[01:18] evand: Ok from the discussion I had earlier with Mario, it seems the GUI in only-ubiquity is not dropping to the user because its conditional on environment variables that don't exist in only-ubiquity mode. I'm going to have a dig to see if I can come up with a sane fix. [04:14] Argh! Even on the official dailies, Orca is zombifying again!! [05:52] ah, SUDO_UID/GID don't exist when running in only-ubiquity mode. I should've caught that. [05:52] I'll work on a fix in a few minutes. [05:52] evand: Ok no problem. [05:52] or have you already? [05:53] Not yet, just attempting to play around with some ideas, but with orca zombifying, even on the official daily, I'm kinda stuck. [05:54] yeah, I haven't looked yet but it's odd that it started doing that again. Nothing really changed there unless I somehow committed old code back to ubiquity-dm. [05:54] Yeah it is odd. [05:55] The log doesn't show such. I remember trying the daily after you added settings groups to ubiquity-dm, and things worked. [05:55] revision 2580 [05:56] However, changes in the surrounding environment can sometimes be enough to cause it to zombify. [05:56] The next daily may be fine. [05:57] only another hour and a half until that. [05:58] well, two counting build time [05:58] Yeah. [05:59] is r2580 a typo? Ubiquity doesn't have that many revisions yet. [06:01] he probably [06:02] No its correct. I'm looking at ubiquity trunk. [06:02] brb [06:02] curious [06:03] oh whoops, I was on the core-dev branch. [06:09] haha [06:10] I'm happy to simply grab the fix from bzr and test in a custom iso. [07:54] (currently being stalled by my fight with m-a over capb escape, but I'll get the SUDO_* fix in as soon as I'm done here) [08:40] no rush [09:04] I can't find the installation guide from either doc.u.c or help.u.c... [09:04] 7.10 doesn't have the link at all, and the dev version has a broken one [10:46] hm, I'm taking a crack at fixing bug 64147.. seems silly that 16:10 screens are not natively supported [11:40] Hi, we have problems adding a 7.10 CD. Last two lines are E: Sub-process gpgv returned an error code (2) [11:40] W: Signature verification failed for: /cdrom/dists/gutsy/Release.gpg [11:42] We already have 7.04 installed, along with latest updates, and we're just trying to 'add' the CD to the repository. Is that what we should be doing? [14:21] ubiquity: cjwatson * r2596 ubiquity/ (debian/changelog scripts/install.py): [14:21] ubiquity: * Remove packages in the restricted section from the installed system if [14:21] ubiquity: apt-setup/restricted is false. [14:29] Is modifying something in coreutils two weeks before release considered evil? install when used with -D, -o and -g creates the directories as the user it's launched with, rather than the user specified in -o. [14:29] oem-config: cjwatson * r442 oem-config/ (debian/changelog finish-install.d/07oem-config-user): * Fix ownership of /home/oem/Desktop (LP: #209683). [14:29] hahaha [14:29] nevermind then :) [14:29] * cjwatson laughs. Synchronicity [14:30] I just made it do install -d separately [14:30] sorry to step on your toes [14:30] no worries at all [14:30] I'm not sure whether it's a coreutils bug on not [14:30] or not [14:31] maybe send mail to their bug list and ask? [14:31] bug-coreutils@gnu.org [14:31] yeah, I was wondering the same thing. Will do. [14:31] Thanks [14:33] evand: (I'd have left you to it except you unassigned yourself from the bug) [14:33] oem-config: cjwatson * r443 oem-config/ (d-i/manifest debian/changelog): [14:33] oem-config: * Automatic update of included source packages: console-setup 1.21ubuntu5, [14:33] oem-config: tzsetup 1:0.20. [14:35] oem-config: cjwatson * r444 oem-config/debian/changelog: releasing version 1.33 [14:47] oem-config: cjwatson * r445 oem-config/ (configure configure.ac): bump to 1.34 [14:48] One of my pkgsel/include packages is returning unmet dependencies, is there a way I can get more information? I have DEBCONF_DEBUG=5, so I think I'm getting all the possible information in syslog. might it be possible to try to run a