
I'm trying, and failing, to understand why ForeignPtr is defined the way it is. It all seems rather complicated and redundant. I was thinking we might want to switch to something simpler and more explicit, like this: data ForeignPtr a = ForeignPtr { fptr :: !(Ptr a) -- What we point to froot :: MutableByteArray# RealWorld fwk :: MutVar# RealWorld Any -- The finalizers } froot represents the object the ForeignPtr points into. When created by one of the malloc variants, this will actually be the allocated byte array. Otherwise, it will be a 0-length array. fwk should be seen as having type fwk :: MutVar# RealWorld (Weak# (MutableByteArray# RealWorld)) but we can't express that directly right now. fwk is keyed on froot, and in the malloc case also points to it, to prevent it from being freed prematurely. It seems that this representation requires a couple extra primitive features from Weak#. I think that's okay: Weak# already has some primitive features designed to support ForeignPtr. 1. An operation to add a Haskell finalizer to a Weak#, similar to the one that adds a C finalizer. 2. A way to deal with mixed finalizers: either make the finalizer-adding primops report distinguishable failure on mixing or arrange to run the C finalizers after any Haskell finalizers (since C finalizers are much more likely to actually invalidate the pointer). 3. A documented guarantee about the order in which the finalizers attached to a particular Weak# run, compatible with the ForeignPtr documentation.

Interesting.
Can you summarise how ForeignPtr is defined right now? And point to any documentation about the specification of ForeignPtr, or Notes about its implementation?
So that we can compare with what your proposal.
What’s the connection with weak pointers?
Simon
From: ghc-devs

Wouldn't this mean that unpacking a ForeignPtr is one word longer than it
is now? That would have a big impact on ByteString.
What do you think is redundant in the current representation?
Cheers
Simon
On Wed, 5 Sep 2018 at 00:23, David Feuer
I'm trying, and failing, to understand why ForeignPtr is defined the way it is. It all seems rather complicated and redundant. I was thinking we might want to switch to something simpler and more explicit, like this:
data ForeignPtr a = ForeignPtr { fptr :: !(Ptr a) -- What we point to froot :: MutableByteArray# RealWorld fwk :: MutVar# RealWorld Any -- The finalizers }
froot represents the object the ForeignPtr points into. When created by one of the malloc variants, this will actually be the allocated byte array. Otherwise, it will be a 0-length array.
fwk should be seen as having type
fwk :: MutVar# RealWorld (Weak# (MutableByteArray# RealWorld))
but we can't express that directly right now. fwk is keyed on froot, and in the malloc case also points to it, to prevent it from being freed prematurely.
It seems that this representation requires a couple extra primitive features from Weak#. I think that's okay: Weak# already has some primitive features designed to support ForeignPtr.
1. An operation to add a Haskell finalizer to a Weak#, similar to the one that adds a C finalizer.
2. A way to deal with mixed finalizers: either make the finalizer-adding primops report distinguishable failure on mixing or arrange to run the C finalizers after any Haskell finalizers (since C finalizers are much more likely to actually invalidate the pointer).
3. A documented guarantee about the order in which the finalizers attached to a particular Weak# run, compatible with the ForeignPtr documentation.

I realized that the current representation is kind of balanced on a knife
edge, efficiency-wise. There's probably no way to make one thing faster
without slowing down another: there are just too many things a ForeignPtr
could be trying to represent.
On Thu, Sep 6, 2018, 3:15 AM Simon Marlow
Wouldn't this mean that unpacking a ForeignPtr is one word longer than it is now? That would have a big impact on ByteString.
What do you think is redundant in the current representation?
Cheers Simon
On Wed, 5 Sep 2018 at 00:23, David Feuer
wrote: I'm trying, and failing, to understand why ForeignPtr is defined the way it is. It all seems rather complicated and redundant. I was thinking we might want to switch to something simpler and more explicit, like this:
data ForeignPtr a = ForeignPtr { fptr :: !(Ptr a) -- What we point to froot :: MutableByteArray# RealWorld fwk :: MutVar# RealWorld Any -- The finalizers }
froot represents the object the ForeignPtr points into. When created by one of the malloc variants, this will actually be the allocated byte array. Otherwise, it will be a 0-length array.
fwk should be seen as having type
fwk :: MutVar# RealWorld (Weak# (MutableByteArray# RealWorld))
but we can't express that directly right now. fwk is keyed on froot, and in the malloc case also points to it, to prevent it from being freed prematurely.
It seems that this representation requires a couple extra primitive features from Weak#. I think that's okay: Weak# already has some primitive features designed to support ForeignPtr.
1. An operation to add a Haskell finalizer to a Weak#, similar to the one that adds a C finalizer.
2. A way to deal with mixed finalizers: either make the finalizer-adding primops report distinguishable failure on mixing or arrange to run the C finalizers after any Haskell finalizers (since C finalizers are much more likely to actually invalidate the pointer).
3. A documented guarantee about the order in which the finalizers attached to a particular Weak# run, compatible with the ForeignPtr documentation.
participants (3)
-
David Feuer
-
Simon Marlow
-
Simon Peyton Jones