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


Channel log from 5 December 2008   (all times are UTC)

00:00
<ball>
exploding cuttlefish?
00:01
<ltsppbot>
"cyberorg" pasted "gnome-session-nx" (37 lines) at http://pastebot.ltsp.org/133
00:01
<cyberorg>
warren, that is what we are using, that works ^^
00:02
<warren>
His method adds no security
00:02
hmm, can't connect to pastebot
00:03
<cyberorg>
http://pastebin.com/d448bfce3
00:04
<warren>
#
00:04
nxagent :${SESS_DISPLAY} -D +kb -noxkblock \
00:04
#
00:04
-auth ~/.Xauthority -nopersistent -noshmem -noignore &
00:04
#
00:04
DISPLAY=:${SESS_DISPLAY} /usr/bin/gnome-session
00:04
<johnny>
so warren are you finally better now?
00:04
<warren>
cyberorg: it could run gnome-session before nxagent is fully running
00:04
cyberorg: that explains why it worked only 1/20 tries on my system
00:04
johnny: ?
00:05
<johnny>
you were sick awhile ago.. surgery?
00:05
are you back to normal now?
00:05
<warren>
johnny: doing a lot better, it was rough for a while
00:05
cyberorg: hmm, this script is also racy
00:05
<cyberorg>
warren, so it needs sleep 2 before last line?
00:06
<warren>
cyberorg: sometihng like that, but the while loop is also a race disaster
00:06
cyberorg: between choosing a number, checking and using it, sometihng else could grab it by accident, or intentionally
00:07
<cyberorg>
yes, a very small chance
00:08
<warren>
cyberorg: this is a classic race condition exploit
00:08
cyberorg: that directory is in /tmp with permission 1777
00:08
literally anyone else on the system can screw with another user
00:08* johnny votes to elect warren ltsp security dictator
00:09
<warren>
johnny: you don't want that, I enabled LDM_DIRECTX by default
00:09
<johnny>
well.. i guess that's a fair tradeoff considering the types of machines in use in many installs
00:10
that can make it unusable in many settings
00:10
<warren>
sure it is more secure than XDMCP
00:10
but not much
00:10
<loather-work>
i'd like to see a weak ssh cipher be used
00:10
like des56
00:10
but most distro builds blow away the weak ciphers
00:11
<warren>
I used to patch my sshd with a null cypher
00:11
=)
00:16
<loather-work>
i did that when i was testing things like traffic shaping
00:16
<ball>
double ROT-13
00:16
<loather-work>
haha :) best encryption evar
00:19* cyberorg is happy to have a first normal conversation with warren
00:19ltsppbot has quit IRC
00:20
<warren>
I'll throw in some hostility.
00:20
cyberorg: growl
00:25
<johnny>
aha.. mr dave airlie.. our graphics hero..
00:26ltsppbot has joined #ltsp
00:29
<warren>
johnny: everyone's graphics hero
00:32
cyberorg: oh!
00:32
cyberorg: I bet you've only tested NX with LDM_DIRECTX=false right?
00:40
<cyberorg>
warren, no, without that
00:41
<warren>
cyberorg: you mean unencrypted?
00:41
<cyberorg>
our default is LDM_DIRECTX=false so without that line in lts.conf
00:41
http://en.opensuse.org/LTSP/Tips_and_Tricks#Using_NXagent_to_decrease_X_traffic
00:42
so technically LDM_DIRECTX=false
01:36Loto__ has quit IRC
01:37Loto__ has joined #ltsp
01:43gate_keeper_ has joined #ltsp
02:03Loto___ has joined #ltsp
02:10tjikkun_work has joined #ltsp
02:18Loto__ has quit IRC
02:26ball has quit IRC
02:31Q-FUNK has joined #ltsp
02:31
<thelmara>
can anybody tell me where the ltsp-build-client binary is found in Fedora Core 9 after install?
02:33Q-FUNK has left #ltsp
02:40
<warren>
oh damn
02:40
nxagent fails even after it sends SIGUSR1
02:40
it needs a few seconds after that to not fail
02:46
cyberorg: hmm, ok, got it working. It dropped bandwidth usage of playing a particular youtube video from 60 to 20mbit/sec, however the video is very jerky and not steady, has this been your experience as well?
02:46
it is also impossible to scroll the window, it takes like 10-20 seconds for firefox to respond
02:48
<cyberorg>
firefox scroll is slow any way, so i didn;t notice more slowdown
02:48
<warren>
firefox scroll normally isn't slow for native X here
02:48
as long as you have sufficient bandwidt
02:49main2frame has joined #ltsp
02:49Shingoshi has quit IRC
02:49
<main2frame>
hello
02:51
Can somebody help me to debug ltspfs - client side is ok but serverside /tmp/.ltspfs-username/ mountpoint isn't created and ltspfsmount script fails
02:51
fuse is ok, device is with right permissions, user has fuse group
02:52
when running on terminal ssh -S /var/run/ldm_socket_vt7_192.168.1.253 192.168.1.253 "/usr/sbin/ltspfsmounter /tmp add"
02:53
I get errror
02:53
Error: /tmp/.test-ltspfs/tmp is not mounted
02:54Shingoshi has joined #ltsp
02:55
<cyberorg>
main2frame, what distro and ltspfs release?
