Hi

> > What is happening here ? Does importing Data.Char shadow (in a hidden way) 
> > some
> > timeworn types or methods in the Prelude ? If so, ghc might be still faster
> > than it looks now.
>
> There is absolutely no way this could be correct behavior.

Not true - importing an additional (but apparently unused) module can
introduce new rules, pragmas and instances into scope, which can
effect the optimisation. However, in this case, it looks like a bug -
in either GHC or Joost's testing.

Thanks

Neil
_______________________________________________
Haskell-Cafe mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply via email to