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


Channel log from 14 April 2008   (all times are UTC)

00:09tjpp has quit IRC
00:25ryudo_home has quit IRC
00:30asac_ has joined #ltsp
00:42asac has quit IRC
00:42asac_ is now known as asac
01:15deavid has joined #ltsp
01:41plamengr has joined #ltsp
02:01cpunches has quit IRC
02:29mikkel has joined #ltsp
02:30mccann has quit IRC
02:44ogra has quit IRC
02:44ogra has joined #ltsp
03:15bobby_C has joined #ltsp
03:15plamengr has quit IRC
03:48* gvy is back (gone 62:27:04)
03:49
<gvy>
mornin'! :)
03:58Q-FUNK has joined #ltsp
03:59Q-FUNK has joined #ltsp
04:13cyberorg has joined #ltsp
04:24debayan has joined #ltsp
04:26
<debayan>
say there are more than one ltsp servers on the network.... and one of them is operating at its maximum capacity, can it then "transfer" the thin client to the second server? Is this what a DHCP helper does? I mean a load balancing sort of arrangement... does ldminfod help in this regard?
04:36soneyka has quit IRC
05:11soneyka has joined #ltsp
05:13slipttees has quit IRC
05:29viking-ice has joined #ltsp
06:19TelnetManta has quit IRC
06:33sepski has joined #ltsp
06:33
<Ryudo>
mornin gvy :D
06:34
<gvy>
debayan, looks like you want load balancing, google for ltsp plus that
06:34
Ryudo, morn'
06:34
<Ryudo>
:D hehe gvy i test the new beta version of mozilla firefox
06:34
<debayan>
gvy: cool
06:34
<Ryudo>
FF 3 Beta 5 and openoffice 2.4
06:34
<gvy>
debayan, afaih it's not ported to 5.x (ldm*) yet, just 4.2
06:34
Ryudo, all stable, no deadlocks? :)
06:34
<debayan>
ohh
06:34
<Ryudo>
the problem wich pixmaps persist
06:35
<gvy>
Ryudo, well, then i'm sorry to say i was right
06:35
<debayan>
gvy: who does the porting work?
06:35
<Ryudo>
all deadlockable ehehe no stable :{
06:35
<gvy>
oooo-o-ogra! that was a kernel fix needed :)
06:35
debayan, dunno... i'm only catching up myself last year
06:35
<ogra>
intresting that i get different reports from hardy users
06:35
<gvy>
Ryudo, well, i'll probably do a mini iso this week
06:35
<Ryudo>
yes we need a kernel fix :(
06:35
<debayan>
gvy: ok thanx
06:36
<ogra>
anyway, way to busy to deal with that ...
06:36
<gvy>
Ryudo, is it hardy for you?
06:36
<Ryudo>
i'm using feisty
06:36* ogra goes back to care for the release
06:36
<Ryudo>
feisty 7.04 :D
06:36
<gvy>
ogra, yup, +1 todo for you at best; luck with release :)
06:36
Ryudo, then re-read the log what he told ya
06:36
<Ryudo>
log ? where ?
06:37
<gvy>
you better set up your irc client to log the conversations... so you have answers already told to the questions you already asked at least :)
06:37
for unix xchat it's ~/.xchat*/xchatlogs/*
06:38
if enabled (not by default) :-/
06:38* gvy is away: to care for release, too
06:40
<Ryudo>
its not enable :(
06:41
the pixmaps bug at the moment no the fix correct ? sorry gvy my english is not good :{
06:41
<gvy>
Ryudo, np, i'm not native english squeaker too
06:41
Ryudo, as i told you we _do_ have the fix, and i gave you the iso link :)
06:42
if you can wait a few days, i'm gonna to build smaller iso this week
06:42
and not school specific
06:42
<Ryudo>
:D ohh cool i wait for this !
06:43
gvy
06:44
you know if ltsp 4.2 have problens wich pixmaps cache ?
06:47
<gvy>
Ryudo, it should use less ram and there _is_ xramperc there
06:47
so the problem is there but you have less chances to catch it
06:47
<Ryudo>
xramperc close app correct ? this not solve the problem ehehe :{
06:48jammcq has quit IRC
06:49
<Ryudo>
hummm
06:55
<gvy>
Ryudo, sorta
06:56
<Ryudo>
hi
06:56
<gvy>
"sorta == sort of"
07:00mhterres has joined #ltsp
07:08TelnetManta has joined #ltsp
07:15sepski has quit IRC
07:24cliebow has joined #ltsp
07:24
<Ryudo>
gvy
07:24
ooo developers reply my question
07:25
<gvy>
and? :)
07:26gloubiboulga has joined #ltsp
07:27jammcq has joined #ltsp
07:27
<Ryudo>
you see ?
07:27
"I fear I have no immediate solution for this."
07:27
:(
07:27
<gloubiboulga>
hello, is this the right place for ltsp help ?
07:28
I have used the debian/etch version for some times
07:28
and tried the backported version to get printer support and new features
07:28
but it doesn't mount my nfs share
07:28
it continuously retries
07:29
the nfs server is ok (I tried manually)
07:29
the client gets a network config (every two seconds)
07:29
but never mounts the nfs share
07:30
it eventually stops trying and drops me to a minimal shell
07:30
what can I do to debug that ?
07:31
anybody here ?
07:31
ping ?
07:32
<laga>
patience is a virtue
07:32
<gloubiboulga>
:-)
07:32
I thought I was alone
07:33Guaraldo has joined #ltsp
07:35
<Ryudo>
gvy
07:35
you see ?
07:35Pascal_Debian has joined #ltsp
07:40
<cliebow>
gloubiboulga, what does showmount -e show on server?