02:55hanthana_ has joined #ltsp
02:55Shingoshi has quit IRC
02:56
<main2frame>
distro is ubuntu-8.04 LTS and ltspfs is 0.5.0~bzr20080109-3ubuntu2
02:56
strangely it was working initially - but all of the sudden stopped worked...it worked with single terminal - later with multiple terminals stopped working
02:57
Im just clueless right now how to debug it furter
02:57
further
02:57
aka debugging serverside fuse
02:57
that seems to fail
02:59
does anybody know how this /tmp/.username-ltspfs mountpoint is set up? by ltspfsmount?
03:01
<cyberorg>
main2frame, did you do anything like update something before it stopped working?
03:02
<main2frame>
yes I did apt-get update once - it might be that after that it got non-working state
03:02
<cyberorg>
also the client chroot?
03:02
<main2frame>
the serverside ltspfs package got minor version 3 instead of 2
03:02
no
03:03
but I downgraded now serverside back to 0.5.0~bzr20080109-3ubuntu2
03:03
clientroot won't upgrade the package
03:03
I didnt look but I suppose updates repo description was missing?
03:04
aka doing apt-get in client root wont upgrade automatically to ltspfsd 0.5.0~bzr20080109-3ubuntu3
03:04
<cyberorg>
usually it is a good idea to update client and server packages both at the same time, can you rebuild the client image?
03:04
<main2frame>
right now they are the same version - both are back to 0.5.0~bzr20080109-3ubuntu2
03:05
but still problem persists
03:05
I can upgrade both again
03:05
and try again
03:08
mhm...updates repo description is missing from client root
03:08
shouldn't it be there - is it a bug?
03:10thelmara has quit IRC
03:12jstephan_ has joined #ltsp
03:14
<main2frame>
/etc/apt/sources.list contains no hardy-updates
03:14
<warren>
cyberorg: when is 11.1 due?
03:14jstephan has quit IRC
03:16
<cyberorg>
warren, later this month
03:17dirigeant has joined #ltsp
03:21hanthana has quit IRC
03:24
<main2frame>
ok I updated client root and image
03:25
booted new image
03:25
now ltspfs and ltspfsd packages both are 0.5.0~bzr20080109-3ubuntu3
03:25
but still non-working
03:25
how can I debug serverside ltspfsmounter etc?
03:31
if I try manually to fire ltspfsmounter from terminal console - does it has to be ssh -S or ssh -X -S ?
03:35
ok - now if I try ssh -X -S /var/run/ldm_socket_vt7_192.168.1.253 192.168.1.253 "/usr/sbin/ltspfsmounter /tmp add" I get
03:35
Wrote 1b88652e06424ac252d22a4d0e2ff2c0 size 32, waiting
03:35
fuse: failed to open /dev/fuse: Permission denied
03:35
Error: /tmp/.test-ltspfs/tmp is not mounted
03:36
ls -l /dev/fuse
03:36
crw-rw-r-- 1 root fuse 10, 229 Dec 5 11:23 /dev/fuse
03:36
id test
03:36
uid=1000(test) gid=100(users) groups=100(users),117(fuse)
03:37
/dev/fuse permissions seem ok to me and test user has fuse group access
03:38
now why Im getting permission denied error from manually invoking ssh -X -S /var/run/ldm_socket_vt7_192.168.1.253 192.168.1.253 from terminal???
03:40jstephan_ has quit IRC
03:47
<cyberorg>
main2frame, post on ltsp-discuss mailing list or wait for ogra to come online
03:48
<main2frame>
I have ltsp as openvz containter
03:50
it can be something wrong with fuse device creation inside container also
03:53
ok that was it - I had accidently removed fuse device rules from ltsp openvz CT config file - so fuse device was ok inside CT but not allowed from OpenVZ hardware node
03:54japerry_cat has quit IRC
04:22dirigeant has quit IRC
05:44Egyptian[Home] has joined #ltsp
06:06dirigeant has joined #ltsp
06:10polytan has joined #ltsp
06:10
<polytan>
hi
06:10
is there any ltsp gentoo's dev here ?
06:14dirigeant has quit IRC
06:14dirigeant has joined #ltsp
06:19dirigeant has quit IRC
06:19dirigeant has joined #ltsp
06:31dirigeant has quit IRC
06:40hanthana_ is now known as hanthana
06:47alkisg has joined #ltsp
06:47
<cyberorg>
polytan, johnny
06:54BrunoXLambert has joined #ltsp
07:01dirigeant has joined #ltsp
07:01Appiah has quit IRC
07:10hanthana has quit IRC
07:16hanthana has joined #ltsp
07:18Appiah has joined #ltsp
07:18
<Appiah>
joining the channel again , reading the topic... Gentoo my fav dist "is getting very close". Is there any status on what's "missing" ?
