A more subtle issue is that there's some sort of memory leak that arises when you can't instruct the GC to follow projection functions of data types. I believe the GHC heap representation has a built-in notion of these forwarding closures and the GC follows them when possible, but most VM GCs are just black boxes. It might still be possible to create an artificial field in a representation of your closure that the GC can follow, but it could be tricky.

On Thu, Feb 23, 2012 at 11:59 AM, Tom Murphy <amindfv@gmail.com> wrote:
There are some substantial technical challenges:

http://www.haskell.org/haskellwiki/GHC:FAQ#Why_isn.27t_GHC_available_for_.NET_or_on_the_JVM.3F

Not that it can't be done, but there's nothing ready yet.

Tom (IRC: amindfv)


On 2/23/12, ARJANEN Loïc Jean David <arjanen.loic@gmail.com> wrote:
> Hello,
>
> Does any one knows of an Haskell compiler targeting the JVM ? And of one
> targeting the .Net virtual machine ?
>
> Regards,
> ARJANEN Loïc
>
> _______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe@haskell.org
> http://www.haskell.org/mailman/listinfo/haskell-cafe
>

_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe