IRC chat logs for #ltsp on irc.libera.chat (webchat)


Channel log from 29 January 2009   (all times are UTC)

00:00
<Ryan52>
I wonder if it'll work through inetd.
00:01cyberorg has quit IRC
00:01
<vagrantc>
Ryan52: and there's the problem with iw_IL.UTF-8 still
00:01
<Ryan52>
what's that?
00:02
<ltsppbot>
"vagrantc" pasted "iw_IL issue" (10 lines) at http://pastebot.ltsp.org/220
00:02
<Ryan52>
o
00:03* Ryan52 fixes
00:03
<Ryan52>
how do I get the length of a string?
00:03* Ryan52 will google
00:04
<vagrantc>
Ryan52: len(string)
00:05
<Ryan52>
language-with-name:iw_IL.UTF-8 (ישראל):iw_IL.UTF-8
00:05
problem solved.
00:06
<vagrantc>
hmmm... i think many of the "languages" with ; delimiters aren't translated
00:07
language-with-name:Georgian (საქართველო):ka_GE.UTF-8
00:07
surely they have a native translation for their own language...
00:07
that doesn't even have the ;
00:08
<Ryan52>
language-with-name:Español; Castellano (España):es_ES.UTF-8
00:08
<vagrantc>
maybe some are just missing translations...
00:08
<Ryan52>
yes, that's what I'm assuming.
00:09* vagrantc suspects language-with-name:ግዕዝኛ (ኤርትራ):gez_ER.UTF-8 is amiss
00:10* Ryan52 totally speaks ግዕዝኛ :p
00:10
<vagrantc>
my unicode is a little weak.
00:14
<Ryan52>
hrm. a few languages do that..
00:15
so what do I do? leave it as is and push?
00:15
<vagrantc>
it's way better than what's there now.
00:16
<Ryan52>
now lets see if it actually works..
00:17
<vagrantc>
and now we enter the hostile world of language politics
00:21
<Ryan52>
it worked! /me pushes
00:27zirconiumks has joined #ltsp
00:57Egyptian[Home]1 has quit IRC
01:20vagrantc has quit IRC
01:21Egyptian[Home] has joined #ltsp
01:29
<redspike>
how i get "euro" keys like "aao" to work on the clients?
01:30
ill canot fint the xorg.conf file for the clients :(
01:31
<Ryan52>
XKBLAYOUT A valid xorg Layout unset see xorg documentation for value
01:31
redspike, from the docs, it looks like you just have to set XKBLAYOUT in lts.conf
01:33
<redspike>
Ryan52: what i have read it can have some with the fonts todo to? that the fonts 75 and 100dpi most be :unscaled ?
01:34
<alkisg>
redspice, distro/version?
01:35shrek has joined #ltsp
01:36
<alkisg>
*redspike
01:36
<redspike>
alkisg, ubuntu 8.10
01:37
its not the keys on the keyboard that dosent work, its some console applications that runs in X that dosent displays that type of chars right.
01:37
<shrek>
hi!
01:37
<alkisg>
It should use your server defaults, without any lts.conf modifications. What do you have in /opt/ltsp/i386/etc/default/console-setup ?
01:38
<shrek>
anyone with some success with openchrome on ubuntu 8.10
01:38
?
01:38
<redspike>
alkisg: it works on the "server" but not on the clients :/
01:38
<shrek>
can't get it working despite the presence of openchrome drivers in the chroot
01:40polytan has joined #ltsp
01:40
<alkisg>
redspike: Ah, I see. If it isn't a keyboard issue, then it's probably a locale issue. What does `sudo chroot /opt/ltsp/i386 locale -a` give you?
01:40
...and also `cat /opt/ltsp/i386/etc/environment` ?
01:41
<polytan>
hi
01:42
<redspike>
alkisg: locale -a gives C POSIX sv_SE.utf8
01:43
/opt/ltsp/i386/etc/environment dosent exist
01:43
<alkisg>
And /etc/environment ?
01:44
<redspike>
PATH="/usr/localsbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
01:45
alkisg: ^
01:45
<alkisg>
You should have LANGUAGE=xxx and LANG=xxx entries in both environment files
01:46
I wonder how you did the installation and ended up without them...
01:46
<redspike>
but i dont have them on the server iself in /etc and on the server it works?
01:47
<alkisg>
Yes, but on the server you run the application that installed the language support
01:47
I don't know the details here, but I think that's the problem
01:49
<ltsppbot>
"alkisg" pasted "environment" (14 lines) at http://pastebot.ltsp.org/221
01:49
<alkisg>
redspike: see mine ^
01:50
redspike: I think you need to run update-locale in the chroot.
01:51
redspike: Could you also see the file /etc/default/locale in both the server and in the chroot?
01:53
<redspike>
alkisg: only on the server
01:53
cp when i run update-locale the file come there but its empty
01:54
<alkisg>
For my language, it's: update-locale LANG=el_GR.UTF-8 LANGUAGE=el_GR:el:en_GB:en
01:54
You need to specify the parameters for yours
01:56
<nubae>
alkisg: was it u that was working on ldap integration with ltsp?
01:56
<alkisg>
Nonono :)
01:56
Never used or even seen ldap working... :)
01:56
<nubae>
heh, thats quite a definitive statement
01:57
must have confused with Asmo then
01:57
I just thought u did something recently that was of interest to me
01:57
but cant remember what now
01:57
<alkisg>
Getting something like this in ubuntu is my next goal: http://fedoraproject.org/wiki/Docs/Drafts/AGBeta#Account_Management_-_Using_A_Graphical_User_Interface
01:58
<nubae>
looking...
01:58
<alkisg>
...and right afterwards, adding scripting support for it.
01:59
<nubae>
like?
01:59
<alkisg>
There was some discussion to get it to Ubuntu this summer: http://www.mail-archive.com/ubuntu-desktop@lists.ubuntu.com/msg01544.html
02:00
Scripts? like, select 10 users, go to the "scripts" menu and select "Clear their firefox cache" or lock their panels or clear their thumbnails or copy a file to them... whatever
02:01
Something like the gimp scripts, these would be "user-manager scripts".
02:01
<nubae>
ok... so u have a set of preinstalled scripts to use on mass users, or u can put in your own
02:01
something like that?
02:02
<alkisg>
Yes, both of them, and with a searching facility
02:02
Because I'm thinking there could be hunderds of scripts in there
02:02
<nubae>
from what I can see, the account creation could include LDAP too
02:02
ideally, yeah, although some are quite unique to user scenario cases
02:02polytan has quit IRC
02:02neoshrek has joined #ltsp
02:03
<alkisg>
I had even started a similar app on my own, before seeing the fedora app, and it was also named users-manager! http://img156.imageshack.us/img156/3473/myusersmanagerle6.png
02:03
<nubae>
and it would be nice to have a synch button so it could then be pushed to regular computers or fatclients
02:03
<alkisg>
I'm thinking "import from..." scripts
02:03
<nubae>
the user list I mean
02:04
ok, so import user list
02:04
<alkisg>
Like, import from a csv, from another ldap server, from a web site or whatever the scripter needs
02:04
<nubae>
and export user list to an ip
02:04
that would be really nice
02:05
<alkisg>
And if it's script (plugin) based, it'll be easier for maintainance, because different people could work on different scripts
02:05
<nubae>
push user list to LDAP server: some.ip.number
02:05
<alkisg>
E.g. I would make 100 plugins for classroom needs, but nothing for LDAP. You could make 5-6 scripts for LDAP or fat clients
02:05
<nubae>
man would that make life easier for fatclients
02:06
<alkisg>
I'm willing to spend several hours into it, if I can find someone to help me. I don't think I have enough hours to take it on my own.
02:06
<nubae>
well, only thing I ask is u make a nice looking big prominent button that says export users
02:06
<alkisg>
(hours = hours per week)
02:06
<nubae>
I can then tie that in with the script to do it
02:07
<alkisg>
The menus would also be able to be defined by scripts
02:07
<nubae>
well, I have so many users asking me for fatclient now, that I think I could help u
02:07
<alkisg>
So it would be your responsibility to provide the menu name, the icon, and the action
02:07
<nubae>
I need something more visual and foolproof...
02:07
<alkisg>
All in pygtk, so no compilation anywhere
02:07
I'm downloading F10 now to see it in action! :)
02:08
<nubae>
hmmm, but I already told u the button I want :-p
02:08
cant u just put it in ;-)
02:08
<alkisg>
Sure... but that would destroy the plugin architecture! :D
02:08
<nubae>
not really... its just an exception
02:09
albeit a very important one
02:09
think about it... no other tool does importing and exporting of users in such a nice way
02:09
the app will make u famous!
02:09
:p
02:09
no but seriously... Im willing to help to get that ported to buntu...
02:10
<alkisg>
Sure, but importing from where? There would be an "Import users/groups" menu, but it should be populized by plugins
02:10
E.g. import from csv => maintained by XYZ
02:10
Import from LDAP => maintained by nubae
02:10
<nubae>
well... there arent that many ways... its probably easier to give just a few options
02:10
<alkisg>
Import from website users.sch.gr => maintained by alkisg
02:11
<nubae>
import from .passwd, or import from ldap
02:11
dont think need anything else
02:11
<alkisg>
Import from my joomla db => maintained by asdf
02:11
<nubae>
import from website?
02:11
<alkisg>
Yes, that's what I need.
02:11
I don't care for any other import option! :)
02:11
<nubae>
hmmm, u might be able to do something with openID there
02:12
<alkisg>
That's why I'm saying it should be plugin based, needs vary a lot
02:12
No, I couldn't, because I can't control the sch.gr accounts. I only want to duplicate them locally.
02:13
<nubae>
hmmm, maybe... anything is better than nothing right now... but we should prioritize the options
02:13
I understand u're user case scenario
02:13
even though its probably a rare one... think most people would import user accounts from other servers
02:14
anyway, I'm game...
02:14
<alkisg>
csv should always be used as an intermediate step
02:14
<nubae>
or xml
02:14
<alkisg>
E.g. from LDAP or website or whatever to csv => then import it
02:14
Right, or xml.
02:14
<nubae>
but not from direct user accounts
02:15
<alkisg>
meaning?
02:15
<nubae>
.passwd and .shadow could be copied directly
02:15
<alkisg>
E.g. another plugin would be to merge users from old passwd / shadow files
02:15
<nubae>
right
02:16
would be nice even to have a dialogue allowing u to choose which users and groups
02:16
<alkisg>
...and actually, merging would be in the main "import users" function
02:16
Yes, with lots of dialogs for things that could go wrong (duplicate usernames, existing directories etc)
02:17
<nubae>
oh, and we need a cron to automate if necessary
02:17
<alkisg>
That would be in the main function, not in a plugin
02:17
ok, lets start from somewhere
02:17
<nubae>
its almost like a user server
02:17
clonezilla for users and groups
02:17shrek has quit IRC
02:17
<alkisg>
sbalneav, Lns and LaserJock were interested in this. Should we talk about it sometime in #edubuntu?
02:18
<nubae>
yeah, when do u wanna do that?
02:18
<alkisg>
I'm leaving for today and I'll be back Sunday evening (our time zone). So lets see if we can find them from Sunday onwards...
02:19
<nubae>
ok... we could probably schedule time via a google calendar
02:19
so we know when and what people are working on
02:19
<alkisg>
Sure. The first thing would be to make it work for Ubuntu
02:19
<nubae>
so right now, u are downloading the F10 version, right?
02:19
I'll do the same
02:19
<alkisg>
yes
02:19
To see it in action
02:20
Some months later, we may even look for supporting package kit or whatever...
02:20
<nubae>
hopefully, there is source with not too many deps
02:20
<alkisg>
Have a look at the discussion: http://www.mail-archive.com/ubuntu-desktop@lists.ubuntu.com/msg01544.html
02:20
<nubae>
well, we can start by making debs and putting them in a PPA
02:20
<alkisg>
It says "simple python script"
02:21
So we could start by making it run in Ubuntu... :)
02:22
<nubae>
yep.... I wonder why they chose not to go with it?
02:22
<redspike>
alkisg: Xt Warning: Missing charsets in String to FontSet conversion.
02:22
<alkisg>
nubae: Did they? I don't know, we could ask LaserJock to contact them for details
02:22
<redspike>
alkisg: ill get that when i try to runt the applications with when i run i sv_SE :(
02:22
alkisg: it works local on the server
02:23
<alkisg>
redspike: did you run the update-locale command, and updated the chroot?
02:23
with ltsp-update-image?
02:24
nubae: Maybe they didn't find someone to port/maintain it... I was willing to start this from scratch, so now that we have something to begin with, it'll be much easier.
02:25
<redspike>
alkisg: yes :/
02:25
<alkisg>
So now you have an /opt/ltsp/i386/etc/default/environment file?
02:26
<redspike>
ill think it have some with the fonts todo?
02:26
<alkisg>
Can you see the symbols in _any_ font?
02:26
<redspike>
on the server i can but not on the clients :(
02:27
<alkisg>
So I don't think it has something to do with the fonts, I think your locales aren't correctly installed
02:27
<redspike>
alkisg: in the chroot ?
02:27
<alkisg>
Yes
02:28
redspike: can you see these symbols in ldm, before the users log in?
02:29
<nubae>
alkisg: well the mail was from June 2008, so obviously they haven't done much to port it
02:29
<alkisg>
You should be able to... that's plain chroot, no server involved
02:29
<redspike>
alkisg: now, in ldm i have english :(
02:29
<alkisg>
redspike: if you try to type the euro character in the username field, can you?
02:30
(or any other of your local characters)
02:32
For ldm, I see that an sv.po exists, so you should have and ldm.po for your language in the chroot, and be able to see the text in swedish
02:33
<redspike>
alkisg: character works in ldm
02:33F-GT has quit IRC
02:33
<redspike>
and in xterminal after i have logged in, but not in that application, when i start it i cant se aao in the grafical interface :(
02:34
<alkisg>
The problem is only on a specific application? !!!
02:34
I thought you had a general problem...
02:35
For ldm language, see that you have /opt/ltsp/i386/usr/share/locale/sv/LC_MESSAGES/ldm.mo
02:35F-GT has joined #ltsp
02:35
<nubae>
alkisg: any idea where the source code for the users-manager is?
02:35pasmen has joined #ltsp
02:36
<redspike>
alkisg: where is the xorg.conf for the clients? i have found some ppl saded that remove :unscaled from the font can work.
02:36
<alkisg>
nubae: I'll be rebooting to F10, so I'll be able to send it to you in a little while...
