Fun with gcc.

[Rob Landley]

On Thursday 27 October 2005 19:02, Mike Frysinger wrote:
> On Thu, Oct 27, 2005 at 05:59:41PM -0500, Rob Landley wrote:
> > On Thursday 27 October 2005 15:54, Mike Frysinger wrote:
> > > On Thu, Oct 27, 2005 at 03:20:05PM -0500, Rob Landley wrote:
> > > > We _definitely_ need a platform.h file of some kind to stick that
> > > > kind of thing in.
> > >
> > > we already do that in uClibc just like glibc does, so we could just use
> > > the uClibc header as base
> >
> > We don't want busybox to require uClibc.  (We want them to play nice, but
> > not actually NEED.)
>
> you misunderstood
>
> i said use uClibc as a base as in copy the header file from it and then
> tweak it as we see fit

Sounds good.  Got a URL to the right file in the busybox SVN?

> > > > Also, doing that could potentially make the darn warning in
> > > > traceroute.c (the only one we have left!) go away.
> > >
> > > sure, if you build busybox as a 32bit binary ... build it as a 64bit
> > > and you'll see a ton more warnings (about 40 more warnings iirc on
> > > amd64)
> >
> > I haven't got onea them.  (I have, however, recently downloaded qemu
> > which can fake one of them, and I plan to play around with it as time
> > permits.  But first I downloaded a Red hat 7.2 image, because I'm curious
> > what breaks...)
>
> all you really need is a cross-compiler ... after all, you're trying to fix
> build time warnings atm ...

Actually I like the idea of being able to test that it works.  I'm funny that 
way. :)

Rob