Subject: Re: Building nathanw_sa branch?
To: Michael K. Sanders <msanders@confusion.net>
From: Nathan J. Williams <nathanw@MIT.EDU>
List: current-users
Date: 02/03/2002 00:32:00
"Michael K. Sanders" <msanders@confusion.net> writes:

> "Nathan J. Williams" writes:
> >Be sure to run "make includes" in src/sys, src/include, and
> >src/lib/libpthread before attempting the build of libc.
> 
> "make includes" failed in src/sys/fs, because the Makefile doesn't
> exist on the branch?  Easily fixed, but then...

Hm. I might have missed the new directory in my latest merge of the
trunk onto the branch. 
> cc1: warnings being treated as errors
> /usr/include/sys/ucontext.h:53: warning: no semicolon at end of struct or union
> /usr/include/sys/ucontext.h:58: parse error before `}'
> *** Error code 1
> 
> Stop.
> make: stopped in /src/netbsd/SA/src/gnu/lib/libg2c
> 
> Should I not be trying to do a full build?

Well, not really. The intersting bits are the kernel, libc,
libpthread, and libsyscall. Since the rest of userland doesn't make use
of the SA's or the libpthread features, it's not very interesting.

That said, things should be buildable. The particular problem you're
having there is a known one; it's because dtime_.c is declaring
_POSIX_SOURCE and then including <sys/param.h>, which is a totally
bogus thing for it to do, but which didn't happen to be a
problem. I've been working on a clean solution to this (it's actually
fixed in the gcc sources, but not in a version that will be released
any time soon), but in the meantime, don't rebuild gcc.

        - Nathan