02:36
redspike: intrepid doesn't use a xorg.conf by default
02:36
<nubae>
ah good.. stops me from having to install f10 8)
02:36
<alkisg>
If you want to add one, you can put it in the chroot, and specify X_CONF=path and CONFIGURE_X=true in lts.conf
02:36
<redspike>
alkisg: where can i change the FontPath in X ?
02:36
<alkisg>
Got to go. Bye!
02:37
<redspike>
ok thanx bye
02:37
<alkisg>
Ah, no!
02:37
I have virtualbox! :D :D :D
02:37
<nubae>
redspike: check the docs, there are lots of switches for X stuff
02:37
!docs
02:37
<ltspbot>
nubae: "docs" is For the most current documentation, see http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream
02:37* alkisg slaps himself
02:37* nubae helps alkisg out
02:44shrek has joined #ltsp
02:50polytan has joined #ltsp
02:51
<redspike>
nubae: can i move lts.conf to /var/lib/tftp/ltsp/i386 so i dont need to update image everytime i make changes ?
02:52intellishrek has joined #ltsp
02:52
<nubae>
redspike: absolutely, thats where it should be
02:53intellishrek has left #ltsp
02:54
<redspike>
nubae: i dont think it use it :( i add SOUND = N on the first line and restart the client but the soud still work?
02:55
<alkisg>
nubae, http://users.sch.gr/alkisg/temp/system-config-users.tar.gz
02:55shrek_ has joined #ltsp
02:56
<nubae>
redspike: u need [default] as the first line
02:56
alkisg: thanks
02:56shrek_ has left #ltsp
02:56polytan has quit IRC
03:00neoshrek has quit IRC
03:04tjikkun_work has joined #ltsp
03:09shrek has quit IRC
03:24cyberorg has joined #ltsp
03:37alkisg has quit IRC
03:40elisboa has quit IRC
03:42elisboa has joined #ltsp
03:51leio_ is now known as leio
03:55alkisg has joined #ltsp
04:14randra has joined #ltsp
04:35polytan has joined #ltsp
05:01polytan has quit IRC
05:09Guaraldo has joined #ltsp
05:12polytan has joined #ltsp
05:12
<polytan>
hi
05:12alkisg has quit IRC
05:12
<polytan>
how to use nvidia driver on the thin clien t?
05:12
in order to have 3D acceleraation
05:14
<cyberorg>
polytan, install nvidia drivers in the chroot
05:14
<polytan>
apt-get iunstall nvidia ?
05:16
<cyberorg>
polytan, yes something like that in the chroot, there should be instructions somewhere how to install packages in chroot
05:16shrek has joined #ltsp
05:16alkisg has joined #ltsp
05:18phantom has joined #ltsp
05:19F-GT has quit IRC
05:22polytan has quit IRC
05:29Gadi_eeepc1 has joined #ltsp
05:29Gadi_eeepc has quit IRC
05:33alkisg has quit IRC
05:54alkisg has joined #ltsp
06:22slashdotfx has left #ltsp
06:36phantom has quit IRC
06:37F-GT has joined #ltsp
06:43zirconiumks has quit IRC
06:53Gadi_eeepc1 has left #ltsp
06:57BrunoXLambert has joined #ltsp
07:01hanthana_ has joined #ltsp
07:08hanthana has quit IRC
07:19faustino333 has quit IRC
07:33robp2175 has joined #ltsp
07:34
<robp2175>
My volume very low on thin-clients. Anyone know what the fix is?
07:39evilx has joined #LTSP
07:39din_os has joined #ltsp
07:40Egyptian[Home] has quit IRC
07:42Egyptian[Home] has joined #ltsp
07:44alkisg has quit IRC
07:46alkisg has joined #ltsp
07:51dmaran has joined #ltsp
07:55six2one has joined #ltsp
07:58vvinet has joined #ltsp
07:58ogra has quit IRC
08:04
<robp2175>
I have tried changing all volume controls in lts.conf . The pulseaudio volume was already at 100%. gstreamer-properties appeared correct. Any thoughts? Anyone?
08:13ogra has joined #ltsp
08:28robp2175_ has joined #ltsp
08:29
<robp2175_>
My volume very low on thin-clients. Anyone know what the fix is? I have tried changing all volume controls in lts.conf . The pulseaudio volume was already at 100%. gstreamer-properties appeared correct. Any thoughts? Anyone?
08:38gurtbfrobe has joined #ltsp
08:40Gadi has joined #ltsp
08:40
<sbalneav>
Morning all
08:41robp2175 has quit IRC
08:43
<Gadi>
!s
08:43
<ltspbot>
Gadi: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
08:43vvinet has quit IRC
08:47din_os has quit IRC
08:47
<sbalneav>
Morning Gadi
08:47
<Gadi>
it is, isnt it?
08:47vvinet has joined #ltsp
08:48
<sbalneav>
Hmmm. Trying to learn ratpoison
08:48
It's being confusing : :)
08:53
<Gadi>
just dont lick it
08:53
(#1 rule of electrical engineering, and pretty much life as a whole)
09:04shrek has quit IRC
09:04
<sbalneav>
Oh, I don't know. I've had some pretty pleasent expericences licking things....
09:05
Lollipops, salt rims on Caesers, that sort of thing :)
09:06* dmaran is amazed he spent 4 hours troubleshooting the "Test" ltsp server and the whole time it was a bum network card
09:06
<ogra>
frogs ?
09:07warren has quit IRC
09:08
<sbalneav>
No, I've never licked a frog.
09:08
I've *eaten* frogs legs.
09:08
Maybe I licked my fingers afterwards. Can't remember.
09:09
<ogra>
heh
09:09
<sbalneav>
I've eaten chocolate covered crickets...
09:09
Snake...
09:10
ogra: So, we've got a new lunchcounter/restaurant near us called "Kraut king"
09:10
<ogra>
lol
09:10
<sbalneav>
it's been set up by a bavarian immigrant couple.
09:10
they have flipping AWESOME food.
09:10
<ogra>
selling curry sausage ?
09:10
<sbalneav>
What are the little flour dumplings called?
09:11
<ogra>
knödel
09:11
<sbalneav>
No, they've got schnitle.
09:11
yeah, that's it.
09:11
<ogra>
bah, no curry sausage ?
09:11
then they are not german :P
09:11
<sbalneav>
I had some kind of fried veal cutlet with a paprica sauce, and knodel for lunch the other day.
09:11
Yum.
09:11
They had goose too.
09:11
<cyberorg>
Gadi, the students are exploring usbip, it is already in 11.1, will try to figure out how to integrate that in ltsp if possible
09:12
<sbalneav>
Every time I eat there, I think of you :)
09:12
<ogra>
that called "zigeunerschitzel" i think
09:12
*that's
09:12alekibango has joined #ltsp
09:12
<sbalneav>
What would the curry sausage be called?
09:12
I'll ask for it the next time I'm there.
09:13
<ogra>
well, curry sausage (currywurst in german)
09:13
but if they are bavarian centric they might not have it
09:13
its *the* german food everywhere but in bavaria
09:14F-GT has quit IRC
09:14
<ogra>
(you know i dont consider bavaria germany ;) ... but at least they discovered the concept or democracy last year, so they are improving)
09:14
<sbalneav>
lol
09:14
<ogra>
i'm not kidding
09:14
:)
09:15
<sbalneav>
What, was a holdout of the third reich still governing bavaria?
09:15
Maybe that's why they immigrated to canada :)
09:15
<ogra>
after the weimar republic in 1920 they only elcted adolf ... after WW2 they formed their own party that had 70-90% in each election up to last year
09:16dirigeant has joined #ltsp
09:16
<sbalneav>
Wow
09:16
<ogra>
yep
09:16
<sbalneav>
What was the name of the party? (/me wants to read about this on wikipedia)
09:16
<ogra>
november was significant for bavaria ... there is a second party in power now for the first time of its existence
09:17
<nubae>
they have good beer though
09:17
:-)
09:17F-GT has joined #ltsp
09:19
<ogra>
sbalneav, http://en.wikipedia.org/wiki/Bavaria_state_election,_2008
09:19
sbalneav, FDP ... the liberals
09:21
though they lost more to the free voters but wouldnt build a coalition with them
09:26
<dmaran>
anyone have an idea why proc wont unmount in the chroot or outside
09:26
<ogra>
are you looking at the right /proc ?
09:27
<dmaran>
yeah, I tried from in the chroot and outside as in the recent email to the list
09:27
<robp2175_>
My volume is very low on thin-clients. Anyone know what the fix is? I have tried changing all volume controls in lts.conf . The pulseaudio volume was already at 100%. gstreamer-properties appeared correct. Any thoughts? Anyone?
09:28
<sbalneav>
robp2175_: What kind of audio hardware do you have?
09:28
<nubae>
thats a common problem, I've never had loud enough sound.... but from laptops adimittedly
09:29
<ogra>
didnt Gadi add a fix to adjust that via lts.conf ?
09:30
<nubae>
yeah
09:30
still not loud enough though
09:30
<dmaran>
ogra: always says it is busy even after it is done updating I cannot use -f either
09:30
<robp2175_>
none in the server. It is pulseaudio sound server. I am not sure exactly what the thin client uses. The documentation stinks. It is an HP e90
09:30
<Gadi>
set: VOLUME=100
09:30
<alkisg>
VOLUME, HEADPHONE_VOLUME, PCM_VOLUME, CD_VOLUME, FRONT_VOLUME in lts.conf
09:31
<robp2175_>
already set volume to 100, no worky
09:31
yup, set them all to 100
09:31
<Gadi>
robp2175_: get a shell on the client
09:31
and do alsamixer
09:31Guaraldo has quit IRC
09:31
<Gadi>
see if there are some additional controls
09:31
at the end
09:31* Gadi has come across some chipsets (such as VIA) with extended controls that need to be set, asa well
09:32
<robp2175_>
alsamixer also set to 100
09:32
<Gadi>
dont look at volume
09:32
press right arrow to the end
09:32
see if there is: VIA_***
09:32
or some other channel there
09:32twinprism has quit IRC
09:32
<robp2175_>
no via, just master
09:33
<dmaran>
ltsp-upate-image cannot umount proc either so I have to reboot in order to rebuild the image
09:33
!fail
09:33
<ltspbot>
dmaran: Error: "fail" is not a valid command.
09:34
<Gadi>
robp2175_: are all of the output channels at 100% on the client?
09:35
<robp2175_>
Yes all 100%
09:35
<Gadi>
sounds like a hardware problem
09:35
<robp2175_>
Tried two different clients. Both completely different hardware.
09:35
<Gadi>
scp a wav file to the client from the server
09:35alkisg has quit IRC
09:36
<Gadi>
and try playing it on the client
09:36
<robp2175_>
Tried fedora 9 with ltsp 5 and sound worked perfectly, but had other issues that made me decide to try fedora 10 with ltsp 5
09:36
<ogra>
dmaran, lsof
09:36
have a look what keeps it open, but its unusual, i have never seen that
09:37
<Gadi>
robp2175_: maybe warren has an idea of what changed
09:38
<dmaran>
ogra: I will the next time I am in the chroot
09:38
it is building now
09:38
after a reboot
09:39
<robp2175_>
how would i get in touch with warren
09:40
<ogra>
just hang around here
09:40
he is usually around (no idea why he isnt right now though)
09:40
<robp2175_>
OK will do
09:40
thanks for the help
09:42Appiah has quit IRC
09:52
<dmaran>
built fine after reboot
09:53hanthana_ is now known as hanthana
09:54spectra has joined #ltsp
09:55hanthana has quit IRC
09:56twinprism has joined #ltsp
09:57dirigeant has quit IRC
10:01mistik1 has quit IRC
10:03shrek has joined #ltsp
10:20japerry has quit IRC
10:20japerry has joined #ltsp
10:28
<_UsUrPeR_>
gadi - ping
10:32xxiao has joined #ltsp
10:33shrek_ has joined #ltsp
10:34
<Gadi>
_UsUrPeR_: one sec - ph call
10:34shrek_ has left #ltsp
10:36chrisinajar has quit IRC
10:36chrisinajar_ has joined #ltsp
10:37japerry_cat has joined #ltsp
10:38japerry has quit IRC
10:39
<Gadi>
_UsUrPeR_: hit me
10:39
<_UsUrPeR_>
uno momento
10:40
ok
10:40
gadi: ok, good to go
10:40randra has quit IRC
10:40
<_UsUrPeR_>
I have been trying to use the X_VIRTUAL = "xxxx xxxxx" lts.conf setting in F10 for a bit now
10:41
I am currently using 5.1.49 which I got from http://kojipkgs.fedoraproject.org/packages/ltsp/
10:41
and I'm using that version because 5.1.51 is kind of green
10:42
I am seeing no results from X_VIRTUAL with both versions
10:42* _UsUrPeR_ gasps for breath
10:43
<_UsUrPeR_>
I also want a pony ;)
10:43
<Gadi>
and have you checked the xorg log file and /var/run/ltsp_xorg.conf
10:43
<_UsUrPeR_>
oh, RIGHT
10:43* _UsUrPeR_ does that now
10:43
<_UsUrPeR_>
FYI: SCREEN_0X = shell does not presently work in 5.1.51
10:44
<Gadi>
/var/run/ltsp-xorg.conf
10:44
actually
10:44robp2175_ has quit IRC
10:44
<Gadi>
interesting
10:44
<_UsUrPeR_>
ok. Looking in the file now, I see a virtual setting of 3360 1050
10:44
unfortunately, xrandr sees it differently
10:44
:(
10:46
if I try something akin to xrandr --output VGA --left-of TMDS-1, I get an error that says "xrandr: screen cannot be larger than 1680x1680 (desired size 3360x1050)
10:46
"
10:51
<dberkholz>
apparently it's not listening to your virtual
10:51shrek has quit IRC
10:54
<Gadi>
yeah - you probably cannot get the video driver to set that
10:54
try increasing the video memory in the bios
10:54
or check the specs on the video chipset
10:58sepski has joined #ltsp
10:58
<dberkholz>
i'd check the X log to see if it says anything about virtual
11:01Egyptian[Home] has quit IRC
11:01
<_UsUrPeR_>
gadi: same chipset works properly in ubuntu 8.04 & 8.10. It also works properly if I use an xorg.conf
11:01
except for it doesn't re-position the screens
11:02
<Gadi>
does your xorg.conf just have the Virtual line?
11:02
<_UsUrPeR_>
yes
11:02
<Gadi>
so, it looks the same as /var/run/ltsp-xorg.conf?
