| alkisg | Good morning | 04:23 |
|---|---|---|
| HedgeMage | hi :) | 04:23 |
| stgraber | oh, it's already this late ? ;) | 04:25 |
| HedgeMage | yes, sadly it is | 04:28 |
| HedgeMage | (creeping up on midnight here) | 04:28 |
| alkisg | stgraber: do I have 1 hour before you tag ltsp? I think that "tasksel failed because aptitude returned 100" is a file descriptor problem, and actually means "aptitude returned 100% completed", so I'm just now testing ltsp-build-client to see if it's safe to ignore it at least for now... | 04:30 |
| stgraber | alkisg: you have a bit more than that, I did a lot of changes today in ltsp-trunk and ldm-trunk, so I'll wait until tomorrow (late morning / early afternoon) to tag + upload | 04:31 |
| alkisg | Ah, perfect :) | 04:31 |
| stgraber | I merged that nbd-proxy bit and implemented some initial caching for ltsp_config + some other stuff I had laying around | 04:31 |
| alkisg | stgraber: what's the easiest way for me to test your commits? | 04:32 |
| alkisg | I.e. to make a local ubuntu package out of the bzr branch? | 04:32 |
| stgraber | alkisg: I have an updated ltsp in my PPA (bzr snapshot), I can update it with what's currently in trunk so you can easily test it | 04:32 |
| stgraber | give me 2 minutes | 04:32 |
| alkisg | Merci :) | 04:33 |
| stgraber | sorry, don't know the greek for no problem ;) | 04:33 |
| alkisg | :D (or if you could tell me how you do the bzr => ubuntu package thing, I can do it on my own here...) | 04:33 |
| stgraber | I usually use mkdst (that tool we use to tag/release ltsp) to build me a new tarball, then rename it to mention it's a snapshot and include the bzr revision and make sure the version number is lower than the next upstream version | 04:34 |
| crimsun | ok, seed time | 04:35 |
| alkisg | Thanks, I'll put "learn how mkdst works" in my TODO list :) | 04:35 |
| stgraber | so, for the one I'm doing, it's: ltsp_5.1.99~bzr1517.orig.tar.gz | 04:35 |
| crimsun | (now that I've closed 76 bugs with the latest upload) | 04:35 |
| alkisg | crimsun: hi, will we see that libsdl-pulse inclusion any time soon? | 04:35 |
| stgraber | crimsun: that's a lot of bugs :) | 04:36 |
| alkisg | wow, 76 bugs, yeah... | 04:36 |
| crimsun | alkisg: that's the seed thing mentioned above | 04:36 |
| * alkisg dances :) | 04:36 | |
| stgraber | alkisg: new snapshot build for ltsp, uploaded. | 04:37 |
| stgraber | alkisg: I'll do one for ldm too or you'll have some issue getting to ldm | 04:37 |
| * alkisg thinks Lucid will be the best Ubuntu version ever :) | 04:37 | |
| alkisg | stgraber: the nbd-client patch I reported? don't worry I have a local patch for it here... | 04:37 |
| alkisg | *bug | 04:38 |
| stgraber | alkisg: not only that, nbd-proxy makes nbd-client connect to localhost, so it's introducing a lot of other issues ;) | 04:38 |
| alkisg | Oooh ok, waiting for your ppa upload then | 04:39 |
| stgraber | uploaded | 04:41 |
| stgraber | alkisg: you should have both built in an hour or so, the builders seem a bit slower now than they were this afternoon | 04:42 |
| stgraber | alkisg: I also uploaded a new ltspfs but this one was uploaded directly to archive, not to my PPA (it's an actual release, not a snapshot) | 04:43 |
| alkisg | Btw, here's what I'll be commiting in a few hours. It has "TODOs" in it, but I think it works better than the last version. I'll debug those in the next few days. | 04:43 |
| alkisg | http://ltsp.pastebin.com/m20e9470 | 04:43 |
| alkisg | stgraber: ok, I'll be trying all those in the next few hours | 04:44 |
| stgraber | are you sure your tasksel works correctly " | 04:44 |
| stgraber | ? | 04:44 |
| alkisg | I'm testing as we speak | 04:44 |
| stgraber | I'd have done: chroot $ROOT "tasksel install $desktop || true" | 04:45 |
| alkisg | Ooops | 04:45 |
| alkisg | dammit | 04:45 |
