/srv/irclogs.ubuntu.com/2008/04/08/#ubuntu-installer.txt

CIA-1netcfg: cjwatson * r622 ubuntu/ (3 files in 2 dirs):00:05
CIA-1netcfg: * Support preseeding netcfg/network-manager to false to suppress00:05
CIA-1netcfg:  adjustment of /etc/network/interfaces for network-manager (thanks,00:05
CIA-1netcfg:  Andrew Pollock).00:05
CIA-1netcfg: cjwatson * r623 ubuntu/debian/changelog: releasing version 1.40ubuntu500:06
xivulonbdmurray: any luck?00:19
bdmurrayThey all ran fine afaict00:21
xivulonand after that was / rw?00:22
xivulonin particular after chkroot.sh00:23
bdmurrayit seems to be yes00:28
xivulonhmm you can try the scripts in rc2.d ...but I am getting out of ideas :(00:30
CIA-1ubiquity: superm1 * r2606 ubiquity/ (debian/changelog desktop/ubiquity-mythbuntu.png): update mythbuntu ubiquity icon01:31
CIA-1ubiquity: evand * r2607 ubiquity/ (d-i/manifest debian/changelog):07:00
CIA-1ubiquity: * Automatic update of included source packages: apt-setup07:00
CIA-1ubiquity:  1:0.31ubuntu7, choose-mirror 2.19ubuntu5, clock-setup 0.92ubuntu3,07:00
CIA-1ubiquity:  console-setup 1.21ubuntu6, localechooser 1.42ubuntu4, migration-07:00
CIA-1ubiquity:  assistant 0.6.1, partman-auto 73ubuntu7, tzsetup 1:0.20.07:00
evandLaunchpad seems to be failing in sending me the link to download the PO.  I tried a few times and waited quite a bit, so I'll have to talk to the devs later today as I don't have access to my own mail logs (though LP bugmail is coming in just fine).07:01
CIA-1ubiquity: evand * r2608 ubiquity/debian/changelog: releasing version 1.8.207:31
CIA-1ubiquity: evand * r2609 ubiquity/ (configure configure.ac): Bump to 1.8.307:34
Siliciumhow i can create a preseed line to popup a message like "hello world" or so08:18
grrrreg_matin09:09
joshksoir09:09
=== grrrreg_ is now known as grrrreg
cjwatsonMirv: for those extra couple of strings, I was thinking maybe we could make them translatable for 8.04, and then actually have updated translations for them for 8.04.110:28
cjwatsonat this point I don't think I can realistically get new translatable strings in for 8.0410:29
Mirvcjwatson: ok, that would be fine.10:32
CIA-1debian-installer-utils: cjwatson * r652 ubuntu/ (debian/changelog list-devices):11:19
CIA-1debian-installer-utils: * Backport from trunk:11:19
CIA-1debian-installer-utils:  - list-devices: udev 117 merged all udev tools into a single binary11:19
CIA-1debian-installer-utils:  called udevadm. Check for this and use it instead of udevinfo if11:19
CIA-1debian-installer-utils:  available.11:19
CIA-1debian-installer-utils: cjwatson * r653 ubuntu/ (3 files in 3 dirs):11:25
CIA-1debian-installer-utils: * Backport from trunk (Frans Pop):11:25
CIA-1debian-installer-utils:  - user-params: also exclude debconf variables with '?=' (unset seen11:25
CIA-1debian-installer-utils:  flag). Update testset to catch that. Closes: #470892.11:25
xivuloncjwatson, evand, I summarized the results of yesterday test runs for #204133 (courtesy of bdmurray) here: https://bugs.launchpad.net/wubi/+bug/204133/comments/1111:26
ubotuLaunchpad bug 204133 in wubi "wubi install unusable - Buffer I/O error on device loop0" [High,Confirmed]11:26
xivulonI am officially out of ideas...11:26
CIA-1debian-installer-utils: cjwatson * r654 ubuntu/debian/changelog: releasing version 1.50ubuntu311:29
rajuhow i knew my SDcard is connected with my laptop11:46
cjwatsonraju: I'm afraid I can't understand your question11:48
cjwatsoncould you rephrase?11:48
rajucjwatson, sorry i am hving smart card that i want to use11:50
cjwatsonyou're not trying to install from it, are you?11:50
GreatMandrillhi. I am following the instructions from the wiki to create a custom install CD and i have trouble with the customized ubuntu-keyring package11:51
rajucjwatson, i am not trying to install from11:53
rajucjwatson, tell me the package to install & detect the smart card11:53
GreatMandrillthe debootstrap process doesn't add my customs keys bz "It is not signed with a master key". Any pointers plz?11:54
