Re: Fwd: compiling ImpliciCad

[Ben Gamari]

Vincent Rioux <vincen...@gmail.com> writes:

> ---------- Forwarded message ----------
> From: Vincent Rioux <vincen...@gmail.com>
> Date: 2013/6/6
> Subject: compiling ImpliciCad
> To: ch...@colah.ca
>
>
> Hello,
>
Hi Vincent!

> Thanks for this beautiful project - i already started to use extopenscad
> (it's nice!) but i am curious to try the Haskell API.
>
> I am trying to compile ImplicitCad from dev github repo, on osx 10.8 with
> brew ghc7.6.3
> I keep getting problems with compiling the Show deriving clause in
> Definitions.hs
> I am a complete beginner in Haskell so i have hard time trying to
> understand what could be wrong...
>
These `Show (a -> b)` instances have given us persistent problems. As of
`2e9b23acd5015fe4f7905773bdfef128803d35f4` we use the instance provided
by the `NumInstances` package. It seems that Conal decided (correctly,
IMHO) to drop this instance in `NumInstances-1.3`.

I'm frankly a little weary of `NumInstances` in general and the `Show`
instance was a particularly heinous abuse. I'd be happy if we dropped
the dependency `NumInstances` dependency outright. Either way, it seems
we'll need to either again provide a `Show (a -> b)` instance of our own
or alternatively use our own Show-esque typeclass. I would prefer the
latter as `Show` is too often abused, IMHO.

Cheers,

- Ben