[00:04] cjwatson: thanks for the clarification [00:05] I now get more of what is happening that I wasn't understanding before [00:58] i *love* the new timezone picker, evand, cjwatson [00:58] \o/ [00:58] hugs around [00:58] picking Chicago out of that mess of tiny dots on Lake Michigan has sucked for a very long time [00:59] haha [01:00] kirkland: You should see the timezone picker on a 5" screen :-) [01:00] StevenK: is the new one better? [01:00] I've had no time to test it, sadly [01:00] It's on my list [01:03] StevenK: it's a milion times better, imo [04:08] the current version of the installer fails for me when trying the manual option [04:09] for partitioning [04:09] is this a known bug? [04:13] cause by the addition of line 1092 in the ubiquity/components/partman.py file in revision 3030 [09:36] ubiquity: evand * r3036 ubiquity/debian/po/ (79 files): debconf-updatepo [10:07] ubiquity: evand * r3038 ubiquity/debian/changelog: releasing version 1.11.10 [10:12] evand: did you look into shtylman's bug above at all? [10:12] 04:08 the current version of the installer fails for me when trying the manual option [10:12] 04:09 for partitioning [10:12] 04:09 is this a known bug? [10:12] 04:13 cause by the addition of line 1092 in the ubiquity/components/partman.py file in revision 3030 [10:12] that's the translation mapping stuff [10:12] negative, I missed that entirely. I'll take a look now though. [10:12] maybe a fallback if the text is not in self.translation_mappings would be in order [10:13] ok [10:23] ubiquity: evand * r3039 ubiquity/ (debian/changelog ubiquity/components/partman.py): [10:23] ubiquity: Check to make sure there is a partman-auto translation to map to before [10:23] ubiquity: assuming there is one. [10:23] cjwatson, shtylman: thanks for bringing that to my attention. [10:45] usb-creator: evand * r76 trunk/po/ (POTFILES.in sv.po usbcreator.pot): update-po [10:50] usb-creator: evand * r77 trunk/debian/changelog: releasing version 0.1.12 [11:05] evand: is there a away you could add create .img to usb-creator? [11:33] cjwatson: I may have asked you this before, I asked Keybuk and he wasn't sure, and I've looked all over for an answer, but do you know if it's possible to handle inserting arbitrary-length lists in gettext? That is %s becomes "a, b, and c" and somehow retains the proper grammar of the language. [11:34] Case in point, "This computer has %s on it.", where %s can be a >= 1 list of operating systems from os-prober. [11:35] davmor2: it's been planned, I just have not gotten around to implementing it. Maybe today, if not it will end up in my PPA hopefully before release. [11:36] evand: Cool :) [11:36] evand: no, gettext has no such facility [11:37] best answer I know of is to frame it so that the list is presented one per line after a colon or something [11:37] hrm, that would make the interface far too long unfortunately [11:37] but thanks for clarifying and ending my chase after a fictitious option [11:38] or "This computer has several operating systems on it" if that's appropriate [11:38] so, in principle, you could have a separate set of templates "%s" "%s and %s" "%s, %s, and %s" etc. and substitute that in; but I don't feel that would work very well [11:39] and you could well run into problems of the plural-forms type [11:39] indeed, I had thought of that as well and share the same concern [11:39] and of course debconf translations are not as expressive as gettext, and in particular don't support plural forms at all [11:39] (we must fix that one of these days) [11:40] hm [11:42] BTW, I'm adding password length checking to user-setup; just want to run it by kees before committing [11:42] it'll need a hopefully trivial tweak in ubiquity [11:43] wonderful [11:45] largely I want to ensure that we can get away with just a length check and not something more elaborate [11:45] partman-crypto just has a length check right now, though [11:47] I did look into adding password strength checking to ubiquity, but the only open source algorithm I could find was cracklib, and that seemed rather deficient (and python-cracklib would need to go into main). [11:51] my concern with cracklib is that it works best once you build a database out of dictionaries, and what are we