| stgraber | also, AFAIK that APT_GET_OPTS isn't actually part of apt's environment, it's something we usually append to an apt-get command and I don't see one in your script, not sure it's actually doing anything ;) | 04:46 |
| alkisg | I was fiddling with tty handling a while ago, and I saw tasksel's gui (whiptail) going to 100%, and right after that "aptitude failed (100)" | 04:46 |
| alkisg | So I'm beginning to think it's a descriptor/return value interpretation problem | 04:46 |
| stgraber | and IIRC I actually have a plugin writting a file in /etc/apt/conf.d/ to do the same as --no-install-recommends, so you may want to get rid of it too | 04:46 |
| alkisg | APT_GET_OPTS is used later on to write the "recommends" stuff to apt.conf.d | 04:47 |
| alkisg | So I remove "recommends" to prevent it from being written to apt.conf.d | 04:47 |
| stgraber | oh, interesting, I forgot I actually checked for that variable before writting apt.conf.d ;) | 04:47 |
| alkisg | :) | 04:47 |
| stgraber | btw, that same chroot fix also applies to the last chroot call | 04:48 |
| stgraber | the || true should be in the chrooted command, not outside | 04:48 |
| alkisg | It's not really useful in the script, it's only useful for updates later on | 04:48 |
| alkisg | Does it make any difference? | 04:48 |
| stgraber | chroot $ROOT "gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --set --type list --list-type string /apps/panel/global/disabled_applets \"[OAFIID:GNOME_FastUserSwitchApplet,OAFIID:GNOME_IndicatorApplet]\" || true" | 04:48 |
| stgraber | well, with: "chroot $ROOT something || true", if $ROOT doesn't exist or chroot isn't installed (for some unlikely examples), it won't complain | 04:49 |
| alkisg | In any case I'll try to change that to a gconf file, and put it in the correct dir | 04:49 |
| stgraber | whereas with 'chroot $ROOT "something || true"', you'll still get the most important errors but ignore the return code of the command called inside it | 04:50 |
| stgraber | for your debconf issue, it's actually quite simple to workaround I guess | 04:50 |
| alkisg | Well sure but it'd be weird for chroot to return an error there, if it was working in all the lines above | 04:50 |
| * alkisg is all ears... :) | 04:50 | |
| stgraber | we're using the Passthrough backend which uses that fd that you are getting error for to report status to the debian-installer (at least it was the idea) | 04:51 |
| stgraber | you could set it to NonInterractive for the fat client part and that won't use any fd | 04:51 |
| alkisg | :) | 04:51 |
| alkisg | stgraber: finished succesfully - but it was unable to unmount /proc, and that got me a lot of squashfs compression errors (file in use etc) | 04:52 |
| alkisg | I also got that when I tried the version of the script that's currently in bzr | 04:52 |
| stgraber | OLD_DEBIAN_FRONTEND=$DEBIAN_FRONTEND | 04:52 |
| stgraber | DEBIAN_FRONTEND=noninteractive | 04:53 |
| stgraber | tasksel blah | 04:53 |
| alkisg | Thanks, got it :) | 04:53 |
| stgraber | DEBIAN_FRONTEND=$OLD_DEBIAN_FRONTEND | 04:53 |
| alkisg | Can I force umount -l /proc there temporarily, until I properly debug it in the next days? | 04:53 |
| stgraber | alkisg: check for running process inside the chroot | 04:53 |
| stgraber | alkisg: something must have start a process during the installation of some package and that's what blocking /proc | 04:54 |
| stgraber | if you kill that process, you'll be able to unmount /proc | 04:54 |
| alkisg | I think `sudo lsof /opt/ltsp/i386/proc` is giving me the server's processes... !!! | 04:54 |
| stgraber | sure | 04:54 |
| alkisg | How else would I check to see what's running inside the chroot? | 04:54 |
| stgraber | /proc and /opt/ltps/i386/proc are exactly the same thing ;) | 04:54 |
| stgraber | ps aux and look at the start time | 04:55 |
| stgraber | it's usually the easiest | 04:55 |
| alkisg | Ah. /me feels silly :) | 04:55 |
| alkisg | notify-osd it is | 04:55 |
