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


Channel log from 27 August 2008   (all times are UTC)

00:04vagrantc has quit IRC
00:14vagrantc has joined #ltsp
00:40FuriousGeorge has joined #ltsp
00:41
<FuriousGeorge>
hey all
00:41
first attempt at ltsp
00:41
i get a samsung 720xt, which was impossible to get till last week, and it turns out to be the biggest piece of crap
00:41
not so much the hardware's fault, i dont think. but windows and samsung
00:43
its one of those not-so-thin clients which work off of some form of solid-state memory. first of all not enough free space there to install all the updates. in order to get the audio working you need about 9 files you can only get from another xp pro installation (afaik), ie7 crashes whenever you type in the search toolbar
00:43
and not enough local space to install sp3
00:44
so anyway, the error im getting from the bios pxeboot is:
00:45
PXE-E79: NBP is too big to fit in free base memory
00:45
i learned you shouldnt free base anything, but ill make an exception here :)
00:52
Aug 27 05:52:43 bmy in.tftpd[30819]: tftp: client does not accept options is the only thing i get in my logs
00:52phantom has joined #ltsp
00:52F-GT has quit IRC
00:59
<vagrantc>
FuriousGeorge: sounds like the file it's trying to download is too big ... are you telling it to download pxelinux.0 ? or something else?
01:08
<FuriousGeorge>
something else, and you are right. about 60 seconds ago i saw a thread where someone had the same problem
01:09
apparently that kernel i was trying to make a kernel designed for etherboot or something, and you cant just do that
01:09
now i just need to get my nfs settings working, and we might be getting somewhere
01:12
omg its actually booting
01:12
parse error...
01:15
it didnt like the paramater 128m for swapfile_size parameter, but it said the default was 64m
01:20
now it keeps complaining about the last line of lts.conf, but that line is just a comment
01:21dukai has joined #ltsp
01:23
<FuriousGeorge>
i must have some syntax error before hand that the parser isnt catching till the lst line
01:23
<vagrantc>
!pastebot
01:23
<ltspbot>
vagrantc: "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.
01:24
<FuriousGeorge>
vagrantc: k
01:26
<ltsppbot>
Someone pasted "##############################" (22 lines) at http://pastebot.ltsp.org/51
01:26
<FuriousGeorge>
http://pastebot.ltsp.org/paste
01:26
clever bot
01:27
<vagrantc>
FuriousGeorge: what linux distro and release?
01:27
<FuriousGeorge>
ltspbot: it doesnt like parameters when i input them like X_MOUSE_BUTTONS = 3
01:27
<ltspbot>
FuriousGeorge: Error: "it" is not a valid command.
01:28
<FuriousGeorge>
or X_MOUSE_BUTTONS 3
01:28
gentoo / 4.2
01:28
the parser doesnt seem to like it then
01:28
bot might be too clever :)
01:34
vagrantc: is there a some character sequence i need to end lts.conf with or something?
01:35
<johnny>
4.2? 4.2 is masked
01:35
5 is almost done
01:35
<FuriousGeorge>
johnny: 5 isnt in portage
01:35
for me
01:35
<johnny>
it isn't yet
01:35
<FuriousGeorge>
using desktop profile
01:35
<johnny>
i'm just letting you know
01:35
it's accessible via layman
01:35
<FuriousGeorge>
ill upgrade, im not scared, just wanted to give it a shot
01:35
shoulda thought about layman
01:35
<johnny>
it will be in portage after a few things
01:36
mainly automation stuff
01:36
genkernel doesn't easily let us provide custom configs for the initramfs
01:36
which means it loads a bunch of stuff that we don't need.. and doesn't let us add our own customizations
01:36
robbat2 has a patch to provide a custom arch directory, which i think i will do
01:38
<FuriousGeorge>
johnny: is the lts.conf parsing at all changed?
01:38
<johnny>
you have to setup the dhcp, and tftp stuff manually too, haven't thought of a nice way around that, as other distros don't mind including patches directly to dhcp server packaging
01:38
but gentoo would never do that.. and understandably so
01:39
once my sources are integrated into upstream, and tested once again.. then comes the guidexml file for it, and then we can get it itno portage
01:39
<FuriousGeorge>
are options in lts.conf supposed to start with the work "option" or end with a ';' or something
01:39
why else would i get a parse error on the last line...
01:39indradg_ has quit IRC
01:39
<johnny>
i don't know anything about 4.2 other than it is deprecated, so i never touched it
01:40indradg_ has joined #ltsp
01:40
<johnny>
by the time i got into ltsp, ubuntu already had an ltsp 5 implementation, which is what i deploy
01:41indradg_ has quit IRC
01:50marjan has joined #ltsp
01:56
<FuriousGeorge>
i still get the parse error, but it gets a little farther on now... now it complains that it cant detect my video card and i should specify one in lts.conf, then it says xserver failed
01:57
for some reason it cant open the log file, so i get no log @ /tmp/mnt/xorg.log
01:59
<johnny>
try running getltscfg manually (if such a thing exists for 4.2)
02:00
maybe you'll see where it is failing
02:00
if not.. just keep editing :)
02:04
<FuriousGeorge>
johnny: if it exists its not in my path
02:13Pascal_1 has joined #ltsp
02:17dean_za has joined #ltsp
02:23LaserJock has quit IRC
02:26phantom has quit IRC
02:26F-GT has joined #ltsp
02:28vagrantc has quit IRC
02:34
<Pascal_1>
bonjour
02:45
<FuriousGeorge>
johnny: damn gentoo. i just spent 45 minutes trying to remember how to work layman
02:45
and failed
02:47
<johnny>
uhmm.. you coulda read the man page..
02:48
<lejo>
use a distro *hide* ;-)
02:52martin3z has joined #ltsp
02:58leio_ has joined #ltsp
02:59
<FuriousGeorge>
johnny: reading it is one thing, understanding it is another
03:06leio has quit IRC
03:08Q-FUNK has joined #ltsp
03:13The_Code has joined #ltsp
03:21
<FuriousGeorge>
finally, got it
03:29marjan has quit IRC
03:30dean_za has quit IRC
03:32marjan has joined #ltsp
03:36FuriousGeorge has quit IRC
03:36Nick_Hill has joined #ltsp
03:37
<Nick_Hill>
I had a problem where a laptop would hang for 4 minutes whilst booting LTSP but not with straight Ubuntu. I have found that adding acpi=off for the boot parameters of that kernel fixed it. Where should I report this to help other users?
03:40
It would hang between detecting network card and Clocksource tsc unstable (delta = -3291491518 ns)
03:40
I believe the line Clocksource tsc unstable caused the delay. I found related articles suggesting acpi=off to fix the delay.
03:45tjikkun_work has joined #ltsp
03:56Q-FUNK has quit IRC
04:17
<Nick_Hill>
How should I report my findings (above) to help others?
04:18marjan has quit IRC
04:20
<cyberorg>
Nick_Hill, ubuntu launchpad?
04:23
<Nick_Hill>
cyberorg, Thanks, that is one option, but I know there are fora for LTSP directly, so was wondering if there were anything more appropriate.
04:23
<cyberorg>
Nick_Hill, ltsp-discuss ML
04:31
<Nick_Hill>
cyberorg, thanks
04:34Q-FUNK has joined #ltsp
04:50ace_suares has joined #ltsp
04:50
<ace_suares>
!seen ogra
04:50
<ltspbot>
ace_suares: ogra was last seen in #ltsp 18 hours, 3 minutes, and 58 seconds ago: <ogra> (/usr/bin is the right place)
04:51
<Q-FUNK>
ace :)
04:58mikkel has joined #ltsp
04:59moldy has joined #ltsp
04:59
<moldy>
hi
04:59
on ubuntu 8.04, i need to turn up the sound output volume for the clients -- how do i do this?
05:00
i can do this using alsamixer on the client, but how can i make that setting persist?
05:02Pascal_1 has left #ltsp
05:11moldy has quit IRC
05:12moldy has joined #ltsp
05:26
<johnny>
there's a way to set it in lts.conf perhaps?
05:26
i know warren worked on a patch for newer ltsp
05:26
<moldy>
johnny: that would be great... but i did not find anything in the documentation
05:26
<johnny>
you could set it globally in /etc/asoundrc perhaps?
05:27
i'm not sure if asoundrc allows that, but it might
05:31
<moldy>
johnny: i think i found a way... thanks for your guidance :)
05:37
<ogra>
Nick_Hill, add clocksource=hpet to your pxe config
05:37* ogra will comment on the bug as well .... thats an issue with 2.6.24 and newer on certain cpus
05:38
<ogra>
(known issue btw)
05:56
<warren>
johnny: I aborted working no it because gadi already had something
05:56
we need gadi to post what he has
06:04
<moldy>
some usb sticks don't work well: they do not show up on the desktop -- manually mounting them works fine. what could cause this?
06:05
some other usb sticks work fine
06:06
when i boot the machines locally, all sticks work fine
06:10ogra has quit IRC
06:10ogra has joined #ltsp
06:41
<Nick_Hill>
ogra, I have tried clocksource=hpet, but it makes no difference.
06:45etyack has joined #ltsp
06:46etyack has quit IRC
06:47etyack has joined #ltsp
06:49
<etyack>
morning all
06:51
<martin3z>
etyack, morning? :) for you 2 o'clock is morning?
06:52
<etyack>
martin3z: for me 7:45 am morning :)
06:53
<martin3z>
ok, sorry, i forgot time differences :)
06:58
<etyack>
you know how us Americans are, we think everything revolves around us :)
07:01
<ogra>
Nick_Hill, if you boot with no splash, what clocksource is picked in the end ?
07:02mhterres has joined #ltsp
07:03moldy_ has joined #ltsp
07:04warren has quit IRC
07:11otavio has joined #ltsp
07:16
<ace_suares>
ogra ! how r u. Can I PM you for 5 minutes?
07:16_thomas__ has joined #ltsp
07:17
<_thomas__>
Hi everyone
07:17
<ace_suares>
ogra about https://bugs.launchpad.net/ubuntu/+bug/120375
07:17
<_thomas__>
I'm getting in trouble with tftp timeouts, everyone told-me to include two options on my dhcpd.conf
07:17
<ace_suares>
Q-Funk !!
07:18
<_thomas__>
128 and 129, someone could explain-me what this options do?
07:19moldy has quit IRC
07:19
<ace_suares>
_thomas__: http://www.ltsp.org/twiki/bin/view/Ltsp/NetworkCards i think only needed for ISA cards... you dont use those do yo ?
07:19etyack has quit IRC
07:20
<ace_suares>
_thomas__: it's to set IRQ for ISA cards... i never seen this in 3 yrs or so :-)
07:20
<_thomas__>
ace_suares, no, some onboard lans, and most of them offboard with etherboot
07:20
etherboot works fine
07:20
<ace_suares>
_thomas__: they are most likely pci so these options are not used i think... !
07:21
<Q-FUNK>
:)
07:21
<_thomas__>
ace_suares, and what about nfs not responding
07:21
?
07:21
<ace_suares>
_thomas__: the onboard with pxe or etherboard ?
07:21
<_thomas__>
onboard with pxe
07:21
<ace_suares>
_thomas__: don't know, are you using LTSP 5 ???
07:21
<_thomas__>
no, ltsp4
07:21
problems with printers on ltsp5...
07:21
<ace_suares>
Q-FUNK: he he saw you're 'hi' only now... how are you ?
07:22
_thomas__: can't help you with 4 (and not much with 5 either).
07:22
<ogra>
ace_suares, i cant tell much about #120375, feel free to drop by in #ubuntu-devel or #ubuntu-server and talk to kirkland
07:22
<ace_suares>
ogra things are worse, i have contact with kirkland already. he made a patch.
07:22
<ogra>
_thomas__, ltsp 4 isnt maintained since 3 years and hasnt seen security updates either
07:22
<ace_suares>
ogra just wanted to ask you if you think it's okay that edubuntu hardy can not boot if the root raid disk degrades !
07:22
<ogra>
_thomas__, whats your issue with pronting ?
