[Haskell-cafe] Docs on the current and future constraint solver?

[Corey O'Connor]

On Thu, Jul 15, 2010 at 12:57 AM, Simon Peyton-Jones
<simonpj at microsoft.com> wrote:
> Corey
>
> | On 14 July 2010 18:39, Corey O'Connor <coreyoconnor at gmail.com> wrote:
> | > I believe I have run headlong into issue #3064 in ghc
> | > (http://hackage.haskell.org/trac/ghc/ticket/3064). All I think I know
> | > is this:
> | > * this is a performance issue with the system used to solve type
> | constraints.
> | > * the solver is undergoing an overhaul to resolve performance issues
> | > in addition to other issues.
> | > * An efficient constraint solver is difficult. NP-Complete in the general
> | case?
>
> It would be very helpful to have your code as a test case. Can you boil out a concrete program that demonstrates very poor performance of the type checker, and submit a Trac report?  That way we'll test the new type inference engine against it.  Lacking the example, we won't.
>
> Which is isn't a promise that we'll solve your problem -- but it's much easier to solve if we have a concrete example.  Many thanks!

Understood. I'll work on getting  a nice, self-contained example.

Cheers,
Corey O'Connor