, but rather in a format that is understood by C. That's what Storable does. So it essentially provides way to serialize haskell values to a C friendly format (e.g. think C structs). It also supports the opposite operation, it can deserialize values. This is useful when a C function is called from haskell and returns a complex (i.e. non primitive) value.

The serialization/deserialization happens with the help of poke/peek. sizeOf returns the byte size of the C representation of the value. Note that this mechanism only works for values that have a C representation of fixed size (e.g. structs). It does not support things like C strings, they are treated differently. As for alignment, it used to ensure that memory allocations done in the haskell land satisfy the platform's alignment requirements.