how is this integration of the new type-checker into hugs-1.4 going?
is there an estimated time to completion?
currently the new version of our microarchitecture design library only
works with ghc-4. i cant wait to get back to the interactive hugs world.
cheers :-)
- Byron
On Thu, 29 Oct 1998, Mark P Jones wrote:
> Hi Conal,
>
> | I'm still stuck on this problem. Any hope for a bug fix? Any possible
> | work-around? - Conal
>
> I've been looking in to this, and I can confirm that this is a bug in Hugs,
> or at least a "feature" in the old Hugs typechecker. It is related to the
> fact that the old typechecker uses Gofer types under the surface, which
> can't handle class members with overloaded types properly in some
> situations.
>
> The good news is that the new typechecker doesn't have this problem, and
> will accept the example that you gave (i.e., Hugs 1.3c accepts your code,
> modulo a renaming of map (on an array) to amap, because this version of
> Hugs doesn't overload map on arrays).
>
> The bad news is that the new typechecker isn't yet supported in Hugs 1.4,
> with the module system and graphics primitives that you want. Alastair
> is working hard on this, but it's not ready yet.
>
> In the meantime, I've been trying to figure out a patch or a workaround
> to help out ... (with 1.4 June 1998, which I assume is the version you
> are using). I haven't got anything for you yet, but wanted to let you
> know that I'm on the case. Caveat: it may still turn out that the only
> way to get around this is to use the new typechecker.
>
> All the best,
> Mark
>
