[00:00] well, yes, that should be d-i, but again that ought not to break console-setup or locales in any way [00:00] and indeed the standard task is installed by default even if you don't preseed it [00:00] I've reverted to the wrong version and doing another installation. [00:01] Let's see if the keyboard mapping is broken again. [00:01] It should be. [00:05] I see lang-pack-en-base being configured so that should be sufficient. [00:08] Oh dear! Reverted to the previously broken setup and it's working. Very strange. [00:08] I did over a dozen broken installations before and now I have no idea what I did to fix it. [00:09] Nor do I have any idea of how to break it again. [00:09] That's very disturbing. [00:09] I have the preseed file open in Kate so all I had to do was Undo the changes I had made before to "fix" the problem. [00:10] cjwatson: ^^ [00:28] To further complicate matters, I have to install language support for French, Spanish, and Mandarin but make en_CA.UTF-8 the default. [00:29] The school in which these machines are going to be used is a bilingual (English/French) school and offers Mandarin or Spanish for third language. [00:30] I wonder if the "fix" had anything to do with my apt-cacher cache being primed. [00:36] the standard way to do the latter would be 'd-i localechooser/supported-locales string fr_FR.UTF-8, es_ES.UTF-8, zh_CN.UTF-8', in case you don't already have that [00:37] err, multiselect not string [00:37] (not that that's hugely important) [00:42] cjwatson: Is the reason you didn't list en_CA in supported-locales because it's the default? [00:42] mistake :) [00:43] ah, ok [00:43] So how is the default set? [00:43] debian-installer/locale (or locale= on the command line) [00:43] I can manage in three out of four of those languages but I have a bit of a problem with zh_CN. :) [00:44] pkgsel is probably buggy for not making sure the langpack corresponding to debian-installer/locale is always installed [00:44] Should I explicitly specify the English langpack then? [00:45] d-i localechooser/supported-locales multiselect en_CA.UTF-8, fr_FR.UTF-8, es_ES.UTF-8, zh_CN.UTF-8 [00:45] localechooser/supported-locales is used to decide which language packs to install [00:45] yes, that's what I have [00:45] (if it's set; otherwise pkgsel falls back to debian-installer/locale) [00:45] Oh, so it's not necessary to explicitly specify the langpack once I have supported-locales? [00:45] that's correct [00:54] casper: TheMuso * r551 casper/ (3 files in 3 dirs): (log message trimmed) [00:54] casper: * scripts/casper-bottom/30accessibility & [00:54] casper: ubiquity-hooks/30accessibility: [00:54] casper: - Revert to using gconf keys for starting orca, as this is now what [00:54] casper: orca does when the user chooses to automatically start orca from Orca's [00:54] casper: preferences window. [00:54] casper: - Explicitly set orca as the program to start up for magnification, [00:54] if nobody has any more to do with casper, I'll release/upload. [01:16] I don't [01:19] ok will wait till LP is back up again and push. [01:24] casper: TheMuso * r552 casper/debian/changelog: releasing version 1.148 [01:27] TheMuso: hmm, it looks like evand had already released a 1.148 earlier but forgotten to commit the release [01:28] TheMuso: see bug 274076 - I guess you'll get a rejection [01:28] cjwatson: oh... [01:28] * TheMuso looks. [01:29] * TheMuso sighs. Right, I'll fix the changelog. [01:32] casper: TheMuso * r553 casper/debian/changelog: [01:32] casper: * Move my changes to a new release version, since 1.148 was uploaded without [01:32] casper: being officially marked as released in the bzr branch. [01:33] casper: TheMuso * r554 casper/debian/changelog: releasing version 1.149 [01:57] One more thing to preseed: I just did a complete installation of KDE and various packages, all preseed. On first boot, the X server locks the machine up hard. E.g. the NumLock key has no effect. I can't ssh into the host. All I can do is a hard power-off and power-on. [01:58] How would I preseed the X configuration? I just used the default configuration provided by the example preseed file. [01:59] I'm rebooting the machine in recovery mode so that I can run xfix. [01:59] I've found this to be a