On Wed, Feb 21, 2007 at 04:31:42PM +0100, Reinhard Tartler wrote: > [EMAIL PROTECTED] (Sjoerd Simons) writes: > > > On Wed, Feb 21, 2007 at 03:48:08PM +0100, Reinhard Tartler wrote: > >> Trent Lloyd <[EMAIL PROTECTED]> writes: > >> > This check is done at IP up time because this is in no way a once-off > >> > check, it could vary from network to network > >> > >> Err, yes, there are broken networks in use out there. Is it really the > >> job of random packages to check for random breakage? > > > > Not of random packages. Only of the packages related to it :) > > After a fresh installation of Debian, it was very unobvious that avahi > is the culprit for the boot delay. From the users POV, this was some > random breakage in some random package. (This bug is actually a bug > forward from a user I'm supporting). > > > But well, yeah we must do this unfortunately. There are too many > > broken networks out there and we shouldn't break them.. > > Annoying delays at boot time punishes users who have a non-broken > network. > > > Remember people will accuse Debian for being broken, > > not their precious network! > > I notice that people accuse Debian right now for being broken by having > a 3 sec penalty delay with using avahi. > > Don't get me wrong, I personally enjoy using avahi. This particular > check however is pretty annoying.
But unfortunately "necessary" in order to not break networks relying on .local and at the same time having nss-mdns work correctly. I imagine this check can be fully backgrounded, but I do see one major use case this could break 1) The network comes up, avahi check is backgrounded 2) Some service relying on a .local resolution starts and fails 3) The backgrounded avahi check finishes Cheers, Trent -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
