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


Channel log from 17 October 2008   (all times are UTC)

00:20Hyperbyte has quit IRC
00:30alkisg has joined #ltsp
00:43ltspbot has joined #ltsp
00:57alkisg has left #ltsp
01:04basanta has joined #ltsp
01:07warren has joined #ltsp
01:07
<warren>
Somebody broke ldm's setxkbmap stuff
01:10basanta has quit IRC
01:14feinbein has joined #ltsp
01:16feinbein has quit IRC
01:18
<warren>
ogra: huh?
01:18
ogra: your latest patch for ldm comment makes no sense
01:18
ogra: nobody added a patch for a non-standard theme
01:19
ogra: that patch didn't even go upstream yet
01:56indradg has quit IRC
02:22X0d_of_N0d_ has quit IRC
02:22X0d_of_N3d_ has quit IRC
02:23rcy`` is now known as rcy`
02:27X0d_of_N0d has joined #ltsp
02:30siki has joined #ltsp
02:42Hyperbyte has joined #ltsp
02:45warren has quit IRC
02:57warren has joined #ltsp
03:12warren has quit IRC
03:38mikkel has joined #ltsp
03:43Q-FUNK has joined #ltsp
03:49coolio has joined #ltsp
04:02tjikkun_work has joined #ltsp
04:05cyberorg has joined #ltsp
04:29bobby_C has joined #ltsp
04:36ninou has joined #ltsp
04:36
<ninou>
hi
04:37
dhcpport kernel argument for ubuntu/hardy do not work
04:38
how can i rebuild initrd image for ltsp clients ?
05:18zoredache_ has joined #ltsp
05:19zoredache has quit IRC
05:27coolio has left #ltsp
06:04nubae1 is now known as Nubae
06:15alkisg has joined #ltsp
06:15pmatulis has joined #ltsp
06:34Nubae has quit IRC
06:35bobby_C has quit IRC
06:44nubae has joined #ltsp
06:48alkisg has left #ltsp
07:23cliebow has joined #ltsp
07:37mikkel has quit IRC
07:46likuidkewl has joined #ltsp
07:46likuidkewl has left #ltsp
07:47dmaran has joined #ltsp
07:49toscalix has joined #ltsp
07:56indradg_ has joined #ltsp
07:57six2one has joined #ltsp
08:15
<_UsUrPeR_>
good morning all
08:15
<nubae>
morning _UsUrPeR_
08:16* _UsUrPeR_ yawns and sips a monster
08:16vlt has joined #ltsp
08:16
<nubae>
I managed to get sugar collaboration working properly... ltsp and sugar is a go...
08:16
<_UsUrPeR_>
oh spectacular
08:17
<nubae>
yeah I'm doing a write up and I'll post to the list
08:17
<_UsUrPeR_>
are you going to be setting that up as a kiosk?
08:17
<vlt>
Hello. Does anyone know which type of Etherboot/PXE image I need from rom-o-matic.net to put it directly on a hd to boot PXE from?
08:18
"Binary ROM image .zrom"?
08:25
<johnny>
i used the zdsk and booted off a floppy
08:26Gadi has joined #ltsp
08:27mccann has joined #ltsp
08:32alkisg has joined #ltsp
08:41Guaraldo has joined #ltsp
08:42
<pscheie>
is it accurate to say that it is easier to build multiple chroots, because of differing TC needs, with ltsp 5?
08:42
<ogra>
pscheie, depends
08:43
<pscheie>
or at least provide different kernels since some older TCs dislike newer kernels?
08:43
<ogra>
pscheie, most stuff where your TCs use the same CPU arch and bootmetod should be catchable by lts.conf
08:44
<pscheie>
ah, right, just specify a different kernel in lts.conf
08:44
<ogra>
if you use different kernels yes, that might make sense ... though only because we dont have an easy method to have the modules from different kernel trees in one chroot
08:44
<pscheie>
correct?
08:44
<ogra>
then you would only need to change dhcpd.conf
08:45
no, kernel is dhcpd.conf/tftp ...
08:45
lts.conf only comes into play after the kernel loaded
08:45
<pscheie>
but if a different kernel, same arch, is being used, would one likely need a different chroot?
08:46
or run the risk of things being broken by use of older kernel?
08:46
<ogra>
yes, as i said because we dont have an easy method yet to have /lib/modules available for multiple different kernels
08:46
that is possible but would require manual fiddling
08:47
would be a cool addition though
08:47
since in the end you dont really need different chroots but your kernel needs to find its modules
08:47
<vlt>
johnny: It works with GRUB pointing to a .zlilo image. But I wanted to omit the whole GRUB part ...
08:47
<pscheie>
yeah, perhaps a topic for discussion at BTS
08:47
<ogra>
the apps inside the chroots, the setup etc, everything will be the same
08:48
only /lib/moduleswould differ
08:48
<pscheie>
Is that something that would have to be indicated in dhcpd?
08:48
<ogra>
well, in ubuntu our difference is in the initramfs anyway unless you have non i386 clients
08:48
<pscheie>
or lts.conf
08:48
ah, ok
08:48
<ogra>
thats very distro specific to solve
08:50
fedora would have a different approach here, i would supply a different initramfs.conf file, and be done with it, since the ubuntu core kernel is as stripped as possible
08:50
so everything is a module here anyway
08:52
<pscheie>
ok, tx
08:53
is anyone using flash 10 yet?
08:53
more to the point, any comments about getting sound working on TCs with it?
08:54* ogra will soon
08:54
<ogra>
intrepid had it uploaded yesterday, i will start testing it the next week
08:54
but the alsa modules shold just cope
08:54
err
08:54
asound
08:54
sorry
08:55* pscheie wishes he could recall all of Gadi's very informative lecture on the details of LTSP sound
08:56
<ogra>
TC HW -> alsa modules ... TC mixing -> pulse ... session runs grstreamer and forwards to pulse on the TC
08:57
the forwarding should happen through libasound
08:57
and its pulse module
08:57Guaraldo has quit IRC
08:58
<ogra>
and essentially libasound just uses the PULSE_SERVER env variable which we set in ldm
08:58* ogra implemented that upstream ages ago ... needs an urgent overhaul though
09:07
<alkisg>
With the latest intrepid my TCs work OK with no xorg configuration, but at a low monitor refresh rate. But if I do CONFIGURE_X=True, no matter what else I try (e.g. X_VERTREFRESH=72.0), I get a blank screen. So I want to produce a xorg.conf with the (automatic) working configuration, change the refresh rate and provide it to lts.conf. But the new xorg doesn't create a xorg, so I can't take it!!! Any ideas?
09:08
*doesn't create a .conf
09:09
<Gadi>
good morning, all
09:09
alkisg: are you sure it doesn't create a xorg.conf? perhaps the driver is overdriving the monitor
09:09
or not using a monitor-supported mode
09:10
<alkisg>
Gadi, good morning. I think this is the problem, but I couldn't find an /etc/X11/xorg.conf
09:10nicoAMG has joined #ltsp
09:10
<Gadi>
did you log into a shell on the client?
09:10
<alkisg>
Yes, with ctrl+alt+f1 => root etc
09:10
<Gadi>
crl-alt-f1?
09:10
ah, ok
09:10
in the shell, run: X :10 -configure
09:11
<alkisg>
Oh, it's :10 now???!!! I was wondering about that!!!
09:11
<Gadi>
its arbitrary
09:11
<alkisg>
I'm not near the clients, though. This will show me the configuration options?
09:11
<Gadi>
this will just produce a xorg.conf and a log in Xorg.10.log
09:11
or it will error out
09:11
<alkisg>
I was trying "DISPLAY=:7 xterm" which worked at previous versions, but no joy
09:11
<Gadi>
that only works if the Xserver is running
09:12
<alkisg>
Gadi, thanks, I'll try it on Monday when I'll be at the lab
09:12
<Gadi>
but, I am not sure if your Xserver is up and in the wrong mode for the monitor
09:12
or if it failed entirely
09:12
<alkisg>
The xserver is running afaik, I think it's overdriving the monitor like you said
09:12
<Gadi>
ah, in that case
09:12
belay the first response ;)
09:12
just take a look at /var/log/Xor.6.log
09:13
<alkisg>
(well, I'm not entirely sure, the screens are new, and I don't see a "mode not supported" dialog'
09:13
<Gadi>
*Xorg.6.log
09:13
or Xorg.7.log
09:13
whatever it is
09:13
<alkisg>
OK, will try that! Thanks!
09:13
<Gadi>
even better, scp it to the server
09:13
so you can pastebot it here for help
09:13
the Xorg log tells all
09:13
<alkisg>
The problem is that when I'm at the lab, noones here!!! :)
09:13
<Gadi>
hehe
09:14
well, scp it to wherever you are now
09:14
;)
09:14
or just read through it
09:14
<alkisg>
Yeah, sure!
09:14
<Gadi>
it will tell you what mode it is using
09:14
you may have to choose a better one with X_MODE_0
09:14
<alkisg>
What if it fails entirely?
09:14
<Gadi>
if it fails entirely, then X -configure would fail
09:14
but produce a log
09:15
<alkisg>
(I did try with X_MODE_0 and X_VERTREFRESH, but all combinations I've tried just blanked the screen)
09:15
<Gadi>
so, scp whatever Xorg log you find
09:15
X_VERTREFRESH needs to be accompanied by X_HORZSYNC
09:15
<alkisg>
OK, I see.
09:15
<Gadi>
to work properly
09:15
so, try setting both, like:
09:15
X_VERTREFRESH = "55-75"
09:15
X_HORZSYNC = "30-100"
09:16
<alkisg>
Then maybe that's why it failed, I thought I had tried it last year and succeded, but I may not remember well
09:16
<Gadi>
or some such
09:16
and ranges are better than distinct values
09:16
<alkisg>
Even X_MODE_0=1024x768 failed though
09:16
(only this one, with no horizontal / vertical timings)
09:17
<Gadi>
then, it either is at a refresh rate that is too high for the monitor, or it does not recognize the mode or the entire driver fails
09:17
I would lean to the former
09:17
<alkisg>
OK, I'll scp the log and look at it, and if I can't find anything I'll come for help! :)
09:17
<ogra>
alkisg, just setting CONFIGURE_X=True without any other options should create you an xorg.conf
09:17
to fiddle with then
09:17ninou has quit IRC
09:17
<Gadi>
right
09:17
<alkisg>
ogra, I'm pretty sure I've tried that and saw no xorg.conf in /etc/X11
09:17
<Gadi>
and if not, then the driver is borked
09:17
<ogra>
right
09:17
<Gadi>
and X -configure is failing
09:18
<ogra>
well, nsc is borked
09:18
so it depends if the chroot was created before nsc was dropped from xserver-xorg-video-all :)
09:18
<alkisg>
So, I should also specify a driver? But I can't think why autodetection would choose the right one, and CONFIGURE_X=True would fail...
09:18
<Gadi>
ooh! you dropped it! YAY!
09:18
:)
09:18
<ogra>
well
09:19
i voted against it :)
09:19* alkisg is googling for nsc...