07:44
<gloubiboulga>
cliebow: Export list for phallompe:
07:44
/opt/ltsp *
07:44
07:45* cliebow cliebow has senior moment..
07:47
<gloubiboulga>
cliebow: never mind, I found it, /etc/hosts was notconfigured on the server
07:47
<cliebow>
gloubiboulga, could the network card be unsupported in initramfs
07:47
ohh cool!
07:48
<gloubiboulga>
cool cool, two days to find that stupid mistake...
07:48
:-)
07:48
<cliebow>
nfs can be really whiny...
07:48
<gloubiboulga>
I don't know why ltsp couldn't mount it and another host could
07:48
<ogra>
how can you install a system without proper /etc/hosts ?
07:48
<Q-FUNK>
cliebow: what? you found found your golden age club membership card? :-P
07:48
<cliebow>
cant remember where i left it
07:48
<gloubiboulga>
ogra: I don't know what happend, but it was empty
07:48elisboa has joined #ltsp
07:48
<ogra>
weird, d-i should have filled it
07:48
<Q-FUNK>
heh
07:49
<gloubiboulga>
ogra: it's a xen virtual machine,but everything else works on the network part, so I didn't think about it
07:49
and even other "normal" machines could mount nfs shares
07:50
so I didn't check anything on the server
07:50
<ogra>
well, then your xen build script should have cared :)
07:50
<gloubiboulga>
sure
07:50
xen-tools still need some massaging :-)
07:50
<ogra>
(unless you just bootstrapped whithough configuring anything, but then you will run into more probs anyway)
07:51
<gloubiboulga>
no no no
07:51
it's a full blown instance
07:51
and everything *should* be configured properly
07:51
<ogra>
yeah, then it sounds like a bugreport against xen-tools :)
07:52slidesinger has joined #ltsp
07:52
<gloubiboulga>
right now ltsp in my emergency since my regular application server died
07:52* ogra is happy ubuntu has jeos so he doesnt have to care :)
07:52
<gloubiboulga>
but xen is on my list :-)
07:53
now back to fun, I have to find why the X server doesn't start well :-)
07:54colmar has joined #ltsp
07:54
<colmar>
hi
07:54
there are news about LTSP in the incoming ubuntu release?
07:55
<Ryudo>
ogra
07:55
<laga>
it's gonna rock ;)
07:55
<Ryudo>
morning :D
07:55
i send one email to OOo developers
07:56
and i found ansers to my questions
07:57
<colmar>
laga: :) I mean, news about ldm (performances) or about local media (floppy and CD recorders..)
07:57gloubiboulga has quit IRC
07:57
<Ryudo>
colmar how your terminals config ?
07:58
processor mem etc ?
07:58
<colmar>
P 200 64 MB
07:59
It seems that hardy uses 5.0.40 as version of ltsp and gutsy 5.0.39.. there aren't many changes then:)
07:59
<laga>
colmar: who says that? :)
07:59
<Ryudo>
colmar your users use Ooo ?
07:59
<laga>
colmar: read the changelog in /usr/share/docs/ltsp-server/
08:00
<Ryudo>
Impress Ooo
08:00
?
08:01
<colmar>
laga: yes but not on ltsp, just on thinstation:) but I'll switch soon
08:01alekibango has quit IRC
08:01alekibango has joined #ltsp
08:01
<Ryudo>
locall app O_O ?
08:02
i use ooo here ... and this is terminal freeze monster killer :D
08:03
<colmar>
Ryudo: no remote, and it freezees too
08:03
:)
08:03
I can use writer but not impress
08:03
<Ryudo>
OMFG
08:03
im not alone in this via crusis
08:03
<colmar>
or I can use impress but I can't do presentation
08:03
just writing
08:04
<Ryudo>
colmar i send email to developers Ooo
08:04topslakr has quit IRC
08:04topslakr has joined #ltsp
08:05
<Ryudo>
look in pvt
08:05* colmar looks into his pants
08:06* laga sprays colmar with water
08:06
<colmar>
:P
08:06
Ryudo: I didn't see anything in my pvt
08:06
(OMG)
08:06
<Ryudo>
08:06
I think I need put 256 or 512 on each terminal, turning a fatclient
08:07Topslack has joined #ltsp
08:07
<Ryudo>
08:07
I am sad and disappointed
08:08
<mhterres>
Ryudo: that is not just the memory that makes a computer a fat client :-)
08:08
memory is just one piece
08:09
<Ryudo>
Não ingles nãooooo
08:09
:{
08:10
<colmar>
Ryudo: have you enough server memory?
08:10
<Ryudo>
The Truth is ... A terminal with 64 mb is notcapable to running Openoffice (impress) and firefox
08:10
<mhterres>
hmmm
08:11mikkel has quit IRC
08:11
<Ryudo>
i have AMD Atlhon 3700+ 4 GB DDR
08:11
2 x 80 GB SATA II
08:11
the problem is not my server .... is the alocation of pixmaps in my temrinals (low memory)
08:11
<colmar>
Ryudo: and have you tried with a laptop with 1Gb of RAM as thin client? I done and impress freezes again
08:12
<Ryudo>
AHAUAUHAHUHUAHUAUHAAUHUHAAUHHUAUHA
08:12
OMFG really ?
08:12
<colmar>
yes
08:12
try it
08:12
<Ryudo>
oh man this is a big problem
08:12
<colmar>
I think is a network/X.org problem
08:13
have u a 100Mb lan?
08:14
<Ryudo>
yes
08:15
08:15
I have swapped my switchs and the problem does not solve
08:17
This is a critical problem, which has not yet found solution ... I think I will return the use of windows 98 on my terminal:)
08:18
<colmar>
luck you
08:18
lucky
08:19
btw isn't a LTSP problem
08:19
'cause I've the same on thinstation
08:19
I suppose it's an X.org problem
08:20gloubiboulga has joined #ltsp
08:20
<Ryudo>
it is a Xorg problem
08:20
<colmar>
uh?