11:02Egyptian[Home] has joined #ltsp
11:02
<_UsUrPeR_>
gadi: lemme pastebin it
11:03
<Gadi>
okey dokey
11:04
<ltsppbot>
Someone pasted "Section "Monitor" Identifier "" (38 lines) at http://pastebot.ltsp.org/222
11:04
<_UsUrPeR_>
whoops. forgot the name :)
11:06
<Gadi>
can you paste the ltsp-xorg.conf too?
11:06
<_UsUrPeR_>
gasi: ok
11:06
err gadi
11:06robp2175 has joined #ltsp
11:07
<Gadi>
and the xorg log file, while you're at it
11:07
<_UsUrPeR_>
ok, this will take a second :/
11:07
<Gadi>
ok - just scp them to the server
11:09
<_UsUrPeR_>
doing that
11:09
it was in the bios though
11:09staffencasa has joined #ltsp
11:12
<kc8pxy>
how do i keep my client view from being oversized?? my vbox client is sizing beyond the lcd's res.
11:12
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi: Xorg.1.log" (3063 lines) at http://pastebot.ltsp.org/223
11:13
"_UsUrPeR_" pasted "ltsp-xorg.conf" (4 lines) at http://pastebot.ltsp.org/224
11:16tjikkun_work has quit IRC
11:21
<sbalneav>
kc8pxy: vbox? That a hardware client, or virtualbox?
11:22cliebow has joined #ltsp
11:28
<Gadi>
_UsUrPeR_: try this:
11:28xxiao has left #ltsp
11:28
<Gadi>
edit: /opt/ltsp/i386/usr/share/ltsp/screen-session.d/XS85-virtual
11:29
and change "screen_hacks" to "display_hacks"
11:29
in both places where it occurs
11:30* Gadi wonders if the intel driver only honors the VIRTUAL argument when it is in the display subsection
11:30
<_UsUrPeR_>
rgr, done
11:30* _UsUrPeR_ will test that in a bit
11:31
<Gadi>
(it has required some trial/error with this new piece-meal xorg to know what exactly is *necessary)
11:36
<_UsUrPeR_>
the "Ignore" "True" is necessary for sure.
11:36
for LVDS
11:48shrek has joined #ltsp
11:59
<Gadi>
_UsUrPeR_: cant you turn it off with xrandr?
11:59robp2175 has quit IRC
11:59robp2175 has joined #ltsp
11:59
<Gadi>
_UsUrPeR_: like: xrandr --output LVDS --off
11:59
?
12:00
<_UsUrPeR_>
gadi: yes. Is there an X_IGNORE_OUTPUT function?
12:00
gadi: taht does, indeed, work
12:00
taht = that
12:00
<Gadi>
ah, so do this:
12:00warren has joined #ltsp
12:00
<Gadi>
XRANDR_OUTPUT_01 = "VGA"
12:01
XRANDR_OUTPUT_02 = "TMDS-1 --left-of VGA"
12:01
XRANDR_OUTPUT_03 = "LVDS --off"
12:03
<warren>
Gadi: 01 in that case is only a label?
12:03
<ogra>
warren, oh, does fedora use sendsigs.omit ?
12:03
<warren>
ogra: I never heard of that
12:04
ogra: no file of that name
12:04
ogra: what does it do?
12:04
<ogra>
warren, its prevents processes from being killed on shutdown
12:04
<warren>
ogra: killed by what?
12:04
<ogra>
look in /etc/init.d/sendsigs
12:04
<warren>
we don't have /etc/init.d/sendsigs
12:04
ogra: our shutdown uses killall5
12:04
<ogra>
there is /lib/init/rw/sendsigs.omit.d/ and /var/run/sendsigs.omit
12:05
right, thats what sendsigs executes
12:05* ogra checks where /etc/init.d/sendsigs lives in ... i would have assumed usplash
12:05
<warren>
you have the same killall5 command as us?
12:05
we have no usplash
12:05
<ogra>
yes
12:05
<Gadi>
warren: _01 .. _03 are for different outputs
12:05
<ogra>
err, upstrat
12:06RobertLaptop has quit IRC
12:06
<warren>
ogra: nothing named sendsigs
12:06
<Gadi>
warren: so, if you wanted, say 1024x763 on output 2, you can: XRANDR_MODE_02=1024x768
12:06
<warren>
Gadi: oh
12:06
<Gadi>
which corresponds to XRANDR_OUTPUT_02
12:06
<ogra>
ogra@osiris:/var/build$ dpkg -S /etc/init.d/sendsigs
12:06
initscripts: /etc/init.d/sendsigs
12:06
ah, not upstart
12:07
<warren>
ogra: aside from nbd-client what else do you typically omit?
12:07
<ogra>
i didnt omit anything yet (i wasnt aware you guys just dropped poweroff -fp)
12:10
but we surely have a mechanism for it in ubuntu
12:10
and i would expect fedora actually to have something alike
12:13
<warren>
I was told no.
12:14chupacabra has quit IRC
12:26six2one has quit IRC
12:31
<_UsUrPeR_>
gadi: the display hacks change seemed to produce results...
12:31
give me a second to see what's going on though :)
12:31
<Gadi>
ok
12:35warren has quit IRC
12:38din_os has joined #ltsp
12:40
<_UsUrPeR_>
gadi: X_VIRTUAL = "xxxx xxxx" causes my USB keyboard to stop working?
12:41
though it appears to work :)
12:41
<Gadi>
cool
12:42
did you add the ignore lvds stuff?
12:42
actually
12:43
can you try changing display_hacks back to screen_hacks and see if it works
12:43* Gadi wonders if the magic was in your other changes
12:43
<Gadi>
since X_VIRTUAL was working for stgraber on intel chipset as-is
12:44
<_UsUrPeR_>
I have noticed in XS20-intel-virtual-hack that X_VIRTUAL_AUTO = true
12:44
is this something that would affect my settings?
12:45warren has joined #ltsp
12:46* warren found a problem with ALSA
12:46
<warren>
ltsp-trunk/client/initscripts/ltsp-init-common
12:46
<_UsUrPeR_>
p.s. by "it works" I meant that the virtual display appeared to work, making one monitor appear half-wat. the XRANDR outputs don't seem to be taking effect as of yet...
12:46
<warren>
amixer used to work on the sound devices directly, bringing up the default volume
12:46
but then we added alsa-plugins-pulseaudio to the chroot
12:46
which redirected the default alsa device away from hardware
12:47
so we need to add -c0 to the amixer command line
12:48
<johnny>
so.. since sun bought virtualbox, does usb support come in the free version now?
12:48
<Gadi>
_UsUrPeR_: X_VIRTUAL_AUTO only affects something if you have XRANDR_MODE set
12:49
<_UsUrPeR_>
gadi: all changes disappeared when I changed from display to screen_hacks
12:49
<Gadi>
as in the effect of the changes
12:50
<_UsUrPeR_>
gadi: actually, it was just the virtual effects. The xrandr has not yet worked for me
12:50
<Gadi>
ok
12:50
so, it sounds like we *should* change scrreen to display upstream'
12:52
<_UsUrPeR_>
this is weird. It freezes keyboard input at the login screen
12:52
<Gadi>
warren: sounds good - are there any negative effects to that change?
12:52
<_UsUrPeR_>
I am able to get as far as entering a few chars in the username, then the keyboard stops.
12:52
mouse still works though
12:52
<warren>
Gadi: it will suddenly begin working again, making it a very loud surprise
12:52
<Gadi>
hehe
12:53
I'd say push it
12:53
<warren>
Gadi: it worked in the past prior to us making localapps work
12:53
<Gadi>
we should just stop making things work
12:55
<warren>
I'm adjusting the default for master from 100 to 90
12:56
<robp2175>
warren. How long before the fix for sound will be available?
12:56
<johnny>
hmm.. we need some sort of test suite
12:57
<warren>
robp2175: later today
12:57
<robp2175>
I appreciate the very quick attention to the issue. Truly
12:59vagrantc has joined #ltsp
12:59
<Gadi>
robp2175: you can edit /opt/ltsp/i386/usr/share/ltsp/ltsp-init-common
12:59
and add the -c0 to each amixer command
13:00
<robp2175>
Thats easy, sorry, I came in very late in the conversation and only read he had discovered and planned to fix the problem. Thanks Gadi
13:00
<warren>
robp2175: beware, it will triple or quadruple the software volme
13:01
<robp2175>
Thanks for the warning. I'd rather not lose my hearing. Would make the sound issue moot though.
13:01
<_UsUrPeR_>
gadi: is the XRANDR_OUTPUT = "xxxx xxxx" stuff supposed to be showing in the client's /var/run/ltsp-xorg.conf
13:01
?
13:01
<Gadi>
nope
13:02
<warren>
robp2175: it sounded like somebody else on the list had driver issues though
13:02
<_UsUrPeR_>
/var/log/Xorg.1.log?
13:02
<Gadi>
it executes the xrandr command
13:02
<warren>
robp2175: you are referring to k12osn?
13:02
<Gadi>
upon initialization of X
13:03
_UsUrPeR_: xinitrc.d/I10-xrandr
13:05
<robp2175>
fedora 10 ltsp5
13:06
<_UsUrPeR_>
gadi: ok, I'm in there
13:06
on the client
13:07
<warren>
Gadi: are you setup to test the ltsp-trunk change I pushed? I'm pretty sure that works on all distros, but we should test it.
13:07
<Gadi>
I am not atm
13:09
<warren>
hmm
13:09
Ryan52's change to ldminfod crashes.
13:10
language:nb_NO.UTF-8
13:10
Traceback (most recent call last):
13:10
File "/usr/sbin/ldminfod", line 237, in <module>
13:10
print lang_with_name(lang)
13:10
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe5' in position 23: ordinal not in range(128)
13:10din_os has left #ltsp
13:10
<ogra>
uuh
13:10
that cries for utf8-ification
13:10
<warren>
ogra: of what
13:10
<ogra>
the string handed over
13:11
presumably "lang"
13:12
<warren>
hmm... it crashes when I run it via nc
13:12
but not on the terminal
13:13
<ogra>
does nc mangle the 8bit chars ?
13:15
<warren>
nc is only dumping raw bits
13:15
<ogra>
i guess its environment is different though
13:17
but still, properly converting the string in the code should suffice ... "'ascii' codec" is surely not what you want ldminfod to run with
13:17din_os has joined #ltsp
13:17* ogra goes to look at the code
13:18
<vagrantc>
warren: we've seen it crash with output re-directed via | or > and presumably >>
13:18
though it doesn't actually crash
13:18
it just spits out a traceback and continues on, in my experience...
13:19
though i'm not getting the traceback when run on debian, only on fedora...
13:21
<warren>
vagrantc: it is always crashing on the same language here.
13:21
vagrantc: language:nb_NO.UTF-8
13:22
vagrantc: and it is failing immediately after the crash, so it never gets to session: lines
13:22
<ogra>
no, its crashing on the next one
13:22
<warren>
ogra: you sure?
13:22
<ogra>
at least thats how i read your traceback above
13:22
it spits out language:nb_NO.UTF-8 and then runs into probs
13:22
<warren>
it fails to print out the corresponding language-with-name: line
13:23
<vagrantc>
which would be the nb_NO line...
13:23
<ogra>
right
13:23
<vagrantc>
but there are *two* nb_NO lines
13:23
<ogra>
and contain norwegian umlauts
13:23
<warren>
huh?
13:23
<vagrantc>
i get it with nb_NO.UTF-8 enabled as well
13:23
<ogra>
you need to sanitize the string to proper utf 8
13:24
<vagrantc>
the question is... which string.
13:24
<ogra>
proper would be to do that with all of them :)
13:25
<warren>
ldminfod crashes when redirected, but works when not.
13:26
<vagrantc>
yes.
13:26* ogra points at http://www.reportlab.com/i18n/python_unicode_tutorial.html
13:27
<vagrantc>
i've read half a dozen unicode tutorials, and still haven't really figured out what actually needs to be unicodified.
13:27
<ogra>
you want to change: print lang_with_name(lang) to be: print lang_with_name(lang.encode('utf8'))
13:27
<warren>
Perhaps we should approach this in a different way.
13:28
My ldminfod without the crash transmits 18041 bytes, and reads a TON of files taking half a second every time it is run.
13:28
<vagrantc>
ogra: no, that would just encode the locale string to UTF-8
13:28
<warren>
Ryan52 looked into putting these translations in the chroot instead and transmitting only the LANG codes
13:28
but the .mo files alone would have been 3MB due to binary overhead
13:29
<ogra>
vagrantc, well, that should be done with all strings actually ... and best in a function instead of touching all strings
13:29
<warren>
I think maybe we should export this into a flat text file of similar syntax to ldminfod and parse it in the chroot based upon language codes given by ldminfod
13:29
That would make ldminfod faster and simpler.
13:29
and the flat file would be less than 16KB
13:29
<vagrantc>
we don't need to UTF-8 encode nb_NO.UTF-8
13:30
<ogra>
vagrantc, well, sane py code it to properly encode all strings you print out
13:31
<warren>
any other changes people want to ltsp-trunk?
13:31
<vagrantc>
so the real mystery is why UTF-8 encoding only issues a traceback if run through a pipe or re-directed or some such.
13:31
works totally fine if run manually without any redirects or pipes.
13:31
<johnny>
this won't be necessary with python 2.6 will it?
13:32
<ogra>
i dont think so
13:32
<warren>
vagrantc: I haven't been completely happy about this approach because it takes much longer on the server, adding half a second before hte login screen can pop-up
13:32
<ogra>
warren, if you have a million locales
13:32
<warren>
vagrantc: how do you feel about this alternative I suggested above?
13:32
<vagrantc>
warren: it's going to take a long time on the client, too.
13:32
<warren>
vagrantc: no it wont. It has to read only a single 16KB file
13:33
<vagrantc>
warren: how does it generate that file?
13:33
<warren>
vagrantc: during build of the ldm package.
13:33
<vagrantc>
warren: so if you have a server with a different set of supported languages?
13:33
<ogra>
warren, mimicing old gdm behavior ?
13:33
<vagrantc>
these are server-specific settings. just like /etc/X11/Xsesssion vs. /etc/X11/xinit/Xsession
13:34
<ogra>
vagrantc, he means to keep the lang->printout conversion list in ldm
13:34
<warren>
vagrantc: ldminfod hands only lang codes to ldm client
13:34
<vagrantc>
ogra: yes, i know.
13:34bobby_C has joined #ltsp
13:34
<warren>
vagrantc: ldm client looks at its own flat text file to see if it has translations for the given lang codes.
13:34
<vagrantc>
and if it doesn't?