09:19
<ogra>
but there was no fix to make Xorg -configure survive with it
09:19
<Gadi>
alkisg: nsc is an old driver for geode chipsets
09:19
<ogra>
alkisg, dpkg -l xserver-xorg-video-nsc .... in your chroot
09:19
<Gadi>
that conflicts with the newer geode driver
09:19
<alkisg>
Ah, ok, I got S3 virge
09:19
<Gadi>
that savage
09:19
ur find
09:19
*fine
09:20
<ogra>
if xserver-xorg-video-nsc is installed Xorg -configure cant work
09:20
<Gadi>
well, of course, then there's the mess of via drivers
09:20
;)
09:20
<alkisg>
ogra, no such package
09:20
<ogra>
good
09:20
<Gadi>
good
09:20
<ogra>
do the same for -psb
09:20
<alkisg>
I'm using the latest ltsp-server package
09:20
<ogra>
it has a similar prob and was dropped one day later
09:21
doesnt depend on ltsp-server ... only on xorg
09:21
<alkisg>
None for xserver-xorg-video-psb
09:21
<ogra>
good as well
09:21
then it should crete an xorg.conf
09:21
<alkisg>
It's possible that when I tried this I had an older xorg
09:22
I'll try again on Monday, thanks!
09:22
ogra, no other problems, everything's fine!
09:22
<ogra>
it should overwrite the older xorg.conf
09:22
cool !
09:22
<Gadi>
is it me or is ltsp-discuss becoming like slashdot
09:22
;)
09:22
<ogra>
did they mention edubuntu ?
09:23
or why do you think that ?
09:23
<Gadi>
i dunno - i guess i don't read it often enough, but in threaded view, the subjects go off the screen ;)
09:23
<ogra>
hehe, yeah
09:23
its all that server viagra talk atm
09:24
hardening this, hardening that ... blah
09:24
<Gadi>
hehe - haven't gotten to that thread yet
09:24cyberorg has quit IRC
09:24
<Gadi>
lol
09:25* ogra wonders what to do with the guest button in intrepid
09:25
<Gadi>
invite guests to use it?
09:25
<ogra>
ubuntu has a guest session by default now
09:25
<Gadi>
but only fearless guests
09:26
<ogra>
would be saner to attach them directly
09:26
<Gadi>
how does that work?
09:26
the guest session?
09:26
<ogra>
tmpfs /home
09:26
and no access to anything on the system
09:26
nor user settings
09:26
<Gadi>
can you have multi-guest users?
09:27
<ogra>
hm, in gdm consolekit might handle that... not sure that works in our case though
09:27
since we dont have a dbus connection
09:27
thats really something we need to solve for next release
09:28
dbus gets more and more essential
09:28
like, gnoe will drop bonobo ... no applet will work without dbus interaction
09:28
*gnome
09:28
<Gadi>
dbus runs on the server, no?
09:28
<ogra>
so we will need something like dbus-ssh
09:28
right, thats the prob
09:29
dbus also runs on the client now for hal-input
09:29* Gadi only sees the problem with local apps
09:29
<ogra>
so we should use the clients system bus
09:29
no, there is probs with policykit, consolekit, session management etc
09:29
gconfd ...
09:30
if we have dbus on the client connected to the session bus on the server you can finally (at least for the desktop) log in multiple times
09:30
it wont cause any breakage in gnome
09:31
(doesnt help oo.o or FF, i know)
09:31Appiah has joined #ltsp
09:31
<ogra>
Gadi, btw i'm planning to write a new greeter for jaunty
09:32
and reviewing your timeout patch i saw you actually add hboxes or vboxes dynamically
09:32
never ever do that
09:33
keep the layout space for your items always there but hide the items rather
09:33
else you break the layout
09:33
<Gadi>
my timeout patch?
09:33
<ogra>
(just as a future advice)
09:33
<Gadi>
wasn't me
09:33* Gadi never wrote a timeout patch
09:33
<Gadi>
that was ryan52, I think
09:34
<ogra>
err, sorry
09:34
guest patch :)
09:34
if(allowguest) gtk_box_pack_start(GTK_BOX(vbox), guestbox, FALSE, FALSE, 0);
09:34
thats massively evil
09:34
<Gadi>
again, not me
09:34* Gadi has no such interest in guests
09:34
<Gadi>
:)
09:34
<ogra>
heh
09:34
ok, i thought both came initially from you
09:34
<Gadi>
talk to vagrantc and ryan52
09:34
they are all about being social
09:34
<ogra>
heh
09:35
well, i will rewrite the greeter more robust and a lot more themeable
09:35
<Gadi>
cool
09:35
<ogra>
floating transient popup windows FTW
09:35
<Gadi>
please add dialog boxes or some status box on the greeter itself
09:36
just keep in mind, we have no window manager
09:36
<ogra>
if i get it done how its in my head atm you can add remove boxes on the fly at the position you want them independently
09:36
we do
09:36
its called gtkfixed :)
09:37* Gadi smiles and nods and lets ogra do the coding ;)
09:37
<ogra>
you can specify positions with it on the parent (wallaper window)
09:37
<Gadi>
cool - and move things around?
09:38
<ogra>
it just requires a lot of math in the code to get a good default ... if we have that we can add themeable overrides for each of them
09:38
so you will be able to wite a theme file and assign positions for each element
09:38
<Gadi>
ok - then just keep dialogs modal
09:38
:)
09:38
<ogra>
totally free positions ... even overlaps if you want
09:39
<Gadi>
is this something you picked up working on those little toys of yours?
09:39
<ogra>
though some items need to be grouped, if i have a prototype to look at i'll start a discussion about which of them
09:39
nah
09:39
the constant theme issues just made me dig though the code
09:40* ogra still tries to get his hair in shape again after three days of looking at the different hacks and patches
09:40
<ogra>
ubuntu is in final freeze so i cant do any ubuntu development atm :)
09:41
well, beyond bugfixes ... but mobile has no bugs indeed, since *i* built it O:)
09:41
<Gadi>
freeze is before hackfest, right?
09:41
<ogra>
freeze is since yesterday
09:41
<Gadi>
oh, congrats
09:41
<ogra>
release os on 30th
09:41
*is
09:41
<Gadi>
spooky
09:41
:)
09:41
<alkisg>
ogra, from a teacher perspective, if you're going to write gtkgreeter again, please provide username/password edit boxes simultaneously, the students are getting confused with just one box! :)
09:42
<Gadi>
yeah, that and we should also let [TAB} = [ENTER]
09:42
<ogra>
alkisg, i'll consider that as option, though it breaks the concept and usability
09:42
<Gadi>
lots of folks are used to tabbing after username
09:42
<alkisg>
Gadi, right!
09:43
<ogra>
yeah, that could even get added to the current greeter
09:43
<alkisg>
ogra, usability in my classroom says that 2 boxes are needed! :P
09:43
<ogra>
alkisg, all human interface guides disagree with your classroom ;)
09:43
<Gadi>
alkisg: would TAB=ENTER suffice?
09:43
<ogra>
but yeah, we can likely make it an option
09:43
<alkisg>
Oh, come on, windows also have this! :P :P :P
09:43
<ogra>
huh ?
09:44
<Gadi>
I think from usability it would amount to the same thing
09:44
<ogra>
not since XP
09:44
<johnny>
and many web pages
09:44
i think they are used to it from web pags
09:44
pages*
09:44
<alkisg>
ogra, sure, if you press alt+ctrl+del to show the logon dialog, it's there on xp also
09:44
<johnny>
can anybody tell me a web site off the top of their heads that doesn't put both entries on one page?
09:44
<alkisg>
Gadi, no, they get confused on where to write the password since they don't see a box (before pressing enter, that is)
09:44
<ogra>
well, anyway the current infrastructure desont allow it
09:45
so such a design would require changes on a deeper level
09:45
more than the greeter
09:45
<alkisg>
ok, it's minor, no problem... :D
09:46
<johnny>
it's standard on the web
09:46
<alkisg>
They get used after the second lesson
09:46
<johnny>
so you don't have to make a seperate request
09:46
just to get anothet form field
09:46
<ogra>
i'll start off with the current greeter code but strip out all elements and redesign the layout stuff from ground up
09:47
<johnny>
i consider the greeter the odd one out here
09:47
most programs have them together as well
09:47
<ogra>
greeter_iface in ldm itself would need a new mode to accept a user/pw combo instead of waiting for each of them separately
09:47
<johnny>
not sure how that is good usability
09:47
<ogra>
gdm has it like that
09:47
by default, in all distros
09:47
<johnny>
to contradict an accepted behaviour
09:48
<ogra>
the ldm ui always tired to stay close to the gdm ui
09:48
<johnny>
yeah.. i understand that
09:48
it just doesn't make sense in context of EVERY OTHER PROGRAM and web site
09:48
<ogra>
if someone wants to make a kdm ripoff, feel free :)
09:49
<johnny>
sometimes you have to fire your usability focus group
09:51
<alkisg>
ogra, (not a request) gdm also has a face greeter, which is much more user intuitive than the plain one. It would be useless in labs with 100 user accounts though...
09:51elisboa has quit IRC
09:51nicoAMG has quit IRC
09:51
<ogra>
its uninmplementable in ltsp
09:52
<alkisg>
ok, I'm not requesting, I'm happy with Greek in ldm! :P :D :D
09:53elisboa has joined #ltsp
09:53
<Gadi>
unimplementable? that sounds like a challenge
09:53
:)
09:53
Im sure scotty could do it
09:54
<_UsUrPeR_>
is there a --kiosk option in Fedora 9 like there is in ubuntu when using ltsp-build-client?
09:54
It's error'ing on me
09:54
<ogra>
Gadi, only implementable based on a massve security breach
09:54elisboa has joined #ltsp
09:54
<ogra>
_UsUrPeR_, thats an ubuntu plugin
09:54
<_UsUrPeR_>
k
09:55* ogra always found that to insane to commit upstream :)
09:56
<ogra>
thoug its actually been our first localapp implementation :)
09:56Blinny has joined #ltsp
09:57
<Gadi>
3 cheers for localapps!
09:58
<_UsUrPeR_>
so with the ubuntu ltsp image I just created with --kiosk, is there a way to get it to not be slow?
09:58
<Gadi>
use better hardware
09:58
<ogra>
Gadi, btw, would you mind bringing up xrandr integration by default at BTS ?
09:58
<_UsUrPeR_>
Hardware on what? client or server?
09:58
<ogra>
(since i wont be there)
09:58
client
09:59
<Gadi>
ogra: sure - though not all drivers have caught up with xrandr
09:59
<ogra>
the kiosk runs 100% on the client
09:59
in ubuntu and FC they have
09:59
<_UsUrPeR_>
ORLY
09:59
<Gadi>
wait - hold the phone - ur not coming to BTS?
09:59
<ogra>
nope
09:59
<Gadi>
ok, get me Mark's number
09:59
the one for his Jet's phone
10:02wpgmb has joined #ltsp
10:03elisboa has quit IRC
10:04
<wpgmb>
still struggling with ltsp server: what config file should I be looking at to enable a system (not a client) on the backbone subnet so it can ssh directly into the ltsp-dhcp side subnet?