07:19cliebow has joined #ltsp
07:31hanthana has quit IRC
07:34hanthana has joined #ltsp
07:45elisboa has quit IRC
07:45elisboa has joined #ltsp
07:54dmaran has joined #ltsp
07:58ltsppbot has quit IRC
08:06ltsppbot has joined #ltsp
08:10siki has quit IRC
08:24alkisg has quit IRC
08:24ace_suares has joined #ltsp
08:25
<ace_suares>
!seen ogra
08:25
<ltspbot`>
ace_suares: ogra was last seen in #ltsp 1 day, 15 hours, 34 minutes, and 5 seconds ago: <ogra> thanks a lot !
08:25
<ace_suares>
hi there, anyone can provide me with the command line like ssh -X somehting pulseaudio=ipaddress etc ? I see it in my top often but cant find it now
08:29rjester has joined #ltsp
08:31
<rjester>
hey all, i am having a problem with LTSP 5 and a usb CDRW. i am running on opensuse 11.1. when i plug in a USB stick, it pops up fine, but when i plug in a usb cdrw nothing happens. i can see the device being detected if i run dmesg on the thin client, but i don't know where to go from there
08:31
<rjune>
!seen ogra
08:31
<ltspbot`>
rjune: ogra was last seen in #ltsp 1 day, 15 hours, 40 minutes, and 48 seconds ago: <ogra> thanks a lot !
08:33
<ace_suares>
rjune how r u
08:34
okay never mind i got it
08:34ace_suares has left #ltsp
08:41Q-FUNK has joined #ltsp
08:42Egyptian[Home]1 has joined #ltsp
08:43Gadi has joined #ltsp
08:45
<sbalneav>
Morning all
08:45kalib has joined #ltsp
08:45
<kalib>
message
08:46
I'm using LTSP server here... But some terminals are not starting the X server. I got a message about XDMCP server... something about maximum number of sessions.. How can I check the number of sessions? And how can I correct it?
08:46
<Gadi>
kalib: are you using LTSP 4?
08:47
<kalib>
yeah Gadi
08:48
<Gadi>
gdm?
08:48
<rjune>
!s
08:48
<ltspbot`>
rjune: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
08:48
<rjune>
!g
08:48
<ltspbot`>
rjune: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
08:48
<kalib>
Gadi, yes
08:49rjester has left #ltsp
08:50
<Gadi>
kalib: check your gdm configuration for max number of sessions
08:50
<kalib>
Gadi, wich file?
08:51
<Gadi>
kalib: gdm.conf
08:52
<kalib>
Gadi, done...
08:53
after change it.. need to do something else?
08:54
<Gadi>
kalib: you will need to restart gdm - but that will kick everyone off
08:54
so do it when people are not on
08:54
<kalib>
ok
08:55
got it
08:55
<rjune>
Gadi: what's new?
08:55
<kalib>
Gadi, /etc/init.d/gdm restart, right?!
08:58kalib is now known as kalib-off
08:59Egyptian[Home] has quit IRC
09:06Q-FUNK has quit IRC
09:20spectra has joined #ltsp
09:24dirigeant has quit IRC
09:25spectra has joined #ltsp
09:27Ahmuck has joined #ltsp
09:42
<warren>
stgraber: what exactly is the goal of screen-session.d? could you please discuss it on the list?
09:42
Gadi: hey gadi
09:42
Gadi: could you please send your nx scripts?
09:43
I've already determined that NX is a disaster of code quality and its encoding is sub-par, cannot be relied upon, when it is working, which is rare.
09:43
but I want to make it work upstream anyway
09:44spectra has quit IRC
09:46
<rjune>
warren you referring to freeNX or the commercial version?
09:46
<stgraber>
warren: look at the example I just uploaded
09:46Egyptian[Home]1 has quit IRC
09:47
<warren>
rjune: both
09:47
stgraber: I don't understand the part about xorg.conf
09:47
<stgraber>
warren: if you want to update the xorg.conf for things like the Virtual parameter you need to do it before the xinitrc and we didn't have any rc script executed before running the screen.d script
09:48
<warren>
what is the Virtual parameter?
09:48
<stgraber>
warren: it's also something that will help people drop scripts there to fix some auto-detection mistake, generating a xorg.conf
09:48
<warren>
stgraber: what if your xorg.conf is completely blank?
09:48
<stgraber>
the Virtual resolution, you need it to do dual-head with the Intel video card
09:48
<warren>
stgraber: if it cannot be done without a xorg.conf it is an upstream bug
09:49
<stgraber>
well, the fact that you need Virtual for the intel video card when doing dual-head is AFAIK a hardware limitation
09:50
<warren>
stgraber: what happens with that script if your xorg.conf is completely blank?
09:50
<stgraber>
if it's an intel video card, it'll add a minimal Screen section, if it's not, you still will have a blank xorg.conf
09:50
<warren>
oh
09:50spectra has joined #ltsp
09:51
<warren>
stgraber: hmm, we could use a similar scrip to implement the other X_CONF options?
09:51
<stgraber>
yeah
09:51
<warren>
stgraber: Fedora doesn't ship configure-x.sh because it has debianisms, and we don't want an xorg.conf at all
09:51
but some people keep crying
09:51
err
09:51
X_* options
09:51
<stgraber>
Ubuntu doesn't use configure-x.sh either
09:51
<warren>
we do support X_CONF
09:53
stgraber: _UsUrPeR_ hasn't cried about Ubuntu not using configure-x.sh? There is no other way to enable DVI on his clients at the moment?