13:34
<ogra>
so ldminfo would just work like before and the logic goes into ldm
13:35
<warren>
Ryan52 looked into doing this a different way yesterday, but it was not good because we were using .mo files that ended up 3MB in size due to a large number (one per language) and binary overhead.
13:35
<ogra>
if you steal gdm's file it should be relatively safe
13:35
to not miss any langs
13:36
<warren>
ldminfod's current output proves that ldm need only grow by like 16KB to have all the strings we need
13:36
less than 16KB
13:37litlebuda has joined #ltsp
13:38
<vagrantc>
ok, let's go with something like that, with the addition of an override file for missing or incorrect languages ... i.e. we build the list at install time, check for presence of an override file, and use those if present, falling back to the build-time file.
13:38
<warren>
huh?
13:39
override in what direction?
13:39
<ogra>
add more langs
13:39
<vagrantc>
warren: well, say you build your list and it's missing a language. or the translation for a language is wrong. or the translation for a language is politically unacceptible for a particular deployment. we need to override the built-in list.
13:39
<ogra>
though in ubuntu i know we have translations and locales for 124 languages and gdm has a list of them ... easiest would be to steal from the gdm source
13:40
<vagrantc>
sane defaults with possibility to override the default when it's wrong.
13:40
<warren>
hmm, yes probably best to self contain our own list of lang codes and traslations.
13:40
overrides makes this more complicated and confusing
13:40
Let's just self contain a copy, with instructions how to grab a new update of it.
13:41
<vagrantc>
and that update doesn't get overridden by package upgrades?
13:41
there needs to be some sort of override mechanism...
13:41
<warren>
ldm has its own list in its own source code
13:41
<vagrantc>
weather it's override per line, or override the whole file...
13:42
<warren>
Override it in your own build if you need to.
13:42
We really don't need more complication than this.
13:42
<vagrantc>
that entirely misses the point of my concern.
13:43
<warren>
you mean a user should be able to override without building it?
13:43
<vagrantc>
a system administrator, yes.
13:44
say, an override in /etc/ldm/languages
13:44
<ogra>
so you fear admins might *invent* new locales ?
13:44
<warren>
ogra: does that even work? glibc has to define a locale I thought.
13:44
<ogra>
you should know how many locales are available in your distro
13:44
warren, right
13:44
<vagrantc>
ogra: no, but new locales crop up now and again. you don't have the same locale list you did 5 years ago.
13:45
and the translation for some locales are politically charged...
13:45
<ogra>
but you wont forward port
13:45
or do you suddenly plan to ?
13:45
<vagrantc>
i.e. in some particular country, factionA calls the language FOO and factionB calls it BAR.
13:45
<ogra>
if you take todays locale list it will work for the past ten releases
13:45
<warren>
vagrantc: does gdm provide a way to override this without rebuilding it?
13:46
<ogra>
no
13:46
it even has a hardcoded list
13:46
<vagrantc>
i've raised a concern. i'm done.
13:46
<warren>
I deny the existence of your faction.
13:46
=P
13:46
<vagrantc>
yes, i noticed.
13:47
<ogra>
/etc/gdm/locale.conf
13:47
<warren>
vagrantc: It is a valid concern, however the far more widespread equivalents of this don't support that corner case. Should we complicate our software to do so?
13:47
<ogra>
thats the mapping file
13:47
not sure the new one still has it
13:47
ogra@osiris:/var/build$ wc -l /etc/gdm/locale.conf
13:47
397 /etc/gdm/locale.conf
13:47
<vagrantc>
by putting it in /etc, it's configurable by the local administrator...
13:48
<ogra>
#
13:48
# This is similar to a locale.alias file, but is not one. This file was
13:48
# autogenerated from the GNU libc locales source, with some editing.
13:48
#
13:48
<vagrantc>
that would be sufficient for me.
13:48
<ogra>
thats the header
13:48
so it should cover all currently possible locales libc can support
13:48
<vagrantc>
at least, i know debian has semi-same mechanisms for dealing with changes to files in /etc
13:48
<ogra>
given that you wont forward-port that should be totally safe
13:48
<warren>
vagrantc: I suppose it wouldn't be hard to support after we have the proposed functionality implemented... but I am not certain that it should be supported. Let's get more opinions.
13:48
<_UsUrPeR_>
trying to allow ssh into an F10 client, but it's saying that /etc/ssh/* is locked.
13:49
<warren>
_UsUrPeR_: locked in what way?
13:49
<ogra>
vagrantc, you will never support more locales than libc can define
13:49
there is no point making it configurable
13:49
<vagrantc>
ogra: but it's easy to rename a language with the same locale.
13:49
<_UsUrPeR_>
trying to run /etc/init.d/sshd start gives error: starting sshd: /etc/ssh/sshd_config: permission denied
13:49
<vagrantc>
or fix a typo in a language...
13:50
<ogra>
without renaming it in libc ?
13:50
<vagrantc>
yes.
13:50
<ogra>
i doubt that
13:50
<warren>
en_US English (Sovereign Dictatorship of Warren)
13:50
<vagrantc>
you display to the user FOO, the locale is still ee_FF
13:50
<ogra>
you would do the fix in the translation to user foo's lang
13:51
not in the actual source file
13:51
which needs a rebuild anyway, since you need a new .mo file
13:51
<_UsUrPeR_>
warren: in /etc/ssh/* on client, all files are inaccessible
13:51
can't more/less/vi
13:51
<warren>
ogra: this proposed way of lang code to strings displayed in ldm wouldn't use .mo
13:51
<ogra>
oh, why ?
13:52
<warren>
ogra: Ryan52 tried that earlier, it went from 16KB in total to over 3MB
13:52
<ogra>
on disk
13:52
<warren>
yes
13:52
<ogra>
in ram you only have the translation for $LANG
13:52
who cares about diskspace
13:52
<vagrantc>
warren: exactly!
13:52
en_US Cascadian (Peoples Autonomous Zone of Cascadian)
13:52
<ogra>
and whats 3M
13:52
<vagrantc>
you have to pull it over the network...
13:52
<warren>
16KB vs 3MB is huge
13:52
<vagrantc>
downloading 16K vs. 3MB is quite possibly significant on a large network of thin clients
13:53
<warren>
plus reading a single 16KB file is faster than reading many .mo files
13:53
<ogra>
many ???
13:53
<warren>
It isn't a single 3MB file
13:53
<ogra>
why would you read any .mo file thats not for $LANG ??
13:53
<warren>
ogra: it became huge in aggregate because each .mo file is a minimum of 4KB binary up to 48KB and there are many.
13:54
ogra: because ldm now displays the language name in its own localized language
13:54
<ogra>
ogra@osiris:/var/build$ ls -lh /usr/share/locale-langpack/de/LC_MESSAGES/gdm.mo
13:54
-rw-r--r-- 1 root root 113K 2008-11-14 02:15 /usr/share/locale-langpack/de/LC_MESSAGES/gdm.mo
13:55dmaran has left #ltsp
13:55spectra has quit IRC
13:57RobertLaptop has joined #ltsp
13:58
<warren>
vagrantc: so you like this new idea?
13:58
<vagrantc>
warren: build the list, at build time, stick the file in /etc. works for me.
13:59
<warren>
vagrantc: where in /etc/
13:59* ogra would just mimic gdm
13:59
<warren>
vagrantc: I personally will stick it wherever as long as it is LSB compliant, however I will not package the ability to override it because I think it is useless.
13:59
<ogra>
so /etc/ldm/locale.conf
13:59
<vagrantc>
ogra: yeah, that sounds good enough.
14:00
<warren>
go ahead and add the ability to upstream code
14:00
i just wont use it
14:00
<vagrantc>
warren: i'd rather see it in /usr/share with the possibility of an override in /etc, but you think that is useless.
14:00
<warren>
vagrantc: I would prefer /usr/share for the packaged version, and it can go ahead and check for the existence of /etc/something
14:00
<vagrantc>
warren: that sounds reasonable.
14:01
<Gadi>
ok, now repeat the debate in the 200+ other languages
14:01
go
14:01
<cliebow>
ru fdhr trh
14:02
<ogra>
Gadi, forst *you* generate a .mo file for each ... we'll paste it in there
14:02
<warren>
en_cliebow
14:02
<cliebow>
8~)
14:02
<warren>
en_cliebow.utf-8
14:02
<Gadi>
there goes ogra - talking to himself again...
14:02
;)
14:02
<cliebow>
finally someone Understands me
14:02
<warren>
My keyboard for that language has 500 buttons.
14:02
<ogra>
:P
14:03
<robp2175>
warren. FYI. Audio works perfect now. Thanks
14:04shrek has quit IRC
14:05elisboa has quit IRC
14:06
<warren>
robp2175: I didn't even push a fix yet
14:06
robp2175: you're hearing things =)
14:07din_os has left #ltsp
14:07
<ogra>
can you make sure it doesnt break on other distros first ?
14:07
<warren>
ogra: I talked with upstream ALSA and they said -c0 should always work.
14:07* ogra isnt sure stgraber included the alsa to pulse plugin
14:07
<ogra>
warren, i thought it was depending on the plugin
14:07
<warren>
ogra: no
14:07
<ogra>
ah, fine then
14:08
<warren>
ogra: if the plugin is installed in the chroot or not, -c0 always points at the hardware
14:08
<ogra>
good
14:12
<warren>
vagrantc: hmm, copying the source from gdm or iso-codes would add several megabytes of mostly unneeded strings to ldm source.
14:12
<ogra>
which source exactly ?
14:12
<robp2175>
warren Gadi told me how to fix the file manually
14:13
ltsp-init-common
14:13
<ogra>
warren, only /etc/gdm/locale.conf ?
14:13
<warren>
ogra: I don't have that file, can you paste it?
14:13
<ogra>
sure
14:13
<vagrantc>
warren: build dependencies?
14:13robp2175_ has joined #ltsp
14:14
<warren>
vagrantc: so yes, I'm thinking to build require iso-codes and parse it during ldm build, dumping to our text file to include in the ldm binary package.
14:14
<ltsppbot>
"ogra" pasted "locale.conf" (397 lines) at http://pastebot.ltsp.org/225
14:14
<warren>
vagrantc: then add your optional override file support, if exists then do it
14:14
<vagrantc>
warren: sounds reasonable to me.
14:15
<warren>
ogra: insufficient, lacks localized names
14:15
<ogra>
warren, you can even wipe all the .UTF-8 lines and just go with the first nn_NN part
14:15
give it to the translators :)
14:15
<vagrantc>
don't forget esperonto :)
14:15
which has no country.
14:15
<warren>
ogra: we already have iso-codes with this localized, we can auto-generate from it.
14:16
<ogra>
well, but at the cost of several MB
14:16
<vagrantc>
to the build dependencies.
14:16
<warren>
ogra: no
14:17
ogra: instead of generating our own .mo files we dump only the strings we want into a single text file
14:17
ogra: 16KB vs 3MB
14:19
<ogra>
warren, well, i meant give that file to the translators :) not parse it through gettext ...
14:19
but if you can generate something thats indeed easier
14:19
<warren>
ogra: why not rely on someone else who already did the work?
14:19
<ogra>
thats what i said :)
14:19
if you can generate something similar, doit :)
14:20
though i doubt iso-codes covers all ubuntu languages
14:20
we have a lot
14:20
but i'm fine omitting the corner cases
14:20
unless someone complains
14:22
ogra@osiris:/var/build$ apt-cache search language-pack base|grep -c base
14:22
458
14:23
oh
14:23
ogra@osiris:/var/build$ apt-cache search language-pack base|grep -v kde|grep -v gnome|grep -c base
14:23
156
14:23
thats more like it :)
14:24
how do you generate the iso-codes list so i can compare ?
14:24
<warren>
uh
14:25
I'm not really sure how it works.
14:25
ogra: iso also provides us a measure of plausible deniability. if someone blames us, we can say "the international standards organization" did it.
14:25
<ogra>
lol
14:25
<warren>
I'm not kidding, this has been useful in the past.
14:25
<big scary government>
14:26
<ogra>
well, my concern is rather that ubuntu might have langs not covered in the iso list
14:26
like xhosa ... or zulu
14:26
<warren>
Then fix it in iso-codes?
14:26
it should be fixed upstream anyway
14:27* ogra wonders what Oromo is :)
14:27npman has joined #ltsp
14:28mistik1 has joined #ltsp
14:29
<warren>
ogra: are you in favor of this new proposal?
14:29
<ogra>
yes, fine with me
14:30
lol
14:31
we have langpacks for "French, Middle (ca. 1400-1600)" and "Greek, Ancient (to 1453)"
14:32
<vagrantc>
ogra: it's in there.
14:32robp2175 has quit IRC
14:32
<ogra>
Interlingua ?
14:32
<vagrantc>
two variations
14:32
<ogra>
is that the new esperanto ?
14:33
<vagrantc>
no idea
14:33
<warren>
Broken English is the new esperanto
14:33
<ogra>
heh
14:33
hey and i wasnt aware we have actual latin
14:33
<vagrantc>
looks like 486 languages defined in iso_639.xml
14:34
<ogra>
ah, that should cover all i guess
14:34
<vagrantc>
granted, many of them are variations on a theme, but i think it's reasonable to say that covers a lot.
14:35
<warren>
posted to ltsp-devel
14:36
<ogra>
does it have Klingon ?
14:36
<vagrantc>
wow. Gwich'in.
14:36
<ogra>
we seem to have a langpack for that
14:36
<vagrantc>
name="Klingon; tlhIngan-Hol" />
14:37
one of the languages is, however: name="No linguistic content; Not applicable" />
14:38litlebuda has quit IRC
14:39
<warren>
vagrantc: is that body language? =)
14:40
<vagrantc>
the whole issue around political language distinctions... http://en.wikipedia.org/wiki/Burmese_language
14:41
at freegeek we get people who insist they speak Burmese, and others who speak Myanmar. and refuse to speak with one another.
14:41
linguists cannot distinguish the two. political distinctions in language names.
14:41litlebuda has joined #ltsp
14:42
<vagrantc>
something similar with some nepali languages and such, too.
14:43
<warren>
vagrantc: English speakers refuse to talk to each other too
14:43
"Those damn liberal hippies."
14:43
"Those damn conservatives."
14:44
<vagrantc>
and sometimes even say "those people don't even speak english"
14:46chrisinajar_ is now known as chrisinajar
14:46
<vagrantc>
but that's different than insisting they cannot speak with someone because they speak a different language, when there is no lingustic difference. not even dialect.
