> The behaviour that is implemented in 4BSD (and NetBSD) is described in
> a draft of an amendment to .1 (i think).  It's not yet a "real"
> standard, but it's likely that it will be.

> I'm not comletely familiar with the behaviour described, but Keith
> Bostic went out of his way to explain it in symlink(7).

Apart from trying to delete symlinks in /tmp, the most noticeable
effect seen by ls -lt or ls -rtl in a directory containing symlinks.
The symlinks are typically all gathered at the top or bottom of the
list - confusing at best.  No special casing of symlinks will fix that
- unless ls -r simply leaves them out :-) It just gets better...

Since the concept was (according to McKusick) introduced so that local
filesystems would behave the same as files systems mounted from lame
(my word not his) systems like NT?  I'd vote for introduction of
either lpfs (lame posix filesystem) as a new filesystem to use the new
semantics, or lpsl (lame posix symlinks) as a mount option.  I don't
know about everyone else, but I'm far more likely to NFS mount from a
UNIX system that uses the old semantics... 

In any case, its a trivial patch to produce the old semantics so those
of us that like the old can still have it - for now.


--sjg