07:22
*printing
07:23
<ace_suares>
ogra and I tried to follow steps to get it inculded in hardy-updates but got stopped at step 4.
07:23
<ogra>
ace_suares, edubuntu hardy is an addon of educational packages, nothing more ...
07:23
<_thomas__>
ogra, honestly, i don't know, I just take this project yesterday
07:23etyack has joined #ltsp
07:23
<_thomas__>
other people do this
07:24
<ogra>
ace_suares, ubuntu is what you want to talk about ... edubuntu has no install disk anymore etc
07:24
<ace_suares>
ogra i know but the bug means that the whole server can't boot if the disk degrades !
07:24
ogra okay understood but I tought maybe you could upload the patch so the process is n't stopped.
07:24
ogra kirkland had no uupload rights either
07:24
<ogra>
ah
07:24Blinny has joined #ltsp
07:24
<_thomas__>
I need it working for a few weeks, and when we got some time, i will change to ltsp5
07:24
<ogra>
ask him to ping me for sponsorship then
07:25
_thomas__, well, people who still can support ltsp4 are rare ... yymv
07:25
err
07:25
ymmv
07:25
:)
07:25
<ace_suares>
will do. he says the MAYBE in september he will have time. But I am sure there are others that can move this forward ? I am tearing my hair out that NONE of hardy servers can boot with degraded raid !
07:26
ogra I am *so* ready to write to linux-haters :-)
07:28
<_thomas__>
ogra, unfortunately I'm new in this job, and my word don't matter much yet, so I can't ask for money to do this in the right way, and the business is almost stopped
07:32
<Blinny>
ogra: I have a proof-of-concept of mounting an external /home by simply using automount on the server
07:33
<ogra>
ace_suares, the prob seems to be that he'S the only one knowing about it and is on a backpack holiday atm
07:33
if you would have pinged a month ago it would have been easier :/
07:37
<Blinny>
Is there any reason why a user's /home would continue to be accessed after logout? autofs is not unmounting the user's /home
07:37
<ogra>
gvfs probably
07:38
<Blinny>
Would that show up as a separate mount for that user? I'm not seeing it.
07:38
(I logged out last night and came in this morning to /home still being mounted for that user)
07:41
<ogra>
gvfs mounts are hidden
07:41
~/.gvfs is the mountpoint
07:42
<Blinny>
Ah, ok. Is there any problem with having this mount stay around? My users will change sites pretty frequently and their local home may still be mounted from another location.
07:44
I see other .gvfs mounts in /etc/mtab but not this particular user.
07:46
<ogra>
well, i'm not an autofs guy but i doubt autofs will unmount if the dir is still acessed
07:47
<Blinny>
I'll do some more reading. Thanks.
07:54_UsUrPeR_ has joined #ltsp
07:56
<Q-FUNK>
ace_suares: I'm fine. still unpacking things at the new Artec office and working on the dbe61/dbe62 datasheet.
07:56lupi has joined #ltsp
08:01
<martin3z>
where i found quick tutorial form liveusb for ltsp thin client making? i found only some discussion. i need simply howto.
08:03
i have usb key, and i want to use it as bootloader for thin client.
08:04
my bios doesn't support booting from net
08:04_UsUrPeR_ has joined #ltsp
08:07Nick_Hill has quit IRC
08:08
<_UsUrPeR_>
Hey, anyone know how to reference a xorg.conf file from an lts.conf in Fedora?
08:10
<ogra>
like in any other distro :)
08:10
<_UsUrPeR_>
I'm trying to get dual-monitor support
08:10
<ogra>
X_CONF=/etc/myxorg.conf will point to /opt/ltsp/$ARCH/etc/myxorg.conf
08:10
<_UsUrPeR_>
thanks
08:11Subhodip has joined #ltsp
08:11mccann has joined #ltsp
08:13
<lupi>
martin3z: http://elinux.org/LTSP-HDboot #looks like it might do the trick for you
08:17warren has joined #ltsp
08:17
<_UsUrPeR_>
morning, warren
08:18etyack_ has joined #ltsp
08:28etyack_ has left #ltsp
08:28etyack_ has joined #ltsp
08:30GodFather has joined #ltsp
08:31GodFather has joined #ltsp
08:31etyack has quit IRC
08:32
<cyberorg>
martin3z, http://en.opensuse.org/LTSP/USB_Quickstart
08:33etyack_ is now known as etyack
08:47jammcq has joined #ltsp
08:47
<jammcq>
hola
08:48
<etyack>
jammcq: hey
08:48
<jammcq>
hey et
08:48
etyack: hey, you pocket dialed me last night
08:48
or, one of your kids did :)
08:48
<_UsUrPeR_>
jammcq: were you the one that is making the thincan?
08:48
<jammcq>
I had a nice little conversation with parker
08:49
_UsUrPeR_: not me
08:49
<_UsUrPeR_>
k
08:49* jammcq makes only software these days
08:49
<etyack>
jammcq: was he in the middle of his tantrum
08:49
<jammcq>
etyack: umm, I didn't hear any anger. just sounded like kids having fun
08:50
<etyack>
jammcq: twins are teething again and really happy about it
08:50
<jammcq>
ah, the good old days X 2
08:50
<_UsUrPeR_>
etyack: so great that you take that in stride.
08:51
<etyack>
_UsUrPeR_: yah right
08:52
<_UsUrPeR_>
etyack: the sarcasm is melting my monitor. It's not rated for anything this high.
08:53
<ogra>
_UsUrPeR_, thincan -> Q-FUNK
08:54
<etyack>
jammcq: did you add info about bts to the wiki?
08:58
ogra: do you know if Q-FUNK is in the middle of their office move?
08:58elisboa has quit IRC
08:59
<ogra>
etyack, nope, i dont spy on Q-FUNK :)
08:59elisboa has joined #ltsp
08:59
<etyack>
ogra: oh, my mistake. I thought you were the head spy on #ltsp
09:00
<ogra>
heh
09:04mccann has quit IRC
09:14jonnor has joined #ltsp
09:15
<sbalneav>
Morning all
09:15
<ogra>
!s
09:15
<ltspbot>
ogra: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
09:15
<sbalneav>
hey hey!
09:16
<etyack>
morning Scotty!
09:16
<warren>
sbalneav: is it posssible to run ldm as a stand-alone app in order to test themes?
09:16
<sbalneav>
Morning etyack
09:16
<ogra>
warren, you could run it when the backend wasnt split
09:17
thats how i developed all my themes
09:17
<sbalneav>
warren: Should be able to, I used to.
09:17
<ogra>
not sure thats still possible after scotties split
09:17
<warren>
etyack: hmm, so you're actually able to do dual monitor on ubuntu with the X_* options alone?
09:20
<etyack>
warren: no, we are testing a custom xconf first. then once we know which options must be specified we will try with X_* options.
09:21
<warren>
etyack: ok
09:30K_O-Gnom has joined #ltsp
09:31K_O-Gnom has joined #ltsp
09:33mikkel has quit IRC
09:35martin3z has quit IRC
09:39
<ogra>
etyack, note that the X_MONITOR_OPTION stuff isnt upstream yet (pushing it today) so its only in hardy
09:39
<warren>
ogra: what does that do?
09:39
ogra: and yes, please do push it upstream
09:41
<ogra>
warren, adding Option "key" "value" to the monitor section
09:41
<etyack>
ogra: hardy or intrepid?
09:41
<ogra>
hardy
09:41
<etyack>
k
09:41zoon has joined #ltsp
09:42
<ogra>
i'm about to push the last bits to the tree today/night as i need to roll the final upstream packages for intrepid today
09:42
feature freeze in ubuntu tomorrow
09:42
after that new upstream packages are painful to get in (lots of paperwork)
09:43dukai has quit IRC
09:45
<jammcq>
etyack: also note that different chipsets have different methods of doing multi-head. the ATI's don't work properly with the settings in xorg.conf
09:46
and nvidia has their "twinview" method.
09:47
<etyack>
jammcq: right. dual-head is still a PIA as far as i'm concerned. admittedly, i've gotten lazy and use the nvidia gui.
09:47
<jammcq>
and i'm using the ati way
09:47
reluctantly
09:48bitplex has joined #ltsp
09:48
<etyack>
i'm sure via/openchrome is special in their own way
09:49
<warren>
BTW, what is a safe # of colors to stick with for a LDM background?
09:49
etyack: does the nvidia GUI have command line switches?
09:50
<etyack>
warren: yes
09:51_thomas__ has left #ltsp
09:51Blinny has quit IRC
09:54
<ogra>
warren, 256 should be good to not get stairstepping in gradients (at least if your bg isnt to colorful)
09:55
<etyack>
warren: nvidia-settings command line options: http://pastebin.com/d713d8238
09:59zoon has quit IRC
09:59
<sbalneav>
"You can have any colour background you want, so long as it's black." -- Scott Balneaves channeling Henry Ford. :)
10:05
<ogra>
haha
10:05
<laga>
ogra: yay for blogging! i'm looking forward to reading more
10:06
<ogra>
yeah, was about time ...
10:06
i always said "well, if the new server is up" ... but never found the time :)
10:15mccann has joined #ltsp
10:16tutu has joined #ltsp
10:16tutu is now known as Pascal_1
10:18GodFather has quit IRC
10:21mccann has quit IRC
10:23hp has joined #ltsp
10:23
<hp>
hi ogra, thanks, I got the usb printer working, it was a hostname issue I think
10:23
not really the jetpipe
10:24
:)
10:24
<ogra>
cool !
10:25
<hp>
so good night and thanks again
10:25hp has quit IRC
10:25LaserJock has joined #ltsp
10:25
<sbalneav>
Morning LaserJock
10:26
<LaserJock>
hi Scottie
10:26
<sbalneav>
Hmmmmm.
10:27
Killing off all other processes on login. Just checked out the pkill manpage, and pkill -s 0 -v seems VERY interesting.
10:29
<Q-FUNK>
etyack: we moved last week
10:29
etyack: right now, I'm mostly rearranging my office space
10:29
<etyack>
Q-FUNK: did you receive my email?
10:30
<Q-FUNK>
etyack: yes, I was discussing it with the rest of the management team just now.
10:31
<etyack>
Q-FUNK: great
10:31
Q-FUNK: let me know if you have any questions
10:32staffencasa has joined #ltsp
10:36Pascal_1 has left #ltsp
10:44ltspbot` has joined #ltsp
10:44ltspbot has quit IRC
10:44
<ogra>
not sure how much time my work leaves me to update it after FF beyond ubuntu specific patches
10:45
so i need to get in everything that is left over from hardy
10:46
<_UsUrPeR_>
ok, having problems with using an xorg.conf file to dual-monitor. I'm using Ubuntu, and I am using the openchrome driver. It's confusing because it's giving a segfault on cdpinger.
10:48tjikkun_work has quit IRC
10:48
<etyack>
_UsUrPeR_: do not specify an xorg.conf in lts.conf and boot the client
10:48GiantPickle has joined #ltsp
10:49otavio has quit IRC
10:49
<etyack>
_UsUrPeR_: i thought there is an xorg.conf in /tmp/ or in /etc/X11
10:49
_UsUrPeR_: start with that as your basis instead on the one you are using
10:49
<_UsUrPeR_>
etyack: it boots fine without a specified xorg.conf
10:50
<etyack>
_UsUrPeR_: did you look for an xorg.conf in the paths i mentioned above?
10:50
<ogra>
there shouldnt be one
10:50sbalneav has joined #ltsp
10:50
<ogra>
we delete it deliberately
10:50
<_UsUrPeR_>
etyack: you're talking about a non-chroot xorg.conf right?
10:50Q-FUNK has quit IRC
10:50
<_UsUrPeR_>
Actually I do see one in /etc/X11/xorg.conf
10:51
<ogra>
in the chroot ?
10:51
<etyack>
_UsUrPeR_: in the chroot?
10:51
<_UsUrPeR_>
ogra: no, not in the chroot
10:51
<ogra>
ah
10:51
<_UsUrPeR_>
srry :)
10:51
<ogra>
the one in /etc is what your server uses :)
10:51
<etyack>
ogra: is there a way to stop deleting the xorg.conf in the chroot?
