Being the `sinner' who originally changed the semantics of namei() to
its current behavior, I'd like to point out that this is the Right Way,
speaking in terms of POSIX; this also includes cp(1).

As one of the NetBSD Project's goals is compliance to `reasonable'
standards (POSIX is considered mostly reasonable), it's not desirable
to change the semantics of pathname resolution/lookup to the means
expected by `abusive' scripts (no offence intended); these would break
on POSIX-compliant implementations anyway.


In a nutshell: The current semantics of namei() conform to POSIX
               (which has to be regarded as the right behaviour),
               so please end this discussion on current-users.



-klaus