rajucjwatson, i am using Hp tc4200 model laptop & i nstall ubuntu in that11:55
rajucjwatson, in that one smart card is coming , i want to detect it , how?11:55
cjwatsonraju: unless this is actually preventing you from installing Ubuntu, you're in the wrong place; I suggest #ubuntu or #ubuntu-kernel11:56
cjwatsonGreatMandrill: hmm, give me a minute11:57
cjwatsonGreatMandrill: exactly which wiki page are you following, so that I can compare?11:57
rajucjwatson, ok ., can you give a link about smart card detecting in ubuntu11:58
GreatMandrillcjwatson: https://help.ubuntu.com/community/InstallCDCustomization?action=show&redirect=InstallCDCustomizationHowTo#head-c2ca6d4f437c6eb289c74da5e2de3d07e0e7ec6c11:59
cjwatsonraju: no, I'm afraid not12:01
cjwatsonraju: it's not my speciality, nor is it the speciality of this channel12:01
cjwatsonwe deal with initial installation of Ubuntu, which is a quite complex enough subject by itself :-)12:01
rajucjwatson, you are broadmind to help otheres12:01
cjwatsonGreatMandrill: hmm, I think I see the problem but want a double-check. Let me drag mvo in here12:03
GreatMandrillok, thanks. What is your theory?12:03
cjwatsonthat keys should not be required to be signed with the master key unless they are being updated over the network12:04
mvohello12:05
cjwatsonmvo: hi12:05
cjwatsonmvo: GreatMandrill reported an issue following http://help.ubuntu.com/community/InstallCDCustomization, specifically in producing a custom version of ubuntu-keyring12:05
cjwatsonmvo: apt-key requires all keys to be signed by the master key, even if they're not being fetched over the network12:06
mvoright. I think at some pont this was inentional, but I see that its not a good idea for custom CDs12:07
cjwatsonoh, hmm, there was a reason for that wasn't there12:08
cjwatsonI wanted you in here in case I was on crack :)12:08
* cjwatson re-reads the archive key disaster recovery plan12:08
mvoI think its just there on apt-key update, it should be fine if apt-key add is used. but I'm not sure if that is good enough for custom CDs12:08
cjwatsonlet's set custom CDs aside for a moment12:09
cjwatsonif the archive key is compromised, then we can't trust the archive key to authenticate new archive keys on its own12:09
cjwatsonthe master key needs to be involved to ensure that an attacker can't fool you into believing your system has been re-secured when it hasn't12:09
cjwatsondamn12:09
cjwatsonmvo: sorry, I think I might be wasting your time12:10
cjwatson'apt-key add' indeed doesn't check against the master key, but ubuntu-keyring.postinst doesn't use that12:10
cjwatsonGreatMandrill: I suggest skipping the ubuntu-keyring changes and generating an unsigned CD12:11
mvohm, its a tricky problem. we have apt-key net-update (that is run with the regular apt cron job) so that we fetch the keys from the archive without the need for a deb package12:11
cjwatsonGreatMandrill: we don't test that very often, but it actually ought to work, and it would make the job a lot easier12:11
GreatMandrillso, I just skip the Release.gpg file?12:11
cjwatsonmvo: I think on reflection it's an unacceptable risk to skip that check12:11
* mvo nods12:11
cjwatsonsorry, should have thought about it in more detail before dragging you in12:11
cjwatsonGreatMandrill: right12:11
cjwatsonGreatMandrill: it used to be needed12:12
mvono problem cjwatson, happy to help/discuss it with you12:12
cjwatsonI have a feeling that the instructions simply haven't been updated since12:12
cjwatsonmvo: custom CDs are definitely tricky in the event that you want them to be signed12:12
cjwatsonmvo: we could fix the instructions by having people add their custom key to both the master keyring and the archive keyring12:13
cjwatsonmvo: couldn't we?12:13
cjwatson(on the assumption that the key has a self-sig)12:13
mvoyes, that should work12:13
cjwatson    #   all keys that are exported and have the name12:13