supposed to do - use every dictionary in the archive? [11:52] it seems like a more sensible thing for sysadmins to apply to their users rather than for the distribution to apply to the sysadmin-like user installing it [11:52] ah, good point [13:56] partman-auto-lvm: cjwatson * r212 ubuntu/debian/ (changelog partman-auto-lvm.templates): [13:56] partman-auto-lvm: Talk about "Amount of volume group" rather than "Amount of disk space", [13:56] partman-auto-lvm: per Frans Pop. [14:05] morning [14:14] good morning [14:20] is all well with the autotranslation stuff? [14:20] indeed, all fixed now, thanks :) [15:25] Hi, I am having some problems installing using netboot. It was working fine but recently it has started failing to download packages successfully. Is there any reason why this might have changed? [15:26] depends how your environment is set up; are you using a release image or one from updates? what kind of mirror are you installing from? [15:27] It is using one of the standard mirrors, I think it is either the gb or ie mirror. How do I tell whether it is using a release image or one from updates? [15:29] you will have downloaded the netboot image from somewhere ... what URL? [15:29] also the exact error message would come in handy [15:32] http://archive.ubuntu.com/ubuntu/dists/intrepid/main/installer-i386/current/images/netboot/netboot.tar.gz <- that is the link I used to get the netboot package [15:33] and it fails to download seemingly random packages (and at one point it couldn't find the Release file) [15:34] hmm. that doesn't sound like any of the problems I was thinking of; it sounds more like some kind of localised network problem to me ... [15:35] yeah, that's what I thought at first but we tested it as best we could and it seems to be working fine for other things [15:35] I'm on a university internet connection so that is pretty much rock solid. [15:37] I'll try replacing some cables and retry them. [15:39] I don't suppose it's possible to get a trace of the network activity in question? [15:44] evand: I finished the partition bar stuff for the kde side [15:45] shtylman: wonderful, do you need me to merge it in or is Riddell going to review it? [15:45] if you feel the need for inspiration, or if you want to make the two match up better let me know and I can change the kde one around some [15:45] Riddell will probly review it first [15:45] ok, cool [15:46] and I need to run it through some more tests [15:46] I'll try to take a peak at it today, otherwise definitely tomorrow [15:46] k, you can run the partition bar script without the installer [15:46] its just in the kde_components folder [15:47] let me know if you want to make the two interfaces resemble each other better [15:48] ok, will do [15:54] evand: http://shtylman.com/stuff/part_bars1.png and http://shtylman.com/stuff/part_bars2.png whenever you feel like it so you don't have to pull the source and all [15:56] ubiquity: evand * r3040 ubiquity/ (5 files in 4 dirs): [15:56] ubiquity: * Explain the automatic partitioning options more clearly. [15:56] ubiquity: * Move the "after" partition bar to beneath the autopartition options. [15:57] very nice! [15:57] I like your resize handle [15:57] :) .. I thought it might be a little more obvious to use than just a bar [15:57] absolutely :) [16:01] The above commit makes the partitioning page quite slow due to some poor interaction with the find_in_os_prober code. I'm on it, I just wanted to land that first. [16:09] usb-creator: evand * r78 trunk/ (5 files in 4 dirs): [16:09] usb-creator: Fix the .desktop item i18n support (LP: #331061). Thanks Timo [16:09] usb-creator: Jyrinki! [16:09] evand: Hey, did you merge in my branch? [16:10] lfaraone: sorry, completely fell out of my mind. Reviewing now. [16:10] evand: thanks. [16:19] every time I do bzr merge I have to specify the path to main ubiquity trunk...is there a way around that so it remembers? [16:19] bzr merge --remember [16:19] and then the next time just 'bzr merge' without args [16:19] cool, thanks [16:49] lfaraone: it needs a bit of work. As that check is made is several places (your branch only accounts for one) we'd factor that out to a function that references a list of acceptable busses and call that whenever