=== rgreening_ is now known as rgreening [09:13] ubiquity: evand * r4356 trunk/ (debian/changelog scripts/install.py scripts/plugininstall.py): Make sure /var/lib/ubiquity/langpacks gets generated (LP: #630924). [09:34] ubiquity: evand * r4357 trunk/ubiquity/install_misc.py: Overwrite /var/lib/ubiquity/langpacks. [10:27] ev, ARGH ! [10:28] ? [10:28] ev, what did change in oem-config with the last upload, it doesnt start on the armel images anymore [10:28] seems it looks for an oem user [10:28] logs please [10:29] i get a traceback on the screen for line 94 and 531, i cant easily get logs of that borad [10:29] pwd.getpwnam() [10:30] and it ends with a KeyError ... getpwnam(): name not found: oem [10:30] you don't have USERNAME set in casper [10:31] we dont use casper [10:31] the images are preinstalled, we cant create a user there [10:31] it worked without until yesterday [10:33] crap [10:33] that code hasn't changed [10:33] * Run oem-config-wrapper as the user oem rather than root. [10:33] - Fixes issues with pulseaudio applet not working [10:33] - Prevents root terminals from ctrl-alt-t (LP: #594233) [10:33] there we go i guess [10:34] ah [10:34] * ogra sighs [10:34] superm1, is there a way to not do this ? ^^^ [10:34] we need to run as root on the preinstalled images [10:42] Do we? Can we not create an oem user in jasper? Why do we need root? [10:48] persia, i really dont want to keep any fragments of the oem user [10:48] afaik that user is originally created in the aufs mounted setup [10:48] Oh right, cleanup without layered fs would be painful. [10:49] ev might prove me wrong here (please !) [10:49] Maybe oem-config-wrapper should try to run as "oem" and fall back to "root" if it fails? [10:49] ++ [10:49] good idea [10:52] * ogra adds an oem user to test with [10:53] ev, hmm, the language table design seems to have changed, its only populating 2/3 of the vertical space of the window (all langs are there but i needs to scroll a lot more sideways now) [10:55] heh [10:56] the hostname ends with a hyphen for me ... - [10:56] ah, nice, the progressbar is a lot more informative now [10:57] ogra: hostname> can you pastebin me the output of dmidecode? [10:58] ev, arm board, i doubt i even have it installed ... let me see once oem-config finished (10-15min) [10:59] ugh, taking that back, it already finished [10:59] that cant be right [11:00] yeah, seems debconf didnt apply *anything* [11:00] wrong kbd and language [11:01] ogra, persia: http://paste.ubuntu.com/494660/ - look okay? [11:02] * ogra hugs ev [11:02] awesome ! [11:02] you rock ! [11:02] (and it does look ok :) ) [11:02] ev, Looks clean to me. Thanks for the quick fix even before ogra dug into it :) [11:03] ubiquity: evand * r4358 trunk/ (bin/ubiquity-dm debian/changelog): [11:03] ubiquity: Fall back to root if the regular user cannot be found in [11:03] ubiquity: ubiquity-dm. [11:03] sure thing [11:08] ev, good morning [11:08] hi ara [11:08] good morning [11:08] ev, I started testing ISOs with 2.3.18 as part of the pairwise testing cycle I told you about [11:08] awesome! [11:09] anything come up so far? [11:09] ev, after reboot, I got dropped to grub shell [11:09] gah [11:09] ev, bug 640341 [11:09] Launchpad bug 640341 in ubiquity (Ubuntu) "After installation, I get dropped to a grub shell (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/640341 [11:09] ev, let me know if you need any other information or if you need me to rerun with -d [11:11] sure thing [11:11] thanks [11:12] you're welcome [11:16] that log doesn't show it installing grub [11:17] indeed [11:17] in kvm trying to reproduce and dig at it now [11:17] I suspect it's fallout from the --no-bootloader option [11:23] ev, verified, no dmidecode installed on armel [11:24] ev, hrm, and archdetect only returns armel/unknown [11:25] i guess thats why i dont have anything behind the hyphen ? [11:25] ogra: found