On Mon, Mar 17, 2003 at 05:34:55PM +0100, Ralf Nolden wrote: > I would be interested why sysvinit needed to be hacked - I mean, if it's that > what counts -ejecting a CD - over the system's overall integrity then most > people are probably totally paranoid about their systems :-) agreed. also, i'm not sure that it *has* to be done this way... > Although I can understand that from Klaus' perspective it probably makes some > sense to eject the CD but I mean, the machine shuts off completely with an > open CD tray - this is one of the most notable misbehaviors that I personally > don't like to happen. I'd be much better off installing eject and call that > during a reboot, but certainly not on a complete shutdown :-) well, that could be fixed by another call to eject -t after someone presses enter... how about the following as a solution: we already have the capability to have a shmfs or some other in-memory fs/ramdisk, so why not copy over the eject utility and it's needed libraries into this ramdisk, umount the cd, and then somewhere between S40umountfs and S90reboot we put a script that does this copying, ejecting, waiting for enter, and then reloading of cd tray. this way, init wouldn't need a hack for this at all. would this work? are there any other reasons we need to have a patched init? klaus? sean
Attachment:
pgp0RLQ78Htjg.pgp
Description: PGP signature