Imperative Object Destruction

[Ashley Yakeley]

At 2000-11-13 00:58, Simon Peyton-Jones wrote:

>| C++ provides a convenient mechanism for cleaning up stuff, the 
>| destructor, which is guaranteed to get called when an object 
>| passes out 
>| of scope (or something).
>
>Haskell doesn't but GHC has an extension (finalisers) that does.
>There's a discussion of the issues in our paper about weak pointers
>and finalisers
>	http://research.microsoft.com/~simonpj/#weak

I assume you mean finalisers for ordinary Haskell values after they're no 
longer referenced. That's actually a separate issue, I'm more interested 
in doing finalisation in Haskell's existing monadic imperative model, 
something that shouldn't need any runtime extensions.

-- 
Ashley Yakeley, Seattle WA