[06:39] ubiquity: superm1 * r2497 ubiquity/ (7 files in 5 dirs): drop all references to mythappearance [10:22] Hi! [10:25] I've got a problem installing edubuntu on my brand-new machine (Xeon C2D, 8 GB RAM, 64 Bit Edubuntu 7.10, ARECA 1200/1201 RAID controller (SATA)) [10:25] The guys from #edubuntu sent me over to here :) [10:27] The installer works fine until it asks me to pick a module for drive detection [10:27] For the Areca 1200/1201 this should be "arcmsr". [10:28] The module seems to load (syslog: disk-detect: insmod /lib/modules/2.6.22-14-generic/kernel/drivers/scsi/arcmsr/arcmsr.ko [10:28] but the dialog pops up again, again asking to pick a module for drive detection ... [10:30] one moment [10:30] exactly what is the title of this dialog? [10:30] [!] Festplatten erkennen [10:30] :-) (German) [10:30] (= drive detection, I guess) [10:30] should I reboot into English language mode? [10:31] not necessary, thanks [10:31] (a) I understand German (b) I can look up the exact text in the translation files :-) [10:31] it is "Detect disks" [10:31] Ah, fine :) [10:31] are you sure that's the title of the error dialog? that sounds like the title of the general installer step [10:32] Gruesse auf die Insel! :-) [10:32] I was sort of expecting something like "Treiber werden für Ihr Laufwerk benötigt:" [10:34] but basically the problem is likely to be that even after loading that module no disk devices appear [10:34] could you extract the installer's syslog, using "Debug-Logs speichern" from the main menu, drop it onto paste.ubuntu.com, and post the URL here? [10:42] cjwatson, can you think of any reason wht a USB modem with built-in USB mass storage (CDFS) might jam the boot sequence? [10:43] I'd expect that if the device appears as a valid partition I might try to mount that when I scan for ISO preseed, but do not see why that would jam [10:45] sorry jamming is not what happened, the user was thrown into busybox [10:45] no, sorry - ask the kernel folks === thom_ is now known as thom [10:47] It is an installer step that pops up because ... "No disk drive was detected. If you know the name of the driver needed by your disk drive, you can select it from the list. Driver needed for your disk drive: .... continue with no disk drive // 3w-9xxx // ..." [10:48] allerbest: right, need syslog then [10:48] allerbest: oh, and the output of 'lspci -vvnn' if you can get it [10:48] you can run 'anna-install openssh-client-udeb' from tty2 and then you can scp files to another machine [10:49] _this_ is great! just a moment ... [10:54] cjwatson: http://paste.ubuntu.com/4998/ [10:54] http://paste.ubuntu.com/4999/ [10:54] http://paste.ubuntu.com/5000/ [10:55] http://paste.ubuntu.com/5001/ [10:55] http://paste.ubuntu.com/5002/ [10:55] http://paste.ubuntu.com/5003/ [10:56] that's it [10:57] (including lscpi -vvnn [10:57] ) [11:00] allerbest: last thing, can I see the output of 'parted_devices' and 'list-devices disk'? [11:01] parted_devices: empty [11:01] list-dev: empty [11:05] user-setup: cjwatson * r95 ubuntu/debian/ (changelog user-setup-udeb.templates): * Add the initial user to the fuse group. [11:06] allerbest: ok, I have to direct you to #ubuntu-kernel then; the installer seems to be doing the right thing, and there's no error from the kernel, but the disk device simply isn't showing up [11:07] cjwatson: Thanks for your patience. [11:10] evand: could you merge/push your user-setup changes and start off a 1.16ubuntu4 changelog with my change? I just realised you'd already uploaded 1.16ubuntu3 but it wasn't in bzr [11:53] debian-installer: cjwatson * r880 ubuntu/ (37 files in 2 dirs): * Update help text translations from Launchpad. [11:57] debian-installer: cjwatson * r881 ubuntu/ (build/boot/x86/f4.txt.live debian/changelog): [11:57] debian-installer: * Copy the live CD F4 help text from gfxboot into syslinux display format [11:57] debian-installer: (LP: #195564). [13:16] console-setup: cjwatson * r49 ubuntu/debian/ (changelog config.proto): * Set default layout for Vietnam to 'us' (LP: #191451). [14:00] console-setup: cjwatson * r50 ubuntu/ (Keyboard/ckbcomp debian/changelog): * Treat 'any' as a synonym for 'NoSymbol' in XKB input files (LP: #93077). [14:03] will do, sorry about that [14:03] I thought I bound that branch, apparently not [14:11] console-setup: cjwatson * r51 ubuntu/debian/changelog: releasing version 1.21ubuntu3 [14:14] user-setup: evand * r97 user-setup.ubuntu/debian/ (changelog user-setup-udeb.templates): Forgot to push. Merge with trunk. [14:20] evand: thanks [14:21] user-setup: cjwatson * r98 ubuntu/debian/changelog: releasing version 