[00:02] * shtylman is thinking it is time to try a clean install of the new alpha...oh this will be failue...I can smell it :) [01:37] so I finally been able to do a full installation in my vm... [01:38] update-grub has to be fixed slightly but otherwise things seem ok [01:39] main issue is that I get almost consistent fs corruption (for the loopfile) at each reboot... [01:39] will investigate that in the coming days [01:40] also have an error when inserting i915, but should be orthogonal to wubi and I guess you know about it already [01:40] cjwatson ^ [01:42] thanks soren for the kvm tips [02:46] cjwatson: you around? [03:14] cjwatson: well...im gonna sleep some...but I am in the middle of an alpha 5 install..actually just need to install grub2 at this point. LP #420992 is stopping me. I am still in the install livecd system and would like to help work on or create a fix for installing grub onto dmraid devices. I can use the debbug patch as a starting point if you can point me in the right direction for which source package to work with (aka...the [03:14] source package that does that text-based grub installer) ... thanks and I will see if I can come up with a fix :) [03:14] Launchpad bug 420992 in grub-installer "Cannot install grub2 on dmraid with text-mode installer" [High,Triaged] https://launchpad.net/bugs/420992 [08:36] evand: Hey! I installed UNR via Wubi and it doesn't boot with "Unable to find a medium containing a live file system" [08:39] StevenK: it's my day off, but I have a few minutes. Can you post the casper.log on pastebin? [08:40] evand: It has "/init: line 1: can't open /dev/sr0: No medium found" [08:41] * StevenK makes sure to select the CD is mounted [08:42] select the CD is mounted? I don't follow. [08:43] evand: Sorry, I'm fiddling in VirtualBox [08:43] it shouldn't need the CD drive at all [08:43] it copies the ISO to disk [08:45] evand: How is a good way to debug this? [08:47] StevenK: given that a regular install is working just fine for me, compare what wubi is producing for UNR vs a regular Ubuntu install [08:48] make sure the kernel command lines match up and that all the paths exist [08:48] evand: morning [08:48] davmor2: good morning [08:50] evand: would it be okay to play with wubi on unr today see if we can't get it to work by the end of the day? [08:50] davmor2: sure, but I'm not available to help. Today's my day off and I'm going into town in a few minutes. [08:50] davmor2: I just tried it, it doesn't boot, but it does install [08:51] StevenK: Yeah that was the issue I had before [08:51] evand: okay how about Monday instead then :) [08:51] davmor2: sure [08:51] Be nice to get it working :) [08:52] davmor2, evand: I'll leave it to you two, then :-) [08:53] evand: out of interest the grub issue was it in the wubi package? and did you upload it so I can pull it and try it out. Or is it still in the works? [08:53] davmor2: it'll be on the next CD spin [08:53] alternatively you can pull it from http://people.ubuntu.com/~evand/wubi/karmic/ [08:53] evand: Okay cool :) [08:54] I'll have a play and let you know monday :) [08:55] good deal [08:55] thanks [09:04] shtylman: well, how about I make that the next bug I try to fix? [09:05] hmm, I thought somebody said there was a patch for dmraid in the Debian bug, but it's just for multipath AFAICS [09:07] * cjwatson breaks out Luke's test dmraid qcow2 images [09:25] evand, cjwatson left a msg yesterday night ^^ [09:26] have also played a bit with loopback function of grub2 [09:27] I can list the content of the disk image, which is encouraging [09:27] (using loop ext2 grub modules) [09:29] but if I try to access a dir therein e.g. ls (loopdev)/etc => error: out of disk [09:29] does that sound familiar? [09:36] no - I can't reproduce it with a small test image [09:36] (using grub-emu) [09:36] if there's any way I can reproduce this without a full wubi installation, that would be good [09:37] what's the exact text of the error message? [09:37] "out of disk" [09:37] ah, there it is [09:38] so that actually means out of range [09:38] It might be related to the ext4 corruption problem mentioned earlier though, will try again tonight ensuring that I run this on a clean partition [09:38] do you know whether LBA is enabled on the containing disk? [09:39] I narrowed my problem to "apt-get install ^" not finding any task, even "minimal" [09:39] hmm didn't check