[03:41] <Symmz> Hey i was wondering if someone could help me with something
[03:42] <Symmz> im getting this error message while installing
[03:43] <Symmz> ok the error message is "udeved[75}: woker [147} failed while handling /devices/pci0000:00/0000:00:01.0/0000:01:05.0"
[03:43] <Symmz> does anyone know what this means lol im stumped
[08:18] <CIA-45> ubiquity-slideshow-ubuntu: evand * r375 ubiquity-slideshow-ubuntu/ (8 files in 2 dirs):
[08:18] <CIA-45> ubiquity-slideshow-ubuntu: Convert HTML entities to their unescaped forms to help translators.
[08:18] <CIA-45> ubiquity-slideshow-ubuntu: Thanks Dylan McCall.
[08:18] <CIA-45> ubiquity-slideshow-ubuntu: evand * r376 ubiquity-slideshow-ubuntu/po/ (6 files in 6 dirs): generate-pot-files
[08:21] <CIA-45> ubiquity-slideshow-ubuntu: evand * r377 ubiquity-slideshow-ubuntu/debian/changelog: Fix changelog.
[09:37] <CIA-45> ubiquity-slideshow-ubuntu: evand * r378 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 44
[09:45] <CIA-45> ubiquity: evand * r4943 trunk/ (152 files in 3 dirs): Update translations from Launchpad.
[09:47] <CIA-45> ubiquity: evand * r4944 trunk/ (d-i/manifest debian/changelog):
[09:47] <CIA-45> ubiquity: Automatic update of included source packages: flash-kernel
[09:47] <CIA-45> ubiquity: 2.28ubuntu31, partman-base 151ubuntu1.
[09:55] <CIA-45> ubiquity: evand * r4945 trunk/src/wallpaper/Makefile.in: Add missing wallpaper Makefile.in
[10:47] <CIA-45> ubiquity: evand * r4946 trunk/debian/changelog: releasing version 2.7.29
[13:02] <Wellark> cjwatson: hi
[13:03] <Wellark> well, at least the grub-setup problem can be easily workaround by dropping to installer shell, chrooting to /target and running grub-install and update-grub manually
[13:15] <cjwatson> Wellark: yeah, it sounds like a cache coherency bug
[13:15] <cjwatson> not sadly the first such
[13:18] <Wellark> cjwatson: do you want me to report a new bug or is #838981 "close enough"?
[13:20] <cjwatson> please file a new one
[13:20] <cjwatson> it's easier to mark bugs as duplicates than it is to split them up
[13:29] <stgraber> ev: Just saw your fix for the wallpaper in ubiquity. Sadly this won't work for edubuntu as our default wallpaper is /usr/share/backgrounds/edubuntu_default.png. What'd be the best way of making it work?
[13:30] <ev> stgraber: probably factoring out the wallpaper path as an argument to the binary
[13:31] <stgraber> then check for /usr/share/backgrounds/edubuntu_default.png in ubiquity-dm?
[13:40] <ev> correct
[13:45] <stgraber> ok, I'll do that once I'm done doing some more work on friendly-recovery
[13:49] <ev> okay, cool
[13:49] <ev> thanks
=== kentb-afk is now known as kentb
[15:15] <Wellark> cjwatson: if you have time could you explain the cache coherency problem?
[15:16] <Wellark> coherency problem when changing files outside of chroot and inside chroot?
[15:20] <Wellark> shall I tag with something?
[15:21] <cjwatson> no, tags are useless :-)
[15:21] <Wellark> OK XD
[15:21] <cjwatson> (for this anyway)
[15:22] <cjwatson> people tag grub2 bugs with all kinds of irrelevant crap like "boot", gets on my nerves :-)
[15:22] <Wellark> I feel you :)
[15:22] <cjwatson> we're writing files to the partition using ordinary POSIX system calls, and then trying to read them back using GRUB's built-in filesystem drivers reading directly from the disk device
[15:23] <cjwatson> there's a caching layer in between those and if it isn't properly flushed you can get this kind of symptom
[15:23] <Wellark> I think the tags thingy is pretty much undocumented and people don't simply know how to use them
[15:23] <Wellark> so, "sync" is not enough? :(
[15:23] <Wellark> it _should_ be
[15:24] <Wellark> bug #851012
[15:24] <ubot2> Launchpad bug 851012 in ubuntu "alternate installer: grub2 can't be installed on a partition" [Undecided,New] https://launchpad.net/bugs/851012
[16:25] <cjwatson> Wellark: "it's a bit more complicated than that"
[18:35] <Wellark> cjwatson: I'm sure it is :)
[18:35] <Wellark> I'm just trying to learn something here:D
[18:36] <Wellark> because AFAIK "sync" should force the actual buffers to be flushed to the permanent storage so why does that not happen or if it does indeed happen why doesn't grub see the data?