10:11cliebow has quit IRC
10:11mikkel has joined #ltsp
10:15_UsUrPeR__ has joined #ltsp
10:17_UsUrPeR__ has joined #ltsp
10:17tjikkun_work has quit IRC
10:19_UsUrPeR_ has quit IRC
10:20_UsUrPeR__ has quit IRC
10:20
<six2one>
wpgmb: does this system have 2 NICs?
10:20
<wpgmb>
yes
10:20_UsUrPeR_ has joined #ltsp
10:21
<wpgmb>
six2one: sorry - clarification: the ltsp server has two nics, yes
10:21mighty-d has joined #ltsp
10:21
<mighty-d>
Hi
10:21
<six2one>
you will need to set up routes so the outside (non ltsp network) can connect to the inside (ltsp dhcp network)
10:21
is there a reason u need to ssh to the ltsp side of the server?
10:22
<mighty-d>
i want to deploy a 70 thin clients setup, i was thinking on 8GB RAM and a Xeon Quad Core processor for the server, do you think this will be enough?
10:22
clients will be runing openoffice, mozilla and rdesktop most of the time
10:23
<wpgmb>
six2one: I thought so. Some had suggested IP forwarding would do the trick, but it doesn't. Yes - reason: I want to be able to ssh into /home on the LTSP server from a system outside the ltsp network
10:24
<mighty-d>
wpgmb, as six2one said, you might have issues with routes on other subnets, try adding an static route to your default gateway
10:24
<wpgmb>
mighty-d: that's a lot of power. It will depend on the actual usage. Ie - if all 70 are watching youtubes with sound... it may bog down speed
10:24
<six2one>
mighty-d: your network will be important as well....
10:24
mighty-d: enabling local apps will take some load off
10:25alkisg has left #ltsp
10:25
<wpgmb>
mighty-d: I don't think that's the issue: I can ssh into eth0 (gateway nic) and I can ping it. From there, I can ssh into eht1 (dhcp nic). Once I'm there, I can ssh into my system where it all started from
10:25
mighty-d: and clients can access the 'Net just fine, without a static route
10:25
<Gadi>
wpgmb: echo "1" >/proc/sys/net/ipv4/ip_forward and edit the corresponing line in /etc/sysclt.conf
10:26
<six2one>
wpgmb: what version of ltsp u runnin? 5?
10:26
<mighty-d>
wpgmb, pelase explain the problem to me, i just got here so i am missing everything
10:26
<Gadi>
*/etc/sysctl.conf
10:26
<wpgmb>
Gad1: been there, done that. Not effective
10:26
<mighty-d>
six2one, i cant run local apps
10:26
<wpgmb>
six2one: yes, I installed the Hardy LTSP version
10:26
<Gadi>
wpgmb: do you have NAT running?
10:26
or any other firewall rules?
10:27
<wpgmb>
six2one: it boils down to this: what config file should I be looking at to enable a system (not a client) on the backbone subnet so it can ssh directly into the ltsp-dhcp side subnet?
10:27
<mighty-d>
six2one, but do you think this set up is appropiate or should i go for two xeon dual core or quad core isntead?
10:27
<six2one>
wpgmb: you will need to get static routes set up, the "outside" has no idea that the ltsp network exists...just the internet facing nic
10:28
<Gadi>
wpgmb: you need to enable routing, disable NAT, and set a route that uses the backbone interface address as the gateway to the thin client subnet
10:28alkisg has joined #ltsp
10:28
<six2one>
mighty-d: i would go as big as you can, 70 "THICK" clients will be some load...you will need a good raid array as well....
10:28
<mighty-d>
wpgmb, or set up rip or ospf, i know it brings more complexity but you can scale your setup nice
10:29
six2one, you think i can go with lvm instead of RAID ?, should i go with raid hardware or can i use software?
10:29
six2one, i want to know if what i thinking is reasonable or if i should think bigger
10:29
maybe 12 GB RAM ?
10:30
<six2one>
mighty-d: i ran 50 clients from a single amd optie (forget the speed) with 4gb ram...ran well...HOWEVER...all the clients were doing were IBM client access term sessions back to the server (terms on their own network) and then from the server to mainframe
10:30
mighty-d: i would get a GOOD 3ware card
10:30
<loather>
depends on what your goal is out of the RAID
10:30
if you want performance, then software raid is not for you
10:31
areca and 3ware both make excellent hardware that is fully-supprted by the linux kernel
10:31
<mighty-d>
loather, ok, ill take your advice
10:31
<six2one>
mighty-d: loather is right...with 70 clients, u wont want a slow array and u need to take a failure and still keep ur shit up....70 clients might notice downtime
10:31
lol
10:31
<loather>
plan to spend somewhere between $500-$1000 dependent on how many disks you have in your array
10:32staffencasa has joined #ltsp
10:32
<mighty-d>
loather i was thinking on raid 5 with 750 GB disks
10:33
<loather>
how many 750s? you can do raid 5 with 3 disks on up
10:33
<Blinny>
I have an 8GB quad-core dual-processor box, and the most I've had it up to is 60% RAM usage when 16 thin clients were on-board. They're heavy though, FF w/ flash, Thunderbird, and OOo with full gnome
10:34
I think 70 clients would be better served by multiple servers.
10:34
<mighty-d>
http://www.z-a-recovery.com/art-raid-estimator.htm
10:34
<wpgmb>
Gad1:"you need to enable routing, disable NAT, and set a route that uses the backbone interface address as the gateway to the thin client subnet" - makes sense, I think. Thanks
10:34
<Blinny>
Really, the biggest problem you'll run into is maxing out a GigE NIC
10:35
<mighty-d>
Blinny, i was thinking on multiple NICS to do the trick
10:35
i dont trust assymetric switchs
10:35
<Blinny>
Definitely.
10:35
<loather>
there's bondage, erum, bonding, to fix that thougH :)
10:35* mighty-d had bad experiences
10:35
<ogra>
there are also fiber switches that support interface trunking
10:36
<loather>
copper switches too
10:36
<Blinny>
If it were my network I'd do multiple failover round-robin application servers with a fiber back connection to NFS boxes sharing /home and /usr
10:36
<ogra>
you can bundle several fiber lines with tham to get multi Gbit
10:36
<six2one>
mighty-d: yea, ull need to team the nics for sure....ltsp5 is WAY heavier than 4.2
10:36
<Blinny>
Depends on how you subnet.
10:36* ogra wouldnt say 'way'
10:36
<ogra>
but its for sure heavier
10:37
<six2one>
ogra: well i have been upset recently, my good ol' trusted dell gx100s dont work well anymore
10:37
<ogra>
how much ram do they have ?
10:37
<six2one>
ogra: 512
10:38
<mighty-d>
Blinny, how will you do the trick to log in to your fail-over array, since you dont have xdm, does ldm support this already?
10:38
<six2one>
ogra: in my experience the via edens are sexy and work great
10:38
<ogra>
thats way more than you need
10:38
<six2one>
ogra: so ima use them
10:38
<ogra>
yeah
10:38
<six2one>
ogra: it SHOULD be more than i need, but they eventually crash and burn...multiple firefox tabs FTL!
10:39
<Blinny>
mighty-d: I'd have to work on the automagic failover - at first it'd just be 'Oops bad things happened, you need to re-log in'
10:39
<ogra>
64M are good for booting and no extra features, 128M for features, 256M are perfect
10:39
<six2one>
ogra: and since they are like celeron i cant do local apps
10:39
<ogra>
six2one, what distro/release ?
10:39
<mighty-d>
Blinny, suppose i acquire two servers for 35 users each
10:39
<six2one>
ogra: 8.04.1 ltsp5
10:39
<mighty-d>
btw, i think your box is pretty high for 16 users
10:40
<six2one>
they worked great back in the day with 4.2
10:40
<Blinny>
something like http://peopleplaces.org/net.png
10:40* mighty-d has a 25 users box with a dual core 2.2 ghz (core 2 duo) and 4 gb ram running hardy
10:40
<ogra>
hardy had some issues ... like the firefox urandom breakage
10:40
there are fixes in the queue
10:40
<Blinny>
that's not exactly it of course - pulled that from a different setup
10:41
Yeah I know - it also does database stuff. Eventually the box will serve ~30 thin clients
10:41
but I haven't migrated them all yet
10:41
<ogra>
six2one, hardy will likely see a good bunch of extra fixes after the interpid release
10:41
<Blinny>
ogra: YEA!
10:41
<ogra>
currently all devs are busy with the release, after that hardy SRUs will get more attention again
10:41
<six2one>
ogra: dude, i couldnt care less!!! this is WAYYYYY better than the old school fedora 4 days when i started this stuff!
10:42
<ogra>
i personally have a list for ltsp already
10:42
six2one, well, still i'm a tad unhappy that hardy gets so much bad press compared to gutsy so i have interest to have it rock ;)
10:43
<six2one>
ogra: ill gladly feed it better terms, im so glad they jumped off the fedora ship...it gets old when the server goes stale after like 6 months
10:43
<Blinny>
mighty-d: You may just get better reliability/redundancy out of a few smaller servers than one or two bigger ones. You could even go raid sata
10:43
<six2one>
ogra: .deb FTMFW!
10:43
<ogra>
i want it in a state that Gadi recommends it to people over any other release :)
10:43
and bribing didnt help yet
10:44
<six2one>
haha
10:44
<ogra>
so i have to apply fixes it seems
10:44
<Gadi>
hehe
10:44
<ogra>
or i'm not bribing good enough
10:44* Gadi can be a ball-buster
10:45
<mighty-d>
Blinny, i am looking at your deployment, its kinda nice but i dont think my budget allows for such a redundancy, ill have to go with something modests
10:46
Blinny, the lvs router its only a failover for routes?, or you propose a cluster for the servers as well?
10:47
<Blinny>
Cluster everything!
10:47
Really it's up to your & your budget
10:47
and how you can tolerate downtime or slowness if a server goes out
10:48
For my particular setup, I wanted NO downtime, so I went hardware RAID with notifications of failures, multiple power supplies, backup NICs etc
10:48
but it's time for me to go exercise - g'luck be back in a bit if you have more questionS!
10:49
<loather>
ipvs is great. it's a full layer 4-7 switch
10:49
<mighty-d>
Blinny, for my deployment hardware raid sure, multiple power supplies definately, backups NICs doable, an extra server just to sit there and wait for a downtime not possible
10:49
<loather>
well, it'd be load-balanced
10:49
you actually make the connection to the ipvs box, which directs it to the server with the least load
10:49
<mighty-d>
thanks Blinny
10:50
thanks six2one and ogra and wpgmb, gotta go now too
10:50* ogra goes back to play with a new themeable ldm greeter
10:50
<six2one>
mighty-d: np...im usually here
10:51
<loather>
ogra: oooh, sexy! Where can I score source? :)
10:52
<ogra>
loather, not yet, still in a conceptual phase
10:54
<loather>
i see. when you're ready i'll help test. I have to build a new ltsp box here shortly
10:54
<ogra>
loather, trying out some new concepts to put elements in transparent floating windows and the like
10:54
so everything becomes freely positionable
10:54
<loather>
oh, that's pretty cool. ultimate control over where everything goes.