the problem [11:25] I think [11:26] well, it should return /omap4 at least .... /me looks into archdetect [11:27] debian-installer: cjwatson * r1363 ubuntu/ (7 files in 2 dirs): Move to 2.6.35-22 kernels. [11:28] ubiquity: evand * r4359 trunk/ (debian/changelog ubiquity/plugins/ubi-usersetup.py): Properly handle dmimodel() returning the null string. [11:28] ogra: ^ see if that fixes it ending in a - for you [11:29] will do [11:29] if you have a chance [11:29] thanks! [11:29] I thought we were falling back to laptop-detect when dmidecode didn't work. [11:30] we are [11:30] the code was broken though [11:30] it was testing for a null string where one never existed [11:30] Aha! [11:37] ARGH [11:39] ubiquity: cjwatson * r4360 ubiquity/bin/ (oem-config-remove oem-config-remove-gtk): mark executable [11:42] superm1: would you care to review Sebastian's proposed version of oem-config-remove-gtk in bug 637517? since you wrote it to start with ... [11:42] Launchpad bug 637517 in ubiquity (Ubuntu) (and 2 other projects) "oem-config-remove-gtk crashed with AttributeError in _run_transaction() (affects: 2) (heat: 14)" [Undecided,New] https://launchpad.net/bugs/637517 [11:42] ubiquity: evand * r4361 trunk/ (debian/changelog scripts/plugininstall.py): Fix bootloader installation test (LP: #640341). [11:42] *sigh* [11:42] libdebian-installer: ogra * r213 libdebian-installer/ (debian/changelog src/system/subarch-arm-linux.c): add oamp4 pandaboard support [11:43] grr, i hate if i typo commit messages [11:43] Trick is not to use bound branches, so you can uncommit :) [11:44] if you're uploading libd-i, please take care with the non-VCSified stuff in the source package [11:44] persia, i didnt push yet ... but CIA exposes the embarrasment world :) [11:44] err, crap i *did* push [11:45] cjwatson, you mean the autoreconf run ? [11:45] or is there anything else i'm missing [11:46] I mean that you need to debdiff carefully against the previous source package and ensure that there is only the change you made [11:56] ev, I have seen that you found the issue with grub [11:56] ev, anyway to recover my installation? :) [11:56] ara: indeed, thanks for pointing that out so quickly [11:56] would rescue mode do the trick? [11:57] cjwatson, ah, k [11:58] ara: go into a live CD, mount your root partition on /mnt, bind mount sys, proc, and dev into /mnt, then run sudo chroot /mnt grub-installer /dev/sda (where /dev/sda is the disk you want to install the bootloader to the MBR of) [11:59] thanks, will try that [12:13] ev, is this bug going to happen to everybody? or it is only certain installations? [12:13] ara: everyone [12:13] I'll upload a new ubiquity shortly [12:14] and ask if I can do a respin [12:14] ev, OK, I will send an email to avoid people testing this one [12:15] cool, thanks! [12:32] ubiquity: evand * r4362 trunk/ (3 files in 3 dirs): [12:32] ubiquity: Set a the respective timezone, both internally and visually, when an [12:32] ubiquity: entry is selected from the geonames autocompletion. [13:11] libdebian-installer: ogra * r214 libdebian-installer/ (debian/changelog src/system/subarch-arm-linux.c): add support for the OMAP4 Blaze board [13:13] libdebian-installer: ogra * r215 libdebian-installer/debian/changelog: releasing version 0.76ubuntu2 === rgreening_ is now known as rgreening [13:22] ubiquity: evand * r4363 trunk/debian/changelog: LP bug reference. [13:32] ubiquity: evand * r4364 trunk/debian/changelog: releasing version 2.3.19 [13:40] cjwatson, does that look right to you ? http://paste.ubuntu.com/494729/ i'm using a fresh maverick chroot for the autofoo but not ending up with 2010 in the copyrights seems weird [13:46] ogra: please just copy (cp -a to preserve timestamps) the files in from the previous source package [13:46] oh, ok [13:46] the last autotools update was done in a Debian system [13:46] so it won't match exactly, I shouldn't