[18:37] <Wellark> for example in reiser4 case the fact that programs did trigger sync (e.g. vi and emacs) and that caused tremendous performance problems as the reiser4 kernel module indeed was forced to sync to the disk
[18:38] <Wellark> and that completely trashed the performance of the algorithms
[18:40] <kyleN> superm1, hi. my project with oem-config is not installing lang packs for the user selected language even though the apt cache knows where they are at oem-config run time. do I need to turn that one somehow?
[18:41] <kyleN> (he had discussed I think that it was automatic and the packs installed are those returned by check-language-support -l <user-selected-lang>)
[18:41] <kyleN> we had, i mean
[18:41] <kyleN> my packs are in an on disk archive from which I can install them with apt-get install, but no success on the oem-config path.
[18:42] <kyleN> nothing in /var/log/oem-config.log (except some keymap warnings)
[18:49] <kyleN> ev, do you know about this ^ ?
[18:56] <kyleN> superm1, I also have another question: what's the best way to cut down the list of languages displayed on the first page of oem-config (without modifying source if possible).
[18:58] <superm1> kyleN, you need to preseed only-show-installable languages to cut down the list
[18:58] <superm1> as for it not installing them, can you confirm that you can actually apt-get install them manually with how they're set up in your apt cache?
[18:59] <kyleN> superm1, ok I can check on the preseseding
[18:59] <kyleN> superm1, I just did manually install them with apt-get from the on disk archive
[19:00] <superm1> kyleN, then you'll  need to run oem-config in debug mode and see what is happening when it tries to install them i think
[19:01] <kyleN> superm1, how. pass -debug arg and launch with sudo from normal termimal/desktop?
[19:01] <superm1> kyleN, there is a kernel command line option for it
[19:01] <superm1> debug-oem-config I think
[19:01] <kyleN> superm1, ok. thanks
[19:03] <superm1> kyleN, you can ask chris townsend to look at one of the oneiric ISOs that I have working with the full process if you want to compare, he can point you at it
[19:03] <kyleN> superm1, thanks
[19:26] <kyleN> only-show-installable-languages worked regarding cutting down display. (I also set find-installable to true), superm1
[19:27] <superm1> kyleN, cool, so it only offers the ones in your pool now right?
[19:27] <kyleN> superm1, right. but now I click Continue and it hangs and I con't move to the next page
[19:28] <kyleN> Continue button never gets enabled properly
[19:28] <superm1> you have to pick a language for it to get enabled
[19:28] <superm1> maybe the sensitivity setting isn't working right?
[19:29] <kyleN> superm1, I think this is something unrelated to stock oem-config. I'll let you know if there's an issue
[21:30] <kyleN> superm1, something else I notice is that after session login, LANGUAGE var is unset and is not set in /etc/default/locale.
[21:31] <kyleN> that's the main var now I think for controlling the user language for translation purposes
[21:31] <superm1> kyleN, i think you need to investigate the debug logs to see what happened
[21:31] <superm1> hopefully they're informative
[21:31] <kyleN> superm1, do you expect LANGUAGE to be set?