10:51
<_UsUrPeR_>
nothing in the chroot for xorg.conf
10:52
<ogra>
etyack, why would you do that ? its generated during chroot build with the servers /proc mounted (so only applies to the server)
10:53
with hardy you actually dont even need an xorg.conf at all aprt from keyboard setting if you are in a no english env
10:53
with intrepid we wont generate one at all
10:53
<etyack>
ogra: we are working on dual display and want to start with a base lts.conf that works for the thin client
10:53
then add the dual display options
10:53
<ogra>
right, have a look at the configure-x.sh script
10:54
thats responsible for creating an empty xorg.conf during client boot
10:54
<etyack>
s/lts.conf/xorg.conf
10:54
ok
10:54
<ogra>
and then sed's thorugh that generic file to apply lts.conf values
10:55
(its pretty ugly ... makes sure you dont wear glasses, the code might make them burst)
10:55
<_UsUrPeR_>
crap.
10:55
<etyack>
i'll put my contact in
10:55* _UsUrPeR_ takes off glasses and receives a blast of code in the face
10:55
<_UsUrPeR_>
it burns!
10:56
<etyack>
_UsUrPeR_: another way is to install Ubuntu locally on CF or USB hd and work from there.
10:57
<_UsUrPeR_>
etyack: that's a great idea! I'm going to give that a shot
10:57
<etyack>
once you create a working xorg.conf, use that for the ltsp chroot
10:59etyack has quit IRC
11:01etyack has joined #ltsp
11:01
<_UsUrPeR_>
wb!
11:05
<ogra>
use a USB cdrom and boot a liveCD helps too to get an xorg.conf
11:06
<_UsUrPeR_>
ogra: downloading now. Will do.
11:06
<etyack>
does the live cd support via openchrome well?
11:06
<ogra>
no idea, but i'D guess so
11:06
<_UsUrPeR_>
etyack: where is that external cdrom?
11:07
nm, got it
11:11
<etyack>
_UsUrPeR_: made sure you have liveCD ver 8.04.1
11:11
<_UsUrPeR_>
downloading newest release now
11:11LaserJock has quit IRC
11:16mccann has joined #ltsp
11:20The_Code has quit IRC
11:21
<ogra>
warren, did you get your recent kernel booing in vbox ?
11:21* ogra just fund the magic option to at least make it boot
11:21
<ogra>
slow though
11:25etyack has quit IRC
11:26LaserJock has joined #ltsp
11:27GodFather has joined #ltsp
11:30spectra has joined #ltsp
11:32
<warren>
ogra: I never have used vbox
11:32FuriousGeorge has joined #ltsp
11:33
<FuriousGeorge>
hey all
11:35
topic says "gentoo is getting close" but i found 0 documentation for version greater than 4.2... I did find an overlay with ltsp5 in a 3rd party repo, with a package called ltsp-server, which is supposed to be ltsp5, but that fails like this:
11:36
* ${EBZR_UPDATE_CMD} ${EBZR_OPTIONS} "${repository}" || die "${EBZR}: can't merge from ${repository}.";
11:36
* The die message:
11:36
* bzr.eclass: can't merge from bzr://devel.localmomentum.net/ltsp-gentoo.
11:36
seems like the source is 404
11:37
<warren>
FuriousGeorge: ask dberkholz and johnny for the latest
11:38
<FuriousGeorge>
actually, i think i emailed one last night. its always nice to see the names in the bug reports also on irc
11:39* FuriousGeorge pokes dberkholz with a long stick from a safe distance
11:49
<FuriousGeorge>
for "small scale" deployments, what is the advantage of using ltsp over setting up plain old diskless nodes?
11:50Q-FUNK has joined #ltsp
12:01
<_UsUrPeR_>
q-funk: finished testing the thincan.
12:01
the thing is great!
12:03
<Q-FUNK>
hm?
12:03
nice :)
12:03
<warren>
my only complaint is lack of power button
12:04moldy_ has quit IRC
12:04
<_UsUrPeR_>
I was actually going to ask about that...
12:04
<warren>
well, that and the driver needs work
12:04
<_UsUrPeR_>
warren: driver?
12:04
<Q-FUNK>
warren: what work?
12:04
<warren>
Q-FUNK: all the rendering issues with latest X.org
12:04
<Q-FUNK>
adding a power button would add a movable (and therefore breakable) component. it's not gonna happen.
12:05
warren: I don't see any of that here on ubuntu.
12:05
<warren>
Q-FUNK: add two exposed pins at the surface that could be shorted? =)
12:05* warren loves rebooting computers with a screwdriver.
12:05
<warren>
Q-FUNK: X.org server too old
12:05* Q-FUNK loves to screw female drivers
12:05
<warren>
huh
12:06* _UsUrPeR_ drops his monacle
12:06
<warren>
monacle? heh
12:06
<Q-FUNK>
monocle :)
12:06
<_UsUrPeR_>
touche
12:06
<warren>
Q-FUNK: idea!
12:07
Q-FUNK: where the power plugs in, sell an optional attachment
12:07
Q-FUNK: a button that sticks out the top that cuts the power
12:07
<Q-FUNK>
hehe
12:07
<warren>
Q-FUNK: plug the PSU into the (thing), and the (thing) into the power port on the back of the thincan.
12:07
I'm series
12:07
serious
12:08Pascal_1 has joined #ltsp
12:08
<warren>
Q-FUNK: that would solve the only complaint with the thincan, and you could make a few bucks selling this as an optional add-on.
12:08
Q-FUNK: and you can send me $0.20 per unit for the royalty.
12:08
<Q-FUNK>
:D
12:09
<warren>
want a mock up? =
12:09
=)
12:09
<_UsUrPeR_>
q-funk: that's a good deal. I would take that
12:09
<warren>
_UsUrPeR_: you think people would pay a few bucks per button?
12:10Pascal_1 has left #ltsp
12:10
<Q-FUNK>
really, we expect those who absolutely insist on shutting it down every day to plug it into a power bar annd to flick the switch from there
12:10
<_UsUrPeR_>
I think q-funk would be hard-pressed to put a power button on after I fully understand what he was going for in the design. "no moving parts" is admirable.
12:10
<warren>
OK then, *I'm* going to market a thincan button.
12:10
<_UsUrPeR_>
lol
12:11
<warren>
I just have to find someone to manufacture it.
12:11
hm
12:11
<_UsUrPeR_>
radio shack: 10 for $3
12:11
<warren>
Q-FUNK: could you manufacture it for me? =0
12:11
=)
12:11
<Q-FUNK>
heheh
12:11
contract manufacturing works
12:11
it'
12:11
<_UsUrPeR_>
DIY install for power switch
12:11
we can create a new web site based solely on thincan mods
12:11
<warren>
Q-FUNK: dude, I'm totally serious, many of your existing and future customers want this as an option.
12:11
<Q-FUNK>
s our primary buisness, too: custom design and manufacturing
12:12
<_UsUrPeR_>
next on the list: a thincan with a build-in lcs display
12:12
lcs = lcd
12:12
<warren>
I have other devices that could use the same button.
12:13
Q-FUNK: all linksys products for example use the same plug
12:13
and they have no power button
12:13
<Q-FUNK>
"at Togami accessories & co., we know how to push your buttons"
12:14
<_UsUrPeR_>
q-funk: maybe a tilt switch like in a pinball machine ;)
12:14
<FuriousGeorge>
let's say, i have a regular old diskless node, and i decide i want to add 100 more... can i convert to ltsp?
12:14
<_UsUrPeR_>
flip it upside down, and it turns off
12:14
<warren>
Q-FUNK: you really don't like this idea?
12:14
Q-FUNK: I seriously think many thincan customers want this, plus other customers will find other uses for it.
12:14
<Q-FUNK>
now that you mention it, it sounds like an obvious one
12:15
<warren>
_UsUrPeR_: heh, that's cute.
12:15
<_UsUrPeR_>
call it a thin etch-a-sketch or something. thin-a-sketch?
12:15
<warren>
_UsUrPeR_: someone could accidentally turn it off that way though.
12:16
<_UsUrPeR_>
warren: that keeps them on their toes, see. It'll be like the old nintendo. one wrong move and you crash the system, along with hours of gaming.
12:16
true story
12:16
<warren>
_UsUrPeR_: just blow on it
12:16
<Q-FUNK>
hheh
12:16
<_UsUrPeR_>
warren: lol, yeah. That always worked
12:17
<warren>
Q-FUNK: I'm willing to negotiate a lower royalty, like $0.00
12:18
<_UsUrPeR_>
that is, indeed, a generous offer.
12:18
<warren>
The product can be called "The Missing Button"
12:18
<ogra>
warren, a powerbutton ? are you mad ? that gives the user control over the device !
12:18
<warren>
ogra: who cares, if they're willing to pay for it.
12:18
<ogra>
now who would want *that*
12:19
:)
12:19
<warren>
alcohol has worse implications
12:19
but people are willing to pay for it
12:19
<ogra>
heh
12:20
well, it is fun until next morning
12:21
<_UsUrPeR_>
bah. If done properly, it stops being fun the same evening
12:21
<Q-FUNK>
until your employer sues you for accidentally flicking a switch
12:22
<warren>
Q-FUNK: you seriously wont make "The Missing Button"?
12:22
<Q-FUNK>
we clould
12:22
it would be easy
12:22
<_UsUrPeR_>
included in the new switched thincans, you will need to include a "switch protocol" addendum to the manual, in order to avoid catastrophic energy savings.
12:22
<Q-FUNK>
heheh
12:23
<warren>
themissingbutton.com is not taken
12:24
<_UsUrPeR_>
oh man, reg that shit. create a product called "thincan w/button"
12:24
<ogra>
bah, to trivial
12:25
<warren>
Q-FUNK: oh, some of the newest clients out there are shipping with integrated SD reader
12:25
Q-FUNK: that's pretty killer awesome for a thin client
12:27
<Q-FUNK>
that's a corporate liability
12:27
<warren>
I guess some people want it (like schools) but others don't
12:27
<ogra>
oh, seems we'll actually ship with 2.6.27
12:28
<warren>
it'll be ready in time?
12:28otavio has joined #ltsp
12:28
<ogra>
no idea
12:28
but apparently ready enough after lots of testing by the opinion of our kernel team
12:29
<warren>
uh
12:29
2.6.27 right now is a disaster in rawhide
12:29
works great on most machines, but others... oh man
12:29
<ogra>
so seems they'll switch this week with option to switch back to .26
12:29
well, .26 isnt much better imho
12:29
<warren>
26 is a lot more stable on two of my systems
12:30
<ogra>
the madwifi replacement crap doesnt work on any of my machines
12:30
<Q-FUNK>
26 is stable but it introduces too many changes
12:30
<ogra>
luckily we still ship madwifi as fallback
12:30
<Q-FUNK>
ogra: I'm betting udev issues
12:30
<ogra>
Q-FUNK, no, the module simply doesnt work
12:31
main issues seem to be nvidia and fglrx for .27 atm
12:31
but there is work going on to fix that to not leave ati and nvidia users in the cold it seems ...
12:32
though they might only get in as SRU
12:32
<Q-FUNK>
what? canonical is shipping free non-ati/non-nvdia substitutes? :D
12:32
<warren>
good thing we don't care about nvidia or gflrx
12:32
fglrx
12:32
<ogra>
lfrxg ?
12:32
fxgfl?
12:32
<laga>
"fglörx"
12:32
</blech>
12:34
<ogra>
warren, well, we know most cheap lappies have them still ...
12:34
so we do care for these users :)
12:34
<Q-FUNK>
we're gonna be laaaaaaaaaaaaaaaaaaap dancing
12:35
<ogra>
Q-FUNK, you really shouldnt have that much beer that early in the evening :)
12:35
<Q-FUNK>
it's 20:35 here
12:35
<ogra>
oh, youre 1h ahead
12:35* ogra forgot
12:35
<Q-FUNK>
and hellhunt.ee is a really great place
12:35
<ogra>
Q-FUNK lives in the future
12:36
<warren>
http://lwn.net/Articles/294891/
12:37
<Q-FUNK>
meh
12:37* ogra sighs about all the FUD
12:37
<ogra>
and goes out ...