10:55
<ogra>
right
10:55
<nubae>
hmm why does tee need /dev/null on the end when u are writing input to a file?
10:55
<ogra>
and no more issues with logosizes etc
10:56
<loather>
logo too big? move the box down. :)
10:58wpgmb has quit IRC
11:01sbalneav has joined #ltsp
11:01
<sbalneav>
Morning all
11:02
Hmm, I need to figure out what to put in my .irssi to make it reconnect after a netsplit
11:04toscalix has quit IRC
11:05
<nubae>
for anyone curious, this is how to get sugar, ejabberd and collaboration working perfectly on LTSP: http://www.nubae.com/sugar-on-ltsp-ubuntu-intrepid-ibex
11:05
I wonder how much memory sugar needs... could be a good minimal install that requires less ram on TCs
11:07
<sbalneav>
ltspbot: learn sugar as for anyone curious, this is how to get sugar, ejabberd and collaboration working perfectly on LTSP: http://www.nubae.com/sugar-on-ltsp-ubuntu-intrepid-ibex
11:07
<ltspbot>
sbalneav: The operation succeeded.
11:07
<sbalneav>
!sugar
11:07
<ltspbot>
sbalneav: "sugar" is for anyone curious, this is how to get sugar, ejabberd and collaboration working perfectly on LTSP: http://www.nubae.com/sugar-on-ltsp-ubuntu-intrepid-ibex
11:09
<ogra>
nubae, :( where are the pictures ?
11:09* ogra wants screenshots and photos
11:10
<nubae>
:-) ok... I'll add them
11:11
<_UsUrPeR_>
lol
11:11
<nubae>
its super cool... walk into a lab with a laptop, plugin into the lan, and bam! sugared environment
11:11
now tell me what else can do that ;-)
11:13
<Q-FUNK>
;)
11:22wwx has quit IRC
11:25wwx has joined #ltsp
11:26Subhodip has joined #ltsp
11:40alkisg has left #ltsp
11:49
<nubae>
ogra: when u say photos, u mean of the terminals themselves?
11:52syyp has joined #ltsp
11:52
<ogra>
or beautiful wimen, make your pick :)
11:52
<syyp>
Ehlo.
11:53
<nubae>
lol
11:53
or beautiful women using thin terminals
11:54
<_UsUrPeR_>
sotp that right now
11:54
<nubae>
:p
11:54
<_UsUrPeR_>
:)
11:55indradg_ has quit IRC
11:56
<syyp>
What's the current wisdom on setting up local apps?
11:56X0d_of_N3d_ has joined #ltsp
11:56
<_UsUrPeR_>
syyp: what OS are you using?
11:56
<syyp>
Ubuntu 8.04
11:57
<_UsUrPeR_>
do you have LTSP currently installed and working?
11:57
<syyp>
IIRC, there was some talk of sshfs and other weirdness.
11:57
Yes.
11:57
<_UsUrPeR_>
syyp: not that I am aware of... It's a pretty streamlined process now
11:57
chroot to the image you are working with
11:57
<syyp>
Oh, ok. Is this link then considered the most up-to-date instructions? http://www.ltsp.org/twiki/bin/view/Ltsp/LocalApps
11:58* _UsUrPeR_ pulls up link
11:59
<syyp>
I'm primarily interested in local apps, not local homedirs so, I'm guessing the SSH root@ltsp without NIS and/or Password section is what I need?
11:59
<_UsUrPeR_>
syyp: give me one moment
12:00* syyp gives _UsUrPeR_ two of them.
12:01CAN-o-SPAM has joined #ltsp
12:02keithclark has joined #ltsp
12:02* syyp brandishes a can opener at CAN-o-SPAM
12:02
<syyp>
;)
12:03
<_UsUrPeR_>
I am able to use localapps, and there has been no changes to my present lts.conf
12:03
jwas just checking
12:03
<syyp>
meaning...
12:04syyp has quit IRC
12:04
<_UsUrPeR_>
ok
12:04
so here's what I have done recently
12:04
first: "mount --bind /proc /opt/ltsp/i386/proc"
12:05
second: "chroot /opt/ltsp/i386/
12:05
"
12:05
third: "apt-get install firefox"
12:05
fourth: "exit"
12:06
fifth: "ltsp-update-image"
12:06
sixth: "?????????????"
12:06
seventh: "profit"
12:06
:D
12:07
the sixth and seventh are actually "turn on the client" and "xrexec firefox in terminal"
12:07
respectively
12:10zoredache_ is now known as zoredache
12:11Lns has joined #ltsp
12:11
<_UsUrPeR_>
oh. he left
12:11* _UsUrPeR_ has learned a valuable lesson today pertaining to talking to himself in public forums
12:12indradg_ has joined #ltsp
12:12
<zoredache>
what was the lesson?
12:13
<_UsUrPeR_>
I just typed instructions out on how to install firefox as a localapp to an individual who had left minutes prior
12:14
<nubae>
its good practice _UsUrPeR_:p
12:14
<_UsUrPeR_>
I think I got it right :)
12:14
<zoredache>
ah, so you weren't talking to yourself. The other person was just no longer listening
12:15makghosh has joined #ltsp
12:15
<_UsUrPeR_>
precisely
12:15
<zoredache>
welcome to IRC
12:15
<_UsUrPeR_>
I used to use IRC waaay back in the day
12:16
<nubae>
yah, I almost failed uni because of it
12:16
<_UsUrPeR_>
for less... couth things. Pertaining to ye olde warez
12:17vagrantc has joined #ltsp
12:17
<_UsUrPeR_>
ahh to be young again. No fear of corporate reprisal due to my legal status as a minor.
12:18jammcq has joined #ltsp
12:19
<jammcq>
hello friends
12:19* vagrantc waves
12:19
<_UsUrPeR_>
hey jamm
12:19
<jammcq>
hey
12:23
<ogra>
mumble ...
12:28
<jammcq>
mumble ?
12:28
<Lns>
hi jammcq
12:28
<jammcq>
hey lns
12:28
what's shakin?
12:28
<Lns>
just my ltsp term 1220 ;)
12:28
I love that thing
12:29
ogra: your mumble reminded me of the Troll from Zork I... not sure why...I think he mumbles if you try to talk to him...hehe
12:30
<ogra>
that was just a unmotivated mumbling while playing with GtkFixed on the rootwin for a new ldm greeter
12:30
<Lns>
oooo
12:35* vagrantc mutters about hppa and arm holding up ldm
12:36
<ogra>
heh
12:36
hppa
12:36
i'm alwas amused if my mobile builds fail on it
12:36* ogra tries to imagine a hppa mobile device
12:38
<vagrantc>
what's really, really annoying about hppa is the buildd actually sucessfully built the package ... 2 days ago ... and it still hasn't been uploaded.
12:38
at least arm hasn't yet built it.
12:38
of course, it's my own stupid fault for not correcting a mistake i made in the previous upload as soon as i saw it
12:40
ogra: so you've been looking into rewriting the greeter?
12:40
<ogra>
vagrantc, yes
12:41
having one rootwin and adding all items as GtkFixed
12:41
that way they completely are placeable without distrubing each other
12:41
and the theme shouldnt interfere anymore ... you can even overlap items if you like
12:43
by default it will resemble the current look, but you shojld be able to override each and every position in the end
12:43
its a lot of math though ...
12:43* Gadi hands ogra a slide rule
12:43
<Gadi>
and a book about how to use a slide rule
12:44nubae has quit IRC
12:44
<Gadi>
and another about why you don't need to use a slide rule
12:44
<ogra>
GtkFixed is evil, you have to use absolute sizes for everything so i need to compute all values dynamically based on screensize
12:44
<vagrantc>
eeyk.
12:44
<johnny>
didn't we already fix that on the web.?
12:44
and yet it still exists???
12:44* johnny attacks
12:45
<vagrantc>
yeah, lets implement a web frontend to LDM!
12:45
<ogra>
good idea 1
12:45
!
12:45
i guess that would work fine btw
12:45
<johnny>
well.. gnome is seeming to go web everywhere
12:45
with webkit usage
12:45
<ogra>
you just need to make greeter_iface a cgi
12:45
what ?
12:45
<Lns>
hmm...that means you'd have to run FF as a localapp right? ..
12:45
<ogra>
what makes you think that
12:45
<vagrantc>
evil.
12:46
<johnny>
ogra, it's the word on the street
12:46
<ogra>
Lns, yeah, that would be a requirement :)
12:46
<johnny>
Lns, more like webkit-gtk
12:46
<Gadi>
or lynx
12:46
<ogra>
johnny, ah, lucily we walk on different streets then :)
12:46
<Lns>
hehe...lynx would be cool
12:46
<johnny>
gnome almost dropped xulrunner support completely from epiphany
12:46
<ogra>
johnny, i thought that was abandoned since months again
12:46
<johnny>
i'm using epiphany webkit right now
12:47
<ogra>
its an option but not the default
12:47
<johnny>
that's cuz gtk webkit isn't done yet imo
12:47
that it seems to bve the overwhelming opinion of the folks in gnome-hackers
12:48
and now there are js bindings to glib
12:48
<ogra>
anyway, back to greeter_new.c :)
12:48* ogra dives into his coding pit
12:49
<johnny>
personally.. i'd be all for it. if they ever went as far as to implement gtk on the web
12:49
via webkit
12:49
instead of all this clutter stuff and all
12:54nubae has joined #ltsp
12:58
<ogra>
you cant compare that
12:59
clutter is deeply tied in with your graphics cards GL and composite functions ... webkit is just a renderer
12:59six2one has quit IRC
13:02six2one has joined #ltsp
13:08
<Lns>
wow, Mobile FF Alpha 1 just released: http://tech.slashdot.org/article.pl?sid=08/10/17/1656231
13:08
<nubae>
anyone know when chrome is coming to linux?
13:14
<johnny>
who cares
13:14* johnny kicks chrome in the face
13:14* zoredache looks into a magic 8-ball - 'The outlook is unclear'
13:14
<johnny>
nubae, why would you need chrome?
13:14
just use another webkit based browser
13:15
you'll get all the benefits
13:15
and none of the drawbacks :)
13:15spectra has joined #ltsp
13:15
<zoredache>
actually, not all of the benifits... The new javascript engine is unique to chrome right now
13:15
<johnny>
not for long i imagine
13:19
<nubae>
http://www.nubae.com/sugar-on-ltsp-ubuntu-intrepid-ibex <-- wid screen shots, especially for ogra
13:20
johnny: I dunno, it does kinda look neat
13:20
I meant... javascript running 10 times faster..
13:20
<johnny>
the new mozilla has plenty fast js
13:20
<nubae>
dragging tabs into their own windows... I like the idea
13:20keithclark has quit IRC
13:21
<johnny>
nubae, gnome terminal does that now :)
13:22
seems like firefox easily could..
