>>(I guess for pkgsrc NetBSD will always be priority -- after all, it is
>>"the NetBSD packages collection".)
>
>Exactly. As you acknowledge that simple fact, you make my point.
>
>This leaves room for a package system that explicitely supports several
>platforms on an equal base -- like NetBSD has only first class citizens
>when it comes to supported platforms, contrary to, say, Linux. When

The analogy isn't quite valid.  If NetBSD developers dropped NetBSD for
some reason and began feeding in their work to, say, FreeBSD, much of the
material would probably be discarded.  (And, even if it were accepted,
it'd probably have to have a ton of #ifdef's to keep from compromising the
FreeBSD goals.  But, #ifdef's compromise the NetBSD goal of having a
clean, unified system.)

You couldn't merge two OS projects because they tend to have seperate
goals.  But the pkgsrc projects really have the same goal and
infrastructure support for adding a new OS should be relatively small.
(Compared to the overall structure; pkgsrc.tar.gz is over 9MB, I think.)

If Zoularis support is less than spectacular, I can't believe that it's
because anyone is discarding clean fixes.  Instead of patches being
filtered, I tend to suspect it's because there aren't that many clean
patches being generated.


On the other hand, I like plurality.  And a small, experimental package
system may be able to try out things that an established workhorse would
balk at---but if the new system works well, some of its ideas may be
appropriated.  I just wouldn't raise my hopes too high, if I were you, for
OpenPackages (or whatever) to have more volunteer support for Solaris than
does pkgsrc.


  ``I probably don't know what I'm talking about.'' --rauch@math.rice.edu