cjwatson    #   "Ubuntu Archive Automatic Signing Key" must have a valid signature12:13
cjwatson    #   from a key in the ubuntu-master-keyring12:13
mvoI don't think we enfore the name checking12:14
cjwatsonmvo: that comment doesn't seem (a) accurate or (b) a good idea even if it were accurate ;-) (specifically the limitation to a key name)12:14
mvoI think it refelcts what we initially speced, not what I implemented ;)12:14
cjwatsonheh12:14
* mvo corrects12:14
cjwatsonI don't generally like limiting things to key names, particularly since apt will honour the key no matter what its name12:14
cjwatsonmvo: that said, somebody who changes the ubuntu-keyring package can deliver new master keys if they want12:15
cjwatsonmvo: so maybe the restriction on packaged updates is snake oil12:16
mvowe can not protect against this in case of a archive compromise, can we? if a package from a compromised archive gets installed a attacker may run any maintainer script as root and/or ship any content12:17
cjwatsonright12:17
cjwatsonso it makes absolute sense for net-update, but is just an inconvenience for update12:17
mvohm, but its true that this check then makes less sense12:17
mvoyes, I agree12:17
mvocjwatson: I can update apt-key now12:21
cjwatsonsounds good12:21
GreatMandrillthe new update-key will not have the master-key signature restriction?12:24
mvocould you please review the comment in http://paste.ubuntu.com/6614/ for clarity?12:28
GreatMandrillit seems clear to me but i am not a english native speaker12:30
cjwatsonmvo: clear enough, yes12:32
mvothanks12:32
* mvo commits12:32
GreatMandrillcjwatson, mvo: thanks for the support12:35
cjwatsonyou're welcome12:36
mvocjwatson: is it the installer that creates /etc/default/locale ? we don't have it on dapper->hardy upgrades and I was wondering what the best course of action is (we used /etc/environment in dapper)14:39
mvoif we just keep it as it is, gdm needs a (trivial) patch to look there too for the default locale14:40
cjwatson/etc/default/locale is the new world order14:48
cjwatsonbut /etc/environment is indeed needed for upgrades14:48
cjwatsonwhich does gdm use at the moment?14:48
cjwatsonoh, /etc/default/locale14:48
mvoyeah14:48
mvoI have a fix (trivial)14:48
mvothat I'm about to upload14:48
cjwatsonI'm *sure* I told Seb that we needed to keep on supporting /etc/environment for compatibility14:48
cjwatsonplease do sort it out14:49
mvothanks, will do14:49
CIA-1ubiquity: evand * r2610 ubiquity/debian/po/ (78 files): Update translations from Rosetta.15:18
mario_limonciellevand, according to the manifest of the DVD generated this morning it looks like ubiquity 1.8.2 didn't squeeze in.  could you re-generate it so that it gets in?15:35
evandmario_limonciell: sure thing, will do now15:37
CIA-1debian-installer: cjwatson * r904 ubuntu/ (7 files in 4 dirs):16:20
CIA-1debian-installer: * Put fat-modules back in various cdrom, netboot, and speakup initrds16:20
CIA-1debian-installer:  (removed in 20061102ubuntu1), since it was split out from16:20
CIA-1debian-installer:  fs-secondary-modules again.16:20
evandargh, forgot to build the livefs.16:29
cjwatsonoops16:29
evandon it now16:30
evand(and rerolling the daily live while I'm there)16:30
cjwatsonanyone know of a GTK container widget that largely acts like GtkHBox, but that will wrap items to the next line if they get too long?16:55
mario_limonciellyou can't handle that with an alignment widget?16:56
cjwatsonit won't make the hbox wrap, will it?16:59
cjwatsoncontext: I made the rest of the advanced partitioner translatable, and was testing it; I used dummy translated strings that were a bit longer than the English text, and noticed that it caused the whole installer to get wider because the buttons along the bottom of the partition list view were all wider and nothing made them wrap17:00
cjwatsonhttp://people.ubuntu.com/~cjwatson/tmp/wide-partitioner.png17:02
evandheh, cute17:04