12:38
<warren>
This stance was harder to argue a few years ago. But all the hardware vendors are now caving toward us.
12:38
except nvidia
12:38* ogra sighs about all the FUD
12:38
<ogra>
and goes out even more :P ...
12:38
<warren>
have fun
12:39
<Q-FUNK>
I'm coming out! I want the world to know, gotta let Hardy show!
12:40
<FuriousGeorge>
that's appropriate theme music if i've ever heard it
12:41johnny has left #ltsp
12:41
<Q-FUNK>
:D
12:50Blinny has joined #ltsp
12:50
<_UsUrPeR_>
I never suspected that I would get a song stuck in my head due to IRC
12:51johnny has joined #ltsp
12:51
<Q-FUNK>
:D
12:53bobby_C has joined #ltsp
12:55Q-FUNK has quit IRC
12:55GodFather has quit IRC
12:56X0d_of_N0d has quit IRC
12:57X0d_of_N0d has joined #ltsp
12:59GodFather has joined #ltsp
13:00johnny has left #ltsp
13:10etyack has joined #ltsp
13:13
<_UsUrPeR_>
so... dual screens......
13:20johnny2 has joined #ltsp
13:21johnny2 has left #ltsp
13:22Subhodip has quit IRC
13:23vagrantc has joined #ltsp
13:24* ogra is back
13:24
<ogra>
warren, http://lwn.net/Articles/294681/
13:25johnny2 has joined #ltsp
13:25
<warren>
ogra: i think the point is lost
13:25
ogra: I was attacking only the philosophy of accepting proprietary drivers
13:25
ogra: it is dangerous, and the entire community would be in a lot worse position if Red Hat had made that decision years ago.
13:26* ogra wonders ig warren has ever tried ubuntu on a system with nvidia or ati card
13:27
<warren>
I've tried to install ubuntu once
13:27
it failed
13:27
not sure why
13:27* ogra neithe
13:27
<ogra>
r
13:27
:)
13:28
bu you probably should get familiar with stuff you rant about befor you rant about it :)
13:30mhterres has quit IRC
13:33johnny2 has joined #ltsp
13:35johnny2 has left #ltsp
13:35johnny has joined #ltsp
13:37mhterres has joined #ltsp
13:37
<_UsUrPeR_>
envyng-gtk is my friend.
13:38
<ogra>
well, envy isnt shipped ... but jockey is ... and in intrepid the two will be integrated afaik
13:38
but warren would know that we dont install any nonfree stuff at all if he had ever tested it :)
13:41
<johnny>
except firefox.. depends on how hardcore you are on that issue
13:41
<ogra>
and we never did
13:41
well, i dont consider FF nonfree
13:41
its naming probably :)
13:41gbolte has joined #ltsp
13:42* ogra wonders, does FC install iceweasle ?
13:43
<warren>
no
13:43
iceweasel is a fork of old firefox unfortunately
13:43
<ogra>
its not
13:43
its a patchset
13:43
<warren>
Note, the FSF does not consider the trademark restriction to be non-free in itself
13:43
<ogra>
that applies to any version ...
13:43
<vagrantc>
well, there's actually two iceweasels.
13:43
<warren>
Debian is the one that disagrees
13:43
<vagrantc>
there's the one from the gnu project, and the one debian ships
13:44
<ogra>
debian also disagrees about firmware :P
13:44
which is just silly
13:44
<warren>
Debian, Fedora and FSF have slightly differing opinions on a few matters
13:44
<vagrantc>
if you have to ok patches against upstream in order to ship, i don't see how that isn't free.
13:44
<warren>
vagrantc: nothing forces you from changing the name to ship it anyway you want
13:45
vagrantc: FSF does not consider Fedora's trademark restriction to be non-free
13:45
<vagrantc>
warren: other than impracticality.
13:45
<ogra>
right, thats what the iceweasle patchset does
13:45
<warren>
vagrantc: Fedora lets you call your derivative Fedora as long as it contains Fedora components. If you stray from that, you must call it something other than Fedora.
13:45
<ogra>
changing the name only
13:45
<warren>
not only name
13:45
iceweasel also rips out questionable stuff like firefox
13:45
like the part that download data from google
13:46
<ogra>
ubuntu is different here ... you have to add the term remox to the name but using the trademark is fully allowed
13:46
*remix, heh
13:46
so people can distinguish it from the official release
13:47
<vagrantc>
my take on firefox is that the trademark restrictions, while technically free software to the letter, involve technical restrictions which violate the spirit of free software.
13:48* ogra adds --no-install-recommends to APT_GET_OPTS ...
13:48
<ogra>
that recommends business is just to messy
13:48
<vagrantc>
nothing restricts debian from making long elaborate patch sets just to rename it, so that they can ship it with patches, but it sure is a lot of stupid work.
13:50
<warren>
vagrantc: I wouldn't mind personally going with a renamed firefox in Fedora, and that always remains an option if mozilla angers us enough.
13:50
vagrantc: we make it *really* easy to rebuild our firefox with a different name
13:51
<ogra>
hmmm ... do i drop configure-x.sh completely ... =
13:51
?
13:51
<vagrantc>
why mozilla doesn't make it easy to rebuild with another name, logo, etc. is beyond me. then the whole issue would essentially be moot.
13:51
<warren>
vagrantc: but that being side, Fedora has a similar trademark restriction, and FSF approves of it
13:51
ogra: if you don't want the X_ options to work anymore
13:51
<ogra>
warren, well, making it optional indeed
13:51
<warren>
ogra: I'm thinking how to implement a mini-configure-x.sh that does only the X_* options
13:51
<ogra>
hal-set-property :)
13:51
<warren>
that doesn't work for most of the X_* options
13:52
<ogra>
you want something like this:
13:52
http://paste.ubuntu.com/40402/
13:52
sure it does
13:52
well, it should at least
13:52
i didnt test yet
13:52
<warren>
what options work with this?
13:52
borrowed from fedora?! =)
13:52
<ogra>
but it should work for monitor and driver options
13:53
thats the hal callout script i'm working on for touchscreens
13:53
similar stuff will work for all input devices
13:53
just need to adapt it
13:53
yeah, its derived from the fedors kbd script
13:53
*fedora
13:53
<warren>
well, the X_ options i'm referring to are not input related
13:54
<ogra>
no, but it should work as well afaik
13:54
i'll check that after feature freeze
13:54
for now i need to make decisions how my upstream tarball looks :)
13:54
everything beyond that can happen later
13:55
<warren>
you have a different idea of "upstream" if you're deciding that
13:55
=)
13:55
<vagrantc>
use a one-line ascii-art for the name of the tarball
13:55
<ogra>
warren, upstream as in "ubuntu plugins" and installed packages
13:55
vagrantc, <grin>
13:56juanpaul has quit IRC
13:56Q-FUNK has joined #ltsp
13:57johnny has left #ltsp
13:59johnny has joined #ltsp
14:03
<sbalneav>
ogra: Didn't you say the "other people's devices" showing up on all desktops got fixed?
14:03
Because one of my users just plugged something in, and it's on my desktop.
14:04
I'm on hardy.
14:07
<ogra>
sbalneav, using a default ubuntu wih the default group scheme ?
14:07
(i.e. one group per user as default group)
14:08
<gbolte>
ogra, how is that supposed to work in a networked environment where users primary group might be something like "Domain Users" or something
14:08
<sbalneav>
ah, no
14:08
All our users are in the same group.
14:08
<ogra>
gbolte, its a workaround until upstream fixes it poperly
14:08
<vagrantc>
gbolte: i'm sure we'd accept patches to improve it, if they didn't break anything.
14:08
<sbalneav>
That's gonna be the problem.
14:09
<ogra>
which upstream refused to do since he fears access() on a networked fs might hang the system
14:09
<gbolte>
yeah network users are always going to be broken the way it stands it seems
14:09
<ogra>
vagrantc, i'd rather see them applied upstream
14:09
its a general prob upstream has to solve and upstream knows about
14:10
<vagrantc>
ogra: agreed. but i'd also like to have something working for our users.
14:10
<ogra>
gbolte, not with a proper group scheme in your ldap server
14:10
<vagrantc>
i.e. if the patch is not too intrusive or hackish, i don't see a real reason not to apply it.
14:10
<ogra>
vagrantc, it works if you dont break groups
14:10
<vagrantc>
but this is entirely hypothetical...
14:10
<ogra>
people using ldap simply dont folow the default scheme
14:11
vagrantc, its trivial to fix
14:11
<gbolte>
ogra, depends on the ldap
14:11
<ogra>
just add a group per user, make that group the default
14:11
will work in every ldap i know on linux
14:11
<lupi>
anyone know how to disable swap in ltsp5? this the right place to ask?
14:11
<sbalneav>
Who upstream are we talking about for what? ltspfs or ldap?
14:11
<gbolte>
ogra, yeah I suppose if you added a group per user
14:11
lol
14:11
but thats messy
14:12
<vagrantc>
sbalneav: i think gvfs or gnome-vfs or thereabouts
14:12
<ogra>
no thats the defaul in debian and ubuntu
14:12
since debian exists
14:12
ubuntu just inherited that scheme
14:12
<gbolte>
ogra, but its not cross distro
14:12
<sbalneav>
ah, ok, thought ogra was taking a poke at me :)
14:12
<ogra>
nothing messy about that
14:12
sbalneav, me ? on you ?
14:12
sbalneav, BFF remeber ? i'D never do that !
14:13
:)
14:13
<sbalneav>
Oh, fah dee da. I need a poke from time to time to keep me honest :)
14:13
<vagrantc>
gbolte: i think that is also default on fedora... if it works for debian, ubuntu and fedora ... i can't say it's not cross-distro.
14:13Q-FUNK has quit IRC
14:13
<johnny>
ogra, so.. my users are all in their own groups, i'm not using ldap, yet i see all the floppies
14:13
<vagrantc>
it's maybe not absolutely universal, but nothing is.
14:14
<ogra>
gbolte, feel free to discuss and fix upstream ... http://bugzilla.gnome.org/show_bug.cgi?id=526320
14:14
<vagrantc>
and then there's johnny, who manages to have it broken anyways :)
14:14
johnny: you sure it's their default group?
14:14
<ogra>
sbalneav, https://bugs.launchpad.net/ubuntu/+source/glib2.0/+bug/210379 see my last comment here
14:14
<johnny>
how can i tell what is default vagrantc ?
14:15
<vagrantc>
johnny: getent USERNAME
14:15
<warren>
lupi: nbdswap specifically?
14:16
<vagrantc>
johnny: getent group $(getent passwd vagrant | cut -d : -f 4)
14:16
s,vagrant,$USERNAME,
14:16
johnny: think that should do it
14:17
<johnny>
then yes..
14:17
they are all there
14:18
this is an install that has been around since at least dapper.
14:18
it didn't always have ltsp
14:18
<ogra>
your ltspfs is up to date ?
14:19
<_UsUrPeR_>
q-funk: are you back?
14:19
flpxj nm. :|
14:21
<johnny>
Installed: 0.5.0~bzr20080109-3ubuntu3
14:21
hmm..
14:21
<ogra>
well, that should fix it
14:21
how are the permissions in /media/$USR ?
14:21
*USER
14:21
<sbalneav>
Yikes, that is NOT going to work in our environment.
14:22
I'll have to think of something else. Is this because of gvfs?
