[03:16] seems xnet don't charge for moving anyway (i think the thing I saw about a charge on their website is only if you're connected via fibre) [04:03] so Precise comes with a local dns resolver by default? [04:04] yes, dnsmasq [04:05] ajmitch: that's not installed on my box... [04:05] at least according to: dpkg -l dnsmasq [04:05] package is dnsmasq-base [04:06] afaict networkmanager starts it up & passes the resolvers that dhclient gets through to dnsmasq [04:06] ah i see, it's just a dns proxy [04:07] hmm, I need to do precise at some stage [04:08] yeah I'm not sure of the details of what it can do :) [04:08] there seems to be a new thing to manage resolv.conf (called resolvconf) too [04:08] it's not that new, is it? [04:09] debian/changelog says it first made it into experimental in 2003 [04:10] dunno, seems to have only started appeared in my ubuntu-based appliances since I upgrade to precise [04:10] or maybe it worked differently before [04:10] it was moaning that /etc/resolv.conf isn't a symlink [04:10] so I just added it to the list of packages to uninstall [04:10] yeah, that part's new [04:11] not sure when it started being used by default in ubuntu [04:12] generally it seems a good idea to have things to manage this, rather than N different things all fighting to update it [04:12] the appliance situation is a bit special really [19:35] morning [19:41] morning [20:23] morning [22:22] morning [22:33] ibeardslee: zareason coreboot-based laptop, fuck yeah! [22:38] yeap [22:38] wat? [22:38] what is coreboot? [22:38] their bios is already 'open' [22:38] and customised [22:39] https://twitter.com/ZaReasonNZ/status/210488042597777408 [22:39] http://www.coreboot.org [22:40] * ibeardslee goes looking for the bios zareason use [22:40] would be nicer than having to turn off secure boot :) [22:40] morning [22:41] thumper: coreboot is a free software bios. it used to be called linuxbios [22:41] actually if coreboot was trusted, would that solve the problem? [22:41] yeah, i guess zareason aren't too worried about getting the 'made for windows 8' sticker [22:41] apparently boots to a linux console in like 3 seconds or something [22:42] ibeardslee: coreboot wouldn' [22:42] ibeardslee: coreboot wouldn't be trusted unless it only booted to kernels that were signed [22:42] true [22:43] fmarier: booting to a desktop in < 10 seconds (with autologin) should be possible [22:43] the fundamental problem with all this is that if joe random free software dev can make a kernel you can boot, so can joe random haxxxxor [22:44] ajmitch: except that legacy bioses usually waste up to 10 seconds setting up the hardware for DOS [22:44] ajmitch: about 50% of my boot time is bios currently i think [22:44] the "trusted" part seems to really mean "trusted to maintain microsoft profit levels" [22:44] http://www.insydesw.com/products The InsydeH2O is what the Alto I had a look at was running [22:45] fmarier: right, so you'd probably be able to get under 10 seconds with that coreboot bios, is what I meant [22:48] ajmitch: oh, sorry i misunderstood what you said. having a cold boot take as long as resume-from-suspend-to-ram would be pretty neat :) [22:49] ojwb: i believe that the "trusted" in "trusted computing" has always meant "trusting microsoft to act in your best interest" [22:49] I know that ubuntu booted from grub to login in ~10 seconds on some hardware for some releases, I think it could be possible to be faster now [22:49] esss esss deee [22:50] yup [22:50] actually, i need to reboot after an update so i'll time grub -> login [22:50] brb :) [22:52] booting is certainly not < 10 seconds for me, I've got far too much running at startup, like postgresql :) [22:52] about 12s [22:53] I pushed Oneiric 64bit to my netbook (and then upgraded to precise) tempted to drop back to 32bit .. does seem slower [22:56] speed of 64-bit vs 32-bit depends on quite a few things [22:57] there's talk of a 32-bit userspace ABI that can use the extra registers, it might make apps use a bit less RAM [23:00] i'd like someone to obsess about restoring from hibernation as much as they seem to about a clean boot [23:01] for some reason hibernate seems to be even more problematic than suspend these days [23:01] there are "fast suspend" kernel patches around at least [23:01] fast hibernate I mean [23:02] e.g. http://tuxonice.net/ [23:03] handily comes with patch versions for jaunty, karmic, lucid, and maverick [23:03] 3 of which are now EOL IIRC [23:04] apparantly there's a PPA for more recent versions [23:06] yes, I was just looking [23:06] the kernels all say there's a newer version, but I was failing to find the actual versions to compare with [23:06] not a great sign though [23:08] AIUI, the standard hibernate essentially just swaps out everything, and on restore stuff just gets paged in on demand [23:08] so the order of loading is largely random [23:08] morning [23:08] probably works much better for an SSD [23:10] * ajmitch stabs LP timeouts [23:10] most things work better with an SSD don't they? .. .. well apart from my wallet [23:10] ojwb: the kernel is only out-of-date with respect to precise-proposed [23:11] ah [23:11] that's rather misleading on launchpad's side then [23:11] ibeardslee: having lots of data doesn't [23:11] which rather ruins them for me [23:11] the Newer Version link says the newer kernel is in -proposed [23:12] oh, i missed that link [23:12] or I'm just too used to how LP lays things out :) [23:12] d'oh [23:12] no, it's very obvious [23:14] looking at the upload dates, it is actually fairly actively maintained for supported ubuntu releases