--
Alain Coetmeur, Informatique-CDC DTA
mailto:[EMAIL PROTECTED]
> -----Message d'origine-----
> De: Brian Servis [mailto:[EMAIL PROTECTED]]
>
> I really like the power of program mounts(even though I am not really
> taking advantage of them yet). I would be really interested in seeing
> some (pseudo) code from others on this list for some of there more
> 'creative' program mounts.
>
> How about a bragging rights contest for the user with the most
> insane/useful/robust/creative/etc program map. hpa can be the sole
> judge and use the (pseudo) code as an example in the source
> release with
> mention of who's it is. This sounds good to me! Although I won't be
> making a submission, =).
look at this preiminary package
http://www.paralline.net/projets/autofs_slashnet/index.en.html
I've talk about it in another message here
(RE: supporting /net (was: Submounts revisited))
the /net/<host>/<mountpath> map is just using showmount to check
wether an asked mount exists, or is ambiguous
and need an autofs submount...
the /smb/<host>/<share> is simpler since it is just maping to
a smb type mount
I'm tweaking version 2 of this maps.
/net now check the existence of a NFS
host with "ping" to avoid showmount -e 120sec timeout.
note that if someone know how to control this (portmapper)
timeout, and reduce it to a few seconds, I'll be happy...
in V0.0.2 I'm also working on a
/smbuser/<user>/<host>/<share>
but I'm stuck to smbmount limitation in 2.0.5a...
I try to protect the /smbuser/<user> directory
from other users with chown/chmod but I need
gid/uid control that no more exist in smbmount 205a...
version 2 is not public, since /smbuser
does not works properly, but if you're interested...
by the way, I'm not very proud of these scripts
since I suppose you can make them crash or
use them to became root with magic automount name.
moreover sh is not very efficient and perl or C++
should be prefered.
If some one could take just the idea, and implement it
a clean way, it will be very cool.
