Some suggested changes to Windows makefiles with respect to experimental toolchain

What should the next step be? Should I just post the text of the
result of diff here? Should I e-mail someone a patchfile, the six
revised files, or both? Is this something to list at
<https://bugs.r-project.org/bugzilla3/>?

We should hold off optimization tweaks until we have a stable build.
Kevin and me are still trying to debug the rcpp longjmp problem which
is more complex than we thought. Adding optimization in the mix at
this point might further complicate things.

Perhaps we can make a branch in the svn (or github mirror) to store
these features so that they don't get lost, and we can easily test and
merge them later on?