common problem with K/Ubuntu, btw. [02:01] xfix did the trick but I'd rather not have to go through that step for each of the ~ 50 machines in the school. [02:02] K/Ubuntu's monitor detection mechanism doesn't seem to be as good as Mandriva's, for example, because the screen resolution is never quite right. [02:02] It never seems to detect the monitor either. I just get a generic monitor. [02:02] I'm using fairly old ViewSonic and Dell units that are detected first time, every time by Mandriva. [02:02] ... or Fedora. [02:03] E.g. my screen resolution is set to 1280x768, which is some weird mode for a conventional 15" ViewSonic CRT. [06:02] xivulon: done [06:02] TheMuso: sorry about that, I normally use bound branches but didn't have that one set up as such. [06:03] evand: np. [07:57] evand: ping [07:58] davmor2: pong [08:00] evand: Morning Dude. Alt upgrade wasn't it? and will it need internet access or do you want it removing? [08:00] alt upgrade? I don't follow. [08:00] good morning [08:02] oh was it mvo who asked then :-/ gets confused now. Someone asked me yesterday if I could do a cd upgrade today I thought it was you. [08:02] xivulon, the network manager fix should be in the archive now, so wubi/mythbuntu frontend should be sane [08:02] xivulon, if I was to do some more testing with it, what's the latest .exe I want to try from? [08:03] davmor2: I asked if you could re-test a Kubuntu desktop install and see if you're still not seeing the finished dialog. [08:04] yes done seems okay but I'll double check today to be certain :) [08:05] evand: the partitioner ui now is more sane :) and I'll check out manual again today too and check that that fix is in :) [08:06] hooray [08:52] usb-creator: evand * r39 usb-creator/ (bin/usb-creator setup.py): Bump to 0.1.8 [09:01] usb-creator: evand * r40 usb-creator/debian/changelog: releasing version 0.1.8 [09:57] ubiquity: evand * r2907 ubiquity/ (debian/changelog scripts/install.py): Remove broken CD-ROM entries from fstab after install (LP: #150872). [09:57] cjwatson: ^ can you review the above when you have a moment. [11:17] superm1 you need to use wubi rev 512 [11:19] davmor2 can you please test the latest wubi, the latest ISO should hopefully have all the fixes in (assuming wubi rev 512) [11:20] Would someone mind looking at lp:~persia/casper/trunk ? It's not critical, but I'd like to get it included in intrepid. [11:27] evand: the only comment I have is was it intentional to discard compatibility with python 2.4? [11:27] evand: otherwise I think it should be fine [11:28] evand: try: except: finally: is new in python 2.5; in earlier versions you had to use constructions like this instead: [11:28] try: [11:28] try: [11:28] ... [11:28] (oops, pretend that ... was indented again) [11:28] except FooException: [11:28] ... [11:28] finally: [11:28] ... [11:29] evand: ah, interesting, you have the same construction in generate_blacklist - I think we should probably fix that, our dependencies still say python 2.4 and I suspect there are a few backporting maniacs [14:03] ah indeed. I was aware that finally was 2.5, but I didn't think we cared :). Noted, I'll fix that. [15:30] persia: merged [15:30] cjwatson, Thank you. Mostly cosmetic, but should improve the experience until I can sort out supporting multiple users. [15:31] persia: bet you didn't test it though ;-) [15:31] I just fixed a syntax error [15:33] Yeah :/ I need to sort out how to build images properly locally. My test involved applying a patch and updating the initramfs in a chroot, and I suppose I messed up the timing of patch application. [15:33] Thanks for fixing it :) [15:34] Ah. s/:/;/ again. [15:57] Hi! Is the source of the netinst cd images available somewhere? Or instructions about how could I build such a CD image? [16:01] abli: please install the installation-guide-i386 package and read through that (/usr/share/doc/installation-guide-i386/) [16:02] Ubuntu has no such thing as a netinst CD image; that's a term with a very specific meaning in Debian [16:02] we have a netboot CD image, which may be what you are thinking of; it's built by the debian-installer source package [16:03] ah, yes, that was what I was looking for. [16:08] is it possible to run the text-mode installer (that the netboot cd image runs) from a live cd (maybe a custom live cd)? or how much work would be to create a custom version of that installer to be able to run like this? (I assume this is already how it runs: the CD image boots some minimal system and the installer runs in that system. How difficult would it be to make it run in a more complete system?) I am basically looking for a text-mode version [16:08] of the default "install from live cd with graphical installer"; the idea is that I want to be able to run the installer in 'screen' [16:08] as a guideline, ubiquity took approximately six man-months of development effort to become even barely acceptable [16:09] so no, it's not trivial to just run d-i in a live session [16:09] that's what ubiquity is for [16:09] but ubiquitiy is gui-only, right? [16:09] d-i expects to own the system entirely, not to be running inside a normal init-based environment [16:09] not inherently; it has multiple frontends [16:10] there isn't one at the moment, but you could write a text frontend [16:10] I mean, you can try to do it with d-i, I just predict you will run into some interesting problems :-) [16:11] and you'd have to add something like Debian's live-installer component to copy the live filesystem across, which has never been tested on Ubuntu and may not interact with casper very well [16:11] Debian have been trying to do this, I don't really like the approach though [16:13] Ok. Then what is your suggestion for implementing the following: I would like to be able to boot from a CD (or USB), format and partition disks if needed, by hand, then launch the installer to in a screen session, then be able to attach to that screen session over ssh. Write a text frontend for ubiquity? [16:14] honestly, I'm not sure which of (d-i + porting live-installer + debugging undefined set of issues) or (writing a text frontend for ubiquity + figuring out how to make it look reasonable) would be easier. :-) [16:14] do whichever you prefer [16:15] Ok. I'll think a bit more about this. Thanks for your suggestions [16:16] sorry to be brief but I'm trying to figure out intrepid bugs at the moment :) [16:32] oem-config: cjwatson * r543 oem-config/ (debian/changelog lib/components/language.py): [16:32] oem-config: Reset debian-installer/country if the language is changed, as otherwise [16:32] oem-config: we can end up defaulting to an inappropriate country for the language. [16:34] oem-config: cjwatson * r544 oem-config/ (debian/changelog lib/components/timezone.py): [16:34] oem-config: Stop defaulting to the timezone from /etc/timezone or /etc/localtime, as [16:34] oem-config: this has a nasty habit of overriding the usual defaults for countries [16:34] oem-config: with multiple timezone choices (LP: #283861). [16:43] before you release that, would you mind if I squeezed an update to zoommap.py from ubiquity in? [16:57] not at all [16:57] I was waiting for a translations export anyway [16:57] oem-config: cjwatson * r545 oem-config/ (configure configure.ac): bump to 1.52 [17:00] evand can you please double check the URLs in wubi/data/isolist.ini? [17:03] Xubuntu is missing the current symlink [17:03] Mythbuntu throws a 404 [17:03] But Ubuntu and Kubuntu look ok (for the fallback, obviously the first link wont work with the final 8.10 images not being up) [17:04] if mythbuntu is throwing a 404 still, xivulon can you come into #ubuntu-mythtv-dev and bug tgm4883 to fix that? [17:06] I'm convinced it's near-impossible to get a working DVD image out :/ [17:07] oem-config: evand * r546 oem-config/ (debian/changelog lib/zoommap.py): [17:07] oem-config: * Pull timezone map changes from ubiquity: [17:07] oem-config: - Iterate through a list of nearby timezones on click, rather than [17:07] oem-config: selecting the absolute closest timezone to the pointer. [17:07] oem-config: - Stop scrolling the timezone map once the mouse is outside its [17:07] oem-config: boundaries (LP: 251231). [17:13] evand, tgm4883 just fixed something that should have helped for the 404 on mythbuntu, can you try once more [17:14] evand, if it's not working, can you give tgm4883_laptop a little more details/context which part about it doesn't work? [17:15] it's not, I'll message him [17:15] frick, it's not going to work [17:15] er oh [17:15] here you are [17:15] metalink=http://www.mythbuntu.org/releases/mythbuntu-8.10-desktop-amd64.metalink [17:15] metalink2=http://www.mythbuntu.org/devel/dailies/current/intrepid-desktop-amd64.metalink [17:15] is what xivulon has in data/isolist.ini [17:16] bah [17:16] let me fix that then real quick, it should still fail, but much later [17:17] ok, now that i've got both metalink files there [17:18] that part should work, but IIRC, the iso download will fail because the daily iso isn't where it's supposed to be [17:18] cjwatson: have you given any thought to the problem of the live filesystems rarely getting regenerated on account of broken packages? Nag emails on failure to the appropriate package maintainer (which fails on ubuntu-devel@) has been the best idea I could come up with. [17:19] not really; the way live filesystems are lashed up in cdimage right now makes proper notifications a bit tricky [17:20] hmm, indeed [17:20] we should maybe be a bit more aggressive about failing CD builds when the livefs fails [17:20] that would punt the problem up one more layer and make it easier to see [17:20] Perhaps it's just a matter of being more aggressive about archive-consistency cleanup generally? [17:21] well, yes, but it does help when we're notified quickly of problems [17:22] Indeed, I think failing CD builds would be a good start. [17:25] Is this is worth a proper discussion at UDS? I suspect the fact that we haven't had a working DVD since the 7th of October is causing a lot of headaches. [17:25] err s/working/updated/ [17:33] * cjwatson adds "proper notifications on live filesystem build failures" to his UDS list [17:33] I think it might be a bit light for a full session though, unless there's more similar stuff [17:34] evand, is wubi still failing on the mythbuntu installs?\ [17:34] indeed, perhaps we (or I, given the circumstances) could just gather interested parties during lunch or talk about it as part of one of the morning sessions. [17:35] tgm4883_laptop: I've never tried a Wubi mythbuntu install. My Windows VMs have been broken for ages. [17:35] But that link still doesn't wokr. [17:35] work* [17:35] ah I see it now [17:35] yea the links wrong [17:35] Does it need a UDS session? I'd think a mail to ubuntu-devel@ ought be enough, just to see if anyone finds value in updating the images when the livefs isn't updated. [17:35] well [17:36] it's kinda wrong I guess, as the naming doesn't follow the ubuntu daily name [17:36] persia: good point, I'd be satisfied with that [17:36] should be http://www.mythbuntu.org/devel/dailies/current/mythbuntu-desktop-amd64.metalink [17:36] not http://www.mythbuntu.org/devel/dailies/current/intrepid-desktop-amd64.metalink [17:36] tgm4883_laptop: noted, will change [17:36] persia: or a bug on ubuntu-cdimage ... [17:36] ok, thanks [17:37] evand, wait, thats wrong too. http://www.mythbuntu.org/devel/dailies/current/mythbuntu-8.10-desktop-amd64.metalink [17:37] sorry about that [17:38] cjwatson, That works too, although I'd like ubuntu-devel@ would have a wider audience. [17:38] should be that for i386 too [17:38] persia: *nod* [17:39] evand, http://www.mythbuntu.org/devel/dailies/current/mythbuntu-8.10-desktop-amd64.metalink http://www.mythbuntu.org/devel/dailies/current/mythbuntu-8.10-desktop-i386.metalink and in case it matters http://mythbuntu.org/devel/dailies/current/mythbuntu-8.10-metalink.md5sum [17:45] ok, I'll start a thread then [18:02] evand Xubuntu doesn't seem to have a "current" symlink going at the moment, not much I can do myself [18:02] mythbuntu seems to be working though, unless they fixed it after your comment [18:02] most important is that the first metalink for each distro is correct [18:03] no current symlink> which means CD builds have failed for days at a time [18:04] hmm, something bizarre there with mirror syncing, I'll look later [18:10] xivulon: are you sure mythbuntu is working? Both links in isolist.ini for it are 404s [18:13] ah I was looking only at the dir of the URL http://www.mythbuntu.org/devel/dailies/current the actual file name is different, I was using the standard naming convention intrepid-desktop-i386.metalink [18:14] ah, well I've updated the