needed. I've done this much working off of your original code, but it's not picking up my SD card and I don't have time at the moment to investigate any further. [16:49] Let me pastebin what I have. [16:51] http://paste.ubuntu.com/119738/ [16:55] <_MMA_> Has the ext4 format option been removed in the installers lately? I have some Studio users testing dailies tell me so. I'm about to test myself but I figured it would be faster to ask. [16:57] _MMA_: no, it should be there. [16:58] cjwatson: possibly as it goes through another pc on the way to the internet [16:58] <_MMA_> evand: Hmm... We have pending seed changes but I wouldn't think that would effect the installer. Ill keep going with testing todays disk. [16:59] definitely should be there and does not especially depend on seeds [17:00] <_MMA_> As I thought. [17:12] <_MMA_> cjwatson: It is indeed there. [17:26] ah, my code to not call os-prober over and over again is failing to work properly. [17:30] heh...thats unfortunate [17:30] fixed [17:30] ubiquity: evand * r3041 ubiquity/ (debian/changelog ubiquity/misc.py): Stop calling os-prober multiple times if the first run returns nothing. [17:41] ubiquity: evand * r3042 ubiquity/debian/ubiquity.templates: Update the time zone comment. [17:42] ubiquity: evand * r3043 ubiquity/debian/po/ (79 files): debconf-updatepo [17:53] evand: recommendations of the use of os-prober? [17:59] ubiquity: evand * r3044 ubiquity/debian/changelog: releasing version 1.11.11 [17:59] shtylman: can you elaborate on your question, I'm not sure I understand what you're asking [17:59] when should I use it [17:59] and what will it return to me? [17:59] the os name on a partition if it has one? [17:59] exactly [18:00] and it takes in '/dev/sd?#'? [18:00] yes [18:00] or '=dev=sd?#' [18:00] cool, ok [18:01] cjwatson: Heya, in 8.10 what determines whether server or desktop version will be installed in a netboot? [18:01] shtylman: do bug Riddell to review your stuff soon, as feature freeze is tomorrow at some unknown time :) [18:02] DogWater: if you want a server installation, you need to preseed it [18:02] cjwatson: what I mean is, the desktop iso vs the server iso [18:02] when do you a netboot how do you tell it which edition you want? [18:02] I answered your question [18:02] evand: will do...he has looked at it some already..currently the oxygen people are looking at the partition bar to see what they might want to change [18:02] alrighty [18:02] shtylman: very cool [18:02] but really, the only practical difference is which packages are installed [18:03] evand: I am new to the whole dev process, but a feature freeze...what exactly will that entail? [18:03] the server kernel, and the fact that you get server tasks offered [18:03] shtylman: after that point we can only upload bug fixes [18:03] so no more development on new features can be done [18:03] DogWater: you can put base-installer/kernel/override-image=linux-server on the kernel command line to get the server kernel, and pkgsel/language-pack-patterns= to stop it installing language packs [18:03] I see...so would a change to the partition bar interface be considered a bug? if the actual partition bar interface was already in place? [18:04] DogWater: and then it should offer you task selection so you can do the rest from there [18:04] DogWater: the server CD just has a different set of packages on it and starts with a different preseed file (look for file= in the command line) [18:04] and lets say I don't get around to using your find_in_os_prober to display pretty names...would I be able to change that after the feature freeze? [18:05] shtylman: you can always ask for an exception [18:05] https://wiki.ubuntu.com/FeatureFreeze [18:05] cjwatson: i'm assuming you mean this tasksel tasksel/first multiselect standard [18:05] ahh, gotcha...so basically, I can continue work on things/fix bugs and then propose a patch afterwards? [18:06] and if it is deemed worthy it will be brouht in [18:06] DogWater: no, I didn't mean that [18:06] ahh..that page is helpful [18:07] DogWater: where did that come from, anyway? [18:07] preseed documentation [18:07] what documentation where? [18:07] for package group selection [18:07] from debian [18:07] don't use Debian documentation for