Start of better error reporting in FFI

[Brian Rice]

I've applied some clean-ups, and bootstrapped a new VM and image set  
which are now uploaded to the alpha area. Probably more fixes are  
needed, but nothing huge breaks.

I also added a stub for the Windows branch of the VM sources, to  
avoid breakage on that platform.

On Jul 6, 2005, at 3:17 PM, Slava Pestov wrote:

> Hi all,
>
> I was debugging an FFI problem with Brian on IRC, and noticed that  
> dlerror() was not being called to produce more detailed error  
> messages in the case that symbol lookup failed. This patch adds a  
> new primitive to extprim.slate, and some condition signalling code  
> to extlib.slate.
>
> Slava
>
> <foo.patch.gz>
>

--
-Brian