[Haskell-cafe] Re: seq does not preclude parametricity (Re: IO is not a monad)

[Robert Dockins]

On Wednesday 24 January 2007 20:20, Stefan Monnier wrote:
> > FYI, don't try to run this in GHC, because it gives the simplifier fits.
>
> You mean it triggers a bug in the inliner?

http://www.haskell.org/ghc/docs/latest/html/users_guide/bugs.html

Third bullet in secion 12.2.1.


I gather that GHC HQ has decided that the problem is pathological enough to 
sweep under the rug.  I can't say I blame them.  Really, the only reason to 
construct custom fixpoint combinators is to show that it can be done :-)  
Using the built-in facilities for recursion is far easier and almost 
certainly results in better code.


>         Stefan