08:21
how you can say that
08:21
<Ryudo>
http://www.mille-xterm.org/en/Terminal_Memory_Usage
08:21
pay attention eheheh
08:21Gadi has joined #ltsp
08:21
<Ryudo>
gadi morrrrning :D
08:22
this arcticle explain the problem colmar
08:22
http://www.mille-xterm.org/en/Terminal_Memory_Usage
08:22
The solution is simple, so programs like ooo and FF stop doing cache pixmaps in memory of the terminal
08:23
<ogra>
colmar, thats a two bladed sword :) its xorg allowing to use more ram than available as well as oo.o claiming pixmap cache where none is available anymore
08:23ikonia has quit IRC
08:24
<ogra>
most of the oo.o and firefox parts of that bug got fixed (not all though as it seems)
08:24
<colmar>
so?
08:24
<Ryudo>
ogra it not true
08:24
<colmar>
I don't want to know the problem, I want the solution!:P
08:24
:D
08:25
<Ryudo>
i test the openoffice 2.4 and Firefox 3 beta 5
08:25ikonia_ has joined #ltsp
08:25
<Ryudo>
NOT HAS FIXED
08:25
i send email to developer of ooo
08:25
and he said
08:25
<ogra>
Ryudo, read what i wrote before shouting please
08:25
<Ryudo>
Hi,
08:25
<Ryudo> the fix concerned a leak (some pixmaps were not freed that are now). However in this case it seems that simply the Xserver runs out of resources. If OOo is to do anything about that we would need to teach applications to release their bitmaps earlier. Which will cause a performance problem when scrolling a documents containing many images. I fear I have no immediate solution for this.
08:25
<Ryudo> Kind regards, pl
08:25
<colmar>
ogra: do you know if there is a work around?
08:25hersonls has joined #ltsp
08:26
<ogra>
colmar, as i said parts were fixed (as you can see in teh oo.o response Ryudo just posted)
08:26ikonia_ is now known as ikonia
08:26
<ogra>
where they are not fixed i fear there is no work around yet until oo.o comes up with something
08:26
<colmar>
W abiword:)
08:26
<Ryudo>
the impress still freezing :D
08:27
the solution is
08:27
not use FF
08:27
<ogra>
colmar, doesnt provide presentations
08:27
<Ryudo>
use opera !
08:27
opera dont have problens wich pixmaps !
08:27
<ogra>
the pixmap fix seems to affect writer only
08:27
<colmar>
maybe the applied the patch just to writer
08:27Q-FUNK has quit IRC
08:27
<ogra>
(where the pixmap bug was actually found initially)
08:28
colmar, migh be, ask oo.o
08:28
<colmar>
the dev team of oo.o isn't too responsive..
08:28
<ogra>
it not an ltsp problem
08:28
*it's
08:28
<Ryudo>
i know
08:28
<ogra>
nothing we can do about apart from offering the XRAMPERC variable to avoid the hard freeze
08:28
<Ryudo>
this is a problem of software
08:29
<ogra>
and thats what we do since gutsy
08:29
<colmar>
ogra: XRAMPREC?
08:29
<ogra>
colmar, it limits the amount of ram X has available
08:29
<Ryudo>
Ogra you think if I assign more 256MB on each temrinal the problem solve?
08:29
<ogra>
so the kernels oom scheduler doesnt kick in
08:30
that kills the app requiring so much ram but leaves the rest of the desktop and client alone
08:30
<colmar>
I filled a bug a year ago about openoffice base and ODBC and the problem is still there.. and it had many confirmations by other ppl..
08:30
<ogra>
Ryudo, no
08:30
<colmar>
ogra: this is good
08:30
<Ryudo>
I did some tests with the xrestop and presentation(impress) are consuming 135 mb O_o omfg
08:30
<ogra>
Ryudo, it will make it occur less often but wont fix anything
08:30vagrantc has joined #ltsp
08:31
<Ryudo>
:(
08:31
<colmar>
at least it doesn't freezes the client
08:31
I prefer to see my openoffice disappear instead that having my client sleeping like a child
08:31deavidsedice has joined #ltsp
08:32
<ogra>
yeah
08:32
<Ryudo>
08:32
Not freezing the client, but it can be said that you not use impress anymore
08:32
:D
08:32
<ogra>
depends how graphics intense your persentations are :)
08:33
<Ryudo>
117 k presetation for exemple
08:33
65 mb
08:33
<ogra>
compressed ...
08:33
<Ryudo>
on xrestop
08:33
<colmar>
in fact
08:33
<Ryudo>
:{
08:33
<colmar>
btw isn't a good idea to use impress on network...
08:33
<Ryudo>
08:33
Detail, my terminal has only 64 mb
08:33
:)
08:33
<colmar>
it will cause a big network load
08:34
<ogra>
worst case you could try running MSOffice in wine ... (but make sure to have proper multiseat licenses)
08:34
<Ryudo>
The good wishes to provide a suite of applications of high quality as is the openoffice
08:35
omg msoffice dont make pixmaps !
08:35
good idea
08:35
:D
08:35
other good idea is ...
08:35
use windows 98 again
08:35
:(
08:35
<ogra>
not really
08:36
you will have to care for a whole bunch of machines instead of only one
08:36
<Ryudo>
Before everything works ogra .... Today only hear complaints of users ... Frezee frezee frezee
08:37
<ogra>
i know, you hand them directly through to the channel :P
08:37
<Ryudo>
thats is the true
08:37
i like ltsp
08:37
but developers of FF and OOO
08:37
need help us
08:37
:{
08:38
<ogra>
well, ff3 is known to work well in hardy, i didnt have any reports of crashers yet
08:38
(but there might also be xorg fixes, not sure which side helped here)
08:38
<Ryudo>
ogra you can test here ?