| alkisg | Hmm, no, looking.... | 04:56 |
| alkisg | I actually don't have any such process running. Maybe it's some daemon that's being reused from some program in the chroot? | 04:59 |
| stgraber | it shouldn't anyway, you can do a "umount -l" on it, it'll solve your squashfs issue | 05:00 |
| alkisg | OK, so temporarily I'm going to umount -l && remount it, with a # TODO to properly solve it later | 05:00 |
| stgraber | going for some sleep, last night was really short due to that livedvd issue ;) see you all in 7-8 hours. | 05:14 |
| alkisg | goodnight :) | 05:15 |
| alkisg | !info gimp lucid | 05:30 |
| ubottu | gimp (source: gimp): The GNU Image Manipulation Program. In component main, is optional. Version 2.6.7-1.1ubuntu1 (lucid), package size 4285 kB, installed size 12480 kB | 05:30 |
| ma3x | hello | 12:29 |
| ma3x | i installed ubuntu and edubuntu as extra. can i make the distro edubuntu or at least how can ichange the splash screen to edubuntu | 12:29 |
| ma3x | helloooooo | 12:32 |
| highvoltage | stgraber: good morning to you | 13:11 |
| highvoltage | stgraber: did yesterday's image become the alpha 2 or do I need to do another sync? | 13:11 |
| alkisg | highvoltage: I'm having problems to find a *correct* way to prevent gdm from being installed while keeping everything else in place. Is it a problem if it gets installed instead, but of course while keeping ldm as the default display manager? | 13:12 |
| alkisg | (i.e. `echo "/usr/sbin/ldm" > $ROOT/etc/X11/default-display-manager` before installing ubuntu-desktop) | 13:13 |
| stgraber | highvoltage: it became the alpha-2 | 13:14 |
| highvoltage | alkisg: I wonder whether ldm could get a Provides field so that it provides GDM and KDM (that should solve it on Kubuntu too then, if it's installed before the desktop main metapackage) | 13:49 |
| alkisg | highvoltage: I don't think it's a bad idea to make ldm able to work beside gdm | 13:49 |
| highvoltage | stgraber: yay! | 13:49 |
| alkisg | It could also be used in the future in standalone workstations | 13:49 |
| stgraber | alkisg: I suspect your LTSP boot didn't work, right ? | 13:50 |
| alkisg | E.g. I setup ubuntu normally, then install ldm, and I have something like a fat client but with centralized administration... | 13:50 |
| stgraber | alkisg: I just found this bug in my current package: http://pastebin.com/f7b8424bd | 13:50 |
| alkisg | stgraber: it works but only when I manually install the needed packages | 13:50 |
| alkisg | Ah, ok, I used IPAPPEND 3 to bypass this | 13:51 |
| alkisg | Thanks | 13:51 |
| alkisg | highvoltage: I think with a small patch in ltsp_config, the client should be able to boot even with gdm installed | 13:52 |
| alkisg | And gdm won't start if it sees '/usr/sbin/ldm' in /etc/X11/default-display-manager | 13:52 |
| * alkisg tries all these... | 13:52 | |
| alkisg | highvoltage: Yup, with a 2 lines-patch in ltsp_config, now ldm works fine for me with gdm also installed in the chroot | 14:06 |
| alkisg | highvoltage: could you tell me about the other "blacklisted" packages in your fat client script? | 14:06 |
| alkisg | (I mean, why they were blacklisted?) | 14:06 |
| alkisg | Supposedly, the correct way to install ubuntu-desktop is with `tasksel install ubuntu-desktop`. I tried that, but I ended up with a chroot without e.g. lspci, which broke many xorg scripts. | 14:13 |
| alkisg | Should I try `tasksel install ubuntu-desktop` AND apt-get install ubuntu-standard? | 14:13 |
| alkisg | Or should I just do `apt-get install ubuntu-desktop` and get on with it? :) | 14:13 |
| sbalneav | Morning all | 14:43 |
| stgraber | hey sbalneav | 14:44 |
| stgraber | alkisg: that's fixed in the new package, I depend on lspci | 14:45 |
| stgraber | alkisg: the reason why it didn't work is that it was dropped from ubuntu-minimal and so was dropped from the chroot | 14:45 |
| alkisg | Ah, ok | 14:45 |
