I *think* the patch below works correctly: sdl-config gets the -march
flag and SDL builds fine. Perhaps there's a better way to get the
-march flag into both CFLAGS and SDL_FLAGS w/o dragging along more
cruft, but adding it to BASE_CFLAGS seemed the best bet to my eyes.

I did some digging, and I think I've found a better solution. By
including some additional .m4 macros in the acinclude directory
(http://www.gnu.org/software/autoconf-archive/ax_check_compiler_flags.html,
http://www.gnu.org/software/autoconf-archive/ax_gcc_archflag.html and
http://www.gnu.org/software/autoconf-archive/ax_gcc_x86_cpuid.html)
then using the patch below (same as previous patch + one line change
to configure.in), SDL will build correctly out of the box.

The arch can then be overridden by --with-gcc-arch=<arch> (most will
probably want to build with i686 there for general distribution) or
--without-gcc-arch, to disable setting the arch all together.

On my Atom netbook, I get -march=PentiumPro by default, which works
great, as does --with-gcc-arch=i686.

Fixed building SDL test programs

That's actually okay, since gcc has the builtin atomic intrinsics. :)

This allows all SDL code to take advantage of the atomic intrinsics on Windows, but doesn't cause applications just including SDL.h to pull in windows.h

Changed the CAS return value to bool, so it's efficient with OSAtomicCompareAndSwap32Barrier()
Added an atomic test adapted from code by Michael Davidsaver

Implemented SDL_SetWindowTitle(), which turned out to be fairly complex