14:49robp2175_ has left #ltsp
15:03Gadi has left #ltsp
15:08cliebow has quit IRC
15:16litlebuda has quit IRC
15:17litlebuda has joined #ltsp
15:19din_os has joined #ltsp
15:25Guaraldo has joined #ltsp
15:26
<Guaraldo>
!seen jammcq?
15:26
<ltspbot>
Guaraldo: I have not seen jammcq?.
15:26
<Guaraldo>
!seen jammcq
15:26
<ltspbot>
Guaraldo: jammcq was last seen in #ltsp 6 days, 17 hours, 39 minutes, and 56 seconds ago: <jammcq> way cool stuff
15:35bobby_C has quit IRC
15:35
<warren>
hmm, has anyone seen ldm segfault in recent memory?
15:35
It used to here... but I don't see it anymore
15:36
oh
15:36
I think that was the race condition
15:36
with X starting
15:38
<Guaraldo>
warren: :-P
15:50* ogra only has good memories of ldm :)
15:51klausade has joined #ltsp
16:03BrunoXLambert has quit IRC
16:08din_os has left #ltsp
16:17RobertLaptop has quit IRC
16:30
<warren>
vagrantc: ogra: stgraber: I'm tagging ltsp-trunk again.
16:31
<ogra>
feel free ... though dont we slowly run our of version numbers ? :P
16:31
<warren>
ogra: I look forward to 5.1.472
16:31* vagrantc hopes 5.2 comes out first
16:32
<vagrantc>
didn't we even talk about doing a 5.2 release sometime this month?
16:32
<ogra>
\o/
16:32
<warren>
vagrantc: if so, let's agree on a feature freeze...
16:32
vagrantc: after we finally nail down a ldm localization plan that everyone is happy with
16:33
<vagrantc>
ldm is it's own project ... probably won't require changes to ltsp.
16:42* Guaraldo can work on ldm translation...
16:42
<Guaraldo>
to pt_BR
16:42
:-D
16:43Guaraldo has quit IRC
17:02sepski has quit IRC
17:41Gadi has joined #ltsp
17:42
<Gadi>
evening, all
17:43
fyi: I'm hacking on ltsp-genmenu a bit more so that it doesn't write into the user homedir at all
17:43
if it works, I'd like to push it before everyone gets too comfy with the current ltsp-genmenu approach
17:44
<vagrantc>
Gadi: YAY
17:44
<Gadi>
basically, it will put the files in the server's /tmp directory (within a mktemp dir) and set XDG_DATA_DIRS upon ldm login
17:45
based upon suggestions from alkisg and nubae
17:45
<vagrantc>
oh, fun!
17:45
<Gadi>
that way, if it should fail to cleanup, it should not impact the homedirs or concurrent sessions from multiple places or anything
17:46
<vagrantc>
excellent
17:46warren has quit IRC
17:46
<Gadi>
stay tuned...
17:46
:)
17:51npman has quit IRC
17:56
<ogra>
Gadi, in per user subdirs under /tmp ?
18:04
<vagrantc>
ogra: made with mktemp
18:04
<Gadi>
ogra: yeah
18:04
mktemp /tmp/ltsp-localapps-${USER}-XXXXXX
18:04
<ogra>
sweet
18:04
<Gadi>
is what I currently have
18:06
it will affect ltsp-trunk and ldm-trunk, tho
18:06
as ldm needs to set a client variable for the session
18:06
<ogra>
indeed
18:06
<Gadi>
so, it would be another one of these must update both
18:06
for it to work
18:13adrianorg has joined #ltsp
18:14
<Gadi>
Ryan52: ping
18:15adrianorg has left #ltsp
18:17
<Gadi>
does anyone know if the entire environment from before ldm is launched by the screen script is exported to the rc.d/ environment?
18:17
or if it is only certain vars?
18:17
<Ryan52>
Gadi: pong
18:17
<Gadi>
ah, Ryan52 i figured you might know
18:18
^^
18:18
<Ryan52>
it should still all be there.
18:18
<Gadi>
so, if I export a var in the ldm screen.d script, it should be present in rc.d/X99-...
18:18
right?
18:19
<Ryan52>
yep
18:19
<Gadi>
cool
18:22
<Ryan52>
ogra, vagrantc, I heard my name in a conversations...should I ready them? or is it useless boring stuff?
18:22
<vagrantc>
don't recall...
18:23
Ryan52: we made yet another plan for the ldm translated locale stuff
18:23
Ryan52: warren summarized it on ltsp-developer
18:25
<Ryan52>
how do you know it will only take 16K?
18:26
that xml file and the translations may have more than just all of the possible locales.
18:26
so if you're basing it on the size of ldminfod output, don't.
18:26
<vagrantc>
Ryan52: because it's basically doing what you already did for ldminfod, but instead doing it at package build time.
18:27
<Ryan52>
how do we get a list of locales?
18:27
"locale -a" won't work anymore.
18:27
because that doesn't work for Debian/Ubuntu.
18:28
<vagrantc>
on debian, we've got /usr/share/i18n/SUPPORTED
18:28
<Ryan52>
eww. ok.
18:28
<vagrantc>
fedora's got locale -a ... no idea with ubuntu...
18:28* Ryan52 nods
18:28
<vagrantc>
but yes, we'll need to find a way to get a list of available locales
18:29HrdwrBoB has joined #ltsp
18:29
<HrdwrBoB>
morning all
18:30
is there an easy way, from within an LTSP session, to know the IP of the terminal?
18:30
<vagrantc>
looks like debian lenny has 233 supported UTF-8 locales in /usr/share/i18n/SUPPORTED
18:30
<HrdwrBoB>
before I go and hack around it
18:30
<vagrantc>
should be in the $LTSP_CLIENT or $SSH_CONNECTION variables
18:31
<Ryan52>
$ locale -a | grep utf8 | wc -l
18:31
237
18:31
hrm.
18:31
<Gadi>
ls /usr/share/i18n/locales/
18:31
<HrdwrBoB>
excellent, thanks vagrantc
18:31* Ryan52 goes to diff them
18:32
<Gadi>
$ ls /usr/share/i18n/locales/|wc -l
18:32
277
18:32
<Ryan52>
$ ls /usr/share/i18n/locales/|wc -l
18:32
274
18:32
<Gadi>
(thats on intrepid)
18:33
<Ryan52>
that's on all Debian.
18:33
<vagrantc>
Gadi: i'm guessing some of those aren't actual locales
18:33
<Ryan52>
(not stable)
18:33
<Gadi>
maybe
18:33
<Ryan52>
de_LU@euro, translit_combining
18:33
<vagrantc>
Gadi: translit* ...
18:34
<Gadi>
$ ls /usr/share/i18n/locales/|grep translit|wc -l
18:34
12
18:34
as long as there's more stuff and not less
18:34
you can always throw away
18:34
<vagrantc>
heh
18:36
<Ryan52>
so should I work on the new plan?
18:37
<vagrantc>
Ryan52: can we stop you? :)
18:37
<Ryan52>
why?
18:38
<vagrantc>
Ryan52: well, as long as you're happy to work on it, i don't see any reason not to. :)
18:38
though sometimes it's good to let an idea settle a bit.... :)
18:47
Gadi: i don't see why you'd need to touch ldm-trunk to implement your master plan ... just need to move ltsp-trunk/localapps/ldm-rc.d/S02-localapps-menu to XNN-localapps-menu, i'd think ...
18:47
and of course, change the localapps-menu code a bit
18:48
and set CLIENT_ENV="$CLIENT_ENV XDG_MENU_DIRS= ..."
18:49
with a somedir=$(ssh -d mktemp -d /tmp/foo-barXXX) no?
18:49japerry_cat has quit IRC
18:49
<Gadi>
ah - I can set my own CLIENT_ENV...
18:49
didnt see that part
18:49* vagrantc wonders how local_apps_menu and blacklisting can interact...
18:49
<vagrantc>
or is it whitelisting
18:50
<Gadi>
well, I still want to set the mktemp dir name to an env var rather than store it in a file for later use - and that means setting it in the screen script
18:51
<vagrantc>
Gadi: i thought you were running mktemp on the server ... ?
18:51HrdwrBoB has left #ltsp
18:51
<vagrantc>
Gadi: wouldn't it have to be unique per login?
18:51warren has joined #ltsp
18:51
<Gadi>
yeah
18:51
Ill show you what I am crafting in a bit
18:51* vagrantc is curious about what Gadi has brewing
18:51
<Gadi>
hopefully it'll be clearer
18:52
but, thx for the clientevn idea
18:52
*env
18:52
<warren>
Ryan52: you around? Time for us to build ltsp into Fedora, so you can learn how.
18:52
<Ryan52>
oh yay! :)
18:53
<warren>
Ryan52: install mkdst on fedora (it's in the standard fedora repo)
18:54
Ryan52: I already bumped to 5.1.55 and ran "mkdst tag"
18:54
and pushed that
18:54
well, I usually
18:54
bump release.conf
18:54
mkdst tag
18:54
bzr viz
18:54
Just to visually inspect that it looks right.
18:54
mkdst tar --from-tag=ltsp-5.1.55
18:55
<Ryan52>
okay, I don't have an ltsp-trunk check out on my Fedora machine, so give me a few minutes to wait for the normal bzr lag..
18:55
<warren>
Ryan52: I pushed this hours ago
18:55
Ryan52: oh, and if you push and pull directly using bzr+ssh you don't see the lag.
18:56
<Ryan52>
I mean the lag of waiting for "bzr branch" to finish.
18:56
<warren>
oh
18:56staffencasa has quit IRC
18:57
<warren>
Ryan52: did you setup your cvs env to do /cvs/pkgs checkouts?
18:57
Ryan52: what I do is
18:57
mkdir /path/to/fedora-cvs-pkgs/
18:57
<Ryan52>
no, I don't even have a cvs check out over ssh...I needed to ask for help :)
18:57Egyptian[Home] has quit IRC
18:58
<warren>
Ryan52: I have a little shell script in there.
18:58
#!/bin/bash
18:58
export CVSROOT=:ext:wtogami@cvs.fedora.redhat.com:/cvs/pkgs
18:58
export CVS_RSH='ssh'
18:58
bash
18:58
Ryan52: so the shell that it ran can do checkouts of /cvs/pkgs
18:59
cvs co packagename
19:00
Ryan52: of course replace my name with your account name
19:01
<Ryan52>
ugh. lp doesn't know about this ssh key.
19:02
<vagrantc>
Ryan52: just bzr push it from somewhere on your local network ...
19:02
<warren>
you can change the ssh key in one or the other location
19:02
<vagrantc>
Ryan52: and then check it against the one on launchpad
19:02
<Ryan52>
ya, I know, I got it working.
19:05Lns has quit IRC
19:05
<Ryan52>
okay, I have it checked out, and I have ltsp-trunk checked out...now what?
19:05
<warren>
Ryan52: symlink ltsp/F-10/ltsp.spec into ltsp-trunk/
19:05
Ryan52: (just for convenience)
19:06
<Ryan52>
ok
19:07
<warren>
Ryan52: so if you're working on ltsp-trunk
19:07
Ryan52: so you pulled new upstream stuff after the last tag and you just want to test it.
19:07
Ryan52: edit ltsp-trunk/ltsp.spec
19:07
Ryan52: make Version match release.conf's Version
19:08
Ryan52: comment out the Release line and uncomment the Release line with the snapshot datestamp.
19:08
Ryan52: then "mkdst rpm" will build RPMS of your system's arch with whatever is in the ltsp-trunk directory, even if you haven't checked it in yet, with a unique datestamp.
19:09
Ryan52: understand?
19:09
Ryan52: give it a try
19:09
<Gadi>
vagrantc: lp:~gideon/ltsp/ltsp-trunk-genmenu
19:09
let me know what you think
19:09
you were right - I did it without touching ldm-trunk
19:09
<warren>
Ryan52: oh, there is a slightly improved mkdst-trunk that should be tagged and released.
19:10
Ryan52: do the same with mkdst-trunk, it is good practice because you do exactly the same thing to release that and build into Fedora.
19:10
<Gadi>
bbiab
19:11
<warren>
Ryan52: bzr+ssh://wtogami@bazaar.launchpad.net/~ltsp-upstream/mkdst/mkdst-trunk/
19:11
Ryan52: substitute your name
19:11* Ryan52 nods
19:11
<Ryan52>
ok....am I gonna have to actually test ltsp-trunk or can you do that? cause I don't have a working test environment for Fedora atm...
19:11
(or have you done that already?)
19:11evilx_ has joined #LTSP
19:11
<warren>
Ryan52: I already tested this ltsp-trunk and tagged it
19:11
<Ryan52>
oh, right.
19:11
okay..
19:11
<warren>
Ryan52: but why don't you tag and release mkdst-trunk now.
19:12
Get mkdst-trunk and /cvs/pkgs/mkdst
19:12
Add the symlink
19:12
Do a quick test build just for sanity
19:12
mkdst rpm
19:13
Then bump the Version in release.conf (in that case it is referring to another file, go find it.)
19:13
<vagrantc>
Gadi: my only comment is that there should be a separate mktemp -d call server-side.
19:13
Gadi: or it's susceptible to a symlink attack
19:13evilx has quit IRC
19:14evilx has joined #ltsp
19:14
<warren>
Ryan52: commit, mkdst tag, bzr viz to inspect for sanity, then bzr push
19:14
<vagrantc>
Gadi: and i'm just assuming TMP_XDG_DIR works :)
19:14evilx has quit IRC
19:15
<warren>
Ryan52: mkdst tar --from-tag=mkdst-0.11
19:15
Ryan52: let me know when you got that far.
19:17
<Ryan52>
how do I get "bzr viz"?
19:18
nevermind, I got it.
19:18
<warren>
Ryan52: yum install bzr-gtk
19:22Egyptian[Home] has joined #ltsp
19:26
<Ryan52>
bzr: ERROR: Could not acquire lock "(remote lock)"
19:26* Ryan52 waits
19:26
<warren>
Ryan52: which repo?
19:26
Ryan52: I've *never* seen that error message
19:26
<Ryan52>
the one on lp.
19:27
I tried to push, but pressed ctrl-c
19:27
and now launchpad is all broken.
19:27
If you're sure that it's not being modified, use bzr break-lock lp-45673808:///~ltsp-upstream/mkdst/mkdst-trunk/.bzr/branch/lock
19:27
<warren>
you might have done exactly the wrong timing
19:27
<Ryan52>
[ryan52@newbie mkdst-trunk]$ bzr break-lock lp-45673808:///~ltsp-upstream/mkdst/mkdst-trunk/.bzr/branch/lock
19:27
bzr: ERROR: Unsupported protocol for url "lp-45673808:///~ltsp-upstream/mkdst/mkdst-trunk/.bzr/branch/lock"
19:27
-.-
19:27
<warren>
yeah, i'm pretty sure that nobody is touching it
19:27
<Ryan52>
it's held by me.