08:38
use xrestop and open a cmplex web site
08:38
wich many figures
08:39
look
08:39
http://www.carteretcountyschools.org/bms/teacherwebs/sdavenport/artgallery6.htm
08:40
terminal killer test :D
08:40
open this web site in your Firefox 3 Beta 5
08:40Q-FUNK has joined #ltsp
08:40
<Ryudo>
and cry :(
08:40
<ogra>
no, i wont
08:40
i know the page
08:41
and dont have the time to wait until it filled my disk cache
08:41
<Ryudo>
if this page crash my browser ... the problem has not solved correct ?
08:41
HUAuhauh :{
08:41
or the webpage has a bug ?
08:45
<ogra>
the webpage has 200M of jpegs
08:45
indeed that buggy
08:45
*thats
08:45
2600000 359 224 1 128 316 59018K 22K 59040K 27599 Mrs. Davenport's Web - Mozilla Firefox 3 Beta 5
08:45
60M
08:45
(from xrestop)
08:46
<Ryudo>
ok
08:46
<ogra>
thats actually not bad
08:46
<Ryudo>
on terminal iwch 64 mb of ram
08:46
<ogra>
it used to eat about 300M in gutsy
08:46
<Ryudo>
is the killer website :D
08:46
<ogra>
so FF improved a lot here
08:46
<Ryudo>
O_o
08:46
but ogra
08:46
you have opera here ?
08:46
<ogra>
nope
08:46
<Ryudo>
ope this web site on opera
08:46
<ogra>
and i wont clutter my system with it
08:47
<Ryudo>
xrestop reporta 4775
08:47
5 MB ;)
08:47
<ogra>
i know it handles images completely different to other browsers, no need to prove that
08:47
<deavidsedice>
I've tested that URL in Konqueror... ¡ouch!
08:47
<Ryudo>
uahuhaa :{
08:47
<ogra>
deavidsedice, yeah, its the worst example of bad html habits
08:47
<Ryudo>
ahahaha
08:48
;-)
08:48
<deavidsedice>
1,5Gb of RAM eaten by KHTML :-S
08:48
<Ryudo>
O_O
08:48
<deavidsedice>
now trying with FF 3 beta 3
08:48
<Ryudo>
post results here :D
08:49
<deavidsedice>
FF -> 512Mb or more...
08:49
<ogra>
yeah, isnt it beautiful how you can drag tiffs from your scanner directly into your frontpage webpage in windows ? :P
08:49
who cares about filesize :)
08:50
<deavidsedice>
Ok, I can see the webpage with Firefox 3 beta 3.
08:50
All the images are copied to some disk cache
08:50
and they're placed onto the RAM only when needed
08:50
is a bit slower..
08:50
but it's ok.
08:50
<ogra>
yeah
08:50
as i said, they improved it a lot
08:51
<deavidsedice>
now, opera.
08:51
wow. Opera is going a lot faster than FF
08:51
about 256Mb of memory only
08:52
can be less (I'm seeing graphs on a KDE applet)
08:52
Opera does not load images when they are out of my view
08:53
when I scrolled, Opera tried to get them from internet
08:53
ouch. In opera, ever I scroll, Opera reloads the images.
08:54K_O-Gnom has joined #ltsp
08:54
<deavidsedice>
It's like freeing the images that are out of the window, and fetch them after, when the user scrolls.
08:54
the page it's never completely loaded.
08:54
:-(
08:55
PID USER PR NI VIRT RES SHR S PU %MEM TIME+ COMMAND
08:55
1144 deavid 20 0 319m 281m 14m S 0.3 18.5 0:46.01 opera
08:56
ogra: and what about galeon?
08:57
Galeon gives beter results on that URL :-)
08:58
<ogra>
galeon is dead, no ?
08:59
<deavidsedice>
ogra: I don't know, but galeon does not need to store images in cache, and only uses about 500Mb. And it's not like opera: Galeon does not fetch the image every time I scroll.
08:59
<ogra>
right, but it uses libnspr which is dead since xulruner exists
08:59alekibango has quit IRC
09:00
<ogra>
if it gets ported it willl behave exactly the same as firefox and epiphany (whic use xulrunner)
09:00
<vagrantc>
so the fix would be to xulrunner?
09:04
<ogra>
well, there was massive fixing already for ff3
09:05
the beta4 looked very good here
09:05
the complete pixmap handling was reworked
09:06
vagrantc, btw, did you notice that nbd does autoreconnect now ? so failover setups get possible now
09:07
<vagrantc>
ogra: i haven't noticed...
09:11viking-ice has quit IRC
09:11* warren confused as to why ldm isn't displaying any translations
09:11
<vagrantc>
warren: have you tried putting LANG in lts.conf ?
09:11
<warren>
vagrantc: LDM_LANGUAGE or LANG?
09:11
<vagrantc>
warren: i replied to your email, but haven't sent it yet.
09:12
warren: LANG
09:12* warren tries
09:12
<vagrantc>
i think LDM_LANGUAGE is for the login session, but i don't really know
09:13
<warren>
vagrantc: we should have a system fallback
09:13
vagrantc: like in gadi's cases where he wants the login screen to work even if it failed to connect to ldminfod
09:13
I also see it randomly fail to connect here
09:13
<ogra>
yeah, its the selection from the list
09:13
(LDM_LANGUAGE)
09:14alekibango has joined #ltsp
09:14
<warren>
defaultlanguage in ldminfod somehow gets "en_US" instead of "en_US.UTF-8"
09:14
<ogra>
weird
09:14
<warren>
not sure if that's causing any issues
09:15
<ogra>
it likely will, you need the correct locale name ... but ten it only affects the session
09:15
not the ldm ui
09:15
<vagrantc>
it would be nice to make the language selection and greeter language in sync
09:15
<ogra>
ldm ui will only show whats set in the chroot as default locale
09:15
indeed
09:15
vagrantc, but that would mean you need to generate *all* possible locales during client build
09:16
<warren>
which is what we do
09:16
<ogra>
locale-gen alone takes about a minute for en* here
09:16
dont ask for asian
09:16
<warren>
what exactly does locale-gen do?