13:22
but i guess the main thing.. is running each as a process
13:22
not sure when firefox is gonna try that.. or not
13:23
<nubae>
yeah indeed, tabs crashing instead of the whole thing is much better
13:23
but why dont u like chrome?
13:23
cause its google monopoly plan?
13:23
<johnny>
something like that
13:23
<nubae>
:p
13:24
yeah lately, they took back their famous 'we are not evil' keyphrase
13:24
Director said it was no longer appropriate
13:24
<johnny>
i limit my usage of all things google
13:24
not that they don't do some good things
13:25
cuz they do..
13:25
<nubae>
they do better than the rest, but still could do much more
13:25
<johnny>
they are a for profit company tho.. can't give away too much
13:26six2one has quit IRC
13:26six2one has joined #ltsp
13:29jc2it has quit IRC
13:31
<nubae>
I can't believe the 2 ssh discussion is still going...
13:32
<johnny>
what is the question?
13:33
i cna't get sf to thread any kind of decent way
13:33radoeka has joined #ltsp
13:33
<johnny>
to be able to use that interface
13:33radoeka has left #ltsp
13:35
<vagrantc>
google retracted it's "don't be evil" ? that's like saying ... we're intending to do evil.
13:35
<nubae>
yeah, was in the latest wired
13:36
let me see if I can find the ext thing they said
13:36
<_UsUrPeR_>
sounds like google realized where the true money was
13:37keithclark has joined #ltsp
13:37ogra has quit IRC
13:38ogra has joined #ltsp
13:38
<_UsUrPeR_>
"Google's latest press release states that they will be charging for access to the search engine on a monthly basis, using the new "Google Integration(tm)" corporate platform. Every paid user will be allowed to keep their gmail account."
13:39
<nubae>
http://www.wired.com/techbiz/media/magazine/16-10/st_15googleevil
13:40
_UsUrPeR_: access to search, where did u read that?
13:40
<Q-FUNK>
??!
13:40
<_UsUrPeR_>
lol. Just made it up ;)
13:40
<Q-FUNK>
what? bait and switch?
13:40
<nubae>
oh... hehe
13:40
<_UsUrPeR_>
hahaha
13:40
<Q-FUNK>
DOH!
13:41
<_UsUrPeR_>
we could perpetuate this rumor and make a killing on the stock rebound if you guys want though ;)
13:41
<Q-FUNK>
meh
13:44
<Lns>
ugh...ok flash 10 isn't ready for ltsp/linux yet i think...massive CPU hogging going on vs. flash 9
13:44
at least through adobe provided ubuntu 8.04 .deb
13:47jarias has joined #ltsp
13:47
<nubae>
unsurprisingly... could have predicted that one
13:47
<_UsUrPeR_>
color me surprised.
13:48
<nubae>
People tell us: "What Edubuntu needs is content!! Lessons!!
13:48
Data!! Videos!! Manuals!!", we say, "Fantastic! Give us what you've got!",
13:48
and all that comes back is the sound of crickets. - sbalneav
13:48
love it... :D
13:51
<zoredache>
what we need is a some kind of version control system that works well for curriculum content, and a site similar to sourceforge to host it
13:53
<nubae>
zoredache: that already exists...
13:53
more than 1 actually
13:54
ah wait. version control...
13:54
yeah good idea... launchpad springs to ming
13:54
mind
13:54
<zoredache>
I know somethign exists... i just dont think the 'works well' bit is done
13:55
<nubae>
right... there's pretty much nothing, I was on the lookout for stuff to include in the XS server's Moodle for OLPC
13:55
curriki is the closest thing, and door as a repo for exchanging data, but both suck really
13:56Subhodip has quit IRC
13:56
<nubae>
standardisation of content seems to be the last thing on people's minds
13:56
<zoredache>
I don't think content needs to or always should be standardized... I think such a system would probably need to be more distributed
13:57
allowing each teacher to branch off someone elses stuff, add there own bits, then share that back
13:57
<johnny>
so.. who wants to pay me to develop it :)
13:58
lol
13:58
what about MIT's open courseware
13:58
what do they use?
13:58
<zoredache>
I don't think lots of forks of content would be as painful as it is in the OSS world...
13:58
<johnny>
it seems like you could aggregate lots of info for edubuntu tho
13:58
like links to open courseware, creative commons
13:58
<nubae>
well it is if you want to put it into a virtual learning environment like moodle
13:58
<johnny>
archive.org
13:58
etc
13:59
<nubae>
johnny: yeah, if only there were more hours in the day ;-)
13:59
<johnny>
there's alot of want.. but people don't understand how to contribute back
13:59
somebody told me not to bother giving any presentations on open source tho
13:59
that the market was already saturated
14:00
<nubae>
its difficult to get involved sometimes
14:00
saturated? what part of the market is saturated? seems the most new growing market there is
14:00
<johnny>
is there any web page that helps people get involved with open source and navigating the system ?
14:00
i'll let the person remain nameless..
14:00
maybe they didn't understand that it's not current developers who need to be told stuff
14:01
ubuntu's motu thing would be nice tho
14:01
but not just for ubuntu obviously..
14:01elisboa has joined #ltsp
14:01
<nubae>
right ubuntu's system clearly works...
14:01
as opposed to gentoos :p
14:02
<johnny>
huh?
14:02
<nubae>
the politics
14:02
<johnny>
i didn't say anything about gentoo
14:02
<nubae>
no I did ;-)
14:02
<johnny>
not sure why you brought it up
14:02
it has nothing to do with what i'm talking about
14:02
i am not a gentoo developer
14:02
even
14:02
<nubae>
hmm... u made ltsp for it right?
14:02
<johnny>
or anyway affilliate with gentoo
14:03
<nubae>
doesnt that make u a developer?
14:03
<johnny>
lots of people make ebuilds who are not developers
14:03
<nubae>
anyway, I was just comparing the 2 political structures... gentoo council and motu council
14:03
one works, the other doesnt
14:03
<johnny>
i have no idea
14:03
about how or if gentoo's council works
14:04
i'm completely out of the loop on that stuff
14:04
<nubae>
come on there were releases delayed by months just cause they couldnt agree on stuff
14:04
<johnny>
my only concern was certain trolls not being kicked out
14:04makghosh has quit IRC
14:04
<johnny>
nubae, i agree on the new idea
14:04
automating release media
14:05
nubae, the real problem is that nobody wanted to be on releng
14:05
not political at all
14:05
nobody wants to do that unsexy kinda work
14:05
<nubae>
releng?
14:05
<johnny>
release engineering
14:05
the people who roll the releases
14:05
<nubae>
ahz
14:21dmaran has left #ltsp
14:23
<_UsUrPeR_>
for ubuntu kiosk mode, I'm running a LTSP Term 1422 PXE with a gig of RAM and a 1.2gHz proc. Why does ubuntu kiosk take so long to start? It's been 6 minutes and I still am waiting for firefox to pop up :|
14:33siki has quit IRC
14:39Blinny has quit IRC
14:40rcy` has quit IRC
14:50jarias_ has joined #ltsp
14:55jarias has quit IRC
14:55jarias_ is now known as jarias
14:59
<dberkholz>
nubae: council has nothing to do with it. it's lack of a single leader for life, imho
15:02ogra has quit IRC
15:06ogra has joined #ltsp
15:06
<nubae>
dberkholz: totally agree
15:11pmatulis has quit IRC
15:12K_O-Gnom has joined #ltsp
15:15keithclark has quit IRC
15:16jarias_ has joined #ltsp
15:16jarias has quit IRC
15:16jarias_ is now known as jarias
15:17keithclark has joined #ltsp
15:21
<nubae>
what does LDM_TIMEOUT_ENABLE do exactly?
15:23
vagrantc: ?
15:24
<johnny>
i assume timeout for autologin..
15:24* johnny guesses
15:25
<nubae>
any idea how much timeout?
15:25
<johnny>
lol.. grep the src
15:25
not sure
15:26
<vagrantc>
nubae: it enables the timeout
15:26
nubae: it requires also setting LDM_TIMEOUT_TIME ...
15:26
why it wasn't coded with a default timeout is beyond me
15:26
<nubae>
timeout of what exactly, being able to login via LDM?
15:27
ie, what happens when timeout happens
15:27
<vagrantc>
nubae: it does autologin after the timeout
15:27
<nubae>
oh ok... understood
15:27
<vagrantc>
might want to rename those variables
15:28
<nubae>
LDM_LOGIN_TIMEOUT would sound better
15:28
<vagrantc>
nubae: is that a boolean or a number? :P
15:29
<nubae>
well if u made it a number, it wouldnt require the ldm_timeout_time too :p
15:29alkisg1 has joined #ltsp
15:29
<vagrantc>
that's the tricky part with this option ... to be consistant with LDM_AUTOLOGIN and LDM_GUESTLOGIN, there should be a boolean to enable/disable it, but then it also needs a way to configure the timeout
15:30
<johnny>
consistancy isn't worth it ..
15:30
<nubae>
its superflous
15:30
<johnny>
if it isn't there.. it doesn't timeout
15:30
if it is.. and it's not 0..
15:30
then it does
15:30
<nubae>
right
15:30
<vagrantc>
patches accepted :)
15:31
<nubae>
johnny: ? I dont work upstream, just docs... ;-)
15:31
<vagrantc>
that's actually what i had suggested when Ryan52 was implementing it, but someone wanted a boolean too
15:31
<johnny>
lol
15:31
just saying that you could find out the workings by looking at the src..
15:32
<nubae>
so we agree on LDM_LOGIN_TIMEOUT with no on off switch (the on off switch only makes sense if there is a default time out)
15:33
?
15:33CAN-o-SPAM has quit IRC
15:33
<vagrantc>
nubae: sounds good to me.
15:34CAN-o-SPAM has joined #ltsp
15:34
<vagrantc>
shouldn't be too hard to patch.
15:34CAN-o-SPAM has quit IRC
15:36Egyptian[Home] has quit IRC
15:36
<ogra>
nubae, is there a howto in the docs to set up key based auth ?
15:37Egyptian[Home] has joined #ltsp
15:37
<vagrantc>
i think it's as simple as setting /root/.ssh/id_rsa without a password
15:37
or any of the other default ssh keys
15:37
<ogra>
vagrantc, but then only with autologin, right ?
15:37
<vagrantc>
ogra: no
15:37
ogra: any login will attempt to use that key
15:38
there's no special code in ldm to use keys
15:38
<nubae>
ogra: no there isn't, I was thinking about that the other day
15:38
<ogra>
heh, not massively secure though
15:38
<nubae>
also, there is nothing about local apps either
15:38
<ogra>
unless you have a chroot per client ans restrict nbd access
15:38nubae has quit IRC
15:38
<vagrantc>
ogra: it allows you to set up accounts without passwords
15:39
so that the user can;'t change the password
15:39
<ogra>
right
15:39
but everyone would use the same key
15:39
<vagrantc>
that's about all you get
15:39
indeed.
15:39
it's only useful for anonymous login anyways
15:39* ogra notices he took his mouth to full in ltsp-discuss
15:40nubae has joined #ltsp
15:40
<ogra>
damned hardening thread
15:40
<nubae>
I know...