imagine [13:46] ah, that explains it [14:05] cjwatson, sure i'll take a look today [14:06] thanks [14:10] ogra, if you run oem-config as root, you're going to see some weird behaviors in the panel still [14:10] so it would be better if you could set up jasper to create an oem user [14:10] (or some other place that's appropriate) [14:10] superm1, well, i dont want to add users to the preinstalles FS [14:11] what weird behavior beyond sound would that be ? [14:11] (i dont mind disabling the sound applet) [14:11] NM seems to work fine [14:13] well as long as NM works fine i guess that works [14:14] k [14:15] evan already added a fallback in case oem doesnt exist [14:15] so we should be fine again [14:19] is there a problem with ubiquity in today's daily build that installs a system that's not bootable? [14:19] (at least, that's the problem I have on Edubuntu :) ) [14:20] hi ev, it seems that the last ubiquity upload introduced new strings, which gives no time to translators to translate them before today's deadline. May I ask you to fetch translations later? I've filed bug 640482, it would be great if you could comment on that. Thanks! [14:20] Launchpad bug 640482 in ubiquity (Ubuntu) (and 1 other project) "Please download translations for new strings after NonLanguagePackDeadline (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/640482 [14:20] dpm: yes [14:20] I planned to send a mail out to the translation team about this [14:20] I'm happy to upload a new ubiquity with just translations before the LanguagePackDeadline [14:23] The 10.10 amd64 beta ISO doesn't boot on a Macbook Pro. Depending on how I boot "Windows" mode or EFI, I get a black screen w/ the text "Select CD-ROM Boot Type" or a kernel panic, "unable to mount root fs on unknown block". Is this expected behavior? [14:24] ev, cool. Yeah, if you could send that e-mail as well, that would be great [14:24] will do [14:26] smagoun: the intention was for the EFI work not to change behaviour on Macs [14:26] smagoun: I think somebody with the hardware is going to have to debug this, though [14:26] Macs are weird and their behaviour is hard to predict remotely [14:28] cjwatson: "not change behavior" = it's supposed to be broken? I haven't booted from an ISO since Jaunty or so on this machine, so I'm not sure what the expected behavior is these days. [14:29] cjwatson: Is there anything I can do to help debug? The i386 install works fine, not sure if that's useful data or not [14:30] obviously it isn't supposed to be broken :) [14:30] debian-installer: cjwatson * r1364 ubuntu/debian/changelog: releasing version 20100211ubuntu27 [14:30] i386 doesn't have the EFI bits, deliberately [14:30] I don't know how to debug Mac booting [14:31] I know that there are some Macs that refuse to boot from multi-catalog El Torito CD-ROMs due to buggy firmware [14:31] if yours is one of those, i386 may be the only useful fallback [14:32] though that seems odd given that apparently it gives you a choice? what level is the choice at? [14:32] cjwatson: This machine is 3 years old, so certainly might fit into that category. [14:33] cjwatson: Apple's firmware boot chooser gives me a couple options. From the CD I can choose "windows" or "EFI boot". The former dumps me at a prompt that I think is from the firmware, it says "Select CD-ROM boot type". The prompt is a dead-end, the machine doesn't accept input at that point. [14:34] what happens if you press 1 and enter? [14:34] from the prompt [14:34] cjwatson: The "EFI boot" option in the boot chooser gets me to a grub menu that includes "install ubuntu" and "check disc for defects". I choose 'Install Ubuntu", and a few seconds later I get a kernel panic [14:34] http://jowie.com/blog/post/2008/02/24/Select-CD-ROM-Boot-Type-prompt-while-trying-to-boot-from-Vista-x64-DVD-burnt-from-iso-file.aspx may be related [14:35] generally speaking I'm not