09:53
<_UsUrPeR_>
wha
09:54
I'm crying about something completely different right now. It's difficult to switch cry topics.
09:54
<warren>
_UsUrPeR_: configure-x.sh is responsible for X_* options to work
09:55alkisg has joined #ltsp
09:55
<warren>
_UsUrPeR_: so I continued to implement whitelists yesterday, then discovered something even more horrifying about how localappsd works...
09:55
_UsUrPeR_: then I got distracted by something else shiny
09:55
<stgraber>
if you want X_* to work (before they are re-implemented using screen-session.d) you need to set CONFIGURE_X=True in lts.conf, then have the X_* options
09:55
<warren>
_UsUrPeR_: i'll likely finish the whitelist today or during the weekend
09:56
stgraber: wont work here
09:56
we don't ship configure-x.sh, and it never did work on fedora
09:56
<stgraber>
warren: right, because you don't have configure-x.sh at all (I'm not even sure it works in Ubuntu, haven't used it for a long time)
09:56
<_UsUrPeR_>
warren: no prob
09:58mccann has joined #ltsp
10:08alkisg has quit IRC
10:08gate_keeper_ has quit IRC
10:10alkisg has joined #ltsp
10:13warren_ is now known as elwarreno
10:14
<elwarreno>
stgraber: is francis still alive?
10:14
<warren>
oh
10:14
stgraber: did all the lobster brain kill him?
10:17
<stgraber>
elwarreno: yeah
10:18
warren: nope, he's still weird but nothing outside the ordinary :)
10:18
warren: btw, was that you who had the photo(s) of francis eating the lobsters ?
10:20
<warren>
You are what you eat!
10:26
stgraber: could you implement some of the X_ options in your new .d? I need to better understand it.
10:26
stgraber: then I might be able to implement the rest
10:27staffencasa has joined #ltsp
10:28
<stgraber>
warren: Gadi was mainly discussing the X_ implementation, I'll see if he can implement some feature properly so we can then have a good example of how to do it
10:29
<warren>
and how does screen-session.d make sense as a namE?
10:30
stgraber: btw, I was thinking, we need .d an ldm .d K script for removing xauth tries.
10:30
otherwise xauth entries just pile up
10:30
it is not 100% obvious a safe way to do it though
10:30
you have two different types of xauth entries
10:30
with directx or no
10:31
and then there are concurrency problems
10:31tjikkun_work has quit IRC
10:35nubae has joined #ltsp
10:45dirigeant has joined #ltsp
10:47rjune__ has quit IRC
10:49japerry has joined #ltsp
10:51
<Gadi>
hi, guys
10:51
sorry - busy day so far
10:52
warren: I can get you what I have for nx, thats no problem
10:52
I think these days, tho, I have taken to running a local window manager with it so I don't need uglier acrobatics to kill the prcos it forks
10:53
but, I think I separated the wrapper script from the loop with the wm
10:53
so, whatever
10:53
I need to make it coherent for you
10:53
:)
10:53
thats mostly why its not yet in upstream
10:54
.. in terms of screen-session.d, I will try to make some examples as well
10:54
one would be to reimplement X_HORZSYNC and X_VERTREFRESH for naughty monitors
10:55
another would be to force a video driver
10:55
(XSERVER)
10:55krishna_ has joined #ltsp
10:55
<Gadi>
stgraber and I share something in common - real immediate customers that cannot be satisfied with "that's an upstream bug and there's no way to fix it for you until they do"
10:56
:)
10:56japerry has quit IRC
10:56
<Gadi>
but, some of these things are not even bugs
10:56
some are to manage naughty peripherals
10:56chrisinajar has joined #ltsp
10:58
<warren>
Gadi: the screen-session.d is the priority, NX I already know is a dead end
10:59
<Gadi>
ah, well thats encouraging
10:59
:)
10:59
whats up with NX?
11:01alkisg has quit IRC
11:06alkisg has joined #ltsp
11:08
<cliebow>
jeez:nx works awesome for me
11:10
<warren>
Gadi: it sucks
11:11
<Gadi>
ah, sounds like a discrepency
11:11
:)
11:11japerry has joined #ltsp
11:11
<Gadi>
is "sucks" a professional opinion?
11:11
:)
11:12
<cliebow>
how about a pissing match 8~)
11:13
<warren>
Gadi: I could also use the word "disaster"
11:13
Or "segfault"
11:13
Or "The people who wrote it did it wrong."
11:14
<Gadi>
hmm
11:14
I recognized one of those as a technical term
11:14
:)
11:14
does it "dump mad core"?
11:17
<warren>
I found workarounds for most of its problems
11:17
but not all
11:17
oh shit, gotta go now
11:17
bbl
11:24
<cliebow>
i meant a missing patch ;-]
11:27
<EAG>
are Fujitsu's FUTRO thinclients worth looking into?
11:29Lns has joined #ltsp
11:29kalib-off is now known as kalib
11:34
<johnny>
Appiah, polytan : hi
11:34
Appiah, i've been here the whole time :)
11:35
as in.. since you've started coming around :)
11:45klausade has joined #ltsp
11:54LTSP has joined #ltsp
11:59siki has joined #ltsp
12:03randra has joined #ltsp
12:04CAN-o-SPAM has joined #ltsp
12:05alkisg has quit IRC
12:07alkisg has joined #ltsp
12:11ogra has joined #ltsp
12:12
<Appiah>
johnny: yes?
12:12
<johnny>
you said something about gentoo before?
12:13
<Appiah>
yes , I asked
12:13
what's missing from gentoo to go from "very close" to done
12:14
<johnny>
fixing genkernel to load an exterenal config to get it to include all the network drivers we need
12:14
right now it only includes two, and you have to manually regenerate it with good network modules
12:15
it also involves using a git version of genkernel elsewise
12:15
there is a patch for it, it's just not included yet
12:15
it also won't do nbd
12:15
but then again.. neither does fedora
12:16
<Appiah>
I'm a person who prefers gentoo , and thinking of trying to contribute more
12:16
anything I can help with?
12:16dmaran has left #ltsp
12:17
<Appiah>
I wanna help in LTSP in general too I been told o join ltsp-docs
12:18vagrantc has joined #ltsp
12:19
<johnny>
Appiah, i'll ask you to test it again shortly, after i re merge
12:19
but it will involve you manually updating genkernel intramfs
12:20
you could look into quickstart
12:20
which is what i'm using to generate the chroot
12:20
<Appiah>
quickstart?
12:21
<johnny>
http://dev.gentoo.org/~agaffney/quickstart.php
12:21
<Appiah>
oh
12:21
holy crap
12:22
the installer sucked , right ;)
12:22
tried it once and went back to the normal way
12:22
So are you are making a quickstart for ltsp?
12:23
a sort of similar Ubuntu-Alternate -> LTSP Server install ?
12:24
<johnny>
i already made quickstart for ltsp :)
12:25
if you layman -a ltsp and emerge ltsp-server and look in /etc/ltsp/quickstart/profile.qs , you'll see it
12:25
<Appiah>
awesome
12:27
Well I'll get a testing envoirment up and try it
12:27
later
12:27
<Gadi>
layman? what a name... :)
12:28
<Appiah>
;)
12:28
<Gadi>
written by men with big egos
12:28
:)
12:28* ogra waves to east coast Gadi from SF
12:29* Gadi waves to ogra - welcome to the US
12:29* Gadi is happy to know they still let you in
12:29
<ogra>
hehe
12:30
<Gadi>
btw - if fdisk ever tells you to zero out the first 512 bytes of a dos parition - do it - it aint kidding
12:30
:)
12:30
<ogra>
i always look for the line with female officers and flirt a lot to make them let me in
12:30
<Gadi>
I always thought it was a suggestion
12:30
ogra: they are men - you know that right?
12:31
;)
12:31
<ogra>
not all of them ;)
12:32* Gadi tries to remember the name of that movie from the 90s
12:32
<Gadi>
damn
12:33
foiled again
12:34* Lns waves to ogra from 50 miles north of SF
12:35* ogra waves back
12:35
<Lns>
ogra, so you're traveling south to stay at that one hotel right?
12:36
<ogra>
right
12:36
<Lns>
I'm hoping to actually go to SF tomorrow for a float
12:36
<ogra>
we're currnetly at google campus
12:36
<Lns>
oh nice =)
12:36
<ogra>
where we usually spend the days
12:44
<vagrantc>
stgraber: any reason not to use run_parts_list function from ltsp-common-functions for the screen-session.d stuff?
12:45
stgraber: it also runs all scripts in both the before and after... is that intentional?
12:45
ah, nevermind.
12:45
i see the K/S handling.
12:47
though it still would be nice to use the same logic for which files are allowable.
12:50
<stgraber>
ogra: we have xrandr dual-head support in ltsp upstream !
12:51
<ogra>
cute !
12:55randra has quit IRC
13:13hanthana has quit IRC
13:13hanthana has joined #ltsp
13:23
<rjune>
ogra: !
13:25
ogra: I have a tracking # for you. it's due to be delivered on the 10
13:25
<warren>
stgraber: crap, there appear to be numerous bugs in Xorg 1.5 where if you specify Virtual in xorg.conf X fails
13:27chrisinajar has quit IRC
13:30elwarreno has quit IRC
13:30krishna_ has quit IRC
13:30krishna_ has joined #ltsp
13:32
<Gadi>
warren: upstream driver bugs? ;)
13:34
<warren>
Gadi: hmm.... it fails especially in the i810 driver
13:34
<Gadi>
i810 is old
13:34
intel is the supported one
13:35
I doubt i810 even support xrandr1.2
13:37
<warren>
no
13:37
<rjune>
ogra: you about?
13:37
<warren>
i810 is the driver name that contains intel
13:39hanthana has quit IRC
13:43ogra has quit IRC
13:46
<johnny>
it's not the driver name anymore..