14:22
<ogra>
sbalneav, well, convince davidz to accept the access() call
14:22
yes
14:22
<johnny>
ogra, root:$USER
14:22
<sbalneav>
crumb
14:22
<johnny>
for ls -l /media/*
14:22
<ogra>
sbalneav, http://bugzilla.gnome.org/show_bug.cgi?id=526320
14:22
<johnny>
err ls -ld /media/*
14:23
755
14:23
<ogra>
doesnt look like you use the fixed lbmount
14:23
weird
14:23
it shoudl be created 750
14:23
<warren>
the remaining problem with it is if a user's primary group is shared by other users
14:24
<ogra>
right
14:24
<warren>
actually
14:24
<ogra>
warren, like for AD users
14:24
<warren>
my earlier proposal would have avoided that problem
14:24
but vagrantc didn't like it
14:24
<gbolte>
ogra, thats our problem
14:24
:)
14:24
AD group
14:24
<warren>
700 permission with fsacl's instead of 750
14:24
<gbolte>
Domain Users
14:24
<johnny>
would it have changed the privs ogra? those directories have been around
14:24
from before
14:24
<ogra>
gbolte, right, convince upstream to accept the access() patch we used to use in gnome-vfs
14:24
<johnny>
as i still see directories
14:25
<warren>
Hmm, I'm going to test my fsacl and 700 permission idea.
14:25
<johnny>
for users that don't even get used
14:25
<gbolte>
ogra, eh I am not really too concerned with it anymore as we took matters into our own hands
14:25
<johnny>
what script would have changed the privs for those ogra?
14:25
<gbolte>
lol
14:25
<ogra>
johnny, they should be removed
14:25
johnny, usually by ldm on logout
14:25
<gbolte>
too much arguing and not enough DOING
14:26
<warren>
ogra: btw
14:26
+ if (g_access (path, R_OK|X_OK) != 0) {
14:26
<ogra>
gbolte, the fix is clear
14:26
<warren>
ogra: is there a W_OK flag? wouldn't that work as well?
14:26
<johnny>
ogra, my usesr never logout
14:26
<warren>
oh
14:26
<ogra>
warren, no idea if teher is a W_OK
14:26
<warren>
I guess that wouldn't help because you do sometimes want to see readonly mounts
14:26
<johnny>
relying on somebody not to just press the power button is not cool
14:26
<ogra>
johnny, you might have used a pre release version that didnt clean them up for example
14:27
<johnny>
yep
14:27
that'll do it
14:27
thanks ogra
14:27
i'll test it out tonight
14:27
<warren>
W_OK would solve this problem
14:27
<ogra>
just wipe the dirs once all are logged out
14:27
<johnny>
or just chmod them..
14:27
which is what i plan on doing
14:27
<ogra>
warren, right
14:27
warren, would you comment on the bug ... i'm not sure if pitti didnt do aynthing already
14:28
<warren>
who is pitti?
14:28
<ogra>
he's on vacation until monday so i cant ask him now
14:28
<warren>
ogra: W_OK would solve it, but it would be wrong.
14:28
<ogra>
warren, did you read the bug ?
14:28* johnny tells warrent that is W_NOT_OK
14:28
<johnny>
warren*
14:28
<ogra>
pitti is the guy who proposed to fix it
14:28
<warren>
what filesystem is /media/username?
14:29
<johnny>
hmm?
14:29
<sbalneav>
Looks like I'll have to patch gvfs locally
14:29
<warren>
exactly that
14:29
<ogra>
warren, http://bugzilla.gnome.org/show_bug.cgi?id=526320#c35
14:29
<warren>
yes I saw that
14:29
<johnny>
sbalneav, sounds like fun..
14:29
<ogra>
warren, ok
14:29
<warren>
what filesystem is /media/username?
14:29
on your distro?
14:29
<johnny>
what does that question mean?
14:29
<ogra>
warren, ??
14:29
<johnny>
it's in /
14:29
so it's whatever / is
14:29
<warren>
like on my distro, / is squashfs read-only
14:30
and /media/username is tmpfs
14:30
<ogra>
ext3 by default ... but can also be reiser whatever
14:30
<johnny>
aha
14:30
<ogra>
warren, your server is squashfs readonly ?
14:30
<warren>
ooh
14:30
shit
14:30* ogra is confused
14:30
<warren>
ok
14:30
<ogra>
:)
14:30
<warren>
then/media/username is ext3 here then
14:30
I'm going to test my non-W_OK idea
14:30
<ogra>
warren, david wanst concerned about that
14:31
davidz i mean
14:31
he was scaed about other mounted FSes
14:31
<johnny>
FuriousGeorge, i never said genkernel had disadvantages per se.. the steps are nearly the same.. just a minor edit needs to be made, i plan on using genkernel anyways
14:31
<_UsUrPeR_>
am I doing something wrong when specifying an xorg file in lts.conf? My client keeps coming up with a cdpinger segfault.
14:31
<ogra>
like /media/$USER/mynfsmount
14:33jammcq has quit IRC
14:34
<gbolte>
ogra, are you using gnome?
14:34
<FuriousGeorge>
johnny: i c, thanks for the response
14:35
<warren>
_UsUrPeR_: cdpinger segfaults on its own for many people
14:35
_UsUrPeR_: it hasn't caused any visible problems that I can see though
14:38
<johnny>
gbolte, he is a gnome.. big difference :)
14:38
<_UsUrPeR_>
warren: oh. Huh. Must not have noticed it prior
14:38
<gbolte>
er I guess it could be anyone running gnome, anyway we have a strange thing happen where we get the device showing on the desktop but then we also get another icon that says its location is the correct mount point for the device but its mime type is application/octet-stream
14:38
<johnny>
yeah.. my next step is actually making them mount again.. :)
14:38
<warren>
_UsUrPeR_: this wasn't the first thing that segfaulted that made it seem like it was the problem when it really was something else
14:38
_UsUrPeR_: (your earlier ldm problem)
14:39
<_UsUrPeR_>
warren: yeah, I need to get used to seeing those :)
14:39
<warren>
if somebody wants to look at cdpinger and figure out why it is crashing, that could eliminate some false positives
14:39
<gbolte>
and when you click on the icon it pops an error saying could not display "/media/gbolte/atacd-cdrom" there is no application installed for this file type
14:40
the other icon works fine though
14:40
pops up a nautilus window to browse the contents of the drive
14:42Blinny has quit IRC
14:42
<gbolte>
nobody knows what I am talking about I bet
14:43johnny has quit IRC
14:44johnny has joined #ltsp
14:44
<johnny>
oops.. ghosted myself
14:46
<gbolte>
lol
14:48
<lupi>
warren: I want all swapping off completely on my client
14:52
<warren>
lupi: on fedora NBD_SWAP=off works for me
14:52
lupi: why do you want to turn off swap?
14:52
<lupi>
warren: I ran ltsp4 previously and it booted really fast
14:52
warren: now it's *really* slow on my 32MB P133
14:53
<warren>
lupi: that might be a problem that LTSP5 just requires a lot more RAM
14:53mccann has quit IRC
14:53
<warren>
lupi: it might even require nbdswap with LTSP5 to be stable with a 32MB client
14:53
<lupi>
warren: oh, ok. perhaps I'll go back to ltsp4 then. thanks very much!
14:53
<warren>
you can easily oom die
14:54
lupi: what hardware is the client?
14:54
ogra: let me know what you decide on to replace configure-x.sh, i'm thinkinga bout that right now as well.
14:56
<ogra>
warren, well, having something that checks if any non input X option is used and then fire it off would be right i guess
14:56
for a start
14:56
but for now i first need to get hal running
14:57
*SIGH*
14:57* ogra installed hal but forgot dbus in his current image
14:59
<warren>
huh
14:59
you don't have hal anymore?
14:59
<ogra>
we never had hal in client chroots
14:59
<warren>
oh
15:00
I was forced to add it to my F10 client chroot
15:00
<ogra>
waste of resources
15:00
but now we need it
15:00
<warren>
yeah
15:00
<ogra>
and i just forced apt to ignore recommends ...
15:00
so hal is installed without the dbus package it recommends
15:00
which indeed breaks it
15:01* ogra builds another image
15:03
<warren>
yeah, i had to go through a few iterations to get it to work
15:03
<ogra>
silly progress
15:04
why cant we all use abacuses
15:04
<_UsUrPeR_>
anyone know what this means? E: main.c: GID of user 'pulse' and of group 'pulse' don't match.
15:04
<warren>
ogra: nvidia abacus?
15:04
<ogra>
lol
15:04
<warren>
I have a wicked physics engine on my abacus
15:04
<_UsUrPeR_>
it's appearing just below *Starting LTSP client...
15:04
<warren>
what distro?
15:05
<_UsUrPeR_>
debian
15:05
<warren>
ok
15:05
<_UsUrPeR_>
err.. no
15:05
sorry, ubuntu
15:05
strange because there has never been a user/group called pulse
15:07
<ogra>
and you dont need that group at all
15:07
pulse on the client is started as system service
15:07
<lupi>
warren: stoneage AST PC, no idea what mobo, Pentium133, 32MB EDO RAM, s3virge (2M I think)
15:07
<ogra>
lupi, 32m ??
15:07
<_UsUrPeR_>
ogra: oh. What is it's function?
15:07
<ogra>
you cant switch off swap then
15:07
_UsUrPeR_, its the sound server
15:08
<_UsUrPeR_>
ogra: huh. weirder and weirder
15:08
<warren>
lupi: ooh. s3virge isn't even supported in the latest X anymore.
15:08
the driver died i think
15:08
<ogra>
lupi, a recent 2.6 kernel with initramfs needs about 48M to boot
15:08
in ubuntu swap is automatically used if it detects less than 48M
15:08
<lupi>
ogra: oh, oops. think i was running 2.4 before
15:09
<ogra>
so there is no way to disable it without rendering your system unbootable
15:09
<lupi>
cool
15:09
<warren>
ogra: where is the code to detect < 48MB?
15:09
<ogra>
i'm currently applying the changes to ubuntu intrepid to make booting without swap possible though ... thrugh using compcache
15:09
<warren>
< 48MB isn't upstream
15:09
<ogra>
warren, ltsp_nbd
15:09
<warren>
ogra: how much of your stuff isn't upstream?
15:09
<ogra>
it is
15:10
warren, thats ages old and upstream
15:10
i think about 20 LOC of configure-x.sh arent upstream
15:10
because they were added shortly before release ... and i'm pushing and comminting them right now
15:10
<warren>
you pushing those 20 upstream?
15:10
<ogra>
nothing that should concern you
15:10
<warren>
k great
15:11
hmm
15:11
I guess I dno't need the 48MB detection
15:11
<ogra>
it helps
15:11
<warren>
I'm adding swap even on clients with 2GB of RAM
15:11
<ogra>
but compcache is better :)
15:11
<warren>
ok sorry
15:11
<ogra>
i still have requests from users for 24M clients :P
15:11
<sbalneav>
Sigh, looks like I'll have to completely re-do my groups here at legalaid. That's a huge PITA I didn't need.
15:12
<warren>
sbalneav: I might have a fix for you to test tonight.
15:12
<ogra>
sbalneav, isnt that a ten line script onyl ?
15:12
<warren>
sbalneav: (if you are referring to the local dev issue)
15:12* ogra wouldnt imagine that to be hard
15:12
<warren>
sbalneav: in fact, I hope you can test my patch
15:12
sbalneav: with your known broken setup
15:12
<sbalneav>
Patch to what? lbmount, or gvfs?
15:12
<ogra>
gvfs
15:13
<sbalneav>
yeah, but it's not going to make it into hardy, is it?
15:13
<ogra>
it is
15:13
<sbalneav>
ok.
15:13
<ogra>
but only if pitti agrees with it
15:13
pitti wanted to work on it for 8.04.2
15:13
<sbalneav>
ok, I'll hold off for now then.
15:14
<warren>
I talked with another GNOME developer about this
15:14
<ogra>
if warrens patch is accepted upstream pitti will agree though :)
15:14
<warren>
he says that using g_access() is wrong
15:14
W_OK would be wrong as well
15:14
the patch i have in mind uses neither
15:14
although it is a big question mark if it is possible
15:14
<ogra>
pitti made a proposal somewhere
15:14
<sbalneav>
ogra: it won't be that hard, but it a) has to be done after hours, and b) means changing a ton of our custom usercreation scripts.
15:14
<warren>
brb, I need to finish a few other things
15:14
ogra: where?