19:28
cause I pressed "ctrl-c" at first.
19:28
<warren>
still stuck?
19:28
<Ryan52>
yah, I'll go bother the launchpad admins.
19:30* vagrantc still uses bzr+ssh
19:30
<warren>
I use bzr+ssh for both push and pull
19:30
<vagrantc>
or actually, sftp, it looks like.
19:32
<Ryan52>
okay, so now what?
19:32* warren looks at it
19:33
<warren>
ok, so you got the tar.bz2 from mkdst tar --from-tag=blah?
19:33
<Ryan52>
yep
19:34
<warren>
copy it into /cvs/pkgs/mkdst/devel/
19:34
Ryan52: cd into there
19:35
Ryan52: make upload FILES=whatever-version.tar.bz2
19:35
Ryan52: edit sources and .cvsignore to get rid of the old one
19:36
Ryan52: (you can use make new-sources FILES=whatever-version.tar.bz2 instead which wipes out whatever is already in sources and .cvsignore, but some packages you do want to keep other things in there)
19:36
Ryan52: edit mkdst.spec
19:37
Ryan52: make sure you have the real Release number instead of the datestamp one
19:37
Ryan52: set Version number
19:37
Ryan52: and write your own %changelog entry
19:37
Ryan52: cvs diff to sanity check
19:38
Ryan52: make prep (unpacks and applies patches if defined, in this case there are none)
19:38
Ryan52: make local (builds the RPM locally in place based upon whatever is in the ltsp.spec and sources file)
19:38
Ryan52: cvs ci
19:38
Ryan52: write your commit message
19:38
Ryan52: make tag && make build
19:38
<Ryan52>
what does "make upload" do?
19:39
<warren>
Ryan52: adds the tarball to the official repository of binary blobs, and writes it into sources and .cvsignore files in your directory along with sha1sum
19:39
<Gadi>
vagrantc: how can there be a symlink attack if mktemp is client side?
19:40
mktemp itself should prevent symlink attacks
19:41
<warren>
Ryan52: before you make tag, let me know and I'll look it over.
19:41
<Ryan52>
is there something that will automatically add the "* Thu Jan 29 2008 Ryan Niebur <ryanryan52@gmail.com> - 0.11" to the %changelog section?
19:42
<warren>
Ryan52: you can define a vim macro and I think people have emacs scripts, but I never used either.
19:42
<Ryan52>
should the changelog entry be useful?
19:43
or should I just do "- 0.11" like you did?
19:43
<warren>
Ryan52: in this case 0.11 is fine
19:43
Ryan52: I only write something there if I think it is important that somebody know
19:43
in this case I'm the only person using mkdst in Fedora
19:43
well, now you.
19:44
<Ryan52>
error: %changelog not in descending chronological order
19:44
<warren>
Ryan52: you probably didn't bump year?
19:44
<Ryan52>
oh, heh, it's 2009 :p
19:46
okay, I think it committed.
19:46
<warren>
Ryan52: I will soon need to refocus on energy on other parts of the distro to prepare for RHEL6. I will occasionally help if necessary to keep ltsp-upstream compatible with Fedora, but I'm hoping you and another person I find to train can do most of it.
19:47* warren checks
19:47
<warren>
If you and (somebody else I find) can't keep Fedora updated with upstream, then I will only be able to resync every 2-3 weeks.
19:47
instead of daily
19:48
Ryan52: your release number is 2, did you want it 1 or 2?
19:48
Ryan52: probably 1 in this case.
19:48
<Ryan52>
...what's that mean?
19:48
<warren>
Name: mkdst
19:48
Version: 0.11
19:48
Release: 2%{?dist}
19:48
<Ryan52>
oh, I see.
19:48
<warren>
doesn't match your changelog
19:49* Ryan52 fixes
19:50
<warren>
Ryan52: So devel is currently targeting Fedora 11 which is a few weeks away from beta freeze. Since ltsp isn't critical to the core distribution you can go ahead and update components in devel all the way until the hard freeze before RC.
19:50TheProf has joined #ltsp
19:51
<warren>
Ryan52: I typically test things on the side and only after lots of testing, rebuild the latest version for all current fedora distros. Yes, this is different from Debian. I never backport anything.
19:51
I'm pretty careful about changes that don't break older releases.
19:52
<Ryan52>
okay...which reminds me, why are you having me touch ltsp in fedora 10 and not devel?
19:52
okay, so now do I "make tag && make build" mkdst?
19:52
<warren>
Ryan52: I haven't been bothering to update devel for ltsp yet because fedora 11 is broken for netboot at the moment.
19:52
<Ryan52>
ah.
19:52
<warren>
Ryan52: yeah, go ahead
19:53
https://bugzilla.redhat.com/show_bug.cgi?id=481078
19:53
<TheProf>
Hello. I was having some issues with printing and I believe I have isolated the problem but not the solution: I am missing the jetpipe application. It wasn't installed as part of CentOS. Where can I get it from?
19:53
<warren>
They rewrote how it does dhcp in Fedora 11 and broke it in the process.
19:54
TheProf: you're referring to an ancient version of LTSP
19:54
<Ryan52>
how nice of them.
19:54
<warren>
TheProf: there isn't a good solution to run a modern LTSP on CentOS.
19:55
<TheProf>
warren: I think you are referring to LTSP 4.x something?
19:55
<warren>
TheProf: yes
19:55
TheProf: https://fedorahosted.org/k12linux/ you can use modern LTSP on Fedora 10 and upgrade into CentOS6 later.
19:55
<Gadi>
vagrantc: your wish is my command (update ltsp-trunk-genmenu)
19:55
<TheProf>
oh man. This after spending 2 months learning k12linux :)
19:55
<Ryan52>
1092396 build (dist-f11, /cvs/pkgs:rpms/mkdst/devel:mkdst-0_11-1_fc11) completed successfully
19:56
<warren>
TheProf: k12linux?
19:56
<TheProf>
sorry, k12ltsp
19:56
my mistake.
19:56
<warren>
TheProf: k12linux is actually a bit easier to setup, try the K12Linux live server media
19:56
TheProf: it does a lot of K12LTSP automatically
19:57
<Gadi>
TheProf: LTSP 4.2 didnt use jetpipe
19:57
it used lp_server
19:57
<TheProf>
I was always used to K12ltsp since FC 4 I think because it was boot and go. Everything on the system is working fine for the small setup we have. The only thing that seems to be missing is this jetpipe
19:57vagrantc_ has joined #ltsp
19:57
<Gadi>
same thing - different executable
19:57
<vagrantc_>
Gadi: because you need to call mktemp on the client as well as the server
19:57
<warren>
Ryan52: ok, copy mkdst.spec sources .cvsignore into F-10 directory, cvs diff to sanity check, commit, tag and build
19:58
Ryan52: let's not bother with F-9 since nobody will use it
19:58
<Gadi>
vagrantc_: I just pushed a change to do it as you suggest
19:58
<Ryan52>
okay.
19:58
<warren>
Ryan52: https://admin.fedoraproject.org/updates/ while that's building login here and poke around
19:58
<TheProf>
Gadi: I was reading the instructions online and I thought the wiki was referring to using the jetpipe to identify the socket for usb printers?
19:59
<Gadi>
TheProf: perhaps you read a wiki for LTSP5?
19:59
<TheProf>
possibly yes
19:59
<vagrantc_>
Gadi: i think it would be simpler if you generated it in a local directory, and then created a new directory, and scped the whole thing over all in one go.
19:59
<warren>
TheProf: Got a USB stick? You should try K12Linux Live Server. You can demo it entirely from the USB stick to determine if things would work smoothly with a real install.
19:59
<Gadi>
TheProf: if you get a shell on the client, do a ps -ef|grep lp
19:59
<warren>
TheProf: no need to touch your hard drive.
20:00
<Ryan52>
warren, how do I revert a change with cvs?
20:00
<TheProf>
Gadi: I've got shell - I will check that now.
20:00
<warren>
Ryan52: there's no good way, just undo it and commit again
20:00
<Ryan52>
no, I mean equivalent of "svn revert file" or "git checkout file"
20:00
<vagrantc_>
Gadi: LOCAL_TMP=$(mktemp -d) ; SERVER_TMP=$( ssh mktemp -d) ; generate_in $LOCAL_TMP ; scp $LOCAL_TMP $SERVER:$SERVER_TMP
20:00
<Gadi>
vagrantc_: thats what I did at first
20:00
<TheProf>
warren: you are correct in that it probably is an easy process. I'm just on a time crunch where" every day that is delayed is a day that it could have been done if it was windows" mentality.
20:00
<warren>
Ryan52: delete the local file and 'cvs up filename'
20:01
<vagrantc_>
Gadi: but you didn't run mktemp on the server ... you merely copied the locally generated directory over to the server's /tmp
20:01
<warren>
TheProf: you are using software that hasn't been developed for like 4+ years now
20:01
<Gadi>
vagrantc_: ah, right
20:01
<warren>
TheProf: I actually haven't used the old LTSP in maybe 7 years myself.
20:01
<Gadi>
well, in the one I just pushed, I just do it all on the server
20:01
<TheProf>
warren: it's a step up :) until December I was running k12ltsp based on FC 5
20:02
<warren>
TheProf: I have absolutely no clue how printing worked back then.
20:02
<vagrantc_>
Gadi: so the first attempt was actually closer to what i was suggesting, with one glitch.
20:02
<TheProf>
if it wasn't broke I didn't change it.
20:02
<Gadi>
ok
20:03
<TheProf>
Gadi: that line on the client shell returns: sbin/lp_server -n 9100 -w -d /dev/usb/lp0
20:03litlebuda has quit IRC
20:03
<Gadi>
TheProf: so, you are fine
20:03
<TheProf>
sweet!
20:03
<Gadi>
why cant you print?
20:03
<Ryan52>
warren, should I merge your changelog entry from F-10 branch into the devel branch?
20:03
<TheProf>
well I know it prints fine when the printer is directly into the server
20:03
<Ryan52>
or would that destroy the world?
20:03
<warren>
Ryan52: oh
20:03
<Gadi>
TheProf: usb printer?
20:04
<TheProf>
but when I moved it to a client and tried to point it to the client, it keeps asking for the jetpipe
20:04
Gadi: yes
20:04vagrantc has quit IRC
20:04
<Gadi>
TheProf: can you telnet to the thin client IP port 9100?
20:04
<warren>
Ryan52: oops, I should have checked that devel was actually equal to F-10 before
20:04
<Gadi>
(what asks for jetpipe?)
20:04
<warren>
Ryan52: not a big deal, just make it in sync later.
20:04
<TheProf>
Gadi: sorry I should be clear.
20:05
<Ryan52>
warren, what's that mean? should I commit your changes into the devel branch?
20:05
<TheProf>
any print job sits there in the queue processing for a while
20:05
<warren>
Ryan52: or bump devel to release 2 and make it equal now.
20:05
<TheProf>
Gadi: then it suddenly says the job is complete but nothing comes out
20:05
researching through the mailing list archives shows this is a symptom of a lack of jetpipe
20:05
<warren>
Ryan52: after you "make tag" there is no going back and changing it again, so bump the release number
20:05
<TheProf>
searching through the archives about that shows it was missing from recent releases.
20:05
Hence the situation.
20:06
<Gadi>
TheProf: gotta love the internet - more useless info than anywhere else
20:06
TheProf: can you telnet to the thin client IP port 9100?
20:06
and did you make the print queue on the server as a jetdirect queue to port 9100 for that IP?
20:06
<TheProf>
Oh yes. xkcd has a strip exactly about that idea.
20:07
I did point the print queue on the server to a jetdiret queue to port 9100
20:07
the exact line I used is:
20:07
socket: 192.168.2.259:9100
20:08
<Ryan52>
okay, I'll just leave it alone then..
20:08
<TheProf>
as per wiki. I will try to telnet now
20:08
sorry, 250:9100
20:08
<Ryan52>
warren, so what do you do if you want to make changes but not actually upload them? like if they aren't high enough priority to need an upload?
20:09
okay, committed in F-10...should I tag and build?
20:09
<TheProf>
Gadi: yes I can telnet to 9100
20:09
<warren>
Ryan52: you can commit changes but not tag, or tag but not build.
20:09
<Gadi>
TheProf: did you set up the queue in the GUI?
20:09
<Ryan52>
ok.
20:09
<warren>
Ryan52: "make upload FILES=stuff" permanently forever stores stuff on the server, so try to minimize doing it
20:10
<TheProf>
Gadi: once gui, once cups, and once via modifying the printers.conf file
20:10
<Gadi>
heh, no wonder nothing works ;)
20:10
<TheProf>
each time I erased and recreated the printer :)
20:10
<Gadi>
doesnt centos have JetDirect as an option in the printer wizard gui thing?
20:11
<TheProf>
it does yes
20:11
but it seems the script is missing that it is trying to call
20:11
<Ryan52>
warren, so should I "make tag && make build" now?
20:11
<Gadi>
what script is that?
20:11
<warren>
Ryan52: yes
20:11
Ryan52: you're pushing the F-10 build as an official fedora update
20:11
<TheProf>
whereis jetpipe shows nothing
20:12
<warren>
Ryan52: so build it
20:12
<Gadi>
forget jetpipe
20:12
<warren>
Ryan52: then go to https://admin.fedoraproject.org/updates/
20:12
<TheProf>
ok - forgotten
20:12
<Gadi>
what gui on the server knows about jetpipe?
20:12
<warren>
Ryan52: normally we push updates as "testing" first, but in this case the risk is nothing, so push it directly to stable.
20:12
<Gadi>
I am talking straight printer admin gui
20:12
<TheProf>
Gadi: alright loading it up - do you prefer that or cups gui?
20:12
<Gadi>
whichever is prettier
20:12
:)
20:13
<TheProf>
loading up system-config-printer
20:13
got my printer here
20:13
device URI
20:13
<Gadi>
on Ubuntu, if you choose JetDirect it asks you: Host, Port Number
20:13
quite easy
20:13
yours makes you put the full URI?
20:13
<TheProf>
it asks for it in stages.
20:14
First stage it is asking me for connection type
20:14
options are: appsocket/hp jetdirect, ltp 1, serial, ipp, lpd, windows samba, other
20:14
<Gadi>
socket://192.168.2.250:9100
20:14
that should be your URI
20:15
if you need it
20:15
<TheProf>
ok so I've picked that first one appsocket/hp direct, and for the hostname I've put the IP, and the port was prepopulated with 9100
20:15
done.