13:48
<warren>
oh
13:48
I see now
13:49
both i810 and intel are provided by the driver package though
13:52jammcq has joined #ltsp
13:52
<jammcq>
hello #ltsp
13:52
<warren>
hellooooo
13:54chrisinajar has joined #ltsp
13:54elwarreno has joined #ltsp
13:55
<CAN-o-SPAM>
hi jammcq!
13:56
<jammcq>
hey can-o
13:56
<Lns>
hey jammcq
13:57
<cliebow>
jammcq:!
13:58
<jammcq>
hello gents
13:59spectra has quit IRC
14:13cliebow has quit IRC
14:15Gadi has left #ltsp
14:34artista_frustrad has quit IRC
14:34krishna_ has quit IRC
14:35artista_frustrad has joined #ltsp
14:42
<vagrantc>
so, is ~/.dmrc "going away" sometime soon? what is it likely to be replaced with?
14:42
supporting persistant session selection is probably the biggest missing feature in LDM that i can think of.
14:44
<warren>
vagrantc: upstream seems unable to make decisions. I personally would lean towards just doing our own thing until there is a clear upstream standard.
14:45
vagrantc: a future ldm version can see the existing ~/.ldmsomething and migrate to ~/.upstreamsomething and remove .ldmsomething
14:46
<vagrantc>
warren: i don't really like the idea of future migrations, and ~/.dmrc has been around for a while now, no?
14:46
<stgraber>
I'd just go with .dmrc for now and then if there is a change for both gdm and kdm then do it
14:47
<warren>
that seems fine
14:47
<stgraber>
it's a feature I have on the list of things I need to implement for some customers (mainly bi-lingual school districts)
14:47
<vagrantc>
stgraber: default language selection and all that?
14:47
<stgraber>
and they currently use both desktop computers and thin client so .dmrc seems to be a good choice as it'd then work on both
14:47
vagrantc: yeah, language and session
14:50kalib has quit IRC
15:01
<vagrantc>
warren, johnny, cyberorg: is there a consistant URL from each of your distros where i can download your source versions?
15:01
<warren>
vagrantc: given that I don't do any deviation from upstream...?
15:02
vagrantc: I don't fork and patch older versions
15:02
<vagrantc>
i.e. for debian, it would be http://ftp.us.debian.org/debian/pool/main/l/ltsp/
15:02
warren: given even that
15:02
<warren>
http://kojipkgs.fedoraproject.org/packages/ltsp/
15:02
http://kojipkgs.fedoraproject.org/packages/ldm/
15:03
http://kojipkgs.fedoraproject.org/packages/ltspfs/
15:04
<vagrantc>
warren: does the .src.rpm contain the upstream sources?
15:04
<warren>
vagrantc: yes
15:04
vagrantc: oh, my only patch is an alternate layout for ldm
15:05
vagrantc: I'm waiting for the magical ldm layout configurability
15:05
<vagrantc>
i'm not concerned about patches, just upstream versions
15:05
<warren>
if that doesn't come, then I'm including my patch as an IFDEF
15:06
<vagrantc>
warren: those URLs are not likely to change?
15:06
<warren>
njo
15:06
no
15:06Eghie has joined #ltsp
15:06
<vagrantc>
warren: thanks!
15:10user1 has joined #ltsp
15:10
<user1>
hey all
15:10
i have a java based app that works normally on the server, but on the thin clients it takes MUCH longer and is MUCH slower, especially when it comes to opening new windows
15:10
client find is unworkable
15:11
sometimes an 8-10 second delay in opening a window
15:11
are we SOL?
15:11Egyptian[Home] has joined #ltsp
15:11
<vagrantc>
if you don't care about security AT ALL, you could try enabling LDM_DIRECTX=true in lts.conf
15:12
user1: oh, and:
15:12
!release
15:12
<ltspbot`>
vagrantc: "release" is please mention the linux distro and release you're using :)
15:12
<user1>
sorry
15:12
kubuntu-8.04
15:12
<warren>
vagrantc: I'd like to rename it to LDM_INSECURE in the docs and examples, but continue to support the old name.
15:12
<user1>
i wanted to check the resources on the thin client but it keeps telling me me account is expired
15:13
<vagrantc>
warren: i'm supportive of that, but i suspect others won't like it
15:13
<user1>
though they are 1.5ghz amd and 256mb ram, one would think that would be enough for any thin purpose, right?
15:14
<vagrantc>
localapps support wasn't in hardy, but that might be a way to do it, with machines that powerful.
15:14
<warren>
that amount of RAM is a bit small
15:15
<user1>
vagrantc: is it in ibex?
15:15
especially small for a java app, right?
15:16
<vagrantc>
depends on how much ram the java app takes, but yes, that would be my main concern.
15:17
<user1>
vagrantc: i would like to check the resources on the thin client, but it keeps telling me the account is expired, even when i chroot in and set the pw
15:17
<sbalneav>
user1: You're probably being bitten by:
15:17
https://bugs.edge.launchpad.net/ubuntu/+source/ia32-libs/+bug/277069/
15:19
<user1>
sbalneav: wow, thanks for that...
