Let's get this finished

[Simon Marlow]

> On Mon, 15 Jan 2001, Simon Marlow wrote:
> 
> > 	data UnsafeCString
> > 	withUnsafeCString :: String -> (UnsafeCString -> IO a) -> IO a
> > 
> > where an UnsafeCString is valid only in an argument position of an
> > unsafe foreign import.
> 
> Much of the conversion stuff would have to be done in a different way.
> An additional internal interface to conversions implemented in C
> (with ByteArrays instead of Ptrs), pointer arithmetic moved from
> Haskell to C, and in future hard to mix with Handle I/O done on
> Ptrs.

Hmm.  You're right, unfortunately :(  This is depressing.

> > I still can't think of a good way to do this in general.  Perhaps
> > enhancing the garbage collector so that it could "pin" objects - but
> > you've still got the problem of keeping the lifetime of the 
> ByteArray
> > in sync with the Ptr.
> 
> Lifetime is not a problem:
>     withByteArray :: ByteArray -> (Ptr a -> IO b) -> IO b
> 
> (And now we know that the ByteArray type should be parametrized!)

Well, the reason I said this was a problem was because I was trying to
avoid the overhead of the exception handler or continuation in
withByteArray.

Actually we can do this right now: you allocate a large ByteArray (>= 4k
ish) to make sure it gets allocated in immovable space, and then use
touch# to keep it alive.  You get through a lot of memory this way, but
it should be faster than malloc/free.  I'll do some experiments.

Cheers,
	Simon