Supermount (NOT Mandrake!)

[rmoser <mlmoser@comcast.net>]

Okay, Mandrake's supermount is marked as stable, and it's broken
and a piece of shit.  However, supermount is useful on the iPaq and
convienient in other places as well.  I'm going to work out a somewhat
workable alternative to that, and try to impliment it.  Once I fail at that
(some outlook on life I have, huh?), I'll put the proposal up on here for
you all to take a shot at.

One of the things I'm thinking about is multiple partition devices.  For
example, Zip disks.  Some of us make a Zip disk a single filesystem,
with no partition table: `mount /dev/sda /zip`.  Others leave the 4
partitions on the Zip disk when they get it:  `mount /dev/sda4 /zip`.
What to do, what to do.

Of course the answer's simple.
`mount supermount /zip -o device=/dev/sda,user`, and if we have 4 partitions,
/zip becomes mod a-w, and 4 new directories appear:  /zip/1 /zip/2 /zip/3
/zip/4.  Then supermount mounts each partition on each of those, if
possible.  If it's a broken partition (/dev/sda[1-3] on a new zip disk), it's just
marked mod a-w.

Question:  Can I do this?  I've never programmed in the kernel before but i've
tried several times.  I know about the automounter.  Can I control it from a
virtual filesystem device?  Like, on accessing /zip, could those dirs be created
virtually, then suddenly a supermount mounted on each of them?  I have no
idea what I'm doing but I'm trying!

I'm gonna send this before I feel too stupid to.

--Bluefox Icy