cjwatsonI basically just want them to wrap like text would17:07
evandI don't know of any such widget (short of hacking up one), but perhaps you'll get a timely response from the gtk ML?17:07
evandindeed17:07
cjwatsonnot keen on dropping in a new widget at the last minute17:07
cjwatsonmaybe I could shorten the undo_everything text and just use "Undo"17:08
cjwatsonthough it's "Undo all" rather than step-by-step undo17:08
evandyeah, I'm not sure how well "undo all" translates, but that seems reasonable.  As you say, "undo" might be confusing to those who expect it to undo the last action.17:09
cjwatsonor I could just unconditionally move Undo to its own line17:09
cjwatsonmight be better - I suspect some languages will have fairly long translations anyway17:10
evandugly, but that works too :)17:10
Siliciumhow i can fix the locales problem in a new chroot environment? i remember there exist a package to fix17:11
evandlocales problem?17:11
cjwatsoninstall language-pack-en17:12
Siliciumthanks17:12
cjwatsonor language-pack-<whatever>17:12
xivulonCreating a new row wouldn't push Next/Cancel buttons further down potentially creating problems in low res?17:13
evandnot on that page17:13
cjwatsonthe partition list would get shorter instead17:13
xivulonah great17:14
cjwatsonexcessive width is also a concern for low res ...17:14
xivulonabsolutely17:14
xivulonI would have suggested to shrink the to part next... :P17:14
Siliciumyea now both works17:15
Siliciuminstall CD and the install script for debootstrap17:15
Siliciumi created a script that runs debootstrap and then install the package with chroot17:15
mario_limonciellevand, it looks like there is a significant chance that the build is still going to fail.  liblaunchpad-integration1 caused an error in the livefs generation17:38
evandmario_limonciell: the livefs built and daily-live CDs are up.  I just started the dvd generation again a few minutes ago.17:39
mario_limonciellevand, oh perhaps I was looking at an old log at http://people.ubuntu.com/~ubuntu-archive/livefs-build-logs/hardy/ubuntu-dvd/20080408.1/livecd-20080408.1-i386.out17:40
cjwatsonah, that's better17:45
evandwhat is?17:55
cjwatsonmoving Undo to the next line17:56
evandfantastic17:56
cjwatsonhttp://people.ubuntu.com/~cjwatson/tmp/not-so-wide-partitioner.png17:58
cjwatsonin fact if you compare that side-by-side with wide-partitioner.png you can see that the window height is identical17:58
CIA-1ubiquity: cjwatson * r2611 ubiquity/ (6 files in 4 dirs):18:00
CIA-1ubiquity: * Make all text in the partitioner translatable (LP: #144741).18:00
CIA-1ubiquity:  Unfortunately we don't have time to get translations in place for 8.04,18:00
CIA-1ubiquity:  but there ought to be time to do this for 8.04.1. This also involved18:00
CIA-1ubiquity:  moving the "Undo changes to partitions" button to a second line in order18:00
CIA-1ubiquity:  to avoid excessive horizontal growth.18:00
cjwatsonevand: could probably do with a changelog entry for the translation update; it's useful to be able to glance down the changelog and see when it last happened18:01
evandThis may be obvious, but why do you end up with [de] Text when using de?18:01
cjwatsonoh, that's just my sample text18:02
evandcjwatson: I went straight to bzr log for a template to use when making the change as that's how debconf-updatepo is done, but will do.18:02
cjwatsonthose were cases where I couldn't be bothered to reach for the dictionary to figure out what the real translation would be18:02
cjwatson(or where I couldn't remember it)18:02
evandheh18:02
cjwatsondebconf-updatepo is less significant, so I haven't generally bothered changelogging that18:02
cjwatsonbut translation updates can actually matter for bug reports18:02
evandah, indeed18:04
CIA-1ubiquity: evand * r2612 ubiquity/debian/changelog: * Update the translations from Rosetta.18:06
CIA-1ubiquity: cjwatson * r2613 ubiquity/debian/ (changelog ubiquity.templates):18:08