09:16
spit out .mo files?
09:17
<ogra>
Compiled locale files take about 50MB of disk space, and most users only need few locales. In order to save disk space, compiled locale files are not distributed in
09:17
the locales package, but selected locales are automatically generated when this package is installed by running the locale-gen program.
09:17
<Q-FUNK>
I always wondered why C (en_US.UTF-8) is not included by default. I only need C to send bug reports and I really don't need the dozen of englush slangs in my gdm menu.
09:17
<ogra>
from the manpage
09:17
<warren>
we don't even have a tool called locale-gen
09:17
<ogra>
Q-FUNK, C always works in ubuntu at least
09:18
<Q-FUNK>
ogra: only because the full collection of english localisation and proof reading tools are installed by default.
09:18
<vagrantc>
ogra: well, we need to generate all possible supported locales during chroot build, yes. but in the default case, i don't think it's an unreasonable default to assume whatever is installed on the server should be in the chroot.
09:18
<ogra>
but ubuntu is pretty unique in handling translations anyway
09:18
<warren>
what you decide to install in your chroot is your own choice
09:19
our distro chooses to install all possible languages in both server and client
09:19
<vagrantc>
as long as it handles properly when a locale isn't available.
09:19
<ogra>
vagrantc, i didnt talk about the default :)
09:19
<warren>
we make it possible to strip it down though
09:19
<ogra>
but if we want to switch on the fly all possible locales need to be available
09:19
<Q-FUNK>
I don't need e.g. south african and australian english, but those come along for the ride the minute the default english packages needed to get C output are installed.
09:19
<vagrantc>
ogra: well, in the exceptional or unusual cases, it's always going to require some effort :P
09:20
<ogra>
warren, you dont do constant translation upgrades during a support lifecycle of a release, right ?
09:20
<vagrantc>
ogra: switching on the fly only requires that all the locales you *want* to be available.
09:20
<warren>
vagrantc: ok, LANG worked, but it appears to be displaying it in the wrong encoding (looks like garbage)
09:20
<ogra>
vagrantc, all the ones that could come down the drain from ldminfod
09:21
<vagrantc>
ogra: and i think we can't be perfect, but in the typical cases, i think we can do reasonably well.
09:21
<ogra>
warren, go a UTF-8 capable font ?
09:21
<warren>
ogra: actually, we do update translations on an as-needed basis
09:21
checking
09:21
is ldm is using pango to render its text?
09:22
<ogra>
warren, well, its one of ubuntu basic policies that we update translations constantly on a two or three month schedule ... thats why translations are so heavly decoupled from everything for us
09:23
well, it understands pango ML ... it doesnt explicitly pull any extra pango stuff in beyond what gtk uses
09:23
(the labels are all written with us_markup=true, at least they were when i wrote the C frontend, not sure what scottie dropped from it)
09:24
<vagrantc>
on debian, i've had to specify LANG=es_ES.UTF-8 in lts.conf to get spanish as spoken in spain.
09:24
<ogra>
s/us_/use_/
09:24
right, you need the encoding
09:25
<vagrantc>
haven't tried with ja_JP.UTF-8, but will try soon.
09:27
<ogra>
for me ldm with dynamic translation switching means my chroot grows by about 600M
09:27
<warren>
ah, chroot is missing fonts
09:27
ogra: translations alone or everything including fonts?
09:28Q-FUNK has quit IRC
09:28
<ogra>
all language-pack-$lang ackages (containing translations, onts and input methods)
09:28
*fonts
09:28
<warren>
221441 bytes installed with 9 .mo files
09:28
ogra: you don't need input methods for only the login screen
09:28
<ogra>
i'm not talking about .mo files :)
09:29
well, i cant just split the ackages :)
09:29
<warren>
I guess we do have these packages well split
09:30
<Pascal_Debian>
hello
09:31
<ogra>
warren, we did until we got so many asian users that we had to find a better system :)
09:33Blinny has joined #ltsp
09:36Pascal_Debian has quit IRC
09:41mistik1_ has joined #ltsp
09:41mistik1 has quit IRC
09:41Pascal_Debian has joined #ltsp
09:41mistik1_ is now known as mistik1
09:43savetheWorld has joined #ltsp
09:48Bengoa has joined #ltsp
09:49
<vagrantc>
gvy: one thing i forgot to mention ... seems like there were a few files that were renamed to .in (presumably modified at package build), and it would be better to use "bzr mv" instead of "bzr rm FOO" "bzr add FOO.in" ...
09:49Pascal_Debian has quit IRC
09:49
<vagrantc>
keeps the diffs smaller, and makes it easier to see the historical relationships.
09:52Pascal_1 has joined #ltsp
10:04* gvy is back (gone 03:26:42)
10:05
<gvy>
vagrantc, ouch! forwarding, that's important habit
10:05
thanks
10:05
<vagrantc>
gvy: of course :)
10:05gloubiboulga has quit IRC
10:05
<vagrantc>
it's not a *huge* deal, but it's worth mentioning.
