[haskell-llvm] Crash in BrainF: interface change to memset intrinsic

[Henning Thielemann]

On Tue, 12 Oct 2010, Henning Thielemann wrote:

> The crash in BrainF was caused by the memset intrinsic, that got a new
> parameter isvolatile in LLVM-2.8. Two questions:
>
> What is the right Haskell type for i1 arguments?
>
> How can we avoid silently breaking of llvm client code? BrainF locally
> declares the memset intrinsic and the same way other LLVM applications
> may do. Should we provide a module that declares the LLVM intrinsics?
> LLVM notes mention an Intrinsic::getDeclaration method, but I do not
> know, whether this would help the Haskell wrapper.

Any ideas on these question?

If these issues are fixed, we could upload a new version of the llvm 
interface package to Hackage.