A few Questions ...

[Peter Tanski]

On Oct 26, 2006, at 3:37 AM, Simon Marlow wrote:

> Peter Tanski wrote:
>> Do you keep GranSim and PAR in the main GHC code for
>> compatibility or
>> are they leftovers that haven't been cleaned up yet?
>
> They are in the tree pending a future update of the PAR/GRAN code; but
> it now looks like the GPH developers would prefer that the old code be
> removed from the tree entirely, so at some point we'll do that.  It's
> not a big deal, actually I'd prefer if it was left in for now because
> I'm making changes to the GC and if I can avoid conflicts it'll  
> save me
> time.

I'll gladly stay away from it then.  Less work to do :)

>> For that
>> matter, do you want to keep .NET code around for future reference or
>> would someone (say, me) be able to clean that up as well?
>
> You mean the ILX stuff, or foreign import dotnet?  ILX I think is  
> dead,
> I'm not so sure about foreign import dotnet.

I actually meant both.  In a patch from 04 October (this year), with  
the name beginning "Remove ILX from the GHC altogether ..." Simon  
Peyton-Jones removed ILX stuff from the Haskell-code in the compiler:
     M ./compiler/main/CodeOutput.lhs -27
     M ./compiler/main/DriverPhases.hs -4
     M ./compiler/main/DynFlags.hs -4 +1
     M ./compiler/stgSyn/CoreToStg.lhs -13
     M ./compiler/typecheck/TcForeign.lhs -3 +1
In a message you sent on 18 May 2005 to the FFI mailing list (at  
http://www.mail-archive.com/ffi@haskell.org/msg01814.html), in  
reference to the FFI syntax "foreign type..." you mentioned the  
experiment to have GHC generate code for .NET.  (I am not sure what  
you meant--maybe bytecode, so the .NET-FFI is still live?)

Cheers,
Pete