15:19keithclark has joined #ltsp
15:20
<user1>
# ... please create:
15:20
# /var/lib/tftpboot/ltsp/i386/lts.conf instead for your changes
15:21
<Lns>
user1: see https://help.ubuntu.com/community/UbuntuLTSP/ClientTroubleshooting
15:21
(for unlocking root account)
15:22
<user1>
Lns: thanks
15:22
<Lns>
np
15:23
<user1>
sbalneav: it makes sense that it would be the bug if every other app seems to work normally, right?
15:23
<sbalneav>
Yeah, as soon as you mentioned java, that's what my mind went to.
15:24
<Lns>
user1: I experience that same issue w/Java.. there's a good workaround though
15:24
<user1>
sbalneav: then again one of the people mentioned slow typing in open office but i dont get that
15:24
then again open office is mostly c++ these days, or am i wrong about that
15:25CAN-o-SPAM has quit IRC
15:26alkisg has quit IRC
15:27
<sbalneav>
I beleive it's C++
15:29Egyptian[Home] has quit IRC
15:32
<johnny>
vagrantc, my src version is bzr
15:32
atm
15:32
it pulls HEAD from bzr atm
15:33
<vagrantc>
johnny: but there's some file that they download which tells it how to build, no?
15:33
<johnny>
oh.. sure
15:33
<vagrantc>
and version information can be extracted from that?
15:34
<johnny>
http://git.overlays.gentoo.org/gitweb/?p=proj/ltsp.git;a=tree
15:34
version information ? no
15:34
<vagrantc>
so the package is unversioned?
15:34
<johnny>
the official version in tree is *-9999
15:34
yes.. unversioned atm
15:34
until i find a place to put tarballs :)
15:34
hehe
15:35
wasn't planning on forcing the issue until it goes in tree
15:35
instead of just an overlay
15:35Egyptian[Home] has joined #ltsp
15:38
<vagrantc>
johnny: what's "in tree" vs. "overlay" ?
15:40
<johnny>
overlays .. same thing as external repos in debian
15:40
same functionality at least..
15:42
<vagrantc>
what's an external repo in debian?
15:42
<johnny>
stuff that goes in /etc/apt/sources.list that isn't the official debian upstream
15:42
or ubuntu
15:42
<vagrantc>
ah.
15:42
got it.
15:44user1 has quit IRC
15:46BrunoXLambert has quit IRC
15:47japerry has quit IRC
15:48
<Lns>
aka "third party unsupported" i'd assume =)
15:49
aka "It's an RPM...so it might work...but..umm...yeah."
16:00user1 has joined #ltsp
16:01
<user1>
sbalneav: i believe i owe you a debt of gratitude
16:02
sbalneav: for finding that bug... i gotta run, but thanks
16:02user1 has quit IRC
16:03
<vagrantc>
heh. i figured out an ugly hack to parse ~/.dmrc
16:03
getltscfg -c /path/to/.dmrc -n Desktop -a
16:03vvinet has quit IRC
16:03ogra has joined #ltsp
16:15keithclark has quit IRC
16:17Ahmuck has quit IRC
16:19ccherrett has quit IRC
16:20Ahmuck has joined #ltsp
16:25jammcq has quit IRC
16:29SSLater has joined #ltsp
16:39ccherrett has joined #ltsp
16:50keithclark has joined #ltsp
16:59vagrantc has quit IRC
17:09johnny has left #ltsp
17:12ogra has quit IRC
17:13Shingoshi has joined #ltsp
17:22johnny has joined #ltsp
17:22johnny is now known as Guest1673
17:23ogra has joined #ltsp
17:26Lns has quit IRC
17:26ltsppbot has quit IRC
17:31ltsppbot has joined #ltsp
17:32keithclark has quit IRC
17:33Guest1673 has quit IRC
17:33johnny has joined #ltsp
17:42mccann has quit IRC
17:47otavio_ has joined #ltsp
17:48otavio has quit IRC
18:04otavio_ is now known as otavio
18:06Eghie has quit IRC
18:10ogra has quit IRC
18:14ccherrett has quit IRC
18:15yanu has quit IRC
18:15yanu has joined #ltsp
18:17ccherrett has joined #ltsp
18:20vagrantc has joined #ltsp
18:21ltsppbot has quit IRC
18:26ltsppbot has joined #ltsp
18:53staffencasa has quit IRC
18:54
<Ryan52>
vagrantc, if only we had getltscfg on the server :)
18:56
<vagrantc>
Ryan52: well, i was thinking of just downloading the ~/.dmrc to the thin-client from and ldm rc.d script
18:56
that way we don't need to support server-side stuff.
18:56
<Ryan52>
hrm. ya, that could work. what about making changes, though?
18:56
<vagrantc>
push it back
18:57
<Ryan52>
well, obviously, but how do we actually make the change?
18:57* Ryan52 wonders if it lets you have multiple "[Desktop]" sections
18:57
<vagrantc>
edit the file client-side...
18:57
push it to the server...
18:58
<Ryan52>
okie dokie. :)
18:58
<vagrantc>
sound evil?
18:58
<Ryan52>
not really.
