The Revenge of Finalizers

[George Russell]

Simon Marlow wrote:
> 
> > > I don't know how to achieve the same goal with
> > > atomicModifyIORef.
> >
> > I do.  To modify ioRef1 and ioRef2 "simultaneously", write
> >
> > atomicModifyIORef ioRef1 (\ contents1 -> unsafePerformIO
> > ioRef2 (\ contents2 -> blah blah))
> >
> > The actual modification will take place when the result or
> > contents of ioRef1 or ioRef2 get evaluated.
> 
> I don't understand how this works.  Unless I'm missing something, you'll
> see the contents of ioRef1 at the point at which the first
> atomicModifyIORef takes place, but the contents of ioRef2 from the time
> at which the unsafePerformIO is evaluated, which is some unspecified
> time later.

Perhaps we're talking about different sorts of atomicModifyIORef?  I am now
thinking of atomicModifyIORef as having the type

IORef a -> (a -> (a,b)) -> IO b

Thus when it is called the effect is to construct (1) a thunk of type (a,b) containing
the application of the function to the thunk previously in the IORef; (2) a thunk containing
fst $  the thunk from (1) of type (a,b), which is put in the IORef; (3) a thunk
containing snd $ the thunk from (1), to be returned.  Subsequent evaluation of the result of 
atomicModifyIORef will force the thunks in (1) and (3) to be evaluated, meaning the
unsafePerformIO gets evaluated.