| alkisg | stgraber: I'm trying to do this: apt-get install ubuntu-standard ubuntu-desktop | 14:46 |
| alkisg | (or any other user defined desktop) | 14:46 |
| alkisg | I.e. I assume that fat == those packages only... | 14:46 |
| alkisg | And I only left those in the blacklisted packages: etwork-manager modemmanager ubufox apport jockey-common | 14:47 |
| mhall119|work | what is ldm? | 14:54 |
| alkisg | The LTSP display manager | 14:55 |
| sbalneav | mhall119|work: The bit that manages the login process for thin clients | 14:56 |
| mhall119|work | ah, cool | 14:57 |
| mhall119|work | thought it was maybe an lxde display manager | 14:57 |
| mhall119|work | highvoltage: I've been reading about how to customize the new GDM and XSplash | 14:57 |
| mhall119|work | it doesn't look too hard, but it's not nearly an flexible as the old GDM and USplash | 14:58 |
| alkisg | ...almost ...there :) | 15:43 |
| alkisg | stgraber: I commited the modified script, could you ping me when you upload a new ltsp version to Ubuntu, so that I can test again from scratch? | 17:07 |
| stgraber | alkisg: ok | 17:09 |
| alkisg | ty | 17:09 |
| sbalneav | alkisg: Two months ago you couldn't even spell "Upstream Free Software Developer", and today you is one :) | 18:59 |
| sbalneav | \o/ for alkisg | 19:00 |
| alkisg | And who was the one that almost "pushed" me upstream? Thanks to sbalneav!!! :) | 19:01 |
| alkisg | sbalneav: but really, I've been a floss developer for decades. Just in assembly/Delphi, away from Linux :) | 19:02 |
| sbalneav | Well, I think being made a member of upstream is, if nothing else, a recognition of skill, and dedication. | 19:06 |
| sbalneav | One of the things I've always loved about Free Software is, it's one of the true meritocracies in the world. | 19:07 |
| sbalneav | You were making a significant contribution. | 19:07 |
| sbalneav | You *should* be recognized. | 19:07 |
| alkisg | Yup, being part of a good upstream project is a new experience for me, and a very good one | 19:08 |
| alkisg | Thank you guys for accepting me :) | 19:08 |
| * alkisg has to leave the keyboard again because his daughter needs to play ikariam :P :D | 19:08 | |
| stgraber | alkisg: I have two guys here debugging an issue with nbd-proxy, I expect to tag + upload once it's fixed. | 19:49 |
| alkisg | stgraber: np, thanks, what's nbd-proxy all about? what will it solve? | 19:49 |
| alkisg | Will clients need more ram for it? | 19:50 |
| stgraber | no they won't | 19:50 |
| stgraber | it's just something that makes sure you'll never get any squashfs or input/output error | 19:50 |
| alkisg | Cool :) | 19:51 |
| stgraber | it's handling the NBD reconnect in place of the current nbd kernel module and does a lot more checks | 19:51 |
| stgraber | I actually am testing it by doing a checksum of all files in the chroot while killing the server once every second ;) | 19:51 |
| alkisg | Wow!!! | 19:52 |
| alkisg | Sounds rock solid indeed :) | 19:52 |
| stgraber | haven't seen any dataloss or corruption yet but I sometimes have a process hanging and forcing me to reboot | 19:52 |
| stgraber | so that's what's being fixed now | 19:52 |
| stgraber | (something about sending a byte too much when reconnecting, leading to squashfs crashing ;)) | 19:52 |
| alkisg | So is that also included in the initramfs, completely replacing the kernel module? | 19:54 |
| mgariepy | sbalneav, ping | 20:03 |
| alkisg | sync -zthhP rsync://cdimage.ubuntu.com/edubuntu/releases/10.04/alpha-2/lucid-dvd-amd64.iso ./lucid-dvd-amd64.iso | 21:22 |
| alkisg | @ERROR: Unknown module 'edubuntu' | 21:22 |
| alkisg | rsync error: error starting client-server protocol (code 5) at main.c(1521) [receiver=3.0.6] | 21:22 |
| alkisg | How can I rsync to see if what I have is == alpha 2? | 21:22 |
| alkisg | Hmmm pitivi recommends hal... | 21:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!