RE: compile wrapper script with C++ file

[David Byron]

On November, 15, 2009 Ralf wrote: 

> I'm not sure about the best way to proceed here, or what level of
> emulation to add to `compile', but going just one step seems less
> attractive than the current status quo or going even further.

It's not clear to me how to proceed either.  My goal is to work with the
autotools without any wrapper script like cccl, and yet use the native MS
tools.  At the moment, with the pr-msvc-support branch of libtool, I've got
things working without a wrapper with the exception of this
compile/AM_PROG_CXX_C_O change.  So this particular one-more-step seems way
more attractive than the status quo.

I've used a heavily modified version of cccl for awhile but it needed a lot
of help to play with libtool and it seemed to make more sense to ditch it
and dive into the autotools themselves to improve the chances for "proper"
support of native MS tools.  Another approach that might have some chance of
living a fruitful life is if cccl became part of autoconf but that doesn't
seem so likely.

As for what going even further could mean...Teaching autoconf to detect when
it's using a native MS compiler and then how to use /Fo, /Fe sounds like a
great idea.  If that's the kind of patch you've got queued up, I'd love to
try it out.  In the meantime though the patch to compile all by itself or in
addition to an AM_PROG_CXX_C_O macro would help.  Each one makes one less
thing I have to do/document for my build environment.

If there aren't any unfinished patches for native MS support, I may be able
to dive in and create some on my own.  Pointers on where to look and what it
would take to make the patches acceptable greatly appreciated.

Thanks again.

-DB