On Thu, 3 Nov 2005, Dieter Baron wrote:

> > >   If we really want support for required options, I can add explicit
> > > support for this to the options framework (via the new variable
> > > PKG_OPTIONS_REQUIRED_OPTIONS, for example).
> >
> > That would be a nice addition.  However, there's still the PKG_FAIL vs.
> > PKG_SKIP difference to worry about.
>
>   Why would you rather set PKG_SKIP than PKG_FAIL?  Would you prefer
> PKG_SKIP over PKG_FAIL for empty required groups as well?  If not,
> what is the distinctive difference?

The difference mainly shows up in bulk builds.  Packages which are skipped
with PKG_SKIP_REASON do not directly show up as "broken", as make will exit
with a successful result value.  That's why, for instance, USE_X11 with a
lack of X11 libraries is a failure (because it's fixable), but lack of a
usable JVM on a platform is a skip (because it's typically not fixable
without major work).

But the more I think about this, a failure is OK.  The whole fail/skip thing
probably needs a rework anyway.  :-/

-- 
-- Todd Vierling <tv@duh.org> <tv@pobox.com> <todd@vierling.name>