10:05
<gvy>
yeah
10:06* gvy knows the difference
10:06* vagrantc loves peer review
10:07colmar has quit IRC
10:07
<gvy>
vagrantc, ah, he says he tried to cheat around bzr diff :-) it is in fact broken enough, seems impossible to generate diff -Naur
10:07
drops some options if one tries to specify external diff program, and others if "as usual"
10:08
don't remember the details but i've looked from behind the shoulder when he got angry with that and the behaviour was weird indeed
10:08
should have copypasted for a bugreport...
10:08
(the culprit was passing -N iirc)
10:13Q-FUNK has joined #ltsp
10:13Pascal_1 has quit IRC
10:14
<ogra>
gvy, next time 'bzr :) they might be a good help ressource :)
10:14
err #bzr
10:15
(before he gets angry :) )
10:17
<gvy>
ogra, but then again i knew that and just be less lazy eh? :)
10:17
thanks anyways :)
10:17
the diff shrank to 11k (from 60k last time from 80+k a while ago)
10:18slidesinger has quit IRC
10:19viking-ice has joined #ltsp
10:24staffencasa has joined #ltsp
11:23spectra has joined #ltsp
11:25ogra has quit IRC
11:30mistik1_ has joined #ltsp
11:31mistik1 has quit IRC
11:31mistik1_ is now known as mistik1
11:37
<Q-FUNK>
ogra_cmpc: is there anything that can be done to speed up the sync of -geode from debian?
11:40
<warren>
if [ -r /etc/sysconfig/i18n ]; then
11:40
. /etc/sysconfig/i18n
11:40
export LANG
11:40
fi
11:40
ogra_cmpc: vagrantc: any objection to something like this in the screen common?
11:42
<vagrantc>
warren: i think it belongs elsewhere.
11:42
warren: overall, this is the case where i think we need distro-specific functionality
11:43
as in debian it's /etc/default/locale ... not sure what exactly ubuntu does
11:43
and i'm sure there's others
11:45praveer_cool has joined #ltsp
11:46
<johnny>
gentoo it's /etc/env.d/02locale
11:46
and it is always available once logged in via /etc/profile.env env-update handles all the exporting
11:47
<laga>
"env.d" looks like someone might put it into "/etc/env.d/0200-foolocale" as well
11:47
<johnny>
sure..
11:47
it'll still be parsed anyways
11:47
<vagrantc>
what about things run from init?
11:49
<johnny>
i'm not exactly sure how early it is, but it seems early enough atm
11:49
prolly once we are able to use bash
11:49
so.. pretty early
11:51
this is one the few places you use numbers for ordering
11:51
the init system doesn't have anything like that
11:52
i did hear of an interesting meeting that is going to happen between roy marples (the guy who cleaned up our init system) , the upstart guy, and initng guy
11:52
at least such is rumored to be happening
11:53
we've been using dependency based init scripts for some years now
11:53Ryudo has quit IRC
11:53mccann has joined #ltsp
11:53
<johnny>
4 or 5 years i'm guessing
11:53
hmm.. whoa.. touchin 6 now
11:54
<warren>
vagrantc: ok, what method of distro-specific setting should we use?
11:54
vagrantc: I mean, where it should be done
11:55
<johnny>
are there going to be pre and post screen-x-common ? or just screen-x-$distro ?
11:56* warren dislikes screen-x-$distro because I don't want to rely on distro name anywhere
11:57viriato has joined #ltsp
11:58
<vagrantc>
warren: well, it depends on how your distro is calling the screen scripts ... for debian, it would make sense to put it in our init script.
11:59savetheWorld has left #ltsp
12:02EvilBob has left #ltsp
12:04
<warren>
vagrantc: oh, I see.
12:05
<vagrantc>
warren: and this is the classic "try to avoid making unecessary filesystem calls" issue
12:05
since it happens client-side.
12:07
<warren>
ok, so each distro needs to do it their own way
12:07
gotcha
12:07tarzeau has quit IRC
12:07
<dberkholz>
in gentoo, locale ends up being set by /etc/profile so any login shell gets it
12:07
not sure exactly how the init process works in that respect
12:08tarzeau has joined #ltsp
12:08daduke_ has joined #ltsp
12:08daduke has quit IRC
12:12
<vagrantc>
warren: so our options are to try and figure out if there's a distro-independent place to hook that code into, or if there's something else already distro-specific that makes sense to plug into.
12:12
<warren>
there seems to be no distro-neutral standard
12:14
<vagrantc>
right, but we can have an infrastructure, similar to /usr/share/ltsp/ltsp-vendor-functions which i hope to someday populate with distro-specific implementations with common function names, like the architecture code...
12:15bricode has joined #ltsp
12:15cpunches has joined #ltsp
12:17Q-FUNK has quit IRC
12:20
<vagrantc>
maybe a function like "get_ltsp_locale" would work here.
12:24Q-FUNK has joined #ltsp
12:29ogra has joined #ltsp
12:32
<lns>
I guess colmar and Ryudo left, but if anyone's interested I have kind of a 'summary' page for the FF & OOo pixmap issues, with bugtracker URLs and all (like we all haven't heard enough about it) - http://lns.wikidot.com/firefoxpixmapcaching - http://lns.wikidot.com/openofficepixmapcaching
12:36
yack...right as i get an e-mail from a teacher saying 'we're still using windows for a lot of our computer labs because OOo and Firefox lock up a lot"
12:36
was this kind of thing an issue "back in the day" when X11 was just X11, and was designed for network usage?
12:37
or did the advent of fat clients spoil us all into these kinds of coding practices
12:51
<hersonls>
jammcq, tomorrow ;)
12:59
<vagrantc>
lns: i think you're onto something there... X11 was designed to be network transparent, but many apps neglect to prioritize along the same lines...
