Re: support for translators in tar

[Miles Bader]

Kalle Olavi Niemitalo <kon@iki.fi> writes:
> A file system can save only one passive translator for one node;
> this is implicit in the file_set_translator RPC.  The original
> file system saves only the passive translator that is closest to
> it; that first translator is then responsible of saving the
> second passive translator, etc.
> 
> Now, where does the first translator save the second passive
> translator so that it will survive reboots?  It doesn't fit in
> the metadata of the underlying node.

Why should tar worry about that?  It seems both unnecessary and contrary
to the way tar generally works.

If the `first-level' passive translator needs any extra data, then the
person that's running tar needs to make sure that data is somehow saved
as well, if it's desirable in that case.  Note that the situation is
exactly analogous to that of symlinks: tar makes no attempt to do
anything magic with the file pointed to by a symlink, it just saves the
symlink, and any extra magic is the user's job.

The only case that seems annoying to me is that a person might wish to
have tar save certain types of passive translators in the tar file, and
`work through' others (for instance not stopping at normal filesystem
mount points).  Perhaps this sort of thing could be handled with an
option specifying exactly which passive translators should be
saved/not-saved.

-Miles
-- 
`Cars give people wonderful freedom and increase their opportunities.
 But they also destroy the environment, to an extent so drastic that
 they kill all social life' (from _A Pattern Language_)