[00:19] <slangasek> zyga: I don't agree at all that editing /etc/hosts is unreliable.  On the contrary, I think throwing an extra DSO in the NSS stack is more of a reliability hit.  Furthermore, libnss-myhostname has the behavior of mapping the hostname to the current interface addresses, which is a definite behavior change and one I don't think we want
[04:31] <nerdistmonk> Can someone who built this thing tell me how to get control of OnDemand
[19:25] <maxb> Hi. Does anyone know where I could look for more information about the state of https://blueprints.launchpad.net/ubuntu/+spec/core-1311-dmraid2mdadm ?
[21:28] <zyga> slangase`: how would you reliably edit /etc/hosts if someone has made local changes that, say, don't parse? or how expensive and complex would a parser that retains all the current formatting be? I think this is a solution looking for a problem. Perhaps the current NSS-based approach is imperfect but I would argue that it is on the more-right track as it is immune to /etc/hosts being a free-form text fi
[21:28] <zyga> le that cannot be reliably patched in all cases.
[21:28]  * zyga wonders if irssi has a buffer it splits somehow?
[21:29] <zyga> slangase`: I'm curious why debian needs the local hostname in /etc/hosts? fedora somehow doesn't do that, do you know the reason?
[21:32] <JanC> zyga: you must be able to resolve the local hostname
[21:32] <JanC> so it has to be defined somewhere
[21:35] <JanC> or maybe "some software expects to be able to resolve the local hostname" is more correct
[21:37] <JanC> or do you mean why it needs a local IP address?