Re: Windows Server 2008 64 bit

[Aharon Robbins]

> Date: Thu, 19 Feb 2009 21:41:40 +0200
> From: Eli Zaretskii <address@hidden>
> Subject: Re: Windows Server 2008 64 bit
> To: Aharon Robbins <address@hidden>
> Cc: address@hidden, address@hidden
>
> > Date: Thu, 19 Feb 2009 08:28:13 +0200
> > From: Aharon Robbins <address@hidden>
> > Cc: address@hidden, address@hidden
> > 
> > > > and I also don't see any reason why a 64 bit binary could not be
> > > > compiled for Windows Server 2008 64 bit.
> > >
> > > Because there's no MinGW GCC port to 64-bit Windows yet?
> > 
> > Cygwin is probably an option
>
> I'm not sure Cygwin supports 64-bit Windows builds, but I'm not an
> expert on Cygwin.
>
> > as is attempting to build with the Microsoft compiler.
>
> I'm not even sure it will build, as all signs are this was not
> attempted nor tested for a long time, even as 32-bit build, let alone
> a 64-bit one.

So, this is worth the trouble to do. I will try sometime soon.

> One thing that should be kept in mind is that 64-bit Windows use a
> different model than 64-bit Unices (a long is still 32 bit wide on
> 64-bit Windows).  Thus, code that compiles and works on 64-bit Unix
> machines will not necessarily do so on 64-bit Windows.

Weird. Is the long long a 64 bit value, or only the _int64 (or whatever
it's called)?

Since long is 32 bits, I would think things would compile and run and
work ok; I don't think gawk has any dependencies on sizeof long, nor
does it use long long.

It will make an interesting experiment.

I suspect that once Windows 7 is relatively widely deployed, that
there will be more interest in native 64-bit versions of GNU utilities
compiled for Windows, since it looks like 64 bit will be the way most
people deploy it.

Thanks,

Arnold