Hi Andrey,

On Wed, Jul 18, 2001 at 11:34:45AM -0700, Andrey Petrov wrote:
> > Now I want to do a make build of the 1.5.1 system but make includes
> > produces a corrupt rpc/rpcb_prot.h. The "%#elif __STDC__" in
> > rpcb_prot.x produce correct "#elif __STDC__" statements in rpcb_prot.h,
> > but the "%#ifdef __STDC__" results in "#ifdef 1" which makes the
> > compile fail in a later phase of the build.
> > The same source works on my sparc (32) boxes and rpcgen generates
> > "#ifdef __STDC__" as expected on those machines.
> > I recompiled rpcgen but this didn't help...
> > 
> > Did someone else run into this problem? Or other ideas?
> 
> Can you give some details on what you're doing. Ideally I'd like to
> have a test which I can reproduce myself. If that fpu problem I might
> have a patch.

All I did was a fresh 1.5.1 install, checked out the 1.5.1 sources and
did a make build (without NOGCCERROR=1). At a very early stage, I got
complains about '#ifdef 1' in /usr/include/rpc/rpcb_prot.h and the
make failed. After some investigation I found the rpcgen behavior
described in my mail last week.
If the compiler warnings are not treated as errors (Thanks Martin!)
the make build runs fine until the end (~ 6hrs on the Ultra1/140) so I
now have a complete 1.5.1 userland and 1.5W kernel.
Currently I'm struggeling with the missing libstdc++ which seems to be
needed by pkgsrc/databases/postgresql-lib :-/

so long,
 Oliver
	
-- 
   opp@MuFFiN.Org       http://www.muffin.org/~opp/      +49-173-2313691

"Stay away from drugs, gangs, and cyberporn on the Internet and you
can be President some day."	    Elwood Blues in "Blues Brothers 2000"