18:59
why don't we move out the ssh command to it's own shell script? then we could implement hooks like we just did for before and after X...
18:59
I don't see any reason why it needs to be ran from the c program.
18:59
it's kinda messy, I guess..
19:01
<vagrantc>
i don't really see why nearly everything but the greeter isn't written in shell...
19:01
but i haven't been the one to make those calls.
19:02
<Ryan52>
then I would be sad, cause ldm is the only time I ever get to write real C :p
19:03
<vagrantc>
heh
19:08
Ryan52: so, i'm not quite sure, but i envisioned an ldm rc.d script to download the ~/.dmrc and maybe ask if they want to set the new defaults and such ... and then somewhere in the C code, and then something in ldm.c between rc_files("start") and x_session() to set values as appropriate ...
19:08
i.e. if the session type is default, don't muck with the .dmrc, but read the default session from it.
19:09johnny has quit IRC
19:09
<Ryan52>
but we already determind that there is no way to make a shell script change a C programs environment.,
19:09
<vagrantc>
if ~/.dmrc and the selected session differ, then prompt to update it...
19:09
Ryan52: so you need to parse the downloaded ~/.dmrc ...
19:09
<Ryan52>
in c?
19:09
then what's the point of an rc.d?
19:09
<vagrantc>
i guess not much.
19:10
<Ryan52>
exactly! :)
19:10
<vagrantc>
could just do it all in C, i guess.
19:10
<Ryan52>
ya, that's what I'd vote for.
19:11
<vagrantc>
i just like to keep as much of it not over my head as possible :)
19:11
though i'm slowly learning.
19:11
<Ryan52>
just learn C already! :)
19:11
<vagrantc>
pfft.
19:11* Ryan52 <3 C :p
19:12
<vagrantc>
i have a hard enough time learning something like ruby
19:12johnny has joined #ltsp
19:12
<vagrantc>
and it'd be a stretch to say i even know python.
19:14
<Ryan52>
wait, what? you know ruby?
19:14
<vagrantc>
enough to make simple patches here and there
19:14
<Ryan52>
oh.
19:15* Ryan52 has never seen vagrantc write any ruby
19:16
<vagrantc>
Ryan52: i think there's some rudimentary commits in a couple of the smaller ruby scripts at freegeek
19:16
anyways... off to aikido!
19:16
<Ryan52>
ok, have fun! :)
19:36vagrantc has quit IRC
19:47johnny has quit IRC
19:47johnny has joined #ltsp
19:50Loto___ has quit IRC
20:05Loto__ has joined #ltsp
20:18johnny has left #ltsp
20:19rjune_ has joined #ltsp
20:25japerry has joined #ltsp
20:37japerry has quit IRC
20:42japerry has joined #ltsp
20:44petre has joined #ltsp
20:49japerry has quit IRC
20:54F-GT has quit IRC
20:57johnny has joined #ltsp
21:09petre has quit IRC
21:32ltsppbot has quit IRC
21:36ltsppbot has joined #ltsp
21:46
<stgraber>
Ryan52: having .dmrc support is very high on my list of thing to do but as you mentioned, this one needs to be able to change the parent's environment. The easiest way I thought we should implement it is to move the actual session open from the C code to a shell script, then make all ldm rc.d scripts to be sourced instead of simply being called.
21:47
they'd then be able to change every environment variable they want making things like .dmrc handling possible
21:47japerry has joined #ltsp
21:47
<Ryan52>
ya.
21:49
tho, we need tihngs like the pid in ldm..
21:49
pid of ssh
21:49
so that gets...complicated.
21:49
so we might as well just reimplement ldm in shell :(
21:49
wait, nevermind.
21:50
it connects in one ssh command and then runs the session in another.
21:50
so we just move the session one out -- easy.
21:53
stgraber: are you intending on working on that this weekend? mind if I "steal" it from you? I intend to work on ldm Sunday anyway, since I want to improve the ldminfo xsession stuff..
21:53* Ryan52 isn't sure of "very high on my list of thing to do" means things you want done or things you want to do or both..
21:53
<Ryan52>
s/of/if/
21:54try2free has joined #ltsp
21:54
<elwarreno>
how do you do a private message in irc?
21:54
i'm currently using konversation...
21:54
as a client
21:54
<Ryan52>
elwarreno: not sure about the GUI way, but "/query user" should open up a new window to talk to them in.
21:56
oh, we do need the pid.
21:56
hrm.
21:58alkisg has joined #ltsp
21:59
<Ryan52>
well, shell has a "wait" so maybe we don't need it :)
22:00
ya, that'll work fine afaict. yay! easy! :)
22:01CaScAdE^FarAway has joined #ltsp
22:02try2free has left #ltsp
22:08CaScAdE^1arAway has quit IRC
22:10
<stgraber>
Ryan52: feel free to work on it, I'll be flying to UDS on Sunday so won't do much work on it then :)
22:27ltsppbot has quit IRC
22:31ltsppbot has joined #ltsp
23:16siki has quit IRC
23:35LTSP has left #ltsp
23:53beakburke has joined #ltsp
23:59hanthana has joined #ltsp