expecting the EFI option to work on Macs, since AFAIK they're still not UEFI-compliant [14:35] cjwatson: I found that, haven't tried to remaster a disc yet [14:35] ok, I believe that. So I should be trying to get around the "select cd-rom..." prompt. [14:36] the article seems a bit confused and not obviously directly relevant [14:36] obviously we're not using Microsoft's boot sector [14:36] agreed [14:37] it's possible isolinux has a similar problem, though [14:39] though it does seem to at least try to strip off the version [14:40] cjwatson: this seems directly related, not especially promising though (bug is still 'NEW')" https://bugzilla.redhat.com/show_bug.cgi?id=527443 [14:41] bugzilla.redhat.com bug 527443 in syslinux "boot.iso does not boot with MacBook Pro 2,1 or iMac 2006 late" [Medium,New] [14:41] I suppose there's an outside chance that it might be worth trying to get the EFI option to work, although I suspect that will run into video problems later [14:42] yeah - no acceleration, from what I understand. Not sure that's worth it for the user. [14:43] Would be nice if there were a, "sorry, we already know this isn't going to work. Try the i386 CD" type of message from grub. [14:43] no grub [14:43] oh, for efi? um [14:43] not sure that the amount of hacking required to do that is worth it [14:44] and I don't know exactly which models are affected [14:44] agreed. I do know that newer MBP models boot fine from the amd64 ISO, so there's some cutoff of machines from within the last 3 years. [14:44] sounds like a bunch of work [14:45] it might be worth trying remastering with isolinux-debug [14:48] ok [16:09] ubiquity: superm1 * r4365 ubiquity/ (bin/oem-config-remove-gtk debian/changelog): Rework oem-config-gtk to use a mainloop (LP: #637517) [16:12] superm1: maybe the changelog should give credit to Sebastian? [16:12] cjwatson, it does [16:12] assuming you used his code, I didn't check :) [16:12] ah, ok [16:12] it's got his name right above at least [16:12] I assumed you used debcommit so the commit message was identical [16:13] ah, debcommit doesn't show the [ ... ] bit [16:13] ok [16:13] debcommit strips the [ user ] stuff [16:13] yeah [16:19] oh wow, gconftool.set doesn't actually work, and probably hasn't ever [16:20] * ev fixes [16:23] ubiquity: evand * r4366 trunk/ (debian/changelog ubiquity/gconftool.py): Call gconftool-2 --set with --config-source. [16:23] i'm sure it was working before [16:23] you can't use a terminal in the current dailies because of it... [16:24] perhaps it's just due to the way I'm bootstrapping ubiquity [16:24] missing an environmental variable or something like that [16:25] but this does seem correct [16:26] without config source it's supposed to just use the default path, so is there a scenario that the path that should be in use isn't the default? [16:28] hm, not sure [16:28] I'll play with it in a minute [16:28] see if I can get a better answer [16:38] ubiquity: evand * r4367 trunk/ (debian/changelog ubiquity/frontend/gtk_ui.py): [16:38] ubiquity: * Better lock down the installer environment when past partitioning. [16:38] ubiquity: * Lock down the installer environment when at partitioning commit, [16:38] ubiquity: rather than just after automatic partitioning (and not manual [16:38] ubiquity: partitioning) (LP: #640604). [16:50] ubiquity: evand * r4368 trunk/ubiquity/frontend/gtk_ui.py: Add user switching to the list of things to disable in lockdown_environment. [18:47] ubiquity-slideshow-ubuntu: evand * r287 ubiquity-slideshow-ubuntu/ (debian/changelog slideshows/ubuntu/slides/link/general.css): Match the Ubuntu theme's color for the header. [18:53] ubiquity-slideshow-ubuntu: evand * r288 ubiquity-slideshow-ubuntu/ (5 files in 5 dirs): [18:53] ubiquity-slideshow-ubuntu: Update forgotten translation templates. Already has a freeze [18:53] ubiquity-slideshow-ubuntu: exception (LP: #636945).