Re: Summary: experiences with NanoBSD, successes and nits on a Soekris 4801

On Mon, 20 Jun 2005, M. Warner Losh wrote:

> In message: <Pine.GSO.4.43.0506201507010.11816-100000_at_sea.ntplx.net>
>             Daniel Eischen <deischen_at_freebsd.org> writes:
> : On Mon, 20 Jun 2005, M. Warner Losh wrote:
> :
> : > In message: <Pine.GSO.4.43.0506191610170.7472-100000_at_sea.ntplx.net>
> : >             Daniel Eischen <deischen_at_freebsd.org> writes:
> : > : How about NO_FOO[_INSTALL], where NO_FOO = no build and no install,
> : > : and NO_FOO_INSTALL just prevents the install.  In theory, you could
> : > : build the complete system, then use NO_FOO_INSTALL instead of rm(1).
> : >
> : > What's wrong with making sure that NO_FOO will work in the install
> : > case to not install foo when it is set, even if it was unset in the
> : > build process?
> :
> : If it works or can be made to work, then nothing.
>
> Actually, looking at the code, it would cause devd to be built, but
> not installed without changes.  Since NO_GXX is defined in the above
> scenario.  I've started to think about how this might be fixed.  It
> really is a 'don't build this because of toolchain depends' as a
> 'don't build his because I don't want this feature' intertwinglement.

Also, what about dynamic executables that need libstdc++, but you
still don't want the build tools?

-- 
DE