15:15
<ogra>
warren, not sure i'm digging but i think david agree to that ... looping over filesystems in /proc/mounts or some such
15:16
sbalneav, if you know that nfs mounts dont show up in /media just using access() is fine
15:16
as a quickfix
15:17
the prob is that with the new design gfvs will completely hang if you have stale nfs mounts in /media ... that wasnt possible with gnome-vfs
15:17
(imho thats a design flaw in gvfs but upstream disagrees :) )
15:17
<warren>
because gnome-vfs had HORRIBLE hacks to avoid hanging
15:17
forking off a thread to do g_access()
15:17
<ogra>
well, it worked
15:18mccann has joined #ltsp
15:19
<ogra>
andi dont see why forking off a thread is wrong *if you know* the mount is nfs
15:19gonzaloaf_laptop has joined #ltsp
15:19
<ogra>
yu could have special cases for special FSes
15:19
<warren>
what if it checked for ownership of the directory instead?
15:20
that wouldn't require an access() check
15:20
<ogra>
which directory ?
15:20
<warren>
the one that it currently looks for R_OK
15:21
<ogra>
well, these dirs are usually all owned by the user
15:21
only ltsp is different here
15:21
but gnome-mount (or whatever kde uses there) creates them $USER:$USER usually
15:22
normally gvfs would look at /media/device
15:23
<warren>
hmm
15:23elisboa has quit IRC
15:24
<warren>
ogra: 0700 + chown alone might fix this
15:24
ogra: 0700 was rejected earlier before we had the idea to chown it? I dunno.
15:24
<ogra>
heh
15:24
no because lbmount is suid root
15:24
<warren>
ogra: but R_OK check is done as the non-root user
15:24
<ogra>
so the dir is root:$USER in any case
15:24
<warren>
so that's OK
15:25
I thought we added a chown
15:25
OH wait
15:25
that was ldm
15:25
<ogra>
right
15:25
<warren>
I'll play with the things other than access() recommended by the GNOME guy...
15:36
<ogra>
warren, http://bugzilla.gnome.org/show_bug.cgi?id=526320#c20
15:36
Would it be acceptable for you to only do an access() check
15:36
if the mounted volume is a local device? (by looking at /proc/mounts, or
15:36
better, the device in hal)
15:37
that was pittis proposal
15:37
<warren>
I see
15:37
he's right that the patch doesn't help normal GNOME at all
15:37
<ogra>
and matthias seems to have been fine with it
15:37
<warren>
normal GNOME doesn't think about multi user
15:37
who will write that patch though?
15:37
<ogra>
which is massively dumb
15:38
<warren>
nod
15:38
<ogra>
gnome is a desktop for UNIx
15:38
what do they think ...
15:38
<warren>
I fight that with the entire desktop team on a regular basis
15:38
<ogra>
why do we have CK then =
15:38
etc etc
15:38
rant rant rant
15:38
:)
15:39
well, thats the patch pitti proposed to write
15:39
<warren>
ogra: actually, #20 helps the situation but not sbalneav's case
15:39
<ogra>
but as i said he is targeting 8.04.2 which isnt scheduled
15:40
it helps gnome only i guess
15:40
<warren>
it would still be good though
15:41
<ogra>
we could special case ltspfs
15:42LaserJock has left #ltsp
15:42
<ogra>
to use access there too
15:42
that would help sbalneav as well
15:42
since we know it wouldnt hang on broken fuse mounts
15:43
<warren>
keep in mind
15:43pimpministerp has joined #ltsp
15:43
<warren>
it *is* using access() there
15:43
R_OK and X_OK are satisfied because of the group
15:45
<ogra>
ah, right
15:45
but not on gvfs level
15:45
gvfs doesnt use g_access() anywhere atm
15:46
it would do so in all cases but known broken filesystems like nfs with pittis suggestion
15:46spectra has quit IRC
15:46
<ogra>
we wuld just have to whitelist ltspfs and even could drop the 0750
15:46
since it will be solved on gvfs level then
15:47warren has quit IRC
15:47* ogra arghs
15:47
<ogra>
vbox cant send ctrl-alt-backspace if hal isnt started ...
15:47
sigh
15:47
<gbolte>
lol
15:47
:/
15:48
<ogra>
so my idea of starting hal and restarting X is moot
15:48
<gbolte>
ssh in and kill X?
15:48
:D
15:48
<ogra>
nah i have a shell
15:48
but still
15:49
sillyness of the day
15:49
X without hal is totally useless nowadays
15:49
<gbolte>
true
15:49
:P
15:49
<ogra>
i.e. if you have a desktop and hal doesnt start because of any stupid error, you are stuck in X
15:50
cant even click the reboot button in gdm
15:50
or switch to console
15:50* gbolte is still trying to investigate the mystery 2 icons on the desktop for 1 device
15:50* ogra builds his 10th client
15:50
<gbolte>
10 eh
15:51
<ogra>
yeah
15:51
<gbolte>
congrats
15:51
:D
15:51
<ogra>
i need to make sure ltsp-build-client DTRT
15:51
so i cant just poke in the running client but need to verify with a full uild
15:51
*build
15:51
<gbolte>
think we have 26 or 27 clients in use right now
15:52
<ogra>
i mean my tenth image
15:52
doing my stuff in vbox ...
15:52
<gbolte>
oh
15:52
ok
15:52
<ogra>
i dotn have any thin clients in my living room and am not keen to sit the night in my lab
15:53
<gbolte>
lol
15:53* gbolte is at work
15:53
<ogra>
work==one stairs up for me
15:53
<gbolte>
:D
15:54
<ogra>
but i have my TV downstairs and still didnt get around to do my mythbuntu install to share one dvb-s card all over the house
15:54
<gbolte>
what kinda environment are you simulating in vbot
15:54
er
15:54
vbox
15:54
<ogra>
intrepid
15:54
simple ltsp server and client
15:54
<gbolte>
ah
15:54
ok
15:54
<ogra>
nothing special
15:54warren has joined #ltsp
15:55
<gbolte>
we are using edirectory+Active Directory with opensuse 11 as our terminal server
15:55
so users all use their network login on the clients
15:56
<ogra>
warren, is gdm/xorg upstream aware that yu are totally stuck if hal doesnt start by whatever error ?
15:56
<gbolte>
and we have a windows application server for the pesky have to have windows apps
15:56primeministerp has quit IRC
15:56
<warren>
ogra: they don't care
15:56
<ogra>
warren, i.e. no ctrl-alt-backspace ... no console switching
15:57
no clicking on the halt/reboot buttons in gdm
15:57
it just struck me that your only resort id the powerbutton
15:57
*is
15:57
<gbolte>
ogra, they probably state it somewhere that you have to have a working hal demon
15:57
<warren>
I don't agree with a lot of their decisions
15:57
<ogra>
imho X should fall back on something
15:57
<warren>
I hate it that X wont have any keyboard at all if you don't have hal
15:58
<ogra>
right
15:58
<warren>
it should fallback to SOMETHING
15:58
because after all, it still works without hal if you have an xorg.conf
15:58toscalix has joined #ltsp
15:58
<ogra>
it doesnt anymore after recent changes
15:58
<gbolte>
and if you mention it as a bug to them the will most likely say you are trying to use it in an unsupported setup or something
15:58
<ogra>
currently input devices are ignored
15:58
<warren>
really?
15:58
<ogra>
there is a patch pending for that
15:58
yeah
15:59
wont stay that way
15:59
but in FC and ubuntu-dev you are currently screwed, even with xorg.conf
15:59jonnor has quit IRC
16:00
<johnny>
hmm.. but what about when hal goes devicekit..
16:00
<warren>
i'm going to file a bug about this
16:00
<ogra>
http://gizmodo.com/5038298/leaked-dell-inspiron-910-mini-note-specs-and-release-date
16:00* ogra grins
16:02
<gbolte>
nice find ogra
16:03
<ogra>
http://sylvaniacomputers.com/products.php?p=meso is not bad either
16:03
and was on NBC today afaik ...
16:03
intorduced with ubuntu UNR
16:03
though i'd prefer the case and keyboard of the dell
16:07
<gbolte>
ogra, have you ever seen a local device showing 2 icons on the gnome desktop?
16:08
<warren>
oh god
16:08
now it is a 10 to 1 (me) flamewar
16:08
"why would you want to turn off hal?"
16:08
<ogra>
lol
16:08
<warren>
"you can't possibly use anything without hal"
16:08
<gbolte>
lol warren
16:08
<ogra>
warren, ask them what happens if hal fils to start and you want to debug it
16:09
*fails
16:09
<warren>
you can ask
16:09
I give up
16:09
<gbolte>
so wait what happend that hal is not running
16:09
<ogra>
did you open a bug ?
16:09
<warren>
no
16:09toscalix_ has joined #ltsp
16:10
<warren>
they're telling me that they're actually ripping out the ability to do it without hal in xorg.onf next
16:10
because it is unneeded code duplication
16:10mhterres has quit IRC
16:10
<johnny>
so that means the chroot needs hal?
16:10
<warren>
yes
16:10
<ogra>
johnny, i'm just adding it
16:10
<warren>
perhaps dberkholz knows more about this
16:10
<ogra>
should work fine now
16:11
(in ubuntu)
16:11
<warren>
you need hal anyway if you want touch screen and otehr stuff to work
16:11
<johnny>
hmm.. maybe they should have waited for devicekit before requiring such
16:11
<ogra>
right
16:11
<warren>
I just wanted the ability to turn hal off for low memory clients
16:11
<ogra>
well, turn on compcache instead
16:11
<warren>
is compcache going upstream?
16:11
<ogra>
no idea
16:12
we have it in ubuntu now
16:12etyack has quit IRC
16:12
<ogra>
i'll turn it on by default for all clients
16:12
25% compcache
16:12
lest see how that goes
16:12
<johnny>
swhat is compcache ?
16:12
<warren>
http://code.google.com/p/compcache/
16:12
<ogra>
johnny, a virtual swap device sitting in ram
16:12
like a tmpfs where you put a swapfile on
16:12
but compressed
16:13
so effectively you extedn your ram through compressing bits
16:13* gbolte likes squashing bits
16:13
<gbolte>
:D
16:13K_O-Gnom has quit IRC
16:13
<ogra>
that should make 16 or 24M clients possible again
16:13
<gbolte>
woah
16:14
<ogra>
current lowes i get working in ubuntu is 32M
16:14
without sound and localdev though and using xdmcp instead of ldm
16:14
but works
16:16
<warren>
hmm
16:17
<ogra>
hmm ?
16:17
<warren>
haldaemon + messagebus use ~3MB more physical memory
16:17
that's less than I expected.
16:17
<ogra>
not that much
16:17
and i bet you can cut that down if you disable a bunch of hal addons
16:17
<warren>
yeah
16:17
I don't
16:19toscalix has quit IRC
16:24toscalix_ has quit IRC
16:40bobby_C has quit IRC
16:41
<warren>
anyone pushing anything to ltsp-trunk soon?
16:41
I have to tag later today
16:41
<ogra>
i will
16:41
<warren>
k
16:41
<ogra>
just got my first hal-input session working
16:41
and dont even need any XKB settings anymore
16:41
german out of the box
16:42
<warren>
i'm keeping the XKB stuff just in case the user wants to override
16:43
<ogra>
well, i'll change it a bit for ubuntu to change /etc/default/console-setup i guess
16:43
thats where hal reads the kbd settings from
16:45
<warren>
I need to know how it works on F10
16:45
I don't get it at all
16:45
I want to do the mini-configure-x.sh and LDM_SESSION proposal soon
16:45
the latter would require autoconf knowledge that I don't know
16:45
oh wait
16:45
no it wont
16:46
it can fallback to detecting the location
16:46
which sucks, but it only falls back if you have an old ldminfod
16:46
<ltsppbot>
"ogra" pasted "ubuntu/debian hal-input kbd script borrowed from fedora" (21 lines) at http://pastebot.ltsp.org/52
16:46
<ogra>
find the file :)
16:46
<warren>
ogra: what is /etc/default/console-setup ?