15:40* ogra wishes jammcq could just delete it :)
15:40
<nubae>
like 60 replies already
15:40
Even Lns lost interest :-p
15:41
<Lns>
heh
15:41
I blew all my steam out...no more steam
15:41
<jammcq>
delete it??? I just add a .procmail rule so I never see it again
15:41
I've got about 160 such rules in my .procmailrc file
15:42
<johnny>
lol
15:42
what is this thread?
15:42
<Lns>
Can any of you guys vouch for it NOT being trivial to simply add a secondary sshd_config file w/different options, and launch with it? Is there more to it?
15:42
<johnny>
is there a good link to view it
15:42
<jammcq>
I'll never ever see another argument about why scsi is better than sata :)
15:42
<nubae>
now only if folks could be so passionate about helping with community stuff
15:43
<jammcq>
Lns: it's an init script and a config file. trivial in my opinion, and anybody administering a Linux/Unix system outta be able to handle that one
15:43
<nubae>
jammcq: can u add the link to the irc logs in the topic...
15:44
<jammcq>
ummm, where are the logs?
15:44
<Lns>
k, yeah i forgot to mention the init script...but...man. It really doesn't get more simple than that.
15:44
<nubae>
they are here: www.nubae.com/logs/
15:45
<vagrantc>
ogra: do you care if we change LDM_TIMEOUT_ENABLED (boolean) and LDM_TIMEOUT_TIME (# seconds) to simply LDM_LOGIN_TIMEOUT (# seconds) ?
15:45
<ogra>
not at all
15:45
<johnny>
can somebody please show me a nice threaded view of htis topic
15:45
so i can feel dumber please?
15:45
<ogra>
t wont get into intrepid, so we might need to keep the documented option
15:45
<Lns>
johnny: follow the ltsp-discuss link on the ubuntu wiki
15:45
johnny: https://help.ubuntu.com/community/UbuntuLTSP/
15:46ChanServ sets mode: +o jammcq
15:46
<jammcq>
:)
15:46
<ogra>
vagrantc, i would apreciate if the layout mangling didnt hapen btw
15:46
if(allowguest) gtk_box_pack_start(GTK_BOX(vbox), guestbox, FALSE, FALSE, 0);
15:46
thats just evil
15:46
<vagrantc>
ogra: i have no idea what any of that is about.
15:47
<ogra>
the box need to always be there
15:47
vagrantc, someone submitted that code
15:47
it was definately not me
15:47
<vagrantc>
ogra: most likely Ryan52
15:47jammcq changes topic to "LTSP-5 is the current version that you should be using. Check out Ubuntu, Debian, Fedora and Suse for awesome integration. Gentoo is getting very close. #ltsp IRC logs at: http://www.nubae.com/logs/"
15:48
<ogra>
if you see such code in review, ask the coder to keep the layout element persistent and only hide its contents (like we do in all other code thats similar)
15:48
jammcq, urgh
15:48
<vagrantc>
ogra: it's all gibberish to me, so i have no idea how to review it.
15:48
<jammcq>
ogra: what's the problem my friend?
15:48
<vagrantc>
ogra: i tested it, it didn't seem to break anything.
15:49
<ogra>
jammcq, so you learn all the omnious kiwi commands to help people out ?
15:49
<jammcq>
ogra: I thought they already handled that little issue
15:49
<ogra>
jammcq, no
15:49
<jammcq>
ok
15:50
<ogra>
you need to exactly know the kiwi command syntax and its config file to apply a change to their ltsp
15:50
what made you think they changed anything ?
15:50jammcq changes topic to "LTSP-5 is the current version that you should be using. Check out Ubuntu, Debian and Fedora for awesome integration. Gentoo is getting very close and Opensuse has kiwi-ltsp. IRC logs at: http://www.nubae.com/logs/"
15:50
<jammcq>
ogra: they changed the name of it
15:50
<ogra>
where ? o_O
15:51
<jammcq>
they dropped the 'kiwi' part of the name
15:51
it's something like opensuse-ltsp5
15:51
<ogra>
i didnt get any answer from zonker ever and there was definately no change in the code
15:51
<Gadi>
uh oh - irc logs - gotta watch what I say
15:51
although, this means my IRC humor can be retained for posterity
15:51
<jammcq>
Gadi: don't worry, we'll just strong-arm nubae to delete any incriminating comments
15:52
<ogra>
so you need to know the kiwi.cfg, the options and their names as well as what they do
15:52
<Gadi>
is it just me or are his calendars sideways?
15:52
<jammcq>
ogra: ok, changed it already, happy now?
15:52
<ogra>
and the kiwi commands to actually make them apply
15:52
jammcq, no, but thats nothing you can solve :)
15:52
i'll be happy if you can do:
15:52
!docs
15:52
<ltspbot>
ogra: "docs" is For the most current documentation, see http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream
15:53
<Gadi>
why would they make sideways calendars?
15:53
<ogra>
and what you get applies 100% to suse ltsp
15:53
<jammcq>
hmm, /me wonders if there's a limit to the size of the topic
15:53
<ogra>
there is
15:53* Gadi tries to think of a programmatic reason
15:53jammcq changes topic to "LTSP-5 is the current version that you should be using. Check out Ubuntu, Debian and Fedora for awesome integration. Gentoo is getting very close and Opensuse has kiwi-ltsp. IRC logs at: http://www.nubae.com/logs/, LTSP Docs: http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream"
15:53
<ogra>
but you only use a third of that yet or some such
15:54
<nubae>
:-)
15:54
<jammcq>
Febriary
15:54
hmm, makes me wonder where Nubae is from
15:54
<ogra>
thats japanese ;)
15:55
<jammcq>
neat-o
15:55* ogra is kidding indeed
15:55
<ogra>
yay
15:55
<jammcq>
ok, if we're done modifying the topic, can anybody tell me how to de-op myself?
15:55
<nubae>
Nubae is actually an old African Tribe... the Nubae
15:55* ogra has a greeter with free floating logo and input field
15:55* jammcq doesn't like walking around with his fangs exposed
15:55
<nubae>
but hey, I didnt know that until recently, I just like the way it sounds
15:56
<vagrantc>
Ryan52: would you be interested in the cleanup ogra wanted, and possibly changing the LDM_TIMEOUT_* to LDM_LOGIN_TIMEOUT ?
15:56
<ogra>
vagrantc, its fine as is now
15:56
i'll soon come up with a completely new greeter
15:57
<vagrantc>
oh
15:57
well, still would be nice to get the LDM_LOGIN_TIMEOUT changes ...
15:57
<ogra>
thats going fast enough here that i even might be ready before ubuntureleases
15:57
do what you like in the current upstream
15:57
i'm unlikely to pull anything from teher before release
15:58
<Gadi>
ok, kids - have a good weekend
15:58
<ogra>
and i will base the new greeters functionallity on upstream code so the sooner it changes the better
15:58
<Gadi>
this sideways calendar is telling me its time to go ;)
15:58
<ogra>
xrandr FTW :)
15:58
just rotate your screen
15:59
<Gadi>
hehe
15:59Gadi has left #ltsp
15:59Tshital has joined #ltsp
15:59* ogra ponders to put the "username" string above the input field by default in the new version
16:00
<Tshital>
Any one tell me how to pass video mode numbers to kernel?? i have ASUS tiny laptop (7" screen) which is conflicting with boot screen, when i give kernel option vga=786 i get some blurred cursor while booting, can any one tell me exact number?? i also get this http://pastebin.com/m1faf0768 while booting, if i give either of 0 to 6 it will boot fine
16:01six2one has quit IRC
16:03jarias has quit IRC
16:05
<Lns>
Tshital: I think you can probably just edit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default for kernel options
16:06Q-FUNK has quit IRC
16:06ltsplogbot has quit IRC
16:11makghosh has joined #ltsp
16:14ltsplogbot has joined #ltsp
16:29Tshital has quit IRC
16:35johnny has left #ltsp
16:38spectra has quit IRC
16:40K_O-Gnom has quit IRC
16:41
<nubae>
hmmmm... I need to rename a large set of files to their date created - 1 day.... any easy way to do that?
16:44
<zoredache>
it would probably be a pretty easy python/perl script to write. I don't know of anything that already exists to do something like that
16:45
<stgraber>
or as a bash script (for file in *, then get the date from the file, do the -1 and rename $file to that)
16:45
<nubae>
yeah I was hoping not to have to think... but guess its not super hard
16:47indradg__ has joined #ltsp
16:49
<Lns>
nubae: haha.. "Can't someone just make me a script?"
16:52
<nubae>
Lns: :-) I've almost got it...
16:53
its to convert the archived #ltsp logs
16:53
<Lns>
oh sweet
16:55
<nubae>
For file in *
16:55
do
16:55
FILE=$1
16:55
NOW=$(ls -l --time-style="+%b %e, %Y %k%M" $FILE | awk '{ print $6"-"$7"-"$8 }' | sed -e 's/,//g')
16:55
cp $FILE "$NOW.$FILE"
16:55
echo "File $FILE == $NOW.$FILE"
16:55
done
16:55
more or less right?
16:56
<Ryan52>
nubae: you should use stat instead of ls.
16:56
<nubae>
ah wait, it needs FILES=/path/to/files and then for file in $FILES
16:57
right?
16:57
<Ryan52>
nubae: FILE=$1 is wrong...I think...what are you trying to do there?
16:58
the name of the current file in the for loop is in the variable $file already.
16:58
<nubae>
yeah, should be different variable I guess... I'm trying to rename all files in a dir to their datestamp
16:59
<stgraber>
for file in * <--- that means $file will contain a different filename each time the do part is called
17:00
<nubae>
right
17:00
<stgraber>
so you can get rid of FILE=$1 and just use $file
17:00
<nubae>
yeah makes sense
17:01* nubae is still new at bashing
17:04indradg_ has quit IRC
17:14johnny has joined #ltsp
17:16
<stgraber>
for those of you who want to try ltsp-cluster, I have put my howto for a demo server online: https://wiki.stgraber.org/LTSP-Cluster/Documentation/TestSetup
17:16
the doc should be right as I used it to install a couple of server, it still lacks a lot of explanation on how thing works though
17:17
(and also documentation on how to add more application servers, add NX support, ... well lots of thing I have on my todo list)
17:17johnny has left #ltsp
17:18johnny has joined #ltsp
17:18
<stgraber>
packages are also taken from our PPA so although we do some QA on them, it's not perfectly stable and still contains quite a lot of bugs (from typing mistakes to more severe design mistake that we are working on)
17:19
oh and the documentation doesn't cover the centralized user accounts part as our demo servers are usually using autologin with user accounts generated right before they connect
17:20
<ltsppbot>
"vagrantc" pasted "cargo cult patch for LDM_LOGIN_TIMEOUT" (48 lines) at http://pastebot.ltsp.org/47
17:20
<vagrantc>
i'm not really sure i understood how it was working ... but maybe that should work
17:22
<stgraber>
vagrantc: isn't that exactly the same code but using an extra function (why ?) and a different env variable ?