preseeding [18:07] use Ubuntu documentation for preseeding Ubuntu [18:07] help.ubuntu.com, look for installation guide [18:08] Debian and Ubuntu are *not the same* and you can't use the same documentation for both, I'm afraid [18:08] okay, i'm not using that in my current preseed, i'm just trying to figure out how to tell it to include the 'server' packages during the install. [18:08] which packages? [18:08] remember that the server CD hardly installs any extra packages by default [18:08] I guess whatever makes the server 'cd' the server cd vs the desktop cd [18:09] the only task it installs by default is the server task, but mostly the point of the server CD is that it offers a number of different tasks and you get to select [18:09] we don't install gnome, that sounds like a 'gnome' thing [18:09] sure, the desktop CD preseeds 'tasksel tasksel/first multiselect ubuntu-desktop' [18:09] that's actually the only thing the desktop CD preseeds at all [18:10] the server CD preseeds 'tasksel tasksel/first multiselect server' but there isn't actually a lot in the server seed. For any realistic system it's better to ask [18:10] we use the 'text' and 'skipx' in our kickstart [18:10] so i dont believe it installs the desktop at all [18:11] neither of those directives have any effect in Ubuntu [18:11] they are literally no-ops [18:11] package selection is controlled by the %packages section (in RH too, as I understand it) [18:11] if you want to install the server task, put @server in the %packages section [18:11] if you want to add a mail server, add @mail-server [18:11] etc. [18:12] https://help.ubuntu.com/8.04/installation-guide/i386/automatic-install.html [18:12] :q [18:12] (oops) [18:48] cjwatson: so...I have hit a roadblock...the debian installer user-setup currently being used is hard coded to the /etc/kde3 directory [18:48] [13:47] what would I need to do to change that? [18:48] [13:47] I have found the user-setup script I need to change...but beyond that.. [18:49] Is it intentional that migratino assistant comes up in ubiquity even if there are no other OSs to migrate settings from? [18:52] migration assistant even [18:53] cjwatson: that just returns couldn't find package @server (by the way) [18:56] cjwatson: and it says this in the documentation: Most package groups. As special cases, the "Ubuntu Desktop" and "Kubuntu Desktop" groups install the standard Ubuntu or Kubuntu desktop systems respectively, and any group name not containing a space (for example, "ubuntu-desktop") causes packages with the corresponding Task: header in the Packages file to be installed. [18:57] so it appears that package groups are missing [18:57] "most" [18:59] So I guess is there a package group preseed command? [19:20] cjwatson: oh by the way earlier when i was talking about the tasksel preseed i assumed there was a server 'task' that i could tell it to install which would take the place of putting @server in %packages since %packages is incomplete [19:49] DogWater: oh, sorry, it's '@ server' not '@server' [19:50] there is a server task, but there's very little in it. You should think of the default server install as a minimal subset [19:51] shtylman: please file a bug, we need to fix that [19:51] ok...I submited a merge request [19:52] through launchpad [19:52] (couple min ago)...seemed like the right thing to do at the time [19:52] cjwatson: ^^ [19:54] ok [19:54] thanks for that, I'll do something with it when it comes through [19:54] k [19:58] cjwatson: i had people complain because i wasn't installing 'server' [19:58] so i guess, whatever [19:59] how do I cross compile a deb package...if thats even possible? [20:00] with difficulty :) [20:00] :( [20:00] well...thats unfortunate [20:00] in full generality, Debian packages are entirely permitted to execute things they build during the build [20:00] you could try dpkg-cross [20:00] totally missed the fact that my virtual machine is 32 bit and my local comp is 64 [20:01] it may work for some packages [20:01] gotcha [20:01] DogWater: the server seed currently consists of: patch screen landscape-common wireless-tools wpasupplicant w3m ubuntu-serverguide [20:02] so not nothing, I guess, just not lots [20:03] cjwatson: apparently dpkg-buildpackage takes a -a