Re: Results of test runs

[Leif Ekblad]

> This test has $host and $build i686-pc-linux-gnu, but we would like to
> see how it fares on RDOS.  So if RDOS is a cross-target, you would use
> something similar to
>   configure --host=i686-rdos-whatever
> (adjust to the appropriate value/your toolchain prefix) and possibly
> specify some of them also manually (CC, LD, AR, ..)

It says "checking host system type... Invalid configuration 'rdos' : machine
not
recognized"

> and if you can build natively on RDOS, just do that, it's bound to be
> simpler.

No, that is not possible. It might be some day, but not right now.

> I'm not sure what your goal is; your recent question indicates that you
> need the libtool changes in order to be able to create a GCC that
> targets RDOS.  Is that correct?

Yes. GCC uses the libtool.m4 file, and I need RDOS to be
defined there.

> Have you managed to either get them
> to apply your libtool-related changes, and/or have you maybe managed
> to just apply the yourself to a local GCC source tree and build from
> that, including regeneration of the generated files.

I've added them manually to GCCs libtool.m4, and then the build
process of GCC + libstdc++-v3 works just fine. I'm trying to get
the patches accepted for GCC.

> (Note I don't
> know exactly how that works within the GCC source tree, but I guess
> they have some documentation that explains how to do this?)

A guy refered me to the libtool project, so I assumed it was done in
a similiar way as config.sub and config.guess, some files being imported
to GCC, but it doesn't seem like this is how it works. It turns out there
are significant differences between the GCC and libtool version of
libtool.m4.

Leif Ekblad
www.rdos.net/rdos