> > Can someone post a summary of what's yet to be done to build 64 bit binaries
> > on win64?
> > Is there anyone who could complete whatever remaining tasks on a consulting/task
> > basis for $$?
> It's pretty far off from being possible. There are patches pending on
> the binutils list to add x64 PE support to the assembler and linker, but
> they have yet to be reviewed and based on the rough formatting and large
> size I suspect it will take a few rounds of back and forth before they
> are committed.
> And that's only the assembler/linker, someone needs to port gcc as
> well. AFAIK there has been no mention at all of this on the gcc list so
> far, but I could be wrong. Getting things into gcc is traditionally a
> much longer process, since gcc follows a staged development process and
> new features are only supposed to be added during stage 1. This means
> the current mainline (what will become 4.2 soon) will have to be
> branched first and the nearest release anyone could expect a new port
> would be in version 4.3, which could be a year off or more. Of course
> someone could always do the port in a private branch or as unofficial
> patches against HEAD or some existing release, but this is very
> suboptimal compared to actually getting them contributed according to
> the official plan so that they will be in official releases.
> On top of that is the fact that there are still enough issues with the
> 4.x series on windows that mingw and cygwin both still only offer
> versions from the 3.4 series due to stability. There seems to be a lack
> of momentum/developer interest in windows fixes compared to other
> platforms, so it just takes a little longer.
> Regarding paying someone to do the work, that is always an option. If
> you wanted the absolute fastest path to getting x64 PE support in
> official FSF then I would say contact Codesourcery, as they employ a
> number of gcc developers they also use/support mingw for their
Thanks for the info. I've contacted Codesourcery.
I wonder if there's enough interest from the community to split the
cost of such an
The cost is likely in the many k$'s.