metalinks for mythbuntu [18:14] and will put a new wubi up [18:15] ok, the other route is to change the daily iso names on mythbuntu servers to match ubuntu naming convention (not that I am a big fan of it...) [18:17] new wubi is up [18:18] thx [19:02] artwork suggested to remove the shadow in wubi graphics, can I add that in? [19:03] xivulon: I have no objections, but you should probably take it to the release and documentation teams. [19:04] if I can find the svg will do so [19:07] I am wondering how syncio in ntfs affects swap performance, does caching have a role in swap file writes? [19:15] xivulon, i got someone who tried doing a mythbuntu/wubi with the latest daily of mythbuntu and wubi 5.12, it's bailing out on grub4dos for him though. i'm not really sure how to help debug it [19:15] would you be able to come take a look? [19:24] superm1 have to go now, can you tell him to post in the wubi forum? [19:24] http://ubuntuforums.org/forumdisplay.php?f=234 [19:24] xivulon, sure [19:26] console-setup: cjwatson * r78 ubuntu/debian/ (16 files in 2 dirs): Run debconf-updatepo to update translation files. [19:38] oem-config: cjwatson * r547 oem-config/debian/ (18 files in 2 dirs): Update translations from Launchpad. [19:40] user-setup: cjwatson * r123 ubuntu/debian/ (13 files in 2 dirs): Update translations from Launchpad. [19:41] user-setup: cjwatson * r124 ubuntu/debian/changelog: correct changelog [19:43] debian-installer: cjwatson * r975 ubuntu/ (33 files in 2 dirs): Update help text translations from Launchpad. [19:43] user-setup: cjwatson * r125 ubuntu/debian/changelog: releasing version 1.20ubuntu10 [19:45] oem-config: cjwatson * r548 oem-config/ (d-i/manifest debian/changelog): Automatic update of included source packages: user-setup 1.20ubuntu10. [19:46] oem-config: cjwatson * r549 oem-config/debian/changelog: fix bug-closing syntax [19:49] oem-config: cjwatson * r550 oem-config/debian/changelog: releasing version 1.52 [20:03] debian-installer: cjwatson * r976 ubuntu/ (15 files in 5 dirs): [20:03] debian-installer: Use unversioned nic-firmware and scsi-firmware packages instead of the [20:03] debian-installer: old versioned ones (LP: #220917). [20:11] debian-installer: cjwatson * r977 ubuntu/debian/changelog: releasing version 20080522ubuntu20 === ago_ is now known as xivulon_ [20:40] cjwatson, re the X freeze, I do not end up with much of an xorg.conf at all... [20:41] http://paste.ubuntu.com/58489/ [20:41] is this expected/known or do I need to dig further? [20:42] that's normal nowadays [20:42] I suggest asking #ubuntu-x for help anyway [20:42] but nowadays X is meant to detect most things itself [20:43] * xivulon_ feels old [20:44] I'll see if I can reproduce in vm [21:03] usb-creator: evand * r41 usb-creator/ (debian/changelog scripts/install.py): Write syslinux configuration to text.cfg as well. [21:59] X in vm is perfectly fine [22:10] in a normal installation what is the default grub timeout? [22:11] I get 10 secs in wubi, seems a bit high [22:29] xivulon_: 3 seconds if no other operating systems are installed, 10 if they are [22:29] I think that makes sense in general, it's not ridiculously long but it gives people a chance to think [22:36] I see, of course windows will be detected that is why it is 10 secs, even though in wubi case that is going to be the second boot menu [22:36] cjwatson: 275233 was already targeted afaik, at least I remember seeing it in the list of bugs in the meeting on Wednesday. [22:37] since ntldr is used to choose between ubuntu and wubi, menu.lst is mostly used to choose a different kernel or rescue mode [22:37] s/between ubuntu/between windows/ [22:37] yeah, I'm not honestly sure it's worth the complexity to spot that though [22:37] true [22:38] woops meant for devel but oh well. === TheMuso_ is now known as TheMuso [22:38] TheMuso: no, I think it was a different bug [22:38] on pulseaudio rather than libcanberra [22:38] cjwatson: Right. [22:39] unless it got reassigned but I don't see that from the activity log [23:08] debian-installer: cjwatson * r978 ubuntu/ (build/config/lpia.cfg debian/changelog): Move lpia to 2.6.27-4 kernels. [23:31] debian-installer: cjwatson * r979 ubuntu/debian/changelog: releasing version 20080522ubuntu21