[21:32] <superm1> kyleN, LANG should be set in /etc/default/locale
[21:32] <kyleN> superm1, LANG is. LANGUAGE is not and should be
[21:33] <superm1> i dont think there is a LANGUAGE variable in /etc/default/locale by default
[21:33] <kyleN> true, but there needs to be I believe
[21:33] <superm1> i don't think so
[21:33] <kyleN> it is THE var that sets the language for translation purposes now
[21:33] <superm1> i just checked a recent spanish install and it's unset
[21:34] <kyleN> at session time is it set? (must be)
[21:35] <kyleN> LANG is now used for formats (number, currency, etc) only I believe
[21:35] <superm1> i don't see it set in the session no (looking at export output)
[21:35] <kyleN> and I bet you are not seeing all the translations in the UI you would expect
[21:35] <superm1> well all the apps are in spanish it seems
[21:35] <superm1> unity is in spanish
[21:36] <superm1> so if it's supposed to be, that might be an unrelated problem to lang packs getting installed
[21:36] <kyleN> i think it is unrelated
[21:36] <superm1> there is something wrong with lightdm's screen not showing up properly in zh_CN i think though
[21:41] <superm1> according to scripts/localechooser-apply LANGUAGE will only be set if the languagelist is one of  them that has alternatives
[21:41] <kyleN> superm1, I am surprised. there was definitie movement away from LANG to LANGUAGE but oneiric seems to be using LANG for translations
[21:41] <superm1> so it's possible
[21:41] <kyleN> it is true that LANGUAGE sets a hierarchy
[21:41] <superm1> maybe scripts/localechooser-apply needs to be updated still
[21:41] <kyleN> LANGUAGE=fr_FR:en
[21:41] <kyleN> kind of thing
[21:42] <kyleN> superm1, if I set LANG in terminal to en_US.UFT-8 and launch gedit, it is in en
[21:43] <kyleN> if I export LANGUAGE=fr_FR and launch gedit it is in french
[21:43] <kyleN> my guess is that if LANGUAGE is unset it falls back to old-style lang
[21:43] <kyleN> LANG
[21:43] <kyleN> but LANGUAGE overrides, as it should
[21:44] <superm1> looking at localechooser/post-base-installer.d/05localchooser it does the same thing, LANGUAGE is only set if languagelist is a list of languages with alternatives, otherwise it's useless
[21:44] <superm1> so if that is no longer the case i think both localechooser and ubiquity need to be updated
[21:44] <kyleN> so, I assert LANG should not be used for translation language.
[21:44] <kyleN> I can ask david planella and martin pitt
[21:44] <superm1> ok
[21:45] <kyleN> I'll email them and cc you
[21:45] <superm1> Ok
[21:45] <kyleN> email?
[21:45] <superm1> superm1@ubuntu.com is fine
[21:45] <kyleN> thx
[21:48] <kyleN> sent
[21:49] <kyleN> now, why don't my lang packs install :)
[21:51] <kyleN> superm1, does misc_install.py's do_install function always run? i inserted some code to write the value of the langpacks arg to file and it was never written...
[21:51] <kyleN> sorry, that should be install_misc.py
[21:52] <superm1> it's called from plugininstall.py's install_language_packs
[21:52] <kyleN> ok
[21:52] <superm1> so maybe is self.langpacks not being set?
[21:52] <kyleN> maybe
[21:52] <superm1> oem-config/install-language-support
[21:53] <superm1> did you preseed that true?
[21:53] <kyleN> ah
[21:53] <kyleN> don't think so
[21:53] <superm1> oh i think i remember ev decided  to not set it true because not all oem scenarios will have an on disk apt pool, that's right
[21:53] <superm1> (by default)
[21:54] <kyleN> that's ok as long as we know about it :)
[21:56] <kyleN> yes, its false by default
[23:23] <kyleN> superm1, progress. except for one bit. I had to add "--force-yes" to  the code in /usr/share/ubiquity/plugininstll.py or else it fails. Does this mean i did not set up my on-disk apt-cache correctly?
[23:24] <kyleN> the code I refer to is the debconf-apt-progress line that actually does the installation of the pkgs
[23:25] <kyleN> oh it still fails :(
[23:26] <kyleN> no wait, the lang packs ARE (finally) installed. but I do get a big error message at the end of oem-config (in spanish)
[23:27] <kyleN> so there is progress but not solved yet. the error is that the debconf-apt-progress that does that apt-get install exits with status of 100.
[23:29] <kyleN> I can press OK and continue, but checking in another tty shows the lang packs are indeed installed
[23:31] <kyleN> I'll have to take this up later, but to summarize:
[23:32] <kyleN> to install lang pack debs I have in an on disk archive the apt-cache knows about (and from which I can install from at the command line):
[23:33] <kyleN> Oh, and to only show the langs in oem-config  for which I have lang packs in that apt cache:
[23:33] <kyleN> set debconf 'ubiquity ubiquity/show-only-installable boolean true'
[23:34] <kyleN> set debconf 'oem-config oem-config/install-language-support boolean true'
[23:35] <kyleN> edit /usr/share/ubiquity /installplugins.py to use the apt --force-yes option
[23:36] <kyleN> (which may mean I set up my on disk archive wrong? ^ )
[23:36] <kyleN> and I get an error pop-up after the lang packs are successfully installed indicating installplugins.py exited with status 100. but I can press OK and continue and all the lang packs are indeed isntalled