Re: Re: Does gcc support compiling for windows x86-64?

[NightStrike <nightstrike@gmail.com>]

On Nov 23, 2007 6:31 PM, Howard Chu <hyc@highlandsun.com> wrote:
> I've downloaded a couple of the binary tarballs from the mingw-w64 project
> page. Had a lot of trouble getting usable code out of them. I finally figured
> out that I had to compile without any optimization to get anything to run.
> It's not clear whether this is a problem specific to the win64 support, or if
> it's gcc 4.3.0's immaturity. I was also frustrated by the lack of a working
> debugger, giving up half way through building gdb. I'm thinking it may be
> quicker to write a tool that converts the gcc stabs stuff to rudimentary PDB
> format to provide function and variable names for WinDbg.

GDB support isn't done yet.  Optimization issues will be more
difficult to debug, also.  Please remember that the project has only
two or three people on it, only one of which really understands this
stuff.  If you'd like to help out, we would be very grateful.

Also, please try a more current version of the project.  As soon as I
get through with some very pressing family emergencies, I will be
uploading new versions of everything.  Anything that you can test and
find problems with will be very helpful.

> The cross-compiler runs pretty well hosted on Linux but for some reason some
> of the configure scripts I ran were accessing my Linux header files and so
> detecting features they shouldn't have. My only other choice was to run under
> Cygwin on the Windows side, but shell scripts run about 100 times slower
> there, making configure/libtool/etc unbearable. (Normally I would use MSYS but
> the last one I tried just crashes immediately on Win64.) And it looks like
> current bash on cygwin doesn't handle case/esac constructs correctly, so e.g.
> the configure script for BerkeleyDB 4.6.21 fails there.

I made a binary release that runs on i686-pc-mingw32.  That may allow
you to step away from cygwin.  Regarding the linux release, however,
can you describe in more detail what you are seeing in terms of
accessing the linux header files?  It's entirely possible that I am
building the sysroot incorrectly (hey, we all make mistakes :) ).  If
you could provide more feedback, I'd love to try to fix it.  You can
email me directly, post to the mingw-w64 mailing list, file a bug
report on the mingw-w64 project, etc etc.