CIA-1ubiquity: * Adjust po-debconf flags in ubiquity/install/copying_error/md5, so that18:08
CIA-1ubiquity:  ${FILE} doesn't get translated.18:08
evandah, thanks18:09
cjwatson(just happened to notice that in a .po file)18:10
CIA-1partman-basicfilesystems: cjwatson * r559 ubuntu/ (3 files in 2 dirs):18:46
CIA-1partman-basicfilesystems: * Disable automounting unless partman/automount is preseeded to true18:46
CIA-1partman-basicfilesystems:  (works around LP #106209).18:46
ubotuLaunchpad bug 106209 in partman-basicfilesystems "fsck Unable to resolve UUID" [High,Confirmed] https://launchpad.net/bugs/10620918:46
mario_limoncielllooks like 20080408.2 still used the old livefs :(19:36
evandargh yeah:   language-support-cy: Depends: language-support-translations-cy but it is not installable19:49
xivulonbdmurray: have added a comment to the bug, can you review it?22:09
xivulonI now also have I/O problems, looks like a kernel issue...22:09
xivulonerrors are random in my case, generally in latest stages of rc2.d22:10
evandxivulon: cking left for the evening it seems.  I'll shoot him an email.22:11
* evand out for a bit22:11
xivulonhmm if it is a kernel issue that is going to be quite annoying...22:12
xivulonI managed to get some logs and saved them to22:58
xivulonbug #20413322:58
ubotuLaunchpad bug 204133 in wubi "wubi install unusable - Buffer I/O error on device loop0" [High,Confirmed] https://launchpad.net/bugs/20413322:58
TheMusoxivulon: What exactly is the problem?22:59
xivulonloop0 errors23:00
xivulonthat result in / becoming r/o or not accessible at all23:00
xivulonloop0 in my case is the file hosting /...23:01
TheMusoright23:01
xivulonbdmurray was the first to experience something similar23:02
bdmurraylucky me23:03
xivulonbdmurray can you review the comments in the log?23:03
xivulonin the bug23:04
bdmurrayYes, I've been busy with bug day but will do.23:04
* TheMuso will look after breakfast.23:04
xivulonin your case you always had a crash before single user shell correct?23:04
xivulonthanks a lot TheMuso, can do with some help23:04
mario_limonciellevand, it looks like language-support-cy is available in apt now.  perhaps can you try the DVD again?23:12
=== ebel_ is now known as ebel
=== ebel is now known as _ebel
=== _ebel is now known as ebel
xivulonbdmurray one thing to try might be to disable /etc/init.d/lupin-sysctl23:22
xivulonbdmurray: disabling lupin-sysctl does the trick for me or at least I did not experience any crash23:35
bdmurrayxivulon: that might have worked for me too, still confirming23:39
TheMuso_Mand that question only gets asked if a network is available.23:39
TheMusowoops23:39
bdmurrayxivulon: yes I've logged in and the system seems usable23:43
xivulon\o/23:43
xivulonah the new kernel might have issues with some of the sysctl settings23:43
xivulonI23:44
bdmurrayI was using -1223:44
xivulonwill need some kernel dev to go over them23:44
xivulonbdmurray not sure what you mean23:44
bdmurrayxivulon: you said "new" kernel so I wasn't sure new compared to what23:45
xivulonah it's the kernel23:45
xivulonit might be possible to uninstall lupin-support completely...23:46
xivulonwell thanks a lot bdmurray, very very helpful23:46
bdmurrayxivulon: your welcome, I hope it gets sorted23:46
xivulon(finger crossed) yes23:47
bdmurrayWhat provides that lupin-sysctl?23:47
xivulonit was used to remount /host rw which at the moment is pointless since /host is mounted rw to begin with...23:48
xivulonand it had sysctl settings to improve recovery on hard reboot (as suggested by szaka)23:48
xivulonbut those seem to create other issues now23:48
xivulonso no much left...23:49
bdmurrayOkay, but which package provides that file?23:49
xivulonlupin23:49
bdmurrayheh23:49
xivulonevand we might consider to remove lupin-support installation23:58
xivulonI tested booting without it and it works well23:59
xivulononly thing to consider is if some setting in lupin-sysctl still makes sense23:59

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!