13:00
firefox and openoffice being almost as projects for windows platforms as GNU/Linux
13:00
almost as much
13:00Skarmeth has joined #ltsp
13:06
<lns>
vagrantc, it just seems obvious doesn't it, that the origins of X have kind of been disregarded in favor of fanciness
13:07
thus breaking a lot of functionality people would assume from it
13:07
anyway, i'm not a coder so i have no clue.
13:09
<vagrantc>
i don't really know the details, but i suspect you're right.
13:14BGomes has joined #ltsp
13:15
<johnny>
lns, there's been alot of toolkit evolution since then
13:16
odd things have grown in certain places
13:16
so.. refactor refactor refactor
13:16
as more goes into the toolkits, the easier it will be to possibly solve the problem for many apps at one time
13:18
<cpunches>
hi cliebow
13:18
cliebow: sorry i couldn't make it to that lug; i didn't wanna risk it with the hail-- I found out that the hail actually stopped after i went back to bed lol
13:19
s/sleet/hail
13:19
wtf they call it up here lol
13:22Pascal_Debian has joined #ltsp
13:25bricode has quit IRC
13:28Pascal_Debian has quit IRC
13:29ogra has quit IRC
13:30indradg has joined #ltsp
13:47Bengoa has quit IRC
13:49
<lns>
johnny, w/toolkits do you mean things like specifying "ENV=ltsp" or "ENV=local" to determine how to handle things like pixmap caching, etc?
13:50
or just how gfx are handled in general
13:51
<johnny>
how gfx are handled in general
13:51
so the fixes are required in less places, and hopefully in smaller parts
13:53
<lns>
that's good
13:53
=)
13:54
i would assume these things happen in non-ltsp thinclient environments as well ?
13:55praveer_cool has quit IRC
13:55
<lns>
(not trying to start another discussion on what's pretty obvious at this point)
14:00
<vagrantc>
lns: they happen, but they're often mitigated by sufficient swap space. low-end hardware, thin-clients, embedded devices, or any sort of restricted environment all expperience issues with it.
14:00
<johnny>
lots of folks are working on making it perform better in embedded environments at least
14:00
<vagrantc>
i.e. a pixmap cached to swap that you're never going to use again has relatively little impact on system performance.
14:01* vagrantc wonders if vagrantc uses "i.e." correctly
14:01ogra has joined #ltsp
14:03
<lns>
johnny, vagrantc, what I still don't totally get is why the pixmap mem DOESN'T get swapped out
14:03
<pscheie>
i.e. =~ "in other words", not to be confused with e.g., which means "for example"
14:04
<ivazquez>
"Id est", "That is"
14:04
<lns>
I.E. = "In Essense" I thought
14:06
<pscheie>
'In Essence' is a gneumonic (sp) for recalling what i.e. means, but technically it does stand for the latin Id Est
14:08
<johnny>
mnemonic
14:09
<lns>
wow...ya learn something new every day i guess =p
14:14
blah..i don't like it when compiz crashes randomly and my window borders disappear =p
14:18Pascal_1 has joined #ltsp
14:21Q-FUNK has quit IRC
14:32viking-ice has quit IRC
14:36elisboa has quit IRC
14:42viriato has quit IRC
14:50cliebow has quit IRC
14:51deavid has quit IRC
14:51joebaker has quit IRC
14:59debayan has quit IRC
15:00Gadi has left #ltsp
15:02Blinny has quit IRC
15:07ogra_cmpc has quit IRC
15:12Pascal_1 has quit IRC
15:21TelnetManta has quit IRC
15:30viking-ice has joined #ltsp
16:00joebaker has joined #ltsp
16:02dtrask has joined #ltsp
16:04Skarmeth has quit IRC
16:05K_O-Gnom has quit IRC
16:11mhterres has quit IRC
16:13BGomes has left #ltsp
16:14hersonls has quit IRC
16:37J45p3r has joined #ltsp
16:38dtrask has quit IRC
16:42slipttees has joined #ltsp
16:45deavidsedice has quit IRC
17:09Guaraldo has left #ltsp
17:17mnemoc has quit IRC
17:37jammcq has quit IRC
17:48BGomes has joined #ltsp
17:53ccherret1 has joined #ltsp
17:55ccherrett has quit IRC
17:56monteslu has quit IRC
17:58hersonls has joined #ltsp
18:02bobby_C has quit IRC
18:11topslakr has left #ltsp
18:18hersonls__ has joined #ltsp
18:28TelnetManta has joined #ltsp
18:32ccherrett has joined #ltsp
18:33ccherret1 has quit IRC
18:37hersonls has quit IRC
18:45hersonls has joined #ltsp
18:47hersonls__ has quit IRC
18:49ryudo has joined #ltsp
18:49ryudo_home has joined #ltsp
18:49ryudo_home has quit IRC
19:00staffencasa has quit IRC
19:08hersonls has left #ltsp
19:45
<lns>
wow I was really the last one to say something here huh? =p I come here for some after-work pleasure reading on #ltsp and i get nuthin! =p
19:46petre has joined #ltsp
19:50jammcq has joined #ltsp
19:52gentgeen__ has quit IRC
19:53
<petre>
hey jammcq
19:57
<jammcq>
hey petre
19:58
<petre>
did you go to Brazil yet?
19:58* petre can't remember the dates
19:58
<jammcq>
leaving tomorrow
19:59
<petre>
cool!
19:59
ah, to be self-employed & have control over one's time like that
20:01
<jammcq>
oh yeah, it's just nothing but fun
20:02J45p3r has quit IRC
20:03
<BGomes>
petre: where are u from?
20:03
<petre>
Minneapolis (US)
20:03
you?
20:04
<BGomes>
petre: Brasil (MA)
20:04
<petre>
brb
20:04
<jammcq>
BGomes: what city?
20:05
<BGomes>
jammcq: Imperatriz (MA)
20:05
<jammcq>
hmm, where is that?