20:15
<Gadi>
cool
20:16
<TheProf>
should I print a test page?
20:16
<Gadi>
yup
20:17
<TheProf>
hmm message I'm seeing on the status page is Attempting to connect to host 192.168.2.250 on port 9100
20:18
<Gadi>
TheProf: there is also a /var/log/cups/error_log you can check
20:18
<TheProf>
checking now
20:19
<Ryan52>
warren: "If you feel that community testing is unnecessary for your update, you can choose to push it straight to the stable fedora-updates repository instead."
20:19
warren: how do I choose that?
20:20
<warren>
Ryan52: https://admin.fedoraproject.org/updates/
20:20
<Ryan52>
ugh
20:20
nothing like "make update"? :(
20:20
do I still need to do "make update" first?
20:20
<warren>
what are you reading?
20:20
I've never used make update.
20:20
<Ryan52>
http://fedoraproject.org/wiki/PackageMaintainers/UpdatingPackageHowTo
20:21
it was linked from the page you just gave to me.
20:21
<TheProf>
Gadi: wow that's a long file. I'm just trying to find the relevant timestamp
20:21
<warren>
Ryan52: oh, i've never used the text interface to bodhi before
20:22
<Ryan52>
warren, anyway, the build is finished, what do I do now?
20:22
<warren>
Ryan52: I suppose try it both ways just to be sure you know what's going on.
20:22
Ryan52: push it as a stable update for F-10
20:22
<Ryan52>
how?
20:22
<warren>
Ryan52: try it the web page way first
20:22
Ryan52: https://admin.fedoraproject.org/updates/
20:23
<Ryan52>
oh, that page makes more sense when you're logged in...:p
20:24
<warren>
Ryan52: due to the client/server nature of ltsp, going through the regular updates-testing procedure recommended for standard fedora updates is almost useless for us.
20:24
Ryan52: I came up with a different procedure altogether
20:24
<TheProf>
Gadi: so this is the most complex log I've ever seen. I also think that it's taken the print job and then poof it's gone. Just checking the last bit
20:24
<warren>
Ryan52: it is complex and only in my head, so I'll have to explain it tomorrow, I'm super exhausted now.
20:24
<Ryan52>
okay.
20:24
is this bugfix or enhancement?
20:24phpstar has joined #ltsp
20:24
<warren>
Ryan52: go ahead and use the same procedure to build ltsp-5.1.55 into F-9 and F-10
20:24
Ryan52: bugfix
20:25
<TheProf>
Is there a pastebin for this channel?
20:25
<phpstar>
hello everyone
20:25
<Gadi>
!pastebot
20:25
<ltspbot>
Gadi: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
20:25
<TheProf>
Gadi: that's just cool.
20:25
<Ryan52>
warren: "Enable karma automatism"?
20:25
whatever that means..
20:25
<warren>
Ryan52: uncheck it for this one
20:25
Ryan52: you're not using the testing process
20:25* Ryan52 nods
20:25
<phpstar>
can i use LTSP to deploy Win XP on thin clients ??
20:25
<warren>
Ryan52: I might put a pidgin update into updates-testing with 6 karma points to auto-push it to stable.
20:25* Ryan52 clicks teh button
20:27
<warren>
Ryan52: actually do this 1) go ahead and use the same procedure to build ltsp-5.1.55 into F-9 and F-10 2) Copy sources .cvsignore, spec file into devel and build it there to match. the cvs diff will be huge in that case but look it over.
20:27
Ryan52: and copy F-10 to devel for ldm and ltspfs too.
20:27
Ryan52: ldm is special because it has more stuff in sources than just the source tarball, so watch out.
20:28
Ryan52: ldm is also a simplest example of a patch in /cvs/pkgs to apply on the pristine tarball, so look at how that works.
20:28
Ryan52: if you add another file to the directory referenced as a SOURCE or PATCH line in spec, you need to cvs add it before checkin.
20:29
<ltsppbot>
"TheProf" pasted "Relevant part of cups error log. The job was submitted at 21:17:05 as job no 17" (10015 lines) at http://pastebot.ltsp.org/226
20:29
<warren>
Ryan52: many fedora packages differ between fedora releases, but some packagers put effort into making the sources and spec files exactly identical.
20:29
<TheProf>
Warning- it's a long file
20:29
<warren>
Ryan52: "cvs co pidgin" for a very complicated example... it supports every Fedora release and RHEL releases since like 2004
20:30
Ryan52: all from the same spec
20:30
<Gadi>
vagrantc_: pushing fix that I think is what you mean
20:30
<TheProf>
It starts right when the job was submitted. I don't see an errors.
20:31
<warren>
Ryan52: https://admin.fedoraproject.org/updates/mkdst-0.11-1.fc10 good job
20:33
<TheProf>
Gadi: it does seem to be going to the right location: D [29/Jan/2009:21:17:05 -0500] [Job 17] envp[21]="DEVICE_URI=socket://192.168.2.250:9100"
20:33
<warren>
Ryan52: understand so far?
20:33
<Ryan52>
yep
20:34
<Gadi>
TheProf: on the printer properties, can you set any options wrt bidirectional?
20:34
<TheProf>
checking now
20:35
<warren>
Ryan52: let me know when the ltsp F-9 and F-10 builds are done, I'll pull them into my test thing...
20:35
Ryan52: brb shower
20:36
<Gadi>
TheProf: http://h10025.www1.hp.com/ewfrf/wc/fastFaqLiteDocument?lc=en&cc=uk&dlc=en&docname=bpj06101
20:36
<TheProf>
Gadi: nothing default but there is an option to add a job option
20:36
<Gadi>
is your printer on that list?
20:36phpstar has left #ltsp
20:36
<TheProf>
Checking now
20:37
Gadi: yes, 6th one down:HP LaserJet 1000
20:38
<Ryan52>
warren, ok, I'm taking a break to eat dinner first...I haven't eaten a full meal yet today.
20:38
<Gadi>
well, you may be out of luck, then
20:38
read the introductory paragraph
20:39
<TheProf>
hmmmmm
20:39
<Gadi>
unfortunately, we live in a day and age where, just like modems, printers too may not behave like real printers without the presence of a PC
20:39
:)
20:39
<TheProf>
so how does it print when directly connected to the server?
20:39
is it because of the communication link between the two?
20:39
<Gadi>
read the intro paragraph
20:40
some printers cannot be jetdirect'd
20:40
<TheProf>
*sigh* they HAD to go and buy the cheapest laser printer out there.
20:40
<Gadi>
Host-based (sometimes referred to as GDI printers). A printer that relies on the computer's processor to rasterize the print job. Most host-based printers use the GDI interface built into Windows (hence the "GDI printer" term), so there is no need to convert the data to PostScript(R), PCL, or other printer language. PPA (Pri
20:40
doh
20:41
with JetDirect, the printer just gets raw data
20:41
<TheProf>
and then it transforms it onboard.
20:41
<Gadi>
TheProf: you can try *not* using the postscript driver
20:41
<TheProf>
this one can't do it.
20:42
<Gadi>
maybe another driver would work like hpijs, but I dont think so
20:42
I dont think you can send it any raw printer language
20:43
<TheProf>
I did have it printing before -- I did it by hooking it up to a windows 98 Pentium II and sharing it to the LTSP server. I was hoping to gain an extra thin client by doing it this way
20:43
<Gadi>
I faced that issue once before with a client that got cheap printers
20:43
<TheProf>
argh
20:43
<Gadi>
took forever to figure it out
20:44
<vagrantc_>
Gadi: looks pretty good. though you probably want "mktemp -d", no?
20:44
<Gadi>
prolly
20:44
lemme change that
20:44
<TheProf>
Gadi: I appreciate the help. At least it resolves this headache even if it doesn't get me another client.
20:45
Gadi: BTW, this was the comic I was referring to when you were talking about the internet full of useless info http://xkcd.com/214/
20:45
<vagrantc_>
Gadi: my only fear is with: CLIENT_ENV="${CLIENT_ENV} XDG_DATA_DIRS=${TMP_XDG_MENU}:\${XDG_DATA_DIRS}"
20:45
<Gadi>
TheProf: np
20:46
<vagrantc_>
Gadi: will it have XDG_DATA_DIRS available to pre-pend TMP_XDF_MENU ?
20:46
<Gadi>
vagrantc_: which part the backslash?
20:46
yeah, I was curious about that myself - and whether XDG_DATA_DIRS is overwritten by Xsession
20:46
<vagrantc_>
Gadi: just wondering what sets XDG_DATA_DIRS
20:46* Gadi investigates
20:50
<vagrantc_>
that could easily be the stumbling block for this idea...
20:50* vagrantc_ checks it out too
20:51
<TheProf>
Gadi: Sorry. A problem I just realized. Prior I used to have the server in the same room as the computer lab. So there was one NIC for the clients, and one NIC for the internet. The Windows machine plugged into the switch that had the NIC internet and so was able to communicate that way. Because of the kids literally kicking the server, I've moved it out of the lab so there's only the one NIC for the LTSP clients coming into t
20:52
<warren>
Ryan52: if you throw "koji build --scratch whatever.src.rpm" at koji and the build succeeds, then you can grab your scratch builds from http://koji.fedoraproject.org/scratch/ryan52/
20:53
Ryan52: your official builds after 'make tag' and 'make build' go into http://kojipkgs.fedoraproject.org/packages/mkdst (you can see your build there right now)
20:53
<Ryan52>
how do I make a scratch build with "make blah"? is there a way?
20:53
<warren>
hmm
20:54
Ryan52: look in ../common/Makefile.common, it defines all the make targets. Lots of them I've never used before.
20:54
<Ryan52>
and for ltsp, which way do you want me to build it?
20:54
<warren>
Ryan52: I usually "make test-srpm" which spits out a .src.rpm of whatever is in the current directory (even uncommitted) and throw that into "koji build --scratch dist-f10 filename.src.rpm"
20:55
Ryan52: build official ltsp-5.1.55 in F-9, F-10 and devel, based upon the F-10/ltsp.spec
20:55
<Ryan52>
okay.
20:55
<warren>
Ryan52: do the same thing from ldm/F-10 to ldm/devel
20:55
(be careful the contents of ldm is more complicated with extra sources and a patch)
20:55
Ryan52: and ltspfs/F-10 to ltspfs/devel
20:56
<Ryan52>
and just upload those?
20:56
<warren>
you aren't changing the tarball in ldm and ltspfs
20:56
<Ryan52>
without testing? 0.o
20:56
<warren>
you're only copying F-10 to devel
20:56
<Ryan52>
oh.
20:56
<warren>
and building devel
20:56
<Ryan52>
ok.
20:57
<warren>
Ryan52: changelog entry for 5.1.55 should summarize the only important change since .54, which was set default volume higher
20:58
Ryan52: OK I'm sleeping real soon. One last thing for today.
20:58
<vagrantc_>
Gadi: i don't see anything that sets XDG* ...
20:58
<Gadi>
me neither
20:59
<warren>
Ryan52: I plan on doing a new media release after ltsp-5.1.55, possibly with ldm-2.0.30-2.fc10 because it is pretty proven stable. Maybe even tomorrow if testing looks good with ltsp-5.1.55. But I can delay a few days if ldm-2.0.31 with the proposed localization change works well.
20:59
<Gadi>
prolly just appends a pre-existing var
20:59
to whatever is built in
20:59
<warren>
Ryan52: This is the first major release of LTSP that will hit fedora-announce-list, the main project list to over 200k subscribers.
20:59
<Ryan52>
warren, do I want to put the f-10 and f-9 ltsp updates into the bodhi system? or should I just let you do that after you test?
21:00
<warren>
Ryan52: don't put ltsp/ldm/ltspfs into bodhi yet, I'll show you my oddball non-standard testing method tomorrow.
21:00
<Ryan52>
ok.
21:01
warren, I intend on redoing the language localization tonight, so ya..
21:01
<warren>
Ryan52: cool, I look forward to testing it
21:01
<vagrantc_>
Gadi: where'd the documentation for using XDG_MENU_DIRS come from?
21:01
<warren>
Ryan52: if you're really sure it is in good shape just tag and build into F-9 through devel, we can always tag again later if necessary.
21:01
<Gadi>
XDG_DATA_DIRS
21:02
freedesktop.org
21:02
<warren>
Ryan52: those builds don't go anywhere so don't worry about building something not proven
21:02
<Gadi>
http://standards.freedesktop.org/basedir-spec/basedir-spec-0.5.html
21:02
<warren>
Ryan52: well, devel goes into daily F11 snapshots of rawhide now, but netboot doesn't work at anyway so wont hurt anybody right now.
21:02
<vagrantc_>
Gadi: yeah, reading that...
21:02* Ryan52 nods
21:03
<Ryan52>
warren, do people actually use rawhide? should I upgrade my system?
21:03
<vagrantc_>
Gadi: If $XDG_DATA_DIRS is either not set or empty, a value equal to /usr/local/share/:/usr/share/ should be used.
21:03
<Gadi>
vagrantc_: perhaps I should set to: $TMP_XDG_DIR:/usr/local/share/:/usr/share/
21:03
<vagrantc_>
Gadi: that's exactly what i was thinking...
21:03
<Gadi>
mine is: XDG_DATA_DIRS=/usr/local/share/:/usr/share/:/usr/share/gdm/
21:03
on intrepid
21:03
<warren>
Ryan52: not recommended at the moment, we're not even at Alpha yet
21:04
<Ryan52>
ok.
21:04
<vagrantc_>
Gadi: unless, of course, the distro or sysadmin does something weird... and then we've got issues.
21:04
<warren>
Ryan52: even I'M afraid to upgrade to rawhide
21:04
right now
21:04
<vagrantc_>
Gadi: no value set on lenny, logging in from ldm.
21:04
<Gadi>
vagrantc_: env|grep XDG
21:04
ah, right - I am logged in from gdm
21:05
ok
21:05
<warren>
Ryan52: only upgrade to rawhide if you're interested in figuring out exactly what went wrong, filing bugs and manually grabbing new rawhide builds and trying it to report back before the next rawhide nightly so it fixes for other people. Very time intensive process...
21:05
<Gadi>
Ill do that
21:05
<warren>
Ryan52: although after we fix mkinitrd/dhclient, it is possible to run F-11 /opt/ltsp/i386 on F-10 host and that is very safe.
21:07
<Ryan52>
ok.