that [09:40] strace seems to show that it is looking inside the Package files, unlike tasksel [09:40] xivulon: it's more likely to be straightforward filesystem corruption, or a bug in grub2's ext2 code, but it's also possible that it's outside the range it can read with non-LBA accesses [09:41] yann|work: it relies on correct Task fields being present in the Packages file; if you regenerated Packages and you weren't careful, you may have broken those [09:42] cjwatson, will try again tonight and provide more informative feedback [09:43] cjwatson: I did not touch the original ones, they do have their Task: fields - only the Packages file for my "extra" component does not have those [09:44] you might like to use 'apt-cache show util-linux' (as an example) to check that it shows the Task field [09:46] Ah, it does not show - and apt-cache policy shows that the same version exists in both main and extras - so this is where I shot myself in the foot, my filter does not work [09:46] cjwatson: thx [09:47] re LBA, the bios is the stock KVM one [09:50] cjwatson what were the settings to keep author changelogs together in dch? [09:51] cjwatson: btw, how are the task fields added ? I did not find any info about that when looking for it - I would have thought there was some option in apt-ftparchive config file, but did not find one [09:52] xivulon: I use this in .devscripts: [09:52] DEBCHANGE_RELEASE_HEURISTIC=changelog [09:52] DEBCHANGE_MULTIMAINT_MERGE=yes [09:52] yann|work: look for "extra overrides" [09:53] cjwatson thx! [09:55] cjwatson: ah ok - and there is some tool to generate extra-override file from ubuntu-tasks.desc ? [10:00] yann|work: it's generated, although not that particular way round [10:00] our debian-cd instance spits out an extra overrides file, if nothing else [10:01] based on germinate output [10:02] yann|work: but you may also be able to use the override.*.extra.* files in the /ubuntu/indices/ directory of an Ubuntu mirror, as a short-cut [10:04] cjwatson: ok, thanks [10:05] I did not know germinate, useful tool indeed :) [12:07] cjwatson: shouldn't "apt-ftparchive -o BinDirectory::ExtraOverride=myfile" be enough to add Task fields ? [12:26] cjwatson: ok... I can test any potential fixes if you need ... I am curious though...does the grub-pc source package make all the grub-* suite of tools? [12:36] yann|work: maybe Tree::ExtraOverride; it's been a while since I looked at any of this [12:36] shtylman: yeah [14:24] damned, Tree::ExtraOverride does not work either :( [14:33] cjwatson: FWIW, options for the config file are ignored when passed on commandline, and unrecognized options are silently ignored [16:12] cjwatson: it might be possible that the alternate image kernel wasn't updated: "The installer cannot find a suitable kernel package to install." [16:13] I don't think that's the cause. Full logs please? [16:14] that means it couldn't even find an image of the right type, never mind ABI version [18:04] cjwatson, I'm looking at console_setup.py, where it modifies xorg.conf. It looks like it only modifies anything if there's a 'Driver kbd' section. Is that expected? [18:05] I think so, yes [18:06] why? [18:06] cjwatson, well, I thought that was the primary way we set default layout. If user doesn't have an InputDevice or Device kbd section, they don't get new layout? [18:07] X will pick it up from /etc/default/console-setup if it's not separately configured in xorg.conf [18:07] cjwatson, ah, that makes sense then, thanks. :) [18:07] cjwatson, I see that bit of code now [18:07] see /usr/lib/hal/debian-setup-keyboard [18:08] or possibly something newer for halsectomy, but that's the general idea [18:08] :) [18:08] Nice, thanks. [19:23] ubiquity: superm1 * r3429 ubiquity/debian/ (changelog oem-config.init): [19:23] ubiquity: Use the usplash init script to stop usplash in the oem-config init script [19:23] ubiquity: It seems to do a better job, and prevents a black screen on boot with some [19:23] ubiquity: drivers that aren't as usplash friendly. (LP: #403021) === nxvl_ is now known as nxvl [21:18] superm1: oh, cool, thanks for figuring that out [21:19] no prob. it was actually totally accidental. i was debugging a different problem on a freshly installed system, and found out that oem-config was working without usplash, so put two and two together :) [21:19] reassigned the bug back so that it gets auto-closed properly