command from the busybox shell to try to reinstall the package and perhaps get more verbose output there? [14:49] 'chroot /target apt-get install ' should do [14:49] cjwatson: before I embark on this task, any objection to me modifying debconf-set-selections to include an option to use the escape capability? I can get m-a to work with spaces in the key name via escape, but wubi fails on debconf-set-selection as its not using the capability. [14:49] spaces in the *key* name? [14:49] cjwatson: ok, I tried in-target apt-get install and nothing was output [14:49] urk [14:49] cr3_: don't use in-target by hand [14:49] evand: can we avoid using spaces in question names? I really don't think it's a good idea ... [14:50] cjwatson: cheers man, I managed to get exactly what I needed to know with that chroot command [14:50] cjwatson: I don't think I can remove the Windows username from the keys in question, so my options are to either use spaces or continue to replace them with a colon and have wubi carry that wart. Would you prefer the latter? [14:51] I think it's definitely preferable to replace them with something [14:51] ok, works for me. I'll have Ago make the change and leave m-a as is. [14:51] So, what do we need to store in a variable template? Of course we need a [14:51] name to identify the template. Template names are made up of components [14:51] separated by the character `/' (slash). Each component is limited to [14:52] alphanumerics and `+' `-' `.' (plus, minus, full stop). [14:52] /usr/share/doc/debian-policy/debconf_specification.txt.gz [14:52] in fact many templates also use _ [14:52] hmm, I feel a policy bug coming up; but in any case I think spaces are right out :) [14:52] I can't use _ as it's a valid username character in Windows iirc. [14:52] heh [14:53] so you think it should be allowed in the future or explicitly disallowed? [14:53] _ should be allowed; space should be disallowed [14:53] ah === cr3_ is now known as cr3 [15:07] evand, did you get a chance to follow up on the copying of the locales to the destination install? [15:10] superm1: debian-installer/locale not getting copied when using oem-config? [15:10] yeah [15:11] r2594 in ubiquity [15:12] ah great. I hadn't updated my local bzr in a few days [15:14] as always, let me know if you come across any other issues [15:22] well the other thing that came to mind - currently that last step, where is calculates packages to remove - it takes forever on the dvd image. I'm guessing because so much is preinstalled in that live fs? [15:24] that would be my guess as well [15:26] probably not for this 8.04 time frame, but for 8.10, maybe something to allow seeding the exact packages to remove instead of calculating them could be an improvement [15:27] hrmm, that sounds fragile. But I'm sure we can profile that phase of the install for Intrepid [15:27] profile and optimize* [15:27] yeah that's what i was thinking [15:33] evand: bug 203292 is obsolete now, isn't it? [15:33] Launchpad bug 203292 in oem-config "Freeze exception: zoommap changes port to oem-config" [Undecided,Confirmed] https://launchpad.net/bugs/203292 [15:34] indeed, marked as fixed [15:34] thanks for catching that [16:31] choose-mirror: cjwatson * r585 ubuntu/ (Mirrors.masterlist.ubuntu choose-mirror.c debian/changelog): * Use ports.ubuntu.com for all countries (LP: #176672). [16:39] migration-assistant: evand * r74 migration-assistant/ (debian/changelog ma-apply ma-ask): [16:39] migration-assistant: * Use + instead of : for a replacement character as it is explicitly [16:39] migration-assistant: allowed by debconf policy. [16:40] choose-mirror: cjwatson * r586 ubuntu/debian/changelog: releasing version 2.19ubuntu4 [16:51] ubiquity: evand * r2597 ubiquity/ (3 files in 3 dirs): [16:51] ubiquity: * Use + instead of : for a replacement character for m-a as it is [16:51] ubiquity: explicitly allowed by debconf policy. [16:58] at what point in the preseed file could I prepare the pre-pkgsel.d directory? [16:58] I guess preseed/early_command would be appropriate [17:02] yes [17:04] after /target/etc/apt/sources.list.apt-setup is moved to /target/etc/apt/sources.list, should I run apt-get update? if so, in-target or chroot if running from a script under pre-pkgsel.d? [17:05] pkgsel should take care of that [17:13] would it make sense to wget a udeb in the early_command and install that? could this be cleaner in the long term than either building a script in early_command or wgeting a script? [17:16] that can work, but why not use anna/choose_modules=? [17:16] oh, well, only if the udeb in question was in the Ubuntu archive [17:17] cjwatson: I could specify it under local0, that wouldn't it. the content of the udeb would then generate a script under pre-pkgsel.d to do the sources.list thing [17:17] err, "that would do it" [17:17] not for udebs, no [17:18] wget and udpkg -i is probably better [17:18] cjwatson: excellent, thanks, you're saving me lots of trial and error time :) [17:42] for only-ubiquity and oem-config, I think it might be sensible to disable the minimize action [17:43] people might get a bit confused if they accidentally minimize it and don't know how to get it back [17:44] already done for ubiquity [17:44] ah good [18:04] evand, would you be opposed to adding a setterm -blank 0 to ubiquity's init script? setterm isn't available in the initramfs, but it is in the live filesystem [18:05] not at all [18:05] okay i'll add it in [18:07] much appreciated [18:09] ubiquity: superm1 * r2598 ubiquity/debian/ (changelog init): disable console blanking in Ubiquity init script [19:11] oem-config: evand * r446 oem-config/ (3 files in 3 dirs): * Port hiding the minimize button on the main window from Ubiquity. [21:36] Hi, I need help preseeding ubiquity [21:37] someone alive? [21:42] console-setup: cjwatson * r63 ubuntu/ (Keyboard/ckbcomp debian/changelog): * Handle some special cases for the KPDL key (LP: #189529). [21:43] mariodebian: about to go and spend time with my wife, but please ask your question and somebody will get back to you when they're around [21:43] I ask in mailing list last week [21:43] indeed, I'm here [21:43] I will aks today again, thanks !!! [21:44] evands, can you have some time now? [21:44] sure [21:44] oks [21:44] I don't speak English very well but I try to explain better I can [21:45] I don't see a mail from you on ubuntu-installer@ [21:45] perhaps you asked on some other list [21:45] https://lists.ubuntu.com/archives/ubuntu-installer/2008-March/000171.html [21:46] better read this mail [21:48] ah, I missed that [21:48] replying now [21:48] thanks ;) [21:55] another questin appending this mail [21:55] i have preseed some values but some text on ubiquity appears in English and some in Spanish [21:56] for example summary text is in english and some parted strings [21:56] ubiquity.template is full translated [22:03] evand, perusing components/console-setup.py, it appears the proper way to preseed a keyboard is not console-setup/layout and console-setup/variant, but rather console-setup/layoutcode and console-setup/variantcode correct? [22:10] I'm installing packages from pkgsel/include and getting the error: The following packages cannot be authenticated! [22:11] the reason is that I created the file /target/etc/apt/sources.list.apt-setup containing my own repository which is somehow not authenticated [22:11] I guess I need to find a way to authenticate it by retrieving the key file on my site [22:14] mario_limonciell: correct, but I'm investigating why preseeding those isn't working for ubiquity. [22:15] oh it's not working :S. I wasn't sure :) [22:24] mariodebian: I have to run out for a few hours, I'll finish responding to this when I get back [22:24] oks thanks [22:51] cr3: there's apt-setup/local0/key or some such [22:51] was documented in the installation guide last I checked [22:52] cjwatson: yeah, that's what I used to do, but now I seem to need to run apt-key because the repository is defined in sources.list.apt-setup by a pre-pkgsel.d script. no worries, I should have confirmation shortly if it worked. [23:17] cjwatson: worked like a charm!