>> Perhaps this would be a good use for the sticky bit on a symlink.
> Didn't you previously suggest using the sticky bit on a symlink to
> turn on magic interpretation of variable expansions in symlinks?

Quite possibly.

> Given that the bit can't be used for more than one purpose, magic
> variable expansion strikes me as more needful of per-symlink control
> than magic ".." handling.

True.  I don't see any consensus that either one is worth doing,
though, so I'm not sure there is any real conflict.  (I think set-ID
bits would be enough more useful that I'd implement them first anyway.)

> (If I wanted magic ".." handling, I'd almost certainly want it
> globally, whereas I want per-symlink control over magic variable
> expansion.)

I'm not sure I'd want it globally, but on reflection I do think
per-symlink is the wrong granularity for it.  Per-process, maybe?

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B