On Tue, 2004-10-12 at 13:59 +0200, Adrian 'Dagurashibanipal' von Bidder wrote: > On Monday 11 October 2004 20.24, Marco d'Itri wrote: > > On Oct 11, Joerg Sommer <joerg@alea.gnuu.de> wrote: > > > why ppp provides its own mechanism of telling programs when the > > > interface is coming up or down? Many programs register for the ppp > > > mechanism, but not for the network mechanism. Where is the difference > > > and why both isn't > > > > Historical reasons? Also, many programs are interested only the ppp > > interface (it could be argued that they actually care about the > > interface associated to the default route, but at this point it would be > > a bit complex to do something about this). > > I agree with Jörg, though, that this should be unified for sarge+1. > > Jörg: file a wishlist bug with ppp and/or programs using ppp/if-*.d, > perhaps? First off the semantics of using ifupdown ifup.d scripts for ppp interfaces need to be made the same as using ppp ifup.d scritps for ppp interfaces. ifupdown - normally interface is created before ifup.d scripts are run, and does not go away again, except maybe after ifdown is run (modulo users deliberately doing stuff like rmmod). for ppp interfaces, they do not exist when the ifup.d scripts run, and may disappear and reappear without warning (thing persist mode)- without the ifupdown ifup.d scripts being run. In short, the scripts run once when the 'user' runs ifup, and once on ifdown. ppp's ifup.d scripts however are called for each and every time the interface finishes connecting .. and disconnects. In short, the scripts may run many times for on pon, and then once for poff. Thats quite different - I'd love for this to be consolidated and addressed though. Rob -- GPG key available at: <http://www.robertcollins.net/keys.txt>.
Attachment:
signature.asc
Description: This is a digitally signed message part