21:07
<warren>
Ryan52: that being said, this rawhide has been strangely quiet with very little broken... we were so surprised that we pushed our alpha freeze a week later to grab a lot more newer packages, without changing our release date.
21:07
but immediately after the Alpha, we push gcc-4.4 and rebuild all packages
21:07
so things will explode again
21:07
<Gadi>
vagrantc_: pushed
21:10
<vagrantc_>
Gadi: looking good. :)
21:11
<Gadi>
now, if only we can find some unsuspecting sucker to test...
21:11
<vagrantc_>
Gadi: i was figuring with all the critiquing i'm doing, i'd best do some testing, too :)
21:11
<Gadi>
btw: I noticed no Makefile for localapss/ldm-rc.d
21:12
so, I assume it is up to the packagers to note that change?
21:12
(ie from S* to X*)
21:13
<vagrantc_>
Gadi: oh, i noticed another thing. shouldn't LDM_SOCKET and LDM_SERVER already be exported... in the calls to ltsp-genmenu, it seems like it shouldn't need to be explicitly added there...
21:14try2free has joined #ltsp
21:14
<vagrantc_>
Gadi: in X02-localapps-menu and X99-zkill...
21:14
<Gadi>
ltsp-genmenu is run in an su -
21:14
so, the vars dont migrate
21:14
<vagrantc_>
ah, got it.
21:15
<Gadi>
we were bit by that with the LOCAL_APPS_MENU_ITEMS
21:15
:)
21:16
<vagrantc_>
i gotta do a few things, but should be able to test in a bit...
21:16* Gadi reflects on how he loves how configurable ldm has become with all the rc_files
21:16
<Gadi>
Ryan52: remond me to thank you for adding X*
21:16
:)
21:16
*remind
21:18
<warren>
Ryan52: ok.... sleep... ttyl
21:18
<Ryan52>
:)
21:18
warren, bye
21:22
<TheProf>
Gadi: sorry to bother you again. I've almost got the system set up. I've got the windows computer on the same switch as the thin clients. I can see the shared folders on the windows computer from the ltsp server via smb. I just can't see the shared printer. Any thoughts?
21:25ball has joined #ltsp
21:26
<ball>
If I find a machine that can boot from USB, is there a version of LTSP that will let me turn that into an X terminal?
21:28
<Gadi>
TheProf: not sure - you sure the printer is shared?
21:29
brb
21:31
<TheProf>
Gadi: OK. It says it's shared. I'll unshare and reshare it
21:41japerry has joined #ltsp
21:46try2free has quit IRC
21:51
<TheProf>
Gadi: Shared, unshared it, reshared it rebooted 4 times (it's a windows machine :) ) but it's not showing up on the linux machien
22:00CaScAdE^1arAway has joined #ltsp
22:02CaScAdE^FarAway has quit IRC
22:10
<vagrantc_>
Gadi: gah. i don't have a chance to test this tonight. i'm sick and need to turn off my brain.
22:11RobertLaptop has joined #ltsp
22:26crosvera has joined #ltsp
22:27japerry has quit IRC
22:27crosvera has quit IRC
22:27ball has quit IRC
22:42
<Ryan52>
warren, ltsp uploaded to all three, ltspfs uploaded to devel, and I'll do ldm once I get my new upstream release finished :)
22:44
warren, and please double check the changes that were merged from F-10 to devel for ltsp.spec ... I think I understand what it's doing, but I don't understand why that changed (or more of why it was originally that way..) with the "%if 0%{?fedora}"s disabling client stuff in devel..
22:45
<johnny>
lol Ryan52 .. i should have made you a gentoo dev instead :)
22:45
Ryan52, altho they are nearly as resistant to certain changes as debian is
22:46
at least the base system was in good shape before they got to squabbling like debian
22:46
Ryan52, we've been doing dependency based init for years.. waaay before this upstart business
22:47
altho of course that means it will be harder for gentoo to decide to use upstart themselves .. as they are very happy with their init system now
22:47
<Ryan52>
heh
22:47
<johnny>
ala debian
22:47
<dberkholz>
it's hard to justify switching when you had a decent setup to start with
22:47
something where dependencies don't go into the comments.
22:48
god, that disgusts me
22:48
<johnny>
the reason i'm using fedora on this laptop is because they are more acceptable to change
22:48
dberkholz, for real..
22:48
dberkholz, i've still yet to see a true upstart config file tho
22:50shrek has joined #ltsp
22:51
<Ryan52>
with bzr is there any reason for me to bother with "bzr merge" for reverting commits? or can I just use "bzr diff | patch -R"? :)
22:52
<johnny>
Ryan52, no!
22:52
please don't do that
22:53
<Ryan52>
ugh.
22:53
<johnny>
we had a nasty situation like that.. and it nearly broke our repository system
22:53
altho it was monotone not bzr
22:53
<Ryan52>
why?
22:53
<johnny>
the important thing is what happens to the history
22:53
<Ryan52>
my question is does bzr actually pay attention?
22:53
<johnny>
yes
22:53
monotone paid attention to more ..
22:53
<Ryan52>
"bzr log" doesn't make it seem like it does.
22:54
<johnny>
Ryan52, in monotone commits are signed by default with your key
22:54
<Ryan52>
but *shrug*
22:54
johnny, gpg? 0.o
22:54
<johnny>
definitely much more strict
22:54
ssh keys actually
22:54
<Ryan52>
signed with an ssh key?
22:54
wierd.
22:54
*weird
22:54
uhhh...how?
22:54
<johnny>
actually.. they are monotone keys.. but it works via ssh-agent
22:55
<Ryan52>
insanity imho.
22:55
<johnny>
$ mtn ssh_agent_export ~/.ssh/id_monotone
22:55
enter passphrase for key ID [user@example.com]:
22:55
enter new passphrase for key ID [user@example.com]:
22:55
confirm passphrase for key ID [user@example.com]:
22:55
insanity?
22:56
you do create a new key for monotone
22:56
and then when you commit, you sign your revisions
22:56
it's not optional like in bzr
22:56
<Ryan52>
does it use the author or the pusher?
22:56
(for the signing)
22:57
<nubae>
johnny: did u end up setting up a habari blog in the end?
22:57
<johnny>
Ryan52, the pusher
22:58
<Ryan52>
okay, then it makes sense.
22:58
<johnny>
nubae, sorta..
22:58
nubae, i need to keep up with it.. but i kinda don't like the internal design
22:58
<nubae>
Its just such cool software
22:58
<johnny>
wordpress is worse.. but habari could be nicer on the inside..
22:58
they do things in a weird way
22:58
<nubae>
really? I think the abstraction of code is amazing
22:58
and very easy to tweak anything
22:59
<johnny>
we had a much better setup in xaraya
22:59
<nubae>
yes, its diferent
22:59
<johnny>
altho our urls could have been ten times better
22:59
<nubae>
but once u get it, it makes coding feel like eating cake
22:59
<johnny>
we had a nice tempating system.. much better than habari
22:59
xmllike.. so it doesn't look out of place
22:59
<nubae>
check out www.nubae.com com now
22:59
<johnny>
we even had the templates parsing as xslt
22:59
nubae, some folks are really talking about how cool jekyll is
22:59
<nubae>
check out the plugins I've put in.. oh and let me know if it all seems to much, which ones I should take out
23:00
<johnny>
altho totally different way of doing things
23:00
<nubae>
jeckyl... I'll check that out
23:00
<johnny>
it uses a git repository
23:00
altho i guess it could use bzr just as easily..
23:00
<nubae>
I'm just fascinated at the mass integration that these tools all use
23:00
<dberkholz>
xslt for websites is the worst
23:00
<johnny>
you post by commiting named files
23:00
dberkholz, not the way we did it
23:01
<dberkholz>
the one person who knows xslt leaves, nobody else can touch it
23:01
<johnny>
we don't write in xslt.. we write in a much simpler language
23:01
<nubae>
yeah, habari works with svn and trac
23:01
<johnny>
nubae, huh?
23:01
<nubae>
makes it easy to keep updated
23:01
<johnny>
no
23:01
i'm saying that your posts are stored in git
23:01
not in a db
23:01japerry has joined #ltsp
23:01
<nubae>
oooh...
23:01
<johnny>
and when you post, it runs a commit hook on the server to rebuild the html based on templates
23:02
so.. it's mostly static except what js you insert
23:02
<nubae>
thought u were talking about the synching of files so u dont have to keep updating manually
23:02
<johnny>
just a simple templating language
23:02
i kinda find that idea nice..
23:02
<nubae>
anyway... can u check my site and tell me what u think?
23:02
<johnny>
you write in textile or markup
23:03
<nubae>
does keeping twitter, delicious, identi.ca, linked in seem to much?
23:04
<johnny>
maybe..
23:05vagrantc_ has quit IRC
23:06
<nubae>
they all link with openid, and each does something a little different... I've become a sucker for these apps
23:06
specially twitter and identi.ca
23:07
<johnny>
too bad identi.ca doesn't do the right thing with xmpp
23:07
maybe they'll learn one day
23:08
<nubae>
yeah xmpp is too new still
23:08
do u know jaiku?
23:09
thats another alternative to identi.ca which has much better integration with other services including xmpp
23:09
<johnny>
no.. xmpp is not too new
23:09
10 years is not new :)
23:09
people have just recently come to visit us
23:09* johnny blames them
23:09
<johnny>
:)
23:09
<nubae>
well, I'm not talking about its specification
23:10
Im talking about putting it into practice
23:10
<johnny>
yes.. truly
23:10
<nubae>
collaboration is a totally new concept
23:10
<johnny>
i'd like to see a popular social networking site that is non -profit..
23:10
<nubae>
theres been some advances on abiword collab integration I heard recently
23:11
<johnny>
yes there has
23:11
i have it now.. just nobody to test with?
23:11
<nubae>
well identi.ca is FOSS
23:11
<johnny>
i know
23:11
<nubae>
oh, where did u download it, I'll install and we can play :-)
23:11
<johnny>
but somebody still has to host stuff
23:11
<nubae>
yeah true
23:11
<johnny>
uhmm.. from my repositories?
23:12
<nubae>
yah gentoo... thought u were on ubuntu for a second there
23:12
<johnny>
no.. fedora :)
23:12
fedora on this laptop, gentoo on server and fast desktop, ubuntu on my gf's pc and at work
23:13
<nubae>
ok, so which version is it?
23:14
<johnny>
2.6.6 is what i have
23:14
but there might be another package for the jabber support
23:15RobertLaptop has quit IRC
23:15
<Ryan52>
johnny, how much does "locale -a" output on a typical gentoo system?
23:16
<johnny>
typical.. i wonder
23:16
the docs mention setting locale.gen, so many users will have only 1
23:16
or two
23:16
<Ryan52>
ah.
23:16
<johnny>
i don't know
23:16
i don't have an unmodified system to check
23:17
oh.. my ltsp chroot :)
23:17
<Ryan52>
crap, so both debian/ubuntu and gentoo will need an alternate way to determine all of the possible locales.
23:17* Ryan52 likes Fedora a lot right now
23:17
<johnny>
huh? what's the problem with having less?
23:17
how does that affect you?
23:18
<Ryan52>
ldm will get the list of localized names generated at build time.
23:18
how does it know which ones to put in the list?
23:18
debian/ubuntu has /usr/share/i18n/SUPPORTED
23:18
Fedora has "locale -a"
23:19
<johnny>
don't you have locale -a?
23:19
<Ryan52>
yes, but it only outputs a few languages..
23:19
in the build chroots, probably only C and POSIX
23:19
<johnny>
and how is that a problem?
23:19
<Ryan52>
nothing will be translated.
23:19
languages in ldm will appear as en_US
23:19
<johnny>
so.. why don't you give the chroot the same languages as the serveer?
23:20
<Ryan52>
how?
23:20
<johnny>
copy it over? add the packages to install the support?
23:20
<Ryan52>
the packages which contain the info are too big for the chroot
23:20
and it's too much to push over ldminfod
23:21
we've discussed it, just trust me :)
23:21
<TheProf>
!SolvePrinterProblem!
23:21
<ltspbot>
TheProf: Error: "SolvePrinterProblem!" is not a valid command.
23:21
<Ryan52>
heh
23:21
<TheProf>
drat.
23:22
<Ryan52>
johnny, so....no way to get a list of possible locales?
23:22
<TheProf>
Thought it might have an idea :)
23:23
<Ryan52>
*shrug*. if not, then it will just be fr_CA (instead of "Francais (Cananda)") for Gentoo users..we can work something out later.
23:23
<johnny>
possible is locale -a ? i'm not sure what you mean
23:23
<Ryan52>
"locale -a | wc -l"
23:23
what's that say?
23:24
if you get something around 700, then good. if you get something around 3, then bad.
23:24
<johnny>
700.. but that's cuz we build glibc in the chroot.. so it happens by default
23:25
Ryan52, we do have a locale option to only generate the requested locales
23:25
i think it would be fair if the not selected locales just showed up as their code
23:25
i don't consider that a problem
23:25
<Ryan52>
ok.
23:25
<johnny>
so as long as it falls back to that.. i don't care
23:26* Ryan52 nods
23:26SDuensin has quit IRC
23:36rjune_ has quit IRC
23:37rjune_ has joined #ltsp
23:40* nubae is compiling abiword 2.6.6 from source now
23:43
<TheProf>
!BlowupPrinter
23:43
<ltspbot>
TheProf: Error: "BlowupPrinter" is not a valid command.
23:43
<nubae>
TheProf: try !kickprinter
23:43
<TheProf>
Picky bot.
23:43
<nubae>
:-)
23:43
<TheProf>
!BuyMeLinuxNetworkPrinter
23:43
<ltspbot>
TheProf: Error: "BuyMeLinuxNetworkPrinter" is not a valid command.
23:44
<TheProf>
cheap bot too! :)
23:44
<nubae>
your warranty was not valid cause it thinks u tried to put it near a windoze machine :p
23:45
<TheProf>
:) this printer has a grand total of 32 kb onboard ROM. it barely knows it's a printer without a windows machine
23:47* nubae plays the abiword dependency game
23:47
<nubae>
oh what fun building from source is
23:47
I cant believe gentooers do this for the sheer pleasure of it
23:48
<dberkholz>
you're doing LFS if you're playing dependency games
23:48
gentoo takes care of the hard part
23:49
<nubae>
well its just abiword... so yes, its not the gentoo way, but the journey is more or less the same distance
23:50
this is actually where Fedora is a bit nicer... its packages are named more rationally
23:58
<cyberorg>
all distros should be forced to use the same package names, makes everyone's life so much easier