>>Did I say this is the only solution you should implement?  No.
>
>No; you said it's the solution we *should* implement.

No; I said it is one of the solutions you *could* implement.

>>Did I say that it would work for everyone?  No.
>
>Did you say it would work for NetBSD? I thought that's what you said.

And for most of your mirror infrastructure and many of your
end users, it certainly could.

>>What I said was that it is the most efficient solution available. If you
>>want to design a solution to this problem that ignores what FreeBSD has
>>learned by doing it first, or insist that all of your users get their
>>source using a single method restricted by the common denominator, so
>>be it.
>
>That's not my choice. If I had the choice, I'd rather FreeBSD had had
>the clue to do something more portable than a hard-to-port (and
>arguably dead) language like modula-3.

CVSup was written by an individual who chose to use modula-3.  FreeBSD
didn't have anything to do with making CVSup happen other than be fortunate
enough to have someone affiliated with the project decide that writing
it would be fun.

Sure, modula-3 has not been ported to as many platforms as C, but it 
certainly is portable.

>If i had the choice, I'd look
>for something with a real protocol spec that's passed an
>interoperability test with an independent implementation, so that we
>know the protocol spec is complete and workable.

This never happened with SUP.  You've used SUP haven't you?  The CVSUP
protocol happens to be well documented too, which is more than I 
can say for SUP.

>If you really had a solution, I'd use it. But for me, and for many
>other NetBSD users, FreeBSD doesn't *have* any solution. Not in any
>way, shape or form. Nada. Zip. That's all I'm trying to get across,
>but it doesn't seem to be working.

This is silly.  If you can't use CVSup to get FreeBSD source, you can
get it via anoncvs.  If you can't use anoncvs, you can use ftp.  The
truth of the matter is that FreeBSD offers several way to get our
CVS tree.

--
Justin