1.16ubuntu4 [14:23] evand: any idea what's going on with bug 194727? [14:23] Launchpad bug 194727 in ubiquity "[Hardy alpha 5] Live CD does not display release notes when clicked in installer mode" [Undecided,New] https://launchpad.net/bugs/194727 [14:24] looking at it now [14:29] yay zombie process [14:34] /srv/cdimage.ubuntu.com/bin/build-image-set: line 167: /srv/cdimage.ubuntu.com/bin/run-germinate: Permission denied [14:35] cjwatson: are you already aware of this? ^ [14:36] no ... [14:37] wonder how that happened? fixed [14:37] thanks [14:37] whoa, it's wrong in bzr [14:37] * cjwatson will fix harder [14:37] superm1: looks like you changed the permissions on run-germinate in your branch; please pull my changes once I fix it [14:39] superm1: (done) [14:40] cjwatson, okay thanks [14:42] cjwatson: the Firefox issue appears to have been related a previous state of ubiquity-dm, but it's fixed with the latest changes to that file. [14:42] I'll update the bug [14:42] sorry about that [14:42] hrm, although [14:46] odd. i didn't explicitly change them ever, that got pulled in last time i merged with mainline [14:48] bzr log -v definitely blames your branch, but never mind :) [15:05] evand does the new ISO fix kubuntu issues in wubi? [15:06] xivulon: the new ones will [15:06] i.e. tomorrow [15:06] dailies were not generated yet today [15:06] not necessarily [15:06] ok [15:07] just started ubuntu daily live CDs, I'll do a run of Kubuntu next. [16:19] so we have a somewhat interesting problem in that we don't have a way of talking to network manager in only-ubiquity mode. [16:20] I started looking at prodding it into selecting the first interface it finds using dbus, but that seems evil and unworkable in a number of circumstances. [16:20] keybuk pointed out that we could just use a lightweight notification area and actually launch nm-applet [16:21] which lets the user decide, as well as lets them input ESSIDs and passphrases if need be [16:21] I'm looking at the various options for that now [16:22] (the problem was discovered by trying to view the release notes in only-ubiquity mode) [16:22] FWIW I'm seeing the same problem outside only-ubiquity mode [16:23] a lightweight notification area doesn't seem unreasonabel [16:23] unreasonable [16:24] but I wonder what's wrong with what I have here - I had to add preexec_fn=drop_all_privileges but even then it breaks wih no obvious error [16:24] cjwatson: wget ubiquity-dm from the latest ubiquity in the initramfs [16:24] that fixed it from me [16:24] http://evalicious.com/tmp/ubiquity-dm to save you some time [16:24] *outside* only-ubiquity mode [16:24] for* [16:25] ah, misread [16:25] hrm [16:25] I can see it execing firefox [16:28] ah, it has HOME=/root [16:29] perhaps drop_all_privileges should set HOME? I think you want that in all cases where drop_all_privileges is used [16:31] sounds reasonable [16:33] if SUDO_USER in os.environ: os.environ['HOME'] = '/home/%s' % os.environ['SUDO_USER'] [16:33] something like the above seem ok? [16:34] I just tested os.environ['HOME'] = pwd.getpwuid(uid).pw_dir in the SUDO_UID block above, and that worked [16:34] works for me [16:34] (and 'import pwd' at the top obviously) [16:34] indeed [16:34] so where would we need to talk to network-manager? do you mean to get the network set up at all? [16:34] indeed [16:34] so that they can view the release notes, download language packs, etc [16:34] right [16:35] * cjwatson reproduces bug 193329 [16:35] Launchpad bug 193329 in ubiquity "[Hardy] ubiquity hangs after language selection, until languagechooser is -KILL'ed" [Undecided,New] https://launchpad.net/bugs/193329 [16:36] gah, it's back. [16:38] I think I see the problem [16:51] ubiquity: jriddell * r2498 trunk/debian/changelog: changelog [16:53] how on earth...well, something is bringing up a network interface in only-ubiquity on the absolute latest live CDs [16:54] though I imagine the point about having a notification area still stands. They could only have access over a WPA encrypted link. [16:56] ubiquity: cjwatson * r2499 ubiquity/ (debian/changelog ubiquity/components/language.py): [16:56] ubiquity: * Split iso_3166.tab lines by tabs, fixing an infinite loop on countries [16:56] ubiquity: whose names contain more than one word (LP: #193329). [16:56] I think it's only WPA where this is a problem [16:56] AFAICS the bug that was filed is about firefox not showing up at all [16:57] do you have the change for that in your working tree or shall I commit it? [16:57] you can commit it [16:58] assuming you're talking about setting $HOME [16:59] aye [16:59] well, that