17:22jammcq has quit IRC
17:22
<vagrantc>
stgraber: almost
17:22mccann has quit IRC
17:22
<vagrantc>
stgraber: converting two different variables with different types into a single variable
17:24
<stgraber>
hmm, I'm not really familiar with C (other than debugging) but shouldn't get_timeout_boolean() actually return timeout_enabled ?
17:24
<vagrantc>
actually, it should probably just return TRUE or FALSE
17:24
<stgraber>
yeah
17:24
<Ryan52>
vagrantc: no need for the if.
17:24
vagrantc: the getenv_bool returns true or false.
17:25
you can just change the function to a void and have it set the timeout_enabled var, since that var is a global.
17:25
<vagrantc>
Ryan52: oh, that's wrong then ...
17:25
Ryan52: it's not a boolean value
17:25
so can't use ldm_getenv_bool
17:26
<Ryan52>
oh, that's true. blah.
17:26
so just atio it. if the return value is 0 set it to false...
17:26
<vagrantc>
my intention was just to check if the environment variable was set
17:26
<nubae>
stgraber: dd
17:26
did u say cluster uses ldap for authentication?
17:26
cant recall
17:26
<vagrantc>
Ryan52: what's atio? how's it work?
17:27
<Ryan52>
vagrantc: timeout_enabled = (atio(getenv("LDM_LOGIN_TIMEOUT")) > 0 ? TRUE : FALSE)
17:27
atio converts a string to integer.
17:27
<stgraber>
nubae: depends on our customers. When they don't have an existing network, we use LDAP (openldap), otherwise we use what they have so either Novell or Microsoft
17:27
<Ryan52>
vagrantc: wait no. that's wrong.
17:27
timeout_enabled = (getenv("LDM_LOGIN_TIMEOUT") == NULL ? TRUE : FALSE)
17:27
getenv will return NULL if it's not set.
17:28
wait that should be a !=.
17:28* Ryan52 needs to think before pressing enter.
17:28
<stgraber>
nubae: basically ltsp-cluster is about load balancing and having a central place for the thin clients configuration. Then you can choose any authentication system on the application servers as you'd with a standard workstation
17:30
<nubae>
k, will have to check it out some time... not every day one needs clustering though... how big a setup do u use it for normally?
17:31mighty-d has quit IRC
17:31
<Ryan52>
vagrantc: try this: http://slexy.org/view/s21wFdBnih
17:32
<stgraber>
nubae: we use it for our own network (25 clients) but most of our customers have over 1k
17:36
<vagrantc>
Ryan52: isn't that just setting timeout_enabled in the update_timeout_boolean function ?
17:36
<Ryan52>
yes
17:37
it's a global variable.
17:37
<vagrantc>
ah.
17:37* vagrantc wondered why it gets set twice
17:37mikkel has quit IRC
17:37
<Ryan52>
follow the code. it makes sense.
17:37
<nubae>
ok, date script worked, but of course the 1 of the month changed to 0 of the month :p
17:38
<vagrantc>
Ryan52: i can only follow so far without understanding C :)
17:38
<alkisg1>
nubae, http://lists.clug.org.za/pipermail/clug-tech/2003-October/011245.html
17:39alkisg1 is now known as alkisg
17:40
<Ryan52>
it sets it back if you empty the entry box after typing stuff, I think.
17:40
or something like that.
17:40* Ryan52 only has your diff in front of him, not the actual code :P
17:40
<ltsppbot>
"Nubae" pasted "convert all files to date -1 day" (7 lines) at http://pastebot.ltsp.org/48
17:40
<nubae>
alkisg: I did that, and that worked, but the the -1 part affected the 1 of the month files and put them to 0
17:41
just a couple files though so I can do that part manually
17:41
<alkisg>
nubae, I think you should convert the date to seconds since epoch and subtract 86400, then back to date. I'm new to bash too though... :)
17:43
<nubae>
strained my head enough for now... but thanks...
17:44
<alkisg>
nubae, e.g. in your case I'd do find . -type f -printf "mv\t%f\t%Ad-%Am-%Ay\n" and then edit the output with openoffice calc to subtract 1 day! :D :D :D After that I'd just save and run the produced script!
17:45
<nubae>
the -1 day part was easy... just $7-1
17:46
<alkisg>
Yes, but oo wouldn't put 0 for days
17:46
<nubae>
ah right :-)
17:49
<ogra>
there we go
17:49
http://people.ubuntu.com/~ogra/ldm-new-greeter.png
17:49
both items are freely placeable
17:51
<alkisg>
ogra, very nice! The bounding box is part of the background or a gtk panel? (if it's called that way, I'm mostly a Delphi programmer)
17:52
<ogra>
the boudng box is tied to the input box and a semi transparent png
17:53
currently people replacing it need to use the same aspect ratio, but you could do a shadow that drops in or something else
17:54
thats the only constraint ...
17:54
<alkisg>
Why don't you leave the shadow/transparency to the background designers?
17:54
<ogra>
sure, they can do what they want
17:54
<alkisg>
Oh, background scales based on screen size...
17:54
<ogra>
indeed
17:54
it did that before
17:55
the wallpaer code is just copy pasted
17:55
the important part is that the elements wont influence each other anymore
17:55
and that they are freely placeable
17:56
<alkisg>
Very good, we'll be seeing ldm themes soon!!! ;)
17:56
<ogra>
you can put the logo in the top left corner and the input in the bottom right or whatever you desire
17:57
i just need to port all items from the old one first, then merge the greeter code and we have something fully themeable ...
17:57
<vagrantc>
ogra: backwards compatible with old themes?
17:57
<ogra>
beyond that it will be a lot easier to add new items
17:57
vagrantc, not completely
17:57
<Ryan52>
ogra: very nice!
17:57
<ogra>
i want to make the bounding box mandatory, so you would have to copy in the boundingbox.png
17:57* Ryan52 thinks of the mess of that fedora patch, and thinks of how much simpler it is now.
17:58
<vagrantc>
fedora patch?
17:58
<Ryan52>
vagrantc: fedora's stuff is completely rearranged.
17:58
<ogra>
vagrantc, i plan something similar (png themed bar) for teh bottom bar
17:58
so these two pixmaps need to go in
17:59
yeah, that will be easily possible in the new greeter
17:59
<Ryan52>
vagrantc: http://ryan52.info/~ryan52/K12linux-loginscreen-mockuprh.png
17:59
<ogra>
warren just needs to supply a file that holds the positions of stuff
17:59keithclark has quit IRC
18:00
<vagrantc>
heh
18:02
<ogra>
thats only a first shot though, i'll eed to apply a lot math still, currently i set pixel values to move the items off center, that needs to become percentages
18:03
so it scales properly with the screen size
18:06
sbalneav needs to supply me a sensitivity command from greeter_iface and i will add a fullscreen fade-to-grey option during password verification etc
18:06
<nubae>
gah... math.. :-)
18:12* ogra just realizes that it wil be possible to hide the input field and just show a guest button as well :)
18:12
<johnny>
that would be useful to me
18:12
<Ryan52>
why wouldn't you just skip the greeter in that case?
18:12
<stgraber>
ogra: rocks !!!
18:12
<johnny>
to allow people to shut down Ryan52
18:12
<Ryan52>
ah
18:13
<johnny>
then again.. the timeout will be good enough
18:13
<ogra>
right
18:13Lns has quit IRC
18:13
<ogra>
yeah, but there are environments where you might not want people to tinker with the input and just have them click to log in
18:13
internet cafes etc
18:14
LDM_USERNAME/PASSWORD predefined and a locked down session
18:14
<nubae>
what would be cool is a timer for internet cafes
18:14
something that automatically logs out after set time
18:15
<johnny>
that's soemthing my people want
18:15
<ogra>
i moved all UI parts out of the main() function and into subfunctions, so you can just en/disable parts based on variables
18:15
<johnny>
but.. you also need to allow a way to stop it.. i haven't figured out how
18:15
since if nobodoy is waiting, people are allowed to stay as long as they want
18:15
<ogra>
thats something you should probably do through Xsession.d
18:16
start a counter process with the session
18:16
<nubae>
kinda needs a backend on the server, otherwise its kinda useless
18:16
<ogra>
updating a file every mintue from it or so
18:17
<nubae>
although school I was working at wanted to limit usage for each user to 5 minutes during break time...
18:17
<ogra>
easy to do from such a process
18:18
<nubae>
yeah, so many ideas, so little time
18:18
:-)
18:18
<ogra>
that takes probably 2h to write in python
18:18
or even less
18:18
<nubae>
for super cow, yes... not for a calf like me
18:19
<ogra>
well, for you probably a day in shell :)
18:19
<alkisg>
vagrantc, for the code you pasted in pastebot, timeout_enabled = ldm_getenv_bool("LDM_TIMEOUT_ENABLED"); should become timeout_enabled = atoi(getenv("LDM_LOGIN_TIMEOUT")) > 0;
18:20
<ogra>
does taht cope with LDM_LOGIN_TIMEOUT="oeizcn" ?
18:20
<vagrantc>
alkisg: yeah, i figured we could add additional error checking later
18:20
<alkisg>
ogra, yes, it's considered 0
18:20
<ogra>
ok
18:20* ogra rarely uses atoi
18:20* alkisg too, but I tested this! :P
18:20
<ogra>
cool
18:22
<Ryan52>
alkisg: no.
18:22
alkisg: because getenv returns null if it's not set.
18:22
so just check if it's not NULL.
18:23
<alkisg>
Ryan52, right
18:23
<vagrantc>
so there should be another check that checks if the value is valid
18:25
Ryan52: does getenv treat an empty value as NULL ?
18:25
i.e. LDM_FOO=""
18:25
<Ryan52>
iirc, it treats it as an empty string...
18:25Barbosa has joined #ltsp
18:26
<alkisg>
So, in a oneliner, timeout_enabled = atoi(getenv("LDM_TIMEOUT_ENABLED") != NULL ? getenv("LDM_TIMEOUT_ENABLED") : "") > 0; but it should probably be stored in a variable to avoid calling getenv 2 times.
18:28
<vagrantc>
need to make sure it's an integer....
18:28* Ryan52 thinks an if should be used instead of : ?
18:28
<Ryan52>
vagrantc: atio makes it into an integer.
18:28
err, atoi
18:28
<alkisg>
vagrantc, even if it's a string it gets 0
18:28
<vagrantc>
ah.
18:29* vagrantc wonders if the code handles negative values ...
18:29
<Ryan52>
vagrantc: which code?
18:30
<alkisg>
vagrantc, from a quick look, the old code would need 2^64-1 seconds to login for timeout == -1... :P
18:31
with the new semantics, if timeout_enabled checks for negative values, it's enough
18:31
<Ryan52>
that's a good feature :). a quick way to set a really long timeout. :P
18:32* Ryan52 notices that we are back to the LDM_TIMEOUT_ENABLED variable name all of the sudden...
18:32
<alkisg>
My bad, copy/pasted...