20:05
<BGomes>
jammcq: Imperatriz is my city see at maps.google Imperatriz,MA,Brasil
20:06
jammcq: it is long travel to Porto Alegre
20:06
<jammcq>
ah, I'll be flying right over that tomorrow night
20:08
<BGomes>
jammcq: I fly at 3 am (UTC) tomorrow
20:08
<jammcq>
are you UTC-4 ?
20:08
or -3 ?
20:09
<BGomes>
yes
20:09
-3
20:10
sorry.. I am fly 9 am (UTC)
20:13
jammcq: What time you do flying to here?
20:13
<jammcq>
I will arrive at 12:20pm on Wednesday at POA
20:13
<BGomes>
12:20(UTC-3)
21:05monteslu has joined #ltsp
21:09Egyptian[Home] has quit IRC
21:10Egyptian[Home] has joined #ltsp
21:18eharrison has joined #ltsp
21:39topslakr has joined #ltsp
21:50mccann has quit IRC
21:57indradg has quit IRC
21:59petre has quit IRC
22:08BGomes has left #ltsp
22:08spectra has quit IRC
22:10
<johnny>
dberkholz, should i be creating /var/lib/tftpboot in ltsp-server ? i'm not sure how much setup we're trying to do in the ebuild
22:10
oh.. and also symlinking pxelinux.0
22:10
<dberkholz>
johnny: try a pkg_config() function to do that kinda stuff
22:11
people run `emerge --config ltsp-server` to use it
22:11
johnny: that makes it so the config step can be rerun over and over without reinstalling
22:14
<johnny>
so.. about initramfs..
22:14
should we go for initramfs-tools ?
22:15
to do otherwise.. i think would require some patching of genkernel
22:15
and updating genkernel's busybox
22:15
which has mdadm support that is not upstreamed
22:16
<dberkholz>
johnny: we should definitely use genkernel
22:17
whether that ends up using initramfs-tools on the backend is another story
22:17
<johnny>
genkernel's initramfs support is very limited .. the only part maintained is enough to handle livecds
22:18
it's not modularized like genkernel4
22:18
<dberkholz>
johnny: what did you want to do with it?
22:19
<johnny>
do with which?
22:19
<dberkholz>
johnny: with the initramfs
22:20
<johnny>
i am not sure, i'd like to use genkernel, but it is going to require a bit of work
22:21
it'd be nice if genkernel could use initramfs-tools .. or try to get the genkernel4 initramfs integrated (but created in a saner manner)
22:21
<dberkholz>
johnny: well, if you'd get your other stuff in, i could take a look at it
22:21
<johnny>
i've got almost everything in
22:22
it does require manual interaction with genkernel to get it to work
22:23
you do have to patch the file that lists all the modules that get autoloaded for busybox
22:23
to add all the network modules
22:23
i don't think it is a good idea to make users do that tho, thus the reason i brought up this discussion
22:24
since even if we did get the network modules added to the default module list , it still wouldn't support nbdroot, or allow us to use unionfs, or allow us to use uclibc
22:25
the other thing i'm currently testing, is that LDM_XSESSION idea you mentioned
22:26
<dberkholz>
yeah, seems like there's some ways to make the whole module autoloading work
22:26
<johnny>
i don't understand why it does those manually
22:26
<dberkholz>
maybe some runtime option to pass to genkernel to list modules to be autoloaded
22:26
<johnny>
genkernel4 autoloads the modules, but it does have a directive to add a module.. or class of modules to the initramfs
22:27
which is why i was so excited about it before
22:27
but it is too fragile atm
22:27
and would lead to much confusion until the name change
22:27
plus.. i think such a thing should be used by gentoo proper, not hidden off for ltsp users
22:28
<dberkholz>
johnny: yeah. we should be doing things the gentoo way, not going off on tangents
22:29
<johnny>
there is no gentoo way for this yet :)
22:29
imo.. genkernel4 embodies the gentoo way
22:29
maybe that's too strong
22:29
how about.. "comes closest"
22:31
<ryudo>
johnny
22:31
you use ltsp ?
22:31
how your terminal configs ?
22:31
<dberkholz>
johnny: it doesn't because gentoo isn't using it
22:32
johnny: we're trying to follow the ltsp philosophy here of offloading work onto other people, and that means using what gentoo uses =)
22:32
<johnny>
but it is not good if what gentoo uses .. sucks
22:32
<dberkholz>
then we have to fix it or get them to replace it
22:32
<johnny>
let's go with #2
22:33
<dberkholz>
won't happen when gk4 hasn't yet seen a release
22:33
<johnny>
i've got it down to around 6 extra steps once you run ltsp-build-client, to make it work
22:34
dberkholz, but then agaffney and wolf32o1 have to accept it
22:34
is it possible for you to talk to them and explain our situation?
22:35
<dberkholz>
they care about their situation, not ours ... so what we would have to do is figure out how to put things in terms of what they care about
22:35
<johnny>
that's the problem i'm having :)
22:36
<ryudo>
johnny :(
22:37
you use ltsp 5 correct ? how your terminals config ?
22:39
<johnny>
?
22:45
dberkholz, any ideas on that?
22:46
<dberkholz>
johnny: sure, figure out what problems there are with gk3 that gk4 fixes
22:48
<johnny>
easy
22:48
<dberkholz>
johnny: then narrow that down to problems that affect people trying to build gentoo releases
22:51
<johnny>
if you look at the output of genkernel4.. it seems like 70% of the stuff they added would be useful for such people
22:51
output of genkernel4 --help that is
22:54pscheie_ has joined #ltsp
22:57ryudo has quit IRC
23:07pscheie has quit IRC
23:08eharrison has quit IRC
23:10gonzaloaf has quit IRC
23:18abadger1999 has joined #ltsp