flag now [20:03] just fyi [20:04] and it will compile for that architecture [20:04] ah, ok [20:10] cjwatson: hm, is there a group that includes the server kernel? [20:12] cjwatson: i dunno i think 'they' expect 8.04 to say 'lts' or something and 8.10 to say server when its installed. [20:13] not sure why this is such a sticking point [20:31] I know this isn't #mysql-install, but... "apt-get install mysql-server" prompts for an admin pw - how can I script that? (seed the value or ignore it and config the thing later) [20:31] DogWater: no, you need to use the base-installer thing I mentioned above for that [20:31] CarlFK: find out what debconf question it's asking, preseed that in the usual way? [20:32] you can use debconf-set-selections in the installed system [20:32] it may not be preseedable though [20:33] do you know if this coming from apt, or a custom script that's part of the package? http://dev.personnelware.com/carl/a/mysqlpw.png [20:34] I don't need to look at that to know that it doesn't come from apt :) [20:34] heh [20:34] it's mysql-server-5.0's config or postinst script [20:35] both - it asks, and if you leave it blank, it asks again :-/ [20:37] honestly, not something I can really help with - try #ubuntu-server [20:38] no prob - thanks for the pointer. [20:45] back to u-installer... if I have "d-i apt-setup/local1/repository ..." but not a "apt-setup/local0..." (because I #commented it out) it seems to ignore the local1 - any idea how to deal with that? [20:45] rename it to local0 :-) [20:46] I knew you were going to say that [20:46] cdebconf doesn't really have any "iterate over all question names" support that's accessible to clients [20:46] so the only thing that apt-setup can do is to try local0, local1, etc. and keep going until one of them doesn't exist [20:46] Ill just un #comment the other one. won't kill me [21:38] user-setup: cjwatson * r155 ubuntu/ (4 files in 2 dirs): [21:38] user-setup: Add weak password detection (purely length-based for now, matching [21:38] user-setup: partman-crypto). [21:43] user-setup: cjwatson * r156 ubuntu/debian/po/ (63 files): debconf-updatepo [21:45] user-setup: cjwatson * r157 ubuntu/debian/changelog: releasing version 1.23ubuntu12 [22:17] ubiquity: cjwatson * r3045 ubiquity/ (debian/changelog ubiquity/components/usersetup.py): Handle user-setup's new weak password detection. [22:53] oem-config: cjwatson * r607 trunk/ (debian/changelog lib/components/console_setup.py): Only call setxkbmap if DISPLAY is set. [22:55] oem-config: cjwatson * r608 trunk/ (debian/changelog lib/components/user.py): [22:55] oem-config: Don't show the "Who are you?" page just because the auto-login question [22:55] oem-config: is asked and not preseeded (LP: #328281). [23:02] oem-config: cjwatson * r609 trunk/ (65 files in 4 dirs): Handle user-setup's new weak password detection. [23:06] oem-config: cjwatson * r610 trunk/debian/ (61 files in 2 dirs): Add forgotten oem-config/text/network_heading_label template. [23:06] I see nobody but me is testing oem-config on the server CD :-P [23:14] oem-config: cjwatson * r611 trunk/ (65 files in 4 dirs): [23:14] oem-config: Add a "tasks" component (only usable in the debconf frontend at the [23:14] oem-config: moment) that runs tasksel. [23:15] oem-config: cjwatson * r612 trunk/ (debian/changelog scripts/netcfg-wrapper): [23:15] oem-config: Check whether /usr/lib/network-manager/ifblacklist_migrate.sh exists [23:15] oem-config: before running it. [23:38] oem-config: cjwatson * r613 trunk/ (debian/changelog finish-install.d/01oem-config-udeb): [23:38] oem-config: Fix copying of debian-installer/framebuffer to /target in [23:38] oem-config: oem-config-udeb: debconf-dumpdb returns "(null)" as the value if the [23:38] oem-config: question has never been set. [23:39] oem-config: cjwatson * r614 trunk/ (d-i/manifest debian/changelog): Automatic update of included source packages: user-setup 1.23ubuntu12. [23:43] oem-config: cjwatson * r615 trunk/debian/changelog: releasing version 1.54.6 [23:43] ubiquity: cjwatson * r3046 ubiquity/ (d-i/manifest debian/changelog): Automatic update of included source packages: user-setup 1.23ubuntu12. [23:55] ubiquity: cjwatson * r3047 ubiquity/debian/changelog: releasing version 1.11.12