RE: Re[2]: ANNOUNCE: GHC version 6.6.1

2007-04-30 Thread Simon Peyton-Jones
| Making it possible to use Haskell in mixed language projects with C++
| and Java is obviously a good thing, but it's not really a scalable
| solution to distribute them all together. We should look at how to
| improve cabal+ghc to make it easier to use them as a component of a
| larger system. One example of this would be not requiring that ghc is
| used to do a final link, we should be able to make static libs and then
| link them using ordinary gcc.

Indeed.  I don't think there is any reason in principle why one should not use 
GHC and C++ together, and there is certainly some customer demand for this to 
be smooth and easy.  But it's not at the moment, and

we lack a motivated colleague to help make it happen

It needs someone who is familiar with C++ and its various strange naming, 
calling, and linking conventions; and who is willing to learn a bit about GHC's 
strange naming, calling, and linking conventions (with help from us).

We are ready to spend time explaining the GHC end.  But we really need someone 
to figure out what happens in the GHC/C++ link step, and make it all work 
smoothly.

Please do volunteer!

Simon
___
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users


Re: Re[2]: ANNOUNCE: GHC version 6.6.1

2007-04-29 Thread Duncan Coutts
On Sun, 2007-04-29 at 10:38 +0400, Bulat Ziganshin wrote:
 Hello Albert,
 
 Sunday, April 29, 2007, 2:51:24 AM, you wrote:
 
  Is it just me who thinks this is a silly idea? Why should GHC include a
  C++ compiler?
 
  .NET literates, will benefit from the many libraries available in .NET.
  Can we also include a .NET runtime, a .NET documentation suite, all .NET
 
 you are lame.

Hia Bulat,

It's really much better to criticise ideas rather than people
personally, though I do understand that Albert was making a joke at your
expense. In that situation its much better to try and ignore it rather
than flaming. We do want to keep the various Haskell mailing lists
civil :-)


 java/c# libs can't be used with current ghc, so 99% of
 libs we may need are written in c/c++. making porting these lubs as
 hard as possible and then heroically rewrite them in pure haskell is
 one way, good for PhD and other pseudo-scientific activity. building
 bridges to the world of existing software is the way to the real
 haskell usage in big projects

Making it possible to use Haskell in mixed language projects with C++
and Java is obviously a good thing, but it's not really a scalable
solution to distribute them all together. We should look at how to
improve cabal+ghc to make it easier to use them as a component of a
larger system. One example of this would be not requiring that ghc is
used to do a final link, we should be able to make static libs and then
link them using ordinary gcc.

Duncan

___
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users


Re: Re[2]: ANNOUNCE: GHC version 6.6.1

2007-04-28 Thread Neil Mitchell

Hi


 Is it just me who thinks this is a silly idea? Why should GHC include a

alternative way may be inclusion of instructions how to add c++ compiler
to ghc installation - it's not trivial because one need to know which
gcc version should be used and which files copied


Since I know that Bulat is the one always asking for a size reduction
in the installer, why not have a minimal installer (without c++), and
an extra-utils installer (with c++). Once GHC no longer requires C on
Windows, we can move C compilation from minimal to extra.

Of course, this requires installer hacking.

Thanks

Neil
___
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users


Re: Re[2]: ANNOUNCE: GHC version 6.6.1

2007-04-26 Thread shelarcy
Hello Bulat,

I think that builders/distributers use make binary-dist to
make binary distribution normally.

http://hackage.haskell.org/trac/ghc/wiki/Building/Using#BootstrappingGHC


make binary-dist calls prep-bin-dist-mingw to make Windows friendly
distribution now.

http://www.haskell.org/pipermail/cvs-ghc/2007-April/034978.html

And before this change, we use prep-bin-dist-mingw explicitly to
make Windows friendly distro.

So I sent follow previous message. Because leaving incorrect
prep-bin-dist-mingw lack some files that is important C++
for compilation.

http://www.haskell.org/pipermail/glasgow-haskell-bugs/2007-March/008738.html


On Fri, 27 Apr 2007 01:29:26 +0900, Bulat Ziganshin [EMAIL PROTECTED] wrote:
 On Thu, 26 Apr 2007 22:07:47 +0900, Bulat Ziganshin
 [EMAIL PROTECTED] wrote:
 can you please include in win32 distro c++ compiler, as it was done
 before and as it requested by trac ticket
 http://hackage.haskell.org/trac/ghc/ticket/1024 ?

 This problem is already fixed.

 http://www.haskell.org/pipermail/cvs-ghc/2007-April/034976.html

 the problem i said can't be fixed by anyone except than *builder* of
 windows distro. the problem is lack of files of c++ compiler and libraries in
 distro, not the ghc behavior

OK, builders is important role to make windows distro.
But from above reasons, I think prep-bin-dist-mingw (this change
make binary-dist behavior, not to change ghc behavior) also important
role to make windows distribution.


Best Regards,

-- 
shelarcy shelarcyhotmail.co.jp
http://page.freett.com/shelarcy/
___
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users