18:33
I meant LDM_LOGIN_TIMEOUT
18:36
<nubae>
cool the logbot managed to integrated sbalneav's irc logs: http://www.nubae.com/logs/
18:53
<vagrantc>
ok, well, the patch without error checking works fine.
18:53
i even tried it with LDM_LOGIN_TIMEOUT=foo
18:53
which properly sets it to 0
18:54
though it displays "will automatically log in in 0 seconds"
18:55
<Ryan52>
vagrantc: so use: timeout_enabled = atoi(getenv("LDM_LOGIN_TIMEOUT") != NULL ? getenv("LDM_LOGIN_TIMEOUT") : "") > 0
18:55
if your still doing "timeout_enabled = the_function_just_added();", don't.
18:55
set timeout_enabled in the function.
18:56
<rjune_>
!seen jammcq
18:56
<ltspbot>
rjune_: jammcq was last seen in #ltsp 3 hours and 35 seconds ago: * jammcq doesn't like walking around with his fangs exposed
18:56staffencasa has quit IRC
18:57
<vagrantc>
Ryan52: in the update_timeout_boolean function?
18:58
<Ryan52>
yes, put that one liner in the update_timeout_boolean.
18:59* vagrantc doesn't understand the ? syntax
18:59Egyptian[Home] has quit IRC
18:59
<Ryan52>
conditions ? value_if_true : value_if_false
19:01
<vagrantc>
yeah, still doing timeount_enabled in the function
19:01
can just put a conditional in there, no?
19:01* Ryan52 is confused as to what vagrantc is asking
19:02* vagrantc will post diffs
19:04
<alkisg>
vagrantc, better do it right. Just declare a gint login_timeout at the top, it'll make the code a lot easier to read and no function will be required. Can you post your current greeter.c to pastebot?
19:04
<Ryan52>
alkisg: why would a function not be required?
19:04
it's better than having two copies of the code.
19:05
<alkisg>
So only a getenv will be executed
19:05
<Ryan52>
oh. I see what you mean.
19:05
<alkisg>
Ryan52, timeout_enabled = login_timeout > 0;
19:05
Much easier to understand
19:06
<Ryan52>
ya, and then just do this once: login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT") != NULL ? getenv("LDM_LOGIN_TIMEOUT") : "")
19:06
right?
19:06
<alkisg>
No, not at all!!!
19:06
Ah, ok, but with if :)
19:06
<Ryan52>
ya
19:07
really, then we don't even need a timeout_enabled.
19:07
we can just test "login_timeout > 0" everywhere where we are using it..
19:07
oh, no, nevermind. it's updated based on other conditions.
19:08
<alkisg>
No, login_timeout is not updated, (it's a new var I just proposed), timeout_left is the one being updated
19:08
<Ryan52>
ya, I get it, timeout_left and timeout_enabled get changed throughout the execution as they currently do.
19:09
<alkisg>
yeap
19:09
<ltsppbot>
"vagrantc" pasted "patch for LDM_LOGIN_TIMEOUT without much error handling" (48 lines) at http://pastebot.ltsp.org/49
19:09
<vagrantc>
so, i tested that, and it mostly works.
19:09
based on current ldm-trunk
19:10
<Ryan52>
vagrantc: so at the beginning of main (where we do other getenv and such) add this: login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT") != NULL ? getenv("LDM_LOGIN_TIMEOUT") : "")
19:10
<vagrantc>
start over, or just tweak the current patch?
19:11
<Ryan52>
umm...just start over.
19:11
then change the 'timeout_left = atoi(getenv("LDM_TIMEOUT_TIME"));
19:12
' to setting it from login_timeout.
19:12
and change the 'timeout_enabled = ldm_getenv_bool("LDM_TIMEOUT_ENABLED");'s to timeout_enabled = login_timeout > 0;
19:13
<alkisg>
vagrantc, also declare gint login_timeout; at the beginning, right before gint timeout_left;
19:14keithclark has joined #ltsp
19:14* Ryan52 thought he said that already...Ryan52 needs to multitask less.
19:14
<alkisg>
heh!
19:15
<vagrantc>
ok, so i think i've got everything but where to put ... 17:10 < Ryan52> vagrantc: so at the beginning of main (where we do other getenv and such) add this: login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT") != NULL ? getenv("LDM_LOGIN_TIMEOUT") : "")
19:16
<Ryan52>
vagrantc: look for the big set of ldm_getenv_bools...
19:16
<vagrantc>
i think that's in src/ldm.c rather than greeter.c
19:17
somewhere at the top of main, though ...
19:17
<Ryan52>
anywhere before use the login_timeout and before we set the timer to call the timeout function
19:18
<alkisg>
vagrantc, replace timeout_enabled = ldm_getenv_bool("LDM_TIMEOUT_ENABLED"); with what Ryan said
19:18
<vagrantc>
yeah, i got that.
19:18
<alkisg>
and delete one of the two continuous 'timeout_left = 0;' below, it's reduntant
19:18
<Ryan52>
alkisg: he did that earlier today.
19:19
<alkisg>
no, I meant with the initialization, that's where it should go
19:19
In main:
19:19
timeout_enabled = ldm_getenv_bool("LDM_TIMEOUT_ENABLED");
19:19
timeout_left = 0;
19:19
timeout_left = 0;
19:19
<vagrantc>
already commented those out
19:19
<alkisg>
The getenv should go there
19:19
<vagrantc>
i actually removed the second timeout_left = 0 earlier today :)
19:20* Ryan52 has noticed those 2 lines for a while, but never remembered to take them out.
19:20
<Ryan52>
bzr needs a "git add -i"
19:20
(lets you commit part of the file, but not all)
19:20
<vagrantc>
Ryan52: i use bzr shelve to handle that
19:21
<ltsppbot>
"vagrantc" pasted "use login_timeout" (41 lines) at http://pastebot.ltsp.org/50
19:21
<vagrantc>
how's that look?
19:22
oops ...
19:22
<Ryan52>
timeout_left = timeout_left;
19:22
oh, ya.
19:22
<vagrantc>
other than that?
19:22
<Ryan52>
+ login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT") != NULL ? getenv("LDM_LOGIN_TIMEOUT") : "")
19:22
missing semicolon
19:23
otherwise, looks right.
19:23Tshital has joined #ltsp
19:24
<vagrantc>
timeout_enabled = login_timeout > 0; ????
19:24* Ryan52 nods
19:24
<vagrantc>
set's it to NULL if not grater than 0 ?
19:24
<alkisg>
vagrantc, it sets it to false (= 0)
19:24
<Ryan52>
no, sets it to false.
19:25
<vagrantc>
close enough :)
19:25
<alkisg>
if you don't like the ? : syntax, you may also use this:
19:25
if (getenv("LDM_LOGIN_TIMEOUT") != NULL)
19:25
login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT");
19:25
else
19:25
login_timeout = 0;
19:25
<Ryan52>
alkisg: use curly braces
19:25
<alkisg>
*missing a )
19:25
<Ryan52>
if you don't you'll have to listen to ogra nag at you
19:25
<alkisg>
Ryan52, even with single statements? !!! wow... :)
19:25
<vagrantc>
i like the readability of using an actual if statement, yes.
19:26
<Ryan52>
alkisg: ya, I knwo.
19:26
err, know
19:29
<vagrantc>
if (getenv("LDM_LOGIN_TIMEOUT") != NULL) ... works properly if FALSE ?
19:30
<Ryan52>
ya...
19:32
<ltsppbot>
"vagrantc" pasted "timeout patch with if statement ..." (46 lines) at http://pastebot.ltsp.org/51
19:32
<alkisg>
vagrantc, a ) missing
19:33
login_timeout = atoi(getenv("LDM_LOGIN_TIMEOUT"));
19:33
<vagrantc>
fixed.
19:33
<Ryan52>
you still didn't fix teh: + timeout_left = login_timeout;
19:33
nevermind
19:33
<vagrantc>
heh
19:33* Ryan52 misread that
19:34* vagrantc compiles
19:34
<alkisg>
maybe it should be named ldm_login_timeout to resemble the environment variable... :P
19:34
(kidding)
19:35
<Ryan52>
hehe
19:37
<vagrantc>
i wouldn't be opposed to that ...
19:48
looks like it's working to me ...
19:48
tested with setting LDM_LOGIN_TIMEOUT to 0, 10, 5 and foo
19:48johnny has left #ltsp
19:49
<vagrantc>
of course, if they set it to True or something, it will silently behave as if 0 were set ...
19:49
but what can you do?
19:49ltsplogbot has left #ltsp
19:49ltsplogbot has joined #ltsp
19:51ltsplogbot has left #ltsp
19:52ltsplogbot has joined #ltsp
19:53johnny has joined #ltsp
19:53
<alkisg>
LDM_LOGIN_TIMEOUT=True? !!! Maybe then it should be named LDM_LOGIN_TIMEOUT_*SECS* :P :P :P (or they could just shoot themselves)
19:54johnny is now known as Guest46659
19:54
<alkisg>
g'night, everyone
19:54
<vagrantc>
alkisg: thanks for your help!
19:55
<alkisg>
vagrantc, me thanks for ltsp! :)
19:55twinprism has quit IRC
19:55alkisg has left #ltsp
19:57
<vagrantc>
well, can easily implement backwards compatibility in the screen script ...
19:59jammcq has joined #ltsp
20:00
<jammcq>
hey homies
20:00Guest46659 has left #ltsp
20:01
<vagrantc>
ah hell... i forgot to give alkisg and Ryan52 credit in my commit ...
20:01
sorry about that.
20:03vagrantc has quit IRC
20:09
<rjune_>
jammcq !
20:09
did you get a phone call?
20:09
<jammcq>
rjune_: nope
20:09
nothing yet
20:09
<rjune_>
you probably wont' then
20:09
I got the offer letter today
20:09
<jammcq>
hmm, oh?
20:10
that's a good thing, right?
20:11
<rjune_>
I think so
20:11
DOD clearance is next step
20:11
<jammcq>
oooh
20:11
hope that goes well for you
20:14
<rjune_>
me too
20:14twinprism has joined #ltsp
20:16
<rjune_>
don't think you'll get a call for that one though
20:16
which is unfortunate
20:17shogunx has quit IRC
20:21shogunx has joined #ltsp
20:25shogunx has quit IRC
20:26shogunx has joined #ltsp
20:31makghosh has quit IRC
20:51nubae has quit IRC
20:53nubae has joined #ltsp
20:55johnny has joined #ltsp
20:56johnny is now known as Guest80354
21:09Guest80354 is now known as johnny
21:10johnny is now known as Guest45287
21:11makghosh has joined #ltsp
21:26Guest45287 is now known as johnny
21:27johnny is now known as Guest4169
21:44Guest4169 is now known as johnny
21:44johnny is now known as Guest49120
21:46Guest49120 is now known as johnny
21:58zoredache_ has joined #ltsp
21:59rcy has joined #ltsp
21:59zoredache has quit IRC
22:10keithclark has quit IRC
22:38makghosh has quit IRC
22:40nubae1 has joined #ltsp
22:41nubae has quit IRC
23:02indradg__ has quit IRC
23:34cyberorg has joined #ltsp