16:46
<ogra>
the script is 1:1 from fedora
16:47
the defaults for console keyboards, fonts and encodings in debian/ubuntu
16:47
<warren>
what is it called in fedora?
16:47
<ogra>
no idea
16:47
but the script is the same apart from the file
16:47
so look in your FC 10
16:47
in /usr/lib/hal/
16:47
<warren>
hm
16:47
<ogra>
there you should have the same script reading a different file
16:48
grep XKBLAYOUT /usr/lib/hal/*
16:48
or so
16:55
warren, ok, feel free to tag, i pusjed all changes i need for now i think
16:55
<warren>
I might add more fixes before I tag tonight.
16:56
<ogra>
hmm, then i need to tag before you ... i need to roll a package
16:56
<warren>
oh ok
16:56
go ahead and tag, I'll roll a package from .22 as well
16:56
commuting now
16:56mccann has quit IRC
16:56* ogra doesnt even know what version we're at atm
16:57* ogra checks
16:57
<ogra>
ah
16:57
ltsp-5.1.21
16:58
gah, i always forget i need a change if i want to tag
17:00* ogra doesnt use tags anywhere else
17:00warren has quit IRC
17:05
<johnny>
dberkholz, got a chance to look at my stuff yet?
17:05
we're getting more interested users lately :)
17:13
<ogra>
vagrantc, did you include my cdrom naming patch in debian in ltspfs ?
17:14
<vagrantc>
ogra: i included a better variation upstream
17:14
<ogra>
ah, so cdrom devices are called cdrom now ?
17:14
<vagrantc>
ogra: well, hopefully better.
17:14
<ogra>
or cdromX
17:14
<vagrantc>
ogra: cdrom or cdrom[0-N]
17:14
<ogra>
great
17:14
<vagrantc>
depending on what they're called
17:14* ogra requests a sync then
17:14
<vagrantc>
should handle multiple cdroms
17:15
i'm hoping you can sync without any differences ... although the whole ldm versioning debacle is still there
17:15
but is that enough releases ago that you shouldn't need to worry?
17:16
latest upload to unstable is lintian clean :)
17:16
<ogra>
ldm versioning ?
17:16
i switched to your versioning in intrepd
17:17
<vagrantc>
you had maintained conflicts with older ldm versions ...
17:17
<ogra>
i just added the themes
17:17
in hardy
17:17
intrepid uses the epoch
17:17
<vagrantc>
right.
17:17
<ogra>
ogra@osiris:~/Devel/ltsp$ apt-cache show ldm|grep Version
17:17
Version: 2:2.0.8-0ubuntu1
17:17
;)
17:17
<vagrantc>
ogra: i made significant changes to the udev rules ... you'll want to test those out.
17:18
<ogra>
i will
17:18
for now i need to get the syncs done
17:18
2h until freeze
17:18
<vagrantc>
as always, let me know if you have any questions :)
17:20
<ogra>
vagrantc, oh, what happened to the local HDD support ?
17:21
<vagrantc>
ogra: dropped that ages ago
17:21
<ogra>
well, the current rules dont look after "removable"
17:21
<vagrantc>
the rules in 0.5.3-[12] do look at removable ...
17:21
and numerous other factors ...
17:22
<ogra>
ok
17:22
<vagrantc>
my only fear with them is they're a bit more complicated, and i fear something's missing ...
17:22
<ogra>
we'll see
17:22
<vagrantc>
gadi also has some additional changes ...
17:23
<ogra>
i now have two months for bugfixes :)
17:23
sync requested
17:23
vagrantc, ltsp-client-core needs to start at 25 for hal-input btw
17:23
<vagrantc>
the weirdest one was that external USB hard disks often seemed to be marked as non-removable...
17:24
it's ... 32 i think right now?
17:24
<ogra>
30 here
17:24
err
17:24
20
17:24
i just moved it
17:24
for ubuntu
17:24
hal starts at 24
17:25juanpaul has joined #ltsp
17:25* vagrantc hopes for no major breakages
17:25
<ogra>
the opposite :)
17:25
hal-input solves tns and tons of issues
17:25
i.e. it reads directly from /etc/default/console-setup
17:26
no weird workarounds for XKB options needed anymore
17:26
<vagrantc>
i think i ran lessdisks's init script at 19
17:26
sounds nice ...
17:26
<ogra>
it is !
17:26
<vagrantc>
ogra: increase in ram costs?
17:26
<ogra>
i'm working on other scripts for touchscreens etc
17:26
3M
17:27
hal and dbus together
17:27
<vagrantc>
that's not evil ... but every bit hurts a little.
17:27
<ogra>
see my one of last commits
17:27
it needs hal and dbus in the deps an RC2 whitelist addition for both
17:28
ihad to drop the recommends processing fro apt-get though
17:28
<vagrantc>
i'll probably be sticking with the tried and true for lenny...
17:28
<ogra>
somehow our linux-image package recommends lilo
17:28
<vagrantc>
but i've been doing a few experimental uploads, just to not fall completely behind
17:28
heh.
17:28
that's absurd.
17:28
<ogra>
it is
17:29
<vagrantc>
lilo's like... so ... late 90s.
17:29
<ogra>
well, we relied on not instaling recommends by default in the past
17:29
and some weird raid setups on XFS actually need lilo
17:29
<vagrantc>
let em suffer.
17:29
<ogra>
heh
17:29
mdz is one of them :P
17:29
<vagrantc>
speaking of suffering, i finally marked the ldm/tcsh bug as wontfix :)
17:30
<ogra>
yay
17:30
good move
17:30
<vagrantc>
of course, attempting to be diplomatic about it ...
17:30
<ogra>
i'm pretty sure there's more breakage with tcsh
17:30
if /bin/sh points to it at least :)
17:31
<vagrantc>
that's basically what i started finding ... i'd get patches that got it working, but various features or combinations of features would break ...
17:31
i.e. LDM_DIRECTX + LOCALDEV ...
17:31
<ogra>
yeah
17:31
<vagrantc>
well, it was never an issue if /bin/sh pointed to it ... it was if the user had it set as their login shell.
17:31
<ogra>
well
17:32
admins setting all users to tcsh might as well point /bin/sh to it ... you never know ;)
17:32* vagrantc waits for the thousands of tcsh users to complain and fork ldm ...
17:32
<ogra>
haha
17:33
<vagrantc>
i gave it a try, because i'm that kind of person.
17:33* gbolte has always used bash
17:33
<ogra>
i wouldnt expect anything to work
17:33
<gbolte>
I would be so lost on a different shell
17:34johnny has left #ltsp
17:35
<ogra>
dh_installinit -pltsp-client-core --name=ltsp-client-setup --no-start -u"start 32 S ."
17:36* ogra scratches his head
17:36
<ogra>
why the heck did it end up as S20 ?
17:37
oh
17:37
err
17:37
vagrantc, ?? do you have a line for ltsp-client-core in your packaging ?
17:37
thats -setup
17:38
i dont see either of us having a line for ltsp-client-core at all
17:39
<moquist>
ogra: should I bother anymore right now with the moodle package, or just target the next release?
17:40
<ogra>
moquist, lets try it as exception if it gets ready in time before release
17:40
vagrantc, all i see in your packaging branch is another
17:40
dh_installinit -a --no-start
17:40
which just installs with defaults
17:41* moquist replans his evening
17:42
<ogra>
moquist, feature freeze is in 1.5h
17:42
i doubt you can make that
17:42
but i'll bribe slangasek if necessary ...
17:43
<moquist>
NFW, as they say. :)
17:46* ogra pbuilds ltsp
17:48
<vagrantc>
ogra: defaults to 20
17:48X0d_of_N0d has quit IRC
17:48
<ogra>
ah
17:49X0d_of_N0d has joined #ltsp
17:53
<ogra>
urgh
17:53
<vagrantc>
ogra: so either dh_installinit -a --no-start is what's installing it ... or dh_installinit -i --no-start
17:53
ogra: though you might want to look at ltsp-experimental-packaging ... as the ltsp-debian-packaging is still targeted at lenny
17:54
<ogra>
vagrantc, what do you do with client/xrexecd/ldm-rc.d ?
17:54
and why the heck is that in the ltsp tree ?
17:54* ogra grumbles
17:54
<ogra>
that then needs a replaces ldm in debian/control
17:54
sigh
17:54
<vagrantc>
?
17:55
<ogra>
they both will provide the same dir, no ?
17:55
<vagrantc>
ah, a conflicts.
17:55
<ogra>
yeah :(
17:55
who made that silly decision
17:55
<vagrantc>
me.
17:55
<ogra>
that belongs into ldm
17:55
<vagrantc>
absolutely not.
17:55
<ogra>
well, you prefer to have a conflicts/replaces in ltsp ?
17:56
<vagrantc>
it's a plugin *to* ldm, nto a plugin as *part* of ldm.
17:56
<ogra>
it relies on the ldm rc.d dir
17:56
<vagrantc>
i don't need a conflicts in ltsp.
17:56
debian never released with it
17:56
well, the only release that shipped it was in experimental...
17:56
<ogra>
you have two packages claiming the same dir
17:56
<vagrantc>
that's not an issue
17:56
it's if two packages have the same file.
17:56
<ogra>
hmm
17:57
oh, yeah
17:57
late here, sorry
17:57
<vagrantc>
but if you shipped ldm with the localapps plugin, probably should still have a conflicts ...
17:58
my logic in moving it was it depended on features currently in the ltsp packages, was harmless without ldm installed, and if ldm was installed without ltsp, it caused problems.
17:59
i had expressed that opinion at the hackfest when folks asked...
17:59
but folks were eagre to get it committed somewhere...
18:01
<sutula>
vagrantc: Yesterday evening I did a fresh client rebuild with Lenny. Still seeing the password problem (wrong password locks up greeter) which can be worked around by setting ssh options. Is this expected? It would seem to be fairly important for released Lenny.
18:01
<vagrantc>
sutula: hmmm... haven't always seen that
18:01
sutula: if you can reproduce it, please file a bug
18:02
<sutula>
vagrantc: OK...fair 'nuff
18:02
vagrantc: If you're not seeing it, I wonder if I have a combination of packages/cruft on the server from having updated from Etch to testing
18:03
I'll have an opportunity to do a fresh install next week on a diff server, so will check then
18:04
<vagrantc>
sutula: that would be great. i'll try and test myself, also.
18:04
<sutula>
vagrantc: One other difference, I'm using LDM_DIRECTX=True with having older, slower thin clients and a private network
18:04
<vagrantc>
ogra: hrm. i was going to point you to my ltsp-experimental-packaging branch, but there seem to be issues with bzr.debian.org at the moment...
18:05
sutula: i've barely tested with anything but virtualbox on an 866MHz-1.2GHz machine ...
18:05
<ogra>
well, i can change packaging later :)
18:06
<sutula>
I'll look into it more...had assumed it was a known issue...later :)
18:07johnny2 has joined #ltsp
18:08johnny2 has left #ltsp
18:08johnny has joined #ltsp
18:22warren has joined #ltsp
18:25
<ogra>
hmm, whats xatomwait ? do we use it for anything yet ?
18:25
sbalneav, ^^^ ?
18:26* ogra has 35min to get that sorted
18:26
<vagrantc>
ogra: split out the xatoms handling into a c script, and gadi's shell script uses that instead of sleeping in a while loop[
18:26
er, c binary
18:26
<ogra>
i dont see it being called anywhere
18:27
<vagrantc>
xrexecd.sh
18:27
<ogra>
oh, right, there it is
18:27
<vagrantc>
so rather than writing all the logic in C, we can actually keep some of it as shell code ...
18:29* ogra would prefer all of it in shell
18:29
<ogra>
but well
18:29
i'll take whats there
18:29
to at least have the infrastructure
18:31
<vagrantc>
scotty came up with something else that could use xatomwait, too.
18:31
i forget what it was exactly ...
18:31
something related to something i was posting about lately.
18:32
ah yes, he proposed to use it to set an X property to trigger the "kill the ssh connection" stuff.