and dropping privileges [16:59] indeed [17:05] ubiquity: cjwatson * r2500 ubiquity/ (3 files in 3 dirs): [17:05] ubiquity: * Drop all privileges and set HOME before launching a browser to read [17:05] ubiquity: release notes (LP: #194727). [18:15] debian-installer: cjwatson * r882 ubuntu/ (9 files in 4 dirs): * Move to 2.6.24-10 kernels. [18:19] debian-installer: cjwatson * r883 ubuntu/debian/changelog: releasing version 20070308ubuntu30 [18:25] * evand wonders aloud why autogen now requires cvs [18:26] ubiquity: evand * r2501 ubiquity/ (configure configure.ac): Bump to 1.7.13 [18:27] cjwatson: what was the method you used to determine the minimum memory requirement for install in the past? bdmurray brought up the fact that we don't have a minimum recorded for the only-ubiquity mode. [18:28] cvs> autopoint [18:29] I think I just ran it experimentally with mem=256M etc. [18:29] ok [18:29] ubiquity: evand * r2502 ubiquity/ (debian/changelog ubiquity/components/partman.py): [18:29] ubiquity: * partman/confirm_new_label is now partman- [18:29] ubiquity: partitioning/confirm_new_label (LP: #190036). [18:29] I suspect the autogen->cvs requirement is new because I made it actually run autopoint et al [18:31] oops, my bad for not noticing that template change [18:33] ubiquity: cjwatson * r2503 ubiquity/debian/changelog: whitespace that made vim complain [18:33] man, 2500 commits? [18:34] actually, 536 of those were Guadalinex', but still [18:35] heh, still quite fantastic [18:35] ubiquity: evand * r2504 ubiquity/ubiquity/components/partman.py: Add partman-partitioning/confirm_new_label to the regex for clarity [18:49] woo, just got my first real-world demonstration of the md5 checking when a copy failed to produce matching checksums the first time around. [19:31] cjwatson: would it be safe to close bug 89528? They didn't provide the debugging information you asked for. [19:32] Launchpad bug 89528 in ubiquity "Kubuntu installer crashed while scanning disks (step 5 of 7)" [Undecided,Incomplete] https://launchpad.net/bugs/89528 [20:03] evand the chap with amd64 can install fine (except for a grub-install issue...) [20:03] do you still have the amd64 crashes? [20:03] I haven't tried with the latest images, but I'll give it a go in a bit. [20:04] there a couple of reports that seem to imply that menu.lst root is wrong when the user has multiple disk devices [20:04] yikes [20:04] if that is confirmed it is going to be a major pain in the neck [20:04] getting that right is going to be pretty tricky, no? [20:04] indeed [20:07] other bad news, was investigating a crash in the signature checking code, so I compiled with debugging signature... [20:07] when I run it like that it does not crash... [20:07] I hate those days [20:07] I was just going to say exactly that [20:08] ^so I compiled with debugging symbols [20:09] debugging symbols will help quite a bit, assuming I can weasel Visual Studio into giving me a decent backtrace in the cases where Wubi has crashed for me. [20:10] I have enabled debugging symbols for the plugins not for nsis (yet) [20:11] ah [20:11] but in my case the crash was in metadl [20:11] access violation [20:12] even though that can originate from nsis [20:12] for many things nsis is way closer to assembler than to a scripting language... [20:13] lol [20:16] http://nsis.sourceforge.net/Replace_line_that_starts_with_specified_string [20:17] ^ 2008 scripting language... [20:17] hahaha [21:04] xivulon: I think you have a stale lock on wubi on launchpad. Did you cancel a push? [21:06] hm don't think so [21:07] but feel free to kill the lock [21:07] will do, thanks [21:07] I will be home in a 1h [21:47] bdmurray: ugh. ubiquity fails in 256M under those conditions as well (no swap before or during install). [21:47] mind you it probably will still work for that bug reporter as they'll most likely pick one of the autopartitioner options, all of which create a swap partition === ebel_ is now known as ebel [23:21] is ubiquity the same inside the live environment and when you pick the option that just goes straight to install on the hardy live cd? i got different behavior [23:40] ubiquity: cjwatson * r2505 ubiquity/debian/ (changelog control rules): * Explicitly depend on ntfsprogs for NTFS resizing. [23:43] bdmurray: ok, queued up for looking at tomorrow [23:45] cjwatson: I was looking at the ubiquity cluefile for bughelper and a lot of clues say TODO for you. Could you see if those are still accurate? [23:45] when I'm not about to go to bed, sure :) [23:46] cjwatson: Well of course. [23:58] where do I find supported keyboard variant codes?