18:32gonzaloaf_laptop has quit IRC
18:32
<ogra>
RAAAH
18:32* ogra goes mad about single char typos holding him up
18:33
<ogra>
i typoed xrexecd.sh the fifth time now ...
18:33
way to many x'es
18:34
<vagrantc>
i also proposed to rename all that stuff to localapp* ...
18:34
since that's what the feature is called everywhere else ...
18:34
<ogra>
++
18:34
well
18:34
<vagrantc>
(except for LOCALDEV)
18:34
but at least it's *more* similar :)
18:34
<ogra>
i got a building package
18:34
finally
18:34
not sure anything in that thing works at all
18:35
<vagrantc>
my last build didn't build autogen.sh stuff at all.
18:35
that was when xrexec.c had some issues ... i think we'll need to build-dep on auto*-dev
18:35
just for one C program
18:36
<ogra>
ogra@osiris:~/Devel/ltsp/intrepid-packaging/ltsp-5.1.22$ dpkg -c /var/cache/pbuilder/result/ltsp-client-core_5.1.22-0ubuntu1_i386.deb|grep bin/x
18:36
-rwxr-xr-x root/root 5408 2008-08-28 01:34 ./usr/bin/xatomwait
18:36
-rwxr-xr-x root/root 1079 2008-08-27 16:45 ./usr/bin/xrexecd.sh
18:36
all there :)
18:36
<vagrantc>
yay
18:37* ogra wonders if he wants xrexec.sh in /<usr/share/doc/examples
18:37
<vagrantc>
yeah.
18:38
i actually installed it in /usr/bin ... but i know you didn't like that at all :)
18:38japerry has joined #ltsp
18:39
<ogra>
i dont even like it in examples
18:39
but i guess people will complain if i dont do it
18:39
so i will
18:40
it shoudlnt be server side at all
18:42mighty-d has joined #ltsp
18:42
<mighty-d>
Hi
18:43
i have a bunch of old computers with 32 mb ram and 100 mhz, but the new ssh and ldm seems too heavy for this clients on 5.0, is there a way to turn xdmcp on in hardy heiron ltsp 5.0 ?
18:45
<ogra>
yes
18:46
create an lts.conf file in /var/lib/tftpboot/ltsp/i386/
18:46
add:
18:46
[default]
18:46
SCREEN_07=startx
18:46
also with 32M add:
18:46
SOUND=False
18:47
LOCALDEV=False
18:47
that ill do it
18:47
*will
18:48
<mighty-d>
ogra, do you suggest the xdmcp will be enought for this?, or should i downgrade squashfs to nfs as well?
18:49
<ogra>
squashfs should be fine
18:49
you might want to raise ndbswapd's value on the server to more than 32M
18:50
<mighty-d>
ogra, should i change everything to xdmcp?
18:50
<ogra>
(or wait for intrepid, which comes with compcache and is way more friendly to your client ram)
18:50
<mighty-d>
intrepid?
18:50
<ogra>
8.10
18:50
the october release
18:52
<mighty-d>
hmmm.. i see
18:52
what do you think of 16MB thin clients?
18:53
<ogra>
wont boot
18:53
the kernel alone needs more to uncompress nowadays
18:53
<mighty-d>
lol...
18:54
<ogra>
even the 32M only work through a trick
18:54
(we use nbd swapspace for everything below 48M by default)
18:54
in 8.10 24M *might* work ... depending on the amount the kernel needs to uncompress
18:55
(without swap)
18:55
<mighty-d>
what about: booting from hard disk on the client with a low kernel and light OS, and then using xdmcp to get into the server, (i know i wont get cool features, but do you think it could work?)
18:55
and using a swap hard disk space
18:55
<ogra>
sure, but its a lot of maintenance work
18:55
what about adding some ram ;)
18:55
<mighty-d>
ogra, there isnt ram availible for those
18:55
<ogra>
64M would do miracles
18:56
not on ebay ?
18:56
<mighty-d>
i dont even know how it was named, is the tiny socket
18:56
<ogra>
SIM probably
18:56
<mighty-d>
yeah, now i remember
18:56
the issue is i have like 300
18:56
well not i
18:57* vagrantc throws away pounds upon pounds of SIMMs weekly.
18:57
<ogra>
mighty-d, there you got someone :)
18:57staffencasa has quit IRC
18:57
<vagrantc>
mighty-d: what's the project?
18:57
<mighty-d>
the orphans house that is requesting me to get those old machines to live
18:58
<vagrantc>
mighty-d: http://freegeek.org/grants/
18:58
<mighty-d>
ok, there is a low resource orphans house, with 300 or so clients
18:58
they had an school or something
18:58
<vagrantc>
oops.
18:58
mighty-d: trailing / matters ...
18:58
mighty-d: http://freegeek.org/grants
19:00
<mighty-d>
vagrantc, thing is they want to know if they should trash those old computers away or make anything of it
19:01
<vagrantc>
mighty-d: hard to say... 100MHz is getting pretty slow these days
19:02
<mighty-d>
yah, i know
19:02
but they will be happy even if it works just to process some text on openoffice or abiword or whatever
19:02
<warren>
vagrantc: perhaps we need a pop-up message saying "you have FOO amount of RAM"
19:02
<vagrantc>
ogra: looks like the tag for ltsp-5.1.22 is off by one revision
19:02
<ogra>
vagrantc, ?
19:02
i didnt push after i tagged
19:03
or commit
19:03
<vagrantc>
ogra: the revision you tagged is actually 850, in which release.conf is 5.1.21 still
19:03
<ogra>
hmm ?
19:03* ogra checks
19:03
<vagrantc>
bzr log -r tag:ltsp-5.1.22
19:03
850 Oliver Grawert 2008-08-27
19:03
configure-x.sh, fix quoting of EXTRAMOUSE settings
19:03
<ogra>
ogra@osiris:~/Devel/ltsp/ltsp-trunk$ bzr push
19:03
Using saved location: bzr+ssh://ogra@bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/
19:03
No new revisions to push.
19:04
i tagged and pushed the release.conf change then
19:04
<vagrantc>
somehow, you managed to tag the previous version...
19:04
*somebody* tagged the previous version.
19:05
<ogra>
well likely me
19:05
weird
19:05
<vagrantc>
maybe you tagged before you committed?
19:05
<ogra>
i definately tagged and committed the release.conf change
19:05
no
19:06
i'm sure i didnt ...
19:06
<vagrantc>
well... hmmm... wonder if we can fix it.
19:07
<ogra>
i commited the extramouse stuff, then tagged, then cursed because i cant commit a tag, then remembered the release.conf, changed, commited and pushed
19:07
<vagrantc>
ah, yes.
19:07
you should've committed the release.conf first, and then tagged that, and then pushed
19:07
<ogra>
unless the tag works recursive if i commit that should have been all in the right order
19:07
silly tag business
19:07
<vagrantc>
heh
19:07
<ogra>
can someone remnd me why we use that ?
19:08
i work on about 50 different upstream branches ...
19:08
there is none among tem using tags
19:08otavio has quit IRC
19:08
<vagrantc>
so i can do: bzr di -r tag:ltsp-5.1.17..tag:ltsp-5.1.21 ...
19:09
<ogra>
hmm
19:09
<vagrantc>
rather than looking trhough the entire commit logs to find the commit message saying which version was tagged what...
19:09
and then diffing the versions...
19:10
<ogra>
bzr log|grep release.conf
19:10
or some such
19:10
<vagrantc>
bzr log | grep release.conf tagging ltsp-5.1.22, bumping release.conf
19:10
that gets precisely one revision :P
19:10
and takes longer ... anyways ...
19:11
<ogra>
bzr log --line|grep release rather
19:11
bah
19:11
well
19:11
<vagrantc>
that catches 4 revisions ...
19:11
<ogra>
so we have tags ...
19:11
yeah, i see
19:11
<vagrantc>
and not all of them are versions
19:16
ogra: good luck with all your syncing and merging and whatnot.
19:16
off to aikido!
19:16vagrantc has quit IRC
19:18* ogra is done :)
19:22fie_wr0k has quit IRC
19:35jammcq has joined #ltsp
19:35
<jammcq>
hello friends
19:35
<ogra>
hey jammcq
19:35
<gbolte>
hi jammcq
19:38FuriousGeorge has quit IRC
19:39
<jammcq>
hey, how's it goin?
19:39
ogra: you should be sleeping now
19:39FuriousGeorge has joined #ltsp
19:40
<FuriousGeorge>
johnny: i just turned my computer off with 3 packages to go... that measns i have to start all over or can i restart the script?
19:42ace_suares has quit IRC
19:42
<ogra>
jammcq, i will soon
19:42
feature freeze started 40 min ago
19:42
i managed to do my last ltsp upload 9 mins before :)
19:42
<jammcq>
ah, good job
19:43
<ogra>
well, i dont know if anything of that works atm :)
19:43gbolte has quit IRC
19:43
<ogra>
but the packages build and install everything in the right place
19:43
now two months of bugfixing and fine tuning are ahead
19:43japerry has quit IRC
19:44
<ogra>
but hal-input is sweetness
19:44
we will be able to drop 90% of the xorg hacks and scripts and just modify hal parameters on the fly ...
19:46
<jammcq>
well, I hope Intrepid turns out to be worthy of praise. i'm still not feeling the love from Hardy like I hoped for.
19:47
<ogra>
in ltsp or in general ?
19:49psycodad1 has joined #ltsp
19:49
<jammcq>
in general
19:49
well, on my laptop, which worked perfectly under gutsy
19:49
but... I do have it installed on a couple servers, where it's working fine
19:49
but there's something about my lappy that just isn't happy
19:50ace_suares has joined #ltsp
19:50psycodad has quit IRC
19:50
<ogra>
changing the whole input system in intrpid is likely to have regressions
19:50
especially with upstream bing dense
19:50
like there is no plan at all if hal fails on boot ...
19:52
you are simply left without *any* input
19:52
no console switching etc
19:52
power button is your only resort then
19:52
and according to warren upstream X will even drop the code that you culd use for fallback in such cases to avoid duplication
19:53
<jammcq>
heh
20:00fie_wr0k has joined #ltsp
20:05X0d_of_N0d has quit IRC
20:05
<johnny>
FuriousGeorge, uhmm.. this time..
20:05X0d_of_N0d has joined #ltsp
20:05
<johnny>
the next verson will build reusable packages
20:12LaserJock has joined #ltsp
20:13japerry has joined #ltsp
20:34
<mighty-d>
ogra when i change Screen_07=startx it starts asking me for nfs
20:52lucascoala has joined #ltsp
20:59johnny has left #ltsp
21:00
<warren>
ogra: I'm going to fix the tag
21:00
you might need to rebase
21:01
shit
21:01
wont let me fix the tag
21:01
ogra: please don't do this again.
21:16Egyptian[Home1 has quit IRC
21:17Egyptian[Home] has joined #ltsp
21:19lucascoala has left #ltsp
21:19lucascoala has joined #ltsp
21:20mighty-d has quit IRC
21:38johnny2 has joined #ltsp
21:40johnny2 has left #ltsp
21:40johnny has joined #ltsp
21:41
<warren>
ogra: please use "mkdst tag" next time. it wont let you tag in error.
21:46
<FuriousGeorge>
johnny: im up to * Installing extra packages....
21:46
i didnt do verbose this time
21:47
there isnt any way to know what ebuild its up to right?
22:13LaserJock has quit IRC
22:31
<ace_suares>
wanna hear something 'funny' ?
22:33
well the school that I was supposed to install ltsp was hit by a twister - very rare here, last one was in 1996 - and the roof flew off and all the kids have to go to an emergency school about 12 km further away
22:33
<jammcq>
wow
22:33
<ace_suares>
and thieves cut all the cat5 cable with a knife a week earlier cause they taught it was wires for the alarm
22:34
now how's that for 'funny'. Project deadline extended indefinately.
22:53mccann has joined #ltsp
23:12martin3z has joined #ltsp
23:16
<johnny>
dberkholz, you around?
23:18
ace_suares, :(
23:47martin3z has quit IRC
23:54mccann has quit IRC